W 4757 5‘ , 7" 7 DAYS ‘83!“ U. /00/ f /,\ ‘1_ 1,, 74* Numerical Simulation Analysis of the Interaction of I. Lakes and Ground Water GEOLOGICAL SURVEY PROFESSIONAL PAPER 1001 DEC 1 1973: U?‘ I-L..L5t.§fi 1 Numerical Simulation Analysis : of the Interaction of Lakes and Ground Water By THOMAS C. WINTER ‘ GEOLOGICAL SURVEY PROFESSIONAL PAPER 1001 UNITED STATES GOVERNMENT PRINTING OFFICE, WASHINGTON : 1976 4 J p )- 1 UNITED STATES DEPARTMENT OF THE INTERIOR 4 THOMAS S. KLEPPE, Secretary c A GEOLOGICAL SURVEY . V. E. McKelvey, Director F 6 z- 4(- I l p ‘1 A 1 4 ‘ 1 «1 Library of Congress Cataloging in Publication Data 4 Winter, Thomas C. Numerical simulation analysis of the interaction of lakes and ground water. #4 (Geological Survey Professional Paper 1001) Bibliography: p. A 1. Groundwater flow—Mathematical models. 2. Groundwater flow—Data processing. 3. Lakes—Mathematical models. 4. Lakes— Data processing. 5. Numerical analysis. ~ I. Title. II. Series: United States Geological Survey Professional Paper 1001. GBlOOl.72.M35W56 551.4'82 76—608309 A Q For sale by the Superintendent of Documents, US. Government Printing Office . v Washmgton, DC. 20402 Stock Number 024-001—02902—2 V w FIGURE CONTENTS Page Abstract .......................................................................................................................................... 1 Introduction ..................................................................................... 1 Background of the study, including previous investigations.. 1 Purpose and scope.... . ......... 3 Acknowledgment ............... 3 Ground-water flow systems .............................................. 3 Theoretical background. .............................................. 3 Models of hydrologic sections .......................................................... 3 Digital models of ground- -water flow near l.akes ... ................................................. 5 Practical considerations and assumptions of this study ........................................................ 5 Ground- water-flow diagrams ................................................................................................. 7 One-lake system ...................................................................................................................... 11 Multiple-lake system ............................................................................................................... 29 Moderate regional slope and local relief ........................................................................ 29 Low regional slope and local relief ................................................................................ 31 Quantitative aspects of ground-water interchange with lakes ............ 42 Areal and temporal variations ............................................................ 42 Implications for field studies of the interaction of lakes and ground water 43 Summary .................................................................... 44 References cited .............................................................................................................................. 44 ILLUSTRATIONS 1. Sketch showing physical meaning of mathematical expressions for boundaries of hydrologic-section models ........................... 2. Hydrologic section showing local, intermediate, and regional ground-water-flow systems determined from an analytical solution to the ground-water-flow equation ............................................................................................................................ 3. Hydrologic section showing ground- water-flow systems determined from a numerical solution to the ground- -water-flow equation ..................................................................................................................................................................................... 4. Hydrologic section showing a quasi- quantitative flow net of ground- -water flow near lakes in a multiple- -lake system that does not contain aquifers .......................................................................................................................................................... 5. Hydrologic section showing a quasi-quantitative flow net of ground-water flow near lakes in a multiple-lake system that contains aquifers ........................................................................................................................................................................ 6. Hydrologic section showing distribution of hydraulic head in the ground-water reservoir of a multiple-lake system that does not contain aquifers ........................................................................................................................... 7. Hydrologic section showing the dimensions of features included in the diagrams of the one-lake system .................................. 8. Hydrologic section showing distribution of hydraulic head in the ground-water reservoir of a one-lake system that does not contain aquifers or lake sediments ..................................................................................................................................... 9. Hydrologic section showing distribution of hydraulic head in the ground-water reservoir of a one-lake system that contains lake sediments but does not contain aquifers ........................................................................................................................... 10. Hydrologic section showing distribution of hydraulic head in the ground-water reservoir of a one-lake system that has a low water—table mound on the downSIOpe side of the lake and does not contain aquifers .................................................... 11. Hydrologic section showing distribution of hydraulic head in the ground-water reservoir of a one-lake system that has low water—table mounds on both sides of the lake and does not contain aquifers ................................................................. 12. Hydrologic section showing distribution of hydraulic head in the ground-water reservoir of a one-lake system that has an extensive aquifer of relatively low hydraulic conductivity at the base of the system ......................................................... 13. Hydrologic section showing distribution of hydraulic head in the ground-water reservoir of a one—lake system that has an extensive aquifer of relatively high hydraulic conductivity at the base of the system .................................................... 14. Hydrologic section showing distribution of hydraulic head in the ground-water reservoir of a one-lake system that has an extensive aquifer of relatively low hydraulic conductivity in the middle of the system .................................................... 15. Hydrologic section showing distribution of hydraulic head in the ground-water reservoir of a one-lake system that has an aquifer of limited extent upslope from the lake at the base of the system ......................................................................... l6. Hydrologic section showing distribution of hydraulic head in the ground-water reservoir of a one-lake system that has an aquifer of limited extent beneath the lake at the base of the system .................................................................................. III Page 4 10 12 13 14 16 17 18 19 20 21 22 23 IV CONTENTS Page FIGURE 17. Hydrologic section showing distribution of hydraulic head in the ground-water reservoir of a one-lake system that has an aquifer of limited extent downslope from the lake at the base of the system ......................................................................... 24 18. Hydrologic section showing distribution of hydraulic head in the ground-water reservoir of a one-lake system that has a deep lake and an extensive aquifer of relatively low hydraulic conductivity at the base of the system .............................. 25 19. Sketch showing key to table 1 .......................................................................................................................................................... 27 20. Hydrologic section showing distribution of hydraulic head in the ground-water reservoir of a multiple-lake system that has an extensive aquifer of relatively high hydraulic conductivity at the base of the system ................................................ 33 21. Hydrologic section showing distribution of hydraulic head in the ground-water reservoir of a multiple-lake system that has water-table mounds of various height and aquifers of limited extent in the middle of the system ................................. 34 22. Hydrologic section showing distribution of hydraulic head in the ground-water reservoir of a multiple—lake system that has low water-table mounds and aquifers of limited extent at the base of the system ........................................................... 35 23. Hydrologic section showing distribution of hydraulic head in the ground-water reservoir of a multiple-lake system that has low water-table mounds and aquifers of limited extent in the middle of the system ...................................................... 36 24. Hydrologic section showing distribution of hydraulic head in the ground-water reservoir of a multiple-lake system that has high water-table mounds and aquifers of limited extent at the base of the system .......................................................... 37 25. Hydrologic section showing distribution of hydraulic head in the ground-water reservoir of a multiple-lake system that has deep lakes, high water-table mounds, aquifers of limited extent at the base of the system, and Kh/Kv=l,000 ............... 38 26. Hydrologic section showing distribution of hydraulic head in the ground-water reservoir of a multiple-lake system that has deep lakes, high water-table mounds, aquifers of limited extent at the base of the system, and Kh/Ku=100 .................. 39 27. Hydrologic section showing distribution of hydraulic head in the ground-water reservoir of a multiple-lake system that has low regional slope and does not contain aquifers ............................................................................................................. 40 28. Hydrologic section showing distribution of hydraulic head in the ground-water reservoir of a multiple-lake system that low regional slope and aquifers of limited extent in the middle of the system ...................................................................... 41 TABLES Page TABLE 1. Results of digital simulations of the one-lake system—comparison of the effects of all combinations of the parameters that control ground-water flow on lake-ground water interaction ......................................................................................... 25 2. Summary of simulations of one-lake system .......................................................................................................... 28 3. Summary of simulations of multiple-lake system whose water table has moderate regional slope and local relief ............ .. 32 4. Summary of simulations of multiple-lake system whose water table has low regional slope and local relief ............................. 32 NUMERICAL SIMULATION ANALYSIS OF THE INTERACTION OF LAKES AND GROUND WATER By THOMAS C. WINTER ABSTRACT Because the interrelationship of lakes and ground water is perhaps the least understood aspect of lake hydrology, vertical-section, steady- state, numerical-model simulations were run to evaluate the factors that control the interaction of lakes and ground water. The study is concerned only with lakes encircled by water-table mounds that are at a higher altitude than lake level. Simulations of one-lake and multiple-lake systems in vertical sections show that for many hydro- geologic settings, the line (divide) separating local from regional ground-water flow systems is continuous beneath individual lakes. If the divide is continuous, there exists a point along it at which the head is a minimum compared to all other points along the divide. This point of minimum head is always greater than the head repre- sented by lake level, therefore in such a setting there can be no move- ment of lake water through the lake bed to the ground-water system. In a setting where the divide is not continuous, the lake loses water through part of its bed, but rarely in the littoral zone of the lake. Factors that strongly influence the position, shape, and continuity of the flow-system divide beneath lakes are height of the water table on the downslope side of the lake relative to lake level, position and hydraulic conductivity of aquifers within the ground-water reservoir, ratio of horizontal to vertical hydraulic conductivity of the ground- water system, and lake depth. INTRODUCTION BACKGROUND OF THE STUDY, INCLUDING PREVIOUS INVESTIGATIONS Lakes are a valuable natural resource in many areas of the United States. In addition to being sources of water supply for many communities, they are the focus of recreational activity and their aesthetic qualities are highly valued. That they are a desirable feature is evidenced by the large number of small reservoirs being constructed throughout much of the United States for a wide variety of purposes. Increasingly, lakes are being used as the focal point in planning urban developments. Unfortunately, the popularity of lakes often leads to their deterioration. The increased input of nutrients through the activities of man causes the organisms in lakes to flourish at rates far in excess of natural conditions. This increased productivity in turn causes chemical changes in lakes that result in obnoxious odors, fish kills, and unsightly conditions in the lakes and along their shorelines. The number of problem lakes is sharply increasing. In order to evaluate lake problems and the effect of manage- ment projects, nutrient budgets and water budgets of lakes are needed. Only by thorough understanding of the inter- relationships of lake hydrology, chemistry, and biology can progress be made in alleviating present lake problems and preventing more problems from arising. In recent years, much money and research effort have been expended to understand the principles of lake chemistry, biology, and the nutrient balance of lakes. At the same time, comparatively little money and effort have been devoted to understanding the principles of lake hydrology or the water balance of lakes. With few excep- tions, the relationship of ground water to lakes has been a minor part of the hydrologic studies, and it remains the least studied and least understood aspect of lake hydrology. In most studies of lake hydrology, atmospheric water (precipitation and evaporation) and surface-water inter- change with lakes is measured. Ground water is generally calculated as the residual. This practice can lead to serious misunderstandings about the interaction of lakes and ground water, especially when the error of measurement inherent in precipitation and particularly evaporation is considered. Further, only through careful field tech— niques can streamflow measurement error be kept to relatively low values. Some misconception of the interaction of lakes and ground water has resulted from lack of understanding of ground-water flow, which leads to inadequate instru- mentation and data collection. Poorly executed studies have led to various erroneous ideas concerning the inter- action of lakes and ground water, to the detriment of sound lake management. Some scientists and managers believe all lakes are discharge points of the ground-water system and therefore do not lose water through their beds. Others believe that lakes are points of recharge to the ground-water system. Some believe ground water flows in l 2 INTERACTION OF LAKES AND GROUND WATER one side of lakes and out the other, a flow-through condi- tion. Others believe all three situations exist. Most studies of the interaction of lakes and ground water have been in response to a need for water budget information for a particular lake or groups of lakes. Some studies have not used observation-well instrumentation to define the ground-water flow systems, but instead assumed certain hydrogeologic conditions. Generally those that used wells either used too few wells, that is they assumed a simple relationship between lakes and ground water that could be determined by one or two wells near the shore- line, or they installed too many wells within the immediate drainage basin of the lake. Some studies assumed that definition of the water table near a lake was sufficient to understand the interactions. The relation of ground water to prairie potholes in North Dakota was studied by Eisenlohr and others (1972). Although in the early stages of this study the ground-water component was calculated as a residual, a seepage meas- uring device was subsequently developed. In latter stages of the study, wells were placed near some potholes to deter- mine the relationship of ground-water levels to pothole water levels (Sloan, 1972). The ground-water relation to small lakes in Minnesota was studied by Manson and others (1968) and Allred and others (1971) by placing one or several wells near the lake shores. The study approach was similar to that of Eisenlohr and others (1972). The general conclusion of these studies was that most of the lakes studied had a net loss to ground water. In a study of Lake Sallie, Minnesota(McBride, 1969), the ground-water component of the lake-water budget was a primary interest. Wells were placed within the entire drainage basin of Lake Sallie to define lateral ground- water movement. At some sites several closely spaced wells were completed at different depths to define the vertical component of ground-water flow. Flow nets were then constructed to calculate the quantities of ground water moving into and, in one small area during part of the year, out of Lake Sallie. McBride used a digital modeling tech- nique to define the vertical distribution of hydraulic head in the ground-water system near a lake. His models are of the flow within the immediate drainage basin of the lake itself—from the local ground-water divide to the mid- point of the lake (McBride and Pfannkuch, 1975). Studies of the interaction between lakes and ground water similar to those conducted by McBride (1969), although not using simulation modeling, were done in Wisconsin by Hackbarth (1968), Hennings (1974), and Possin (1973). The lakes in these studies are the flow- through type—ground water enters one side of the lakes, and the lakes leak to the ground-water system on the other side. Meyboom (1966, 1967) studied ground—water flow systems in the vicinities of lakes and potholes in the prairie provinces of Canada. Meyboom’s work is some of the first to examine the problem ofdetermining ground-water flow systems and the strength and position of ground- water divides or absence of such divides beneath lakes. Studies by Freeze ( 1969a, 1969b) consider the relationship of ground-water flow systems within large drainage basins to lakes in the prairie provinces of Canada. The scale of the hydrologic sections, however, is such that the detailed flow patterns by the lakes can not be defined. Williams (1968) examined ground-water flow systems in vertical sections near small depressions, and the relationship of those flow systems to wetlands in northern Illinois. The techniques of investigation and results of the study were similar to the work of Meyboom. Studies similar to and using the techniques of Meyboom were also done in East Germany by Schumann (1973). A general overview of studies of the interaction of lakes and ground water, including a literature review, is given by Born and others (1974). They also discuss and present conceptual diagrams of many variations of ground-water flow systems near lakes. The relationship of ground water to large lakes and reservoirs has been studied by several Russian hydro— geologists. Zektzer (1973) discussed the role of ground- water flow in water and salt balances of Lake Baikal and the Caspian Sea. Zektzer and Kudelin ( 1966) discussed the methods of determining ground-water flow to lakes with special reference to Lake Ladoga, USSR. Payne (1970) used radioactive tracer techniques to study the ground- water aspect of the water balance of Lake Challa, Africa. Haefeli (1972) has calculated the ground water inflow into Lake Ontario from the Canadian side. The study made use of the cross-sectional digital ground-water flow model of Freeze (1969b). Van Everdingen (1972), as part of an intens- ive study of the Lake Diefenbaker, Saskatchewan, site, con- structed a series of piezometer nests (closely-spaced group of wells, each well completed at a different depth) before the dam was constructed. The changes in the potentio- metric level of different aquifers in the ground-water system were observed as the lake filled. The study showed reversals of flow within some of the aquifer zones as a result of the creation of Lake Diefenbaker. A similar study is under way at the Kendrid Lake site, North Dakota, by Downey and Paulson (1974). Careful review of the above literature leads one to the conclusion that much work needs to be done to identify and evaluate the factors that control lake hydrology, especially the factors that control the relationship of ground water to lakes. The general principles underlying the interaction of lakes and ground water need to be defined on theoretical grounds to solve even such basic problems as determining the optimum number and place- ment of observation wells needed to define ground-water flow systems near lakes. A logical first step in defining the general principles GROUND-WATER FLOW SYSTEMS 3 controlling the interaction of lakes and ground water is to use simulation modeling to examine the detailed patterns of ground-water flow around and beneath lakes for a wide variety of hydrogeologic settings. PURPOSE AND SCOPE The purpose of this report is to examine ground-water flow systems near lakes through digital models simulating ground-water flow in vertical sections. Use of such models makes it possible to examine the general principles of the interaction of lakes and ground water. By varying shape and altitude of the water table in relation to lake levels, lake depth, sediment distribution, size and position of aquifers within the ground-water system, ratio of horizontal to vertical hydraulic conductivity, ratio of hydraulic conductivity of aquifers to that of the sur- rounding finer-grained materials, and thickness of the ground-water reservoir, it is possible to locate the posi- tion and strength (explained later) of the divide that exists under a lake or the percentage of the lake bottom through which water moves to the ground-water system if there is no continuous divide. Although the cross sections are hypothetical, the values of the parameters listed above are realistic and representative of physiographic and hydro- geologic conditions in glacial terrane, the geologic environment in which most natural lakes occur. Lakes in this type of terrane are of primary interest in this study, but the models used to define the interaction of lakes and ground water and the general principles evolved from this study should have general application to all types of geologic terrane. ACKNOWLEDGMENT This study constituted part of a dissertation (Winter, 1976a), which was submitted to the Faculty of the Graduate School at the University of Minnesota in partial fulfillment of the requirements for the degree of Doctor of Philosophy. Dr. H. O. Pfannkuch provided overall guidance during the course of the study. With minor modifications the dissertation was released to the open-file (Winter, 1976b). GROUND-WATER FLOW SYSTEMS THEORETICAL BACKGROUND This study is concerned only with two-dimensional, steady-state, non-homogeneous and (or) anisotropic, cross-sectional systems. Thus the basic equation of ground-water flow that is used in this report for defining the head distribution within the ground-water system is 6—6x [K(x,z)%]+6—‘:[I<(x.z)Z—Z]=0, (1) where K(x,z)=hydraulic conductivity in the two coordi- nate directions and M, E: gradient of hydraulic head. 6x dz Equation 1 results from coupling Darcy’s equation with the equation of continuity. It assumes that fluid density is constant and the coordinate axes are aligned collinear with the principal directions of anisotropy. Development of the equation is discussed in most textbooks of ground- water hydrology (e.g., Domenico, 1972; DeWiest, 1965). In order to solve equation 1, it is necessary to define mathematically the boundary conditions for the region of interest. A general model for the type of flow section of interest in this report is shown in figure 1. This figure represents an x-z coordinate system that has the origin in the lower left corner. For any point (P) in the system there corresponds a value of hydraulic head. Because there is no flow across the vertical boundaries projected down from the major divide and the major sink, the head gradient is zero 22:0) along the two sides of the diagram. It is also assumed that the base of the system is impermeable, thus 3%: 0. The pressure along the water table is atmospheric and the head (h) there is a function of x only. The hydraulic conductivity (K) values are different for each geologic unit within the system. For this study, the size and position of zones of high hydraulic conductivity (termed aquifers for convenience) are varied, and the degree of anisotropy is varied. Recent development of numerical methods as approx- imate solutions to the equation of ground-water flow (Remson and others, 1971; Trescott and Pinder, 1975; Cooley, 1974; Prickett and Lonnquist, 1971 ; Freeze, 1969b) has made it possible to simulate complex ground—water flow systems. Although a variety of numerical techniques are being developed, the finite-difference method has been well documented and has been used successfully in a number of studies. The finite-difference models used herein have been developed by the US. Geological Survey over a number of years (Finder and Bredehoeft, 1968; Pinder, 1970; Trescott, 1973; Trescott and Pinder, 1975). The alternating-direction-implicit method (ADI) was used to calculate the distribution of hydraulic potential within some of the cross-sectional simulations but, as more complex models were considered, it was necessary to turn to the strongly implicit procedure (SIP) (Stone, 1968). As an example of SIP’s greater efficiency, some checkruns were made on the same section, using both numerical tech— niques; ADI did not converge after 99 iterations, whereas SIP needed as few as 16. It was found that SIP adequately handled all conceivable simulations except one, where extremely high hydraulic conductivity contrasts existed in close proximity, both in the vertical and horizontal directions. MODELS OF HYDROLOGIC SECTIONS Only two comprehensive theoretical studies appear to have been published on modeling ground-water flow in \ INTERACTION OF LAKES AND GROUND WATER 0 II c'x 6669 \ \ £3on conuummeanPE mo wuimwcson g8 333898 .mucmfiefimE «o mEEmuE F0553 9:305 5835]; 559“— W - oué \/ \ \ MAM—(P mmk<>> 8.9 PRACTICAL CONSIDERATIONS AND ASSUMPTIONS 5 vertical section (Téth, 1963; Freeze, 1969b). Subsequent work based on the report by Freeze has been done by Haefeli (1972) and by Freeze (1969a). These studies have concentrated on ground-water-flow patterns in vertical section in large basins and, although lakes occur along some of the sections modeled, the scale of the sections were such that detailed flow patterns in the immediate vicinity of lakes could not be shown. Téth’s work is based on an analytical solution to the ground—water—flow equation, whereas the work of Freeze is based on a numerical solution. Freeze (1969b) discusses the advantages and disadvantages of each approach. In the analytical-solution approach Téth had to make some rigid simplifying assumptions: The field has to be approx- imated by a rectangle; the ratio of horizontal to vertical hydraulic conductivity must be 1, that is, the porous media must be isotropic; the water table is simulated by super- imposing a sine curve on a low regional slope. The analytical approach requires that equations be set up for each individual case considered according to the boundary configurations. There is no general analytical solution to the ground-water-flow equation that is valid for all boundary conditions at the same time. The advantages of the numerical solution as outlined by Freeze (1969b), on the other hand, are as follows: Many of the restrictions of the analytical solution are removed; the true shape of the water table can be approximated, that is, it is not restricted to the equation for a sine curve; equa- tion 1 can be handled in the numerical solution, that is, anisotropic and heterogeneous conditions can be simulated. The numerical solution available in the USGS program is general; it was designed to handle a wide variety of hydrogeologic problems, including the vertical section simulations used in this study. Both basic studies by Toth and Freeze assume no-flow boundaries at the base of the system and no-flow boundaries beneath the regional topographic highs and lows. In addition, the water table is at steady state. An example of one of Téth’s cross-sectional ground- water flow simulations is shown in figure 2. The simula- tion is of a thick ground-water system that has low regional slope and low local relief. Although Téth presents many such diagrams showing the effects on ground-water flow of many combinations of slope, relief and system thickness, this diagram best depicts his delineation and definition of local, intermediate, and regional flow systems. . An example of the flexibility of numerical models in closely simulating field conditions is shown in figure 3 (Freeze, 1969b). The diagram is of a hydrogeologic setting that has a variable water table in different parts of the section, anisotropic media, and zones of different hydraulic conductivities. DIGITAL MODELS OF GROUND-WATER FLOW NEAR LAKES PRACTICAL CONSIDERATIONS AND ASSUMPTIONS OF THIS STUDY The boundary conditions assumed for this study relate closely to those of Freeze (1969b), that is: (1) The base of the system is considered to be impermeable; (2) the vertical no- flow boundaries at each end of the section are considered to be controlled by the major topographic high and the major drain on the ground-water system; and (3) the upper boundary of the flow system is the water table and it is at steady state. In the model used for this study, the steady- state water table is specified by assigning effectively infinite storage to the water-table nodes. Freeze discusses thoroughly the justification for these boundary condi- tions and particularly the assumption that vertical no- flow boundaries exist at the two end positions. As he points out, many workers have found that vertical no-flow boundaries do not occur beneath every topographic high. This is undoubtedly true, but the assumption that they do exist beneath the major high and low is considered to be valid. Although the sections discussed in this report are hypo- thetical, they closely approximate field conditions for lakes in glacial terrane. The variations in height of the water table relative to lake level are reasonable for this type of terrane. Although the hydraulic conductivities that have been assigned to the geologic matrix in general are representative of silty till, in the simulations of this study, the relative values of hydraulic conductivity are the con- trolling factor, not the actual values. The simulated sections apply equally to geologic settings that have the given relative hydraulic conductivities, whether the setting is sand within till, or gravel within sand. The hydraulic conductivities assigned to the zones of higher relative hydraulic conductivity vary from 100 to 1,000 times greater than the surrounding geologic materials. There are, of course, some zones within glacial drift that have hydraulic conductivity values greater than 1,000 times that of the surrounding materials, but the values chosen are representative of a great many geologic settings. For convenience in this report, the geologic matrix is referred to as “till,” and the zones of relatively higher hydraulic conductivity are referred to as “aquifers.” The hydraulic conductivities assigned to lake sedi- ments are as low as can be assigned with the computer program used. In most lakes, the littoral zone is free of fine grained sediment; therefore, in the models that considered lake sediments, the sediments purposely were not extended to the shore line. Lake water was simulated by assigning very high INTERACTION OF LAKES AND GROUND WATER 4 Ir 4 , v ‘ 1‘ A ‘ I ‘. J .. . 1 , . F I / I. ‘ W ‘ mad mwd mud mod wmd wed ,_-—- - 1 _ _ _ _ L _ _ <32 .5? 82% mfio mm .o .83ng KOELBaBfichm 2: o. act—:9. Eutfimcm as Eod‘vofigaov 2:39.? 30: B~$fic=0hw~mmovwvm «Em 35¢?qu 480— ME??? :omuuum owwo—oéfialfi 550...— ucfi ooodw 5.3%; \sot >33me A 58%»... \so moo Emgmsmgot R604 Vso: U6 c2825 EEK mac.” t \ q 133} OOO'OI. GROUND-WATER-FLOW DIAGRAMS 7 hydraulic conductivity values to the nodes within lakes. This proved to adequately simulate lakes because the cal- culated head values were identical to the initial head values within each lake simulated in this study. The least well known parameter needed in the ground- water-flow model is the ratio of horizontal to vertical hydraulic conductivity (Kh /Kv ). This ratio has generally not been determined in glacial terrane and it is a point around which much of the discussion of the following sec- tion of this report centers. It will be shown that it is critical to the interaction of lakes and ground water. If the ratio is less than 100, lakes rarely lose water to the ground-water system under the conditions simulated in the models and if the ratio is greater than 1,000, lakes lose water under many conditions. If the ratio is between 100 and 1,000, other factors that control ground-water flow become more important in the relationship of ground water to lakes. Weeks (1969) measured Kh /KU within a single outwash sand and gravel aquifer in Wisconsin and found the ratio to be not more than 20. Vecchioli and others (1974) cal- culated K h/ K, for part of the drift section of Long Island, New York as approximately 500. Bennett and Giusti (1971), in using electric-analog techniques to study ground-water flow in the coastal plain of Puerto Rico, show that the ratio had to be 1,000 for the simulations to match field data. The importance of this parameter in ground-water-flow modeling has been recognized because it is the topic of recent papers by Freeze (1972) and Gillham and Farvolden (1974). Freeze concludes that Kh /Kv values of 100 or larger are not uncommon, and in fact were necessary to correlate simulations with field measure— ments in his study of the Old Wives Lake basin in Saskatchewan (Freeze, 1969a). Gillham and Farvolden (1974) tested the sensitivity of the ratio and found it to be particularly important in areas of recharge and discharge. Considering the above studies, the models discussed in the following section of this report use both 100 and 1,000 as lower and upper examples. GROUND-WATER-FLOW DIAGRAMS Much of the remainder of this report is a discussion of ground-water flow systems near lakes, which are best illustrated by diagrams showing ground-water flow in cross section. Understanding of the important features of this type of diagram is basic to the discussion of the simulation results. A ground-water-flow diagram shows the distribution of hydraulic head within the ground- water system. After the head is calculated for each node, lines of equal head, equipotential lines, are drawn. Ground-water flow lines are drawn perpendicular to the equipotential lines, if the medium is isotropic. Equipo- tential lines are the projections of water-table contours into the subsurface (figure 4). The finite difference grid used in this study is a network of uniform rectangles. About 900 nodes were used for the one-lake simulations and about 1,800 nodes for the three- lake simulations. To give the most accurate picture of ground-water flow systems, and to estimate relative quantities of ground water moving through various parts of the ground-water reservoir, a flow net should be drawn. Construction of a quantitative flow net in a medium that is isotropic (Kh/ Kv=1), requires that flow lines and equipotential lines he orthogonal such that curvilinear squares are formed (Harr, 1962). If the medium is anisotropic (Kh /Kv¢l) the squares are deformed according to the degree of anisotropy. In this study, Kh /K,, is either 100 or 1,000. To compensate for the anisotropy, in homogeneous media equation 1 can be transformed to the LaPlace equation by a transformation of coordinates. The scale factor used to do this, which was not applied to the diagrams in this report (discussed later), is the square root of the ratio Kh /Kv (Harr, 1962, p. 29). A requirement for constructing precise flow nets is that the section have no vertical exaggeration (Van Everdingen, 1963). Vertical exaggeration is usually used for convenience in illustrating important details of geology and ground-water movement that would be lost if the true scales were used. The sections discussed in this study have a vertical exaggeration of 80:1. Because the graphical correction of Van Everdingen (1963) was not applied to the sections in this study, they cannot be used as exact quantitative diagrams. The diagrams of Van Everdingen (1963) provide good examples of the effect of the correction on the ground-water-flow fields. The effect basically is that the corrected flow nets are much more rectangular and the local systems extend deeper into the ground-water reservoir than in flow sections not corrected for vertical exaggeration. The flow nets (figures 4 and 5) were drawn as if the porous medium were isotropic and homogeneous merely to show the general concept of flow nets and the different-magnitude flow systems that can occur near lakes and within the ground-water reservoir. Although the two flow nets drawn for this study are not quantitatively precise, they do show relative proportions of flow in different parts of the ground-water system. The ground-water system illustrated in figure 4 consists of several flow systems of different magnitude. The upper part of the ground-water system consists largely of local flow systems where water moves from high points on the water table (water-table mounds) to adjacent lowlands occupied by lakes. Regional ground-water flow occurs deep in the system. Recharge to the regional flow system occurs at the major drainage divide and discharge from this system is to the major drain (stream). A zone of inter- mediate flow (darkly stippled area) is recharged at the water-table mound between lakes 2 and 3 and is dis- charged into lake 1. It should be noted, but will be dis- cussed in more detail later, that much more ground water m0ves through local flow systems than through the deep INTERACTION OF LAKES AND GROUND WATER .muorswm 58:8 .0: 80¢ :2: ESP? 8.263235 m E 3&2 E»: 30: Bumaficzew we 8: 30a gas—Emzcémmzc a 9:305 :ocuwm BuBofiiElé muse—h 8 8 88 00 8 Co 8 8 7u7u7u 7v 7v 7v .6 7y 37:6 7.. .6 3.633333 7g 00 .6 .68 3 Ir Ir 0 0 666 6 8 8 I. I. 999 9 9 97.78838 Ir 0 C._ 8.6 0 9 0 C. O 978 O C._ 0 9 0 9.78 0 0.. 0909090 9 cocoa o a o \oom o8; "€19 mustrm mxm._\—. 3 F d: E “55298 LmnEsc 8.3 % Sc 83th .650 Emum>m 26: v8 Emucaom Eoum>m \socLBmB 938:8 \EtmE / -9505 333,585 *0 26M éqna Em. 92$ bmcmwQ W W 8 .mEmtms & \mEmtS u °_ Egg” 0 hum”. ooom o o o -oa 0:32P»; Exam Po 9.3 ‘ omw .uofl :_ ._m>w_ 3.2 269.... 38; 2 52:52 EEmEU £5 9 vozaam on 2:8 we: 26E .E_oa cozmcmmum 96 32:03 .9550 was“ $.QO 23:3: *0 m_amem om E 5/0 BE mmes. zo_F2|6 .5505 8 9 9 8 8 8 33333 3 3 3 333 3 3 3 3 3 3 3 3 7y 7v Ir Ir 0 0 66668 8 Z I. 999 9 9 .V .V 00 8 7a a C. 0 C. 0 C._ 0 QZlOQ 0 C._ 0 9L0 9 0 C._ O 9 0 9 0 u _ n \ 3 / m L m . ............ .. 1/ / c._ c 0 9 .3 , _ .. , 83 n “Vi my .owm.Z3\a . ooogu\_¥\§ \ an“... k .. 25:2“ 3_3\\F mwwz P a: s 25293 . 0mm \ \ 83:5: 3.3 0% 2m 2:8qu 550 . N W W «V Eofi>m Bo: +0 faucsom u. Ir 9 0 ...................... .n on. C._ \ / 3 3 8 E3828 \StmE 8 8 9 6 0 8 8 3 0 $30.5 Em amt: nutme 9 0 GENES. M3 REES 3E: 82 o 53% 28%: $14..qu m m>onm 38 E ._m_u:8 hwwn. ooom o o o .oa 0:396? .33 *0 mcfi 0mm .33 E ._m>w_ 9.2 m>onm cam: m_ 33:52 £235 25 9 923% on 2:00 9.: EOE .qua cozmcmmum wo >H_:_u_> 3550 $5 $.QO 23:2: *0 23:86 om my ' . 310 Eu”. mmmEs. zo_k:l.o 550.”. Do 8800 8 9 76.6 .6 .6 .6 .66 6 _6 00 6—6.6 Ir 0 66 6 8 I. 99 9 C._ 0 8.60 O 0 .78 0 0 0 .700 0 O INTERACTION OF LAKES AND GROUND WATER \ \ .6 .6 / IQ , \mtztzm oxm._\.‘ 9 9 Nu .1 EnEac 3.3 8 .7 0 8 8v 0 BEE: 88 0 BE 88 o o o .3. E .302 3E 26% gm; 2 89:32 E836 25 o. .5:an 8 2:8 .qua cow—wcmmum ”.0 3:53 6350 was. 8.x“; 23:8. .0 EQENXM om m— .HWU. BE 9352 12 6 .6 6 .V 8 .6 0 O O "| \‘I \ 8 O\ L . O \ \ o8; "$1 .9 F .m: E 85298 Em 3558 .550 63%.? m... .9285 .Eamu Emacs: a 30% $3 5 .523 .3 0:35.35 :33 96 9:4 llll 0mm «53 853 ZO_._.> \ xx I. m I. /\f. \\ It .0 --l}---\ x x m _ \\ _ \\ _ \\\ # P------.. INTERACTION OF LAKES AND GROUND WATER 14 555:9; $13 5 Bummzvm 53:8 .0: sow “m5 889$ 913.25 a *0 :9:qu SE3 .95on 2: E 68: EEGFE mo :25“:me MESS? .5193 ammo—ougild 559m 83 n Q \ ck F a: 5 Bsmaé \ $555. 88 c an 85:8 5:5 \ _l_.l|1_|—|_ \ \ ._.mm H. ooom o o o 6338. & 5:th \ .Ean Egan: \\\ m o>onm 38 5 €ch IIIII \\ .8 232?»... .33 3 9:4 .uwfl E ._m>m_ 9:: goes tam; 2 33:52 OMN .HEOQ cotmcmfim v6 >::_o_> E350 E836 £5 op 3:95 8 2:8 223 .325 \\.ul/ “or: $.QO Bug—m9. *0 mEmew om mp IIIIIIIIII , «/m|_< ._.mm_u_ mmwkmé ZO_._.o_ 3.2 30% v8: 2 52:32 .EBQ cozmcmfim ”.o 5E6; .8250 wmmeS. coo— fill ._.m m n. Doom E836 £5 9 329% 8 2:8 :5 EN; 23:8. ”6 SQmem om humu— m: mmwhwE O 9 4—H \ R r I‘llll 83 u 6: .9 P .mfi E 352.“.me Em 38th .650 $33.5. & E385 .632. 9355 m u>onm «3+ 5 ._m_u:$ .oa 0::Sb>r_ :33 v6 9:]. Ilomm 2an $me ZO_._ronm tau; 3 39:52 .ann cozmcmflm we 5E6; .9350 E23? $5 3 3:an 8 2:8 IImWWWWGII muVu—U SE 8.93. 0523.. we uEmem om mp II‘ Pwmu wawhms. ZO_._.::_o_> E550 @0me 18 wmmhmfi coop ._.m_ m. u Odom E2923 2.: 9 3:93 3 2:8 «St 8.8... 23:8; +0 2&6me 83 n 6: .9 _‘ .mE E UmEmEXm 2a 23“qu 850 o IJLJII. . o o o metms & 5:35 .EBmU “rm—Baum m «>03 38 E ._m_EB .8 232?»: Base +0 «:3 2an .825 cm . E lllllllll ._.m_m_n_ mam—km: ZO_._.m u m m>o m 8 m: .E: u w . _ _ v: n U z . we 2 228553 .qua cozmcafim +0 >250; .9980 E8923 £5 3 323% an 2:8 ||||||||| «Mano!» :2: 83% 23:3: yo Bnmem om m_. . x humu— mmmkmE ZO_._.ml.m~ 552m q CC CC qu C: CC CC CC <41 C C ‘1 Qd‘ C (C CC CG CG CG 0“ Ho 00 0° °° 00 do 0° 00 an an 00 Q0 dd ‘4‘ “ ooooonooooooooooooooooooooaooooooooooooooeoooooooooooooooooeoaaaoooooooooooooo aoooooooooonoooooooooa.ooaoooooonoooooooooococoooboooooooooooaoooeoooaon o 000 one one one one one 900 one oco 000 00090030 000 ooo>ooo one one can coo 00 000 can one one 000 one 000 one can one 000 one oobobooboaoofioaoouoaoono \\ $592 88 o \ \ Ti IO \\ hmmuooom o o 0 mm. x \\ \ \ Swami 25 3 323% 3 2:8 55 8.8». 23:8. *0 2953 cm mp hum“. wrmhws. [x 9‘ 0\ L‘ IIIIII-Illllll ooo..u€\ 8k 83 u a: f F é: E «552ng .95 2.338 550 .mb.t§ m.‘ 3335 .Eamu Eaten: a o>onm 38 E .523 .8 0:383; 338 *0 9:.— 0mm 033 8wa 2023.26}.me 21 ONE-LAKE SYSTEM .889? 2: «0 23:8 «5 5 51:83:59 ”532?: So. E3932 «0 SEEN giants :m mm: :2: 589$ 83.30 m we :0288 $33 .vcsohw 2: E 98: 9258?»: we 552::va MEBOfi :268 ummo~o€>ml.3 550E \ $322 82 ._.w m. u. Doom EEmEU «Eu 9 Um__anw up 238 «m5 $.QO 23:3: v.0 EQmem o2 "flax So; u a: ck F .mrF E toEmEXo Sm 23th $50 63mins & Rim“ 5 .Esme 9.353 a o>onm 38 E .3355 .8 232?»: :33 we 2:... 0mm 298 .325 202.4263me INTERACTION OF LAKES AND GROUND WATER 22 .8393 2: we 33 05 .m 3.2 «5 EC: 323: “538 FEE: *0 Susan nu ma: “m5 829$ 9—295 m «o :oiom: “Suzyécsohm 2: E «.8: 31:23: *0 comazflbmmv Matsocm c288 umw£o€>E|d~ 552m canaooooaooio oco : o o o o o econcoooonoowoowoonoowoooooon 0000000 00 07:9 a can one can a o o 0 on D o o o ooo;u$\ 82 83 n a: .9 F .m¢ E 35298 8a 2.3qu $50 wmmkms. coop o _ _|4 II _ .wEmtS. .2 $5535 .55“. ooom o o 0 .EB% .8353 m m>onm $3 E .528 .oa 232?»; Eng *0 wcfl #8... E ._w>m_ 3.3 m>onm 03; 2 53:52 m 3 $5 .EEQ coamcmflm we 3:53 3550 E292“. 35 9 023% 3 2:8 IIWIIIGII . . ‘1le $5 8.8m 25:3. 3 mamew om mF «(min—v» Hmmu. mam—bus. ZO_._.mldfi 552m 83" $2 any 83 u f \ ck F .m: E UoEmEXm 05 3.3qu $50 \\ 2mm: 82 o IJLIl . \ \ Emu. 88 o o o guts. & 5:35 .838 3323.». m w>onm 32 E .323 .oq 0:329»; :53 we 9:.— 0mm .33 E .35. 9.2 m>onm tau; 3 89:52 .HEOQ cow—Emmy... +0 BEE; .95ch 69533 £5 op Umzaam an 2:8 \III 55 8.8” 232m": we EQmem om mp ~ 0 z rmlx Em: mmmEs. zofimls~ HERVE \ L 5 a I. 8 I» 0 6 O 5 0 0 O \ O 0 n4 \ 5 2 \\ 11111 |\ 2 2 x \ \ 0v \ 0 \ 6 \ \ x0 2 . . .. \ cogukklmx no . , . \ > Or 83 n V: .9 \ /I\ \\ F .m: E 35298 \\ Sm 3.9qu $50 \\ wmmeS. coop o \ Till GENES. & R535 r1 Inllllr\ Em: 88 o o o .532” Egg: m Sons 58 E .635“ be 2:852. :33 *0 «En. Efimfiv £5 3 3:33 an 238 228 .325 Sr: 8.8m 23:54: *0 mEmem om mp ||||||||| Hum”. wmmkmi ZO_._.:|.w~ 559nm an on on no Jim on c an ac at no on as on 44 an a: an an cc .2. ad a: e a o o o 0 con coonoooooooooooooooooofloooooooaaoooooo$ooooooooeeoooo nonoeooeooooooeouoooooococoowooooocoooooooouooouooooo—uooooooo use can oooooooooooouo 000000 on o o oo oo oo on no —¢ k.vono“.eon.ooaoo:oeoenaoaaaaa.o>oeeanoaoooneo coco onooooooooooncooooooocoo 000000 c o o o a cannon aooo o2 uhviumx 83 u 9 \ .9 \ 0 \ mu \ P .mfi E 85293 w \ PE 88th $50 / \ mum—hm: 002 o \ fllJLll . \ / Hmmuooom o o o minim: & E335 E32. Emucfim \ a 269m 33 E .535; \ .2. 2.3235 :33 we $5.. u lllll \\ 0mm 293 .825 EEmmzu £5 9. 323% 3 2:8 um...“ 33% 33.:me ho EQmem om mp .rwwu. mmmhms. ZO_._.uH 3—».— WD ) ) J H4: 5 A” 1 218 [424412491 7 2 3 8 5 0 763881‘ 02 I453 1 9( 5 7 3181/46 3200 uu>oawuauoa ((.(..(..... . . . . . .........( ... . .M . ..(( s . 1 2 h EOHUHEMQUE "“1 “0]“04 00 o 0 1 0 5 1011000 “0 123 0 0 O 0 100201 01”” uuumom ... a) 0 eta 03m. XX XX XX XXXX XXX XXXX X XXXXXXX X XXX XXX XX XX XX X X X rkos 0 \ mace was: In new“ Inga—J XXXX X X XXXXXXXXXXX XXX XX XXX XXXX X X X X X X XX X XX vaM W omvt 03ml XXX X XXXXXX XXX XXXXXXXXXXX XXX XXXX XX XXX XX XX Mama 0 “wumnoamanfi XXXXXX XX XXXXXXXXXXXXXXX XXX XXXX XX X XXX X XXX X earv. 0. H3 XXXXX XXX Nani—Eco lllllll 5 XX nusucwnfl lllll 2. X X 2 $63350 1 _ XXXXXX XX .— XX X X XXXXX dun—0:580 raIIIIII a X P :uuwnun n llllll a B X X H 95.23: E If!“ i xxvmlwnXX XXXX XXX u ~33: M. Ill- 3‘ In... XX XXXXXXXXXXXXX XXXXX l1 muan m. Xxx XX F 8.52 XXXXXXXXXXXX 3 Ktm XXXX XX XX X X XX XX XX XXXXXX X X X X / «F K m XXXX X XX XX XX XXXX X XXX X XXXX XXXXXXXX X XX XXX $0 \ 1 X It IIIII Jill I‘lllllllllll‘lllllllqlllIlillitlllllllllll IIIIIIIII Illll. llllll 3 vm X XX XXXX X X X X X X X X X XXX XX X X XXXXXX X XXX XXX X XX X mm- Knm X X X XXXXXXXXXXXXXXXX X XXXX XX X XXXXXX XXXXXX 1 X a o X X X X .mm N XXXXXXXXX XXXXXXX 2: M 8 XXX XXXXXXXXXXXXXX XXXXXXXXXXXXXXXXXXXXXXXXXX XXXXXXXXXXXXXXXXXX S e v. T 5 h 2. XXXXXXXXXXXXXXXXXXXXXX nab. T mm M XXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXX XXXXXXXXXXXXXXXXXXXXXXXXXX m 0. 11 mt 1140514183916476072728491571985331 8110302723064/45858088670789598300236100918608350 Mum... 7289371549715666175J0613535720874 812648016144576013997900887991774383639119151243 a 1 0 7 1.4 9 0 L 8 7 8 4 9 3 4 2 5 8 l 0 1.0 5 9 7 9 3 9 3 5 6 6 9 3 &.4 L & l L &‘b 0 8 1 4 9 0 Q 5 0 0 6 4 4 8 7 8 6 1 7 7 8 4 6 4 i 9 5 5 5 5 4 1 6 5 6 6 1 7~Lfi 1H 21 1 2 222 21212 1 1 11 1 1211 11 111 1 11 1 1.1111112111121111 r m c n 1 1 2 3 4 4 4 1 1 7 8 7 8 7 8 7 8 4 9 B 9 1 1 8 1 8 0 1 O 4 . ... . . 2 . 4 4 2 7 ... e um 33343443311.41.431.431.331. 35.7.1.1333333394.544.3449...77.j§_7_g_§.7_7_47_s_§_7_477_ u nun-.52 nu“ 123/456789012345678901234567890123 4567890123/45678901235567890123456789012345678901 30“ 111111111122222222223333 3333334144414[4/44/4455555555556666666666777777777788 28 5Aquifer is in contact with base of sedimnts. oes not extend full length of the basin. hickness is 0.20T. Aquifer t 2A 3 MULTIPLE-LAKE SYSTEM 29 Some general relationships concerning the one-lake system can be deduced from table 1. For most hydro- geologic settings as simulated in this study, the following changes in parameters tend to weaken the difference in head between the lake and the stagnation point beneath a lake or tend to cause a lake to lose water: ( l) Lowering the water table on either side of a lake (lowering it on the downslope side has a greater effect than lowering it on the upslope side), (2) increasing Kh /K,,, (3) increasing the hydraulic conductivity of aquifers relative to the till, which has a slightly greater effect than increasing Kh /Kv, (4) increasing the depth of a lake, and (5) raising aquifers from the base to an intermediate level in the ground-water system. An exception ‘to statement 5 is the situation where an extensive aquifer that has low K h /KD and low Kaq /Kt (quadrant I, table 1) is raised from the base to an inter- mediate level. In this case the head at the stagnation point increases slightly. Aquifers that extend the full length of the ground-water basin have a greater effect on the interaction of lakes and ground water than limited aquifers. If no aquifer is present, a lake generally will not lose water even if K h /KU is 1,000 (table 2). Limited aquifers upslope and beneath a lake have little effect on the ground-water divide beneath a lake. Under the conditions simulated in this study, aquifers in these positions will not cause the divide to weaken regardless of the hydraulic conductivity of the medium. Limited aquifers downslope from a lake, whether at the base or at an intermediate level of the ground-water system, have a significant effect on the inter- action of lakes and ground water because under many conditions of high K h /Kv and KM /Kt and low water- table configurations, a lake will have a weak ground-water divide or will lose water. In every simulation run for this study, the stagnation point associated with the ground- water divide is always under the downslope littoral or shoreline zone of a lake. This has particular significance when designing an observation-well program for studying the interaction of lakes and ground water because deter- mining the position and head, or absence, of the stagna- tion point is the key to the relationship. MULTIPLE-LAKE SYSTEM In many geologic environments, lakes occur at different altitudes within major drainage basins. Although it is a common belief that water moves successively from higher lakes to lower lakes through the ground-water system within a major drainage basin—a situation where all the lakes would be the flow-through type, it is shown in the following discussion that this may not be the case. Digital simulations were run of a number of variations in the hydrogeologic setting of three lakes along a regional slope from a major divide to a major drain. The vertical exaggeration of the sections are the same as those in the one-lake system (80:1). MODERATE REGIONAL SLOPE AND LOCAL RELIEF The length and thickness dimensions of the multiple- lake system are different from those of the one-lake system, thus the proportions of the features are different. As before, the dimensions in parentheses are included as examples of realistic field situations. A three-lake simulation set will be discussed later in which the models are of a system that has a water table with very low regional slope, a physiographic condition common in many geologic terranes. One type of lake-ground water setting not simulated in this study is that of a straight-line, sloping water table between adjacent lakes. As will be apparent in the following sections of this report, the height of the water table mound on the downslope side of a lake relative to lake level is an important control on the interaction of lakes and ground water. Lakes in a setting of a straight- line water table simply would gain ground-water inflow on one side and lose water out the other. The thickness of the ground-water reservoir of the three- lake system ranges from T (as much as 250 ft; 76 m) on the upslope end of the section to 0.40T (100 ft; 33 m) on the downslope end. The height of the water table above lake level on the regional upslope sides of the two lower lakes is 0.20T (50 ft; 15 m) and on the downslope side of all three lakes is 0.08T (20 ft; 6 m). The highest lake in the section has the water table 0. 12T (30 ft; 9 m) above lake level on the upslope side. The difference between the level of the highest lake and the elevation of the water table on the upslope side of that lake was purposely made less than the two lower lakes because it is believed that lakes that occur high topographically are more likely to have lower water tables adjacent to them than the lower lakes. The water- table altitudes were varied in some of the simulations. The dimensions of the shallow lakes are about 0.1L wide and 0.20T deep. The aquifers are about 0.04T thick and those of limited areal extent are 0.15L long. As before, the datum for the head values is 100. The positions of the three lakes are as follows: The surface of lake 1 is about 0.65T above the base and 0.32L from the left side of the section, lake 2 is at 0.75T and 0.60L, and lake 3 is at 0.88T and 0.85L. The aquifers are positioned at the base and, if at an intermediate level, at 0.30T above the base. The lateral position of the limited aquifers varies as noted in the text. A simulation of a three-lake system in which each lake is shallow, is underlain by sediment, has a 0.08T (20 ft; 6 m) water-table mound on the downslope side of each, has a ground-water system whose K h /K,, is 1,000, and contains no aquifers, is shown in figure 6. The position and head of the stagnation point on the ground-water divide beneath each lake is shown as in the discussion of the one-lake simulations. Under the given conditions, each lake gains ground-water discharge from its own local ground-water flow system, and each is completely isolated from the others. 30 INTERACTION or LAKES AND GROUND WATER For the three-lake system, comparison of all possible combinations of different water-table altitudes, lake depths, hydraulic conductivity ratios, and size and posi- tion of aquifers, as was done for the one-lake system, would lead to a far more complex summary table. The basic goal of simulating the multiple-lake system is to determine if the lakes have interrelationships with each other that would not be evident in simulations of the one— lake system. The results of the multiple-lake simulations in which the water table has moderate regional slope and local relief are summarized in table 3. The effect on the interaction of lakes and ground water of an extensive aquifer at the base of the system can be seen by comparing row 1 with rows 2 and 3 of table 3. If an extensive aquifer at the base of the system has a hydraulic conductivity 1,000 times greater than till, and the K h / K U of the ground-water system is 100, the difference in head between the lake and the stagnation point beneath lake 2 is very small. In this simulation, there is very little change in the head at the stagnation point for lake 1 compared to the simulation listed in row 1, but there is water loss through the bottom of lake 3. If Kh /K,, is increased to 1,000 (row 3, table 3), holding all other parameters as in row 2 of table 3, a very strong divide is established beneath lake 1, and lakes 2 and 3 lose water through most, or all, of their bed (fig. 20). Evidence that the lakes are independent of each other can be seen by comparing rows 6 and 7 of table 3. In these simulations, Kh/K, and KIM/K, are both 1,000 and two limited aquifers are at the base of the ground-water system at the 0.45L and 0.75L positions. The only difference between the two is that the water—table mound between lakes 2 and 3 was decreased to only 0.04T (10 ft; 3 m) above the level of lake 3, and the upslope side of lake 3 was decreased to 0.08T (20 ft; 6 m) higher than lake level for the latter simulation (row 7). The effect of this change in the water-table near lake 3 on lakes 1 and 2 is very little, whereas the head at the stagnation point beneath lake 3 decreases 0.003T (0.8 ft; 0.2 m). If a limited aquifer is added at the 0.15L position, and all other parameters are as in the simulation listed in row 7 (table 3), the head at the stagna- tion point on the ground-water divide beneath lake 1 decreases considerably, and there is little effect on lakes 2 and 3 (compare rows 8 and 7). The effect on the interaction of lakes and ground water of moving the position of limited aquifers laterally can be seen by comparing rows 14 and 15 of table 3. In the simula- tion listed in row 14, the aquifer at the 0.15L position is downslope from the edge of lake 1 (fig. 5). In the simula- tion listed in row 15 (fig. 21), this aquifer is shifted slightly in the upslope direction (to 0.22L), so that the upslope part of the aquifer is partially beneath the downslope side of lake 1. All other parameters for these two simulations are held constant. This change lowers the head at the stagna- tion point near lake 1 by 0.002T (0.5 ft; 0.2 m). Thus, it is evident that the interaction of lakes and ground water is sensitive to the lateral position of limited aquifers down- slope from a lake. The maximum effect is felt if the upslope part of an aquifer is beneath the downslope side of a lake, and the aquifer underlies the water-table mound downslope from a lake. The effect on the interaction of lakes and ground water of moving the limited aquifers vertically from the base of the system to a middle position is similar to the one-lake situation; that is, the head at the stagnation point is lowered. (Compare rows 6 and 12, and rows 11 and 19 of table 3, and see also figs. 22 and 23.) Note that the middle aquifer in these two simulations is shifted to the right, from the 0.45L to about the 0.53L position, compared to the other simulations discussed to this point. This was done for most of the multiple-lake simulations because of the findings mentioned in the previous paragraph. The effect of deep lakes compared to shallow lakes is also similar to the simulations of the one-lake system. Comparing rows 9 and 22 of table 3, in which all para- meters are similar except lake depth, shows the relatively greater influence on the ground-water flow systems near deep lakes compared to shallow lakes. The ground-water divide beneath lake 1 in the simulation of shallow lakes (fig. 24) has a head at the stagnation point 0.004T (1 ft; 0.3 m) greater than lake level, whereas in the simulation of deep lakes (fig. 25), the lake loses water through two- eighths of its bed. Beneath lake 2, the head at the stagna- tion point decreases from 0.014T (3.4 ft; 1.0 m) greater than lake level in the shallow-lake simulation to 0.008T (2.1 ft; 0.6 m) in the deep-lake simulation. Beneath lake 3, the difference in head between the lake and the stagnation point of 0.004T (0.9 ft; 0.3 m) for the shallow lake, can be compared to the deep lake which loses water through three-eighths of its bed. In most of the simulations of three-lake systems discussed thus far, K“, /K, and Kb /KU are 1,000. These values were chosen because, as shown in the one-lake- system simulations, they tend to create conditions for maximum water loss from the lake to the ground-water system. It is of interest, therefore, to compare rows 21 and 22 (table 3) and figures 25 and 26. The simulation summarized in row 21 has Kh /Kv=100 (fig. 26), whereas in. the latter this ratio is 1,000 (fig. 25). The stagnation points occur very deep in the system in the former and, in fact, the local flow cells around each of the lakes extend nearly the full thickness of the ground-water system and little ground water flows past the middle lake. It is evident from the above discussion that, in most cases, lakes in a multiple-lake system act essentially independently. Changing the height of a water-table mound or the position of aquifers by one lake and not by the others has a considerable effect on that one lake that is similar to the equivalent one-lake simulation. The change may have a minor effect on an adjacent lake, changing the head at the stagnation point, for example, by perhaps 0.0004T to 0.0008T (0.1 to 0.2 ft; 0.03 to 0.06 m). MULTIPLE-LAKE SYSTEM 31 The only situation where a considerable effect is felt on the lakes is in the case of extensive aquifers within the ground-water system, and where K h / K ,, and K a q /Kt are large (row 3 of table 3) (fig. 20). In this case, and there are similar cases in the one-lake settings, there is massive downward movement of water in the upper region of the ground-water basin and massive upward movement in the lower region. The great difference in head between the lake and the stagnation point associated with the lowest lake seems to be established because of the large influence of the water-table mound on the downslope side of that lake. If the water-table mound were not there, there would be massive upward movement on the lower side of the diagram and the hinge line (the point at the water table separating regional recharge from regional discharge) would be near the middle of the diagram. Apparently the mound is sufficiently strong to cause downward move- ment that will not be overcome, but instead, will be responsible for a strong ground-water divide that has a head at the stagnation point of very large magnitude. It is interesting to note in this case that the stagnation point is not beneath the lake. Rather, it is downslope from the lake and is nearly directly beneath the highest part of the water- table mound. LOW REGIONAL SLOPE AND LOCAL RELIEF A series of simulations were run of a multiple-lake system in which regional slope and local relief were kept relatively low (summarized in table 4). The length and thickness dimensions of the multiple—lake system that has low regional slope and local relief are unlike those for the system of moderate slope and relief, thus the proportions of the features again are different. The thickness of the system ranges from T (160 ft; 49 m) on the upslope end to 0.60T (100 ft; 30 m) on the downslope end of the section. The water table on the upslope side of all the lakes is 0.13T (20 ft; 6 m) higher than the adjacent lake level, and on the downslope side is 0.06T (10 ft; 3 In) higher. The difference in altitude between the lakes is 0.06T (10 ft; 3 In). Only shallow lakes 0.06T (10 ft; 3 m) deep are simulated. The aquifers are 0.06T (10 ft; 3 m) thick, and the limited aquifers are 0.13L long. A ground-water system that has no aquifers is shown in figure 27. Ground-water divides of equal size and strength (the difference in head between the lake and the stagna- tion point is 0.02T (3.0 ft; 0.9 m) by each lake) occur beneath each lake. It should be noted that simulations of this system of lower slope and relief differ somewhat from the other three-lake system discussed previously in that the water table relative to lake level by the highest lake is the same as by the two lower lakes. Simulations that have extensive aquifers at the base of the system, K” /K, =100, and Kh/KU=1,000 (row 27, table 4), show no water loss from any of the lakes for any of the simulations. This is true also if the hydraulic conductivity ratios are reversed; that is, K“, /Kt =1,000 and K), /K,, =100 (row 28, table 4). In simulations that have both ratios at 1,000, a ground-water-flow pattern similar to the other three-lake system occurs—there is water loss through the entire bed of the higher two lakes and a divide established beneath the lowest lake (row 29, table 4). In simulations that have limited aquifers beneath the water-table mounds between the lakes, and the up-slope edge of the aquifers are partly beneath the lake, a ground- water divide occurs beneath the lake, although the difference in head between the lake and the stagnation point of each is rather small (fig. 28). This result holds even when the limited aquifer is moved to about 0.13T (20 ft; 6 m) below the bed of the lower lake. In the three-lake system that has low regional slope and local relief, it appears that the ground-water divides tend to be slightly stronger in most cases than where a higher regional slope is simulated. The slope of the water table from the lowermost water-table mound to the left edge of the hydrologic sections in the simulation of moderate regional slope, shows a drop of several tens of feet. The water-table slope in this same part of the section, in the simulations of low regional slope, is essentially flat, resulting in very slow ground-water movement. This very flat water table near the valley of the major drain on the system acts as a “check valve” on the entire system and tends to increase the tendency of the lakes to have stronger local ground-water-flow systems than in conditions of a higher-sloping water table. In none of the simulations does water move from one lake to the next downslope if there is a water-table mound between the two lakes. Even if an upper lake loses water to the ground-water system, the movement generally is deep to the system and the water moves beneath the local flow systems associated with lower lakes. Thus, even in settings where lakes occur at various altitudes down a sloping valley side, each of the lakes, for most settings, acts as an independent entity and has little relationship to lakes either upslope or downslope from it. This] conclusion gives increased importance to the summary table (table 1) of the one-lake system. The effect of reservoir thickness on ground-water-flow systems is extensively discussed by Toth (1963), and to a lesser extent by Freeze (1969b). The general effect of a thin ground-water reservoir is that regional flow does not occur for many settings, especially if regional slope is low. To check the effect of a thin ground-water system on the inter- action of lakes and ground water, a setting was simulated that was 100 feet (30 m) thick on the upper end of the section and 50 feet (15 m) thick on the lower end. The difference between the section and figure 29 is that the local flow systems associated with each lake extend to the base of the ground-water reservoir, and regional flow does not occur (row 33, table 4). INTERACTION OF LAKES AND GROUND WATER 32 .Euumhm uuum3lv§ouw BOHan—wm $333 goal—flan: Me .thiuu Nana—dumua Ma .33.— mgu we nuwuuH :5 3.0qu no: muov uwwH=v<~ .523 2.3 we numnuH HHsu avnuuxm Hafiz—«f «“52 N32 v5» 23?. 35.8%.: 33385 3.‘ mES .533 30:3 5329. $3.33335 go $925353 \0 EaEEsmlfi “EEC. Chm A To 5m 3me x x x x x x . x x x mmaéano mm m o e o x x x x x x w x x x x . . 3.5.2 «é 1~ x x x x x x x S 3.3-... ~m 5H m.H 9H x x x N x N x x x 3.3%le Hm x x x x x x x M x uméuHHnw on w\w a: . mun x x x x x x x N x x 4H.muHH|e ¢~ . m.H m.~ x x x x x x x x x x wméuauw mu .u o a H n N N x x x x x N M x om.:m|© 5N o.m o.n 9n x x x x x x x x x 0H 93H uwum: Hm>uH 33H W H W U W H N .9: E 8 c. E u a w m N mwmcH 26%.. 2125 mwmoH gong ucHoa mmeH 36% 250; 1 1 K H T: K a a d U W W... m “a :5 :oHuuawmum uunu =OHuu=mmum amp—u uoHuuamwum u o u NM. NM aun— uo mo Emu..— von we we v.3: vwn we we vww: anOHm unchnD wnaHm magma: «AOHm wQOHmna qud was viva 3<5~ Inudmul mg a 1 ucwuuwm mo uanmm uanHNm mo uano= uaoouwa we ustwm IESV 15.5% IESn 33..— n W W. a w u wiHuumm .HHEH I. m x A mMHawww H wxuq m 93; N SE.— H «and cowuHmoc nuuH=v< m Ammo—51:5 soummm H33 3 u>HumHunV Ho>uH 35H 93%. .933 Sun: we “53¢: 3:2 :82 3:» 2x0? ~58me 32 3: 33: $33 86$: $399. $3303.53“: \0 EQESBSB .3 3a5§3ml+ HES—l .n 33H Eouu waning: ”was ad “0333 HMHuummm .426 cu Amie Baum quMHF—m hug—Em .me Jinan 3—... m0 535: .35 “:03; you muov hawH=vwH Hw>wH Hm>wH uOHumoHuHunuuH WA yuan: 3—3 953m uuuma 93H 02:? hwy—w: 93H u>onw . u gHumHsaHm 1 munoH usHon nonoH ucHon momoH 252— Sn .0 — ommm 5 c3— SQFQ nunww unsu =0Hun=wuuw uncu QOHuunmuum uwsu cont—Baum wQOHm unOHmaa 2*on magma: oQOHm unOHmn—D ume vwn no mo vac: awn we «0 cum: was m0 «a was: IEEv Inge LES—v LHmHuumm HHHah unwuumm we 33% £533 mo £30: “E35,— wo 239: 53:2— 3:22 n 9:3 N 3:3 H 3—3 m 3G...— N 33H H use..— muH=mom Idlqus—uunu Eoumhm in u a» 933.35 Hu>oH mu—nH u>onm anwu uwuma we ”Emu: 33 MULTIPLE-LAKE SYSTEM .5035 2: “0 fine 2: gm iguuzvcg 9258?»: am“: 32:33“ «0 335m u>mm5§o :m was :2: 829$ BEEQEJE m «o doiumfl $33.95on 2: E 98: 9259?»: Mo #5153me 9:305 :ouuum ammo—SEVIIdN 559m 0 8 . oq a... .- 3 . u.- ..o n: ... .. ... .o u .u u... u . o a a a a n A . .u......u.u.... ........o. o. .c ..u..-...... ..u-.ou.oo:.u... E. .«u.......... . . a . . o u... .. ... . . .u. on: a». an a a n .o I. 3.. a . a . o. . o. .- 3 can OLZ\ . on on com \ \ \\ muEEm 9.3 w 3 .6 _ \ 9 I. .6 . ~ um 69:2: 3.3 8 0 9 coo F u k\ k 9 83 n ci ck P .mc E UmEmExm PE 3358 EEO .mBth. & 3:th .Eamu Emncfim wmwhms. coop o m 30% $3 E .333 _ o o hum”. Doom o .oa 0:323»: 3:00 we 2:1. 0mm .38 E .33. 8:: 963m 68; 2 528:2 m B m .qua cozmcmmuw we >::_o_> 33ch £236 WE“ 3 “5:25 8 238 23 u H >> xii; was“ 3.8..” 25:3; *0 mEmem om mp Illllllll ~ . a WW Em”. wmmEs. zoFo_ 3.2 m>onm gm; 2 39:32 .pan cozmcmfim v5 >u_c_o_> 6850 \l\lll .9: ll\\ 00 7263 a 8 7V 7.26 7v 0 666 9 8 I. 99 9 0 7ulr0 C._ O 0 Iro 0 \ a A 39:5: 373 O mmmkmS. coop o o 0 Hum”. ooom 0 E935 £5 9 623% an 238 as...“ 333 23:8; *0 EQmem om _ my .55“. mum—Hus. m m>onm 58 E -oa 0238.9»; Esau we mg] .688 V8.6 000 |‘\l \Z \ L 0 80; u “v: 8k 83 u Q \ f F .m: E umEmEXm Em 885mg 550 .mENtms & REES .Eam—u Emu—Sum .5598 0mm 2an .533 ZO_.ro_ 3:2 m>onm Um»; m_ .3832 .wan cozmcwmum +0 57:03 3550 \‘utlt a 0.0 u I \ IlI\ .63 Z .6 3.6 .6 RURU AU IL 87v 0 ‘ 0/ \ way/Ina \ \ 9 2.3% 8.3 \F L thEfi—C wxwl— mmwkm—z coop o :Jrloo ._.w m. u. ooom 0 E235 £5 3 3:an an 2:8 Sr: 838 2::me U6 Bamem om mp Hum“. wmmkm—S. \ \ 83 u w: any 83 n Q \ :2 F .m: c_ 3:835 96 2:3qu 550 633g? & \mEmt: .EBmU Emocmu m o>onm $3 5 .523 -oa 0:32P»; :33 we we] 0mm 298. 5:25 ZO_._.m_ $5 26% gm; 2 39:32 .Hqu :ozmcmfim +0 3E5; 3350 £835 25 9 U23% 8 2:8 Ilmmamluhomwgll \l/ S5 $.QO 35:8. we mEmem om mp ~ . — _,w T SE 35% zo_k¥ \ ck 33:5: 3_3\ F .9», E UwEmExm Pa mEfluca 850 63mins & EE& 5 \ m 8 .EEau Panama In W o meL-WS— 00°F 0 N m>ODN Hmmuv F: ._N_HCmH O Hmmn. oo_om o o o be 0:25P»; .macw +0 2:.— 0mm .33 E ._m>m_ 9.2 30% gm: m_ 89:32 0 m m m .E_on cormcmfim ”.0 BEBE 3550 £836 25 3 3:95 B 2:8 3 u ‘_ a >> IIIIIIIIIII \dufll. 55 8.8“ 23:3. *0 BQmem om m— P--\\ ._.m_m_u_ mmmhmé ZO_._.m_ 3.2 36% tam; 2 32:52 m .qua cog—5.3m +0 >t:_o_> .8280 EEmEU £5 9 633% 8 2:8 Ilomuwuuwmwgll \.\.:l :2: 33% 25:3: U6 BQmem om m— .,--.r.m\\ Emu. mmmEE zofim_ 3:: 303m cum: 3 39:52 .qun cotmcmgm v8 >u_c_o_> .8230 .oo~u=M\:v~ van KEEP? 2: we own: 2.: am 23:3 UBME: ~o flufiaca N€555 QESLBQS :wE £81: 3% was :3. 589$ 0:263:35 m «0 :0338 BEBAESBM 9: 5 use: ozsmhsf we :OuSQEmG MEBOfi coauom ummofiefiimldw 559m 8 9 726.6 7.. Ir 0 668 I— 0 0 .700 o \\ \musttm 9.3\ _. \ . \ 82:5: 3.3 wrmhms. 009 o _ _ o Hmwu Doom o 8535 £5 3 vmzuam an 338 «St $.QO 23:8; ho mEmew om kwmu. 7267a a 9.78 7v 000 0 % x Z L 1:09‘ 08 0 0 mp wmwaE \ \\7U .69 90 a \ \ \ ooo..uf\8¥ 2: ">219 F .mz E UmEmExm 2m 33th $50 GENES. & F558 5 .Eamu Emccfim m m>onm 33 E .6593 .oa 2 32?»; Egg +0 wcfi 0mm 033 $me ZO_._.m_ 8.2 30% E8; 2 39:52 . m an .Hqu cozmcmmum v8 3E6; .2230 ””2 ”Mumflmwwwhw‘” cmwmmcflug o.m x) 5 _ o .H ._ 9F _ w om m— :J.\. Em: $5.62 \I'l'|'lllllll 'Illlll'l" ' \ 7.. Ir 0 7v 0 9 83 u a: .9 F .m: E 35298 2m 33th .850 .mEth. fl 5:3 5 .EBmU 935% m 30% 58 E ‘_m_u:$ -oa 0:396; :33 $0 2:.— 0mm 293 .325 ZO_._.EIdN 559m .6 7w 7v 9 .V .7 o 9 Ir 7v 8 C. .6 V Z Z Z Z S 9 7o .6 3 Ir 0 9 Ir 0 7v 9 9 .38 E ._u>u_ 9.2 30% cam; m_ huoEzz dan cozmcmfim we >::_o_> 5.230 0mm 8 ll \F 33 9 .6583: $7.1. mmmth. coop 0 Ti 5588 oo o 683% 25 9 .8:an we 2:8 «m5 8E8 23:8. +0 gamem om mp hum". wmewS. \ Iva why—5.33 wv_ml_ \ C.— O°°~F H 7v Ir 0 \I'|l'|'lll'll| 'l'l’llllll 903 my 83 M Q \ .9 F .m: E 8:635» 95 «53qu $50 .mENtS. & E385 .Eamv Emucmu m u>oam How», 5 .635“ -on 0:323; .250 *0 33 0mm 293 Sam; ZO_._. .5 / “'3 LESW/ HONDURAS C? ‘/1 1‘", IN ‘ 5 ' A{‘_‘:’_L.~- Guula’n \ '1 Q- \ r a??? \ alamoo ‘\ ' .— v I010?“ ’ Joyubaj \,BAJAVERAPAZ /’ 650 ‘1 ZACAPA Y o ‘ — é‘ / 1 \1 o ,/ 069‘ Y_' 02 Q Is 00 2“ ,/\,?R El \ acapa J! O. \ - ‘c—A E onqyaspf' ,\\__’.’—""‘ ( A7 / 7‘ \. ATEMN- -\ J/ Li.) \ Guarpmala‘ I" "1 CHIQUIMULA , C'” JALAPA ./ J mf’p\"j—L--‘ r/ '5." | ./'C"“°"g r JUTIAPA ' ‘\~ '56 [SANTA ROSA 9. / _ , EL SALVADOR TV I fly I I4°oo' 90°00' 89°00' 88°00' FIGURE 1.—Guatemala and its Departments (States), including its general location in Central America. Base map modified from Guatemala, Instituto Geografico Nacional, 1974, 1: 500,000. of the earthquake investigation team. The open-file report and this publication represent the immediate response by our agency to the request of the Guate- malan government. Guatemala has a long history of damaging earth- quakes. Historic chronicles date earthquake occur- rences in Guatemala from the period of the Spanish conquest. The chronicles indicate that the cities of Antiqua and Guatemala City have been badly dam- aged by earthquakes more than 15 times since the early 16th century. The most damaging earthquakes to have occurred before the February 4 event were on December 25, 1917, and on January 3, 1918. These earthquakes and their aftershocks claimed numerous lives and partially destroyed Guatemala City. The February 4 earthquake caused extensive damage, as the following statistics from a few of the cities and Departments (States) show: 88,404 houses were destroyed and 434,934 people left home- less in the Department of Guatemala, about one- half that number in the Department of Chimalte- nango (fig. 1), and all houses were destroyed in the towns of San Pedro Sacatepéquez, El Jicaro, Sumpango‘, Tecpan, and Gualan. From the above, it is obvious that the principal hazard to life from earthquakes in the Republic of Guatemala is the collapse of manmade structures because of ground shaking. There are, in addition, other serious earthquake hazards, especially land- slides in the highlands and surface breakage along active faults. Landslides can interrupt lifelines, in particular, roads, railroads, and communication net- works, and they can dam streams and cause lakes that are subject to sudden release upon failure of the dam. There is much to learn from this earthquake that is directly relevant to the problem of reducing earth- quake hazards in the United States. The fault that produced the earthquake is the boundary between two crustal plates and is tectonically similar to the San Andreas fault in California and to the Fair- weather fault in southeastern Alaska. Many of the modern buildings in Guatemala City are similar to structures in the United States and in particular to those in cities in coastal California. The different papers of this report present (1) the tectonic setting and the seismicity of the region; (2) the local seismic activity before the February ‘4 earthquake; (3) the main event and principal aftershocks as recorded at teleseismic distances; (4) the source parameters of the main event and possible future earthquakes on the fault that slipped on February 4; (5) estimation of strong ground INTRODUCTION 3 motion due to the main event; (6) aftershocks from local data and their relation to primary and second- ary faulting; (7) the geologic effects, including pri- mary and secondary faults, landslides, and relation of faulting to damage; (8) the intensity of shaking and its distribution, casualties and damage, ground- motion parameters at intermediate distances, and source parameters from field observations; (9) damage and engineering implications and earth— quake-resistant design code; and (10) design, con- struction practice, and general observations on dam- age to high-rise structures. The results from this collective investigation will aid in the reduction of earthquake hazards in Guate- mala and in the United States, but the material presented in this report should be considered as preliminary. Some of the details may change, and the conclusions may be modified upon further scrutiny of existing data and upon analysis of new data gathered in later geologic field studies. ACKNOWLEDGMENTS We gratefully acknowledge: Colonel J. Guillermo Echeverria Vielman, Coordinator of the Comite Na- cional de Emergencia; Mr. Jose Asturias, Scientific Coordinator of the Comite Nacional de Emergencia; Dr. Eduardo Gonzalez Reyez, Executive Secretary for Education, Science, and Culture of the Organi- zation of American States; Dr. Gabriel Dengo, Di- rector of the Instituto Centroamericano de Investi- gacion y Tecnologia Industrial; and Ing. F. Her- nandez Cruz, Director General, Instituto Geografico Nacional, for their full support in our investigation. Mr. M. Freundel, Assistant Director USAID/ ROCAP, provided ground transportation and an ex- cellent chauffeur. We wish to acknowledge the sup- port given to us, in the questionnaire canvassing of Guatemala City, by Ing. M. A. Castillo Barajas, Director, and Ing. H. E. Molina, Assistant Director, of the Civil Engineering Department of the Univer- sidad de San Carlos. Also, we are grateful to the university students who participated in this effort and to Dr. S. B. Bonis of the Instituto Geogréfico Nacional, who also obtained a great number of intensity questionnaires for our program. We wish to thank the Camara de Construccion and the Insti- tuto Geografico Nacional for providing us with temporary office space. We also would like to thank the helicopter ofi‘icers and staff of La Fuerza Aerea Guatemalteca for their efficient and enthusiastic support of our canvassing fieldwork. We gratefully acknowledge the support given to us by members of the Guatemala Air Club and par- ticularly Messrs. Rudy Weissenberg 0., Antonio Delgado Wyld, and Francois Berger, who gen- erously made their personal aircraft available for our reconnaissance studies of the Motagua fault. Mr. David Schwartz of the Woodward-Lundgren Company participated with us in some of the field- work along the Motagua fault and freely shared his personal knowledge of the geology in part of this area. We wish to thank Messrs. D. F. Miller and A. Sunde‘rman, engineers with ROCAP, for their support while we were in Guatemala. Also, Dr. 0. Salazar of the Instituto Geografico Nacional gave full support to our efforts in Guatemala. We wish to thank Mr. A. Gregg of the Inter- American Geodetic Survey; the US. Military Group; Mr. C. Urrutia, Director, and/,Mr. E. San- chéz, technician, of the Observatorio Nacional for their support to our seismic network task. The heli- copter support for this program was provided by the US. Army, 114th Aviation Company, from Panama. Mr. J. Irriarte, a geologist with the» Insti- tuto Nacional de Electrificacion, gave us field sup- port and participated with us in the deployment of our seismic equipment. Our most sincere thanks are also due to the Guatemalans, both private citizens and public officials, who helped by relating their personal‘experiences about the earthquake, to those who provided support to our mission, and to those who provided information to our efforts. Without their support, this investigation would not have been as fruitful as it has been. THE GUATEMALAN EARTHQUAKE OF FEBRUARY 4, 1976, A PRELIMINARY REPORT TECTONIC SETTING AND SEISMICITY By WILLIAM SPENCE and WAVERLY PERSON INTRODUCTION The February 4, 197 6, main event occurred within the Motagua fault zone, the active boundary between the North American and Caribbean plates. The tectonic setting for the region of the Caribbean plate, with seismicity for the years 1962—69, is shown in figure 2. The absolute motion of the Caribbean plate has been shown to be nearly fixed, assuming stationarity of the hot-spot reference frame (Jordan, 1975). Thus, the average relative plate motions—Cocos- Caribbean (7.47 cm/yr at 253° azimuth), North American-Caribbean (2.08 cm/yr at 252.4° azi- muth), and Cocos-North American (9.01 cm/yr at 35.0° azimuth)—directly cause the seismicity of the region of Guatemala. These plate motions were cal- culated at lat 15° N., long 90° W. from relations given by Minster and others (1974). The rupture of the February 4 earthquake progressed into the zone of the triple junction between the Cocos, Caribbean, and North American plates and thus assumes a particular significance. SEISMICITY The seismicity of the zone from lat 13°—18° N., long 87°—95° W. for the period 1902—75 is shown in figure 3A. The main event and its nine largest aftershocks, through March 7, 1976, are shown by x’s (Person and others, this report). The Motagua fault zone is notable for its lack of major earth- quakes from 1902 (or earlier) until the February 1976 main event and is a clear example of a seismic gap on the Caribbean-North American plate bound- ary. The one large earthquake epicenter shown near the western end of the mapped Motagua fault (Ms=7.5, 1921) is assigned a focal depth of 120 km (Gutenberg and Richter, 1954) and thus is as- sociated with the subducting Cocos plate rather than with the Motagua fault. The eastward exten- sion of the Motagua fault, the Swan fracture zone, has been seismically active during this period. It is 4 characterized by shallow, left-lateral, strike-slip faulting (Molnar and Sykes, 1969). The maximum seismicity in this figure occurs with the zone from lat 14°—15° N., long 91°-94° W. and suggests that a complex tectonic regime is associated with the Carib- bean-Cocos—North American triple junction. Five cross sections through this Middle America seismic zone, for the period 1962—72, are shown in figure SB (A. C. Tarr, written commun., 1976). These show that the diffuse seismicity in figure 3A is largely attributable to the underthrusting of the Central American continental mass by the Cocos plate. Of the 1,331 earthquakes that occurred dur- ing the period 1902-75, shown in the map portion of figure 3A, the 17 events with MSE7 .0 apparently are associated with the subduction of the Cocos plate. The seismicity of these cross sections is more diffuse than that of seismic cross sections that are based on high-quality local data for island arcs (for example, Engdahl, 1971, 1973; Mitronovas and others, 1969). An increased resolution of the dip of the Cocos plate beneath Nicaragua has been ob- tained by a relocation of earthquakes in that area (Dewey and Algermissen, 1974). An improved pic- ture of Guatemalan regional seismicity as a series of cross sections could provide insight into the nature of the associated triple junction. A major seismic gap has been noted at the west- ern coast of Central America between long 88° and 91° W. (Kelleher and others, 1973). Now that the Motagua fault-lock zone has ruptured, the possi- bility that additional major earthquakes could oc- cur deserves special study. South of the Motagua fault, an eastward relaxation of the Caribbean plate coupled with an accelerated local subduction of the Cocos plate could result in underthrust-type earth- quakes that would fill this seismic gap. North of the Motagua fault, an accelerated westward motion of the North American plate could override the Cocos plate and lead to normal faulting off the southwestern coast of Mexico. TECTONIC SETTING AND SEISMICITY 30. 70° Fig/Gian Peninsula 0 20"" U ‘1 \, RTH‘GAMEIQ gs: . “s\ I°CAN PLA E “2 - A d 'Puerto Rico Trench n. ' ._ . ()ritntc F‘Z' ' Cayman Rise Q I I I 1‘ I ~., "err—N "D f " ' lO'N CARIE 3BEAN PLATE ”A ‘ ' _ : Q9 3 - ,,/ _ U”? \V' 7/ ' % 1 soun— AMERI PLATE Mercator Scale I! Proioction I ‘5' N 90°W FIGURE 2.—Tectonic setting of the Caribbean plate. The Motagua fault zone, Swan fracture zone, Oriente fracture zone, and Puerto Rico Trench are elements of a major transform fault system on ’which the North American plate moves westward, at 2.1 cm/yr, relative to the nearly stationary Caribbean plate. Large arrows indicate direction of plate move- ment. The location of the February 4 Guatemalan earthquake is shown by the large dot. Small squares are previous earth- quake locations. Serrated line indicates subduction. Modified from Jordan (1975) by Tarr and King (1976). HISTORICAL SEISMICITY Historical accounts of damaging earthquakes in Guatemala are numerous and date from the time of the Spanish conquest. A particularly destructive earthquake occurred in 1541 near Ciudad Vieja, the primary early Spanish settlement in Guatemala, and caused the deaths of approximately 150 Span- iards and at least 600 Indians and Negroes. The Spanish settlement subsequently was moved to nearby Antigua, and this city became the original capital of Guatemala. Antigua was extensively dam- aged by earthquakes 11 times between 1565 and 1773. Accounts say that Antigua was destroyed in 1586, 1717, 1773, and 1874. Following the 1773 earthquake, Guatemala’s capital was moved to Guatemala City. Before 1976, Guatemala City had experienced damaging earthquakes in 1917-18, 1863, and 1862. The December 25, 1917, earthquake destroyed or seriously damaged about 40 percent of the houses in Guatemala City. This earthquake was followed by large earthquakes on December 29, 1917, January 3, 1918, and January 24, 1918. The most damaging earthquake of the 1917—18 series was that on J an- uary 3, 1918. A chronological historical record of important Guatemalan earthquakes, from 1526 to the time of the February 1976 main event, for noninstrumental and instrumental data is given on p. 88. Included are earthquakes occurring in or offshore from Guatemala that caused damage or are known to have had a magnitude greater than or equal to 6.0. The source of descriptive information for each earthquake is indicated in parentheses. At this writing, primary sources have not been checked for all earthquakes, and thus the information in the chronological listing may contain occasional omis- sions or errors. In figure 4, the earthquakes in the chronological historical listing that have instrumentally deter- mined epicenters are plotted as triangles, and the larger or better described earlier earthquakes are plotted at their maximum damage zone as squares. A notable recent earthquake series occurred near Mazatenango. The October 23, 1950, earthquake (MS=7.1) was followed, within 2 weeks, by six aftershocks of Msé6.0. The pattern of historical seismicity in Guatemala is similar to that based on reliable epicenter estimates, such as shown in figure 3A. Notably, there is a general lack of major his- GUATEMALAN EARTHQUAKE OF FEBRUARY 4, 1976, A PRELIMINARY REPORT . 833 0.980 was a395%“95cm £28m 52m 30332 9539» mix #53 and mg 343 .movwvmw uamwaoea 0:» mafia "Emu hzgwfimmwm mg 335 .3 au—uswnfidw 2.3 25 mo econ 3°:qu 25. .m.x he, 525m was 654 S 5.32 SMASH“ .mxoonmawfid “£5: 2:: one and «:95 5.58 25. .32: ..§EEoo 5323 .859 .0 .4I<3;<.—m.\ m523». x D < D 4 us A 2 «5 w I wo‘m 06 V .2 o<£ 05m: :3: g on; Vo . Ihmuo 45.0) best located aftershocks of the main event, occurring during the first week, is about 250 km long (Person and others, this report). This zone of the largest aftershocks coincides very closely with the zone of surface fault- ing mapped after the earthquake (Plafker and others, this report). Possibly an additional 50 km of fault rupture could be postulated on the basis of small aftershocks recorded after the earthquake (Langer and others, this report) and from high damage west of Guatemala City (Espinosa and others, this report). For the purpose of the analysis that follows, we shall take the fault length as 300 km. The relationship between seismic moment (Mo), fault length (L), width (w), and average displace- ment (D) is (Aki, 1966) M0: “L105, (1) MAIN EVENT SOURCE PARAMETERS FROM TELESEISMIC DATA 21 O|335 cm-s v FIGURE 13.—Azimuthal variation of the displacement spectral . density normalized to an epicentral distance of 90°. The dashed lines are theoretical nodal lines of the Love-wave radiation pattern ‘for a vertical strike-slip fault, strik— ing N. 65° E. Where ,u. is the rigidity of the faulted medium, here taken to be 3X1011 dynes cm”. The average dis- placement, 5, is inferred from geologic observation to be 100 cm (Plafker and others, this report). Together with our estimate of seismic moment, a fault length of 300 km and an average displacement of 100 cm implies w=29 km. This fault width is apparently greater than those of earthquakes on California’s San Andreas fault; the seismogenic fault width associated with the Cali- fornia earthquake of 1906, for example, is thought to be 10 km (Thatcher, 1975). If we had chosen Mo equal to the mean rather than the median of the moments observed at indi- vidual stations, if we had taken L=250 km rather than L=300 km, and if we had taken 5 less than 100 cm, the discrepancy between the fault width of ' the Guatemala earthquake and that of the 1906 Cali- fornia earthquake would be even greater. There are alternative explanations for this discrepancy: 1. Displacement on the fault at depth may be larger than surface-fault displacement, so that 100 cm would be significantly smaller than the actual E. The difficulty with this explana- tion is that the displacements observed at the surface are quite uniform over long distances (Plafker and others, this report); it is hard to visualize a process acting uniformly over 100 or more kilometres that would retard sur- face-fault slippage relative to slippage at depth. In fact, one might make the contrary argument—that seismic-fault displacement on a long strike-slip fault will have a tendency to decrease with depth from the free surface. 2. Seismic rupture on the Motagua fault may ac- tually have extended to several tens of kilo- metres in depth. Such rupture would have to produce a large amount of long-period energy in order to significantly affect amplitudes of 100-s G-waves. However, the fault rupture at depth need not necessarily have produced a large amount of short-period energy. Like- wise, the shallow depths of aftershocks re- corded by Langer, Whitcomb, and Aburto (this report) do not preclude fault rupture extending several tens of kilometres into the crust, since such rupture could occur com- paratively slowly in a medium that is incapable of producing high-frequency strikeslip earth- quakes. The stress drop, Aa, for the main event may be estimated from . 2 E Aa=—,. <—>(Knopofl’, 1958). (2) 71' w With ,u=3>< 1011 dynes cm—Z, l—)= 100 cm, and w=29 km, Ao= 6.6 bars. A stress drop of 6.6 bars is less than the world- wide average for interplate earthquakes; Kanamori and Anderson (1975) find that 30 bars is typical for such earthquakes. If, as discussed above, w were less than 29 km, the stress drop would be corres- pondingly larger. The epicenter of the main event lay about 90 km from the eastern end of the inferred 300-km-long zone of fault rupture, a position that suggests that the fault rupture propagated from northeast to southwest. The level of shaking near the western end of the fault might be expected, under such circumstances, to be higher than the level of shak- ing at the eastern end of the fault because of con- structive interference of waves from the propagat- ing source. The mantle-wave observations tend to support such a conclusion, the amplitudes being larger for waves leaving the event to the southwest 22 GUATEMALAN EARTHQUAKE OF FEBRUARY 4, 1976, A PRELIMINARY REPORT PTO GZ PTO G3WWWM/VWM G4 l—_—J IOOs FIGURE 14.—Efl"ect of source propagation on G—wave ampli- tudes at station PTO. G2 and G4 left the source at an angle of about 15° from the direction of fault rupture. G3 left the source at an angle of 165° from the direction of fault rupture. Note that the amplitude of G4 is at least equal to that of G3, although G4 has traveled 213.6° farther than G3. than for those leaving toward the northeast. Figure 14 shows a striking case of this phenomenon in the records from Porto, Portugal. The high intensities observed near the western end of the Motagua fault (Espinosa and others, this report) may thus be, in part, an effect of source propagation. FUTURE EARTHQUAKES ON THE MOTAGUA FAULT SYSTEM At the time this paper was written (April 1976) , Guatemala had experienced several moderate (5:.}:s?25 * . / Lo: '2 Amons/ FEB IO, \976 50‘) HONDURAS $35Mala9€11 § F6535 i), I97)6 R t; . MAG _,5°00' Q‘ Chichicastcnonqo AGU A 0. Zacapu J 6", 89" “OT Lo Fruouo I V. El Proqvaso (\ “A 117/7111/1/ \ z’ k ‘ G 7 I Z/J'Mk/i “ME §¢7 ugieya u 7 * Main Event Epicenter " & A} fl Aftershock Epicenter: FEB 6,1976 , . . 5.75 (MAG) I o S'ronrMotIon Slllmoqraph _.,. Y \ |4.6,90.4 — Faun f I , S“ .1th ‘7‘ r, ”$5 ,/ \ .-\_1 EL SALVADOR WOO. 0 mo KI L l l L I J l l I l L 91°00' 90°00‘ 89°00' 88 00' w FIGURE 15.—Strong-motion field network, surface fault breakage (Plafker and others, this report), and epicenters of main event and aftershocks. Date, magnitude, and location are given for each (Person and others, this report). had been dislodged. The ground motion on this rec- ord exceeded the maximum radius of the plate be- fore it was dislodged. The trace on the seismoscope plate of the ground floor (fig. 17) had a maximum relative displacement of Sd=5.3 cm for T1=0.78 s, and a 10-percent damping. Two sections of the maximum excursions of the recorded motion on the ground-floor seismoscope were analyzed in order to recover the levels of ground accelerations. Part of the trace between the two analyzed sections could not be followed, but it is certain that section one was first real time. The following constants were used in the analysis of the deconvolution of the seismoscope plate: T1=0.78 s (natural period of seismoscope), T2=0.055 s (sec- ond harmonic of seismoscope), and S=5.8 cm/rad, (sensitivity of seismoscope). The T2 and S values are average determinations obtained from similar seismoscopes, and the other constants were obtained in the field calibration of these instruments. The results of the above analysis are shown in figures 18 and 19. These two figures show the ac- celeration as a function of time for the north and east direction of motion. The first section was fol- lowed for 2 s, and the second section for 5 s. Maxi- mum accelerations as shown are about 200 gals. A 600-gal acceleration appeared on the north direc- tional component at approximately 1 s after the be- ginning of the second section. Unfortunately, there is insufficient recoverable trace length, and hence the data available do not warrant a spectral-analy- sis evaluation, since the time window is very short. 26 GUATEMALAN EARTHQUAKE OF FEBRUARY 4, 1976, A PRELIMINARY REPORT FIGURE 16.—Seismoscope plate of main event located on a rooftop instrument at the administration building of the Univer- sidad de San Carlos, Guatemala City. Instrument was about 30 km south of Motagua fault surface breakage. Arrow indicates north. The plate is scratched all over. Recording of main event is shown in the middle part of plate and to the sides before dislodging. STRONG-MOTION RECORDINGS OF MAIN EVENT AND FEBRUARY 18 AFTERSHOCK 27 FIGURE 17.——Seismoscope plate of main event located on a ground-floor instrument at the administration building of the Universidad de San Carlos, Guatemala City. Instrument was about 30 km south of Motag'ua fault surface break- age. Arrow indicates north. The plate is scratched all over. Recording of main event is shown in middle part of plate. 28 A B 400 . NORTH ’ EAST 200 2; a a! r ' 5 0 A A A \ e \NUV \ a) a , _ 8 <( —200 VA“ _400 ......................... o 1 2 o 1 2 Time (s) A FIGURE 18.—Deconvolved 2-5 section of seismoscope of the main event. Acceleration as a function of time. A, North component; B, east component. FIGURE 19.—Deconvolved 5-s section of seismoscope of the) main event. Acceleration as a function of time. These 5 s of recordings are later than those shown in figure 18. A, North component; B, east component. FIGURE 20.—Accelerogram of the February 18, 1976, after- shock recorded in Guatemala City by SMA—l, No. 1926, at station IBM building, lat 14.64° N., long 90.51° W. (Sens.=sensitivity; Per.=period; crit.=critical.) Acceleration (gals) GUATEMALAN EARTHQUAKE OF FEBRUARY 4, 1976, A PRELIMINARY REPORT 600 A NORTH 400 200 - 200 -4 00 ‘ ‘ ‘ 600 EAST 400 200 ‘ 01/1 An 1% .t\ . WV - 200 -400 . . I I ”I 0 1 2 3 4 5 6 Time (s) I. SOUTH Sens. = L73 cm/q Per. .038 s :1 __________________ ... _________________ .. ____________________________________________ Damp = 0.60 crit J DOWN ‘l . fl 2 __ Sens. = I.83 cm/g WWW Per. = .038 s I'M? W Damp. = 0.56 crit J EAST ...M Sens. = L76 cm/g Per. = .038 s ._ _ _ ______ Damp = 0.60 crit STRONG-MOTION RECORDINGS OF MAIN EVENT AND FEBRUARY 18 AFTERSHOCK 29 STRONG-MOTION ACCELEROGRAM OF FEBRUARY 18 AFTERSHOCK Several aftershocks have been recorded on the ac- celerographs located at Zacapa, the Observatorio Nacional, and at the IBM building stations. These aftershocks have a Richter magnitude less than 4.0 and are not considered significant. An aftershock was recorded at the IBM building on February 18, 1976, at 03:59 local time. The accelerogram in figure 20 shows a vertical motion of approximately 0.1 9 during the first second. SAN SALVADOR ACCELEROGRAPH OF MAIN EVENT The Observatorio de San Salvador reported that an AR—240 located in San Salvador at the Biblioteca station was triggered by the Guatemala earthquake. The readings obtained from the accelerograms were 0.066 g on the north component; 0.25 g on the ver- tical and 0.053 g on the east component (M. Mar- tinez, oral commun., 1976). The Modified Mercalli intensity rating in San Salvador was V (M. Mar- tinez, oral commun., 1976). THE GUATEMALAN EARTHQUAKE OF FEBRUARY 4, 1976, A PRELIMINARY REPORT AFTERSHOCKS FROM LOCAL DATA By CHARLEY J. LANCER, JEAN P. WHITCOMB, and ARTURO ABURTO Q. INTRODUCTION Aftershocks of the main event were monitored in two phases by single-component portable seismo- graphs from February 9 to February 27. This study represents a combined effort by the US. Geological Survey and the Nicaraguan Instituto de Investi- gaciones Sismicas. Rapid deployment of portable in- strumentation around the Motagua fault zone pro- vides a data base for the first detailed aftershock investigation of a major earthquake (magnitude greater than 7.5) in Central America. Tectonic and seismic aspects of the main event and large after- shocks are discussed in other sections of the report (Spence and others, Person and others). The topic addressed here is hypocentral locations of a repre- sentative sample of locally recorded aftershocks and their relationship to primary and secondary fault- mg. INSTRUMENTATION AND FIELD PROCEDURE Aftershocks were recorded by portable, smoked— paper seismographs, each consisting of a vertical transducer, a high-gain amplifier, and a crystal- controlled clock. The seismograph recorded at a speed of 60 mm/min, and the trace separation was 1 mm, which allowed 48 hours of continuous opera- tion. Precise time corrections were determined with an oscilloscope by comparing WWV radio time with recorder clock times during record changes. Clock drift did not exceed 20 ms/ day. Seismograph magni- fications generally ranged between 50,000 and 100,- 000 at 10 Hz. Amplifier gains were limited by the background noise at the sites, most of which were on unconsolidated soils and close to cultural noise sources. Because of the intense aftershock activity at many of the station locations, the peak-to-peak deflection of the recorder pen was limited to 10 mm. A two-phase aftershock recording program was required because of the great length of fault rupture (more than 240 km), constraints imposed by the available logistical support, and the limited amount of seismograph equipment available. The phase I, 30 or western, network (table 5) was installed on February 9 and 10 and extended approximately 95 km east-west between Sanarate and Chichicaste- nango. Another portable seismograph was installed in Guatemala City after the main event by person- nel at the Observatorio Nacional. This network sur- rounded the western end of the Motagua fault zone and also encompassed many of the northeast—trend- ing secondary faults in the vicinity of Guatemala City, Chimaltenango, and Tecpan. The phase I oper- ation was terminated on February 17 when all seis- mographs, except those in Guatemala City, were re- moved. During phase II, a much broader seismograph network was installed to the east between Guate- mala City and Puerto Barrios (table 5). It covered about 225 km of the central and eastern segments of the Motagua fault and adjacent regions. On February 18, 19, and 20, seismographs were located at eight sites (table 5). The Puerto Barrios station was relocated at a site near La Pifia on February 21 because of the high cultural background noise at Puerto Barrios. Phase II was completed on Feb- ruary 27 when all the instruments were retrieved. DATA AND ANALYSIS Several thousand aftershocks were recorded dur- ing the field investigation (fig. 21). The amount of seismic activity was greatest at the Western end, near Tecpan and Chimaltenango, and did not notice- ably diminish during the 8-day monitoring period of the western network. The unusually high level of observed seismicity in this area is not merely a func- tion of station location or of time, that is, early in the aftershock sequence; the Tecpan-Chimaltenango region is unique to the total aftershock zone in terms of level of seismicity. Arrival times were determined by using a low- power magnifier and were corrected for variations in distance between minute marks. S-phases were easily identifiable in many cases, often at two or AFTERSHOCKS FROM LOCAL DATA 31 TABLE 5.——List of seismograph stations occupied during this study Lati- Longi Eleva- tude tude tion Period of Name Symbol (°N.) (°W.) (metres) operation Western network Chichicastenango _____ CCO 14.950 91.110 1,990 Feb. 9—Feb. 17 Tecpan ______________ TEC 14.766 90.996 2,320 Feb. 9-Feb. 17 Joyabaj _____________ JOY 14.990 90.804 1,400 Feb. 9—Feb. 17 Chimaltenango _______ CHM 14.635 90.818 1,760 Feb. 9—Feb. 17 E1 Chol _____________ ELC 14.958 90.487 995 Feb. 9——Feb. 17 Guatemala City ______ GCG 14.586 90.533 1,497 Feb. 9—present Palencia _____________ PAL 14.664 90.361 1,310 Feb. 10—Feb. 17 Sanarate ____________ SAN 14.784 90.196 770 Feb. 10—Feb. 17 Eastern network Guatemala City ______ GCG 14.586 90.533 1,497 Feb. 6—present San Jeronimo ________ SJE 15.065 90.247 1,005 Feb. 18—Feb. 27 Jalapa ______________ JAP 14.638 90.003 1,370 Feb. 18—Feb. 27 Teleman _____________ TEL 15.339 89.744 65 Feb. 19—Feb. 27 Chiquimula __________ CML 14.801 89.533 360 Feb. 18—Feb. 27 Quirig’ué ____________ ARC 15.273 89.039 70 Feb. 18—Feb. 27 La Esmeralda _______ RIO 15.656 88.994 10 Feb. 20—Feb. 27 Vitalis ______________ VIT 15.312 88.806 120 Feb. 18—Feb. 27 La Pifia _____________ FFF 15.600 88.608 40 Feb. 23-Feb. 27 Puerto Barrios _______ PTO 15.712 88.583 40 Feb. 20—Feb. 22 more stations for the same earthquake. Accuracy of most P-wave times is thought to be within $0.1 s; the selected S-wave readings are believed ac- curate to $0.20 s. Seventy-eight hypocenters (table 6), most of which lie inside or very near to the margins of the temporary seismic networks, were determined by the HYPO71 computer program (Lee and Lahr, 1975). A measure of their solution quality is de— noted by the symbol SQ and ranges between B (good) and D (poor). This SQ rating is dependent upon the number and accuracy of data, station dis- tribution, and crustal velocities. All D-quality solu- tions are a few kilometres outside the network; otherwise they would be rated as B or C. The average root-mean-square (RMS) errors of the travel-time residuals are 0.17 s, which implies that the random errors in reading the P- and S- ar- rivals account for most of the RMS errors. An aver- age of the standard errors indicates hypocentral ac- curacies of about $1.3 km in the horizontal plane and approximately $2 km in the vertical plane. Although the standard errors may not represent actual error limits, particularly for hypocenters outside the seismograph net, S-phase data mitigate the possibility of gross mislocations. Any systematic location error or bias is most likely caused by the six-layer Managua velocity model of Brown, Ward, and Plafker (1973) used in the HYPO71 program. This model was employed in this study because of the absence of velocity data for interior Guatemala. Although the model is an assumed velocity structure for the Managua area, it is representative of vol- canic terrane and therefore may be generally appli- cable to the Motagua fault zone west of long 90.5° W. To the east, where crystalline and marine sedi- mentary rocks are predominant (Bonis and others, 1970), increased velocities would be expected in the upper layers. The Managua model, however, is con- sidered adequate for obtaining preliminary loca- tions. Because the peak-to-peak signal amplitudes were electronically clipped, local magnitudes, ML, are esti- mated from the aftershock coda lengths (Lee and others, 1972). The lower magnitude threshold for hypocentral determinations using either the western or eastern network data is about 2.2. None of the larger aftershocks reported by Person, Spence, and Dewey (this report) occurred within a temporary seismograph net. The largest located event (magni- tude 3.8) is approximately one order of magnitude below the limit for teleseismically locatable earth- quakes in Central America. RESULTS AND DISCUSSION Aftershock epicenters are distributed along the Motagua Valley from the lowlands near the Gulf of Honduras westward to the Guatemalan highlands northeast of Lake Atitlan, a distance of some 300 km. A large number of located events occurred on secondary faults south of the Motagua fault and west of long 90.3° W. (fig. 22). Focal depths range from near surface to about 14 km. In particular, we note the following aspects: 1. The eastern terminus of the causal fault rupture is most likely defined by the cluster of 12 epi- centers southeast of Puerto Barrios. The gen- GUATEMALAN EARTHQUAKE OF FEBRUARY 4, 1976, A PRELIMINARY REPORT 32 .hummmoa 23 E m_ was 533m .NN 3st E :39? .8532 553m 623 .bmlmm mwwgnom How ESMoEmEm E; 533m ”m .33 .NHIoH ban—Sch 78m Ewawofimmwm OMB 533w .v. .mcuooon fifiwio no fish? SE ow and 8258 3:55 .93 .58 N mm 5333me 82B .33 “Avosgwfiw 23 E 3353“ we 355:8 MEBonm mfiauwofiwmmmlém 55lo ~ opjgir mEOmp< 0H3 ENTIO 9:; ©h®rmr9®m NI Of _ I NI m “04 mgoti QU©©L®EEE< mEONp< UHDmVNNMcO @EF ©h©i0rnmu UMP 33 AFTERSHOCKS FROM LOCAL DATA 1. ,. , .e. ifiwsvévgniE anur_: ELmloJ mL@P:L .EHEECSOIAN 55lo @9353: Qu©©LQEQE< 2c0©p< utermi:O@c:b enoergwu mEOmpfl 0H3 OmOchmEF ©.\|®_Xmmnoh_ H_> 34 GUATEMALAN EARTHQUAKE OF FEBRUARY 4, 1976, A PRELIMINARY REPORT TABLE 6.—Aftershocks of the main event located by temporary seisrrwgraph network [1No. sta.--refers to the number of stations used to obtain hypocentral solutions;.ZDMIN--distance to the closest seismograph station; 3RMS—-root mean square errors of travel time residuals; “Standard errors--refers to the indices of precision relating to the values and distribution of the unknown errors in the hypocentral solution where DLAT = error in latitude, DLON = error in longitude, and D2 = error in depth; 5SQ—-a measure that is intended to indicate the general reliability of the hypocentral solution where A = excellent epicenter, good focal depth; B = good epicenter, fair focal depth; C = fair epicenter, poor focal depth; D = poor epicenter, poor focal depth; 6ML-—local magnitude of shock.] Western network Standard errors“ Date Origin Lat. N Long. W Depth No.l DMIN2 RMS3 DLAT DLON DZ (Feb. 1976) (UTC) (deg) (deg) (km) sta. (km) (sec) (km) (km) (km) SQ5 ML5 11 0636 48.69 14.750 91.126 9.2 8 14 0.18 1.1 1.3 1 5 C 3.2 11 0932 20.37 14.808 90.510 10.0 7 17 0.14 0.9 0.6 2 1 B 3.1 11 0953 02.55 14.760 90.980 7.0 7 2 0.11 0.6 0.8 1 2 B 2.9 11 1044 34.78 14.763 91.024 10.3 7 3 0.09 0.6 0 7 0 5 C 2.8 11 1051 03.91 14.767 90.975 4.2 7 2 0.19 1.5 1.5 1 5 B 2.9 11 1142 03.72 14.724 90.998 4.0 8 5 0.22 1.3 1 3 1.4 C 2.9 11 1636 23.01 14.790 90.984 4.9 7 3 0.04 0.3 0 3 0.4 B 3.0 11 2210 58.78 14.637 90.680 4.0 10 6 0.16 1.7 0.4 1 8 C 3.3 11 2253 51.55 14.809 90.596 2.0 7 20 0.24 0.4 0.3 1 2 C 2.4 12 0039 13.82 14.759 90.501 2.0 6 18 0.25 0.6 0.5 1 9 C 2.7 12 0144 35.11 14.861 90.343 12.2 7 18 0.20 1.9 1 7 2.4 C 2.6 12 0215 04.82 14.694 90.468 6.0 8 12 0.11 0.7 0.4 2 2 C 3.3 12 0333 36.40 14.855 90.710 12.0 9 18 0.21 0.9 0.7 1 6 B 2.7 12 0408 15.10 14.745 90.499 13.0 8 17 0.22 0.9 0.8 2 4 B 2.5 12 0439 54.52 14.636 90.668 12.1 7 16 0.11 1.1 0.4 l 4 C 3.0 12 0545 45.58 14.827 90.513 13.0 8 15 0.19 0.9 0.6 1 9 B 2.8 12 0702 02.43 14.859 90.340 10.5 8 14 0.11 0.8 0.5 l 3 B 3.3 12 0743 42.34 14.800 90.544 12.3 7 18 0.19 1.1 1 0 2.7 B — 12 0744 35.68 14.852 90.619 12.2 9 18 0.25 1.0 O 7 4.3 B 3.0 12 1057 35.51 14.714 90.796 11.7 8 9 0.10 0.7 0.4 l 2 B 3.2 12 1203 33.49 14.589 90.625 13.2 7 21 0.09 1.0 0.4 1 0 C 3.2 12 1927 36.00 14.590 91.037 10.0 8 9 0.13 0.8 1.6 1 2 C 3.4 12 2211 59.01 14.674 90.482 12.0 7 13 0.20 0.9 1 0 2.4 C 2.4 12 2250 34.00 14.760 90.355 5.4 7 11 0.28 2.6 l 4 4.8 C 2.9 13 0627 42.32 14.673 90.483 10.0 7 13 0.16 0.7 0.8 5 3 C 3.1 13 0701 32.46 14.684 90.477 11.9 8 13 0.18 0.9 0.7 2 2 C 3.2 13 1344 01.31 14.755 90.987 5.7 8 1 0.16 1.0 2.3 1 7 C 3.2 13 2359 50.50 14.767 91.025 1.1 8 3 0.20 1.5 1 2 1.5 C 3.3 TABLE 6.——-Aftershocks of the main event located by temporary seismogmph network—Continued Western network Standard errors“ Date Origin Lat. N Long. w Depth No.1 DMIN2 R1153 DLAT DLON DZ (Feb. 1976) (UTC) (deg) (deg) (km) Sta. (km) (sec) (km) (km) (km) SQS ML 14 0300 40978 14.858 90.636 11.7 9 20 0.10 0.4 0.3 1.2 B 3. 14 0315 59.79 14.696 90.545 10.0 8 20 0.20 0.9 0.8 2.8 c 3. 14 0424 53.89 14.831 90.319 9.0 7 14 0.25 1.2 1.5 3.7 c 2. 14 0916 38.15 14.711 90.737 10.7 9 12 0.17 1.0 0.7 2.1 B 3. 14 1543 57.80 14.699 90.481 12.0 8 13 0.19 0.6 0.5 1.5 c 2. 14 1757 33.16 14.700 90.514 12.4 8 17 0.32 1.6 1.6 3.7 c 3. 14 1842 41.94 14.754 90.312 10.0 7 11 0.29 0.6 0.7 2.8 B 2. 14 1912 53.22 14.643 90.950 4.0 8 14 0.14 1.2 0.8 0.8 c 3. 14 2036 28.16 14.815 90.583 10.6 8 19 0.14 0.7 0.6 1.9 B 3. 14 2044 04.68 14.743 90.377 6.0 8 9 0.24 0.7 0.8 3.5 B 3. 14 2122 55.03 14.740 91.007 11.4 8 3 0.10 1.1 1.0 0.5 c 3. 14 2219 24.40 14.746 90.355 8.0 7 9 0.24 0.5 0.6 1.6 B 2. 14 2318 26.40 14.741 90.323 5.0 6 9 0.21 1.4 0.5 2.0 B 2. 15 0034 43.75 14.776 90.965 6.2 9 4 0.15 0.6 0.6 1.6 B 3. 15 0436 12.15 14.808 90.551 2.5 10 18 0.27 0.6 0.5 1.1 c 3. 15 0650 31.18 14.728 90.359 2.5 8 7 0.24 0.8 0.9 2.7 B 2. 15 1053 24.11 14.720 90.748 10.0 8 12 0.25 0.5 0.5 3.3 B 3. 15 1308 31.57 14.782 90.980 6.4 9 2 0.10 0.6 0.5 0.9 B 3. 15 2019 59.93 14.792 90.982 3.8 6 3 0.25 2.4 2.3 4.0 c 3. 16 0758 08.62 14.848 90.678 12.2 11 21 0.16 0.7 0.4 1.1 B 2. 16 0911 46.82 14.750 90.998 10.0 7 2 0.23 1.8 1.9 1.2 D 3. 17 0345 47.31 14.708 91.008 7.8 6 6 0.04 0.5 0.5 0.5 c 2. 17 0527 05.94 14.723 90.801 11.8 10 10 0.20 0.8 0.6 2.2 B 2. 17 1549 25.34 14.791 90.974 2.4 6 4 0.09 1.2 0.8 1.4 B 2. 20 0321 50.53 15.152 89.228 1.3 8 24 0.16 0.9 0.7 1.1 c 2. 21 0205 36.04 15.052 89.452 5.9 7 29 0.18 1.0 1.0 4.0 c 3. 21 0752 07.81 14.991 89.627 10.4 8 23 0.20 0.9 0.7 2.0 c 3. 21 1303 52.99 14.971 89.676 11.2 6 24 0.08 0.5 0.5 1.2 B 2. 22 0500 33.55 15.671 88.445 10.0 5 16 0.16 0.8 2.3 1.6 c 2. 22 0642 40.71 15.526 88.520 8.5 6 22 0.07 0.7 1.5 5.5 D 2. 22 1209 58.28 15.275 89.007 10.0 6 3 0.16 2.8 1.0 1.8 0 3. 22 2138 32.95 15.217 89.003 14.0 7 7 0.16 1.5 0.8 0.9 c 3. 23 0503 36.68 15.314 88.906 8.9 6 11 0.16 1.3 0.9 1.6 c 3. 24 0417 00.95 15.670 88.437 10.0 5 20 0.21 0.6 1.4 1.1 c 2. 24 0737 18.11 15.556 88.519 5.1 6 11 0.10 2.2 2.1 2.8 D 2. 36 GUATEMALAN EARTHQUAKE OF FEBRUARY 4, 1976, A PRELIMINARY REPORT TABLE 6.—Afte'rshocks of the main event located by tempera/r1; seismogmph network—Continued Eastern network Standard errors "‘ Date Origin Lat. N Long. w Depth No.1 DMIN2 _ RMs3 DLAT DLON DZ (Feb. 1976) (UTC) (deg) (deg) (km) sta. (km) (sec) (km) (km) (km) SQ5 M 6 24 0807 06.51 14.983 89.635 3.5 8 23 0.11 0.3 0.4 1.6 B 2.7 24 0821 49.24 15.660 88.438 1.6 6 19 0.13 0.6 0.5 0.5 c 3.2 24 1316 05.14 15.485 88.601 10.0 5 13 0.20 0.6 0.9 2.6 c 2.5 24 1337 59.96 15.496 88.599 12.4 6 12 0.27 2.6 2.5 4.2 0 3.0 25 0128 48.30 14.977 89.674 5.9 7 25 0.13 0.3 0.4 1.2 B 3.3 26 0033 22.94 14.964 89.690 7.6 8 25 0.05 0.2 0.2 1.0 B 2.7 26 0510 13.27 15.617 88.437 7.0 5 18 0.19 1.0 0.8 1.1 c 2.4 26 1120 06.00 14.841 89.641 8.6 7 13 0.14 1.1 0.6 2.0 B 2.9 26 1903 20.29 15.561 88.515 8.0 6 11 0.21 0.5 1.1 2.1 c 2.3 26 2216 11.70 14.972 89.612 8.0 8 21 0.17 0.5 0.7 2.0 c 2.8 27 0120 58.22 15.580 88.451 9.0 5 17 0.05 0.9 1.0 1.0 c 3.1 27 0344 29.49 14.972 89.662 10.0 8 23 0.28 0.5 0.7 1.8 c 2.7 27 0458 00.86 15.537 88.565 2.8 5 8 0.14 0.4 3.3 2.8 0 2.5 27 1200 38.49 15.602 88.621 5.7 7 1 0.20 2.0 4.0 2.8 c 2.4 eral trend of the southern group of eight aftershocks is in line with the inferred exten- sion of the Motagua fault (Plafker and others, this report), whereas the four epicenters slightly to the north may be associated with induced movement at the eastern end of the San Agustin fault. 2. Epicenters associated with the western end of the Motagua fault do not extend beyond the mapped fault breakage. Consequently, with the data at hand, the aftershock pattern does not suggest a more precise limit to the pri- mary fault rupture than the obvious diminu- tion of seismicity west of long 90.45° W. Also, there are no located aftershocks that appear to be related to induced movement on the western segment of the San Agustin fault. 3. The distribution of energy release along the Motagua fault proper is roughly uniform, with exception of the concentration of activity west of Zacapa. The group of seven epicenters be- tween long 89.6° W. and 89.7° W. may be a result of fracturing east of Where the Motagua fault bends from a general east-west direc- tion to a northeasterly direction. Three north- east—trending secondary faults (not shown in fig. 22), which cut Paleozoic metamorphic rocks, are mapped in this area (Bonis and others, 1970). 4. The majority of aftershocks located west of long 90.3° W. are directly associated With sec- ondary faulting. Four groups are considered to be of principal interest: a. Tecpan (long 91° W., lat 14.75° N.). The high level of activity observed at the Tecpan seismic station (fig. 21) is re- flected by the dense cluster of epicenters located in this area. Plafker, Bonilla, and Bonis (this report) have defined a lineament that projects through Tecpan and the center of the northeasterly trending concentration of aftershocks. Therefore, on the basis of the epicentral locations, the lineament can be inter- preted as a northeast-striking fault. b. Chimaltenango. Four epicenters occurring in the vicinity of a northeast-striking lineament that runs through Chimalte- nango lend support to the existence of a secondary fault. 0. Guatemala City region. These aftershocks are very likely associated with faults AFTERSHOCKS FROM LOCAL DATA 37 . 9mov new ‘3 x ~ 88°00' 1' cf' i fl H... J, |5°00' > mwa‘»., l EXPLANATION Main event epicenter o Aftershock epicenters (teleseismicolly located) I Aftershock epicenters(regiondlly located) A Portable seismograph Main event causal fault. Arrows show sense of horizontal slip,- borbs on downthrown side; dotted where inferred Fault with known or suspected late Cenozoic displacement Lineoment possibly indicative of foul? V ' SG'W 50 M | LES CAN/555M! SEA 50 Kl LOMETRES l l PACIFIC OCEAN lNDEX}MAP FIGURE 2‘2.—Aftershock epicenters and portable seismograph locations (geology from Plafker and others, this report). See table 5 for station names and their geographical locations and table 6 for aftershock-location parameters. Station code is in the Glossary. (Base map modified from Guatemala, Instituto Geografico Nacional, 1974, 1:500,000.) forming the Guatemala City graben. The Mixco fault, west of the city, rup- tured the ground surface. Some epi- centers appear to correlate with the Mixco fault and also with the northerly extension of the mapped fault bounding Guatemala City on the southeast. d. Agua Caliente (long 90.35° W., lat 14.75° N). A group of epicenters 15 km north of Palencia (station PAL) surround the Agua Caliente Bridge site. Secondary faulting, although not mapped at this locale, is certainly indicated by the after- shock cluster and may have contributed, in part, to the collapse of the bridge. 5. The preponderance of aftershocks lying off the Motagua fault west of long 90.3° W. suggests that induced motion along secondary faults is rare east of long 90.3° W. 6. There is an apparent southerly bias of epi- central locations along the Motagua fault prop- er. The spatial distribution of aftershocks thought to be associated with the primary fault indicates a systematic offset of 2 to 3 km. This offset would suggest that (1) the Motagua fault is dipping steeply to the south in accordance with the main—event focal mech- anism of Dewey and Julian (this report) or (2) there is a large contrast in seismic veloc- ities across the fault similar to that observed by Eaton, O’ Neill, and Murdock (1970) on the San Andreas rift zone near Parkfield, Cali- fornia. THE GUATEMALAN EARTHQUAKE OF FEBRUARY 4, 1976, A PRELIMINARY REPORT GEOLOGIC EFFECTS By GEORGE PLAFKER, MANUEL G. BONILLA, and SAMUEL B. BOle 1 INTRODUCTION This report is based on preliminary field studies of the geologic effects of the earthquake made dur- ing an 11-day period from February 5 to 16, 1976. During this period, we examined the main and secondary faults on the ground, using vehicles and helicopters for logistic support. In addition, several reconnaissance flights were made with fixed-wing aircraft over most of the area that was strongly affected by the earthquake to identify and map landslides, liquefaction phenomena, and damage to communities near the surface faults. Some of our interpretations may be modified by the more de- tailed field investigations that were being conducted by the US. Geological Survey and other organiza- tions at the time this paper was written (March— April). THE MAIN FAULT The main fault along which the destructive main event (Ms=7.5) occurred was identified for 240 km in the Motagua Valley and the mountainous area west of the valley 2 (fig. 23). This fault is of special interest because it is the most extensive surface rupture in the Northern Hemisphere since the 1906 San Francisco earthquake. Identification of this fault permits evaluation of damage relative to the earthquake source and provides critical new infor- mation on the present mode of deformation to a ma- jor tectonic belt of Central America. The eastern part of this major fault, within the Motagua Valley, has been named the Motagua fault (Dengo and Bohnenberger, 1969; Instituto Geografico Nacional, Chiquimula 1:250,000 sheet, 1969), and this name is herein applied to all the fault that slipped during the earthquake. Ground breakage was observed in a discontinuous line extending 240 km from near Quebradas in the lower Motagua Valley on the east to about 10 km 38 east of Patzaj on the west.2 At the closest point, the fault is 25 km north of the center of Guatemala City. The fault could not be identified farther to the west because the area is characterized by young volcanic deposits and rugged terrane in which numerous earthquake-triggered slope failures effectively mask the fault-related surface fractures. At the eastern end, the fault trace is obscured in the lower Motagua Valley by swamps and dense tropical vegetation. However, the occurrence of aftershocks southeast of Puerto Barrios (Langer and others, this report) near the eastern coast sug- gests that the faulting probably extends at least that far. If so, the main break is on the order of 300 km long2. Some of the characteristics of the fault- ing at localities where it was studied on the ground are given in table 7. The fault trace is a well-defined linear zone with a gradual change in average strike from N. 65° E. at the eastern end to N. 80° W. at the western end. It consists of right-stepping en echelon fractures and connecting low compressional ridges that 10- cally form the “mole tracks” that are characteristic of strike-slip faults (figs. 24—28). Individual frac- tures within the zone may be as much as 10 m long; most are tightly closed, but some have spread as much as 10 cm. The fractures are oriented at angles of as much as 35° to the fault trace and have the northeasterly azimuths that are to be expected for sinistral slip. The width of the fracture zone is mostly 1 to 3 m, with a maximum observed width of about 9 m. At one locality near E1 Progreso, Where the fault surface is exposed in a highway cut, the zone of slip is 1 to 3 m wide, and the dip is es- sentially vertical. Displacement across the fault in most places is al- most entirely horizontal and sinistral. The strike- 1 Instituto Geogréfico Nacional de Guatemala. 2Later, more detailed studies indicate that the length of surface fault- ing is 230 km and that the main break from relocated epicenter data is 270‘ km. GEOLOGIC EFFECTS 90°00V 59°00V 39 89°00’ 15°00“ >9 f % K H 1‘ . A, 2» [““Exfl GUATEMALA / L ‘2, 3 i A X w (”J flown P \ - OLOCHI . FAULT N/‘"*\cr’n_‘ ? 5‘ flf~ ‘ Kw¥\_e.~_\m// \ “ sat!" H SALAMK‘KQK“ /r~\ wings. _ “JOYABAJ EXPLANATION /' 7-" '7 ”’2‘” 0 ,3“ _‘ 17/251195! Main event epicenter / CNUA RANC‘OHO ,' f . i . ”7/; .ri: ’ . Aftershock epicenters , r/C ,‘ ‘— n k .1"; V \f ——-m- if OCHMLTE’f‘RBGUNEfiLA 3 ,1 “3“” l l976 earthquake faults ' puma» C'TYj ’ ’1 Arrows show sense of horizontal slipi Jane f barbs on downthrown side; (it {41%* 0 GPABLEIi‘L / ‘~. dotted where interred ( 3 w a» .3.»/ m—.... 5 Fault with known or suspected i ' " JG’ESCU'M”/’ @CU'EAPA *JUHPWZk W‘ [alt $3" late Cenozoic di5placement \ 3/ K ‘V/ ”Kw-r" g3 —_. _._._ {I \ * 2 ‘ *,/ Lineament possibly indicative of fault l g 3 ,fl N» ,/ EL SALVADOR a: / f VMJ f "‘3“ Major stratovolcano i "' "se/ alt ’ d .> /'{&4\[ * * v ./V ,r', 0 5OKM UAW/8554A! L) * I I l I l I MEXICO 4:, V, 5“ O N l8° * PACIFIC OCEAN ....... * INDEX MAP |4°00’ FIGURE 23 .—Relationship of the Motagua and Mixco faults to the main- -event epicenter, the epicenters of large aftershocks, and major structural and volcanic features in northern Central America. Numerals along the Motagua fault refer to lo- calities listed in table 7. Epicenters are from data of the N EIS and Person, Spence, and Dewey (this report); faults and volcanoes are modified from Dengo (1968) and Bonis, Bohnenberger, and Dengo (1970). slip component of displacement appears to increase irregularly from about 73 cm near near Quebradas at the eastern end of the exposed trace to a mea- sured maximum on a single trace of 142 cm in the area due north of Guatemala City.3 At the extreme western end of the observed trace, a single mea- surement suggests that displacement there may de- crease to 68 cm. However, since this locality is in an area of large-scale gravity sliding, the reliability of the measurement is uncertain; The vertical off- sets that we observed along the fault are generally minor (less than 30 percent of the horizontal com- ponent) and down to either the north or the south. An exception is the 10-km-long segment near Que- bradas at the eastern end of the observed surface “Later studies indicate that maximum sinistral displacement is as much as 325 cm in the area between E] Progreso and Chuarrancho. trace where the vertical displacement is consistently down to the north and locally as much as 50 percent of the sinistral component. Subsidiary faults and splays appear to be rela- tively scarce along the Motagua fault; the only two occurrences noted in our preliminary reconnais- sance are near El Progreso and Chuarrancho. J ust northeast of El Progreso, a subsidiary fault about 1 km long with 20-cm sinistral displacement is ori- ented roughly parallel to, and 400 m south of, the main fault trace. Near Chuarrancho (north of Gua- temala City), a prominent surface break splays off the main trace in a northeasterly direction at an angle of about 15°. This splay has a sinistral offset of 28 cm and was estimated from the air to be about 250 m long. 40 GUATEMALAN EARTHQUAKE OF FEBRUARY 4, 1976, A PRELIMINARY REPORT TABLE 7.——Characte’ristics of earthquake fractures along the Motagua fault [Measured aggregate displacement: N, estimate; (?), measured dis- placement probably not true value; >, measured displacement probably minimum value; S, sinistral or left-lateral; V, vertical. Leaders indicate no data. Observations by George Plafker, S. B. Bonis, and M. G. Bonilla, February 6—13, 1976] Station (fig. 23) Trend of fault zone Approx. width of fault Average trend zone (m) of fractures Measured displacement Sense of (cm) displacement Ground surface Remarks 10 ll 12 13 N70E N6SE N65E N65E N61E N70E N75E NSOE N7SE N7SE N75E 1.5 ------ 3 NSZE l NSOE 5 N46E l N61E 3.25—9 NN7SE 4 N40E -- N4OE 5 NN7OE 72 S Pasture Down—to—north. North- facing scarp as much as 5 m high along part of fault. See fig. 27. >33 S Dirt road Minimum displacement, measured across the largest fracture in a zone contain— ing 7 fractures. (?)107 S Railroad Offset railroad tracks. embankment Unreliable measure- ment due to gentle curve in tracks. 93 S Concrete— Good displacement lined canal measurement. 89 S Soccer field Good displacement measurement on offset sidelines. See fig. 26. >60 S Asphalt Displacement may be highway minimum value if highway fill par- tially decoupled from ground. (?)120 5 Dirt road Poor displacement measurement due to curve in road. ~90 s Plowed field ____________________ 20 S Pasture Subsidiary parallel fault 1 km long, and 200 m south of main break. 100 S Pasture Fair displacement measured on off- set cactus fence. 105 S 20 V Cultivated North side down. field See fig. 25. 142 S Pasture Fair displacement on offset path. Splay to north off main fault trends NGOE with 28 cm sinistral, and 17 cm vertical displacement. (?)68 S Dirt road Poor measurement. Probably masked by landsliding. GEOLOGIC EFFECTS 41 FIGURE 24.-—Oblique aerial view looking south towards linear trace of the Motagua fault (arrows) in farmland west of Cabanas. Furrows with sinistral offset may be seen in the field at left. See figure 23 for location of Cabanas. Faulting during the February 4 earthquake coin- cided closely with the previously recognized fault on the southern side of the Motagua Valley in the area east of El Progreso (Dengo and Bohnenberger, 1969; Bonis and others, 1970). Locally, however, the faulting of February 4 was as far as 1 km from the Motagua fault as it was mapped before the earth- quake. Moreover, faulting related to this earthquake has shown that the Motagua fault extends 85 km beyond its previously recognized western limits. Much of the Motagua fault trace is marked by linear stream valleys, minor scarps, shutter ridges, and sag ponds that are suggestive of repeated geo- logically youthful tectonic activity along this fault. Earthquakes that destroyed Omoa, Honduras, in 1859 (Montessus de Ballore, 1888) and caused dam- age at Quirigua (near Las Amates) in 1945 (Seis- mological Society of America Bulletin, v. 35, p. 194) and at Puerto Barrios in 1929 (Seismological So- ciety of America Bulletin, v. 19, p. 55) may have been generated along the Motagua fault or its off- shore extension. However, because surface breaks were not observed and because the epicentral loca- tions are not well constrained by the seismological data, it is not possible to preclude the alternative that these earthquakes were caused by movement on other faults in the area. RELATIONSHIP OF FAULTING TO DAMAGE The Motagua fault break caused extensive dam- age to buildings, roads, and the railroad. In Gualan, Cabanas, Subinal, and several smaller communities, structures that were astride the fault were damaged by the tectonic displacements. The most intense damage from shaking is within 40 km of the Mo- tagua fault trace (Espinosa and others, this report) and is predominantly in areas of thick pumiceous ash-flow deposits of Pleistocene age. These poorly consolidated deposits may have amplified ground motions. However, other factors, such as lateral 42 GUATEMALAN EARTHQUAKE OF FEBRUARY 4, 1976, A PRELIMINARY REPORT FIGURE 25.—View towards the north showing rows in a cultivated field west of E1 Progreso (station 11, table 7) that are offset 105 cm in a sinistral sense by the Motagua fault. See figure 23 for location of El Progreso. variations in energy release along the fault, con- struction practices, topography, and movement on subsidiary faults, undoubtedly influence the distri- bution of damage resulting from seismic shaking. SECONDARY FAULTS Secondary faults (faults which underwent sur- face displacement approximately concurrent with that on the main fault but which at the surface do not join the main fault) ruptured the ground sur- face in the Mixco area, in the western part of Gua- temala City, and in the area between those cities. Data presently available indicate that the secondary faults occurred as much as 30 km from the main fault. This distance from the main fault is one of the longest ever documented for secondary faults associated with historic strike-slip faults, and this fact alone makes the study of these faults of inter- national interest. The faults are particularly im- portant to Guatemala for at least two reasons. First, they traverse an urban area, and their future behavior should be considered with regard to pre- sent and future land use adjacent to them. Second, the generally north-trending faults near Guatemala City, on some of which the 1976 secondary ruptures occurred, may themselves be capable of producing damaging earthquakes. Even though such earth- quakes probably would be of smaller magnitude than those produced by the Motagua system of faults, they could damage Guatemala City because of their proximity. The following description of the distribution and trends of the secondary faults relies on data gath- ered by the Sociedad Geolégica de Guatemala, which in a remarkably short time prepared a map of the faults. The description of details of the fault- ing at particular places is based on field examina- tions by the writers. The secondary faults can be grouped into the three zones indicated in figure 29; individual faults are not shown, primarily because of the scale of the figure. The description of the faults will proceed from west to east. GEOLOGIC EFFECTS 43 FIGURE 26.—Oblique aerial view of Motagua fault trace crossing a soccer field at Gualan (station 5, table 7). Note char- acteristic right—stepping en echelon fractures and sinistral offset (89 cm) of white sideline stripe at right. See figure 23 for location of Gualan. MIXCO ZONE The 1976 surface faults in the Mixco area have been mapped only in reconnaissance, and the gen- eral trend, length, and width of the zone of rup- tures are uncertain at present. Faults are well de- veloped northeast of La Brigada at locality 1 (fig. 29), and a rupture crossed Highway CA1 just south of Mixco (loo. 2, fig. 29). The bearing between these two points is about N. 27 ° E.; however, en echelon faults occur in a broad band that extends to the area east of Ciudad Satellite (loc. 3, fig. 29). If the band is all considered part of the Mixco fault zone, then the trend may be more northerly than N. 27° E. For the present, the whole band (A, fig. 29) will be considered part of the Mixco zone, but further mapping may requiresubdivision into two or more zones. Individual faults in the zone com- monly strike between N. 10° E. and N. 30° E.; their lengths range from about 100 m to 3.5 km. The total length of the Mixco zone of faults is greater than 10 km. Three of the faults were examined at and north- east of locality 1 (fig. 29). One of these, traceable for 1.2 km, out several paved roads and ruptured the curbs and pavement at each crossing (fig. 30). Maximum measured displacement consisted of about 12 cm vertical slip (down to east) combined with about 5 cm of right slip, or about 13 cm of ob- lique slip. The vertical component was conspicuous at each street crossing, and the right-lateral com- ponent, though not conspicuous, could be seen at nearly all the crossings. The fault was essentially vertical, and rifting (displacement of the walls of a fracture perpendicular to the walls (Gill, 1972)) was very minor. In bare ground the fault appeared as a zone of discontinuous cracks, locally en echelon (stepping left), vertical displacement being dis- tributed over a zone of variable width ranging up 44 GUATEMALAN EARTHQUAKE OF FEBRUARY 4, 1976, A PRELIMINARY REPORT FIGURE 27.—View looking east along fault trace at the most easterly locality visited on the ground (station 1, table 7). Fault trace trends along base of 5-m—high scarp in foreground and through the fallen tree in the distance, which has a base diameter of more than 5 m. The tree was split and toppled by fault movement of about 72-cm sinistral displacement and 37-cm displacement down to the north. The north-facing steep scarp was probably formed by many repeated earlier move- ments along this same trace. to about 2 m. Along part of its length the fault coincided with moderately inclined slopes that may be degraded fault scarps. A second fault, nearly parallel to the first, could be followed for about 1 km. This fault intersected and displaced a high garden wall, passing about 4 m from a house without damaging it. The wall, made of brick with a reinforced concrete beam at midheight, was displaced vertically 13 cm, down on the. east; slight right-lateral separation was noted. Over most of its length, this rupture was near the base of a moderately inclined east-facing slope that may be a degraded fault scarp. A third fault in this general area could be traced for 1.7 km. Its principal displacement was also ver- tical, down to the east. The vertical displacement measured at a severed garden wall was 12 cm; no strike-slip was noted. The fault passed through a group of concrete-block houses (Husid and others, this report), where its course was marked by verti- cally displaced roofs and foundations (fig. 31), broken windows, and severely damaged interior and exterior walls (fig. 32). Aerial photographs taken in 1966 before construction of the houses show an east—facing break in slope, probably a degraded fault scarp, that the 1976 rupture followed in part. VILLA LINDA-CASTANAS ZON E A zone of faults (zone B, fig. 29) trending N. 20° E. extends more than 8 km from Colonia Villa Linda (10c. 4, fig. 29) to Colonia Castafias (loc. 5, fig. 29). Individual faults in the zone range in length from perhaps 100 m to 3 km and commonly strike between N. 18° E. and N. 31° E. One of the faults (near loc. 6, fig. 29) was examined briefly from the air and on the ground. It vertically dis- GEOLOGIC EFFECTS 45 FIGURE 28.—Part of the village of Subinal, 7 km west of El Progreso, showing the destruction of adobe structures near the Motagua fault trace. The fault is a broad zone of ground cracks that cuts diagonally across the lower right corner of the photograph (arrows). placed the highway called Anillo Periferico (not shown on fig. 29) by as much as several centi- metres, crossed an open field, cracked a masonry wall, and damaged at least two houses to the extent that they had to be vacated. In the open field the fault followed a small steepening in a gentle slope, which suggests that displacements of the same sense (down to the southeast) had occurred along the same line before 1976. In the same area were some deceptive artificial “scarps” resulting from shallow excavations, apparently to obtain topsoil. This fault was followed for about 0.5 km, but others nearby and parallel to it that we did not examine are much longer. INCIENSO-SANTA ROSA ZONE A zone of discontinuous faults (C, fig. 29) ex- tends from the area north of Incienso Bridge (loc. 7, fig. 29) to the vicinity of Colonia Santa Rosa (loo. 8, fig. 29), a distance of more than 7 km. In- dividual ruptures in the zone are generally less than 1 km long; strikes generally cluster around N. 19° E. but vary widely. One fault in the zone displaced the highway northwest of Incienso Bridge about 4 cm, relatively up on the east side. An excellent ex- posure in the roadcut there shows that the 1976 displacement occurred on a preexisting fault zone that had earlier displacement in the same sense. The earlier displacement, as well as 1976 displace- ment, was apparent reverse movement, up on the side toward the deep canyon spanned by the bridge. The cumulative displacement was not measured but is clearly several times larger than the 1976 dis- placement. A paleosol is cut by the fault, and some evidence suggests that the topsoil may have been faulted before 197 6 also; thus, it suggests geologi- cally young, pre-197 6 movement on the fault. 46 90° 30' w . San Pedro = Socuvenéquez ““0 |4° 40‘ N jfi |4° 35' 0 - 5KM l_l_l___l_|_l “- Villa \. __Nuevo _ FIGURE 29.—Zones A, B, and C (hatched areas) of secondary faults, and numbered localities discussed in text. A, Mixco zone; B, Villa Linda-Castafias zone; C, Incienso-Santa Rosa zone. REGIONAL TECTONIC RELATIONS OF EARTHQUAKE FAULTS The Motagua fault is part of a complex zone con- sisting of four major subparallel arcuate fault zones that trend in a general east-west direction across Guatemala and northern Honduras. As used in this paper, these are the Motagua and San Agustin faults in the Motagua Valley; the Polochic zone to the north, comprised of the Polochic and Chixoy (not labeled in fig. 23) faults; and the Jocotan Jocotan and Chamelecon faults (fig. 23). For con- venience, this broad group of faults is referred to zone to the south, which consists primarily of the herein as the Motagua fault system. The nature of the faults in this system and their relationship to the Cayman Trough (also referred to as Bartlett Trough) and the tectonics of the Ca- ribbean region have been the subject of much study and speculation. Most workers agree that the faults in the Motagua system are old fundamental breaks that have undergone recurrent displacement at least since the late Paleozoic. Some have postulated large sinistral displacements on the faults in this zone during the Cenozoic, although significant vertical movements occurred during the earlier history of GUATEMALAN EARTHQUAKE OF FEBRUARY 4, 1976, A PRELIMINARY REPORT FIGURE 30.—Fault displacement of road at locality 1 (fig. 29). Note right-lateral component of displacement. the zone. Excellent recent comprehensive sum- maries of the onshore geologic data relevant to the tectonic development of the region, including exten- sive bibliographies, have been given by Dengo (1968), Dengo and Bohnenberger (1969), Malfait and Dinkelman (1972), and McBirney and Bass (1969). Data on seismicity and marine geology and geophysics in the Caribbean and their relationship to plate-tectonics models have been presented by Molnar and Sykes (1969) and Jordan (1975). The secondary faults of the Guatemala City- Mixco area are part of a system of predominantly dip-slip faults in Guatemala, western Honduras, and El Salvador that lie between the Motagua fault and the chain of stratovolcanoes that passes through the highlands of Guatemala and E1 Salva- dor (Dengo, 1968; Williams and others, 1964; Wil- liams and McBirney, 1969). As shown in figure 23, these secondary faults group roughly into three sets that may in part reflect reactivated fractures in the crystalline basement rocks. The dominant set trends generally north to north-northeast; in a number of places, faults in this set bound prominent structural depressions such as the graben in Which Guatemala City is located, the Ipala Graben of eastern Guate- mala and western El Salvador, the Ulua Graben in western Honduras, and a series of grabens along the Chamelecén-Jocotan fault zone. A second im- portant set of faults is located along, and approxi- mately parallel to, the northwest-trending chain of stratovolcanoes (fig. 23) that comprise the Middle America volcanic arc. This set of faults becomes GEOLOGIC EFFECTS 47 FIGURE 31.—Fau1t damage to a house in Guatemala City. The roof, foundation, and sidewalk have been displaced vertically. increasingly prominent towards the southeast, where it bounds the central trench of El Salvador and the broad Nicaragua Depression (Williams and others, 1964, fig. 5; Dengo, 1968, fig. 9). A third set of oblique faults, not shown in figure 23, strikes northeast; it is locally well developed in the south- eastern part of Guatemala and is present in much of the adjacent area to the southeast (Williams and others, 1964). Although detailed studies of the dis- placement histories of these faults have not been published, there can be little doubt that many of them are geologically youthful features. Most of them offset upper Tertiary or Quaternary deposits, in places they are marked by prominent scarps that border topographic depressions, and some of them serve as conduits for Quaternary volcanic eruptions. Available data on the 1917—18 series of moderate- sized earthquakes that heavily damaged Guatemala City raise the possibility that those earthquakes may have originated on faults south or southwest of the city. The description by Vassaux (1969, p. 18—22) shows that Amatitlan and Villa de Guada- lupe, both south of Guatemala City, sustained more damage than the city proper in the November 17, 1917, earthquake, which initiated the destructive series. Vassaux (1969, p. 21) concludes that there were at least two centers of activity during the series, including Petapa (about 15 km south-south- west of Guatemala City) and Escuintla (45 km southwest of Guatemala City). However, we suggest that themost likely cause of these earthquakes was a series of fault displacements on the Mixco system and, perhaps, on the extension faults that bound the graben in which Lake Amatitlan is situated (about 16 km south of Guatemala City). LAN DSLIDES The main event and some of the large aftershocks triggered numerous landslides throughout abroad region of central Guatemala parallel to the main fault and extending as far westward as long 91°30’ W. (fig. 33). The landslides, numbering in the thousands, were mainly falls, slides, and flows in- volving thick pumiceous pyroclastic rocks, but they also included slides of consolidated bedrock (figs. 34 and 35). The overwhelming majority of the 48 GUATEMALAN EARTHQUAKE OF FEBRUARY 4, 1976, A PRELIMINARY REPORT FIGURE 32.—Fault damage to the rear exterior wall and roof of a house in Guatemala City. Vertical displacement near the front of the house was 12 cm. slides occurred along the steeper slopes of the deeply incised drainages in the Guatemalan high- lands and at the larger road and railroad cuts and fills. They blocked many transportation routes, in- terrupted surface communication lines, and in places damaged structures built in their paths (fig. 36). Some of the larger slump blocks and rotational slumps observed contain several million cubic metres of material. A number of these larger slides (indicated in fig. 33) have formed natural dams behind which lakes are developing (fig. 37). Such lakes are potentially hazardous because, when the dams are overtopped, rapid erosion and relatively sudden breakout could cause catastrophic flooding of inhabited areas and communication routes down- stream. Many highland drainages are choked with land- slide debris, particularly those areas of intense landsliding shown by a distinctive pattern in figure 33. This debris could become sufficiently water sat- urated during the rainy season (June through 0c- tober) to become mobilized and move as debris flows either naturally or as a result of aftershock activity. Such flows could pose a major hazard to communi- ties and transportation routes situated downstream. LANDSPREADING, FISSURING, SUBSIDENCE, AND SAND MOUNDS Landspreading, involving near-horizontal move- ment of mobilized or liquefied water-saturated granular deposits toward free faces, occurred at a number of localities in the Motagua Valley, along the Atlantic coast in Guatemala and Honduras, and along some lake shores in the highlands (fig. 33). Extension cracking and subsidence that accom- panied the spreading damaged structures in many of these areas. Throughout the lower Motagua Valley and part of the Chamelecén Valley in Honduras, widespread fissuring and settling of sediments in areas of high water table have caused damage to irrigation facili- ties, roads, levees, railroads, buildings, and pasture lands. At many localities, the compaction of satu- GEOLOGIC EFFECTS 49 59°00! as'oo' SALAMKQL IR“ \ \u/ .__~ / S ANTA ©DEL aul _ RU} HONDURAS j‘j‘fi’UAL/A‘N /41 \ g, ; "a \\ «I ,r Worms») \______ I ’49' {@ ZACAPA ? «a 7 “gm/(“M , ,1 EXPLANATION \# >_ {View Ma’DWU/ ’W W "“’ R K 2 / 9 QTPQJ.‘ Am: (9:1. eRosasso >fi\~w Emu \ / > CHUARRAN‘CHO‘Of— C \\ 5," CHIQL9ULAg< \ \ /@// ‘ \.\ r_/ \ R K / . . -. . l M? l w 3; ALTENAN /‘ ( 46:11AM K L ") Approximate limits of earthquake LAKE ; ©8PTATEM‘L“ ’ «/,/ ,\i induced landslides. Pattern indicates ATIrLAN / )9\‘”"°°‘\ ,/’ ,/ areas of intense landslide activity, l I "/ .c—\ . . - - ff; l /.x2§y =1 (1’ Solid triangle Indicates //5/,/K / (Lg; ,, Li}: a \ (“\u" /'\\ landslide-dammed drainage r x / , / l - R «— E‘SCU 7.. ,,~‘ . . ~\,\x , / / E}; \ \é )@ k\ w :/*I [figsfltfl‘ .UT APA‘QM, gig”; ‘XX— X / ‘ « e __,- f M‘ \t j ,2 / Area of abundant ground cracks f 7 ' ’ . - H II - . / a t,’ R f 1/ EL SALVADOR related to liquefaction. X rindlcotes } / / l small area of cracking % 50 Ml CARIBBEAN 50 KM ., 5“ O PAC/F16 OCEAN INDEX MAP FIGURE 33.——Areas of earthquake-induced landslides and of ground cracks probably related to liquefaction of unconsolidated deposits. Landslide distribution is from a preliminary study of post-earthquake aerial photographs by Edward Harp, Ray C. Wilson, and Gerry Wieczorek of the U.S. Geological rated materials was accompanied by ejection of water or water-sediment mixtures and the form- ation of sand mounds (figs. 38 and 39). Similar effects of sediment liquefaction were observed in the delta on the northern side of Lake Amatitlan near Guatemala City and along the shore of Lake Atit- lan; they reportedly occurred as far away as Lake Ilopango in El Salvador. VOLCANIC ACTIVITY There is no indication that the main event or any of its aftershocks were related to volcanic activity. One of the writers (Bonis), who studies the active volcanoes of Guatemala on a continuing basis, be- lieves that the amount of ash erupted from the Survey. Volcano Pacaya, located south of Guatemala City, may have increased slightly but that the apparent increase is well Within the limits of the prequake variations in the volcano’s activity. Numerous reports have been received of steam suddenly venting from the ground after the earth- quake or of changes in hot-spring activity. Because this area is characterized by Widespread and abun- dant thermal activity related to the volcanoes or their deposits, such reports are to be expected. Those reported anomalies that have been checked by geologists, however, suggest that there is no evidence of dramatic new volcanic activity initiated by the earthquake that could be a hazard to life or property. I5°oo’ M°oo’ 50 GUATEMALAN EARTHQUAKE OF FEBRUARY 4, 1976, A PRELIMINARY REPORT FIGURE 34.—Landslides in steep road cut in stratified pumice and ash deposits at San Cristobal, west of Guatemala City. FIGURE 36.—Aerial view of landslide in pyroclastic deposits near the edge of a steep-sided gully (barranco) in Guate- mala City. Slides such as this (as shown by arrow) and their associated headwall cracking caused extensive dam- age to homes, roads, and other facilities in the northern part of the city. FIGURE 35.—Aerial view looking northeastward along Rio Pixcaya, due north of Chimaltenango, showing numerous landslides in pyroclastic deposits. The river was partially dammed by a major landslide, shown by arrow in the middle: distance. (Also see fig. 37.) GEOLOGIC EFFECTS 51 FIGURE 37.—Landslide-dammed lake along Rio Pixcaya. The toe of the dam had been breached by the river by the time this photograph was taken on February 13. FIGURE 39.—View of linear sand mounds and circular to ellip- tical craters in area of water-saturated unconsolidated deposits along the Motagua River. FIGURE 38.—Aeria1 view of ground cracks and sand mounds (white patches) in unconsolidated alluvial deposits along the Motag'ua River north of Quebradas in the lower Mo— tag'ua Valley. Ground cracks are believed to result from liquefaction of water-saturated sediments and spreading towards the river channel. THE GUATEMALAN EARTHQUAKE OF FEBRUARY 4, 1976, A PRELIMINARY REPORT INTENSITY DISTRIBUTION AND SOURCE PARAMETERS FROM FIELD OBSERVATIONS By ALVARO F. ESPINOSA, RAUL HUSID, and ANTONIO QUESADA 4 INTRODUCTION The earthquake of February 4 in Guatemala was felt over an area of at least 100,000 kmz. It origi- nated on the Motagua River Valley to the east of Los Amates and propagated west along the Mota- gua fault through Gualan and El Progreso to Chuarrancho (details of the surface faulting map- ping are given by Plafker and others, this report). The sense of motion from field observations as well as from instrumental seismic determination (De- wey and Julian, this report) is a left-lateral strike- slip fault. This paper is a preliminary report on the earth- quake-damaged area studied during the period Feb- ruary 6—22. The purpose was to obtain information in Guatemala City and along the Motagua fault area to delineate the distribution of intensities (Modified Mercalli), damage to adobe-type struc- ture, strong motions, and other related phenomena. The ground movement in the fault zone was very severe, and numerous estimates of the time duration of strong shaking range between 30 and 40 s. The first movement was vertical and was followed by a strong horizontal ground motion, which was so strong that it hindered people from getting out of bed, and in many instances people were thrown down or were unable to walk. In many areas of the country, a second intense horizontal motion was re- ported nearly a minute after the main disturbance. In one particular case illustrating the last report, a man tried to get out of bed and failed. He waited several seconds and tried again,‘ failing for the second time. He stayed in bed for about 30 s, and then he was able to get up, pick up a child from a crib, and go out. As he was going out, he felt the second severe horizontal ground motion, which col- lapsed his house. 4 Organization of American States, Washington, DC. 52 CASUALTIES AND DAMAGE The statistics for casualties and damage are given in table 8 by Departments and in table 9 by municipalities. These figures were provided by the Comite Nacional de Emergencia, Presidencia de la Republica de Guatemala. The total number of houses destroyed, as of February 15, 1976, was 254,750, and 1.07 million people were left homeless. From a total population of 3,213,962, there were 22,868 deaths and 77,190 injuries as of March 3, 1976. The total loss in Guatemala is $1,100 million (from Ministry of Finance statistics). INTENSITY DISTRIBUTION IN GUATEMALA The areas of maximum Modified Mercalli inten- sity are concentrated in‘and near the town of Gua- lan, Department of Zacapa, and to the west in the town of Mixco, Department of Guatemala. Maxi- mum intensity in the meizoseismal area was IX. In the Gualan area, however, much of the damage could be classed as VIII. On the Modified Mercalli scale (Richter, 1958) , large landslides, such as those that developed between Guatemala City and El Progreso and also between Guatemala City and Antigua, sug- gest an intensity greater than IX. Another factor that yields higher intensities is surface faulting, examples of which were observed in Gualan (see cover photograph) and along the Motagua fault. The authors visited, by car and helicopter, villages in areas of high, intermediate, and low damage and by using questionnaires gathered data (fig. 40) used to assess the Modified Mercalli intensity rat- ings throughout the nation. These intensities are rated by using the abridged version of the Modified Mercalli intensity scale (Richter, 1958), with the following exceptions: landslides are not rated in this report as intensity X; rails bent greatly are not rated as intensity XI; and destroyed bridges are not rated as intensity XI. INTENSITY DISTRIBUTION AND SOURCE PARAMETERS FROM FIELD OBSERVATIONS 53 TABLE 8.—Casualties and damage, by Departments Percent Department (state) Population Deaths Injuries damage Guatemala 1,681,736 3,370 16,549 68.82 E1 Progreso 78,364 2,028 7,767 90.43 Sacatepéque: 105,210 1,582 8,855 71.00 Chimaltenango 214,290 13,754 32,392 88.00 Santa Rosa 20,591 40 291 1.60 Solola 30,707 110 300 10.00 Totonicapan 162, 678 27 89 34 . 00 Que zaltcnango 79, 241 14 228 1. 00 Huehuetcnango 34,362 10 so N.A.1 Quiché 150,073 843 5,722 73.00 Baja Verapéz 49,820 152 718 82.50 Alta Verapaz 59,664 18 953 67.50 Izabal 183,370 73 379 40.00 Zacapa 107,148 693 1,998 72.86 Chiquimula 76,603 50 378 50.00 Jalapa 88,802 91 473 31.67 Jutiapa 91,303 13 48 10.00 1N.A. - Information not available. These exceptions have been made because the Modi- fied Mercalli intensity scale is used to represent the intensity of an earthquake based on purely vibra- tional effects as well as on the damage sustained by structures from the earthquake. The above effects are of a secondary nature to the seismic energy release. In the area of heavy landsliding, many adobe houses sustained no damage. Landslid- ing implies intensity X, but undamaged adobe houses suggest much lower intensities (fig. 41). Also, numerous houses near landslides along the highways toward the Pacific were not damaged. Rails bent greatly are not related directly to ground shaking, but this effect is related to ground movement due to faulting, as seen in Gualan (fig. 42A) and near El Jicaro (fig. 423), or to ground compaction, as observed in Puerto Barrios (fig. 420). Another factor that yields higher intensities is surface faulting, examples of which are observed in Gualan (see cover photograph) and along the Motagua fault, near Las Ovejas (fig. 42D). The Agua Caliente Bridge was destroyed, and the Benque Viejo Bridge was at the verge of collapse owing to large ground displacement in those areas. The displacements sustained by the Agua Caliente Bridge were larger than those planned in the origi- nal design for the structure. The damage to these structures gives an indication of the severity of the ground deformation but does not indicate the level or the time duration of the seismic disturbance. The Benque Viejo Bridge is similar in construction to some of the highway overpasses in the San Fer- nando Valley of California, which collapsed as a re- sult of the 1971 San Fernando earthquake. The isoseismal map shown in figure 43 repre- sents a preliminary Modified Mercalli intensity dis- 54 GUATEMALAN EARTHQUAKE OF FEBRUARY 4, 1976, A PRELIMINARY REPORT TABLE 9.—Casualt'ies and damage, by municipalities in a department (as listed in table 8). [*This consecutive number identifies the total number of municipalities] Municipality Population Deaths Injuries % Damage l.* Chimaltenango 20,194 600 3,000 25% 2. San Jose Poaquie 9,795 1,000 2,657 90% 3. San Martin Jil‘otepeque 33,066 2,920 5,000 100% 4. Zaragoza 7,317 300 1,000 100% 5. Patzicia 10,585 811 2,248 90% 6. Sta. Cruz Ba1anya 2,903 100 500 80% 7. Tecpan 24,181 3,023 7,000 100% 8. Patzun 18,900 309 390 85% 9. Parramos 3,237 200 900 90% 10. E1 Tejar 3,039 50 900 85% 11. San Andres Itzapa 8,447 150 728 90% 12. Yepocapa 10,457 87 289 90% 13. Comalapa 18,163 3,050 5,000 95% 14. Sta. Apolonia 4,182 900 844 85% 1. Guatemala 700,504 1,195 5,550 45% 2. San Pedro Sacatequepez 10,714 720 1,667 100% 3. San Juan Sacatepequez 43,116 720 2,400 100% 4. Chuarrando 6,985 42 1,789 60% 5. Sn. Raymundo 9,225 118 1,543 60% 6. San Pedro Ayampuc 10,481 54 316 90% 7. Mixco 129,878 346 2,400 80% 8. Amatitlan 26,412 16 80 20% 9. Palencia 18,982 68 157 85% 10. Villa Canales 31,774 2 100 20% 11. Sn. Miguel Petapa 8,078 2 140 70% 12. Sta. Catarina Pinula 12,934 9 7O 75% 13. Chinautla 32,763 50 15 80% 1. Progreso Cabecera 11,048 1,300 3,500 95% 2. E1 Jicaro 6,197 372 2,538 100% 3. San Agustin Acasagustlan 17,344 126 917 50% 4 Morazan 7,080 134 570 100% 5. Sanarate 15,253 69 137 70% tribution of the main event in Guatemala. The iso- seismal for an intensity rating VII follows the gen- eral trend of the mapped Motagua fault. The iso- seismal VIII, and higher, in the Departments of Sacatépequez, Chimaltenango, Guatemala, and the southern part of Quiché, follows the general trend of maximum adobe-damaged areas. The high intensities attenuate faster in the east- ern part of the country near Los Amates. However, as one progresses west, from E1 Jicaro to near Sa- narate, the intensities increase in a narrow area, and then, outside Sanarate, there is a sudden in- tensity decrease for the next 35 km and again a rather large increase to Modified Mercalli inten- sities of VIII and IX in the Mixco area. Guatemala City, as it appears on this map, has been assigned an average intensity of VII and, in the northern part of the city, an intensity rating of VIII. A detailed mapping of the intensity distribution in Guatemala City associated with the February 4 earthquake is now being done and will be presented in a sub- sequent separate report. A study of intensity distributions unexpectedly showed that a number of small villages near the causative fault sustained no damage. The intensity ratings attenuate rather rapidly in a north-south direction in the eastern part of the country. The epicenter was located west of the town of Los Amates, approximately 12 km away. The highest in- tensities were in Gualan and 145 km due west in Mixco. In Guatemala City, the intensity was IX in the center of the city. To the northwest of the city, the intensity was VIII along the strike of some faults mapped after the earthquake. The intensity VII isoseismal has an east-west trend, from Los Amates, parallel to the Motagua INTENSITY DISTRIBUTION AND SOURCE PARAMETERS FROM FIELD OBSERVATIONS 55 TABLE 9.—Casualties and damage, by municipalities—Continued Municipality Population Deaths Injuries Z Damage 1. Sacatepequez 26,945 277 1,251 257. 2. Sumpanjo 10,232 315 1,300 1002 3. Magdalena Milpas Altas 2,921 135 584 507. 4. Jocotenango 3,426 118 582 50% 5. San Lucas Sacatepequez 4,344 157 1,170 407. 6. San Antonio Aguas Calientes 3,866 113 544 50% 7. Pastores 4,592 127 567 30% 8. Sta. Domingo Xenaxoj 2,759 57 560 70% 9. Sn. Miguel Duenas 4,215 7 524 307. 10. Santiago Sacatepequez 7,943 218 1,247 40% 11. San Maria de Jesus 7,144 2 218 20% 12. San Bartolome Milpas Altas 1,513 27 246 407. l. Quiche 35,147 56 175 2. Joyabaj 32,134 600 5,497 95% 3. Chinique 4 , 353 35 4. Chichicastenango 45,733 140 1. Jutiapa 54,680 18 2. Asuncion Mita 29,071 13 30 1. Zacapa 34,703 198 475 50% 2. Gualan 23,375 187 550 99% 3. Rio Hondo 9,637 95 281 80% 4. Cabanas 5,817 89 240 95% 5. Huite 3,941 67 152 75% 6. Usumatlan 3,771 26 150 507. 7. Teculutan 5,933 31 150 60% 1. Baja Verapaz 21,913 119 377 75% 2. Rabinal 20,393 33 341 907. l. Izabal 38,903 30 167 50% 2. Los Amates 45,537 14 158 22 3. Morales 52,677 29 54 l. Totonicapan 52,688 10 507. 2. St. Maria Chiquimula 15,161 3 10 3. Momostenango 43,398 3 11 4. San Cristobal Totonicapan 16,623 3 5. San Fco. e1 Alto 19,329 21 55 507. 1. Chiquimula 38,872 10 110 2. Esquipulos 19,304 20 110 1% 3. Sn. Jacinto 3,851 20 158 fault for a distance of 150 km to near San Antonio La‘ Paz in the Department of E1 Progreso. From San Antonio to Zaculeu in the Department of Quiché, an east-west distance of another 85 km, the intensity VII isoseismal broadens considerably to 72 km in width. The intensity VIII and IX isoseis- mals follow a trend parallel to the trend of surface faulting. The dashed line is questionable for the in- tensity VIII isoseismal continuation to the west be- tween Sanarate and to the east of Guatemala City. The number of landslides in this area was very high, on the average one landslide per kilometre. From Guatemala City toward El Progreso, there were 32 landslide areas as far as Kilometre 29 near the town of El Chato and a total of 54 landslide zones in the first 48 km on this main highway to- ward the Atlantic Ocean. A landslide zone consists of one to three large landslides obstructing the high- way. On this road there were two bridges that suffered considerable damage. The Agua Caliente Bridge col- lapsed and impeded traffic, and the Benque Viejo Bridge was on the verge of collapse (Husid and others, this report). Severe landslides occurred also 56 GUATEMALAN EARTHQUAKE OF FEBRUARY 4, 1976, A PRELIMINARY REPORT TABLE 9.—Casualties and damage, by municipalities—Continued Municipality Population Deaths injuries Z.Damage l. Solola 25,819 110 300 1. Jalapa 45,425 27 254 50% 2. San Pedro Pinula 23,846 9 97 25% 3. Mataquescuinla 16,145 55 122 202 1. Sta. Rosa 14,127 40 291 l. Alta Verapaz 43,505 15 700 60% 2. Sta. Cruz Verapaz 3,508 3 253 70% 1. Quezaltenango 65,526 14 228 1. Aguacataman 18,492 2. San Sebastian Huehuetenango 7,824 3. San Pedro Necto 11,371 4. San Miguel Acatan 15,011 5. Concepcion 8,102 6. Newton 12,613 7. St. Ana Huista 4,755 8. La Libertad 14,756 9. Colotenango 9,458 10. San Gaspar Ixchil 3,058 1. Villa Nueva (Guate) 5 12 8% 2. Acatenango (Chimaltenango) 22 Total 22,525 74,027 along the main highway to the Pacific Ocean, be- tween Guatemala City and Antigua. Landsliding in- terrupted road traffic along these two main through- ways and also disrupted railroads near Las Ovejas, Gualan, E1 Progreso, Rio Hondo, and Puerto Bar- rios. The preliminary intensity distribution in Guate- mala (fig. 43) suggests that the shaking intensity was greater in the western part of the country. This isoseismal pattern suggests a fault propagation rupture from east to west. The isoseismals broaden to the west, a phenomenon similar to a Doppler effect, which creates a constructive interference pat- tern to the west. Several small villages were located near and at intermediate distances from the causa- tive fault; for example, adobe construction in Jones, about, 8 km from the fault, sustained no damage. Also, in several communities south of the Motagua fault, such as San Pedro Pinula nearly 25 km from the fault, adobe construction sustained no damage. Numerous small villages in which adobe buildings sustained no damage were observed 8 to 30 km from the causative fault. Other towns, such as Entre Rios approximately 37 km due east of Morales, near the extension of the Motagua fault, had an intensity rating of only V. The pattern of isoseismals displayed in figure 43 may be the effect of a moving source in the near field. This effect is shown schematically in figure 44 (Benioff, 1955) to be the progression of a discrete number of points. The initiation of the fault mo- tion is near Los Amates, at point 0, and terminates at point 8, toward Guatemala City. The largest circle represents a wavelet at point 0, which in the time domain is shown at the lower part of this dia- gram and is identified with a 0, and the succesive circles represent the wavelet position as it propa- gates from points 1, 2, 3. . ., and so on. There is a time delay between these points, as is seen in the two lower diagrams. The lower left diagram repre- sents the signal from each point as seen at a station west of the fault, and the lower right diagram re- presents the signal as seen at a station east of the fault. The composite signal for each direction of propagation is shown as the resultant in the lower part of figure 44. The energy can be measured as the square of the velocity amplitude; hence, the re- sultant wavelet traveling to the west, shown at the 57 l / /l 5 I ‘k/‘u—w'i / ,/ 2 II/ I/ (2 [J éLa Mesilla ’/ ~\\ . , " Chaju| , / o _ I i} Q /‘ Cobayn . P I , O , San Sebastian \ I anzos , W ' ° 9 ‘H‘ .. ° / , ,\,...\ ' ‘\-- . ' ' Los Amates ( k/ \\ HNJ f “1’ VF‘L”? ,L‘ f 0 // - \V- , '/:5u la'n '\ \ -J1-—’ \fg, Santa Cruz \\ Salgna /,A R‘Td “’3“ a. . . ~' 0 > K i Del Qutc.he . $/’ . m on n' Zacapa _ l5 N — ' Totonicapa’n o \. , \ '@ o ( J 2 0 —' ,‘_ . El Jicaw) /I f . \_,: O :\9"Af\” (3 .Las OvejasA--‘z A-fi’ \ an h II “ ' 1 Ch" Sitgea";°i'°'omi 54‘5" fii'fieifi «h @C 1”” a\ b/ Ima g OEI Chato Ping. I . ( '@Guatemala 9 ' ° ' Jalapa r , \ . EsculntLa/ o J PAC/NC OCEAN I I City FA / Cuilapa Li\". Jut©iapa \‘ JJ/ 0 50 KM L____J l 92° 9|° eo°w 89° FIGURE 40.—Intensity sampling distribution. Each dot represents the location where one or more questionnaires was com- pleted during a survey taken in Guatemala. Largest circle indicates epicenter location of main event. (Base map modified from Guatemala Instituto Geografico Nacional, 1974, 1:500,000.) lower left in figure 44, has a larger concentration of energy than the slightly dispersed wavelet traveling in the opposite direction. The possibility of a fault rupture traveling from east to west (suggested in fig. 43) could be verified with teleseismic data at western azimuths. Also, the suggestion of a double rupture or a multiple earth- quake from the isoseismal distribution is plausible. If this second alternative is adopted, the first earth- quake will be constrained to the observed surface faulting from Los Amates to Kilometre 15 south- west of El Progreso. The second event could be as- sociated with the secondary faulting observed in the Mixco area. Other factors that may enter into the intensity distribution pattern shown in figure 43 are seismic- wave amplification effects, topographic seismic- wave amplification, influence of the surficial soil conditions, and depth of the water table. The high-level isoseismals VI, VII, and VIII re- present the shape of the radiation pattern in the near‘ field, assuming the rupture started near the epicentral region and propagated west. A similar suggestion made by Hanks (1975) correlates the intensity VI and VII isoseismals from the San Fer- nando earthquake with the radiation pattern for 8-s Rayleigh waves and also with the azimuthal varia- tions of the amplitude ground displacements. The isoseismal map was plotted on a 1:500,000 geologic map of Guatemala (edition by Bonis, Bohnenberger, and Dengo, 1970), and no simple correlation was foundbetween the gross surficial geology and the intensity distribution. There is a correlation between the fault that slipped during the February 4 earthquake and the intensity dis- tribution. West of Guatemala City is the Mixco fault, which has a north-south trend. In this re- gion, the intensity rating attains a maximum of IX. The earthquake was felt by nearly everyone over an area of at least 93,125 kmz, suggesting that an intensity V or higher extended over that area. The areas felt for intensities VI and higher are given in table 10. 58 GUATEMALAN EARTHQUAKE OF FEBRUARY 4, 1976, A PRELIMINARY REPORT FIGURE 41.—One of the many landslides obstructing the main highway from Guatemala City toward E1 Brogrerso. _ TABLE 10.—Felt area. for Modified Mercalli intensities VI and higher Modified Mercalli intensity Felt area (km 2) VI and higher ________________________ 32,697 VII and higher _______________________ 6,437 VIII and higher ______________________ 2,495 IX (small area) ______________________ 125 DISTRIBUTION OF ADOBE DAMAGE The map shown in figure 45 represents the over- all distribution of adobe damage in the Republic of Guatemala. This map was prepared with informa- tion from a variety of sources. First, we studied reports from the Comite Nacional de Emergencia, which gathered a large amount of information through the local committees. These reports were supplemented by detailed lists of damage and by our assessment of the sustained damage in the field. Newspaper clippings from the national press were also consulted. Other sources of information were gathered by the Universidad de San Carlos and by a team of the Camara de Construccion in Guatemala City. The adobe damage distribution map shows the amount of damage to a given Village. On this map, a scale of 4 means that 91 to 100 percent of all the adobe houses in a village collapsed (fig. 46); a 3 implies 76 to 90 percent; a 2 means 51 to 75 per- cent; 1 means 26 to 50 percent; 0 in the scale im- plies negligible to 25 percent damage to adobe houses. The maximum adobe damage was found in FIGURE 42.—A, Rails bent in Gualan, Department of Zacapa. B, Rails repaired between E1 J icaro and Las Ovejas, De— partment of El Progreso; also shown in B is the surface faulting with an east—west trend. This photograph was taken from a helicopter in. a westward direction. C, Rails bent on the Puerto Barrios wharf. D, Aerial view of fault trace near Las Ovejas. See figure 40 for town loca- tions. 60 GUATEMALAN EARTHQUAKE OF FEBRUARY 4, 1976, A PRELIMINARY REPORT lSN PAC/F/C‘ OCEAN i l Intensity Distribution 0 50 KM l_i___|_l_l__l l i 92° 91° 90°w 89° FIGURE 43.—Modified Mercalli intensity distribution in Guatemala from the main event. Circle indicates epicenter location of the February 4 earthquake; dashed Line indicates approximate isoseismal. (Base map modified from Guatemala Insti- tute Geogréfico Nacional, 1974, 1:500,000.) the Departments of Chimaltenango, Guatemala, and lower Quiché. These regions of the country are the highlands of Guatemala. The affected region of 51- to loo-percent damage covers a surface area of ap- proximately 8,725 kmz. There are two well-defined areas of maximum damage; one follows a trend parallel to the Motagua fault located in the Depart- ments of Izabal, Zacapa, and El Progreso. A gap is found between the 4 and 2 contours for nearly 40 km. Within the next 35 km there is another con- centration of high damage (fig. 45). The damage ,pattern diminished rapidly to the west into the state of Totonicapan. TABLE 11.—Damage to adobe-type structures Damage Percent area Scale 1 damage (km 2) 2 and higher ___________ 51—75 8,725 3 and higher ___________ 76—90 2,825 4 _____________________ 91—100 1 ,7 20 1 Scale values taken from figure 45. A number of cracks, fractures, and short fault scarps were found due west of Guatemala City in the Mixco area. Damage decreases more rapidly to the east than to the west of the zone of strongest shaking. The areas damaged for different percen- tiles of collapsed adobe houses are given in table 11. Adobe structures are apt to collapse more easily than bajareque (Husid and others, this report) and wooden-type construction. The latter construction is found throughout the Motagua River Valley in the lowland near the coastal area of Guatemala. In Puerto Barrios and in Santo Tomas, many of the houses are built of wood. The wharf of Puerto Bar- rios collapsed because of structural failure and pos- sible ground compaction (Ray Wilson, oral com- mun., 1976) (fig. 47). There was no damage to well-built structures in the port of Santo Tomas, 5 km south of Puerto Barrios. In the port of Santo Tomas, a number of ground cracks were observed, and ground-level changes of 10 to 15 cm in the maritime port were due to ground compaction. Near Puerto Barrios, mud spouts occurred during the main earthquake. The damage in this region to wood-frame and adobe structures was minimal. INTENSITY DISTRIBUTION AND SOURCE PARAMETERS FROM FIELD OBSERVATIONS 61 ///C \\ \\ \ ////:’- \\\\ \\ \ ////:\ \ \ \ \ \ toward //¢/,—\ \ \ \ \\ Guatemala City f//4:\ °\’/- near Los Amates ‘— lll , /l ) l ,s / / / \ /// // active fault segment kt _’ / / // <— direction of fault progression mmamo ___I—\___ g (D —_/—¥_._ -: —_,_\_ = Wavelets g Wavelets 8 Time ——> < Time—> < _—J—\.—_ resultant in direction resultant in reverse direction of fault propagation FIGURE 44.—Schematic of slip progression and its effect on wave amplitudes and shapes due to a moving source (modified from Beniofl", 1955). Fault propagation from east, near Los Amates, to west, toward Guatemala City. The pattern of adobe damage (fig. 45) is similar to the intensity distribution shown in figure 43. The similarity suggests that more energy was released to the west than to the east. It also suggests that seismic energy was released along and near the known fault zone of the Motagua fault. The maxi- mum concentration of damage, mostly in the De- partment of Chimaltenango, is suggestive of a con- structive interference of seismic waves due to a moving dislocation. INTENSITIES IN GUATEMALA CITY The distribution of earthquake intensity was in- vestigated by canvassing Guatemala City in a man- ner similar to the way in which the Republic of Guatemala was canvassed. In the process of canvas- sing Guatemala City, data were obtained from a representative number of questionnaires solicited from each of the 16 zones into which the city is divided. A total of 1,050 questionnaires was com- pleted and collected. A map of intensity distribution for Guatemala City associated with the earthquake is now being prepared. The following cursory comments are the preliminary result of our fieldwork. The maximum Modified Mercalli intensity in the city was IX, as shown by the partial collapse of re- inforced-concrete structures, and there were pockets of high intensities in different zones of the city. A cursory examination of the questionnaires shows that some of the localized high intensities may be related to possible ground-amplification effects of seismic waves. Similar findings were reported else- where on a seismic-zonation study in Lima, Peru (Espinosa and others, 1976). A number of chimneys cracked and collapsed in different suburbs of the city (fig. 48). The varia- tions of the intensity ratings in the city proper varied from VI to IX. A number of wells showed an increase in temperature of more than 2° C, and their water level changed drastically, sometimes more than 60 cm, due to the February 4 earthquake. In the northern part of the city, damage to adobe- type construction was intense. Also, throughout the city, there were a number of damaged reinforced- concrete structures (Husid and others, this report). A number of landslides occurred to the northwest in the Colonia 10 de Julio, Zone 19, on the Barranco de las .Guacamayas. The barrancos have an almost vertical drop of nearly 90 m. Figure 49 shows the head of a landslide in the area, at approximately long 90°33.5’ W., lat 14°40.6’ N. In Zone 7, 13 Calle B. No. 31—14, a one-story brick house sustained a large amount of damage from a ground crack, possibly a small fault with a north- south trend, which crossed through the house. This ground crack could be traced very easily through the suburb, across damaged streets and houses. In other zones, such as Zone 19, 8 Avenue No. 6—96 there was considerable damage because of ground failure. In Zone 11, 8 Calle No. 20—62, Colonia Mi- rador, there was moderate damage to a one-story brick house. These examples show the damage that can be caused by faulting in the immediate vicinity or below some of these houses in Guatemala City. INTENSITIES IN NEIGHBORING COUNTRIES The Modified Mercalli isoseismal V covered an area from Ilepango, Izales, and San Salvador in El Salvador to Santa Barbara, San Pedro Sula, and Puerto Cortes in Honduras. In Tegucigalpa, an in- tensity rating of IV was assigned. To the north, 62 GUATEMALAN EARTHQUAKE OF FEBRUARY 4, 1976, A PRELIMINARY REPORT / /I g‘ "Twp/«v.1 GULF OF / / f) w . HONDURAS , ‘ — |5°00' , ‘ ‘ _|I§®\\\\\3 EL SALVADOR [V' CUILAPAO M) s, / 50km North 1 1:1 r, i /.-\§ l |4oOOI 92°oo' 9I°OO' 90°00' 89°00' 88° 00' FIGURE 45.—Contour map showing damage to adobe—type structures in Guatemala owing to the February 4 earthquake. See figure 1 for Department names. (Base map modified from Guatemala Institute Geografico Nacional, 1974, 1:500,000.) Belice had a IV and Mexico City a IV rating. Poc- tun in the Guatemalan northern territory had an intensity of V. The total felt area for an intensity V or higher was 93,125 kmz. The earthquake of 1773, called the Santa Marta earthquake, caused extensive damage in the Guate- malan highlands down to El Salvador. The Feb— ruary 4 earthquake and the Santa Marta earth- quake affected similar areas in Guatemala and El Salvador, but there was no damage to El Salvador from the February 4 earthquake. GROUND MOTION AT INTERMEDIATE DISTANCES Strong ground displacements were recorded on a seismoscope in Guatemala City (Knudsen, this re- port), but, unfortunately, no accelerographs were operational in the area at the time of the earth- quake. Strong ground motions were experienced in the Mixco area, in particular at the Licorera Mixco, Where a large oil-fired boiler was displaced horizon— tally a maximum of 41 cm. In La Colonia 10 de Julio, a northwestern suburb of Guatemala City, a large water tank sustained damage to its back bracings at the upper level (nearly 12 m from the ground). The shear bracings also were bent. The water tank sustained damage at a height of 4 m from the bottom of the tank and 16 m from the ground, where the bolts tying the steel walls of the water tank were sheared and water leaked out. In Guatemala City, in the chemistry building of the ICAITI (see Glossary) complex, a large machine weighing 1,298 kg was displaced 7 cm horizontally. To the southeast of the city in Zone 10, in the Her- rera’s grounds, four medium—sized marble statutes were thrown 40 cm from their stands (fig. 50). There were some reports that hanging lamps dented the ceilings of homes and that outside hanging lamps left a fan of dents on the walls. In the maritime port of Santo Tomas, a large up- right unloading cargo crane weighing 60 tons was displaced from its rails 5 cm in a northeasterly direction. This port is northeast of Los Amates. SOURCE PARAMETERS FROM FIELD OBSERVATIONS Many major earthquakes have occurred in close association with major faults; as a result, a number of empirical relations that correlate magnitude, length of surface faulting, fault displacement, and magnitude have been derived. King and Knopofi' (1968) developed the following relation: log Lu2=224 M—4.99. (1) I6°OO' FIGURE 46.—Photog'raphs showing the sustained damage in the towns of: A, Joyabaj ; B, Comalapa; C, Tecpan; and D, San Martin Jilotépeque. Using this equation, one determines the expected cm (Plafker and others, this report). This yields a magnitude for the Guatemalan earthquake by sub- magnitude of 7 .4.If King and Knopofi’s expression stituting the field observations of L 300 km and (modified from Tocher) is used, the average horizontal fault displacement 17:100 log L122=2.75 M—8.93, 64 GUATEMALAN EARTHQUAKE OF FEBRUARY 4, 1976, A PRELIMINARY REPORT FIGURE 47.—Puerto Barrios wharf, Department of Izabal, destroyed by February 4 earthquake. Arrows show large ware- house partially submerged. See figure 40 for location of town. one obtains an Ms value of 7.4. These values are in good agreement with the value MS=7 .5 determined for the Guatemalan earthquake from teleseismic ob- servations (Person and others, this report). Having established that Ms=7.5 is a reasonable estimate for this earthquake, we proceed to deter- mine the seismic energy, Es, from the Gutenberg— Richter energy-magnitude relationship given by Richter (1958) : log ES=11.8+ 1.5 MB. (3) This equation yields ES=1.1><1023 ergs for the Guatemalan earthquake of Ms=7.5. The stress drop (Keilis-Borok, 1959; Aki, 1966; Brune, 1970, eq. 30) is defined by ,m 77r Aa=— —, 4 716 () where r is the radius of a circular dislocation, as— sumed in this case to be 150 km, the rigidity repre— sentative of volume around the faulted area is FIGURE 48,—Chimney collapse from a one-story house in Guatemala. City, Zone 11. INTENSITY DISTRIBUTION AND SOURCE PARAMETERS FROM FIELD OBSERVATIONS 65 FIGURE 49.—Head of large landslide (shown by arrow) in the Barranco de las Guacamayas, Guatemala City. assumed to be ,u=3><1011 dyne cm”, and fi= 100 cm; then, Aa=3 bars. To verify the above stress-drop determination, the seismic moment M0 and the energy E (Brune, 1968), defined by M0 = ,ut—LA (5) and E = mA, (6) are combined to obtain 0' ,u E' a=M0 . (7) Assuming ,lL and i as given above and A, the disloca- tion area, to be 300 km in length and 20 km in width (assumed, Brune, 1968, table 2), M0=5.4>< 1027 dyne-cm. Using E, determined from equation (3) FIGURE 50.—A marble statue thrown 40 cm from its pedestal; it is 120 cm high, weighs approximately 200 kg, and is located in Zone 10 in Guatemala City. to be 1.1x 1023 ergs, and substituting these values in equation (7 ), one obtains o= 18 bars. From the above computations, it seems that the February 4 earthquake was a low-stress-drop earth- quake. Independently, Dewey and Julian (this re- port) have found Au=6.6 bars, using information determined from the spectral density of G-waves. Using a Modified Mercalli intensity rating of VII for Guatemala City With an epicentral distance of 157 km, one determines the particle horizontal velocity (Espinosa, unpub. data, 1975) from log 9'6= 1.27—0.79 log A+0.16 1mm (8) to be 4.5 cm/s, where A is the epicentral distance in kilometres and [mm is the Modified Mercalli in- tensity rating. If, instead of the epicentral distance, one uses the distance from the causative fault to Guatemala City (25 km), then one obtains a maxi- mum particle velocity of 19.3 cm/s. The above quan- tities give an indication of the level of ground mo- tion of the main event experienced in Guatemala City. 66 GUATEMALAN EARTHQUAKE OF FEBRUARY 4, 1976, A PRELIMINARY REPORT An earthquake similar to the Guatemala earth- quake was the Varto-Ustukran earthquake of August 19, 1966, on the Anatolian fault system in Turkey (Ambraseys and Zétopek, 1968; Wallace, 1968), and data obtained from it are very similar to the observations made by the authors after the Guatemalan earthquake. These two earthquakes are strike-slip faults, the former right-handed and the latter left-handed. In terms of fault displacement, magnitude, and length of faulting, the February 4 earthquake is similar to the November 26, 1943, Turkish earthquake, which had a magnitude of 7.6, a length of rupture of 280 km, and a relative hori- zontal displacement of 110 cm. The February 4 earthquake had a magnitude of 7.5, a length of rup- ture of 240 km, and an average relative horizontal displacement of 100 cm. THE GUATEMALAN EARTHQUAKE OF FEBRUARY 4, 1976, A PRELIMINARY REPORT DAMAGE AND ENGINEERING IMPLICATIONS By RAUL HUSID, ALVARO F. ESPINOSA, and ANTONIO QUESADA 5 INTRODUCTION In this report, we discuss the damage done by the February 4 earthquake and the engineering impli- cations in greater detail. We report on the damage to selected structures in the capital city and on a few structures in the rest of the affected area but do not attempt to include all the important failures. EARTHQUAKE-RESISTANT DESIGN PRACTICE IN GUATEMALA When the February 4 earthquake occurred, no earthquake-resistant-design code had been enacted into law in Guatemala, and therefore it was not mandatory to design structures to withstand seis- mic forces. Each engineer or architect selected a foreign code and designed accordingly (J. Asturias, oral commun., 1976). The same professional was usually in charge of supervising the construction process. Review of design and construction by specialized structural engineers, independent of the original designer, was not required, as it is in Chile, Mexico, and the United States. According to two local structural engineers, J. Arias and R. Zepeda (oral commun., 1976), many professionals used elements of a version (not neces- sarily the latest) of the Structural Engineers As- sociation of California code. Thus, the structures in Guatemala City were not designed according to common standards. In the short time available for the study, it was difficult to assess, from the con- dition of the buildings, whether the various stand- ards employed are suitable for the local soil con- ditions, quality of construction materials, dynamic characteristics of the structures, and other im- portant factors. It is noteworthy to mention that the material characteristics, such as the strength of steel reinforcing bars, are frequently assumed by the engineer without any supporting technical evi- dence. 5 Organization of American States, Washington, DC. TYPES OF STRUCTURES Guatemala City has many modern buildings; most are reinforced concrete, but a few are high-rise steel structures. The predominant type of modern con- struction appears to be the reinforced-concrete frame structure having flat beams in one or two directions and masonry (reinforced or unreinforced) filler walls. It is common to find filler walls made of poorly reinforced hollow brick or hollow tile. One of the most common forms of construction is adobe, which is used for the majority of houses, churches, and small structures throughout the country. Roofs are generally tile on wood-pole raft- ers. Reinforced mud or bajareque construction is also used extensively in Guatemala. It consists of a wood frame covered with lath, the wall space being filled with mud and plastered. Bajareque is similar to quincha, which is frequently used for building houses in the coastal region of Peru. Quincha con- struction sustained extensive damage in the 1970 Peruvian earthquake (Husid and Gajardo, 1970; Berg and Husid, 1971, 1973). Wooden construction was common in Puerto Ba- rrios and in the port of Santo Tomas. Corrugated- steel and reinforced—concrete grain silos were used in the area affected by the earthquake. Water tanks were predominantly elevated and built of reinforced concrete or steel. DAMAGE SURVEY Although the capital city was not damaged as severely as towns along the Motagua River Valley and some towns in the highlands west 'of Guate- mala City, there was extensive damage in several zones, and some reinforced-concrete and steel struc- tures completely collapsed. The types of construction found outside Guate- mala City are adobe, bajareque, and wood. Adobe construction in many towns sustained the same 67 68 GUATEMALAN EARTHQUAKE OF FEBRUARY 4, 1976, A PRELIMINARY REPORT FIGURE 51.—Damage to a wooden structure in Puerto Barrios, caused in part by ground compaction. Note the large offset of 11 cm shown in the photograph. heavy damage that has been observed after many previous earthquakes in other countries (Husid and Gajardo, 1970; Eisenberg and others, 1972; Husid and Espinosa, 1975; Espinosa and others, 1975). Wooden construction withstood damage well even when extensive damage was caused by ground com- paction (Ray Wilson, oral commun., 1976) beneath the building (fig. 51). Many school buildings were severely damaged by this earthquake, and, if the earthquake had occurred during class time, the death toll would have been larger. The second story of a three-story reinforced— concrete frame structure with masonry walls at the Colegio San Javier collapsed (fig. 52). In the same school complex, a second building, next to the one that partially collapsed, was extensively damaged. There was no available information about the lateral loads used in the design of the school struc- tures. The Instituto Guatemalteco Americano, a five- story reinforced—concrete frame structure with poorly reinforced hollow brick walls sustained ex- tensive damage. This structure has rather large cantilevered slabs in its perimeter supporting very heavy concentrated loads (reinforced-concrete orna- FIGURE 52,—Collapse of the second story of a building at the Colegio San Javier (Zone 12) , Guatemala City. ments and hollow brick walls) at their free end. Most of the walls were on the verge of collapse, and the slabs showed severe cracks in the canti- levered area. A slab on the penthouse partially col- lapsed, and reinforced—concrete columns and beams showed severe damage at the same level. It is im- portant to note that this school building, which sometimes houses more than 2,000 students, has only one stairway. If the earthquake had occurred when 2,000 students were attending classes, many could have been injured as a result of panic and lack of adequate exits. A three-story framed reinforced-concrete struc- ture (fig. 53) was partially collapsed (Zone 12) when columns on its second floor failed. Severe damage to several hospitals in Guatemala City created a serious problem because of the large number of injured people therein. Included were the Hospital Neuro-Psiquiatrico (Zone 7), Sanatorio San Vicente (Zone 7), Hospital Roosevelt (Zone 11), and the Nursery School of Casa del Nifio No. 1 (Zone 1). The Cathedral of Guatemala City, which was par- tially destroyed during the 1917—18 earthquakes DAMAGE AND ENGINEERING IMPLICATIONS 69 FIGURE 53.—Partial collapse of a three-story reinforced-con- crete structure due to failure of columns in its second floor (Zone 12), Guatemala City. (Penney, 1918; Seismological Society of America Bulletin, 1911—1975, vol. 8, no. 1, p. 38—39), was ex- tensively damaged, and both front towers were on the verge of collapse. A new church, Iglesia del Divino Redentor (fig. 54), sustained extensive damage, including collapsed roof and walls (Zone 11). The roof was supported from the bottom chord of steel trusses that failed and caused the collapse of the walls. The overlapping of vertical steel bars was done at the same level, and the separation of ties contributed to the weak- ness of the structure. Other churches were also seriously affected in Guatemala City, and some were completely or partially collapsed. The Hotel Terminal (Zone 4), a reinforced-con- crete frame, flat-slab, six-story building, collapsed when several columns at the third floor failed (fig. 55). Simple ties were used, and they were widely spaced. This column failure appears to be similar to failures seen in the upper story of the Student Union Building of the Universidad Agraria Na- cional in Lima, Peru, after the October 3, 1974, earthquake (Husid and others, 1976). Numerous gas-station structures, designed in the form of inverted pendulums, collapsed when the welding failed (fig. 56). The International Airport building sustained minor damage, and, in one area, ground compaction occurred beneath the western part of the structure. An inspection of the airport building was made at the request of the airport commander. The two side concourses (fingers) off the main structure were designed as two-story inverted-pendulum-type struc- tures. These fingers had one stair exit, a severe haz- ard in case of fire or earthquake. Various elevated steel water tanks collapsed in Guatemala City and vicinity. Figure 57 shows one type of tank that suffered complete failure. Figure 58 shows the engineering plan of another type of elevated steel water tank (50,000-gallon capacity) that was used extensively throughout the city and failed in several locations. Three water tanks of this design that were full of water at the time of the earthquake collapsed (fig. 59). Two other tanks of this design that were empty sustained extensive damage; the anchor bolts had sheared, and most of the diagonal bracing failed. A few corrugated-steel grain silos collapsed (fig. 60), and some that remained standing sustained ex- tensive damage Where they were connected to the foundation. An inspection was made of the XAYA—PIXCAYA project, which, when it is completed, will provide water for Guatemala City. The sedimentation tank and the Plant Lo De Coy are located near Mixco. The main event, or possibly its aftershocks, created a new system of fractures that crosses beneath the sedimentation tank and other structures (fig. 61). These fractures may be related to renewed move- ments on an old fault. The existence of steep slopes in this area creates a potential for landslides, es- pecially if cracks should develop in the sedimenta- tion tank and water should leak into the ground. Surface breakage on north- to northeast-striking secondary faulting (Zone 19) in Colonia San Fran- cisco locally generated extensive damage to local residences (fig. 62). In San J osé Rosario, a subdivision of Guatemala City, surface breakage with both vertical and right- lateral displacement occurred on a secondary fault parallel with the San Francisco fault. This surface rupture was in an area where plans have been made for construction of high-cost new homes (fig. 63). Three central spans of the Agua Caliente Bridge, about 36 km northeast of Guatemala City, collapsed, but the piers remained undamaged. The spans were pinned at one end and had rocking rollers at the other. The failure appears to have been generated by the local failure of the supports caused by large 70 GUATEMALAN EARTHQUAKE OF FEBRUARY 4, 1976, A PRELIMINARY REPORT FIGURE 54.—Collapse of a new church, Iglesia del Divino Redentor, Guatemala City (Zone 11). displacements of the deck (fig. 64). An old steel- truss railroad bridge near the Agua Caliente Bridge sustained no damage. The Benque Viejo multiple-span steel—truss bridge across the Platanos River was near collapse because of the large relative displacements between the trusses and their supports. The rollers have snapped flat between the upper and lower bearing plates (fig. 65). GENERAL OBSERVATIONS During the damage survey, it was found that similar problems recurred many times. Some of them are briefly described below. The use of nonstructural masonry walls in rein- forced-concrete framed structures was frequently neglected in the design, as has been observed in previous earthquakes, for example, the 1970 and 1974 Peruvian earthquakes (Husid and Gajardo, 1970; Berg and Husid, 1973; Husid and others, 1976). When lateral displacements occur in struc- tures, resisting elements are loaded in proportion to their stiffnesses. Hence, short columns will be loaded with far greater shear forces than long columns. A column having a free height H has a stiffness approximately eight times greater than that of a column of equal cross section and height 2H. Hence, the short column carries a lateral load approximately eight times larger than that carried by the larger column. One of the many examples of this kind of failure observed in Guatemala is shown in figure 66. DAMAGE AND ENGINEERING IMPLICATIONS 71 FIGURE 55.—-Collapse of the Hotel Terminal, caused by the failure of reinforced-concrete columns in its third story. This building is located in Guatemala City (Zone 4). The lack of reinforced-concrete columns framing masonry walls has been shown to be responsible for heavy losses during earthquakes. Experimental studies concerning the behavior of masonry walls subjected to lateral loads have shown that, when such walls are designed with integral framing or edge members, their behavior is almost ductile, even after cracking. When the masonry walls are not so framed, their lateral failure is extremely brittle and sudden, even for smaller lateral loads (Jor- quera, 1964). A common practice in Guatemala is to build non- structural unreinforced masonry walls in tall struc- tures. The behavior of these walls was very poor, as in previous earthquakes. An example is shown in figure 67, where brick falling from the 10th and 11th stories of a building destroyed the slab roof and part of the supporting structure of the first floor. A similar problem was reported by Husid, Espinosa, and de las Casas (1976) at the Industrial Bank of Peru in downtown Lima, after the October 3, 1974, earthquake. Several one-story reinforced-concrete framed structures collapsed completely, and the quality of concrete did not seem to be a factor that could justify the failure. It was observed that distances between ties in columns were rather large and ex- ceeded the American Concrete Institute specifica- tions, especially close to the slab and ground levels (fig. 68). Heavy parapets located in the upper part of front facades, which are severe hazards for the popula- tion during earthquakes, collapsed in many areas throughout the capital city. 72 GUATEMALAN EARTHQUAKE OF FEBRUARY 4, 1976, A PRELIMINARY REPORT FIGURE 56.—Collapse of inverted-pendulum gas-station structure, Guatemala City (Zone 7). Although chimneys are not common on homes in Guatemala City because of the mild weather, several of those that did exist collapsed. Adobe construction is not earthquake resistant; some examples of the unsatisfactory performance of adobe is shown in figures 69A, B, C, and D. Con- sidering the fact that it is not economically feasible to eliminate adobe construction in Guatemala, it would be desirable to make an inventory of the dif- ferent types of adobe used countrywide and pre- pare recommendations for simple modifications that might improve the strength of adobe houses sub- jected to earthquakes. A common design for multistory structures in Guatemala City utilizes principal frames in the transverse direction only. The horizontal loads in the longitudinal direction are resisted with a “pseudorigid” frame in which the normal beams are replaced by wide strips of slab (flat beams) that join the tops of the columns. As an example, figure 70 shows a five-story reinforced-concrete structure that sustained serious damage when flat beams were almost destroyed; the building had to be evacuated. Buildings of this design are usually very flexible and have long fundamental periods. It is probable that the behavior of the multistory build- ings would have been less satisfactory if they had been subjected to a stronger ground shaking or to shaking of similar amplitude but longer periods. A close look at the effects of the February 4 earthquake in Guatemala City shows the following aspects: 1. Strengths of reinforced-concrete lateral-load- resisting elements were often unrelated to their stiffnesses. 2. Masonry filler walls in multistory buildings lacked minimum reinforcement. 3. Numerous heavy parapets collapsed and created a serious hazard in Guatemala City. 4. Reinforced-concrete column ties were frequently too widely spaced and sometimes not ade- quately hooked. 5. Brick walls often lacked reinforced-concrete corner columns, and long walls also lacked intermediate reinforced-concrete columns. 6. Surface breakage on secondary faulting oc- curred in developed areas. 7. Adobe construction sustained heavy damage. Adobe houses did not have any edge mem- bers. 8. Heavy roofs of adobe and unreinforced masonry houses frequently collapsed. 9. Elevated water tanks often failed. From the number of collapses, types of connections, and sizes of resisting elements, it is suspected that the lateral-force seismic coefl‘icient used for the design was too low. 10. Corrugated-steel grain silos frequently failed, and several collapses were observed. DAMAGE AND ENGINEERING IMPLICATIONS 73 mm “‘ J FIGURE 57.—Collapse of an elevated steel water tank in the Instituto Técnico Vocacional in Guatemala City (Zone 13). Foundation Plont SCALE |=75 Tank Cross Section r————1 0.6m 0.6m Hé—6JOm——>i-+i I.I_§.m |.|2m ? i, i l 6.|Om D 1-0.65cm 1 n ‘ 6.|Om k1=o.65cm _ EDDDDDD I22m 1:0.65cm 1.22m _L____ 0.4m |5.0m 6"0’" ’ Plate t=|.3cm 9b 27.3 cm I gb lcm at I6cm .4 2 im eqb 2.2 cm H—8.|5m———>t Detail of Footing Elevation View FIGURE 58.—Engineering plans of an elevated water tank sho (4), diameter.) ll .2cm Both Directions ivn inflizgure 5%.m 74 GUATEMALAN EARTHQUAKE OF FEBRUARY 4, 1976, A PRELIMINARY REPORT w m» ‘mn.»...\wm,~mgwéwg FIGURE 60.—Collapse of a corrugated—steel grain silo in Villalobos, 5 km southeast of Guatemala City. DAMAGE AND ENGINEERING IMPLICATIONS 75 l I I l |<——200m————’————<>| } I \ \ l \‘ Sedimentation Tank \ \ \ Operations ‘ Building % ‘\ W. .__. _. ,___... DE! :IC‘JIZIE 200m A FIGURE 61.—P1an of sedimentation tank and Plant Lo de Coy of the XAYA-PIXCAYA project and cracks (heavy lines) running under and across the sedimentation tank. FIGURE 62,—Severe damage to reinforced-masonry construction} caused by secondary faulting in Colonia San Francisco (Zone 19). The fault strikes north to northeast. 76 GUATEMALAN EARTHQUAKE OF FEBRUARY 4, 1976, A PRELIMINARY REPORT FIGURE 63.—Rupture across the San J osé Rosario subdivision. Fault with 13 cm vertical and 5 cm right—lateral displace— ments (Zone 19). FIGURE 64.—Collapse of three central spans of the Agua Caliente Bridge, Kilometre 36 on the road to the At- lantic Ocean. This bridge was constructed in 1959. FIGURE 66.—Fai1ure of short column in three-story framed , reinforced-concrete structure in Guatemala City (Zone after fallure of supports, shown by arrow. 7) FIGURE 65.—Benque Viejo Bridge, on the verge of collapse DAMAGE AND ENGINEERING IMPLICATIONS 77 FIGURE 67.—Partia1 collapse of protruding first-story structure of the Cruz Azul ll—story reinforced-concrete building caused by falling (shown by arrows) of masonry walls from the two topmost stories, in Guatemala City (Zone 1). FIGURE 68,—Collapse of one-story reinforced-concrete framed structure in the Licorera Mixco, Mixco. 78 GUATEMALAN EARTHQUAKE OF FEBRUARY 4, 1976, A PRELIMINARY REPORT FIGURE 69.—A, Massive destruction of some adobe houses in Antigua, Guatemala, near the center of town. B, Large-scale de- struction of adobe houses in Guatemala City, Calle 2 and Avenida 9A (Zone 2). C, Collapse of adobe houses in Guate- mala City, Calle 22A and 34 Avenida (Zone 5). D, Collapse of adobe houses in Guatemala City, Calle 22 and Avenida 32 (Zone 5). DAMAGE AND ENGINEERING IMPLICATIONS 79 FIGURE 70.——Severe damage to flat beams and slab in the five-story reinforced-concrete Edificio ELGIN, 1n Guatemala City (Zone 14). THE GUATEMALAN EARTHQUAKE OF FEBRUARY 4, 1976, A PRELIMINARY REPORT STRUCTURAL ENGINEERING OBSERVATIONS IN GUATEMALA CITY By KARL V. STEINBRUGGE INTRODUCTION This paper reports on earthquake damage to buildings in Guatemala City and places special em- phasis on earthquake-resistant structures and other building types having engineering relevance to construction in the United States. The author inspected in detail the interiors and exteriors of about 25 major structures and, in a few instances, the construction drawings. His engineer- ing colleagues, who formed a team under the aus— pices of the Earthquake Engineering Research In- stitute, collaborated on data collection, and they have had some input to this report. The views ex- pressed in this report, however, are those of the author. This report is preliminary, and some details may change With further evaluation upon comple- tion of the data collected. LIFE LOSS AND CONSTRUCTION The population of Guatemala City is about 700,000, about 1,200 of whom may have died in the earthquake, according to one source (table 9). The fatality rate of 1 person in about 600 for Guatemala City contrasts sharply With the very high rates for some of the villages and small cities located much closer to the fault rupture. The deaths in Guate- mala City resulted essentially from the collapse of adobe construction; this type of construction seems to be more prevalent in the northern and north- western sections of the city. These zones were closer to the fault and to the source of seismic energy. In contrast to the damage patterns found in some loca- tions near the fault (Espinosa and others, this re- port), widespread flattening of city blocks of adobe construction was not found in Guatemala City, which lies about 25 km from the fault break (Plaf- ker and others, this report). In contrast to the heavy mass construction such as adobe, light-mass all-metal structures, including 80 one-story warehouses and aircraft hangars, per- formed excellently. Any comparative wind versus earthquake analysis would show that, if these light- mass structures can withstand a moderately strong Windstorm, they can survive a major earthquake. DESIGN AND CONSTRUCTION PRACTICES As in many Latin American countries, Guate- mala has no effective building code or enforcement procedures comparable to those found in the United States. Each design professional is allowed to estab- lish his own design criteria and supervise his own construction without independent scrutiny, on the basis of his status as a registered professional. This practice does not preclude good construction, but, in effect, it places no limitations on the extent of poor construction. On the other hand, the local struc- tural engineers and architects are, in general, ex- cellent design professionals and comparable to their US. counterparts. Many Guatemalan professionals have been competently trained, as was seen from a brief review of several sets of construction draw- ings and from discussions with them. This is not to say, however, that poor design does not exist. The quality of construction appears to be gen- erally good for the newer buildings. High-strength concrete seems to be common. The quality of the reinforcing steel, however, may be open to ques- tion in some instance, since this steel does not ap- pear to be locally tested and the strength character- istics (and other qualities) of imported steel often are not known. There apparently are only two structural-steel multistory buildings in Guatemala City, the Finance Ministry Building (fig. 71) and the National Thea- ter (which is only partially steel frame). In these two known instances, special quality-control efforts were made during fabrication and erection, and no significant damage occurred to them. STRUCTURAL ENGINEERING OBSERVATIONS IN GUATEMALA CITY 81 FIGURE 71.—The Finance Ministry Building (in Guatemala City) has 19 stories above ground plus 3 below. The undamaged steel-frame building is in the final stages of construction. (Photograph by W. H. Smith, American Iron and Steel Institute.) Reinforced-concrete floors and roofs in major multistory buildings normally were of “waffle” con- struction. Forms for the waffles were customarily wooden rather than metal or plastic. Lateral-force resistance was commonly provided by frame action that utilized the waffle slabs and concrete columns, but some designs utilized reinforced-concrete shear walls. Exterior walls were normally of reinforced-unit- masonry construction and were considered to be nonstructural for design purposes. These unit- masonry panel walls were strengthened by the in- clusion of reinforced-concrete-bond beams and equivalent vertical members. This bracing system did what it was designed to do. There was no major fallout of these panels where the building remained intact. The exterior wall panels that fell from the Seguros Building were not reinforced. Since the large majority of multistory buildings have been built in recent years, design and con- struction practices are modern. Guatemalan prac- tices generally follow U.S. practices, and therefore many Guatemalan buildings will perform like thoSe in the United States. For the usual building design, the Uniform Building Code appears to be the norm for seismic design. SPECIFIC DAMAGE OBSERVATIONS Tall buildings that were in close contact pounded together. Structural damage was minimal in most cases, and architectural damage was usually not excessive. However, whenever a building consisted of two independent structural units, such as a stair and/or elevator tower that was structurally inde- pendent from the rest of the building, then the elec- trical, plumbing, and mechanical equipment, and so on was broken at the floor lines, and the functional capacity of the building was reduced. One example of pounding damage was observed between the stair tower and the main structure of the 13-story Segu- ros Building where, at the roof, a 7-cm separation was observed. Folded plates and shells performed well, and no instances of significant internal damage are known. (The author, however, understands that there were one or more instances of such damage, but they re- main unconfirmed.) The International Airport shown in figure 72 is one example of a folded-plate structure where minor spalling occurred at the tops and bases of columns, footing rocking was observed, some glass broke (see open windows in fig. 72), structural separations showed significant differen- tial movements, and the stairs acting as diagonal braces were damaged, but the folded-plate roof sus- tained no damage. A second example of folded plates and shells is 'the Universidad del Valle de Guatemala y Colegio Americano de Guatemala. The one-story hyperbolic paraboloid roofs performed well, but certain rein- forced-concrete columns supporting these roofs were badly damaged. This damage can be attributed to the “structural performance of non-structural” brick panel walls that framed against these columns. These well—built brick walls did not fail, but they stiffened the columns to the point that they re— sisted a disproportionate share of the lateral forces. Clearly, a better understanding of the importance of “non-structural” walls is vital to Guatemalan engineers as well as to many US. architects and engineers. Brick-infill walls between exterior col- umns as well as brick interior “partitions” used for tenants’ improvements within a building were often well constructed and of rather substantial strength. Not only did these “non—structural” walls change the dynamic characteristics of the building, but often they also led to column failure or other sig- nificant damage. Another example of this kind of damage was observed at the two-story administra- tion building adjacent to the airport control tower. 82 GUATEMALAN EARTHQUAKE OF FEBRUARY 4, 1976, A PRELIMINARY REPORT FIGURE 72.—Internationa1 Airport at Guatemala City. Note that some windows are broken. FIGURE 73.—Hote1 Terminal in Guatemala City. Columns in second story collapsed. Restated, a brick “non-structural” panel wall is an effective shear wall in a concrete frame as long as the “non-structural” wall remains intact. Few instances of precast-concrete construction were inspected. One hospital structure, being a one-story “shear wall” building, had no damage to its roof, which consisted of a long span of precast concrete planks (“Spancrete”). There was no inter- connection between the planks. A precast-concrete double-tee roof at the Universidad del Valle col- lapsed, but this building was in the course of con- struction, and its final bracing system was incom- plete. In general, building performance ranged from good to excellent, but several spectacular failures oc- curred. For example, in the Hotel Terminal (fig. 73), columns failed in the second story (fig. 74). Another example is the dormitory-classroom unit of the Catholic Boys School (Colegio San Javier), which collapsed owing to failure of its second-story columns (fig. 75). Figure 76 shows a collapsed column at the second floor of this unit. The time of night at which this earthquake occurred pre- vented major life loss in this building, since the first and second floors are used as classrooms. A large number of hospitals were evacuated in Guatemala City owing to structural damage and to functional impairments. Functional problems caused in multistory build- ings were commonly in the form of elevator outages, and most elevators were still nonoperational 2 weeks after the event. Standby power remained in service STRUCTURAL ENGINEERING OBSERVATIONS IN GUATEMALA CITY 83 FIGURE 75.—Collapsed second story in one unit of the Catholic Boys School in Guatemala City. Roofs were originally at the same level. 84 GUATEMALAN EARTHQUAKE OF FEBRUARY 4, 1976, A PRELIMINARY REPORT FIGURE 76.—Detail of column failure in second story at the Catholic Boys School. Story height now measured in inches. at almost all the locations inspected by the author, but, undoubtedly, instances of failure did occur. Batteries required for standby power generally did not shift or, if they shifted, did not fall or break their connecting cables. The nearest American counterparts to Guate- malan adobe buildings are the brick structures in the older sections of almost every US. city, includ- ing San Francisco and Los Angeles. These non- earthquake-resistant brick buildings, with their sandlime mortar that “holds brick apart,” are a major potential source of loss of life. REFERENCES CITED 85 Aburto, Arturo, 1975, Reporte de los temblores ocurridos en Nicaragua los meses de Marzo a Junio de 1975: Buletin Instituto de Inv. Sismicas, v. 1, no. 1, 38 p. Aki, Keiiti, 1966, Generation and propagation of G waves from the Niigata earthquake of June 16, 1964, part 2, estimation of earthquake moment, released energy, and stress-strain drop from G wave spectrum: Tokyo Univ. Earthquake Research Inst. Bull., v. 44, p. 73—88. Ambraseys, N. N., and Zétopek, A., 1968, The Varto-Ustukran (Anatolia) earthquake of 19 August 1966; summary of field report: Seismol. Soc. America Bull., V. 58, no. 1, p. 47—102. Beniofl‘, Hugo, 1955, Mechanism and strain characteristics of the White Wolf fault as indicated by the aftershock se- quence, in Oakeshott, G. B., ed., Earthquakes in Kern County, California, during 1952: California Div. Mines and Geology Bull. 171, p. 199—202. Berg, G. V., and Husid, L. R., 1971, Structural effects of the Peru earthquake: Seismol. Soc. America Bull., v. 61, no. 3, p. 613—631. 1973, Structural behavior of the 1970 Peru earthquake, in World Conf. on Earthquake Eng., 5th, Proc., V. 1: Rome, Edigraf, p. 458—467. Bonis, Samuel, 1976, The 1974 eruption of Volcan de Fuego, Guatemala [abs.]: EOS Am. Geophys. Union Trans., v. 57, no. 4, p. 345. Bonis, Samuel, Bohnenberger, O. H., and Dengo, Gabriel, 1970, Mapa Geolégico de la Repfiblica de Guatemala [lst ed.]: Guatemala Instituto Geografico Nacional. Bowin, C. 0., 1968, Geophysical study of the Cayman trough: Jour. Geophys. Research, v. 73, no. 16, p. 5159—5173. Brigham, W. T., 1887, Guatemala, the land of the Quetzal: New York, Scribner, 453 p. Brown, R. D., Ward, P. L., and Plafker, George, 1973, Geologic ’and seismologic aspects of the Managua, Nicaragua, earth- quakes of December 23, 1972: U.S. Geol. Survey Prof. Paper 838, 34 p. Brune, J. N., 1968, Seismic moment, seismicity, and rate of slip along major fault zones: Jour. Geophys. Research, v. 73, no. 2, p. 777—784. 1970, Tectonic stress and the spectra of seismic shear waves from earthquakes: Jour. Geophys. Research, v. 75, no. 26, p. 4997—5009. Brune, J. N., and Allen, C. R., 1967, A microearthquake sur- vey of the San Andreas fault system in southern Cali- fornia: Seismol. Soc. America Bull., v. 57, no. 2, p. 277—296. Brune, J. N., and Engen, G. R., 1969, Excitation of mantle Love waves and definition of mantle wave magnitude: Seismol. Soc. America Bull., v. 59, no. 2, p. 923—933. Burridge, R., and Knopofl", Leon, 1964, Body force equivalents for seismic dislocations: Seismol. Soc. America Bull., v. 54, no. 6, p. 1875-1888. Carey, S, W., 1958, The tectonic approach to continental drift, in Continental drift, a symposium: Hobart, Univ. of Tasmania, p. 177—355. Carr, M. J ., 1974, Tectonics of the Pacific margin of northern Central America: New Hampshire, Dartmouth 0011., Ph.D. dissert., 181 p. Chinnery, M. A., 1963, The stress changes that accompany strike-slip faulting: Seismol. Soc. America Bull., v. 53, no. 5, p. 921—932. Dengo, Gabriel, 1968, Estructura geologica, historia tectonica, y morfologia de America Central: Mexico, Centro Re- gional de Ayuda Técnica, Agencia para el Desarrollo In- ternacional, 50 p. 1969, Problems of tectonic relations between Central America and the Caribbean: Gulf Coast Assoc. Geol. Socs., 19th Ann. Mtg., Trans., p. 311—320. Dengo, Gabriel, and Bohnenberger, Otto, 1969, Structural de- velopment of northern Central America, in McBirney, A. R., ed., Tectonic relations of northern Central America and the western Caribbean: Am. Assoc. Petroleum Geol- ogists Mem. 11, p. 203—220. Dewey, J. W., 1976, Seismicity of northern Anatolia: Seismol. Soc. America Bull. [In press] Dewey, J. W., and Algermissen, S. T., 1974, Seismicity of the Middle America arc-trench system near Managua, Nicaragua, in Special Papers on the Managua, Nicaragua, earthquakes of December 23, 1972: Seismol. Soc. America Bull., v. 64, no. 4, p. 1033—1048. Eaton, J. P., O’Neill, M. E., and Murdock, J. N., 1970, After- shocks of the 1966 Parkfield-Cholame, California, earth- quake—a detailed study: Seismol. Soc. America Bull., v. 60, no. 4, p. 1151—1197. Eisenberg, A., Husid, Raul, and Luco, J. E., 1972, A pre- liminary report—the July 8, 1971, Chilean earthquake: Seismol. Soc. America Bull., v. 62, no. 1, p. 423—430. Engdahl, E. R., 1971, Explosion effects and earthquakes in the Amchitka Island region: Science, v. 173, no. 4003, p. 1232—1235. 1973, Relocation of intermediate depth earthquakes in the central Aleutians by seismic ray tracing: Nature Phys. Sci., v. 245, no. 141, p. 23—25. Espinosa, A. F‘., Husid, R., and Algermissen, S. T., 1975, Seismological and engineering features of the October 3, 1974, Lima earthquake [paper 141]: European Conf. on Earthquake Eng, 5th, Istanbul 1975, Proc., v. 2. Espinosa, A. F.. Husid,- R., Algermissen, S. T., and de 1as Casas, J ., 1976, The Lima earthquake of October 3, 1974, intensity distribution: Seismol. Soc. America Bull. [In press] Fahquist, D. A., and Davies, D. K., 1971, Fault-block origin of the western Cayman Ridge, Caribbean Sea: Deep-Sea Research, v. 18, no. 2, p. 243—253. Freeland, G. L., and Dietz, R. S., 1971, Plate tectonic evolu- tion of Caribbean-Gulf of Mexico region: Nature, v. 232, no. 5305, p. 20—23. Gill, J. E., 1972, Continued confusion in the classification of faults; Reply: Geol. Soc. America Bull., v. 83, no. 3, p. 925—926. Gutenberg, Beno, and Richter, C. F., 1954, Seismicity of the Earth [2d ed.]: New York, Hafner, 310 p. Hanks, T. C., 1975, Strong ground motion of the San Fer- nando, California, earthquake—ground displacements: Seismol. Soc. America Bull., v. 65, no. 1, p. 193—225. Hess, H. H., and Maxwell, J. C., 1953, Caribbean research project: Geol Soc. America Bull., v. 64, no. 1, p. 1—6. Holcombe, T. L., Vogt, P. R., Matthews, J. E., and Murchison, R. R., 1973, Evidence for sea-floor spreading in the Gay- man Trough: Earth and Planetary Sci. Letters, v. 20, no. 3, p. 357—371. Husid, R., and Espinosa, A. F., 1975, Seismological and engin- eering aspects of the March 6, 1974, Carazo earthquake (Nicaragua) [paper 137] : European Conf. on Earthquake Eng., 5th, Istanbul 1975, Proc., v. 2. 86 GUATEMALAN EARTHQUAKE OF FEBRUARY 4, 1976, A PRELIMINARY REPORT Husid, R., and Gajardo, E., 1970, Aspectos sismologicos y estructurales en el terremoto del Peru del 31 de Mayo de 1970: Simposio Panamericano de Estructuras, Buenos Aires, Argentina. Husid, R., Espinosa, A. F., and de las Casas, J ., 1976, The Lima earthquake of October 3, 1974, damage distribution: Seismol. Soc. America Bull. [In press] Jordan, T. H., 1975, The present-day motions of the Carib- bean plate: Jour. Geophys. Research, v. 80, no. 32, p. 4433—4439. Jorquera, L., 1964, Estudio experimental sobre la resistencia de muros de Albafiileria :sometidos a cargas horizontales: Rev. Idiem, v. 3, no. 3., p. 193—222. Kanamori, Hiroo, and Anderson, D. L., 1975, Theoretical basis of some empirical relations in seismology: Seismol. Soc. America Bull., v. 65, no. 5, p. 1073—1095. Keilis-Borok, V. I., 1959, On estimation of the displacement in an earthquake source and of source dimensions: Annali Geofisica, v. 12, p. 205—214. Kelleher, John, Sykes, Lynn, and Oliver, Jack, 1973, Possible criteria for predicting earthquake locations and their ap- plication to major plate boundaries of the Pacific and the Caribbean: Jour. Geophys. Research, v. 78, no. 14, p. 2547—2585. Kesler, S. E., 1971, Nature of ancestral orogenic zone in nu- clear Central America: Am. Assoc. Petroleum Geologists Bu11., v. 55, no. 12, p. 2116—2129. King, Cyi-yu, and Knopofl', Leon, 1968, Stress drop in earth- quakes: Seismol. Soc. America Bull., v. 58, no. 1, p. 249— 257. King, P. B., compiler, 1969‘, Tectonic map of North America: Reston, Va., U.S. Geol. Survey, 2 sheets. Knopoff, Leon, 1958, Energy release in earthquakes: Geophys. Jour., v. 1, no. 1, p. 44—52. Koch, A. J., and McLean, Hugh, 1975, Pleistocene tephra and ash—flow deposits in the volcanic highlands of Guatemala: Geol. Soc. America Bull., v. 86, no. 4, p. 529—541. Lee, W. H. K., and Lahr, J. C., 1975, HYPO71 (revised)—a computer program for determining hypocenter, magnitude, and first motion pattern of local earthquakes: U.S. Geol. Survey open-file rept., 113 p. Lee, W. H. K., Bennett, R. E., and Meagher, K. L., 1972, A method of estimating magnitude of local earthquakes from signal duration: U.S. Geol. Survey open-file rept., 28 p. Malfait, B. T., and Dinkelman, M. G., 1972, Circum-Caribbean tectonic and igneous activity and the evolution of the Caribbean plate: Geol. Soc. America Bull., v. 83, no. 2, p. 251—271. McBirney, A. R., 1963, Geology of a part of the Central Guatemalan Cordillera: Calif. Univ. Pubs. Geol. Sci., v. 38, no. 4, p. 177—242. McBirney, A. R., and Bass, M. N., 1969, Structural relations of pre-Mesozoic rocks of northern Central America, in Tectonic relations of northern Central America and the western Caribbean: Am. Assoc. Petroleum Geologists Mem. 11, p. 269-280. Minster, J. B., Jordan, T. H., Molnar, Peter, and Haines, E., 1974, Numerical modelling of instantaneous plate tec— tonics: Roy. Astron. Soc. Geophys. Jour., v. 36, no. 3, p. 541—576. Mitronovas, Walter, Isacks, Bryan, and Seeber, Leonardo, 1969, Earthquake locations and seismic wave propagation in the upper 250 km of the Tonga Island arc: Seismol. Soc. America Bull., v. 59, no. 3, p. 1115—1135. Molnar, Peter, and Sykes, L. R., 1969, Tectonics of the Carib- bean and Middle America regions from focal mechanisms and seismicity: Geol. Soc. America Bull., v. 80, no. 9, p. 1639-1684. Montessus de Ballore, Fernard de, 1888, Tremblements de terre et éruptions volcaniques au Centre-Amérique: Dijon, Societé de sciences naturelles de Sa6ne~et-Loire, 293 p. Morgan, W. J ., 1971, Convection plumes in the lower mantle: Nature, v. 230, no. 5288, p. 42—43. Page, R. A., ed., Interim report on the Guatemalan earthquake of 4 February 1976 and the activities of the U.S. Geologi- cal Survey earthquake investigation team: U.S. Geol. Survey open-file rept., 31 p. Penney, W. T., 1918, Guatemala earthquake shifts steel via- duct on its piers: Eng. News-Record, v. 80, no. 21, p. 999. Pinet, P. R., 1972, Diapirlike features offshore Honduras; implications regarding tectonic evolution of Cayman Trough and Central America: Geol. Soc. America Bull., v. 83, no. 7, p. 1911—1921. Richter, C. F., 1958, Elementary seismology: San Francisco, W. H. Freeman and Co., 768 1). Savage, J. C., 1971, A theory of creep waves propagating along a transform fault: Jour. Geophys. Research, v. 76, no. 8, p. 1954—1966. Seismological Society of America, 1911—1975, Seismological notes: Seismol. Soc. America Bull., v. 1—65, pagination varies. Stoiber, R. E., and Carr, M. J ., 1973, Quaternary volcanic and tectonic segmentation of Central America: Bull. Volcanol., v. 37, no. 3, p. 304—325. ‘ Tarr, A. C., and King, K. L., 1976, Puerto Rico seismic pro- gram: U.S. Geol. Survey open—file rept., 23 p. Thatcher, Wayne, 1975, Strain accumulation and release mechanism of the 1906 San Francisco earthquake: Jour. Geophys. Research, v. 80, no. 35, p. 4862—4872. Uchupi, Elazar, 1973, Eastern Yucatan continental margin and western Caribbean tectonics: Am. Assoc. Petroleum Geologists Bull., v. 57, no. 7, p. 1075—1085. U.S. Coast and Geodetic Survey, 1960, United States earth- quakes: Washington, D.C., Govt. Printing Office, 90 p. 1965, United States earthquakes: Washington, D.C., Govt. Printing Office, 91 p. Vassaux, P. J ., 1969, Cincuenta afios de sismologia en Guate~ mala: Guatemala City, Observatorio Nacional, p. 85-98. Wallace, R. E., 1968, Earthquake of August 19, 1966, Varto area, eastern Turkey: Seismol. Soc. America Bull., v. 58, no. 1, p. 11—45. Ward, P. L., Endo, E. T., Harlow, D. H., Allen, Tex, Marquez, Dan, and Eaton, J. P., 1974, Development and evaluation of a prototype global volcano surveillance system utilizing the ERTS—l satellite data collection system: U.S. Geol. Survey open~file rep., 154 p. Williams, Howel, and McBirney, A. R., 1969', Volcanic history of Honduras: Calif. Univ. Pubs. Geol. Sci., v. 85, 101 p. Williams, Howe], McBirney, A. R., and Dengo, Gabriel, 1964, Geologic reconnaissance of southeastern Guatemala: Calif. Univ. Pubs. Geol. Sci., v. 50, 62 p. Wilson, J. T., 1963, A possible origin of the Hawaiian Islands: Canadian Jour. Physics, v. 41, no. 6, p. 863—870. CHRONOLOGICAL HISTORICAL RECORD OF DAMAGING EARTHQUAKES IN GUATEMALA, 1526-1976 [Reference sources listed at end of table. Dashes mean no data. h is the depth of focus for the earthquake. Asterisks mark earthquakes plotted in fig. 3. Station codes are given in the Glossary] NONINSTRUMENTAL DATA Date Time (local time) Place Reference 1526 July 19 or 20 ________________ 3 Damaging earthquake that caused removal of Spanish settlement to Ciudad Vieja. 1541 September 10* ________ Ciudad Vieja 1, 3 Two destructive earthquakes killed approximately 150 Spaniards and at least 600 Indians and Negroes. Two days of rain prior to the earthquakes added damage from avalanches and landslides. The primary Spanish settlement then moved to Antigua. 1565 February ________________ 3 Series of violent earthquakes caused extensive damage in and near Antigua. 1575 ______________________ 1, 3 Several large shocks caused damage in San Salvador and also in Antigua, Guatemala. 1577 November 30 ________________ 3 Earthquake swarm. Largest shock caused much damage in Antigua. _ 1586 December 23* ________________ 1, 3, 6 Long earthquake sequence beginning January 16, 1585, and ending with largest shock on above date. Accompanied by eruption of Fuego Volcano. Antigua was destroyed, causing many deaths. 1607 April* ________________ 3,6 Many buildings collapsed, killing a number of people in Antigua. 1651 February 18 13:00 ________ 1,3 Extensive damage in Antigua. 1681 July 22 ________________ 3 A swarm of earthquakes caused extensive damage in Antigua. 1684 August ________________ 3 Earthquake swarm caused notable damage in Antigua area. 1689 February 12 ________________ 1, 3, 6 Earthquake swarm caused extensive damage and loss of life in Antigua area. Stronger than shock of 1651. 1702 August 4* ________________ 3,6 A strong earthquake caused extensive damage in Antigua. 1717 September 29 19:00 and 9:00 ________ 1, 3,6 and 30* Antigua damaged on September 29, destroyed September 30; loss of life was extensive. Large aftershock on October 3. Earthquakes accompanied by violent eruptions of Fuego. 1751 March 4 8200* ________ 1, 3, 6 Antigua damaged. Cathedral dome destroyed. 87 88 CHRONOLOGICAL HISTORICAL RECORD—Continued NON INSTRUMENTAL DATA—Continued Date Time (local time) Place Reference 1765 April 20* ________________ 6 Fifty killed and many injured; many towns destroyed in the De— partment of Chiquimula. Earthquake may have originated on the Motagua fault. 1765 October ________________ 1 The earthquake of “San Rafael” severly damaged many towns in Guatemala. 1773 July 29 15:45* ________ 1, 3, 6 This major event was part of an earthquake swarm beginning in May and continuing until December. Very strong shocks occurred on June 11 at 5:00 and 17:00 hours (local time). Large after- shocks occurred on September 7 and December 14. Antigua was completely destroyed, and many deaths resulted. The capital was then moved to Guatemala City. These earthquakes were felt even more strongly in Chimaltenango and Quezaltenango, nearer to the Motagua fault, and thus the Motagua fault may have been the source of these earthquakes. 1830 April 1 ________________ 3 Swarm, similar to that of 1773, destroyed many buildings in Antigua. Major aftershock on April 23. 1852 May 16 ________________ 1 Damage in the vicinity of Quezaltenango. 1853 February 9* ________________ 3 Major earthquake caused great alarm in Quezaltenango. Also strong- ly felt in Antigua and Amatitlan. 1855 January 1—26 ________________ 3 Swarm with main events on the 18th and 26th. Damage at Cantel and Zunil. 1859 December 8 20:15 ________ 1,3 Major earthquake near E1 Salvador-Guatemala border. Houses were shattered ‘in Escuintla and Amatitlan. Tsunami at Acajutla, El Salvador. 1860 December 19* ________________ 6 Extensive damage to churches and homes in Escuintla. Aftershocks continued until December 31. 1861 August 27 ________________ 3 Damage to homes and churches in Conquaco and J alpatagua. 1862 December 19* ________________ 1,3 Antigua, Amatitlan, Escuintla. Tecpan, and neighboring towns were severely damaged. Damage to many churches and ancient con- structions. Slight damage to old churches in Guatemala City; astronomical observatory reported tilt of 3'29". 1863 December 12‘ ________________ 6 The earthquake, centered near Guatemala City, caused changes in flow of springs in the northern part of the city, and earth frac- tures opened in the areas of Jocotenango and El Bosque, causing much panic throughout the city. 1870 June 12 15:00 ________ 1, 3, 6 Extensive damage in the regions of Chiquimulilla, Cuilapa, and Ix- huatan. A later quake at 18:23 (local time) caused serious damage in Cuilapa. Aftershocks continued until the 23d. 1874 September 3 21:00* ________ 1, 3, 6 (or possibly 18) Antigua, Chimaltenango, and Patzicia were destroyed and 200 people killed. CHRONOLOGICAL HISTORICAL RECORD—Continued NONINSTRUMENTAL DATA—Continued Date Time (local time) Place Reference 1881 August 13 12:30 ________ 3 Earthquake swarm felt in San Marcos; possible damage in Chinigue. 1885 November 22 ________________ 3 Strong shocks; damage at Amatitlan. 1885 December 18 17: 36 Amatitlan destroyed (int‘ensitszX, Rossi-Forel). Cracking of ground; new hot springs on shores of Lake Amatitlan. Many fore- shocks and aftershocks into January 1886. Volcano Pacaya in- creased level of activity. 1902 April 19 20:20 lat 14.0" N., long 91.0° W. 4, 5, 6 Magnitude 8.3 (PAS) Destroyed the city of Quezaltenango. Extensive loss of life. Activity continued until September 23, when an earthquake was strongly felt and an eruption of volcano Santa Maria began. 1912 June 12 12:43 42.0 lat 17.0° N., long 89.0° W. 4 Magnitude 6.8 (PAS) 1913 March 9 9249* Strong earthquake felt in the central region of Santa Rosa and also in the Departments of Cuilapa and Santa Rosa of Lima. Felt over large parts of the country. Many deaths and much destruction. 1915 September 7 01:20 48.0 UTC lat 14.0° N., long 89.0° W. 2, 4, 5 Magnitude 7.9 (PAS) Heavy damage in Jutiapa. Felt strongly over large areas of Guate— mala and El Salvador. 1917 December 25 22:20* ________ 2, 6, 8 A series of earthquakes that began on November 17, 1917, and con- tinued into January 1918. December 25 main event had magnitude of 6 plus and maximum Modified M-ercalli intensity of VIII-IX. In Guatemala City, cracks opened in the streets, and about 40 percent of the houses were destroyed or seriously dam- aged. The Colon Theatre collapsed while filled with people; school buildings, churches, asylums, hospitals, sugar mills, the post office, the railway station, and the British and American legation buildings were thrown down, and many occupants were killed or injured. Later destructive earthquakes occurred on December 29 (14h), January 3, 1918 (22:37), and January 24 (19:30). The most destructive of this series was the January 3, 1918, earthquake. INSTRUMENTAL DATA Location Magni- Time —— h tude Station Refer- Date UTC N. Lat W. Long (km) (Richter) code ence 1919 Apr. 17 20 53 03.0 14.5 91.7 ___ 7.0 PAS 4 1921 Feb. 4 08 22 44.0 15.0 91.0 120 7.5 PAS 4,5 1929 Jan. 171 19 00 ___ ___ ___ ___ __ ___- 2 1931 Sept. 26 19 50 30.0 15.0 92.0 60 6.0 PAS 4,5 1931 Sept. 26 20 03 07 15.5 91.5 ___ 6.25 PAS 4,5 1932 May 22 22 40 02.0 14.2 90.0 80 6.0 PAS 4,5 1934 May 19 10 47 37.0 14.8 91.2 120 6.25 PAS 4,5 1939 Sept. 28 14 58 27.0 15.5 91.5 110 6.25 PAS 4,5 1939 Dec. 5 08 30 07.0 14.5 91.5 -__ 6.75 PAS 4,5 1940 July 27 13 32 30.0 14.2 91.5 90 6.75 PAS 4,5 1942 Apr. 11 01 25 12.0 14.7 91.5 140 6.5 PAS 4,5 1942 Aug. 62 23 36 59.0 14.0 91.0 ___ 8.3 PAS 4,5,6 1942 Aug. 8 22 36 34.0 14.2 91.5 ___ 6.5 PAS 4, 5 1943 Aug. 31 16 10 40.0 14.2 91.5 80 6.75 PAS 4,5 1943 Sept. 23 15 00 44.0 15.0 91.5 110 6.75 PAS 4,5 1944 Oct. 2 17 22 00.0 14.5 89.8 160 6.5 PAS 4, 5 1945 Aug. 10" 11 20 03 15.4 88.8 ___ __ ___- 2 1945 Oct. 27 11 24 41.0 15.0 91.2 200 6.75 PAS 4,5 1946 Jan. 5 01 15 11.0 15.0 91.0 210 6.0 PAS 4,5 1946 June 26 07 53 40 14.7 90.8 90 6.5 PAS 4 5 90 CHRONOLOGICAL HISTORICAL RECORD—Continued INSTRUMENTAL DATA—Continued Location Magni< Time _ h tude Station Refer- Date UTC N. Lat W. Long (km) (Richter) code ence 1948 July 16‘ 07 12 28 14.6 91.2 ___ 6.25 PAS 2 07 19 39 14.3 91.2 ___ 6.75 1949 July 8712 40 47.0 14.0 91.5 100 6.0 PAS 4,5 1950 Feb. 03 47 16.0 13.9 90.8 ___ 6.25—6.50 PAS 4,5 1950 Oct. 235 16 13 20.0 14.5 91.5 ___. 7.1 PAS 2,4,5 1950 Oct. 23 17 47 51.0 14.5 92.0 ___ 6.5 PAS 5 1950 Oct. 23 23 38 43.0 14.3 91.7 __- 61 PAS 5 1950 Oct. 24 00 52 03.0 15.0 92.0 ___ 6.2 PAS 5 1950 Oct. 24 O5 50 15.0 14.5 92.0 ___ 6.0 BRK 5 1950 Oct. 28 22’ 15 47.0 14.3 91.7 64 6.2 TAC 5 1950 Nov. 5 16 35 20.0 14.5 92.0 -__ 6.25 PAS 5 1951 Jan. middle of month " 1951 July 25 18 42 19.0 14.5 90.5 64 6.25 PAS 5 1953 Aug. 24 13 21 02.0 14.1 91.4 96 6.5 PAS 5 1953 Nov. 17 7 13 29 52. 0 13. 8 91. 8 ___ 7. 25— 7. 50 PAS 5 1954 Oct. 21 06 51 44.0 13.9 90. 6 _-_ 6.5 PAS 5 1955 Aug. 28 20 13 30.0 14.0 91.0 60 6.75 PAS 5 1955 Sept. 3 12 36 21.0 13.8 90.8 64 6.50—6.75 PAS 5 1957 July 8 15 30 33.0 14.5 91.0 150 6.0 BRK 5 1959 Feb. 20 18 16 20. 0 15. 9 90. 6 48 6.5 PAS 5 1959 Mar. 9 22 03 03.0 15.1 91.0 171 6.3 TAC 5 1960 Apr. 139 12 37 43.0 15.5 92.0 30 6.0 PAS 5,7 1960 Aug. 20 00 19 35.2 14.5 91.5 115 6.0 PAS 5 1961 June 17 15 O7 33. 7 14.2 92.0 85 6.0 PAS 5 1961 Sept. 1 18 50 35.4 13.6 92.5 37 6.5 PAS 5 1965 Aug. 5" 19 05 07.7 14.8 91.0 59 4.0mb CGS 5,7 1966 Aug. 181° 10 33 17.7 14.6 91.7 85 6.0 PAS 2,5 5.9—6.2 BRK 6.0—6.25 PAL 5.9 mi, CGS 1969 Apr. 21 02 19 07.1 14.1 91.0 82‘ 6.0 PAS 5 6.25 BRK 5.5 mi. CGS 1971 Oct. 12 n 09 44 59.3 15.8 91.2 36 6.0 BRK 2,5 517 Ms ERL 5.7 mb ERL 1972 Jan. 2212 13 08 50.3 14.0 91.0 102 6.0 PAS 2,5 5.5 mb ERL 1973 June 7“ 18 32 42.9 14.3 92.0 78 6.2 PAS 2,5 5.5 mb ERL 1974 Dec. 31 13 20 21 09.0 14.1 91.8 39 6.0 BRK 2,5 6.1 Ms CGS 5.7 mi, CGS 1 Considerable damage in Puerto Barrios. 2Felt strongly throughout the central region of Guatemala; alarmed a large part of the population and. caused considerable damage. 3 Numerous earth cracks near Quirigua and significant (moderate) damage in Quirigué. Felt in De- partments of Chiquimula, Zacapa, and part of Izabal. 4 People were tossed from their beds in Guatemala City. 5 Near the coast of Guatemala. Damage in San Marcos area. “A series of strong earthquakes in the region of Ixhuatan in the Department of Santa Rosa that caused considerable damage in the area. 7 Near the coast of Guatemala. sGuatemala—Mexico border; one killed, 14 injured, and damage to the San Marcas area; also damage at Heuhuetenango. 9 One killed and four injured at a hydroelectric project when several workers were buried under dirt. 1° Felt at San Salvador. 11 Mexico-Guatemala region. 12 Felt (intensity V) at San Salvador, El Salvador. 1" Felt in Guatemala City area. References: 1) Brigham (1887). (2) Seismological Notes, Seismological Society of America Bulletin, v. 5—15, v. 8—18, v. 38—48. (3) Montessus de Ballore (1888). (4) Gutenberg and Richter (1954). (5) NOAA Hypocenter Data File (1976), written commun). (6) Vassaux (1969). (7) U.S. Earthquakes (1960, 1965). (8) Claudio Urrutia E. (1976, written commun.). fiUS. GOVERNMENT PRINTING OFFICE: 1976 0—211/167 COVER PHOTOGRAPH: Surface fractures along the Motagua fault crossing a soccer field at Gualan, Department of Zacapa. The fractures, formed during the February 4, 1976, earth- quake, show left-lateral strike-slip displacement; the ground on the right side of the photograph moved west (away from the viewer), and the ground on the left side of the photograph moved east (toward the viewer). Superimposed on this picture is the trace of the horizontal long-period low-gain seismogram, east-west component, recorded at Albuquerque Seis- mological Center, Albuquerque, New Mexico, U.S.A. Epicentral distance is 2,778 km. Effects of Urbanization on Streamflow and Sediment Transport in the Rock Creek and Anacostia River Basins, Montgomery County, Maryland, 1962—74 By THOMAS H. YORKE and WILLIAM J. HERB GEOLOGICAL SURVEY PROFESSIONAL PAPER 1003 Work done in cooperation with the Maryland-National Capital Park and Planning Commission (Montgomery County and Prince Georges County Planning Boards), the Washington Suburban Sanitary Commission, the District of Columbia Department of Environmental Services, the Montgomery County Department of Environmental Protection, and the Maryland Department of Natural Resources UNITED STATES GOVERNMENT PRINTING OFFICE, WASHINGTON : 1978 UNITED STATES DEPARTMENT OF THE INTERIOR CECIL D. ANDRUS, Secretary GEOLOGICAL SURVEY V. E. McKelvey, Director Library of Congress Cataloging in Publication Data Yorke, Thomas H Efl’ects of urbanization on streamflow and sediment transport in the Rock Creek and Anacostia River basins, Montgomery County, Maryland, 1962-74. (Geological Survey professional paper ; 1003) Bibliography: p. Supt. of Docs. no.: I 19.16:1003 1. Stream measurements—Maryland—Montgomery Co. 2. Sediment transport—Maryland—Montgomery Co. 3. Urbanization—Mary]and—Montgomery Co. 4. Urban hydrology. I. Herb, William J ., joint author. 11. Maryland. Maryland—National Capital Park and Planning Commission. III. Title: Effects of urbanization on streamflow . . . IV. Series: United States. Geological Survey. Professional paper ; 1003. G31225.M3Y67 551.3’03 76—608222 For sale by the Branch of Distribution, U.S. Geological Survey, 1200 South Eads Street, Arlington, VA 22202 STOCK Number 024-001—03036—5 CONTENTS III . Page Page Abstract ________________________________________ 1 Sediment discharge—Continued Introduction _____________________________________ 1 Size distribution of suspended sediment ________ 16 Acknowledgments ________________________________ 2 Bedload cqntiiPutim """"""""""""" 17 . . Storm variability ____________________________ 19 Factors afi‘ectlng streamflow and sediment transport- 2 Correlation matrix _______________________ 20 Physmgraphy '. """""""""""""""" 5 Regression equations _____________________ 21 Geology and $0113 """""""""""""" 5 Effects Of land use/land cover ________________ 41 Climate """""""""""""""""""" 5 Average annual sediment yields __________ 41 Land use/ land cover """"""""""""" 6 Construction-site sediment yields _________ 51 Streamflow —————————————————————————————————————— 8 Stream—channel erosion ___________________ 56 Annual and monthly runoff ___________________ 8 Sediment control _________________________________ 63 Duration and low flow ——————————————————————— 9 Field applications ____________________________ 64 Storm runoff and flood-peak discharges ———————— 10 Effects of sediment-control measures __________ 66 Sediment discharge ______________________________ 12 Cost of sediment controls ______________________ 67 Annual and monthly suspended-sediment Summary ________________________________________ 69 discharge __________________________________ 12 Selected references _______________________________ 70 ILLUSTRATIONS PLATE 1. Map of average slope in North Branch Rock Creek and Northwest Branch Anacostia River basins, Montgomery County, Md. _____________________________________________________________ In pocket 2. Land use/land cover in North Branch Rock Creek and Northwest Branch Anacostia basins, Mont- gometry County, Md., 1966 ____________________________________________________________ In pocket 3. Land use/land cover in North Branch Rock Creek and Northwest Branch Anacostia River basins, Montgomery County, Md., 1974 ________________________________________________________ In pocket Page FIGURE 1. Location map of Northwest Branch Anacostia River and North Branch Rock Creek basins _____ 3 2. Graph showing mean monthly rainfall at locations in and near the study area, October 1962—Sep- tember 1974 ___________________________________________________________________________ 6 3-15. Graphs showing: 3. Variation in mean monthly precipitation and runoff, Northwest Branch Anacostia River near Colesville, Md., October 1962—September 1974 _______________________________ 9 4. Flow-duration curves for the Northwest Branch Anacostia River, 1923—39, 1961—74 (A); 1963—74 (B); and 1967—74 (C) __________________________________________________ 9 5. Flow-duration curves for five small streams draining areas of difierent land use/land cover, 1967—74 _________________________________________________________________ 10 6. Relation between storm runofl’ and precipitation for selected subbasins, 1971—74 _________ 11 7. Annual flood-frequency curves for selected subbasins, 1967—74 __________________________ 12 8. Annual variation of precipitation, water discharge, and suspended-sediment discharge, Northwest Branch Anacostia River near Colesville, 1963—74 _____________________ 14 9. Duration of water discharge and suspended-sediment discharge, Northwest Branch Ana- costia River near Colesville, 1963—74 _____________________________________________ 15 10. Mean monthly suspended-sediment discharge, Northwest Branch Anacostia River near Colesville, 1963—74 ______________________________________________________________ 16 11. Relation of sand, silt, and clay content of suspended sediment to water discharge, North- west Branch Anaeostia River near Colesville, 1960—73 ___________________________ 17 12. Particle-size distribution of suspended sediment transported during medium and high flows in the small study basins, 1967—74 _______________________________________________ 18 13. Daily suspended-sediment and instantaneous bedload transport curves for the Northwest Branch Anacostia River near Colesville, 1963—74 ________________________________ 19 14. Example of flow-duration (A) and suspended-sediment transport (B) curves used for com- puting average annual sediment loads _____________________________________________ 48 15. Relation of suspended-sediment yield to percentage of drainage basin under construction ___ 50 IV FIGURE 16. 17—19. 20. 21—23. 24. 25—27. 28. TABLE 1. 10. 11. CONTENTS , 1 Map showing construction sites and average slope conditions, Beli Pre Creek and Manor Run basins, June 1967 _-_________________________._______________‘_ ___________________________ Photographs showing: 17. Extensive erosion in the Manor Run basin, 1967 _____________________________________ 18. Construction activities without (1966) and with (1974) sediment controls, Bel Pre Creek basin ______________________________________________ F ___________________________ 19. Large sediment-stormwater management basin used to trap‘ sediment before it can leave construction sites, Bel Pre Creek basin, 1974 __________________________________ Graph showing relation between construction-site sediment yields (ibmputed from observed data and those estimated by equation 3 ___________________________ L ___________________________ Photographs showing: 21. Grading for street right—of-way within the Manor Run stream channel, 1967 ______________ 22. Construction site in the Lubes Run basin, 1964 (A), 1968 (B), and 1974 (C) ___________ 23. Stream channel-geometry reaches, 1974: Manor Run (A), Northwest Branch Anacostia River at Norwood (B), and Batchellors Run (C) ________________________________ Graphs showing typical cross-section changes between 1967 and 1974 in straight (sections 9, 8, and 4A) and curved (sections 5, 11, and 9) reaches of three typical stream channels in the study area _____________________________________________________________________________ Photographs showing: 25. Typical sediment basin used in 1966 _________________________________________________ 26. Sediment basin before (A) and after (B and C) it was breached for installation of a sewer line _____________________________________________________________________ 27. Small sediment basin requiring maintenance _________________________________________ Graph showing relation between suspended sediment and storm runoff for dormant season (A) and growing season (B) storms, Northwest Branch Anacostia River near Colesville, 1963—74 ___ TABLES Summary of streamflow, suspended-sediment, and recording precipitation stations in the Rock Creek and Anacostia River basins _______________________________________________________ Land use/land cover, in acres, in the Rock Creek and Anacostia River basins, 1963—74 _________ Monthly and annual water and suspended-sediment discharge, Northwest Branch Anacostia River near Colesville, Md., 1963—74 ____________________________________________________________ Particle-size distribution of suspended sediment, Northwest Branch Anacostia River near Colesville, Md., 1960—73 ___________________________________________________________________________ Hydrologic characteristics and related factors for storms in the Rock Creek and Anacostia River basins, 1962—74 ________________________________________________________________________ Simple correlation coefficients of storm-related variables ______________________________________ Regression coefficients, multiple correlation coefficient, and standard error of best equations explain- ing the variation of storm suspended-sediment discharge __________________________________ Example of computation of average annual suspended-sediment discharge ______________________ Average annual suspended—sediment yields and selected land-use/land-cover factors _____________ Construction-site sediment yields and related factors ___________________________________________ Summary of stream-channel surveys _________________________________________________________ CONVERSION FACTORS The following factors may be used to convert the English units published in this report to metric units. Multiply English um’t By To obtain metric unit acres 0.4047 square hectometers (hm2) cubic feet per second (ft3/s) 0.02832 cubic meters per second (ms/s) cubic feet per second-day (ftS/s-d) 2447 cubic meters (m3) 0.002447 cubic hectometers (hm‘) cubic feet per second per square mile 0.01093 cubic meters per second per square [(ft“/s)/mi2] kilometer [(ms/s) /km”] feet (ft) 0.3048 meters (m) inches (in) 25.4 millimeters (mm) miles (mi) 1.609 kilometers (km) pounds per cubic foot (lbs/ft“) 16.05 kilograms Per cubic meter (kg/ma) square miles (miz) 2.590 square kilometers (kmz) short ton 0.9072 metric tons or tonnes (t) tons per acre 2.242 tonnes per square hectometer (t/hm‘) tons per square mile (tons/mi“) 0.3503 tonnes per square kilometer (t/km’) Page 54 55 56 57 58 59 61 62 64 65 66 67 Page 13 16 22 42 46 47 48 53 EFFECTS OF URBANIZATION ON STREAMFLOW AND SEDIMENT TRANPORT IN THE ROCK CREEK AND ANACOSTIA RIVER BASIN S, MONTGOMERY COUNTY, MARYLAND, 1962-74 By THOMAS H. YORKE and WILLIAM J. HERB ABSTRACT Land use/land cover, precipitation, streamflow, and sedi- ment data were collected from nine drainage subbasins in a 32-square-mile (83-square-kilometer) area north of Wash- ington, DC, in Montgomery County, Md. This study was begun in 1962 to define urban runoff and sediment problems and was expanded in 1966 to evaluate response to sediment- control practices in areas undergoing urban development. Land use/land cover varied considerably in the study sub- basins, which ranged in size from 0.35 to 21.1 square miles (0.91 to 54.6 square kilometers). Three subbasins remained virtually rural, while the others underwent urban develop- ment. In 1974, urban land represented from 0 to 60 percent of the land use in the nine subbasins. Urbanization did not affect median and low flows, but did increase storm runoff and peak discharges. Suspended sediment transported from one of the basins that underwent urban development, the 21.1 square mile (54.6 square kilometer) Anacostia River basin, averaged 14,800 tons (13,400 tonnes) per year between 1962 and 1974, while urban construction averaged about 3 percent of the drainage area during this period of time. Bedload was esti- mated to be 6 to 13 percent of the total sediment load. Most of the suspended load was transported during storms, and varied from storm to storm and seasonally. High loads were generally associated with high runoff in the spring and intense thunderstorms in the summer. Cropland, urban land, and construction sites were the major sources of sedi- ment. Annual suspended-sediment yields from land under cultivation ranged from 0.65 to 4.3 tons per acre (1.5 to 9.6 tonnes per square hectometer), compared to estimated yields from forest and grasslands ranging from 0.03 to 0.2 ton per acre (0.07 to 0.45 tonne per square hectometer). Yields from urban land were about 3.7 tons per acre (8.3 tonnes per square hectometer), most of which came from stream-channel erosion. Average annual suspended-sediment yields computed for urban construction sites ranged from 7 to 100 tons per acre (16 to 224 tonnes per square hectometer), The magnitude of the yield from construction sites was significantly affected by (1) the average slope of the sites, (2) the proximity to stream channels, (3) the existence of buffer zones of natural vegetation, and (4) the use of sediment-control measures. Sediment controls, particularly those enforced under a 1971 sediment-control ordinance, decreased construction-site sus- pended-sediment yields by 60 to 80 percent. It was estimated that the suspended-sediment load in the Anacostia River basin between 1962 and 1974 would have been reduced by 50 percent if strictly enforced sediment controls had been used throughout the period. Based on a reported cost of $1,030 per acre (0.4 square hectometer), the cost of sedi- ment controls on the 1,900 acres (769 square hectometers) developed during the period would have been $19 per ton ($21 per tonne). INTRODUCTION Each year thousands of acres of pastures, culti- vated fields, and woodlots are replaced by housing subdivisions, apartment complexes, and shopping centers. Urbanization is particularly evident in the Washington, DC, metropolitan area. Guy and others (1963) estimated that 500 mi2 (1,300 km?) of rural land would be urbanized between the mid- 1960’s and 1980 to accommodate a population in- crease of 2 million. Urban construction activities and the resulting increase in impervious area place a tremendous stress on the local streams. Erosion and sediment transport increase manyfold during construction, and increased runoff from impervious surfaces causes stream-channel erosion and flood- ing after construction. Studies by Guy and Fer- guson (1962), Wolman (1964), Guy (1965), and Vice, Guy, and Ferguson (1969) indicate that an- nual sediment yields for urban construction sites range from 25,000 to 120,000 tons/mi2 (8,800 to 42,000 t/km2) as compared with rural sediment yields of 150 to 500 tons/mi2 (53 to 175 t/kmz) (Wark and Keller, 1963). Anderson (1970) re- ported that urbanization increased flood peaks in northern Virginia by a factor of 2 to 8, depending on relative impervious area and the density of the storm-sewer system. This report presents the results of a study begun in 1962 to define urban sediment problems and which was expanded in 1966 to evaluate response to 1 2 EFFECTS OF URBANIZATION ON STREAMFLOW AND SEDIMENT TRANSPORT IN MARYLAND a program designed to control runoff and sedimenta- tion from urbanizing areas. Facts about land use/ land cover, streamflow, and sedimentation represen- tative of urbanizing areas in the Washington, D.C., area are presented. The effects of different land uses, construction practices, and sediment-control methods on runoff and sediment are evaluated and summarized. This information should provide the basis for land-use planning that would minimize the damaging effects of urban development on streams in the Maryland Piedmont and other areas of the country with similar climatic and physio- graphic characteristics. The study was made in a 32 mi2 (83 km?) area drained by the North Branch Rock Creek and the Northwest Branch Anacostia River. Both streams enter the Potomac River in the District of Columbia (fig. 1). Data from 9 streamflow-sediment monitor- ing stations, 9 recording rain gages, a varying num- ber of nonrecording rain gages, and 14 sets of aerial photographs taken between 1963 and 1974 were used in the study. The locations of the gages are shown in figure 1. The type of recording and sam- pling equipment and the type and length of record available for each site are summarized in table 1. Streamflow and suspended-sediment discharge were measured using the standard procedures de- scribed by Carter and Davidian (1968) and Guy and Norman (1970). Stage was recorded with digital water-stage recorders at a frequency of 48 to 288 readings per day, depending on the size of the drainage area and land use of the basin. Direct measurements of water discharge and indirect de- terminations of one or two peak discharges were used to establish the stage-discharge rating for each site. Suspended-sediment concentration was sampled with a combination of manual and automatic sam- plers. Each site was equipped with a single-stage sampler, which automatically collected samples at preselected levels during storms. Some sites were also equipped with pumping samplers. Hand samples were collected during most storms to provide a check on. the automatic samplers and to monitor sus- pended-sediment concentration during recession. flows. Precipitation was measured with a network of recording and nonrecording‘rain gages. Recording gages were located to provide rainfall duration and intensity data for each subbasin within the study area. Eight sites were equipped with digital re- corders having 5- or 15—minute recording intervals. One site was equipped with a weighing-type gage with a 24-hour continuous chart. Nonrecording gages were located throughout the basins to mea- sure the areal distribution of precipitation. Land use/land cover data were obtained from aerial photographs using grids with a dot density of 64 or 100 dots/in? The scale of photographs available prior to 1966 ranged from 1:23,000 to 1:37,000, and each dot represented 0.87 to 3.4 acres (0.35 to 1.4 hmg), depending on the scale of the photographs and the dot-grid used. After 1966, 1:12,000—scale photographs were obtained at least annually. Each dot represented 0.36 acre (0.15 hm?) on these photographs. Land use/land cover was summarized in five major categories: farmland and parks, rural residential, urban residential, commer- cial, and construction. The amount of cropland, grass, forest land, water, and impervious area was summarized for farmland and parks. The amount of grass, forest land, and impervious area was sum- marized for the residential or commercial categories. ACKNOWLEDGMENTS This report was prepared by the US. Geological, Survey in cooperation with the Maryland-National Capital Park and Planning Commission (Mont- gomery County and Prince Georges County Plan- ning Boards), the Washington Suburban Sanitary Commission, the District of Columbia Department of Environmental Services, the Montgomery County Department of Environmental Protection, and the Maryland Department of Natural Resources, Water Resources Administration. The authors wish to ex- press their gratitude to the following: J. S. Hewins, Inter-Governmental Coordinator, Maryland-Nation- al Capital Park and Planning Commission, and E. R. Keil, former State Conservationist, U.S. So-il Con- servation Service, for their encouragement and their strong support during the project; J. D. Large, Montgomery County Department of Public Works, for providing information on grading and paving permits issued to developers in the study area; R. M. Seely, Montgomery County Department of En- vironmental Protection, for providing records of sediment control inspections; and L. H. Williams, formerly of the US. Soil Conservation Service, for providing information on sediment-control stand- ards and field applications in the study area. FACTORS AFFECTING STREAMFLOW AND SEDIMENT TRANSPORT The streamflow and sediment characteristics of streams in the study area are affected by numerous factors including physiography, soils, climate, and FACTORS AFFECTING STREAMFLOW AND SEDIMENT TRANSPORT 77‘iUS' q). EXPLANATION Continuous-record gaging station Partial-record gaging station Sediment-measurement site 39° 05’ A v <$> - .— Drainage basin boundary Recording rain gage — ' '— Sub—basin boundary Glenmont \i MARYLAN . x990 f -‘ 77°|00’ Area of this report VIRG IN IA 192 O ’3 052* ‘9 CU ‘ NorthwestA Q Q? ' A O «92‘ e Lutes (2:0? , - N/ / / Wheaton Regional Park 39° _05’ 2 MILES J I 3 KILOMETERS i 77°05’ Base from U S Geoioglcai Survey Beitsvilie, Ctarksville, Kensmgton and Sandy Spring V 24000, 1971 FIGURE 1.—Location map of Northwest Branch Anacostia River and North Branch Rock 77°00’ Creek basins. EFFECTS OF URBANIZATION 0N STREAMFLOW AND SEDIMENT TRANSPORT IN MARYLAND 4 .3: fishing“ 4:5 .13»an £223 dun—-3 ................................... ”:2 .uaoMINgH .m=< £me .unwmlvwau .An2 .3235 £055 .556 :2 .unwmlgmfi :32 $333 uuuuuuu ace—ESE .—a>.~3§ £259 .EE-m :3 .anwmlwwm; $02 @2538 :uuuuul..-{Inuuuuullulu{Innilluuuiiw x3982 .3235 non—a din-2 3.3 .unwmluwaH :32 92338 ||||| l ................................ “mama 5534 5.359 msczfiunoo 3.3 .anwmlmwaa >62 38mm; uuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuu 550w “525.52. $353.55 £573 £32 .30 3 Rain 3235 52:5 din-m v52 dnwmlrwaa #32 $5358 ........................................... 3.5330 3533.55 .EEImH .32 #90 3 hora 3:35 A223 .Efipm :3 dammiwwag $02 $838 ............................... 550w magnum 35am .3233 :05:— .:mE-mH 3.3 dnwWIwwmm .8Q 3:.va .............................. «$35qu xownhoz .umuwaouwfi. .Efiuwn £me .30 3 now; 1:35 £955 .Eium v2: .anwmlpwg .Aa2 $3.va ........................................... 33.de £3.30qu— .=mEan £me .30 3 how; 323.: £05.:— .EE-m :2 .unwmlwwaa 5oz 23.va a uuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuu “~an 350 hiozconm Mflmvuooom 6.59% we «6er .02” :oflauw :cmaaum mZO~H18 ESm «mm: .Ewmlcmx: anm .............. 3:5 3:: .Ewmlmmi .80 :N 88ch 5:95 $255.82 .3285 :3 .anlmz: SE. :2 .EowIE: 33. -- - dmflm Swim ............ 2:0 53m 6:: :37me .3; ...... sag—NEE: v85 62: .34133 .2; S. 3338 - $5.5 as =.=m $25 .............. smug :3 .anww $me .30 553A .omsm 015m ............ 3:0 Snow :3 .M:oz .............. 3.25 «:2 .Ewmucme .30 84 $538 .58: 5.x 5.32 .u—uwnhcz haw: xwwao .wuflm 29:» ES min—55 ............ >15 Sham 2.2 .aawméca 462 .............. 3.5 :2 .unomlwwfi .30 35 3:38 goon 552m fiaoz ............ 3.3 Eu3m :3 .anwmlwme .30 4350 Law: 653m 03:? v5. Mina—Em ............. '1: 315 $3 .anwmlwwa duo .............. 3:5 :3 .umwmlwwg .uoo 3.x 56wa Si 35353:? £58353 .282 no wahh 2:802 no 63.8% 6.503 no 0538 when: we vomawm ANMEV .o mam—9.53 8.; :9 WW .5355 055:3.5‘ unwEmvom cod—nonmaw BouEumhm own—:39 .a m mZOF—lflhm HZHE~QHWQHDmemDm DZ< 304h2«.«m«.«« 3.3 8.88 3.8«.« 8.38.3 «8.88; «88; 8.83 $.83“ «v.8 8.x: 8.2.3 8.«««.« ............ «3H «38.3 «3.8 8.58 8.88 b«.««« 8.83 8.83 «mm: 8.8% 882 8.«B 8.8«4 «88H ............ :«H ««.E«.S 8.««« 5.28 3.83 E.«8.« 8.8% «o.««w.m 842 84.8 8.8 8.88; «3:: 8.8 ............ «E: 8885 888 «$85 858 E.«««.« «3.2. 88 «Sm m«.««« «38 88$ 86mm 8.««H ............ 83 888.3 «58 3}. 868 ««.«fi.« 8.82 «8.8 «58; «8.8 «38 8.8«.« «82 «38 ............ 82 «.«88 «.3 3.8.8 «.83 8.: «.83 «.8 825 38 «.«8 «.8 «.«S 8.«2..« ............ 82 «.83: «.53. «.8 «.m3 «.3 $5 «.83 8.««H «.83. #8 m.« 85 «.«38 ............ 8«« H.288 ES «.83. H88 «.8 «.3 5« «83%. Sw§ «.88 «.83 33 «.m« ............ 82 «.88.: «.w: «.«8 $3 «.«8 «.8 «.88 «.«E 8.88 33.8 8.88 «.w««.« 3“ ............ «8H 35.3 «.mg «.83 «3 848.8 «.8 .8 «.8«.« «.83 «in «.3 8.88 gm ............ 82 2:: E 63.38:“ EoEEuTVu-Euamzw «.83. «Em 8.me 52¢ «32 «8.28 «38 888 888 «.88 «83 «38 8.3m ..... E32 «.53. 8% 3.8 «.«S SS «58. 8m. «8; «8 88; «58; :8 :5 ............ EE «saxfi «.m«m 38. «.«E «8 Eon H««.« «8% 83 8o; 85 «.83 :«v ............ «R: «.85: «.8«« 858 81H 8% «8 H 88 H «8 H b««.« «8 «.«mb «.8? «.83 ............ «R: «$8.3 3.8 «4&5 «.Sw «.me 83 8F 8« «.«E.« «E 8:. «.Em «.m«« ............ SE «.83 «.5: 58 «.«8 «.«8 «m» «8 H «8 3:. «.88 «.«o; «.«8 «.ma ............ «5H 28% «.88 w.«H« «.8« «.«mm v.««« «58 58 «88 «.«8 «.mov «.83 WE ............ «8H «.88 «.mfi 33 m.««« «58 $8 3% «.«8 855 S.« «.33 38 E«« ............ 83 p.88 :8 «fl; «.«om 3.8 «8 «8 8«.« «in «8 8.54. «.88 «.88 ............ 83 «.85: E8 «4&5 «AS. «.88 «.88 «d8 «.88 «.84; «.8«« «.8; «.2: «.88 ............ 8.: «.83 «.«E «.98 «.«E 5.8 $8 «8 >««.~ «.«8 :8 «.28 «.«fi 38 ............ 82 8.3.8 «.8 8.8: «.5: H.8« «.88 88 ««« Em H8; «8. «s8 «5: ............ 88 «58.8 «.85 «.Sm ES «.5 «.8« «8m «8%; «.«8 «.88. 3.8 58 «.9: ............ 82 mhuvéncuom you «new 03:0 5 «engagemmw «335 1355‘ Jnow .w=< 35H. wish. .3: ua< flag .nwh duh. .8Q $02 #90 “WNW? £4...»me :vsc 633390 .38: .8pr SQSSSV «Kofism «8:23.82 .wsafiomwfi 8§w§m~§wéww§amsm «58 .8293 338:: ES. Bf:§.~l.m «Em—<8 14 EFFECTS OF URBANIZATION 0N STREAMFLOW AND SEDIMENT TRANSPORT IN MARYLAND 40,000 i- ' Suspended-sediment * 30,000 i. Z _ fl discharge E L“ 30,000 _ E E a) _ m _ DJ 0 z D 2 LU Lu (J) O u) 2 I l— : o o 2 8 .— 3 -_ 20,000 D 2 E 5 20,000 E 3 L a: (L (D (I) < (I) a: D I D < m U ‘0 I .1 (fl .4 0 < o .1 g 9 3 10,000 j: 10,000 2 o z ' z < < 0 0 a: 60 a D: E |:| Precipitation _ E 50 F ‘ _ — m i m W Water dlscharge _ 1200 :E 2 § 2 t] < O 40 — 4_ q E z E k g j E Z_ 30 _ g» 800 .: 2 < u.| < Z l: o l: -. E a: W E Lu 0 < 20 — a U (3 Lu I g a: E § _ e 400 ‘1 § 0 _ SS 0 g 10 ’ \ \ \ \\ \ \ \§5 Egg 2 _ \f ‘ Z 2 :l\ \ \ <2: < 1963 64 65 66 67 88 59 70 71 72 73 1974 FIGURE 8.——Annual variation of precipitation, water discharge, and suspended-sediment discharge, Northwest Branch Anacostia River near Colesville, 1963—74. discharges (fig. 8) ; however, a decreasing trend in suspended-sediment discharge with time was also apparent. This decrease in sediment discharge, as it relates to land use and sediment control, will be discussed in detail in later sections of this report. Most suspended sediment was transported during large storms, as indicated by the duration curve of suspended-sediment discharge (fig. 9). Daily sedi- ment discharge was greater than 100 tons/d (91 t/d) 5.7 percent of the time or 252 days for the 12 years of record shown on figure 9. The sediment dis- charge for those 252 days represents 94 percent of the sediment discharged during the 12-year period. Sediment discharges of 500 and 1,000 tons/d (454 and 907 t/ d) were exceeded 2.2 and 1.0 percent of the time and represented 73 and 54 percent of the total. The 'quantity of sediment carried during storms, particularly large storms, is a significant factor that must be considered in any program de- signed to control erosion and sediment transport. If controls such as diversion berms and sediment basins are underdesigned, a large part of the total sediment eroded fro-m construction sites will still reach the streams. The seasonal variation of suspended-sediment dis- charge is another factor affecting sediment control. Sediment discharge is generally higher between, February and August than it is between September and January. The average monthly load for the period 1963—74 ranged from.670 tons (609 t) in September, to 2,700 tons (2,400 1:) in June (fig. 10). A comparison of the 1963—74 and the long-term pre- cipitation in figure 2, indicates that lower monthly loads would be expected in November, December, SEDIMENT DISCHARGE 15 10'000 — 1 1 1 1 1 | n. : _ \\\\ : T T— 100 1000 : : ~ 1000 d — : g 2 T O O (/1 r— _ O o 2 w w (n L” O T ~ 10 m Lu (/2 1— _ _ a: z x 2 Lu 2 If T m E I- 5 — _ g z s a m __ LL I E — 100 L“ o 100 r : E (D 9. u: a a: m E Z : g < 3 _ m I U 1— 1— U W A D U) z E . o 5 ~ 2 T I —1— 1 Z 5 5 — I “i E a: “-1 1| o u.| < U" “ a: E I Q E. ‘ _ < a 0 Lu I I u: . LU n: E 10 — 1, Water discharge _ E E 10 8 m U) ‘ I : D '3: 8 : I _ E 5 A — “ _ B 3 <2: 2 T ‘ — (n L“ D I ~ 0.1 2 > 5 o — ‘1‘ — 5 :1 Z < Z < E ‘1‘ _ 2 o ‘o‘ g \\ < \ D — 1 1 : \\ : Z _ T _— 0.01 0.1 1 1 1 1 I I 1 — 0.1 0.01 0.1 0.5 ‘I 5 20 50 80 99 99.99 PERCENTAGE OF TIME AN INDICATED DISCHARGE WAS EXCEEDED FIGURE 9.—Duration of water discharge and suspended—sediment discharge, Northwest Branch Anacostia River near Colesville, 1963—74. and June, and higher loads would be expected in January, July, and August. With the slight adjust- ments to figure 10 considered, the distribution of sediment discharge compares favorably to the distri- bution of erosion index values for the Atlantic coast area (Guy, 1964). The erosion index is highest in 16 EFFECTS OF URBANIZATION ON 8 ‘2 5 9. ~ 3000 0 Z I— —~ 3000 — w E LLI . (D —.'.' LIJ E .‘.‘.‘ g a s (L) ~ 2000 (0)) o 2000 1 a — '2 0 Lu P- ; E B 2 (I) ¥ 0 3 1000 _ g 1000 3,) g 8 Lu 0 a 2 U) Lu 3 D. m E 0 ‘3 ° 0 N D J F M A M J J ' A s "’ STREAMFLOW AND SEDIMENT TRANSPORT IN MARYLAND FIGURE 10.—Mean monthly suspended-sediment discharge, Northwest Branch Anacostia River near Colesville, 1963—74. June, July, and August and decreases to lowest in December and January. The variation of erosion potential and sediment discharge during the year creates an opportunity for an effective method of sediment control. Scheduling construction activities to coincide with the period of low erosion potential should result in a signifi- cantly lower sediment discharge. It is not possible or practical to restrict all grading and heavy earth- moving to the period between September through January, but it may be possible to schedule a major part of the grading on critical areas of construc- tion sites during this period of low erosion poten- tial. If major grading work were completed in the fall and protective measures were installed by mid- winter, there would be protection during the high runoff in early spring and during the intense storms in the summer. Also, if construction activities were completed and critical sites were stabilized before the high-erosion periods, less intensive control mea- sures would be required on the remainder of the construction area, resulting in an overall reduction in costs. SIZE DISTRIBUTION OF SUSPENDED SEDIMENT The size distribution of suspended sediment trans- ported by a .stream is an indicator of the sediment— source material and the competence of the stream. In the Northwest Branch Anacostia River near Colesville, suspended sediment transported during medium and high flows averaged 20 percent sand, 45 percent silt, and 35 percent clay (table 4). This approximates the distribution of particles in the predominant silt-loam and silty clay-loam soils found in the basin. The general relationship between water discharge and its capability to transport par- ticles of various sizes in suspension is illustrated by the curves in figure 11. At low discharges, the sus- pended sediment is composed mostly of silt‘ and clay. As the discharge increases, the percentage of the total suspended-sediment load consisting of silt and clay generally decreases, while the percentage consisting of sand increases. The large scatter of points about the generalized relationships in figure 11 probably represents dif- ferences in the type and amount of sediment eroded from various sediment-source areas and the distance between the source areas and the sampling sites. The TABLE 4.—Pm‘ticle-size distribution of suspended sediment, Northwest Branch Anacostia River near Colesville, Md. 1960—73 Suspended- Dat d-wfim sediment. 5:532:53- Percentirsxuzpzeenitlaggzediment e 150 arge concen- discharge (“J/5) my]? (tons/day) Sand Silt Clay Sept. 12, 1960 _____________________ 628 895 1,520 27 62 11 Feb. 26, 1962 ______________________ 247 1,480 987 38 41 21 Mar. 12, 1962 _____________________ 788 1,080 2,300 30 51 19 June 3, 1963 ______________________ 141 4,700 1,790 17 52 31 Aug. 13, 1963 _____________________ 73 15,100 2,980 7 53 40 Oct. 7, 1965 _______________________ 662 8,590 15,400 20 44 36 Feb. 13, 1966 _____________________ 662 5,070 9,060 36 38 26 May 19, 1966 _____________________ 50 9,830 1,330 3 44 53 June 22, 1967 __________________ 213 9,460 5,440 6 46 48 July 22, 1969 ______________________ 895 3,520 8,510 8 43 49 Apr. 2, 1970 ______________________ 71 198 38 6 21 73 Apr. 14, 1970 _____________________ 338 2,000 1,820 35 34 31 Oct. 21, 1970 ______________________ 169 2,070 945 17 55 28 July 29, 1971 _____________________ 197 3,700 1,970 14 45 41 Aug. 27, 1971 _____________________ 790 2,610 5.570 27 47 26 Nov. 29, 1971 _____________________ 213 836 423 24 44 32 Feb. 3, 1972 ______________________ 626 2,360 4,000 21 50 29 Feb. 2, 1973 ______________________ 376 1,660 1,690 24 45 31 Dec. 26, 1973 ______________________ 405 1,140 1,250 36 33 31 Sediment discharge-weighted average __________________ ___ ___- ____ 20 45 35 SEDIMENT DISCHARGE H .q PERCENT SAND (0.062—2.0 MILLIMETERS) PERCENT SILT (0.004—0.062 MILLIMETERS) O 1000 I I 1 1 I I T I I I I I I I a Z 0 ‘ O 8 800 0 “ 20 8 Lu 0 o o ‘F U) m 600 ° ° ° — I E - a 400 0 ° — <0 E o o — 10 E LLI — I— u. ‘_ 8 UJ 9 o o - E m o _ 5 U D 200 _ _ o — g 0 Z 0 — 4 2 g i n: 100 __ 3 < "‘ O: 5 80 g < <2 ° ' o —— 2 5 O 60 — ‘L’ E _ o c: ’2 40 ~ — E 3 —1 < I I I I I I J I I I L I L I I I i I E 0 10 20 30 40 10 20 30 40 50 60 7O 10 20 30 40 50 60 70 PERCENT CLAY (<0.004 MILLIMETERS) FIGURE 11.—Relation of sand, silt, and clay content of suspended sediment to water discharge, Northwest Branch Anacostia River near Colesville, 1960—73. first storms after cropland or construction sites have been opened will generally erode and transport large quantities of fine (silt and clay) particles completely through the stream system. During these first few storms, the larger sand particles, eroded by the impact of rain on the bared soils or by runoff from the land surface, are generally deposited at some place on the site or in the stream channel near the point of erosion. Subsequent storms Will continue to erode fine soil particles and also lift the heavier sand particles and redeposit them at some point farther downstream. If the surface soils are not reworked to expose new sources of fine sediments, the surface soils will gradually become armored with the remaining large particles. Thus, fewer fine particles will be available for transport, and sand particles will represent a greater percentage of the suspended load. An increase in the sand percentage of suspended sediment can also be expected after construction areas have been stabilized with vegetation. A dra- matic reduction in the total load occurs when active source areas are stabilized; however, large quantities of sand previously eroded from construction sites and deposited in stream channels are still available for transport. This is illustrated by the distribution of sand, silt, and clay transported during medium and high flows in the small study basins (fig. 12). Generally, the urban streams transported a higher percentage of sand than the rural basins. The sus- pended load ranged from 7 percent sand for Batchel- lors Run to 28 percent sand for Lutes Run. The higher levels of sand in the suspended load of urban streams can be expected to continue until a new equilibrium is established between the runoff regime, channel size, and bed material. BEDLOAD CONTRIBUTION Bedload was not measured directly during this study; however, the approximate range of its magni- tude was computed for the Northwest Branch Ana- costia River near Colesville, using the Schoklitsch and Meyer-Peter and Muller bedload formulas. Com- posite bed-material samples from several cross sec- tions and discharge measurements ranging from 20 to 800 ft3/s (0.57 to 22.7 mB/s) were used for the computations. The Meyer-Peter and Muller formula converted to English units by the US. Bureau of Reclamation (Sheppard, 1960) is G,,= 1.606 B [3.306 (fig—8) (My/2 dS ”s — 0.627Dm] 3/2 where G,s=total bedload discharge, in tons per day; B=bottom width of stream channel, in feet; 18 EFFECTS OF URBANIZATION 0N STREAMFLOW AND SEDIMENT TRANSPORT IN MARYLAND STATION NUMBER OF PERCENT ANALYSES 80 100 l | l Williamsburg Run (urban) - 5 J Manor Run (urban) 7 ) NW. Branch Anacostia River 4 J at NonNood (rural) Nursery Run (rural) 4 l Batchellors Run (rural) 3 ) Bel Pre Creek (urban) 11 ( Lutes Run (urban) 5 :-:- i l l l l | l | l L l l Sand Silt Clay (0.062—2.0 mm) (<0.004 mm) § //////A (0.004—0.062 mm) E FIGURE 12.—Particle-size distribution of suspended sediment transported during medium and high flows in the small study basins, 1967—74. Values are sediment-discharge weighted averages. Q8=the water discharge determining the bed- load, in cubic feet per second; Q=total water discharge, in cubic feet per second; Dgo=particle size at which 90 percent of the bed material is finer, in millimeters; n8=Manning’s roughness coefficient for the streambed; d= depth of flow, in feet; S=slope, in feet per foot, and Dm=effective size of bed material, in milli- meters. _. EDA}; ’" 100 where D is the geometric mean diam- eter of particles in a given size frac- tion and p is the percent by weight in that size fraction. The nomographs and standard computation forms prepared by the US. Bureau of Reclamation were used for the computations. The Schoklitsch bedload formula for bed material of uniform grain size, adapted from the English conversion by Shulits (1935), is 3/2 , 3745 S (Q— 0.00532wD,0> 9" = DWI/2 34/3 Where g,,=bedload discharge of particles in given size fraction, in tons per day; Dm=median diameter of bed material particles, in inches; S=slope of the energy gradeline, in feet per foot; w=width of the stream, in feet, and Q=total water discharge, in cubic feet per second. The bed material was divided into uniform size fractions and the bedload wascomputed for each fraction. The total bedload discharge was computed by the formula SEDIMENT DISCHARGE 19 _ a1g1+azgz+ ***a,,g,, G 100 where G=total bedload discharge, in tons per day, and an=percent of bed material in size fraction. Bedload transport curves computed by the two formulas and the average suspended-sediment trans- port curve for the period of record are shown in figure 13. The average annual bedload discharges computed for the 1963—74 period were 990 and 2,290 tons/yr (900 and 2,080 t/yr) for the Meyer- Peter and Muller and the Schoklitsch methods, re- pectively. Computed bedloads were 6 and 13 percent of the total sediment discharge for the same period. Bedload discharges for other streams in the study area were not computed; however, since the bed material and channel characteristics of the other streams in the study area were similar to those of the Northwest Branch Anacostia River, the bed- loads probably represent about the same percentage of the total load. With other conditions affecting bedload being equal, the bedload probably would be slightly less in streams draining rural areas and slightly higher in urban streams because of the WATER DISCHARGE, IN CUBIC METERS PER SECOND I 10 L II I ITfIIII I I I ’— EXPLANATION Suspended sediment — —————— Schointsch Meyer-Peter, Muller 1 000 — 1 000 I lllllI l 100 ‘100 SEDIMENT DISCHARGE, IN TONS PER DAY lllIIlII SEDIMENT DISCHARGE. IN TONNES PER DAY 111111—10 1 000 II : Illllll l I 10 100 WATER DISCHARGE, IN CUBIC FEET PER SECOND 10 l l I FIGURE 13.—Daily suspended-sediment and instantaneous bedload transport curves for the Northwest Branch Ana- costia River near Colesville, 1963—74. greater availability of sand-size particles in the urban streams. STORM VARIABILITY The sediment discharge of a stream is highly variable from year to year, season to season, and storm to storm. Many factors affect the erosion of soil particles and transport of sediment to the stream system. Guy (1964) analyzed variables af- fecting storm-sediment loads for seven Atlantic coast streams with drainage areas ranging from 98.4 to 4,571 mi2 (255 to 11,840 kmz). The variables that significantly affected sediment discharge in- cluded storm runoff, long-term mean air tempera- ture as a measure of season, rainfall intensity, and a peakedness index. The number of small drainage basins and the varying land use/land cover within the basins in this project provided an opportunity to expand on Guy’s study to determine if the sediment-water dis- charge characteristics of urbanizing basins respond differently to various storm variables than the char- acteristics of rural basins. A total of 16 hydrologic variables or related factors was determined for each subbasin studied in the Rock Creek and Anacostia River basins. These variables are: Storm suspended- sediment discharge (Q,) Total runoff (Qt) Storm runoff (42,) Peak dis- charge (6).») Antecedent dis- charge (Q0) Antecedent days (Ad) Total precipita- tion (Rt) total suspended-sediment dis- charge transported during the runoff period, in tons. total runoff during days of sur— face runoff, in cubic feet per second-day. total runoff minus estimated base flow, in cubic feet per second-day. (Base flow was estimated as the runoff below a straight line drawn from the point of initial rise to the point where the recession limb of the hydrograph approached a straight line.) instantaneous peak discharge, in cubic feet per second. mean daily discharge the day before the initial rise, in cubic feet per second. the number of days between storms. Average precipitation on the drainage basin, in inches, as 20 EFFECTS OF URBANIZATION ON STREAMFLOW AND SEDIMENT TRANSPORT IN MARYLAND determined by isohyetal maps for each storm, except as noted. maximum rainfall during a given interval, as deter- mined from the closest avail- able recording-rain gage: Rainfall intensity ROS—maximum 5—min rainfall, in inches per hour. Rls—maximum 15—min rainfall, in inches per hour. Rag—maximum 30—min rainfall, in inches per hour. RIF—maximum 1—hr rainfall, in inches per hour. Rsh—maximum 3—hr rainfall, in inches per hour. NOTE—The 5—min, 15—min, 30—min, and l—hr rainfall intensities were determined for the Lutes Run basin. The 3—hr rainfall was substituted for the 5—min rainfall for the other seven basins. Number of peaks (Np) Time (Tm) number of peaks during storm. number of months between the beginning of the record and the storm. Construction (Cp) percentage of basin area under active construction at the time of the storm. discharge-weighted mean sedi- ment concentration. 370.37 , c=__fl_) Q.~ approximated as : Qp_Qa P,.=__ Q7. The basic data compiled for each station are listed in table 5. The storms, which range in number from 26 for Batchellors Run to 93 for Williamsburg Run, represent all storms for which sediment sam- pling was adequate to define the sediment concentra- tion for the period of storm runoff. A stepwise multiple-regression model was selected to evaluate the effects of the different storm vari- ables on sediment discharges. The output consists of a simple correlation matrix of all selected vari- ables and a series of multiple-regression equations. In the first step of the regression program, all the variables are included in the computations. In sub- sequent steps, the variable with the least significant partial-regression coefficient in the preceding equa- tion is eliminated from further computations. The elimination and recomputation continues until one independent variable remains. Sediment concen- tration (C) Peak ratio (Pr) In order to meet the assumptions of a linear re- gression model, hydrologic data generally have to be transformed to logarithms. A review of the study by Guy (1964) and a partial graphical analysis of data used in this study indicated that the following transformations were required: log Q,” log QT, log (2,, log (IOXQa), log (IOXRt), log (10XR05),10g (IOXRis), log (IOXRan), 10g (10xRu.), 10g (10XR3h) , 10g C, and log P... Antecedent runoff and the rainfall factors were multiplied by 10 before conversion to simplify the logarithmic expression of the original values. CORRELATION MATRIX Simple correlation matrices of all selected vari— ables for each station are summarized in table 6. These matrices were examined with three objec- tives: to investigate any bias caused by changes in sampling procedures or climatic conditions, to deter- mine the degree of relation between the dependent variables (sediment discharge and sediment concen- tration) and the independent variables, and to deter- mine the intercorrelation of the independent vari- ables. The relation between time in months (here- after referred to as the chronology factor) and the other independent variables was used as an indica- tor of a climatic or sampling bias during the study period. An examination of the correlation between chro- nology and total storm rainfall indicated that there had been no significant change in the magnitude or intensity of the sampled rainfall events during the study period. Lutes Run was the only exception, ex- hibiting a highly significant (99 percent confidence level) negative correlation between chronology and total precipitation. Apparently, more smaller storms were sampled toward the end of the study period as a result of an improvement in equipment and tech- niques at the Lutes Run automatic pumping sam- pler. In spite of the nonsignificant correlation between chronology and total storm precipitation for indi- vidual storms, there was a generally positive cor- relation between chronology and net storm dis- charge. Correlations were significant at the 95-per- cent confidence level for Williamsburg Run, Manor Run, and Northwest Branch Anacostia River. The positive correlations were probably related to the significant positive correlations between chronology and antecedent discharge, which reflected a trend toward wetter years at the end of the study. This interpretation was supported by the precipitation records for stations near the project area, which SEDIMENT DISCHARGE 21 indicated an increase in average annual precipita- tion of about 10 inches (250 mm) during the second half of the sample period. Apparently, there were: more storms and a shorter time between storms near the end of the project, resulting in higher soil moisture levels and higher base flows at the begin- ning of each storm. Lutes Run and Bel Pre Creek, draining the two most extensively urbanized basins in the study area, were the only streams exhibiting significant positive correlation between chronology and peak ratio. This relationship indicated that the ratio of peak dis- charge to storm discharge had been increasing, probably as a result of the effects of increased im- pervious area and storm sewers in these basins. In the Lutes Run basin, the smaller storm size and disproportionate number of summer storms in 1973 and 1974 also contributed to the significant positive correlation between chronologyand peak ratio. Sum- mer storms with low runoff volumes generally have a single peak of relatively short duration, resulting in a high peak ratio. The effect of each independent variable on sus- pended-sediment discharge and sediment concentra- tion was evaluated in subsets to facilitate compari- son between basins. Basins were considered as urban. or rural and small or large. Basins with less than 10 percent of the area in urban residential or pub- lic-commercial were considered rural and the others were considered urban. The urban basins generally had active construction sites throughout the study period. Basins were classified as small or large de- pending on whether their drainage areas were less than or greater than 1.5 mi2 (3.9 kmz). The relation between each independent variable and both suspended-sediment discharge and sedi- ment concentration was generally consistent. If the variable had a significant positive correlation with sediment discharge, it also had the same relation with sediment concentration. The only exception was storm runoff. There was no consistent trend between sediment concentration and storm runoff for the basins; however, there was a highly signifi- cant positive relation between suspended-sediment discharge and storm runoff. This was expected as runoff transports suspended sediment. Other variables with a significant positive relation with suspended-sediment discharge were peak water discharge, total precipitation, and intensity param- eters. These were significant for both small or large and rural or urban basins. The chronolgy factor and percentage construction were only sig- nificant for the urban basins. That there was no significant correlation between chronolgy and sedi- ment discharge in the rural basins was further evidence that the sediment data were not biased by changes in sampling or climatic conditions. Antecedent discharge, antecedent days, and peak ratio affected suspended—sediment discharges dif— ferently in the large and small basins. Sediment dis- charge had a significant positive correlation with antecedent water discharge and a significant nega- tive correlation with antecedent days on the large basins. The relations were insignificant at the 95 percent confidence level on the small basins. This difference between the large and small basins was probably the result of a higher correlation between runoff volume and sediment discharges for the large basins. The sediment discharge from small basins was less affected by runoff volume and more de- pendent on the intensity or concentration of runoff. This was illustrated by the significant correlation between peak ratio and suspended-sediment dis- charge for the small basins. A number of independent variables were found to be significantly correlated with each other. In particular, the correlations between the rainfall intensity parameters were highly significant. Cor- relations between rainfall parameters, peak dis- charge, and peak ratio were also high. A generally negative correlation existed between antecedent dis— charge and rainfall intensity. This correlation was apparently a reflection of the seasonal variation of storms, that is, of the occurrence of intense convec- tive storms during the growing season when the base flow is generally lower than during the dormant season. Because of the intercorrelation of these variables, the regression model was set up so that these variables would not be considered simultan- eously. As many as 10 runs of the model were made for each basin so that the effect of each of these variables could be evaluated without the influence of the other highly correlated independent variables. REGRESSION EQUATIONS The computer analyses of various combinations of independent variables produced regression models of the form log Q.= b0+ b1x1+ ngg + *** + bum" where Q. =suspended-sediment discharge ; b0 = regression constant; b,=regression coefficient for the corresponding variables (x,,). 22 EFFECTS OF URBANIZATION ON STREAMFLOW AND SEDIMENT TRANSPORT IN MARYLAND TABLE 5.—Hydralogic characteristics and related factors for Mean sedi- Suspended- Total Storm. . Peak Antecedent Period of runoff runoff ment c9“' sed1ment discharge discharge storm runoff (Qt) (Qr) centxéafxon disfgaige p (Qa) (ft3/s-d) (ftS/s-d ) (mg/1) (WIS) (ffi/s) (ft3/s) Williamsburg Run near 1966 Nov. 28—29 _________________ 7.40 5.00 644.44 8.70 20.0 1.00 1967 Jan. 7— 9 _________________ 14.00 6.20 436.08 7.30 17.0 1.70 Jan. 27—28 ................. 15.00 11.00 6,632.98 197.00 71. 0 1.40 Mar. 6— 8 _________________ 70.00 62.00 2,514.93 421.00 242.0 2.40 May 7— 8 _________________ 19.00 15.00 1,604.94 65.00 51. 0 1.30 Aug. 3— 5 _________________ 22.00 20.00 7,296.28 394.00 127.0 .59 Aug. 24—25 _________________ 60.00 58.00 1,749.68 274.00 424. 0 .68 Oct. 10—11 _________________ 2.10 ‘ .83 267.74 .60 6.7 .57 Oct. 25—26 _________________ 4.50 2.70 1,248.28 9.10 26.0 .65 Nov. 2— 3 _________________ 5. 80 4. 20 1,146.38 13.00 19.0 .69 Dec. 10—12 _________________ 30. 00 23. 00 1,127.21 70.00 48.0 1.40 Dec. 28—29 _________________ 25. 00 20. 00 425.93 23.00 64.0 1.70 1968 Jan. 14—15 _________________ 51.00 46.00 418.68 52.00 118.0 .98 Mar. 17—18 _________________ 25.00 18 00 2,386.83 116.00 46.0 3.20 Apr. 24—25 _________________ 5.30 2.10 617.28 3.50 11.0 1.10 May 27—29 _________________ 38.00 33. 00 1,010.10 90.00 84.0 .90 June 19—20 _________________ 12.00 8 50 3,485.83 80.00 69.0 1.40 June 26—28 _________________ 20.00 16. 00 2,824.07 122.00 70.0 .98 Aug. 19—20 _________________ 2.20 1.40 3,174.60 12. 00 15.0 .34 Sept. 6 _________________ 1. 40 1.20 864.20 2. 80 9.6 .17 Sept. 10—11 _________________ 6. 80 6. 20 3,166.06 53. 00 46.0 .18 Nov. 18—19 _________________ 12. 00 9.10 651.20 16.00 34.0 1.10 1969 Jan. 21 _________________ 6.20 5.00 585.18 7.90 15.0 .98 Feb. 24 _________________ 8.20 4.20 520.28 5.90 20.0 6.20 May 20—21 _________________ 3.30 1. 30 1,054.13 3.70 9.1 .77 June 2— 3 _________________ 13.00 12. 00 1,141.97 37.00 96.0 .47 June 18—19 _________________ 11.00 9.80 1,700.68 45.00 90.0 .40 July 20—21 _________________ 8.20 7. 60 1,023.39 21.00 75.0 .17 July 22—23 _________________ 23.00 21 00 1,005.29 57.00 246.0 .66 Aug. 1— 4 _________________ 49.00 45.00 1,045.27 127.00 606.0 .47 Aug. 9—10 _________________ 33.00 31.00 800.48 67.00 259.0 .69 Aug. 18 _________________ 22.00 21.00 1,022.93 58.00 118.0 .69 Sept. 3— 4 _________________ 44.00 42.00 740.74 84.00 331.0 .78 Oct. 2 _________________ 4.30 3.60 1,337.45 13.00 36.0 .60 Nov. 19—20 _________________ 7.30 5.50 875.42 13.00 62.0 .61 Dec. 7— 8 _________________ 5.50 3.60 627.57 6.10 45. 0 .61 Dec. 10—11 _________________ 24. 00 21.00 776.01 44.00 105.0 1.00 Dec. 22—23 _________________ 14. 00 11.00 606.06 18.00 43.0 .90 1970 Apr. 14—15 _________________ 63.00 57.00 1,098.11 169.00 210.0 1.40 May 13—14 _________________ 18.00 14.00 1,693.12 64.00 103.0 1.90 May 24—25 _________________ 23. 00 19. 00 1,423.00 73.00 134.0 1.20 June 16—17 _________________ 7.10 4. 80 2,083.33 27.00 61.0 .85 June 21—22 _________________ 17 00 15 00 3,012.34 122.00 112.0 .93 July 9—10 _________________ 31. 00 29. 00 1,123.88 88.00 194.0 .68 July 20—21 _________________ 44.00 41.00 1,056.91 117.00 427.0 .70 Sept. 10-11 _________________ 6.00 4.60 3,462.15 43.00 67.0 .59 Nov. 3 _________________ 4 20 3. 40 871.46 8.00 26.0 .67 Nov. 4— 5 _________________ 40. 00 37. 00 950.95 95.00 155. 0 4.20 Nov. 14—15 _________________ 22. 00 18. 00 1,008.23 49.00 90.0 1.50 Dec. 16—17 _________________ 14.00 11.00 1,144.78 34.00 57.0 1. 00 Dec. 22 _________________ 18.00 15.00 691.36 28.00 69.0 1. 20 1971 Feb. 7— 9 _________________ 80.00 72.00 1,347.73 262.00 271.0 3.80 Feb. 22—23 _________________ 43.00 35.00 1,576.72 149.00 98.0 2.40 May 13—14 _________________ 34. 00 30.00 2,432.09 197.00 218.0 1.10 May 30—31 _________________ 39. 00 34.00 577.34 53.00 109. 0 1.40 June 2— 4 _________________ 38. 00 28. 00 3,558.19 269.00 345.0 2.30 July 29—30 _________________ 8.80 7. 30 3,754.43 74.00 63.0 .55 Aug. 1— 2 _________________ 24. 00 22 00 2,171.71 129.00 122.0 1. 90 Aug. 3— 4 _________________ 80. 00 76. 00 2,573.10 528.00 1,140.0 17. 00 See footnotes at end of table. SEDIMENT DISCHARGE 23 storms in the Rock Creek and Anacostia River basins, 1962-71. c- Total pre— Maximum Maximum Maximum Maximum gfxgrea Antecedant cipitation 15-minute 30-minute l-hour {El-hour Peak Number of Time (0,.) 3323 (Rt) ”(‘32—‘32 27332 733" "3533“ 3:2? 2232? (“1013.) (223?” (males) (in/hr) (in/hr) (in/hr) (in/hr) r ,. has“; area Olney, Md. 39 1.16 0.32 0.30 0.24 0.14 3 80 1 1 6 4 1 26 .08 .08 07 .03 2.47 1 3 6.4 2 1 02 2.20 1.40 94 .37 6.33 1 3 6.4 10 2 68 .72 .68 55 .41 3.86 1 5 7.0 9 1 92 .76 .70 50 .40 3.31 1 7 7.6 13 2 61 4.24 2.52 1 38 .47 6.32 2 10 7.6 3 3 20 2.00 1.56 1 03 .59 7.30 1 10 7.6 43 2 84 .40 .30 22 .18 7.39 1 12 6.6 14 90 2.16 1.30 75 .33 9.39 1 12 6.2 7 1 00 .52 .42 34 .24 4.36 1 13 6.2 2 2 1.65 .44 .40 38 .27 2.03 1 14 6.2 1 2 1.50 .88 .68 45 .30 3.11 1 14 6.2 9 1.50 52 .48 46 37 2.54 1 15 6.2 3 1.25 40 .38 33 19 2.38 1 17 5.7 18 2 .97 76 .60 40 15 4.71 2 18 5.2 2 3.00 80 .62 50 33 2.52 1 19 4.7 1 .96 2.84 2.28 1.16 .37 7.95 1 20 4.7 5 1.84 2.76 1.98 1.09 .62 4.31 2 20 4.7 1 2 .70 3.00 1.66 .86 .29 10.47 1 22 4.3 17 2 .83 1.12 .98 .79 .42 7.86 1 23 3.8 3 1.57 3.56 2.62 1.58 .65 7.39 1 23 3.8 3 2 1.35 .32 .30 .27 .19 3.62 2 25 3.8 1 .70 .20 .20 .16 .10 2.80 1 27 3.8 0 2 1.09 .16 .16 .15 .13 3.29 2 28 3.8 10 1.46 1.44 .76 .43 .18 6.41 2 31 3.0 11 2.42 2.12 1.84 1.53 .74 7.96 1 32 3.0 14 1.96 4.40 3.12 2.13 .76 9.14 1 32 3.0 30 2 1.81 5.36 3.16 1.78 .60 9.85 1 33 3.0 1 2 1.49 5.88 2.96 1.48 .49 11.68 1 33 3.0 2 3.20 5.60 4.04 2.54 .88 13.46 3 34 3.0 3 2.43 2.72 1.46 .80 .63 8.33 1 34 3.1 6 2 2.08 3.52 2.06 1.07 .36 5.59 2 34 3.1 11 3.51 2.28 2.06 1.22 .60 7.86 2 35 3.1 23 1.20 1.68 1.12 .61 .25 9.83 1 36 3.1 10 1.22 1.16 .72 .47 .22 11.16 1 37 3.1 16 2 .90 .40 .36 .26 .19 12.33 1 38 3.2 1 1.46 .52 .44 .40 .28 4.95 1 38 3.2 10 1.10 .24 .24 .20 .15 3.83 1 38 3.2 11 3.06 .40 .38 .37 .27 3.66 1 42 3.2 18 1.41 2.48 1.26 .63 .38 7.22 1 43 3.3 6 1.57 2.56 1.48 .80 .32 6.99 2 43 3.3 21 1.81 2.40 1.54 .77 .26 12.53 1 44 3.3 2 1.51 1.96 1.86 .93 .39 7.40 1 44 3.3 16 2.71 2.00 1.52 1.04 .50 6.67 2 45 3.9 9 2.60 3.28 2.32 1.36 .50 10.40 1 45 3.9 17 1.26 2.44 1.68 .90 .30 14.44 1 47 3.9 2 2 .57 .68 .48 .33 .15 7.45 1 49 3.9 0 2.04 .60 .52 .45 .36 4.08 1 49 3.9 2 1 05 .44 .44 .37 .21 4.92 1 49 3.9 3 1.04 .24 .24 .20 .16 5.09 1 50 3.9 4 1.01 .36 .34 .28 .18 4.52 1 50 3.9 0 1.86 .60 .42 .30 .25 3.71 2 52 4.5 7 1.69 .52 .30 .29 .17 2.73 2 52 4.5 4 2.09 1.32 1.06 .78 .48 7.23 1 55 5.1 3 1.62 .52 .36 .26 .20 3.16 2 55 5.1 1 1.70 .28 .28 .25 .20 12.24 2 56 5.1 27 21'84 2.16 1.58 1.10 .42 8.55 2 57 5.1 2 2 1.99 1.96 1.80 .96 .35 5.46 1 58 5.1 0 2.25 4.80 3.66 1.99 .77 14.78 1 58 5.1 24 EFFECTS OF URBANIZATION ON STREAMFLOW AND SEDIMENT TRANSPORT IN MARYLAND TABLE 5.—Hydrolog'£c characteristics and related factors for storms Mean sedi- Suspended- See footnotes at end of table. _ Total Storm me t on- sediment .Peak Algtecedent “55:33:. r333.“ r333? Mg?“ “353”?“ “€32?” “15:32:“ (its/s-d) (st-d) (mg/1) “0,1,, (“n/s) (rte/s) Williamsburg Run near 1 971—Continued Aug. 27—28 _______________ 61.00 56.00 800.26 121.00 305.0 0.89 Oct. 10 _______________ 33.00 30.00 1,333.33 108.00 160.0 1.30 Nov. 24—25 _______________ 67.00 61.00 340.01 56.00 258.0 1.60 Nov. 29—30 _______________ 21.00 15.00 345.68 14.00 84.0 2.40 Dec. 7 _______________ 16.00 12.00 586.42 19.00 55.0 2.60 1972 Feb. 3— 4 _______________ 31.00 25.00 1,200.00 31.00 132.0 190 Feb. 13—14 _______________ 42.00 34.00 1,546.84 142.00 139.0 1.80 Feb. 18—19___.‘ ___________ 37.00 30.00 1,160,49 94.00 147.0 2.50 Mar. 16—17 _______________ 39.00 33.00 751.96 67.00 121.0 2.90 Apr. 13 _______________ 22.00 18.00 1,748.97 85.00 59.0 2.10 Apr. 16~17 _______________ 54.00 46.00 2,922.70 363.00 368.0 5.60 May 3— 4 _______________ 41.00 29.00 1,468.71 115.00 198.0 2.60 May 19—20 _______________ 16.00 9.90 635.99 17.00 44.0 2.50 May 22—23 _______________ 10.00 3.20 729.17 6.30 36.0 3.40 June 4— 5 _______________ 13.00 7.30 3,044.14 60.00 85.0 2.20 June 29-30 _______________ 60.00 50.00 1,066.67 144.00 411.0 2.60 July 2— 3 _______________ 87.00 73.00 1,557.58 307.00 982.0 4.60 Oct. 28 _______________ 14.00 13.00 911.68 32.00 51.0 .69 Nov. 8 _______________ 21.00 19.00 604.29 31.00 82.0 .88 Nov. 14 _______________ 39.00 34.00 675.38 62.00 184.0 1.00 Dec. 8— 9 _______________ 52.00 42.00 793.65 90.00 266.0 2.70 1973 Feb. 2 _______________ 31.00 24.00 1,157.41 75.00 86.0 2.40 Apr. 4 _______________ 20.00 15.00 888.89 36.00 112.0 4.10 Apr. 27 _______________ 27.00 20.00 740.74 40.00 138.0 16.00 May 28 _______________ 22.00 17.00 1,089.32 50.00 134.0 3.60 July 3— 4 _______________ 35.00 32.00 2,754.63 238.00 567.0 2.70 July 20—21 _______________ 41.00 37.00 2,072.07 207.00 320.0 .90 Sept. 14 _______________ 16.00 15.00 839.50 34.00 57.0 .45 Oct. 2 _______________ 6.30 5.40 1,028.80 15.00 30.0 .63 1974 Mar. 30 _______________ 67.00 60.00 802.47 130.00 365.0 2.00 Aug. 19 _______________ 11.00 11.00 673.40 20.00 82.0 .49 Sept. 3— 4 ___________________ 10.00 8.30 397.14 8.90 38.0 .43 Sept. 6 —7 _______________ 14.00 12.00 148.15 4.80 48.0 .68 Sept. 28 _______________ 13.00 12.00 305.55 9.90 98.0 .49 North Branch Rock Creek 1.966 Nov. 28—29 _______________ 33.00 20.00 500.00 27.00 63.0 5.20 1967 Jan. 27—28 _______________ 55.00 38.00 3,118.90 320.00 164.0 6.70 Mar. 6— 8 _______________ 308.00 266.00 2,826.51 2,030.00 800.0 10.00 Mar. 15—16 _______________ 65.00 38.00 506.82 52.00 98.0 11.00 May 7— 8 _______________ 90.00 66.00 1,352.41 241.00 187.0 7.10 June 22-23 _______________ 14.00 7.50 938.27 19.00 33.0 2.10 Aug. 3— 5 _______________ 92.00 78.00 5,935.41 1,250.00 317.0 1.70 Aug. 24—25 _______________ 218.00 203.00 3,995.62 2,190.00 823.0 2.50 Aug. 27-28 _______________ 54.00 36.00 6,224.27 605.00 132.0 9.10 Nov. 2— 3 _______________ 22.00 14.00 396.82 15.00 50.0 3.20 Dec. 3— 4 _______________ 68.00 50.00 496.30 67.00 104.0 4.70 Dec. 10~11 _______________ 82.00 62.00 388.29 65.00 163.0 7.20 1968 Jan 14—15 _______________ 204.00 182.00 830.28 408.00 413.0 5.50 Mar 12—13 _______________ 82.00 53.00 663.87 95.00 99.0 6.60 Mar. 17—18 _______________ 113.00 66.00 1,032.55 184.00 139.0 16.00 May 27—29 _______________ 138.00 118.00 1,117.39 356.00 233.0 4.20 June 16 _______________ 17.00 12.00 4,228.39 137.00 123.0 4.00 June 19—20 _______________ 48.00 37.00 3.533.53 353.00 196.0 7.60 June 26—28 _______________ 88.00 74.00 2,007.00 401.00 255.0 4.10 Sept. 10—11 _______________ 30.00 27.00 4,581.61 334.00 126.0 .71 Oct. 19—20 _______________ 12.00 6.20 77.66 1.30 16.0 1.90 in the Rock Creek and Anacostia River basins, 1962—74—Continued SEDIMENT DISCHARGE 25 Construc- T taJ Maximum Maximum Maximum Maximum _ tion area Antecedent .0 - pfe- 15-minute 30-minute l-hour 3-hour Peak Number of Time (Cp) days 0'1"“ng rainfall rainfall rainfall rainfall 1 ratio peaks ( Tm ) (percent (Ad) (. I; ) 5121:.) '(R30) (Rm) {Ran (Pr) (Np) (months) of. me es (in/hr) (in/hr) (in/hr) (m/hr) basm area) Olney, Md.—Continued 7 3.16 1.04 1.04 0.79 0.52 5.43 1 58 5.8 7 1.88 .68 .68 .55 -35 5-29 1 60 5-8 20 2.72 .52 .46 .39 .35 4.20 1 61 5.8 3 2 .78 .44 .34 .25 .20 5.44 1 61 5.8 6 a .91 .40 .28 .22 .12 4.37 1 62 5.8 9 1.23 .40 .34 .29 .17 5.20 1 64 5.8 8 1.57 .28 .28 .26 .19 4.04 1 64 5.8 3 1.80 .28 .26 .24 .20 4.82 1 64 5.8 12 1.52 .84 .48 .27 .20 3.58 1 65 5.8 4 1.19 .44 .30 .23 .21 3.16 3 66 6.5 0 1.48 2.60 1.54 .91 .44 7.88 1 66 6.5 8 1.88 .84 .76 .55 .34 6.74 1 67 7.0 9 21.06 .32 .28 .23 .19 4.19 1 67 7.4 1 2 .42 .84 .64 .39 .13 10.19 1 67 7.4 2 .76 2.68 2.52 1.31 .44 11.34 1 68 7.4 2 2.25 1.96 1.40 .73 .27 8.17 2 68 7.4 0 2.06 2.60 2.26 1.54 .66 13.39 2 69 7.4 8 1.54 1.84 1.20 .73 .31 3.87 2 72 6.5 10 1.52 .60 .52 .45 .30 4.27 1 73 6.5 5 2.00 .72 .52 .39 .34 5.38 1 73 6.5 1 1.94 .40 .38 .33 .30 6.27 1 74 6.5 3 1.21 .40 .32 .26 .18 3.48 2 76 6.5 1 .87 .44 .42 .38 .23 7.19 1 78 6.0 0 2 1.22 .48 .40 .35 .24 6.10 1 78 5.5 2 1.39 .68 .66 .49 .23 7.67 1 79 5.1 0 1.73 3.32 2.86 1.66 .57 17.63 1 81 5.1 15 2.49 2.60 2.48 1.43 .67 8.62 1 81 5.1 10 2.08 .80 .56 .50 .36 3.77 1 83 4.8 17 1.18 .28 .28 .23 .19 5.44 1 84 3.0 8 2.80 .72 .58 .47 .37 6.05 1 89 2.0 1 1.46 1.80 1.34 1.09 .45 7.41 1 94 1.0 4 1.57 .36 .24 .17 .11 4.53 2 95 1.0 1 1.46 .20 .18 .18 .17 3.94 1 95 1.0 20 1.85 1.08 .62 .48 .30 8.13 1 95 1.0 near N orbeck, Md. 23 1.14 0.32 0.30 0.24 0.14 2.89 1 1 2.8 16 .97 2.20 1.40 .94 .37 4.14 1 3 2.8 10 2.67 .72 .68 .55 .41 2.97 1 5 2.8 5 1.02 1.04 .78 .49 .30 2.29 1 5 2.8 9 1.94 .76 .70 .50 .41 2.73 1 7 3.1 43 1.17 1.48 1.28 .93 .35 4.12 1 8 3.1 12 2.42 4.24 2.30 1.15 .47 4.04 2 10 3.1 2 3.06 2.00 1.56 1.03 .59 4.04 1 10 3.1 1 a .76 .52 .46 .27 .09 3.41 2 10 3.1 7 .97 .52 .42 .34 .24 3.34 1 13 2.5 29 ’ 1.34 .64 .56 .49 .32 1.99 1 14 2.5 5 2 1.48 .44 .40 .38 .27 2.51 1 14 2.5 9 1.49 .52 .48 .46 .37 2.24 1 15 2.5 37 1.95 .52 .34 .32 .19 1.74 1 17 2.5 3 1.20 .40 .38 .33 .19 1.86 2 17 2.5 2 2.96 .80 .62 .50 .33 1.94 1 19 2.0 2 s 1.39 1.96 1.18 .63 .14 9.92 1 20 2.0 1 .93 2.84 2.28 1.16 .39 5.09 1 20 2.0 5 1.82 2.76 1.98 1.09 .62 3.39 2 20 2.0 3 1.80 3.56 2.62 1.58 .65 4.64 1 23 2.0 11 2 1.44 .72 .42 .35 .25 2.27 1 24 1.7 26 EFFECTS OF URBANIZATION ON STREAMFLOW AND SEDIMENT TRANSPORT IN MARYLAND TABLE 5.—Hydrologic characteristics and related factors for storms Total Storm Mean sedi- Susp'ended- Peak Antecedent .- . me 1; on- sedlment . . .3323. 73:7 73:? mm... “.33“ “5:33“ (ftfl/s-d) (ft3/s-d) ”(mg/1, (£353, (ft3/s) (us/s) North Branch Rock Creek near 1969 Jan. 21—22 ................. 32.00 17.00 261.44 12.00 42.0 5.20 June 2— 3 _________________ 39.00 35.00 2,677.24 253.00 158.0 1.50 June 18—19 _________________ 37.00 33.00 3,209.87 286.00 167.0 1.40 July 20-21 _________________ 17.00 16.00 1,921.29 83.00 117.0 .42 July 22—23 _________________ 67.00 64.00 2,777.77 480.00 332.0 3.10 Aug. 2— 3 _________________ 142.00 132.00 1,083.05 386.00 676.0 3.80 Aug. 9—10 _________________ 123.00 117.00 2,602.09 822.00 498.0 2.50 Aug. 18—20 _________________ 102.00 88.00 1,094.27 260.00 207.0 2.50 Sept. 3— 5 _________________ 193.00 176.00 2,735.69 1,300.00 706.0 2.40 Nov. 19-20 _________________ 27.00 16.00 671.30 29.00 64.0 3.50 Dec. 10—11 _________________ 86.00 68.00 1,612.20 296.00 255.0 8.20 Dec. 22 _________________ 37.00 31.00 657.11 55.00 106.0 5.60 1970 Apr. 1— 3 ................. 82.00 46.00 1,167.47 145.00 144.0 12.00 Apr. 14—15 _________________ 237.00 208.00 1,572.29 883.00 455.0 9.00 May 13—14 _________________ 69.00 52.00 3,262.10 458.00 255.0 8.10 May 24—25 _________________ 99.00 82.00 4,042.45 895.00 455.0 6.50 June 16—18 ................. 47.00 31.00 812.42 68.00 98.0 3.50 June 21—22 _________________ 63.00 51.00 2,316.63 319.00 287.0 4.70 July 9-10 _________________ 156.00 145.00 2,998.72 1,174.00 438.0 2.50 Sept. 10 _________________ 16.00 13.00 797.72 28.00 79.0 2.10 Nov. 4— 5 _________________ 111.00 95.00 1,255.36 322.00 327.0 11.00 Dec. 16—17 _________________ 50.00 31.00 585.42 49.00 130.0 4.70 Dec. 22—23 _________________ 78.00 54.00 576.13 84.00 165.0 6.50 1971 Feb. 7-10 _________________ 317.00 285.00 1,806.36 1,390.00 606.0 19.00 Feb. 13—14 _________________ 341.00 314.00 1,415.43 1,200.00 1,030.0 8.90 Feb 22—23 _________________ 175.00 147.00 1,345.43 534.00 335.0 14.00 Apr. 6— 7 _________________ 69.00 50.00 377.78 51.00 85.0 7.50 May 13—14 _________________ 148.00 130.00 1,735.04 609.00 498.0 5.40 May 16—17 _________________ 118.00 93.00 637.20 160.00 333.0 9.70 May 30-June 1 ______________ 176.00 138.00 791.73 295.00 415.0 8.00 June 2— 4 _________________ 109.00 70.00 3,910.05 739.00 382.0 15.00 July 1— 2 _________________ 40.00 29.00 2,094.51 164.00 186.0 4.30 July 29—30 _________________ 27.00 19.00 1,754.38 90.00 122.0 2.00 Aug. 1— 4 _________________ 323.00 293.00 1,222.35 967.00 998.0 7.00 Aug. 19 _________________ 22.00 17.00 849.67 39.00 112.0 3.40 Aug. 27—28 _________________ 184.00 171.00 1,407.84 650.00 642.0 3.00 Sept. 17—18 _________________ 25.00 11.00 841.75 25.00 59.0 5.60 Nov. 24—26 _________________ 286.00 239.00 692.70 447.00 657.0 7.10 Nov. 29—30 _________________ 85.00 57.00 487.33 75.00 237.0 14.00 Dec. 7— 8 _________________ 75.00 44.00 404.04 48.00 159.0 10.00 1972 Feb. 3— 4 _________________ 110.00 81.00 2,016.46 441.00 308.0 10.00 Mar. 16—17 _________________ 138.00 108.00 1,080.25 315.00 347.0 14.00 Mar. 22 _________________ 30.00 15.00 740.74 30.00 85.0 13.00 Apr. 13 _________________ 71.00 58.00 1,245.21 195.00 163.0 9.70 Apr. 16-17 _________________ 175.00 136.00 3,867.10 1,420.00 642.0 24.00 May 3— 4 _________________ 161.00 120.00 1,691.36 548.00 460.0 11.00 May 19—20 _________________ 58.00 33.00 549.94 49.00 112.0 9.90 May 30—31 _________________ 213.00 187.00 4,436.52 2,240.00 986.0 8.10 June 4— 5 _________________ 38.00 18.00 2,818.93 137.00 123.0. 9.70 Oct. 28 _________________ 46.00 39.00 1,291.55 136.00 128.0 4.30 Nov. 8 _________________ 75.00 65.00 695.16 122.00 268.0 5.50 Nov. 14—15 _________________ 151.00 128.00 801.50 277.00 398.0 5.10 1979 Feb. 2— 3 _________________ 168.00 126.00 1,055.26 359.00 335.0 13.00 Apr. 1— 2 _________________ 281.00 228.00 2,404.16 1,480.00 612.0 15.00 Apr. 4 _________________ 97.00 78.00 1,158.59 244.00 362.0 23.00 Apr. 27 _________________ 158.00 138.00 939.34 350.00 490.0 97.00 May 28—29 _________________ 144.00 110.00 1,040.40 309.00 415.0 15.00 July 3— 4 _________________ 146.00 127.00 1,971.42 676.00 660.0 31.00 July 20~21 _________________ 223.00 207.00 2,343.89 1,310.00 978.0 4.50 Sept. 14—15 _________________ 77.00 65.00 1,111.11 195.00 242.0 2.70 See footnotes at end of table. SEDIMENT DISCHARGE 27 in the Rock Creek and Anacostia River basins, 1962—74—Continued Construc- Maximum Maximum Maximum Maximum tion area Antecedent T9331 p.re- 15-minute 30-minute l-hour 3-hour Peak Number of Time (Cp) days C'pltatlon rainfall rainfal] rainfall rainfall 1 ratio peaks ( Tm.) (percent (Ad) . ‘ (R15) (Run) (Rm) (Rm) (Pr) (Np) (months) of (Inches) (in/hr) (in/hr) (in/hr) (in/hr) basin area) Norbeck, Md.—Continued 1 0.70 0.20 0.20 0.16 0.10 2.16 1 27 1.7 11 2.57 2.12 1.84 1.53 .74 4.47 1 32 1.5 14 2.14 4.40 3.12 2.13 .76 5.02 1 32 1.5 6 2 1.44 5.36 3.16 1.78 .60 7.29 1 33 1.5 1 2 1.24 5.88 2.96 1.48 .49 5.14 1 33 1.5 3 5 3.12 5.60 4.04 2.54 .88 5.09 1 34 1.5 3 2.36 2.72 1.46 .80 .63 4.24 1 34 1.5 6 2.79 3.52 2.06 1.07 .36 2.32 4 34 1.5 13 3.39 2.28 2.06 1.22 .60 4.00 1 35 1.5 9 1.10 1.16 .72 .47 .22 3.78 1 37 1.5 1 1.49 .52 .44 .40 .28 3.63 1 38 1.5 10 1.09 .24 .24 .20 .15 3.24 1 38 1.5 2 .87 .24 .22 .21 .13 2.87 2 42 1.5 8 2.98 .40 .38 .37 .27 2.14 1 42 1.9 18 1.10 2.48 1.26 .63 .38 4.75 1 43 1.9 5 1.49 2.56 1.48 .80 .32 5.47 1 43 1.9 20 1.74 2.40 1.54 .77 .26 3.05 2 44 1.9 2 1.50 1.96 1.86 .93 .39 5.54 1 44 1.9 16 2.84 2.00 1.52 1.04 .50 3.00 1 45 1.9 17 1.41 2.44 1.68 .90 .30 5.92 1 47 1.9 0 1.92 .60 .52 .45 .36 3.33 1 49 1.9 24 1.00 .24 .24 .20 .16 4.04 1 50 1.9 3 1.06 .36 .34 .28 .18 2.94 1 50 1.9 O 1.70 .60 .42 .30 .25 2.06 2 52 1.9 2 1.97 1.24 .80 .62 .36 3.25 1 52 1.9 7 1.65 .52 .30 .29 .20 2.18 1 52 1.9 16 9 1.53 .20 .16 .13 .10 1.55 1 54 2.2 3 2.14 1.32 1.06 .78 .48 3.79 1 55 2.2 1 1.32 .32 .30 .25 .22 3.48 1 55 2.2 3 1.41 .52 .36 .26 .20 2.95 2 55 2.2 0 1.21 .28 .28 .25 .20 5.24 2 56 2.2 22 1.09 1.12 .94 .84 .40 6.27 1 57 2.2 26 1.85 2.16 1.58 1.10 .42 6.32 1 57 2.2 1 4.70 4.80 3.66 1.99 .77 3.38 4 58 2.2 13 2 1.06 1.24 .98 .71 .31 6.39 1 58 2.2 7 3.06 1.04 1.04 .79 .52 3.74 1 58 2.2 3 2 .58 1.28 .92 .61 .21 4.85 1 59 2.2 20 2.78 .52 .46 .39 .35 2.72 1 61 2.2 2 3 .78 .44 .34 .25 .20 3.91 1 61 2.2 6 2 .88 .40 .28 .22 .12 3.39 1 62 2.2 9 1.13 .40 .34 .29 .17 3.68 1 64 2.5 12 1.47 .84 .48 .27 .20 3.08 1 65 2.5 4 2 .49 .36 .28 .24 .09 4.80 1 65 2.5 4 1.10 .44 .30 .23 .21 2.64 1 66 2.5 2 1.42 2.60 1.54 .91 .44 4.54 1 66 2.5 8 1.90 .84 .76 .55 .34 3.74 2 67 2.5 8 2 1.02 .32 .28 .23 .19 3.09 1 67 2.5 6 2 2.40 3.44 3.38 2.24 .84 5.23 1 67 2.5 2 .65 2.68 2.52 1.31 .44 6.29 1 68 2.5 8 1.60 1.84 1.20 .73 .31 3.17 1 72 2.5 10 1.48 .60 .52 .45 .30 4.04 1 73 2.5 5 2.07 .72 .52 .39 .34 3.07 1 73 2.5 3 1.25 .40 .32 .26 .18 2.56 1 76 2.0 5 1.89 1.48 1.16 .62 .28 2.62 1 78 2.0 1 .83 .44 .42 .38 .23 4.35 1 78 2.0 0 2 1.30 .48 .40 .35 .24 2.85 1 78 2.0 2 1.35 .68 .66 .49 .23 3.64 1 79 2.0 0 2.04 3.32 2.86 1.66 .57 4.95 1 81 2.0 15 2.72 2.60 2.48 1.43 .67 4.70 2 81 2.0 9 2.20 .80 .56 .50 .36 3.68 1 83 2.0 28 EFFECTS OF URBANIZATION ON STREAMFLOW AND SEDIMENT TRANSPORT IN MARYLAND TABLE 5.-—Hydralogic characteristics and related factors for storms Mean sedi- Suspended- Total Storm . Peak Antecedent Period of runoff runoff ment “.m‘ sed1ment discharge diSCharge sborm runoff (on (an “en???“ ”figure (0..) (Qa) (ft3/S-d) (ftfl/s-d) (mg/l) ”0:15) (ft3/s) (ftS/s) North Branch Rock Creek near 1 973—Continued Dec. _________________ 58.00 47.00 606.78 77.00 180.0 4.30 Dec. 20—22 _________________ 259.00 230.00 790.66 491.00 460.0 5.50 1974 Jan. 21 _________________ 84.00 76.00 1,769.00 363.00 322.0 9.30 Mar. 21 _________________ 35.00 27.00 1,536.35 112.00 204.0 7.10 Mar. 30—31 _________________ 363.00 327.00 1,449.77 1,280.00 ‘ 850.0 8.30 May 12—13 _________________ 76.00 55.00 1,245.79 185. 00 240.0 6.90 June 2 _________________ 51.00 40.00 509.26 55. 00 172.0 8.30 Aug. 19—20 _________________ 64.00 59.00 1,494.03 238. 00 465.0 2.20 Sept. 3— 4 _________________ 27.00 20.00 259.26 14. 00 55.0 1.40 Sept. 6— 7 _________________ 41.00 33.00 258.14 23.00 114.0 2.10 Sept. 28 _________________ 46.00 43.00 938.84 109.00 235.0 1.70 Manor Run near 1.966 Nov. 28 _________________ 4.40 4.00 1,203.70 13.00 23.0 0.40 1967' Jan. 27—28 _________________ 8.60 5.90 13,370.97 213.00 57.0 .65 Mar. 6— 8 _________________ 47.00 44.00 3,779.46 449.00 182.0 1.00 Mar. 15—16 _________________ 8.40 6.10 1,578.63 26. 00 26.0 1.10 Apr. 17—18 _________________ 2. 60 1.30 997.15 3. 50 5.0 .63 May 7— 8 _________________ 16. 00 14.00 7,037.02 266. 00 66.0 .78 June 22—23 _________________ 4.60 4.10 11,020.76 122. 00 110.0 .21 Aug. 3— 5 _________________ 8. 70 7. 80 20,750.21 437.00 77.0 .19 Aug. 24—25 _________________ 37. 00 36. 00 14,197.51 1,380.00 364.0 .27 Oct. 25 _____________ ~. - .. 2.50 2. 20 20,033.64 119.00 42.0 .26 Nov. 2 _________________ 2.60 2. 20 4,208.75 25.00 13.0 .31 1968 Jan. 14~15 _________________ 29. 00 26.00 1,695.15 119.00 97.0 .50 Mar. 12—13 _________________ 8. 40 5.90 2,385.43 38.00 21.0 .74 Mar. 17—18 _________________ 13. 00 9. 50 7,914.21 203.00 56.0 1.80 May 27—29 _________________ 19. 00 17.00 7,647.05 351.00 81.0 .43 June 12—13 _________________ 3. 20 2.20 9,764.29 58.00 19.0 .48 June 16 _________________ 10. 00 9.30 20,430.07 513.00 161.0 .43 June 19—20 _________________ 11. 00 9.30 19,633.58 493.00 194.0 1.10 June 26—28 _________________ 10. 00 8.00 12,962.95 280.00 114.0 .67 July 2— 3 _________________ 6. 50 5.20 16,452.97 231.00 95.0 .67 Sept. 10—11 _________________ 9. 50 8.90 24,219.69 582.00 196.0 .25 1969 Jan. 21~22 _________________ 5. 60 2. 60 3,133.90 22.00 10.0 .60 May 20—21 _________________ 9. 50 8. 40 14,814.79 336.00 210.0 .48 June 2—— 3 _________________ 19 00 17. 00 13,137.23 603. 00 252.0 .28 June 8 _________________ 2.30 1.90 4,288.49 22. 00 64.0 .34 June 18—19 _________________ 5.00 3.90 10,826.20 114.00 136.0 .34 July 22—23 _________________ 2,40 1.70 13,943.34 64.00 76.0 .34 July 28 _________________ 3.40 3. 00 11,111.09 90.00 105.0 .28 Aug. 9—10__-____-__‘ _______ 26.00 25. 00 11,703.69 790.00 335.0 .31 Aug. 18 _________________ 11.00 10. 00 8,222.20 222.00 203.0 .31 Sept. 3— 4 _________________ 11. 00 10. 00 8,999.98 243.00 151. 0 .60 Oct. 2 _________________ 3. 40 3.10 8,960.56 75.00 89.0 .25 Dec. 10—11 _________________ 16.00 14. 00 6,296.29 238.00 100.0 .40 1970 Apr. 1— 2 _________________ 8.00 5.30 10,062.88 144.00 54.0 1.30 Apr. 14—15 _________________ 32.00 29.00 4,699.86 368.00 126.0 .79 May 24—25 _________________ 14.00 12.00 11,141.96 361.00 277.0 .52 June 16 _________________ 8.90 8.50 8,932.45 205.00 262.0 .36 June 21—22 _________________ 11.00 9. 60 8,449.06 219.00 198.0 .39 July 20 _________________ 18.00 18. 00 3,703.70 180.00 448.0 .26 Aug. 14 _________________ 11.00 11. 00 4,814.80 143.00 214.0 .27 Sept. 10 _________________ 8.40 8.10 7,956.09 174.00 246.0 .65 Nov. 3 _________________ 2. 40 2. 00 1,851.85 10.00 30.0 .29 Nov. 4— 5 _________________ 24. 00 23. 00 1,465.38 91.00 116.0 2.40 Nov. 14—15 _________________ 12. 00 11. 00 2,491.58 74.00 102.0 1.00 Dec. 22 _________________ 11.00 8.80 2,020.20 48.00 81.0 .74 See footnotes at end of table. in the Rock Creek and Anacostia River basins, 1962—74—C0ntinued SE DIMENT DISCHARGE 29 Construc- Maximum Maximum Maximum Maximum tion area Antecedent T9911 Hre' 15-minute 30-minute l-hour 3-h0u1' Peak Number of Time (Cp) days 0121;33310'1 rainfall rainfall rainfall rainfall 1 ratio peaks (Tm) (percent (Ad) . 1’: (R13) (R30) (RM) {R310 (Pr) (Np) (months) of. ("‘c 95) (in/hr) (in/hr) (in/hr) (In/hr) bamn area) Norbeck, Md.—Continued 3 1.29 0.40 0.36 0.28 0.19 3.74 1 86 1.2 10 2 1.90 .40 .40 .26 .17 1.98 1 86 1.2 7 1.02 .56 .40 .36 .28 4.11 1 87 1.2 33 .85 .56 .44 .30 .22 7.29 1 89 1.2 7 2.78 .72 .58 .47 .37 2.57 1 89 1.2 28 1.38 1.16 .72 .51 .30 4.24 1 91 .5 19 1.64 .44 .32 .28 .22 4.09 1 92 .5 19 1.63 1.80 1.34 1.09 .45 7.84 1 94 .5 4 1.53 1.44 1.14 .74 .31 2.68 2 95 .5 1 1.45 .20 .18 .18 .17 3.39 1 95 .5 19 1.78 1.08 .62 .48 .30 5.43 1 95 .5 Norbeck, Md. 38 1.29 0.92 0.76 0.50 0.25 5.65 1 1 5.2 11 .82 1.20 .68 .37 .18 9.55 1 3 5.2 9 2.61 .96 .86 .67 .47 4.11 1 5 5.2 7 .95 1.32 .90 .74 .35 4.08 2 5 5.2 3 2 .66 1.24 1.00 .59 .26 3.36 1 6 10.8 3 1.98 .68 .52 .46 .37 4.66 1 7 10.8 37 1.50 3.76 2.42 1.33 .50 26.78 1 8 10.8 13 2.18 2.88 1.76 .91 .32 9.85 1 10 10.8 4 3.28 1.64 1.22 .88 .59 10.10 1 10 10.8 14 .97 2.12 1.32 .77 .34 18.97 1 12 11.8 7 .98 .48 .40 .33 .24 5.77 1 13 11.8 8 1.49 .32 .30 .29 .24 3.71 1 15 11.8 10 1.73 .60 .42 .39 .20 3.43 1 17 11.8 3 1.29 .60 .50 .41 .24 5.71 2 17 11.8 2 2.83 .52 .44 .37 .23 4.74 2 19 11.8 13 a .94 .64 .48 .38 .19 8.42 2 20 11.8 2 1.49 3.48 2.14 1.21 .40 17.27 1 20 11.8 1 .88 3.40 1.78 .89 .30 20.74 1 20 11.8 5 1.61 1.76 1.48 .81 .32 14.17 2 20 11.8 3 1.31 1.88 1.34 .91 .32 18.14 2 21 11.8 3 1.99 2.16 1.86 1.28 .53 21.99 1 23 11.8 1 .55 .12 .10 .08 .07 3.62 1 27 9.6 10 1.58 2.72 1.92 .99 .33 24.94 1 31 9.6 11 2.10 3.00 2.02 1.54 .74 14.81 1 32 9.6 4 .53 1.84 1.18 .60 .20 33.51 1 32 9.6 9 1.50 2.64 1.58 .84 .29 34.78 1 32 9.6 1 3 1.11 4.00 2.20 1.15 .42 44.51 1 33 9.6 5 2 1.12 2.28 1.24 .64 .24 34.91 1 33 9.6 5 2.83 2.84 1.50 .77 .75 13.39 2 34 9.6 7 2.50 3.60 2.18 1.15 .39 20.27 2 34 9.6 13 2.59 1.12 1.20 .66 .23 15.04 2 35 9.6 23 1.05 1.64 .98 .51 .22 28.63 1 36 9.6 1 1.45 .56 .46 .39 .27 7.11 1 38 9.6 2 .86 .88 .58 .33 .14 9.94 1 42 7.4 6 2.85 .52 .40 .36 .28 4.32 1 42 7.4 6 1.43 1.96 1.48 .75 .26 23.04 2 43 7.4 9 2.01 4.00 2.52 1.39 .46 30.78 2 44 7.4 3 1.45 2.48 1.28 .67 .37 20.58 3 44 7.4 9 2.38 3.08 2.34 1.39 .49 24.87 2 45 7.4 14 1.57 1.24 1.18 1.00 .45 19.43 1 46 7.4 17 1.76 3.08 2.40 1.35 .45 30.29 1 47 7.4 12 2 .53 .64 .38 .24 .12 14.85 1 49 7.4 1 1.90 .56 .54 .41 .34 4.94 1 49 7.4 2 1.02 .44 .38 .34 .21 9.18 1 49 7.4 4 1.15 .40 .34 .30 .18 9.12 1 50 7.4 30 EFFECTS OF URBANIZATION ON STREAMFLOW AND SEDIMENT TRANSPORT IN MARYLAND TABLE 5.—Hydrologic characteristics and related factors for storms Mean sedi- Suspended- . Total Storm _ d' . Peak Autecedent “5:311;sz :g ff 1215:3115 1:16;)? gigrgfign giggggg‘gte dischzrge dlschirge (fin/s41) (ft3/s-d) mg)“ (toss) (ft-Vs) (ftS/s) Manor Run near 1971 Feb. 7— 8 _________________ 36.00 33.00 2,368.12 211.00 161.0 2.00 May 13 _________________ 18.00 16.00 8,842.58 382.00 210.0 .52 July 29 _________________ 8.40 8.00 6,851.84 148.00 182.0 .35 Aug 1— 2 _________________ 9.60 8.10 3,475.07 76.00 185.0 .81 Aug 3— 5 _________________ 37.00 34.00 2,766.88 254.00 510.0 1.00 Aug- 27 _________________ 30.00 29.00 2,848.01 223.00 313.0 .39 Nov 24—25 _________________ 39.00 37.00 1,301.30 130.00 152.0 .53 1972 Feb. 3— 4 _________________ 17.00 15.00 1,925.92 78.00 105.0 .68 Feb. 13 _________________ 17.00 16.00 1,342.59 58.00 74.0 .57 Mar 16—17 _________________ 18.00 15.00 1,283.95 52.00 69.0 1.10 Apr 13 _________________ 8.30 6.70 829.19 15.00 46.0 .93 May 4 _________________ 12.00 9.80 2,380.95 63.00 134.0 1.00 June 29—30 _________________ 36.00 32.00 1,828.70 158.00 370.0 .89 Aug 27-28 _________________ 11.00 9.80 1,625.09 43.00 169.0 .39 Oct 28 _________________ 9.30 8.70 1,830.56 43.00 73.0 .32 Nov 14 _________________ 17.00 15.00 567.90 23.00 106.0 .44 Dec 8— 9 _________________ 25.00 20.00 648.15 35.00 141.0 1.10 1973 Apr. 1 _________________ 19.00 14.00 3,518.51 133.00 196.0 2.20 Apr. 4 _________________ 15.00 13.00 968.66 34.00 113.0 2.00 July 2 _________________ 13.00 13.00 2,222.22 78.00 240.0 .85 Sept. 14 _________________ 11.00 10.00 296.30 8.00 67.0 .37 Oct. 2 _________________ 2.80 2.30 111.11 0.69 17.0 .37 1974 Jan 21 _________________ 10.00 8.80 925.92 22.00 107.0 .76 May 12 _________________ 12.00 11.00 942.76 28.00 124.0 .53 Aug. 9 _________________ 7.90 7.40 750.75 15.00 124.0 .34 Aug. 19 _________________ 16.00 16.00 1,412.03 61.00 295.0 .28 Northwest Branch Anacostia 1967 Mar 15 _________________ 14.00 11.00 437.71 13.00 44.0 2.00 May I- 8 _________________ 23.00 16.00 4,351.84 188.00 75.0 1.40 June 22—23 _________________ 3.70 2.60 3,133.90 22.00 38.0 .52 Aug. 4— 5 _________________ 14.00 12.00 5,925.91 192.00 205.0 1.90 Nov. 3 _________________ 3.80 2.30 338.16 2.10 13.0 .57 1968 Jan 14—15 _________________ 46.00 42.00 379.19 43.00 139.0 2.20 Mar 12—13 _________________ 18.00 14.00 714.28 27.00 45.0 1.00 Mar. 17—18 _________________ 30.00 21.00 1,040.56 59.00 89.0 2.60 May 27—29 _________________ 39.00 33.00 819.30 73.00 121.0 .86 June 19—20 _________________ 14.00 10.00 2,925.92 79.00 159.0 1.10 June 27—28 __________________ 12.00 7.80 4,083.56 86.00 102.0 .79 1969 June 2— 3 _________________ 11.00 9.60 3,549.38 92.00 109.0 .45 June 18—19 _________________ 6.30 4.90 3,023.43 40.00 107.0 .42 July 20—21 _________________ 5.20 4.70 2,836.88 36.00 109.0 .05 Aug. 2— 3 _________________ 30.00 28.00 2,407.40 182.00 496.0 .49 Aug. 9—10 _________________ 43.00 41.00 2,511.29 278.00 441.0 .54 Aug. 18 _________________ 14.00 13.00 968.66 34.00 155.0 .58 Dec. 10—11 _________________ 32.00 28.00 1,124.34 85.00 162.0 1.50 1970 Apr. 2— 3 _________________ 26.00 20.00 1,962.96 106.00 110.0 2.90 Apr. 14—15 _________________ 83.00 76.00 1,793.37 368.00 322.0 1.80 May 24—25 _________________ 40.00 36.00 1,882.71 183.00 282.0 1.30 June 16-17 _________________ 9.00 5.90 1,067.17 17.00 91.0 1.00 July 9—10 _________________ 42.00 40.00 1,750.00 189.00 382.0 .86 July 20—21 _________________ 34.00 32.00 1,493.05 129.00 362.0 .86 Nov. 4— 5 _________________ 25.00 22.00 471.38 28.00 124.0 1.00‘ Dec. 22 _________________ 16.00 13.00 284.90 10.00 69.0 1.50 See footnotes at end of table. SEDIMENT DISCHARGE 31 in the Rock Creek and Anacostia River basins, 1962—74—C0ntinued Construc- T tal 1‘ Maximum Maximum Maximum Maximum tion area Antecedent .0 'ta?’ 9- 15-minute 30-minute l-hour 3-hour Peak Number of Time (Up) days CIDER )wn rainfall rainfall rainfall rainfall 1 ratio peaks (Tm) (percent (Ad) (. g 5) (R15) (R30) (Rm) (Rah) (19.) (Np) (months) of. ”1° e (in/hr) (in/hr) (in/hr) (in/hr) basm area) Norbeck, Md.—Continued 1 1.99 0.88 0.62 0.41 0.32 4.82 2 52 5.4 4 2.15 1.92 1.30 .76 .49 13.09 1 55 5.4 51 2.19 2.16 1.74 1.48 .56 22.71 1 57 5.4 1 2 1.63 1.72 1.48 .81 .36 22.74 2 58 5.4 1 2 3.30 4.80 3.66 2.00 .77 14.97 2 58 5.4 7 3.18 1.24 1.00 .95 .56 10.78 1 58 5.4 16 2.70 .52 .46 .37 .34 4.09 1 61 5.4 5 1.14 .44 .38 .32 .19 6.95 1 64 3.9 8 1.62 .36 .30 .28 .21 4.59 1 64 3.9 1 1.53 .68 .42 .32 .21 4.53 1 65 3.9 2 1.05 .40 .34 .27 .22 6.73 1 66 3.9 1 2 .64 .60 .58 .37 .16 13.57 1 67 3.9 3 1.74 1.36 .92 .60 .24 11.53 2 68 3.9 19 2 1.73 1.28 1.00 .85 .51 17.21 1 70 3.9 8 1.66 1.80 1.02 .60 .25 8.25 2 72 2.4 5 2.25 .60 .60 .45 .41 7.04 1 73 2.4 1 1.89 .56 .52 .41 .31 6.99 1 74 2.4 1 1.58 .36 .34 .32 .20 13.84 1 78 8 1 .94 .20 .16 .15 .13 8.54 1 78 8 1 a 1.12 1.16 1.02 .64 .34 18.40 2 81 8 7 2.02 .72 .52 .39 .27 6.66 2 83 8 17 1.00 .36 .26 .24 .20 7.23 1 84 8 8 .98 .80 .50 .47 .26 12.07 1 87 9 2 1.40 1.72 1.24 .78 .38 11.22 1 91 9 10 2.14 2.96 2.30 1.38 .59 16.71 2 94 9 9 2.13 3.28 2.84 1.76 .60 18.42 1 94 9 River at Norwood, Md. 7 0.95 4.24 2.14 1.12 0.50 3.82 2 5 1.1 52 2.03 .76 .74 .57 .40 4.60 1 7 1.1 45 1.30 3.16 2.14 1.16 .44 14.42 1 8 1.1 14 2.54 2.92 1.94 1.09 .38 16.92 1 10 1.1 7 1.06 .44 .42 .32 .23 5.40 1 13 .2 16 1.54 .44 .40 .35 .25 3.26 1 15 2 57 1.90 .68 .46 .43 .20 3.14 1 17 2 3 1.56 .68 .56 .46 .25 4.11 2 17 2 69 2.92 .56 .54 .41 .23 3.64 3 19 2 20 1.04 4.12 2.44 1.23 .41 15.79 1 20 2 6 1.75 1.92 1.10 .60 .25 12.98 1 20 2 69 2.10 2.88 1.68 1.13 .61 11.31 1 32 3 14 1.30 3.08 1.80 .97 .34 21.75 1 32 3 30 2 1.35 3.40 2.54 1.31 .44 23.18 1 33 3 3 1.79 1.88 1.48 1.00 .36 17.70 1 34 3 6 2.35 3.12 1.64 1.12 .74 10.74 1 34 3 7 a 1.66 4.16 2.24 1.17 .39 11.88 2 34 3 2 1.46 .56 .42 .38 .27 5.73 1 38 3 3 1.02 .58 .46 .29 .12 5.35 1 42 .4 11 3.16 .68 .60 .56 .39 4.21 1 42 .4 6 1.92 1.88 1.52 .94 .39 7.80 2 43 .4 21 1.64 1.68 1.62 .85 .29 15.25 1 44 .4 17 2.70 1.64 1.38 .97 .53 9.53 2 45 .4 9 2.37 3.20 2.26 1.49 .53 11.29 1 45 .4 12 2.02 .56 .48 .44 .34 5.59 1 49 .4 4 1.00 .36 .36 .30 .18 5.19 1 50 .4 32 EFFECTS OF URBANIZATION ON STREAMFLOW AND SEDIMENT TRANSPORT IN MARYLAND TABLE 5.—Hydrologic characteristics and related factors for storms Mean sedi- Suspended- . Total Storm _ ed‘ 1; . Peak Autecedent .3335... *st? 73?)“ 535%}?311 sisggg‘ge W ”“12“ (fts/s-d) (fts/s-d) (mg/l) (tons) (ft:"/s) (ft/s) Northwest Branch Anacostia River 1971 Feb. 7— 9 _______________ 94.00 84.00 1,009.70 229.00 280.0 5.40 May 13 _______________ 38.00 33.00 2,065.09 184.00 240.0 1.50 July 1— 2 _______________ 12.00 9.20 1,409.02 35.00 122.0 1.00 Aug. 1— 2 _______________ 31.00 28.00 1,058.20 80.00 214.0 2.60 Aug. 3— 4 _______________ 110.00 104.00 1,324.78 372.00 1,500.0 17.00 Aug. 27—28 _______________ 67.00 64.00 729.17 126.00 374.0 .77 Nov 24—25 _______________ 74.00 65.00 837.61 147.00 253.0 1.60 NOV 29—30 _______________ 23.00 16.00 324.07 14.00 84.0 2.90 Dec. 7 _______________ 18.00 15.00 271.60 11.00 70.0 2.10 1972 Feb. 3— 4 _______________ 39.00 31.00 1,851.85 155.00 200.0 2.10 Feb. 13 _______________ 42.00 39.00 1,082.62 114.00 155.0 1.80 Feb. 18—19 _______________ 57.00 42.00 1,772.48 201.00 218.0 6.00 Mar. 16—17 _______________ 50.00 41.00 1,056.91 117.00 141.0 3.20 Apr. 16—17 _______________ 40.00 30.00 1,024.69 83.00 143.0 7.40 May 3— 4 _______________ 55.00 41.00 2,005.42 222.00 280.0 2.80 June 29—30 _______________ 66.00 54.00 1,021.95 149.00 582.0 2.70 July 2— 3 _______________ 57.00 46.00 998.39 124.00 338.0 4.70 July 16 _______________ 70.00 67.00 989.50 179.00 980.0 2.30 Oct. 28 _______________ 7.00 5.80 178.80 2.80 31.0 .86 NOV. 8 _______________ 15.00 13.00 267.81 9.40 67.0 1.10 NOV. 14 _______________ 35.00 30.00 555.55 45.00 209.0 1.30 Dec. 8— 9 _______________ 58.00 48.00 632.72 82.00 308.0 3.10 1973 Apr. 1— 2 _______________ 58.00 39.00 2,174.74 229.00 313.0 4.30 Apr. 4 _______________ 30.00 25.00 874.07 59.00 215.0 4.70 Apr. 27 _______________ 33.00 24.00 586.42 38.00 191.0 9.50 July 3 _______________ 16.00 14.00 1,931.21 73.00 210.0 2.10 July 20—21 _______________ 29.00 25.00 1,244.44 84.00 225.0 1.20 Dec. 20—21 _______________ 78.00 73.00 826.99 163.00 168.0 1.30 1974 Jan. 21 _______________ 16.00 13.00 826.21 29.00 96.0 2.10 Mar. 30 _______________ 76.00 69.00 1,175.52 219.00 330.0 2.40 June 2 _______________ 16.00 12.00 740.74 24.00 81.0 1.90 Aug. 19 _______________ 7.70 6.90 2,039.72 38.00 116.0 .77 Nursery Run at 1967 May 7— 8 _______________ 3.50 2.40 154.32 1.00 9.6 0.41 July 29—30 _______________ 1.90 1.40 2,566.13 9.70 31.0 .14 Aug 3— _______________ 5.10 3.90 1,519.47 16.00 42.0 .16 Aug 24—25 _______________ 13.00 12.00 1,450.62 47.00 120.0 .25 Oct 25—26 _______________ 1.10 .54 411.52 .60 8.7 .16 Nov. 2 _______________ .83 .60 61.73 .10 3.7 .17 1968 Jan 14—15 _______________ 7.10 6.40 567.13 9.80 27.0 .18 Mar 12—13 _______________ 2.00 1.00 148.15 .40 4.6 .26 Mar 17~18 _______________ 3.60 2.40 92.59 .60 6.8 .49 May 27—28 _______________ 3.30 2.30 209.34 1.30 8.7 .21 June 19 _______________ 2.40 2.10 3,350.96 19.00 61.0 .36 June 26—28 _______________ 1.40 .61 60.72 .10 2.9 .24 Sept. 10——11 _______________ .82 .54 137.17 .20 6.4 .10 1969 Jan. 21 _______________ .94 .69 118.09 .22 1.8 .22 June 2— 3 _______________ .91 .64 144.68 .25 6.0 .12 June 18—19 _______________ 1.90 1.50 5,185.18 21.00 35.0 .12 July 22—23 _______________ 1.30 .95 779.73 2.00 14.0 .19 Aug. 2— 4 _______________ 5.40 4.20 3,350.97 38.00 72.0 .19 Aug. 18 _______________ 2.10 1.70 784.31 3.60 25.0 .17 Sept. 2— 3 _______________ 1.40 .79 797.00 1.70 9.0 .15 Sept. 4 _______________ 2.50 2.10 1,164.02 6.60 32.0 .64 Oct. 2— 3 _______________ 1.00 .63 199.88 .34 4.6 .16 Dec. 10—11 _______________ 3.90 2.90 447.00 3.50 18.0 .31 Dec. 22 _______________ 2.00 1.50 246.91 1.00 8.0 .25 See footnotes at end of table. SEDIMENT DISCHARGE 33 in the Rock Creek and Anacostia River basins, 1962—74—Continued Construc- T tal re- Maxi-mum Maximum Maximum Maximum ' Lion area Antecedent .0 . p. 15-mmute 30-mmute l-hour 3-hour Peak Number of Time (0,.) days 0121138310“ rainfall ramfall rainfall rainfall 1 ratio peaks (Tm) (percent (Ad) (. lie ) '(st) .(Rw) (Ru) {Rah} (Pr) (Np) (months) of: ”‘c S (In/hr) (In/hr) (In/hr) (In/hr) basm area) at Norwood, Md.—Continued 11 1.79 0.44 0.40 0.32 0.25 3.27 2 52 .2 4 2.26 1.52 1.12 .74 .50 7.23 1 55 .2 27 1.74 2.40 1.90 1.28 .49 13.15 1 57 .2 2 2 1.99 3.88 2.20 1.22 .49 7.55 2 58 .2 1 2 3.51 4.92 4.40 2.48 .90 14.26 1 58 .2 21 3.69 .96 .84 .76 .62 5.83 1 58 .2 20 3.09 .48 .46 .44 .37 3.87 1 61 .2 3 2 .84 .44 .36 .27 .19 5.07 1 61 .2 7 2 .91 .24 .18 .16 .12 4.53 1 62 .2 23 1.27 .48 .46 .39 .21 6.38 1 64 .1 9 1.69 .24 .22 .22 .20 3.93 1 64 .1 5 2.20 .48 .48 .47 .38 5.05 1 64 .1 12 1.60 .24 .24 .21 .09 3.36 1 65 .1 2 1.18 .64 .48 .36 .22 4.52 3 66 .1 10 2.22 1.40 .88 .63 .35 6.76 1 67 .1 5 1.74 3.24 2.06 1.13 .40 10.73 1 68 .1 1 1.73 2.20 1.98 1.37 .60 7.25 1 69 .1 7 1.87 3.16 2.16 1.27 .54 14.59 2 69 .1 8 1.55 .68 .68 .46 .26 5.20 2 72 .1 10 1.55 .64 .52 .46 .32 5.07 2 73 .1 5 1.90 .68 .50 .41 .31 6.92 1 73 .1 1 2 1.72 .68 .54 .44 .34 6.35 1 74 .1 5 1.65 1.36 .94 .52 .27 7.92 2 78 .9 1 .96 .60 .50 .47 .27 8.41 1 78 .9 15 2 1.03 .56 .46 .32 .20 7.56 1 78 .9 2 1.14 2.64 1.96 1.20 .40 14.85 1 81 .9 16 3.08 3.32 2.38 1.34 .77 8.95 ' 2 81 .9 10 2 2.08 .48 .40 .31 .20 2.28 2 86 .9 9 .88 .48 .38 .36 .24 7.22 1 87 2.5 8 3.01 .84 .68 .56 .46 4.75 1 89 2.5 20 1.81 .48 .42 .32 .25 6.59 1 92 2.5 8 1.32 2.56 1.82 1.25 .46 16.70 1 94 2.5 Cloverly, Md. 3 2.00 0.88 0.80 0.65 0.59 3.83 1 7 0.0 7 2 1.47 3.20 2.56 1.44 .49 22.04 1 9 .0 3 2.98 3.12 2.12 1.32 .47 10.73 2 10 .0 3 3.67 2.52 1.56 1.10 .61 9.98 1 10 .0 ‘5 .95 1.88 1.28 .73 .31 15.81 1 12 .0 7 1.03 .40 .34 .28 .21 5.88 1 13 .0 15 1.61 .56 .46 .40 .17 4.19 1 15 .0 56 1.76 .72 .50 .42 .19 4.34 2 17 .0 3 1.68 .52 .44 .35 .23 2.63 2 17 .0 33 3.15 .48 .42 .38 .24 3.69 2 19 .0 1 1.24 3.80 2.30 1.15 .39 28.88 1 20 .0 5 1.26 1.48 .94 .51 .18 4.36 2 20 .0 3 1.41 .84 .64 .58 .33 11.67 1 23 .0 1 .70 .12 .12 .10 .07 2.29 1 27 .0 11 1.54 1.20 .76 .50 .33 9.19 1 32 .0 14 1.70 5.16 2.94 1.54 .52 23.25 1 32 .0 1 2 1.48 3.48 2.06 1.18 .44 14.54 1 33 .0 4 3.40 3.88 2.86 1.86 .88 17.10 2 34 .0 7 2 1.82 2.88 1.94 1.12 .38 14.61 2 34 .0 12 1.90 1.84 1.76 .91 .33 11.20 1 35 .0 1 1.05 1.52 1.32 .80 .29 14.93 1 35 .0 23 1.15 .68 .54 .31 .20 7.05 1 36 .0 1 1.93 .60 .52 .46 .34 6.10 1 38 .0 9 1.20 .52 .44 .39 .30 5.17 1 38 .0 34 EFFECTS OF URBANIZATION ON STREAMFLOW AND SEDIMENT TRANSPORT IN MARYLAND TABLE 5.—Hydrologic characteristics and related factors for storms Total Storm Mean sedi- Suspended- Peak Antecedent . - ed . . $5,125.25 r333? ‘73:? 322%. 3.59;: W39 “532:” (ftfi/s-d) (fh'l/s-d) (mg/l) (mus) (MB/s) (ft-Vs) Nursery Run at 1970 Apr. 2 _________________ 2.40 1.20 1,728.39 5.60 12.0 0.64 Apr. 14 _________________ 6.00 5.20 619.66 8.70 28.0 .48 Apr. 24 _________________ 1.00 .26 128.20 .09 1.8 .77 June 3 _________________ 4.20 3.90 3,893.63 41.00 113.0 .28 June 16—17 _________________ 2.30 1.30 1,111.11 3.90 26.0 .31 July 9—10 _________________ 3.60 2.60 897.43 6.30 27.0 .22 July 20—21 _________________ 3.90 3.00 2,716.05 22.00 74.0 .22 Nov 4— 5 _________________ 3.60 2.70 589.85 4.30 15.0 .35 Dec 22 _________________ 1.60 .96 246.91 .64 7.1 .35 1971 Feb. 7— 9 _________________ 9.30 7.50 938.27 19.00 28.0 .58 May 13 _________________ 2.00 1.50 839.50 3.40 13.0 .31 July 1— 2 _________________ 2.70 2.10 4,938.26 28.00 72.0 .22 July 29—30 _________________ 1.20 .68 816.99 1.50 9.8 .15 Aug 3— 4 _________________ 7.30 6.20 4,599.75 77.00 260.0 1.90 Aug 27 _________________ 6.00 5.50 1,010.10 15.00 44.0 .19 Nov 29 _________________ 1.60 .91 227.92 .56 6.4 .57 Dec 7 _________________ 1.80 1.00 259.26 .70 5.4 .48 1972 Feb. 3— 4 _________________ 4.00 2.60 1,709.40 12.00 20.0 .45 Feb. 18—19 _________________ 5.90 3.90 769.23 8.10 22.0 .61 June 21—22 _________________ 69.00 66.00 1,402.92 250.00 695.0 .38 NOV 14 _________________ 3.20 2.60 470.08 3.30 16.0 .32 Dec 8— 9 _________________ 4.40 3.00 259.26 2.10 19.0 .50 1973 Apr. 1— 2 _________________ 5.10 3.10 1,911.59 16.00 25.0 .68 Apr. 4 _________________ 2.60 1.80 493.83 2.40 17.0 .78 July 3 _________________ .80 .34 1,633.98 1.50 8.2 .46 July 20 _________________ 2.20 1.90 2,729.04 14.00 25.0 .26 Sept. 14 _________________ 1.40 1.10 175.08 .52 4.3 .18 1974 Aug. 19 _________________ .69 .51 1,016.70 1.40 9.4 .16 Batchellors Run 1967 Aug. 4— 5 _________________ 2.40 1.80 555.55 2.70 19.0 0.20 1968 Jan. 14 _________________ 12.00 10.00 185.18 5.00 41.0 .30 May 28 _________________ 4.30 2.60 455.84 3.20 17.0 1.70 June 19—20 _________________ 6.30 5.40 1,303.15 19.00 133.0 .30 June 27—28 _________________ 1.90 .70 227.51 .43 7.0 .40 1969 June 2— 3 _________________ 2.50 1.90 1.364.52 7.00 20.0 .20 June 18—19 _________________ 1.50 1.00 3,148.14 8.50 21.0 .10 Dec. 10—11 _________________ 6.00 5.10 798.84 11.00 32.0 .20 1970 Apr. 14 _________________ 12.00 10.00 1,407.41 38.00 67.0 .40 July 20—21 _________________ 5.40 4.90 400.60 5.30 102.0 .10 Nov. 4— 5 _________________ 4.20 3.60 226.34 2.20 14.0 .20 Dec. 22 _________________ 3.20 2.80 145.50 1.10 19.0 .20 1972 Feb. 3— 4 _________________ 7.10 5.70 1,039.63 16.00 39.0 .40 May 4 _________________ 4.20 3.50 1,269.84 12.00 54.0 .60 Aug. 27—28 _________________ 2.50 1.80 308.64 1.50 18.0 .20 Oct. 28 _________________ 1.90 1.40 203.70 .77 5.7 .30 Nov. 8 _________________ 2.50 2.20 144.78 .86 ‘10.0 .20 Nov. 14 _________________ 6.60 5.90 238.54 3.80 39.0 .30 Dec. 8— 9 _________________ 11.00 8.30 446.23 10.00 76.0 .60 1973 Feb. 2 _________________ 4.80 3.30 426.49 3.80 14.0 .50 Apr. 1— 2 _________________ 8.20 5.20 1,282.05 18.00 70.0 1.00 See footnotes at end of table. SEDIMENT DISCHARGE 35 in the Rock Creek and Anacostia River basins, 1962—74—Continued Construc- T t 1 , _ Maximum Maximum Maximum Maximum ' tion area Antecedent 9 ’3 p.” 15-minute 30-minuhe l-hour 3-hour Peak Number of Time (012) days “171:2“; on rainfall rainfall rainfall rainfall 1 ratio peaks (Tm ) (percent) (Ad) (indies) (R15) (Rao) (Ru) (Run) (Pr) (Np) (months) of (in/hr) (in/hr) (in/hr) (in/hr) basin area) Cleverly, Md.—Continued 3 1.11 1.04 0.64 0.37 0.18 9.47 1 42 0.0 10 4.32 .72 .68 .55 .43 5.29 1 42 .0 3 .38 .48 .36 .19 .06 3.96 1 42 .0 8 1.97 3.56 3.22 2.37 .80 28.90 1 44 .0 8 1.57 4.05 2.70 1.36 .45 19.76 1 44 .0 6 2.79 1.92 1.40 1.05 .45 10.30 2 45 .0 8 2.42 2.72 2.48 1.54 .54 24.59 2 45 .0 12 2.33 .68 .60 .51 .36 5.43 1 49 .0 4 1.15 .36 .36 .31 .19 7.03 1 50 .0 11 1.75 .56 .42 .33 .25 3.66 2 52 .0 4 2.02 1.20 .84 .66 .42 8.46 1 55 .0 15 2.30 4.60 3.70 1.90 .66 34.18 1 57 .0 26 1.88 2.48 1.58 1.16 .43 14.19 2 57 .0 1 2 2.64 4.52 3.16 1.64 .63 41.63 1 58 .0 7 3.58 .84 .80 .62 .48 7.97 3 58 .0 3 2 .83 .36 .36 .32 .21 6.41 1 61 .0 6 2 1.05 .24 .20 .16 .13 4.92 1 62 .0 9 1.39 .36 .34 .32 .22 7.52 1 64 .0 3 2.10 .20 .18 .18 .17 5.48 1 64 .0 2 2 8.23 2.64 2.50 2.17 1.28 10.52 4 68 .0 5 1.77 .68 .60 .49 .41 6.03 1 73 .0 1 1.92 .56 .52 .40 .32 6.17 1 74 .0 1 1.86 2.48 1.40 .77 .38 7.85 2 78 .4 1 1.08 .76 .62 .62 .34 9.01 1 78 .4 2 .87 2.64 1.50 .87 .29 22.76 1 81 .4 8 2 2.32 2.60 2.40 1.34 .72 13.02 2 81 .4 7 2.29 .56 .56 .44 .34 3.75 1 83 .4 2 1.23 1.76 1.58 .99 .38 18.12 1 94 .0 at Oakdale, Md. 1 2.18 2.92 1.86 1.04 0.36 10.44 2 10 1.7 9 1.55 .44 .42 .37 .28 4.07 1 15 .7 3 2.94 .60 .48 .38 .24 5.88 1 19 .7 1 1.23 2.84 2.28 1.16 .39 24.57 1 20 .7 1 2 .58 1.92 1.28 .68 .19 9.43 1 20 .7 11 2.10 2.72 1.82 1.45 .61 10.42 1 32 .7 14 1.50 3.60 2.50 1.35 .46 20.90 1 32 .7 1 1.43 .44 .42 .35 .25 6.24 1 38 .7 6 3.05 .52 .42 .38 .34 6.66 1 42 .7 9 2.51 1.44 .96 .50 .23 20.80 1 45 .7 1 2.08 .60 .54 .44 .36 3.83 1 49 .7 4 1.06 .40 .36 .30 .18 6.71 1 50 .7 9 1.17 .48 .38 .33 .19 6.77 1 64 .3 1 2 1.55 .60 .50 .35 .23 15.26 1 67 .3 17 2 1.62 1.64 1.10 .88 .50 9.89 1 70 .3 8 1.64 1.56 1.04 .61 .26 3.86 1 72 .3 10 1.57 .72 .66 .57 .36 4.45 1 73 .3 5 1.91 .80 .54 .40 .33 6.56 2 73 .3 1 1.88 1.64 1.04 1.01 .57 9.08 1 74 .3 1 1.05 .40 .32 .26 .18 4.09 2 76 .3 5 1.70 2.20 2.04 1.30 .58 13.27 1 78 .3 36 EFFECTS OF URBANIZATION ON STREAMFLOW AND SEDIMENT TRANSPORT IN MARYLAND TABLE 5.—Hydrologic characteristics and related factors for storms Mean sedi- Suspended- Total Storm . Peak Antecedent .- . - d . . £591.23. W “:53? 53533311 $191333: ““2"“ “:32?“ (Ms-d) (us/Sm (mg )1) (£355), (nu/s) (m5, Batchellors Run at 1 9 73—Continued July 3— 4 _________________ 6.70 5.40 672.15 9.80 122.0 0.80 July 20—21 _________________ 7.60 7.10 573.81 11.00 130.0 .20 Sept. 14 _________________ 1.20 .80 189.81 .41 3.6 .20 1.971; Jan. 21—22 _________________ 3.80 1.90 467.84 2.40 25.0 .60 Aug. 19-20 _________________ 4.00 2.70 1,508.91 11.00 39.0 .50 Bel Prc Creek at 1963 June 2— 3 _________________ 23.00 22.00 286.19 17.00 61.0 0.20 Aug. 20—21 _________________ 20. 00 20. 00 203.70 11. 00 102. 0 .30 Sept. 29 _________________ 4. 50 4. 20 114.64 1. 30 18.0 .10 Nov. 6— 7 _________________ 40. 00 38. 00 136.45 14. 00 80.0 2.00 1964 Jan. 9—10 _________________ 27.00 26. 00 170.94 12. 00 117.0 .40 Aug. 3— 4 _________________ 1. 30 1 2.0 277.78 .90 14.0 .10 Dec. 27—28 _________________ 11. 00 10. 00 137.04 3. 70 31. 0 .20 1.965 Mar. 5— 6 _________________ 75.00 71.00 944.18 181.00 260.0 .20 Aug. 26—27 _________________ 71.00 70.00 4,851.84 917.00 410.0 .10 1966 Jan. 6 _________________ 6.30 4.50 1,646.09 20.00 22.0 1.20 Feb. 13 _________________ 65.00 63.00 2,316.28 394.00 180.0 1.50 Apr. 12—13 _________________ 23.00 19.00 2,358.67 121.00 62.0 .50 Sept. 14 _________________ 59.00 55.00 2,249.15 334.00 370.0 .30 Oct. 18—19 _________________ 35.00 34.00 1,361.65 125.00 84.0 .30 NOV. 28-29 _________________ 13.00 12.00 1,666.66 54.00 63.0 .28 1967 Jan. 27—28 _________________ 9.50 8.30 4,283.79 96.00 56.0 .58 Mar. 6— 8 _________________ 59.00 55.00 2,545.45 378.00 193.0 .98 Mar. 15—16 _________________ 18. 00 15. 00 1,407.41 57.00 52.0 .98 Apr. 17—18 _________________ 3. 80 2. 50 1,629.63 11.00 11.0 .58 May 7- 8 _________________ 34. 00 30. 00 2,037.03 165.00 96.0 1.10 June 22—23 _________________ 10. 00 9.00 4,485.59 109.00 130. 0 .39 July 20—21 _________________ 7. 30 6. 60 3,198.65 57.00 85.0 .17 Aug. 3— 5 _________________ 26.00 24. 00 8,009.25 519.00 158.0 .24 Aug. 24—25 _________________ 67. 00 64. 00 2,575.23 445.00 338.0 .57 Oct. 25—26 __________________ 12.00 11. 00 4,377.10 130.00 135.0 .25 Nov. 2— 3 _________________ 9.30 8. 30 1,383.31 31.00 58.0 .29 Dec 3— 4 _________________ 31.00 28. 00 396.82 30.00 68.0 .41 Dec. 28—29 _________________ 29.00 27.00 644.72 47.00 81.0 .49 1968 Jan 14—15 _________________ 48.00 45.00 510.29 62.00 159.0 .30 Mar 12-13 _________________ 25.00 22.00 993.26 59.00 63.0 .46 Mar. 17—18 _________________ 35.00 28.00 1,574.07 119.00 95.0 1.90 May 27—29 _________________ 44.00 40.00 1,074.07 116.00 106.0 .32 June 16—17 _________________ 17.00 15.00 1,567.90 104.00 137.0 .37 June 19—20 _________________ 16.00 14.00 2,804.23 106.00 178.0 1.50 June 27—28 _________________ 20.00 18.00 5,020.56 244.00 134.0 .40 July 2— 3 _________________ 16.00 15.00 4,395.05 178. 00 114.0 .27 Sept. 6— 7 _________________ 2.70 2.50 962.96 6. 50 29.0 .15 Sept. 10—11 _________________ 9. 70 9. 20 2,173.91 54. 00 104. 0 .11 Oct. 6— 7 _________________ 4. 70 4. 40 690.23 8. 20 34. 0 .17 1.96.9 Jan. 21—22 _________________ 12.00 11.00 505.05 15.00 23.0 .47 May 20—21 _________________ 15.00 14.00 3,412.69 129.00 192.0 .53 June 2— 3 _________________ 29. 00 29. 00 1,915.71 150.00 206.0 .10 June 18—19 _________________ 7. 20 6. 50 3,646.72 64.00 113.0 .21 July 20—21 _________________ 14. 00 14. 00 10,952.36 414.00 214.0 .09 July 22-23 _________________ 6.00 5.10 5,010.88 69.00 57.0 1.20 Aug. 9—10 _________________ 58.00 57. 00 1,858.35 286.00 307. 0 .19 Sept. 2- 4 _________________ 12.00 11. 00 2,794.61 83.00 164.0 .23 Nov.19—20 _________________ 11.00 10.00 888.89 24.00 66.0 .21 Dec. 10—11 _________________ 34.00 33.00 808.08 72.00 153.0 .67 See footnotes at end of table. SEDIMENT DISCHARGE 37 in the Rock Creek and Amcostia River basins, 1962—74—Continued Construc- T 1 _ Maximum Maximum Maximum Maximum . tion area Anwcedent cgtgtagre Iii-minute 30-minute l-hour 3—hour Peak Number of Txme (01)) days DER: )“m rainfall rainfall rainfall rainfall 1 ratio peaks ( Tm ) (percent) (Ad) (. ches) $1315) {1330) {R110 (-13%) (Pr) (Np) (months) Of. m (Ln/hr) (In/hr) (In/hr) (In/hr) basm area) Oakdale, Md.—Continued 3 0.99 2.36 1.92 1.21 0.40 22.44 1 81 0.3 15 2 2.21 2.60 2.48 1.43 .67 18.28 1 81 3 11 1.87 .72 .54 .41 .31 4.25 1 83 3 8 1.08 .60 .50 .45 .30 12.84 1 87 4.0 1 2.23 2.40 2.16 2.01 .82 14.26 1 94 4.0 Layhill, Md. 2 2 2 95 0.44 0.36 0.33 0.22 2.76 1 3 0.0 5 2 2.75 1.94 1.72 1.00 .52 5.08 2 5 .0 12 2 2.10 .56 .40 .39 .26 4.26 1 6 .0 3 5 3.30 .40 .36 .35 .30 2.05 1 8 .0 1 2 1.16 .36 .32 .25 .22 4.48 1 10 0 20 2 1.95 .84 .52 .40 .20 11.58 1 17 0 13 2 .72 .60 .40 .25 .13 3.08 1 21 3 1 7 2 2 95 2 64 1.34 .67 22 3 66 1 24 18.0 3 2 2 81 3 32 2.24 1.40 49 5 86 2 29 18. 88 2 .89 .16 .14 .12 .07 4.62 1 34 14.0 1 2 1.73 1.48 .80 .43 .17 2.83 2 35 14.0 17 ”2.20 .20 .18 .15 .13 3.24 1 37 14.0 19 2 5.48 1.12 1.10 .95 .72 6.72 2 42 14.0 14 2 2.45 .44 .30 .25 .22 2.46 1 43 13.8 35 1 29 92 .76 50 25 5.23 1 44 13.8 18 .75 1.20 .68 .37 .18 6.68 1 46 13.8 12 2.55 .96 .86 .67 .47 3.49 1 48 13.8 7 .85 1.32 .90 .74 .35 3.40 2 48 13.8 9 2 .66 1.24 1.00 .59 .26 4.17 1 49 13.9 3 2.01 .68 .52 .46 .37 3.16 1 50 13.9 33 1.49 3.76 2.42 1.33 .50 14.40 1 51 13.9 16 .96 1.28 .80 .47 .23 12.85 1 52 13.9 3 2.28 2.88 1.76 .91 .32 6.57 2 53 13.9 3 3.69 1.64 1.22 .88 .59 5.27 3 53 13.9 6 1.14 2.12 1.32 .77 .34 12.25 1 55 10.0 7 1.02 .48 .40 .33 .24 6.95 1 56 10.0 30 2 1.42 .40 .34 .30 .21 2.41 1 57 10.0 4 2 1.48 .48 .36 .22 .17 2.98 1 57 10.0 15 1.36 .32 .30 .29 .24 3.53 1 58 10.0 37 1.79 .60 .42 .39 .20 2.84 1 60 10.0 3 1.49 .60 .50 .41 .24 3.32 2 60 10.0 2 2.76 .52 .44 .37 .23 2.64 1 62 9.8 2 2 1.21 3.48 2.14 1.21 .40 9.11 1 63 9.8 1 1.04 3.40 1.78 .89 .30 12.61 1 63 9.8 6 1.70 1.76 1.48 .81 .32 7.42 2 63 9.8 3 2 1.36 1.88 1.34 .91 .32 7.58 2 64 9.8 19 a .76 .64 .48 .37 .19 11.54 1 66 9.8 3 1.50 2.16 1.86 1.28 .53 11.29 2 66 9.8 26 2 1.18 .36 .34 .32 .27 7.69 2 67 9.8 28 .56 .12 .10 .08 .07 2.05 1 70 9.8 10 1.47 2.72 1.92 .99 .33 13.68 1 74 11.1 11 1.92 3.00 2.02 1.54 .74 7.10 1 75 11.1 9 1.34 2.64 1.58 .84 .29 17.35 2 75 11.1 30 2 1.13 3.60 2.14 1.11 .38 15.28 1 76 11.1 1 2 1.11 4.00 2.20 1.15 .42 10.94 1 76 11.1 4 3.16 2.84 1.50 . .77 .75 5.38 2 77 11.1 12 1.48 1.12 1.20 .66 .23 14.89 6 78 11.1 10 1.15 1.04 .70 .48 .27 6.58 1 80 11.1 1 1.45 .56 .46 .39 .27 4.62 1 81 9.7 38 EFFECTS OF URBANIZATION ON STREAMFLOW AND SEDIMENT TRANSPORT IN MARYLAND TABLE 5.—Hydrologie characteristics and related factors for storms Mean sedi- Suspended- “ Total Storm _ d‘ t _ Peak Agtecedent ..£’::.:‘:$.:f.. 73?? 733? 172.122.12.931axschye “7.32;“ (ftp/s-d) (N/s-d) (mg/l) (tons) (mi/s) (ft3/s Bel Pre Creek at 1970 Apr. 1— 2 _________________ 23.00 21.00 1,975.31 112.00 117.0 2.00 Apr. 14—15 _________________ 79.00 76.00 1,310.91 269.00 234.0 .78 July 9—10 _________________ 44.00 43.00 956.07 111.00 265.0 .23 July 20—21 _________________ 32.00 31.00 1,051.37 88.00 313.0 .22 Ang. 14—15 _________________ 12.00 11.00 1,212.12 36.00 116.0 .15 Sept. 10 _________________ 7.80 7.50 1,629.63 33.00 141.0 1.10 Get. 21 _________________ 7.20 6.70 939.74 17.00 61.0 .15 Nov. 3 _________________ 2.90 2.40 324.07 2.10 19.0 .29 Nov. 4— 5 _________________ 37.00 35.00 529.10 50.00 138.0 2.90 1971 Feb. 22—23 _________________ 39.00 34.00 1,209.15 111.00 105.0 1.50 July 29 _________________ 13.00 13.00 5,527.05 194.00 180.0 .12 Aug. 1— 2 _________________ 34.00 32.00 5,925.92 512.00 255.0 3.10 Aug. 3— 4 _________________ 91.00 87.00 2,303.10 541.00 1,030.0 15.00 Aug. 27—28 _________________ 63.00 60.00 2,641.97 428.00 262.0 .33 Sept. 11—12 _________________ 27.00 24.00 5,679.00 368.00 173.0 .33 NOV. 24—25 _________________ 68.00 63.00 1,475.60 251.00 198.0 .38 Nov. 29—30 _________________ 17.00 13.00 1,111.11 39.00 59.0 1.10 Dec. 7 _________________ 19.00 16.00 1,018.52 44.00 65.0 .86 1972 Mar. 16—17 _________________ 50.00 45.001 2,230.45 271.00 160.0 1.50 Apr. 16—17 _________________ 33.00 27.00 2,414.26 176.00 134.0 6.80 June 21—22 _________________ 381.00 368.00 3,713.76 3,690.00 1,930.0 1.80 July 2— 3 _________________ 56.00 49.00 2,494.33 330.00 264.0 2.60 Oct. 28 _________________ 20.00 19.00 6,354.77 326.00 106.0 .46 NOV 8 _________________ 22.00 21.00 1,181.66 67.00 90.0 .45 Nov 14 _________________ 36.00 35.00 761.90 72.00 142.0 .43 Dec 8— 9 _________________ 52.00 50.00 1,022.22 138.00 154.0 1.50 1973 Feb. 2 _________________ 34.00 29.00 1,583.65 124.00 90.0 1.00 Apr 1— 2 _________________ 50.00 40.00 2,527.77 273.00 198.0 3.40 Apr. 4 _________________ 26.00 23.00 1,433.17 89.00 150.0 2.10 Apr 25—26 _________________ 31.00 26.00 740.74 52.00 105.0 1.10 Am 27 _________________ 31.00 27.00 1,001.37 73.00 135.0 21.00 June 29 _________________ 8.20 7.20 2,006.17 39.00 107.0 .65 July 3— 4 _________________ 24.00 22.00 2,912.45 173.00 276.0 3.20 Sept. 14 _________________ 19.00 18.00 1,481.48 72.00 79.0 .29 Oct. 2— 3 _________________ 7.90 6.90 1,610.30 30.00 44.0 .31 Dec _________________ 5.90 4.70 1,260.83 16.00 43.0 .24 1971; Jan. 21 _________________ 19.00 16.00 2,500.00 108.00 131.0 1.00 Mar. 30—31 _________________ 71.00 63.00 1,064.08 181.00 244.0 1.30 May 12—13 _________________ 27.00 22.00 1,060.60 63.00 198.0 .63 June _________________ 9.70 7.70 577.20 12.00 50.0 .81 July 29—30 _________________ 6.30 5.00 2,444.44 33.00 120.0 .11 Aug. _________________ 11.00 9.60 4,089.50 106.00 109.0 1.00 Aug. 19—20 _________________ 13.00 11.00 2,424.24 72.00 128.0 .40 Lutes Run at 1963 June 2— 3 _________________ 9.10 8.10 9,785.07 214.00 50.0 0.30 Nov. 6— 7 _________________ 9.70 8.80 9,301.32 221.00 37.0 .40 1964 Jan. 9 _________________ 7.70 7.10 11,632.73 223.00 42.0 .10 Aug. 3 _________________ 3.40 2.80 10,846.54 82.00 48.0 .50 1965 Feb. 7— 8 _________________ 7.60 5.80 6,066.40 95.00 36.0 .80 Mar 26 _________________ 4.80 3.60 3,909.46 38.00 46.0 1.00 Oct 7— 8 _________________ 37.00 34.00 3,834.42 352.00 350.0 .70 1966' Feb. 13 _________________ 18.00 12.00 8,950.60 290.00 170.0 7.00 Nov 28 _________________ 3.40 3.20 1,851.85 16.00 23.0 .10 See fooflnotes at end of table. SEDIMENT DISCHARGE 39 in the Rock Creek and Amcostia River basins, 1962—74—C0ntinued c- Total pre- Maximum Maximum Maximum Maximum ngitilzl'ea Antecedent cipitation 15-minute 30-minute 1-.hour 3Thoux' Peqk Number of Time (Cp) days (Rt) ramfafl ramfall ramfall ramfall 1 ratlo peaks (Tm) (percent) (Ad) (inches) $1315) I(R:;o) €131») (Run) (Pr) (Np) (months) of: (m/hr) (m/hr) (m/hr) (In/hr) has"; area Layhill, Md.—-Continued 2 1.06 0.88 0.58 0.33 0.14 5.48 2 85 9.7 6 2.92 .52 .40 .36 .28 3.07 1 85 9.7 16 3.18 2.24 1.80 1.23 .52 6.16 2 88 9.7 9 2.46 3.08 2.34 1.39 .49 10.09 2 88 9.7 6 1.40 1.24 1.18 1.00 .45 10.53 1 89 9.7 17 1.48 3.08 2.40 1.35 .45 18.65 1 90 9.7 4 2 1.35 .72 .50 .39 .24 9.08 1 91 9.7 2 2 .53 .64 .38 .24 .12 7.80 1 92 9.7 0 1.94 .56 .54 .41 .34 3.86 1 92 9.7 8 1.65 .40 .34 .30 .18 3.04 2 95 8.9 3 2.21 2.16 1.74 1.48 .56 13.84 1 100 8.9 2 a 1.63 1.72 1.48 .81 .36 7.87 2 101 8.9 0 2 3.30 4.80 3.66 2.00 .77 11.67 1 101 8.9 7 3.37 1.24 1.00 .95 .56 4.36 1 101 8.9 11 3.20 1.72 1.20 .84 .47 7.19 5 102 8.9 16 2.50 .52 .46 .37 .34 3.14 1 104 8.9 3 2 .67 .40 .32 .24 .17 4.45 1 104 8.9 5 2 .97 .32 .24 .18 .12 4.01 1 105 8.9 1 1.55 .68 .42 .32 .21 3.52 1 108 13.7 2 1.03 .72 .50 .28 .16 4.71 3 109 13.7 6 2 8.23 2.40 2.30 1.85 1.48 5.24 2 111 13.7 1 1.72 2.60 2.54 1.57 .65 5.33 2 112 13.7 8 1.65 1.80 1.02 .60 .25 5.55 2 115 13.7 10 1.50 .64 .52 .45 .29 4.26 2 116 13.7 5 2.00 .60 .60 .45 .41 4.04 1 116 13.7 1 1.85 .56 .52 .41 .31 3.05 1 117 13.7 3 1.09 .24 .20 .18 .13 3.07 2 119 15.2 5 1.55 .36 .34 .32 .20 4.86 2 121 15.2 1 .95 .20 .16 .15 .13 6.43 1 121 15.0 14 2 1.94 .68 .56 .38 .22 4.00 2 121 15.2 0 2 1.37 .60 .50 .42 .26 4.22 1 121 15.2 3 .78 1.00 .84 .57 .20 14.77 1 123 15.2 3 2.03 3.24 2.38 1.23 .41 12.40 1 124 15.2 7 1.87 .72 .52 .39 .27 4.37 3 126 15.2 17 1.05 .36 .26 .24 .20 6.33 1 127 15.2 36 .92 .52 .38 .25 .14 9.10 2 129 13.1 8 .93 .80 .50 .47 .26 8.13 1 130 13.1 8 2.63 .64 .52 .45 .36 3.85 1 132 13.1 27 1.50 1.72 1.24 .78 .38 8.97 2 134 11.0 19 1.45 .44 .34 .28 .23 6.39 1 135 11.0 35 1.03 2.68 1.61 .87 .34 23.98 1 136 11.0 9 2.23 2.96 2.30 1.38 .59 11.25 2 137 11.0 9 1.50 3.28 2.84 1.76 .60 11.60 1 137 11.0 Lutes, Md. 2 ”2.93 0.44 0.36 0.33 0.60 6.14 4 5 20.3 3 2 2.80 .40 .36 .35 .72 4.16 2 10 14.3 1 2 1.16 .36 .30 .25 .48 5.90 1 12 14.3 20 2 1.95 .84 .52 .40 1.20 16.96 1 19 11.3 12 ’ 1.38 .24 .22 .20 .48 6.07 1 25 8.3 6 2 1.26 .40 .34 .30 .84 12.50 1 26 8.3 12 .45 1.44 1.30 1.02 2.52 10.27 3 33 6.7 1 2 1.71 1.48 .82 .42 3.60 13.58 3 37 13.3 24 1.17 .92 .76 .50 .96 7.16 1 46 10.3 4O EFFECTS OF URBANIZATION ON STREAMFLOW AND SEDIMENT TRANSPORT IN MARYLAND TABLE 5.—Hydrologic characteristics and related factors for storms Total Mean sedi- Suspended- Storm . Peak Antecedent .~ , . - d . . . ngfig. 1333,“ 13533 $233.32.. 555525-22 ("schjlge “"f3:§”e (ft3/s-d) (ffii/s-d) ("EKG/)1) (:33) (fa/s) (ftfl/s) Lutes Run at Lutes, 1967 Jan 27 _________________ 2.90 2.70 11,659.79 85.00 44.0 0.10 May 7 _________________ 9.10 8.30 2,989.73 67.00 43.0 1.30 June 22 _________________ 5.00 4.80 4,629.62 60.00 189.0 .10 July 20 _________________ 5.90 5.80 4,661.55 73.00 194.0 .10 July 29-30 _________________ 3.70 3.50 2,539.68 24.00 153.0 .10 Aug 3— 5 _________________ 13.00 13.00 2,849.00 100.00 221.0 .10 Aug 19 _________________ 2.70 2.60 4,843.30 34.00 88.0 .10 Oct 25—26 _________________ 3.80 3.60 3,292.18 32.00 101.0 .10 NOV 2 _________________ 3.00 2.90 753.51 5.90 26.0 .10 1968 Jan 14—15 _________________ 12.00 11.00 673.40 20.00 65.0 .20 Mar 17—18 _________________ 8.60 7.30 1,522.07 30.00 51.0 .50 May 27—29 _________________ 10.00 9.10 1,221.00 30.00 75.0 .20 June 19-20 _________________ 6.70 6.20 3,643.96 61.00 270.0 .20 July 2— 3 _________________ 8.90 8.20 2,393.85 53.00 218.0 .20 Sept. 2 _________________ .88 .84 2,425.04 5.50 48.0 .02 1970 Apr 2 _________________ 3.90 3.50 5,396.82 51.00 83.0 .20 1973 June 29 _________________ 4.10 3.90 2,943.97 31.00 137.0 .20 July 3 _________________ 4.50 4.10 3,703.70 41.00 170.0 .40 Sept 14 _________________ 5.20 5.10 275.96 3.80 43.0 .10 Oct. 2 _________________ 3.30 3.20 231.48 2.00 26.0 .10 1974 May 12 _________________ .64 .45 987.65 1.20 176.0 .20 June 16 _________________ 2.80 2.70 905.35 6.60 112.0 .10 June 26 _________________ .45 .30 197.53 .16 9.3 .20 July 5 _________________ .33 .22 538.72 .32 15.0 .10 July 29 _________________ 2.70 2.60 2,136.75 15.00 176.0 .10 Aug. 4 _________________ .26 .21 194.00 .11 12.0 .05 Aug. 9 _________________ 2.90 2.80 1,230.16 9.30 153.0 .10 Aug. 28 _________________ .31 .25 400.00 .27 12.0 .06 Aug. 30 _________________ 1.40 1.30 769.23 2.70 90.0 .06 Sept. 3 _________________ 3.30 3.20 578.70 5.00 95.0 .06 Sept. 6— 7 _________________ 7.30 7.10 156.49 3.00 44.0 .10 1 Maximum 5-minute rainfall for Lutes Run, maximum 3-hour rainfall for all other stations. 5’ From single rain gage in or near the basin. Each model for each station was analyzed, and the best models with one, two, three, and four inde- pendent variables were selected on the basis of the multiple correlation coefficient and the standard error of estimate. The regression equations sum- marized in table 7 were generally consistent with] the results of the correlation analyses. Storm runoff was the most significant parameter affecting sus- pended-sediment discharge. It explained 52 to 72 percent of the variation of sediment discharge for the five sites when it was the most significant inde- pendent variable. Peak discharge was the most sig- nificant independent variable at three sites: North Branch Rock Creek, Manor Run, and Nursery Run. It explained 40, 83, and 86 percent of the variation of suspended-sediment discharge, respectively. There was no apparent relation between the size of the basin or degree of urban development and the significance of peak discharge in the regression equations. The addition of a second independent variable to the regression model for each basin generally resulted in a marked improvement in the multiple correlation coefficient and standard error of estimate. This was particularly true in the basins where storm runoff was the most significant inde- pendent variable. The addition of the third and fourth independent variables improved the rela- tions to a lesser extent. The standard error tended to increase! with the addition of a fourth independ- ent variable at stations having a small number of storms available for analysis. The second and third most significant independ- ent variables in the regression models were general- ly related to storm intensity. Peak ratio, 15—min rainfall, or 30-min rainfall was selected at all sites except Nursery Run. Percentage of construction or SEDIMENT DISCHARGE 41 in the Rock Creek and Anaeosti’a River basins, 1962—74—Continued Construc- Total pre- Maximum Maximum Maximum Maximum . fion area Antecedent . 'tat' 15-minute {SO-minute l-hour 3_—hour Peak Number of Time (Cp) days CIDER, )ion rainfall rainfall rainfall rainfall 1 ratio peaks ( Tm ) (percent) (Ad) (inches) {1315) me») fRn) {R310 (Pr) (Np) (months) of (in/hr) (in/hr) (in/hr) (in/hr) basm area) Md.—Continued 17 0.74 1.24 0.68 0.39 3.60 16.26 1 48 10.3 9 2.06 .64 .54 .46 .96 5.02 2 52 9.3 5 1.39 3.32 2.22 1.23 4.20 39.35 1 53 9.3 4 1.42 2.84 2.10 1.22 3.12 33.43 1 54 9.3 7 .70 3.60 2.16 1.15 3.60 43.69 1 54 9.3 3 2.67 2.88 1.76 .91 4.20 16.99 3 55 9.3 13 2 1.29 2.52 1.30 .75 4.44 33.81 2 55 9.3 5 1.32 2.80 1.80 1.07 3.48 28.03 1 57 9.3 6 .97 .44 .38 .31 .48 8.93 2 58 9.3 14 1.45 .44 .38 .33 .44 5.89 2 60 9.3 3 1.84 .60 .50 .49 .84 6.92 2 62 4.3 2 3.13 .56 .48 .43 .72 8.22 1 64 4.3 1 .70 3.40 1.78 .89 4.44 43.52 1 65 4.3 3 1.98 2.88 1.98 1.20 3.96 26.56 3 66 4.3 11 2.56 1.68 1.10 .56 2.04 57.12 2 68 4.3 3 2 .65 1 24 .74 44 1.24 23.66 1 87 4.3 2 .98 1.96 1.32 .88 1.96 35.08 1 125 1.7 2 2 .78 3.24 2.38 1.23 3.96 41.37 1 126 1.7 6 1.85 .64 .46 .36 .84 8.41 4 128 1.7 1 1.10 .32 .30 .25 .36 8.09 4 129 .7 8 2 .17 1.20 .92 .61 1.20 390.67 1 136 .7 13 2 1.04 .56 .36 .25 .56 41.44 3 137 .7 2 2 .20 .28 .24 .18 .36 30.33 1 137 .7 6 2 .20 .44 .22 .13 .60 67.73 1 138 .7 23 .95 2.08 1.34 .74 3.00 67.65 1 138 .7 4 2 .17 .52 .26 .15 .84 56.90 1 139 .7 4 2 .68 1.76 1.26 .67 2.52 54.61 1 139 .7 1 2 .25 .44 .40 .20 1.08 47.76 1 139 .7 1 .53 1.48 .82 .47 2.40 69.18 1 139 .7 3 1.40 1.00 .62 .46 1.92 29.67 2 140 .7 1 1.50 .28 .26 .25 .36 6.18 1 140 .7 the chronology factor was also selected as the second or third independent variable for most stations. In fact, these were the only factors that showed any difference between the urban and rural streams. Each basin with active construction showed a signi- ficant relation between, suspended-sediment dis- charge and percentage of construction or the closely related chronology factor. These analyses indicate the complex relation be- tween sediment discharge and the factors affecting erosion and sediment transport. As in the study of large Atlantic coast streams (Guy, 1964), storm runoff, rainfall intensity, and storm peak ratio signif- icantly affect the sediment discharge. However, even though a relatively dense network of rain gages was used to determine rainfall amount and inten- sity, the relationships developed for these small basins are no better, and in some cases worse, than the relationships developed for the large basins. The lowest standard error of estimate for the equations in table 7 is 0.221 log units, or about 52 percent. Part of the error is attributable to the inaccuracies inherent in suspended-sediment sampling; however, much of the error is probably attributable to factors not analyzed by the regression model. Factors such as the location of sediment source areas and dif- fering soil and cover conditions, with respect to the areal distribution of rainfall, must be analyzed before individual storm sediment discharges can be defined by regression or other modeling techniques. EFFECTS OF LAND USE/LAND COVER AVERAGE ANNUAL SEDIMENT YIELD‘S Average annual suspended-sediment yields were computed for periods of equal land use/land cover in the study basins using flow duration and sedi- 42 EFFECTS OF URBANIZATION ON STREAMFLOW AND SEDIMENT TRANSPORT IN MARYLAND TABLE 6.——Simple correlation coefi‘icients Mean 8:53:33- sediment Storm Peak Anbecedent Antecedent discharge concen- runoff discharge discharge days (log Q.) tratwn (log Qr) (log Qp) (log 10 On) (Ad) (log C) Suspended-sediment Williamsburg Run ______________ 1.00 ___ 1 0.84 1 0.84 1 0.38 1 —0.37 Manor Run ____________________ 1.00 ___ 1 .50 1 .63 —.07 —.08 N. Br. Rock Creek ______________ 1.00 ___ 1 .84 1 .91 1.26 2—.26 N.W. Br. Anacostia River _______ 1.00 ___ 1 .74 1 .83 1 .29 ——.05 Nursery Run ___________________ 1.00 __- 1.81 1 .95 .20 —-.16 Batchellors Run ________________ 1.00 __- 1 .72 1 .84 ___ ___ Bel Pre Creek __________________ 1.00 ___ 1 .76 1 .84 1 .31 1—.26 Lutes Run _____________________ 1.00 ___ 1.85 1 57 ___ ___ Mean sediment Willlamsburg Run ______________ ___ 1.00 0.11 1 0.27 0.05 —0.12 Manor Run ____________________ ___ 1.00 ——-.21 .17 1 —.33 .10 N. Br. Rock Creek ______________ ___ 1.00 1 .32 1.53 ——-.01 —.13 N.W. Br. Anacostia River _______ ___ 1.00 —.04 2 .33 —.21 .23 Nursery Run ___________________ ___ 1.00 1 .44 1 .79 .10 —.15 Bel Pre Creek __________________ ___ 1.00 .12 1 .45 .08 —.03 Storm runoff, Williamsburg Run ______________ 1 0.84 0.11 1,00 1 0.89 1 0.44 1 —-0.39 Manor Run ____________________ 1 .50 ——.21 1.00 1 .70 1 .32 —.23 N. Br. Rock Creek ______________ 1 .84 1 .32 1.00 1 .92 1 .40 1 —.29 N.W. Br. Anacostia River _______ 1 .74 —.04 1.00 1 .79 1 .56 1 —.27 Nursery Run ___________________ 1.81 1 .44 1.00 1 .82 1 .26 —.11 Batchellors Run ________________ 1 .72 ___ 1.00 1 .80 ___ -__ Bel Pre Creek __________________ 1.76 .12 1.00 1 .80 1.38 1 —.35 Lutes Run _____________________ 1.85 ___ 1.00 1.56 ___ ___ Peak discharge. Williamsburg Run ______________ 10.84 0.27 1 0.89 1.00 1 0.37 1 —0.32 Manor Run ____________________ 1 .63 .17 1 .70 1.00 —.03 —.08 N. Br. Rock Creek ______________ 1 .91 1 .53 1 .92 1.00 1 .33 1 ——.29 N.W. Br. Anacostia River _______ 1.83 ”.33 1.79 1.00 1.35 1—.30 Nursery Run ___________________ 1 .95 1 .79 1 .82 1.00 .17 —.19 Batchellors Run ________________ 1 .84 _-_ 1.80 1.00 __- ___ Bel Pre Creek __________________ 1 .84 2 .45 1 .80 1.00 1.27 1 ——-.34 Lutes Run _____________________ 1 .57 ___ 1.56 1.00 ___ ___ Antecedent discharge. Williamsburg Run ______________ 1 0,38 0.05 1 0.44 1 0.37 1.00 1 —0.43 Manor Run ____________________ —.07 1 ———.33 1 .32 —.03 1.00 1 —.44 N. Br. Rock Creek ______________ 11 .26 ——.01 1 .40 1 .33 1.00 1—.29 N.W. Br. Anacostia River _______ 1.29 —.21 1 .56 1 .35 1.00 1—.40 Nursery Run ___________________ .20 .10 .26 .17 1.00 1 ——.30 Be] Pre Creek __________________ 1.31 .08 1 .38 1.27 1.00 2—.25 Antecedcnt Williamsburg Run ______________ 1 —0.37 —0.12 1 ——0.39 1 —0.32 1 —0.43 1.00 Manor Run ____________________ —.07 .10 —-.23 —.08 1 ——.44 1.00 N. Br. Rock Creek ______________ ’—.26 —.13 1—.29 1—-.29 1—.29 1.00 N.W. Br. Anacostia River _______ —.05 .23 2—.27 1—-.30 1 —.40 1.00 Nursery Run ___________________ —.16 —.15 —.11 —.19 1—.30 1.00 Bel Pre Creek __________________ 11—.26 —.03 1—.35 1 ——.34 ”—25 1.00 Total precepitation, Williamsburg Run ______________ 1 0.64 0.18 1 0.69 1 0.69 0.06 0.05 Manor Run ____________________ 1 .54 .04 1 .73 1 .63 ——.14 .09 N. Br. Rock Creek ______________ 1.56 2 .25 1 .63 1.58 1——.28 .03 N.W. Br. Anacostia River _______ 1 .59 1 .27 1 .54 1.54 —.04 .21 Nursery Run ___________________ 1 .67 1 .38 1 .79 1 .70 —.12 .12 Batchellors Run ________________ .29 ___ .29 .18 __- ___ Bel Pre Creek __________________ 1.51 .07 1.69 1.65 .02 —.20 Lutes Run _____________________ 1 .78 ___ 1.92 1’ .38 _-_ ___ Maximum 5-minute Lutes Run _____________________ 1 0.36 ___ 0.17 1 0.72 ___ ___ Maximum 15-minute Williamsburg Run ______________ 1 0.36 1 0.54 0.08 1 0.37 1 ——0.29 0.08 Manor Run ____________________ 1 .39 1 .48 —.06 1 .49 1 —.47 .20 See footnotes at end of table. SEDIMENT DISCHARGE of storm-related variables 12332321. 111.131.13.121: 11531531132 335.15.131.33 M11131? ”33.13131“ ”“3111" Chgngm £3353; 5:3}; tation rainfall rainfall rainfall rainfall rainfall peaks (7“; )1” tion (log P .) (log 10 Rt) (log 101205) (log 101315) (log 10 Ban) (log 10 Rm) (log 10 Ran) (Np) "1 (C?) I discharge, log Q. 0.64 ___ 1 0.36 1 0.41 1 0.45 1 0.55 0.13 0.12 1 0.27 0.08 1 .54 ___ 1.39 1 .39 1 .41 1 .43 .14 1 —.33 1.42 1 .25 1 .56 ___ 1.35 1.38 1 .38 1 .49 .16 .05 .14 .05 1.59 ___ .21 .22 1.28 1 .34 .01 .09 —-.05 .07 1.67 _-- 1.55 1.59 1.65 1.70 1.33 .16 .01 1.52 .29 ___ .14 .20 ___ ___ ___ —.02 ___ 1.49 1.51 ___ 1.41 1.43 1 .45 1 .49 1.26 1.26 1.51 —.02 1.78 1 0.36 1.33 1.39 ___ ___ ___ 1 —.80 1.70 1 —.47 concentration, log C 0.18 ___ 1 0.54 1 0.57 1 0.56 1 0.44 0.09 —O.20 1 0.23 1 0.36 .04 ___ 1.48 1.43 1 .37 .18 .05 1 —.59 1 .69 1 .48 1.25 ___ 1.57 1 .59 1.57 1.49 .08 —.14 .19 1.47 1.27 ___ 1 .53 1.52 1 .51 1.37 —.19 1 —.28 .12 1.56 1 .38 ___ 1 .71 1 .73 1.73 1.63 .10 .25 .10 1.78 .07 ___ 1 .55 1 .53 1.50 1.34 1 .25 1.31 1 .58 1.46 log Q: 1 0.69 ___ 0.08 0.13 0.19 1 0.39 0.10 1 0.29 0.18 —0.14 1 .73 __- ——.06 .01 .11 1 .40 .14 1.27 1—.28 1—.26 1 .63 ___ .03 .06 .08 1 .32 .18 .19 .04 1 -—.33 1.54 ___ —.18 —.16 —.09 .12 .15 1.36 —-.18 1 —.39 1 .79 ___ .17 .22 1 .33 1 .54 1.49 .01 ——.11 .04 .29 ___ —.24 —.18 ___ ___ __- .05 ___ .08 1.69 ___ .06 .11 .18 1 .39 .15 .08 .20 1 —.48 1.92 0.17 .18 .27 ___ ___ ___ 1 —.59 1 .51 1 —.64 10g Qp 1 0.69 ___ 1 0.37 1 0.42 1 0.46 1 0.59 0.08 1 0.35 0.10 1 0.31 1.63 ___ 1.49 1.55 1 .58 1 .62 1 .22 1.31 —-.18 1.51 1.58 ___ 1.23 1 .26 1 .29 1 .46 .12 1 .28 -—.02 .07 1 .54 ___ 1.32 1.35 1.40 1.47 —.01 1.33 —.12 .25 1 .70 ___ 1 .61 1 .66 1 .73 1 .78 1 .37 .08 —.07 1 .59 .18 -__ .24 ___ ___ ___ ___ ___ ___ 1.66 1.65 ___ 1.50 1 .56 1 .59 1 .67 1 .22 1 .27 1 .24 .14 1.38 1 0.72 1 78 1 81 ___ ___ ___ —.13 .05 .27 log 10 Q.. —0.06 -__ 1 —0.29 1 —0.26 1 —0.27 1 ~0.21 0.01 1 0.27 1 0.44 —0.16 —14 ___ 1——47 1—.44 1—43 1—.34 04 .17 1—27 1—.43 1—28 ___ 1—.44 1—42 1—43 1—.38 —01 1 23 1 25 1—.28 —- 04 ___ 1 —.29 1 — 28 1 —.27 —.19 .09 1 40 40 1 —.38 — 12 ___ -— 21 —’.21 — 22 —.17 —.08 1 30 19 —.12 02 ___ ——.13 —.10 —.11 —-.06 0 1 34 1 21 1 —.25 days, Aa 0.05 ___ 0.08 0.05 0.04 0.02 -—0.08 —0.19 —0.07 0.13 .09 __- .20 .21 1 .25 .19 —.17 —.14 .02 .18 .03 ___ .01 —.01 .02 .01 —.15 —.06 —.02 .06 .21 ___ .03 .05 .06 .06 .04 1 ——.41 .03 —.03 .12 ___ ——.06 —.08 —.06 —.11 .17 —.19 —.14 —.14 —.20 ___ 1—.21 1 —.21 1—.23 1 —.28 —.11 —.11 .05 .08 log 10 Rt 1.00 ___ 1 0.40 1 0.43 1 0.49 1 0.69 0.20 0.12 —0.12 0.08 1.00 ___ 1 .30 1 .37 1.47 1.70 1 .24 .10 —.05 —.02 1.00 ___ 1 .41 1 .43 1 .48 1 .65 1.25 -—.03 —-—.09 —.16 1.00 __- .20 .26 1 .34 1 .54 .10 .01 —.09 —.03 1.00 ___ 1.35 1 .40 1 .52 1 .74 1 .57 .06 —.06 .14 1.00 ___ .06 .06 ___ _-_ ___ —.04 ___ -——.06 1.00 ___ .28 1 .34 1.44 1 .65 1.21 —.10 ——.05 —.18 1.00 ___ .06 .15 ___ __- ___ ___ ___ —.71 rainfall, log 10 R05 ___ 1.00 ___ ___ ___ ___ ___ —0.09 0.05 __- rainfall, log 10 R15 1 0.40 _-_ 1.00 1 0.98 1 0.95 1 0.81 0.11 —0.11 ——_0.11 10.66 1 .30 -__ 1.00 1 .98 1 .93 1 .71 1 .27 —.16 1 .22 1.73 44 EFFECTS OF URBANIZATION 0N STREAMFLOW AND SEDIMENT TRANSPORT IN MARYLAND TABLE 6.—Simple correlation coefi‘icients Sgggggsg‘ sel‘i’lifnagnt Storm ' Peak Antecedent Anhecedent discharge (ancen- runofi dlscharge dlscharge days (log Q.) am“ (102 Or) (log Qp) (log 10 0..) (Ad) (108 C) Maximum 5-minute rainfall, N. Br. Rock Creek ______________ 1 0.35 1 0.57 0.03 1 0.23 1—0.44 0.01 N.W. Br. Anacostia River _______ .21 1.53 —.18 2.32 2—.29 .03 Nursery Run ___________________ 1 .55 1 .71 .17 1 .61 —.21 —.06 Batchellors Run ________________ .14 ___ —.24 .24 ___ ___ Bel Pre Creek __________________ 1 .41 1 .55 .06 1 .50 ———.13 2 —.21 Lutes Run _____________________ 2 .33 ___ .18 1 .78 ___ ___ Maximum 30-minute Williamsburg Run ______________ 10.41 1 0.57 0.13 10.42 2—0.26 0.05 Manor Run ____________________ 1 .39 1 .43 .01 1 .55 1 —.44 .21 N. Br. Rock Creek ______________ 1.38 1.59 .06 2 .26 1—.42 —.01 N.W. Br. Anacostia River _______ .22 1.52 —.16 1.35 2—.28 .05 Nursery Run ___________________ 1.59 1.73 .22 1.66 -—.21 —.08 Batchellors Run ________________ .20 ___ —.18 ___ ___ ___ Bel Pre Creek __________________ 1.43 1.53 .11 1.56 —.10 2—.21 Maximum l-hour Williamsburg Run ______________ 1 0.45 1 0.56 0.19 1 0.46 1 —-0‘.27 0.04 Manor Run ____________________ 1.41 1 .37 .11 1 .58 1 —.43 2 0.25 N. Br. Rock Creek ______________ 1 .38 1 .57 .08 1 .29 1 —.43 .02 N.W. Br. Anacostia River _______ 2 .28 1.51 —.09 .40 2—.27 .06 Nursery Run ___________________ 1.65 1.73 1‘ .33 1 .73 —.22 ——.06 Bel Pre Creek __________________ 1 .45 1.50 .18 1 .59 —.11 2—.23 Maximum 3-hour Williamsburg Run ______________ 1 0.55 1 0.44 1 0.39 1 0.59 2 —0.21 0.02 Manor Run ____________________ 1.43 .18 1 .40 1 .62 1 —.34 .19 N. Br. Rock Creek ______________ 1 .49 1.49 1.32 1 .46 1 —.38 .01 N.W. Br. Anacostia River _______ 1.34 1.37 .12 1 .47 —.19 .06 Nursery Run ___________________ 1 .70 1 .63 1.54 1 .78 —.17 —.11 Bel Pre Creek __________________ 1 .49 1 .34 1 .39 1 .67 —.06 1 —.28 Number of Williamsburg Run ______________ 0.13 0.09 0.10 0.08 0.01 —0.08 Manor Run ____________________ .14 .05 .14 .22 .04 —.17 N. Br. Rock Creek ______________ .16 .08 .18 .12 —.01 —.15 N.W. Br. Anacostia River _______ —.01 —.19 .15 —.01 .09 .04 Nursery Run ___________________ 2 .33 .10 1 .49 1 .37 —.08 .17 Bel Pre Creek __________________ 2 .26 2 .25 .15 2 .22 0 —.11 Chronology Williamsburg Run ______________ 0.12 ———0.20 1 0.29 1 0.35 1 0.27 —0.19 Manor Run ____________________ 1 —.33 1 —.59 2 .27 1 .31 .17 —.14 N. Br. Rock Creek ______________ .05 —.14 .19 1.28 2.23 —.06 N.W. Br. Anacostia River _______ .09 2—.28 1 .36 1 .33 1 .40 1——.41 Nursery Run ___________________ .16 .25 .01 .08 2 .30 —.19 Batchellors Run ________________ —.02 ___ .05 __- ___ _-_ Bel Pre Creek __________________ 2.26 1 .31 .08 1 .27 1 .34 ——.11 Lutes Run _____________________ 1 .80 ___ 1—.59 —.13 ___ ___ Percent Williamsburg Run ______________ 1 0.27 2 0.23 0.18 0.10 1 0.44 —0.07 Manor Run ____________________ 1 .42 1 .69 2 ——.28 —.18 2 -—.27 .02 N. Br. Rock Creek ______________ .14 .19 --.04 —.02 2 .25 —.02 N.W. Br. Anacostia River _______ —.05 .12 —-.18 —.12 .02 .03 Nursery Run ___________________ .01 .10 —.11 —.07 .19 —.14 Bel Pre Creek __________________ 1.51 1 .58 .20 2 .24 2 .21 .03 Lutes Run _____________________ 1 .72 ___ 1 51 .05 ___ ___ Peak ratio, Williamsburg Run ______________ 0.08 1 0.36 -—0.14 0.31 —0.16 0.13 Manor Run ____________________ 2 .25 1 .48 2 -—.26 1 .51 1 —.43 .18 N. Br. Rock Creek ______________ .05 1 .47 1 -—.33 .07 1 ——.28 .06 N.W. Br. Anacostia River _______ .07 .56 1—.39 .25 1 —.38 —-.03 Nursery Run ___________________ 1 .52 1 .78 .04 1 .59 —.12 ——.15 Batchellors Run ________________ 2 .49 ___ .08 1.66 ___ ___ Bel Pre Creek __________________ —.02 1 .46 1—.48 .14 2—.25 .08 Lutes Run _____________________ 1—.46 ___ — 64 27 --— -—— 1 Significant at 99 percent confidence level. ’Significant at 95 percent confidence level. of storm-related variables—Continued SEDIMENT DISCHARGE 222:- 22:22:: M2222 M2222 M22222 M22222 ”2‘2" Chggggggy 22:22:. 35;: tation rainfall rainfall rainfall rainfall rainfall DEaks (T ) tion (log P ‘) (log 10 R1) (log 10 Ros) (log 10 R15) (log 10 R30) (log 10 R11.) (log 10 Run) (Np) "1 (Cp) ' log 10 Rls—Continued 1 0.41 -__ 1.00 1 0.98 1 0.95 1 0.81 0.19 —0.16 —0.07 1 0.51 .20 ___ 1.00 1 .99 1 .96 1 .77 .05 —.23 .03 1 .77 2 .35 __- 1.00 1 .99 1 .95 1 .75 .20 —.02 .09 1 .84 .06 ___ 1.00 ___ ___ ___ ___ —.06 ___ 1.68 1 .28 ___ 1.00 1 .98 1 .94 1 .74 .14 .04 .07 1 .64 .06 ___ 1.00 ___ ___ ___ ___ -__ ___ 1.51 rainfall, log 10 R30 1 0.43 ___ 1 0.98 1.00 1 0.98 1 0.85 0.07 —0.11 —0.08 1 0.67 1 .37 ___ 1 .98 1.00 1 .97 1 .77 1 .27 —.13 .20 1 .72 1 .43 ___ 1 .98 1.00 1 .98 1 .84 .19 —.16 .02 1 .51 1 .26 ___ 1 .99 1.00 1 .98 1 .80 .04 —.20 .02 1 .78 1 .40 ___ 1 .99 1.00 1 .98 1 .80 .22 .03 .08 1 .85 .06 ___ ___ 1.00 ___ ___ ___ —.02 ___ 1.71 1 .34 ___ 1 .98 1.00 1 .98 1 .80 .18 .07 .06 1 .63 rainfall, log 10 Ru. 1 0.49 ___ 1 0.95 1 0.98 1.00 1 0.90 0.05 -—0.09 —0.08 1 0.64 1 .47 ___ 1 .93 1 .97 1.00 1 .86 .22 —-.09 .15 1 .65 1 .48 ___ 1 .95 1 .98 1.00 1 .89 .15 —.16 0 1 .51 1 .34 ___ 1 .96 1 .98 1.00 1 .87 0 —.17 .03 1 .74 1 .52 ___ 1 .95 1 .98 1.00 1 .89 1 .28 .04 .08 1 .82 1 .44 ___ 1 .94 1 .98 1.00 1 .89 .16 .08 .04 1 .58 rainfall, log 10 Km 1 0.69 ___ 1 0.81 1 0.85 1 0.90 1.00 0.02 —0.02 —0.05 2 0.49 1 .70 ___ 1 .71 1 .77 1.86 1.00 .17 .02 .02 1 .36 1 .65 0. 1 .81 1 .84 1 .89 1.00 .08 —.10 —.03 1 .34 1 .54 ___ 1 .77 1 .80 1 .87 1.00 .01 —.05 .07 1 .51 1 .74 ___ 1 .75 1 .80 1 .89 1.00 1 .33 .12 .10 1 .62 1 .65 ___ 1 .74 1 .80 1 .89 1.00 .13 .11 .01 1 .34 peaks, Np 0.20 -__ 0.11 0.07 0.05 0.02 1.00 0.01 —0.03 —0.06 2 .24 ___ 1 .27 11 .27 .22 .17 1.00 .04 0 .13 2 .25 ___ .19 .19 .15 .08 1.00 —.06 .02 —-.16 .10 ___ .05 .04 0 —.01 1.00 0 —.12 —.25 1 .57 ___ .20 .22 1 .28 1’ .33 1.00 .03 .03 —.06 1 .21 ___ .14 .18 .16 .13 1.00 .13 .08 .08 factor, Tm 0.12 ___ -—-0.11 —0.11 —0.09 —0.02 0.01 1.00 —0.15 0.16 .10 ___ ——.16 —.13 —.09 .02 .04 1.00 1 —.85 .09 —.03 ___ ——.16 —.16 —.16 —.10 —.06 1.00 1 —.54 .19 .01 ___ —.23 —-—.20 —.17 —.05 0 1.00 1 .31 —.07 .06 ___ —.02 .03 .04 .12 .03 1.00 1 .49 .13 -——.04 _-_ —.06 —.02 ___ ___ ___ 1.00 ___ .01 —.10 ___ .04 .07 .08 .11 .13 1.00 .37 1 .25 1 —.61 —0.09 -__ —.08 ___ ___ ___ 1.00 1 —.91 1 .57 construction, Cp —0.12 _-_ —0.11 —0.08 —0.08 —0.05 —0.03 —-0.15 1.00 —0.19 —.05 ___ .22 .19 .15 .02 0 1 —.85 1.00 .10 —.09 ___ ——-.01 .02 0 —.03 .02 1 —.54 1.00 ——.17 —.09 ___ .03 .02 .03 .07 —.12 1 .31 1.00 .09 —.06 ___ .09 .08 .08 .10 .03 1 .49 1.00 .02 —.05 ___ .07 .06 .04 —.01 .08 1 .37 1.00 .02 ___ 0.05 ___ ___ ___ ___ ___ 1—.91 1.00 __- log Pr 0.08 ___ 1 0.66 1 0.67 1 0.64 1 0.49 —0.06 0.16 —0.19 1.00 -—-—.12 _-_ 1 .73 1 .72 1 .65 1 .36 .13 .09 .10 1.00 —.16 ___ 1 .51 1 .51 1 .51 1 .34 —.16 .19 —.17 1.00 ——.03 ___. 1 .77 1 .78 1 .74 1 .51 —.25 —.07 .09 1.00 .14 ___ 1 .84 1 .85 1 .82 1 .62 —.06 .13 .02 1.00 —.06 ___ 1 .68 1 .71 ___ ___ ___ .01 ___ 1.00 —.18 ___ 1 .64 1 .63 1 .58 1 .34 .08 1 .25 .02 1.00 1—.71 ___ 1.51 ___ -__ ___ ___ __- ___ 1.00 46 EFFECTS OF URBANIZATION ON STREAMFLOW AND SEDIMENT TRANSPORT IN MARYLAND TABLE 7.——Regression coeflicients, multiple correlation coefficient, and standard error of best equations explaining the varia- tion of storm suspended-sediment discharge Maxi- Maxi— Stand- Number Sto m Peak Total Percent mum mum Chro- 5:11:51} Iv{Jiltile e131; Regres- 3:113:11: “1“ng “3::- ptgeéipi- constrnc- miigube milifllte nology 5:3: dis- correla- of_ Siam variables '(log Q") (10g 5:) (log???) (13):) may? “(iflfgan 31172:); (log P") c7115? 051311111- 16.15:; 5:11:13 10 Rao) 10 R15) 10 Q“) 9°” “5’3 units) Williamsburg Run near Olney 1 ____________ 11.08 ___ ___ -_ _-_ ___ ___ ___- ___ 0.84 0.304 0.387 2 ____________ 11.03 ___ ___ ___ 10.45 ___ ___ ___ ___ .90 .253 .051 3 ____________ 1 .99 ___ ___ 1 0.05 1 .48 ___ ___ ___ ___ .91 .240 ——.166 4 ____________ 1 .94 -__ ___ 2 .04 1 .52 ___ ___ __- 1 0.13 .91 .238 —.241 North Branch Rock Creek near Norbeck 1 ____________ ___ 11.49 ___ ___ -_- ___ ___ ___ -__ 0.91 0.255 —1.22 2 ____________ 11.50 ___ ___ 10.15 ___ ___ ___ -__ ___ .92 .239 —1.53 3 ____________ ___ 11.44 ___ 1.15 ___ 10.23 ___ __- ___ .93 .223 —1.63 4 ____________ 0.30 1 1.13 -__ 1 .14 ___ 1 .29 ___ ___ ___ .93 .221 —-1.46 Manor Run near Norbeck 1 ____________ ___ 10.83 ___ ___ ___ ___ ___ ___ ___ 0.63 0.410 0.334 2 ____________ ___ 11.07 ___ ___ _-- -__ 1—0.01 __- ___ .84 .288 .371 3 ____________ ___ 11.24 ___ ___ ”—0.32 ___ 1—.01 ___ ___ .85 .278 .385 4 ____________ 0.09 1 1.16 ___ ___ —.01 ___ 1 .03 ___ ___ .86 .280 .411 Northwest Branch Anacostia at Norwood 1 ____________ 10.97 ___ -_- ___ ___ _-_ __ ___ ___ 0.74 0.336 0.530 2 ____________ 11.19 ___ ___ ___ ___ ___ -__ 10.88 ___ .84 .274 —.522 3 ____________ 1 1.31 __- __- ___ ___ ___ 1 —0.004 1 .92 ___ .87 .252 -—-.470 4 ____________ 1 1.20 ___ 0.35 -__ ___ _ 2 —.004 1 .86 ___ .87 .250 -—-.743 Batchellors Run at Oakdale 1 ____________ 11.18 ___ _.._ __- ___ ___ __ ___. ___ 0.72 0.372 0.049 2 ____________ 11.13 ___ ___ ___ ___ ___ __ 10.88 ___ .84 .297 —.757 3 ____________ 1 1.07 ___ 0.44 _-_ ___. ___ ___ 1 .90 __- 85 .295 —1.283 4 ____________ 1 1.07 ___ .43 -__ ___ _ —0.001 1 .90 ___ 85 .300 ——1.211 Nursery Run at Cleverly 1 ____________ ___ 1 1.49 -__ ___ ___. ___ ___ ___ 0.95 0.253 —1.33 2 ____________ ___ 1 1.47 ___ ___ __- 2 0.003 -_- ___ .95 .245 —1.45 3 ____________ 0.21 1 1.33 ___ ___ ___ ___ ”.003 _-_ ___ .96 .242 —1.34 4 ____________ .29 1 1.34 —0.22 ___ -_- _-_ 2 .003 ___ -__ .96 .243 ——1.10 Bel Pre Creek at Layhill 1 ____________ 11.12 ___ ___ __- ___ ___ __ ___ __- 0.76 0.402 0.458 2 ____________ 1 1.44 ___ ___ ___ __- ___ __ 1 1.09 ___ .85 .324 ——.791 3 ____________ 1 1.31 ___ ___ 1 0.051 ___ ___ 1 .97 ___ 91 .256 —1.088 4 ____________ 1 1.18 ___ ___ 1 .052 2 0.28 ___ 1 .60 ___ 92 .248 ——.934 Lutes Run at Lutes 1 ____________ 1 1.52 ___ ___ ___ ___ ___ ___ -__ ___ 0.85 0.473 0.438 2 ____________ 11.05 ___ ___ ___ ___ ___1—0.008 ___ ___ .93 .345 1.37 3 ____________ 11.32 ___ ___ ___ -_- ___ 1—.010 1 0.66 -__ .96 .274 .520 4 ____________ 1 1.24 ___ ___ ___ ___ 0.14 1 —.010 .52 ___ .96 .277 .582 1 Signicant at 99 percent confidence level. 3 Significant at 95 percent confidence level. ment—transport curves as described by Miller (1951). The number of yields estimated for each basin varied depending on land-use/land-cover changes in the basins. For example, six average annual yields were estimated for the Williamsburg Run basin because of changes in the amount and location of construction within the basin. In the rural basins, only one average annual yield was estimated since the land use/land cover was essen- tially constant during the study period. A total of 27 yields was determined for the eight basins. The initial selection of periods for which average annual sediment yields were to be estimated was based on the variation of land use/land cover be— tween 1967 and 1974. Periods of equal land use/ land cover were determined by examining aerial photographs, field notes, and grading and paving permits furnished by the Montgomery County De- SEDIMENT DISCHARGE 47 partment of Public Works. Because of constantly changing conditions on construction sites, it was not possible to define exact periods of equal land use/ land cover, but the beginning and ending of each period were selected so that the land use/land cover on aerial photographs would be representative of average conditions during the period. An effort was made to delineate periods by the amount of construc- tion and the location of construction within the basins so that a range of physical conditions on conn struction sites (slope, proximity to stream, and so forth) could be evaluated. Once the periods were selected, sediment-trans- port curves were prepared for the periods having adequate storm data. The curves were developed from logarithmic plots of daily suspended-sediment discharges and the corresponding daily water dis- charges. Particular attention was given to the me- dium and high ranges of water and sediment dis- charge since most of the annual sediment load was transported during large storms. Curves for several periods were not used because the high ends of the curves were poorly defined or because a dispropor- tionate number of storms were sampled during either the growing or dormant season. The sediment-transport curves were used in con- junction with duration curves of water discharge for the corresponding streams to compute the aver- age annual sediment yields. Flow-duration curves for the period 1967—74 were available for all sta- tions except Lutes Run and Batchellors Run. Since only data for days with storm runoff and selected base-flow days were available at these two stations, curves were estimated using graphical-regression methods and the duration curves of the other sites. An example of the computation of the average annual sediment yield is illustrated by figure 14 and table 8. Average water dis-charges for selected time intervals were determined from the flow-duration curves (fig. 14A). Suspended-sediment discharges- corresponding to the water discharges were deter- mined from sediment-transport curves and multi- plied by the duration intervals of water discharge to calculate the average annual sediment discharge. In the example, the average water discharge for 2.5 to 4.5 percent of the time is 14 ft3/s or 0.40 ma/s for Williamsburg Run (fig. 14A). The corres- ponding suspended-sediment discharge is 29 tons or 26 t (fig. 143). Multiplying the suspended-sedi- ment discharge by the time interval for each inter- val in table 8 and dividing the sum of column 5 (table 8) by 100 yields the mean daily sediment dis- charge. The mean daily discharge times 365 days TABLE 8.—Example of computation of average annual sus- pended—sediment discharge [Data from 1967—74 duration curve and January 1972 to December 1973 sediment-transport curve of Williamsburg Run near Olney] Sus~ T‘ “get“ 5115- pegged- 1me . is- s 1- 22.22 2222: ...... 2:22.:- m... time mter- nate (cubic ment dis- (per val (pen- feet dis- charge cent) (5:55) cent) 32:: charge i 13);- ond) (tans) ( val)1 tons (1) (7A) (3) (4) (5) (6') 0.25 0.25 0.125 68 260 65 .75 .50 .50 42 163 81 1.5 .75 1.125 32 119 89 2.5 1.0 2.0 20 61 61 4.5 2 3.5 14 29 58 8.5 4 6.5 6.5 5.2 21 15 6.5 11.75 3.5 .9 5.8 25 10 20 2.5 .3 3.0 35 10 30 1.9 .1 1.0 45 10 40 1.6 0 0 65 10 50 1.4 .._-_ _-_- 75 20 65 1.0 ____ ____ 95 20 85 .65 -___ -_-_ 100 5 97.5 32 ____ ___- Total -_ --_ _--_ ____ __-_ 384.8 Mean daily suspended-sediment discharge: 384.8 + 100 = 3.85 tons. Average annual suspended-sediment discharge: 3.85 X 365 = 1,400 tons. 1 Column 6:column 2 X column 5. equals the average annual suspended-sediment dis- charge. The average annual suspended-sediment yields calculated by this method represent the average annual yields that would have been expected be- tween 1967 and 1974 if land-use/land-cover condi- tions at each site had remained the same as the average during the selected periods. The use of the 1967—74 flow-duration curves assures that the water discharge component of the computations was equal for each period at each site. The one questionable component of computations is whether the sediment data used for developing the individual sediment- transport curves were significantly affected by dif- ferent climatic conditions during the selected periods. A change in the statistical distribution of total precipitation for the sample storms. or a change in the intensity or time of concentration of storm runoff could result in a substantial bias in the com- puted sediment yields. However, the regression models discussed previously indicate no substantial change in storm magnitude or intensity during the period of record. Another factor that could affect the computed yields is the total precipitation during the selected periods. Miller (1951) found that the flow-duration sediment-transport curve method generally over- estimated the long-term load when sediment data were collected during dry years and underestimated it when data were collected during wet years. The estimate may be incorrect because the greater num- ber of storms during wet years exhausts the supply of sediment readily available for transport. Al— 48 EFFECTS OF URBANIZATION ON STREAMFLOW AND SEDIMENT TRANSPORT IN MARYLAND though the total load for the year is higher, the average load per storm is less than during dry years. This explanation may not be valid for urban construction sites where the land surface is re— worked frequently for right-of-way grading, in- stallation of utilities, lot grading, and final grading prior to stabilization. The calculated suspended—sediment yields and land-use/land-cover conditions corresponding to the sample periods are listed in table 9. The basin sedi- ment yields range from 0.27 to 11.3 tons/acre (0.61 to 25.3 t/hmz). Land under construction ranges from 0 to 13.9 percent of drainage area. Cropland ranges from 0 to 43.8 percent, and urban land ranges from 0 to 54.2 percent of drainage area. Regression equations were computed to deter- mine the effect of the land—use factors on suspended- sediment yield. Equation 1, which is the simple TABLE 9.-—Average annual suspended-sediment yields and selected land-use/land-cover factors 1:23}: Percentage of basin area . ment Period yield Crop- U ba (Eon- (tons/ land 1' n StFW acre) 1°“ Williamsburg Run near Olney, Md. Oct. 1966—Mar. 1968 _______ 1.98 43.8 0.1 7.6 Apr. 1968*Ma1'. 1969 -. 1.61 43.3 5.3 4.7 Apr. 1969-Mar. 1971 __ 1.18 42.1 8.1 3.1 Apr. 1971-Dec. 1971 _. ___ 1.16 29.9 12.0 5.1 Jan. 1972—Sept. 1973 _______ .98 29.8 16.0 6.2 Jan. 1974—Sept. 1974 _______ .57 29.0 24.7 1.0 North Branch Rock Creek near Norbeck, Md. 1 Oct. 1966—Mar. 2.07 26.4 0.4 1.8 Apr. 1969—Mar. 2.01 25.9 .8 1.3 Apr. 1971—Dec. _ 1.39 25.3 1.8 1.3 Jan. 1972-Dec. 1.43 25.3 2.7 1.1 Jan. 1974—Sept. 1974 _______ 1.33 25.3 3.7 .3 Manor Run near Norbeck, Md. Mar. 1967—Sept. 1968 _______ 11.26 3.9 2.5 11.3 Jun. l969—Dec. -_ 10.34 4.8 9.4 9.6 Apr. 1970—Dec. 4.92 4.8 14.9 7.4 Feb. 1971—Dec. 2.05 4.8 19.8 4.6 Apr. 1973—Aug. 1.32 4.8 23.8 .9 Northwest Branch Anacostia River at Norwood, Md. Jan. 1968—Mal‘. 1974 _______ 0.96 24.9 0 0.3 Nursery Run at Cloverly, Md. May 1967-Aug. 1974 _______ 0.62 10.3 0 0 Batchellors Run at Oakdale, Md. Aug. 1967—Sept. 1973 _______ 0.32 20.3 0 0.7 Bel Pre Creek at Layhill, Md. Mar. 1963—Dec. 1964 _______ 0.27 6.5 o 0 Jan. 1966-Dec. 3.31 2.2 4.6 13.9 Jan. 1968-Dec. 3.27 2.2 8.4 104 Jan. 1970—Dec. 1971 _______ 2.13 1.8 11.0 9.2 Jan. 1973—Sept. 1974 _______ 1.90 1.7 20.8 13.1 Lutes Run at Lutes, Md. Oct. 1966-Jan. 1968 _______ 5.08 0 51.2 9.8 Mar. 1968—Apr. 1970 _______ 2.79 0 54.2 4.3 June 1973-Sept. 1974 _______ 2.21 0 54.2 1.2 , Sediment yield and land-use/land-cover percentages equal total for North Branch Rock Creek near Norbeck minus the contribution of the Williamsburg Run tributary. linear relation between the logarithm of the sedi- ment yield and percentage of the basin under con- struction, was the best equation for explaining the variability of sediment yield. Log Y,.=0.059 C—0.047 (1) where Y.=suspended-sediment yield, in tons per acre; C=percentage of land under construction. Forty-seven percent of the variation of the sediment yield is explained by this equation. The standard error of estimate is 0.28 log units, or about 62 per- cent. The use of percentage of urban and percent-, age of cropland as additional variables did not signif- icantly improve the explained variation of sediment yield. The addition of the percentage of urban land increased the explained variation to 50 percent, and the addition of the percentage of cropland as a third independent variable increased the explained varia- tion of sediment yield to 51 percent. The standard errors of estimate for both equations are 0.28 log units, the same as the simple relation between sedi- ment yield and percentage of construction. The relatively low percentage of variation ex- plained by the regression equations is illustrated by the scatter of data points in figure 15. This large scatter indicates that factors other than the vari- FI I I I I *I' 5 f I I q _ :‘ l0 1: D _" Z O i f c: E — a W (D a 100 3 3— E D: ”D _ o. E : g g E i I; I” E c. : 2 e 1 “: 2 D ________________ m U 10 _ :_ 8 Z E I? g g : : uJ E — ’: U I g I — —— < L) '— I (L3 _ _: a D 7 a Z ! E ID : E ”35 2 E :7 E 3 fl : a g : _g 0.01 g 01 I l I 5 I I I I I I I 0.01 01 0.5 I 5 20 50 80 95 99 99.8 99.99 PERCENTAGE OF DAYS INDICATED DISCHARGE WAS EXCEEDED FIGURE 14.—Example of flow-duration (A) and suspended- sediment transport (B) curves used for computing aver- age annual sediment loads. A, Williamsburg Run flow- duration curve, 1967—74. B, Williamsburg Run sediment- transport curve, January 1972 to December 1973. DAILY SUSPENDED-SEDIMENT DISCHARGE, IN TONS 100 .4 o _| O 0.1 SEDIMENT DISCHARGE DAILY WATER DISCHARGE, IN CUBIC METERS PER SECOND 0.1 1.0 I I I I I I I I I I I I I I I I I I I I I I I I I | | I I I I | I I I I I I _ L _ .— . —— EXPLANATION o _ — 0 Daily value ° — CI Average of discharge ciass 0 _ _: 100 Z ' :— : o :T _ _— 1o — 1.0 _ —I_ _ _: 0.1 : . i: _ I I I I | I I I II I I I I I I f— 1 10 100 DAILY WATER DISCHARGE, IN CUBIC FEET PER SECOND FIGURE 14.—Continued. DAILY SUSPENDED-SEDIMENT DISCHARGE, IN TONNES 49 50 AVERAGE ANNUAL SUSPENDED-SEDIMENT YIELD, IN TONS PER ACRE 50 EFFECTS OF URBANIZATION 0N STREAMFLOW AND SEDIMENT TRANSPORT IN MARYLAND I DATA TREND POINT LINE I {OVIQODEI I I EXPLANATION Williamsburg Run N. Branch Rock Creek Manor Run Bel Pre Creek Lutes Run NW. Branch Anacostia River Nursery Run Batchellors Run I L~100 _r10 0 5 PERCENTAGE OF BASIN AREA UNDER CONSTRUCTION FIGURE 15.—Rela’cion of suspended-sediment yieldto percentage of drainage basin under construction. 10 15 AVERAGE ANNUAL SUSPENDED-SEDIMENT YIELD, IN TONNES PER SQUARE HECTOMETER SEDIMENT DISCHARGE 51 ables used in the regression equations affect the suspended-sediment yields of the basins. The lines on figure 15, which represent the approximate rela- tion between sediment yield and percentage of con- struction for different periods on individual basins, further illustrate the variation of sediment yields. Apparently, other basin factors such as slope or soil conditions caused substantially greater sediment yields in the Manor Run, North Branch Rock Creek, and Lutes Run basins than in the Bel Pre Creek and Williamsburg Run basins. CONSTRUCTION-SITE SEDIMENT YIELDS To evaluate the variation of suspended-sediment yields from construction sites, a simple arithmetic procedure was used to- successively eliminate the: part of each basin yield attributable to other land use/land cover, so that the remainder repre- sented only the yield from construction sites. Land use/land cover in each period for each basin was first summarized in the following categories: Forest land, grass, crops, rural residential, urban residen- tial, public and commercial, and construction. These categories represent the same general categories of table 2 except that the impervious category of farm- land and parks was placed in the cropland category because most of this area consisted of dirt roads with unimproved roadside ditches. The next step was to estimate suspended-sediment yields. for the different land-use/land-cover cate- gories with relatively low erosion potential. Wil- liams and George (1968) determined that the sedi- ment yield from basins in Pennsylvania with 100 percent forest cover averaged 0.03 (ton/acre)/yr [(0.07 t/hm2)/yr]. Sediment yields from forested watersheds in Mississippi were reported as 0.02 (ton/acre)/yr [(0.04 t/hm2)/yr] by Ursic and Dendy (1965). Wark and Keller (1963) found that forested areas in the Potomac River basin averaged about 0.03 (ton/acre) /yr [(0.07 t/hm2)/yr]. On the basis of these previous investigations, the annual sediment yield from forested areas in the study basins was assumed to be 0.03 ton/acre (0.07 t/hmz). Ursic and Dendy (1965) found the annual sedi- ment yield of grazed pastures and abandoned fields in Mississippi averaged 1.6 and 0.12 tons/acre (3.6 and 0.27 t/hmz), respectively. The grass category in this study is generally well-managed turf on golf courses, abandoned fields, or lightly grazed pastures. These conditions are probably comparable to the abandoned fields in Mississippi; therefore, an annual sediment yield of 0.2 ton/acre (0.45 t/hmz) was as- sumed for the grass category. The yields for rural residential and commercial- public areas within the basins were assumed to be 0.5 and 0.3 (ton/acre)/yr [(1.1 and 0.7 t/hm2)/yr], respectively. The rural residential yield was as- sumed to be considerably higher than the yield for grass because a fairly large part of the rural resi- dential area was used for large gardens and for grazing horses. Also, some of the roads and road- side ditches in these areas were unimproved. The yield for commercial-public land was assumed to be somewhat lower than rural residential because most of this land was in either well-managed turf or im- pervious surfaces. It was assumed to be greater than grass, however, because of accelerated erosion in the stream channels immediately downstream caused by increased runoff from the impervious sur- faces. This subject will be discussed in more detail in the section on stream-channel erosion. Suspended-sediment yields for urban residential, cropland, and construction areas were determined by subtracting the assumed yields for forest land, pas- ture, rural residential, and public-commercial from the total sediment yields of each basin. For example, the sediment yield for cropland was determined for four basins during periods when there were no urban residential or construction areas. Example 1 for Be] Pre Creek during the March 1963 to De- cember 1964 sample period illustrates this method. EXAMPLE 1.—Bel Pre Creek, March 1963 to December 1961, Average annual sediment yield:295 tons; drainage area:l,082 acres. (1) (2) (3) Land use/land cover Sediment Columns Acres yield (2) X (3) Forest _______________ 619 0.03 18.6 Pasture _____________ 375 .2 75 Rural residential _____ 18 .5 9 Crops _______________ 70 __ __ Sum ___________ 1,082 __ 102.6 Annual sediment yield from cropland=(295—102.6) tons+ 70 acres:2.7 tons/acre. The computed annual yields for cropland in the four basins were: Ttms/Ac're Batchellors Run __________________ 0.65 Bel Pre Creek ____________________ 2.7 Nursery Run _____________________ 4.3 Northwest Branch Anacostia River _________________ 3.3 Similarly, the annual yield for urban residential was determined by subtracting other known or assumed yields from the total yields of the basins during periods with no construction or only minimal con- struction activities (Example 2). 52 EFFECTS OF URBANIZATION ON STREAMFLOW AND SEDIMENT TRANSPORT IN MARYLAND EXAMPLE 2.—Lutes Run, June 1973 to September 1971, Average annual sediment yield—2664 tons, drainage area:301 acres. (1) (2) (3) Land use/land cover Acres Sefiggnt (gallgfigj Pasture _________ 17 0.2 3.4 Forest land ______ 54 .03 1.6 Rural residential _ 48 .5 24 Urban residential _ 163 __ __ Commercial-public- 15.5 .3 4.7 Construction _____ 3.5 Assumed 10 35 Sum _______ 301 __ 68.7 Annual sediment yield from urban land: (664—687) tons + 163 acres: 3.65 tons/acre. The annual suspended-sediment yields computed for Lutes Run and Manor Run were 3.6 and 3.9 tons/ acre (8.1 and 8.7 t/hmz), respectively. The computed values of cropland and urban resi- dential were then used with the other assumed values to compute the construction-site sediment yield for basins with active construction. If cropland or urban yields were computed for individual basins, these values were used; otherwise, the average of the yields computed for all basins were used. Construction-site sediment yields computed for the different time periods in the five basins with active construction ranged from 7.2 to 100.8 (tons/ acre)/yr [(16.1 to 226 t/hm2)/yr] and averaged 32.7 (tons/acre)/yr [(73.3 t/hm2)/yr]. The yields are comparable to those found by previous investi- gators. Vice, Guy, and Ferguson (1969) reported that the yield from highway construction in north- ern Virginia was about 100 tons/acre (220 t/hmz). Guy and Ferguson (1962) estimated the yield from residential construction sites near Lake Barcroft, Va., to be 39 tons/acre (87 t/hmz). Several factors could affect the variation of con- struction-site yields. Slope conditions on the sites, proximity to stream channels, buffer zones of nat- ural vegetation, ratio of the largest construction site to total construction area in the basin, and the percentage of construction area with sediment con- trols were factors considered for analysis. Other factors are probably significant, but could not be evaluated properly. Soils, particularly their relative erodibility, could result in substantial variation of sediment yields; however, the soil conditions were generally uniform except for one or two sites. The density of housing could affect the construction methods, but, again, the conditions in the basins were too uniform to detect significant differences. Another important factor that could not be evalu- ated for use in the analysis was the length of time that sites were not stabilized, as the period of non- stabilized construction activity varied among sites within an individual basin. Average slope on the construction sites was deter- mined by using the average slope map shown in plate 1 as an overlay to the aerial photographs of the study basin. The construction area in each slope class was determined with a dot-grid, and a weighted average was calculated. This method aver- aged the slope conditions of different sites within the basins, which could obscure the influence of ex- treme conditions. However, it did provide a relative index of slope for evaluating the variation of yields among the basins. Average slopes ranged from 4.30 to 7.34 percent. The amount of construction within 200 ft (61 m) of stream channels was selected as an arbitrary index of the proximity of construction to the drain- age net. An overlay of streams shown on the US. Geological Survey quadrangles and US. Soil Con- servation Service soil maps of the study basins was used with aerial photographs to measure the per- centage of construction area within this category. Construction within 200 ft (61 m) of stream channels (hereafter referred to as the proximity factor) ranged from 17.4 to 64.8 percent of the total construction area in the basins. The existence of buffer zones of natural vegeta- tion between the construction sites and the stream channels was also evaluated. The percentage of con- struction area without a 200 ft (61 m) strip of natural vegetation between the site and the natural drainage to the stream channels was determined from aerial photographs. This parameter ranged from 52 to 100 percent. The size of the construction sites was also con- sidered as a possible source of variation of site yields. The area of the largest construction site as a percentage of the total construction area within each basin was used as an index of this parameter. The last parameter considered for analysis was the percentage of the construction sites with ade- quate sediment-control measures. This parameter was based on county-wide field surveys, which will be discussed in more detail in the section on sedi— ment control. Some adjustments to the average values determined by the field surveys were made in cases Where observations by project personnel were available. These adjustments to the average values were particularly necessary during the initial periods of development when many sites did not have controls. Sites with control measures con- stituted 0 to 68 percent of the total construction area in the basins. SEDIMENT Construction-site suspended-sediment yields and, the corresponding site factors are summarized in table 10. These data were analyzed graphically and with correlation and multiple regression techniques to determine the relative effect of each site factor on construction-site yields. Graphical and correla- tion analyses indicate that there is a poor relation- ship between site yields and the individual site fac- tors. Correlation coeflicients between the logarithm of site yield and the site factors are: 0.36 for aver- age slope, 0.36 for the proximity factor, 0.28 for the percentage of construction area represented by one large site, 0.18 for construction sites without buffer zones, and —0.67 for the percentage of construction sites with sediment controls. The correlation coeffi- cient for site yield and percentage of controls is the only one significant at the 95 percent confidence level. The multiple effect of the site factors was evalu- ated with a multiple-regression model. The loga- rithm of construction-site sediment yield was the dependent variable and the site factors were the independent variables. Equations 2, 3, and 4 illus- trate the results of the regression analysis. DISCHARGE 53 log S,=1.73—0.010 C, S.E.=0.24 log units (2) log Sy=0.86+0.143 S—0.010 C, S.E.=0.22 log units (3) log Sy=0.67+0.143 S+0.002 B—0.010 C, S.E. =0.22 log units (4) Where Sy=construction site suspended-sediment yield, in tons per acre per yr; C=percentage of construction area with sedi- ment control; S =average slope of construction sites, in per- cent; B=percentage of construction area Without a buffer of natural vegetation. Equation 2 explains 45 percent of the variation of construction-site yield. Equations 3 and 4 explain 59 and 60 percent of the variation, respectively. The use of additional variables did not significantly in- crease the explained variation of site yield. The analysis suggests that sediment controls used to limit erosion and sediment transport from the construction sites significantly reduced construc- TABLE 10.—Construction-site sediment yields and related factors C Ratio C t C t Average Average ”33?.” 3?... ”5253.” 01350;?”- Construc- gamut)! slogfe 911;? Leaf t 951% t 9.11:; Period fill-De: s ylirgignt cogstruc- "2.700 f? «1551:»- “2100311: admuate (acres) ( tons/ 8:2: I: of 1 mam: bufie: 55311131131: acres ) (percent) (59:31:13) total (pezx‘i’cnent) (percent) (percent) Williamsburg Run near Olney, Md. Oct. 1966—Mar. 1968 (1) 109 20.6 5.45 25.3 99.3 96.6 10 Apr. 1968—Mar. 1969 (2) 68 24.4 5.75 20.4 83.3 85.2 36 Apr. 1969—Mar. 1971 (3) 45 22.8 5.97 26.1 39.8 52.0 42 Apr. 1971—Dec. 1971 (4) 74 14.4 4.30 20.0 24.0 88.0 49 Jan. 1972—Sept. 1973 (5) 90 8.6 5.15 17.4 39.0 75.6 60 North Branch Rock Creek near Norbcck, Md.1 Oct. 1966—Mar. 1968 (1) 87 46.4 6.70 52.1 45.1 88.7 5 Apr. 1969—Mar. 1971 (2) 60 63.0 7.64 20.3 74.3 81.4 36 Apr. 1971—Dec. 1971 (3) 63 12.5 7.34 29.2 58.3 91.7 49 Jan. 1972—Dec. 1973 (4) 52 15.1 6.94 7.0 49.2 52.8 60 Manor Run near Norbeck, Md. Mar. 1967—Sept. 1968 (1) 73 96.4 6.38 36.6 87.2 81.7 10 Jan. 1969-Dec. 1969 (2) 62 100.8 6.72 41.7 88.9 100.0 15 Apr. 1970—Dec. 1970 (3) 48 54.4 5.90 37.5 80.0 87.5 27 Feb. 1971—Dec. 1972 (4) 30 21.2 7.28 22.2 46.1 88.6 52 Be] Pre Creek at Layhill, Md. Jan. 1966-Dec. 1967 (1) 150 21.1 5.72 19.3 88.1 65.1 0 Jan. 1968-Dec. 1969 (2) 113 26.4 5.92 32.4 92.4 93.8 35 Jan. 1970—Dec. 1971 (3) 100 16.5 5.56 34.8 83.0 67.8 45 Jan. 1973—Sept. 1974 (4) 142 7.2 5.58 25.0 91.7 97.0 68 Lutes Oct. 1966—Jan. 1968 (1) 29.5 31.8 5.42 55.0 35.0 75.0 0 Mar. 1968—Apr. 1970 (2) 13.0 17.0 6.58 64.8 78.5 38 1 Represents North Branch Rock Creek near Norbeck excluding Williamsburg Run. 54 tion-site suspended-sediment yields. The slope of the construction sites was also a significant factor af- fecting the site yields. Construction on mild slopes resulted in lower sediment yields than construction on steep slopes. The relative size of the largest con- struction site in the basin and the amount of con- struction within 200 ft (61 m) of the stream chan- nel did not significantly afiect the suspended-sedi- ment yields. This outcome could be due to the small number of observations available, or due to the effect of other factors not analyzed in the regression analysis. Figure 16 illustrates the slope factor in two of the study basins. The average slope on construction sites in Bel Pre Creek basin was 5.72 percent dur- ing the January 1966 to December 1967 sampling period, and the average annual sediment yield was 21.1 (tons/acre) /yr [(47.3 t/hm2)/yr]. In contrast, 77°06 T MANOR RUN BASIN\ 39., EXPLANATION 06’ Construction area Average slope conditions in percent /. S H . {‘52}; if ”/4 H 'l ="‘ - , _ 7 -='-7/_-7g/ - E 8-15 "/-b ,—-v——."\\%- ”#25 ”5 =‘7\=':'/ 1:! 11m /A —- - — Basin drainage divide '/|2 J1 MlLE «@xywi’ EFFECTS OF URBANIZATION ON STREAMFLOW AND SEDIMENT TRANSPORT IN MARYLAND the average slope on the Manor Run construction sites was 6.38 percent and the average annual sedi- ment yield was 96.4 (tons/acre)/yr [(216 t/hm2)/ yr] between March 1967 and September 1968. Vir— tually all the construction in Bel Pre Creek was on slopes ranging from 3 to 8 percent, Whereas a large part of the construction area in Manor Run was on slopes between 8 and 15 percent. Some sites had slopes greater than 15 percent. The photograph in figure 17 shows the severity of erosion on the long, steep slopes in the Manor Run basin. The extensive rill and gully erosion on slopes greater than 10 per- cent apparently resulted in the high sediment yield from this basin. The implementation of sediment controls is illus- trated by aerial photographs of the Bel Pre Creek basin for June 1966 and June 1974 (fig. 18). The 1966 photograph shows 148 acres (60 hm?) of land 77°03’ i BEL PRE CREEK BASIN 0 | f I U FIGURE 16.—Construction sites and average slope conditions, i KILOMETER 1 J Bel Pre Creek and Manor Run basins, June 1967. Land slope adapted from Soil Survey of Montgomery County, Md. (US. Dept. Agriculture, 1961) SEDIMENT DISCHARGE 55 FIGURE 17.—Extensive erosion in the Manor Run basin, 1967. under construction with no sediment controls. The 1974 photograph shows about the same amount of construction controlled by a network of five sediment basins. Some of the basins are temporary and others are designed as permanent stormwater management structures (fig. 19). The increased use of sediment basins and other control measures has substantially reduced the construction-site sediment yields of the study basins. For example, the average annual sedi- ment yield delivered to the stream channels from construction sites shown in figure 18 was reduced from 26 to 7 tons/acre (58 to 16 t/hmz) between 1968 and 1974. To further examine the relation between construc- tion-site sediment yields and various site factors, the relation between computed site yields and the yields estimated with equation 3 is shown in figure 20. The scatter about the line of equality is signi- ficant, reflecting the large standard error of esti- mate of the regression equation. The data points with the greatest departure from the line of equality were probably influenced by other site factors not used in the equation. For example, the high com- puted yields for Manor Run sample periods 1, 2, and 3 probably result fro-m grading activities ad- jacent to and within the stream channels (fig. 21). The proximity factor averaged 39 percent in Manor Run during these periods. In contrast, the relatively low proximity factors for period 1 in the Williams- burg Run and Bel Pre Creek basins, 25.3 and 19.3 percent, are reflected by the low computed sediment yields. An earlier study in the same basins found that construction-site sediment yields from summer storms increased as the amount of construction within 100 and 300 ft (30 and 91 m) of stream channels increased (Yorke and Davis, 1972). The lower sediment yields from construction sites farther from stream channels represent the effects of filtering of sediment-laden runoff waters by vege- tation. As runoff transporting sediment from a con- strucfion site passes through a heavily vegetated zone, velocity is reduced by an increase in friction. 56 EFFECTS OF URBANIZATION ON STREAMFLOW AND SEDIMENT TRANSPORT IN MARYLAND APPROXIMATE SCALE 1:24 000 FIGURE 18.—Construction activities, A, without (1966) and, B, with (1974) sediment controls, Bel Pre Creek basin. Circles indicate locations of sediment basins. This action reduces the competence of the water to transport sediment. A less important factor in the reduction in competence is a loss of water through infiltration into the soil of the vegetated zone. As a consequence, much of the sediment is deposited before it enters the stream. The low computed yields in the Lutes Run basin, where the percentage of the construction area with sediment controls was lower than average and the slope of construction sites was higher than average, are indicative of another factor affecting site yields. One construction site in this basin was opened prior to 1963 and remained undeveloped through 1973 (fig. 22). The low yields from the Lutes Run basin, in general, and this site in particular, probably re- sulted from the site being open long enough for natural stabilization to develop. As fine soil particles are eroded and transported from a site, the re- maining large particles armor the soil mass. The gradual increase in the size of remaining soil par- ticles results in decreased sediment yields because the particles are too large to be transported except by the very large storms. In summary, the degree of sediment control and the slope of construction sites are the most signifi- cant factors affecting construction-site sediment yields. Buffer zones of vegetation, proximity of con- struction to stream channels and the length of time the sites are not stabilized affect the annual yield to a lesser degree. STREAM—CHANNEL EROSION In anticipation of significant urbanization during the period covered by the project, three channel reaches were selected to study possible changes in channel geometry. The reaches are located immedi- ately downstream from the gaging stations (fig. 1) on Manor Run, Northwest Branch Anacostia River at Norwood, and Batchellors Run. Drainage areas above the study reaches range from 0.47 to 2.45 SEDIMENT DISCHARGE 57 APPROXIMATE SCALE 1224000 FIGURE 18.——Continued basin FIGURE 19.—Large sediment-stormwater management used to trap sediment before it can leave construction sites, Bel Pre Creek basin, 1974. mi2 (1.22 to 6.35 kmz). The basins were basically rural at the beginning of the study period with only small amounts of impervious area. Stream channels selected for study are incised in Wehadkee silt loam, which is commonly found in stream valleys of the study area. The texture of Wehadkee silt loam varies from a silty clay loam to a light silt loam, and the soil is susceptible to frost action and very erodible (U.S. Dept. of Agriculture, 1961). The Manor Run and the Northwest Branch Anacostia River channels are typical U-shaped channels carved in flood-plain sediments. Low berms defining the low-water channel are generally absent except at the inside of meanders. In contrast, the Batchellors Run channel has distinct berms or ter- races between the low-water channel and the flood plain (fig. 23). Channel-geometry reaches were established and surveyed in 1967. Cross sections in these reaches 58 EFFECTS OF URBANIZATION ON STREAMFLOW AND SEDIMENT TRANSPORT IN MARYLAND ESTIMATED SUSPENDED-SEDIMENT YIELD, IN TONNES PER SQUARE HECTOMETER 20 50 100 200 I I I I I I I I | I 100 _ I I I I I I I I 2.II I I I |_ M 1 __ E 200 ,_ LU — EXPLANATION “ 5 Lu T A WiIIiamsburg Run 2 _ Q g _ o N. Branch Rock Creek 0 _ :: LIJ n: I Manor Run .3 c: 3-] 50 i V Bel Pre Creek '0“ _ ‘3’: L te R 1 O (2 . u 5 un . KQ’ . — 100 (n 0 (numbers refer to penods 0 I: '— — listed in table 10) '\° “— 33 Z — m d 1 § I: _- O s z > e '2 T E UJ . g 1 -‘ 50 3 3 v a U.) 20 — _ : 8 m z D LU Z 2 E 8 g _ U) m a e e I- LIJ E 10 — — 3) O _ —~ 20 8 0 — — o t‘i _ — D D. 2 _ — 0 U 5 I I I I I I I I I I I I I 5 1o 20 so 100 ESTIMATED SUSPENDED-SEDIMENT YIELD, IN TONS PER ACRE FIGURE 20.——Relation between construction—site sediment yields computed from observed data and those estimated by equation 3. were selected to represent typical conditions and variations in stream-channel geometry. Cross sec- tions were defined by determining the elevation of selected points along a line perpendicular to the stream banks to the nearest 0.1 ft (0.03 m) with transit and rod. Elevations were determined at 1- or 2-ft (0.3- or 0.6-m)-horizontal intervals across the streambed, at 5-ft (1.5 m) intervals in the flood plain, and at breaks in slope. The position of the thalweg was located by azimuth and stadia; thalweg elevations were determined to the nearest 0.01 ft (0.003 m). The channel reaches were resurveyed in 1974 to determine changes in the cross sections and total volume of the active channels. Cross sections were located through the 1967 thalweg and perpendicular to the banks. Additional cross sections were mea- sured in some reaches to facilitate computation of the channel volume. Changes in the cross sections typical of straight and curved reaches of the chan- nels for each stream are shown in. figure 24. jm SEDIMENT DISCHARGE 59 FIGURE 21.—Grading for street right-of-way within the Manor Run stream channel, 1967. From the information obtained by the 1967 and 1974 channel surveys, four channel-geometry param- eters were chosen for study. These parameters include channel length, cross-sectional area, channel volume, and channel width. Channel length was de- fined as the total distance between successive thal- weg locations in the study reaches. Cross-sectional area was determined by plotting the channel cross- sections as determined from the survey field notes and planimetering the area below a line drawn be- tween the top bank elevations. Total. channel volume was calculated as the sum of the products of each cross-sectional area times one-half the distance be- tween the thalweg of the section and the thalwegs of the preceding and succeeding sections. Channel width was determined as the horizontal distance, perpendicular to the channel, between the top of the low bank and the wall of the opposite bank. The effects of urbanization and the associated in- crease in flood peaks are illustrated by the differ- ences in the channel-geometry parameters of Manor Run and the other two streams (table 11). Each of the parameters, except for channel length, increased in the Manor Run basin. as the impervious area in- creased from 5.6 percent in 1966 to 13.5 percent in 1974. In the Northwest Branch Anacostia. River, all the channel parameters increased slightly as the impervious area increased from 3.4 to 4.4 percent. In the Batchellors Run basin, channel parameters increased slightly, except for a slight decrease in channel cross-sectional area, as the impervious area increased from 11 to 12 acres (4.5 to 4.9 hmz), 3.6 to 4.0 percent of the basin area. The changes in the stream channel parameters are consistent with changes found in previous investiga- tions. Leopold (1968) indicated that stream chan- nels form in response to the regimen of the stream and will carry, without overflow, a discharge which is somewhat smaller than the average annual flood. Brown (1971) showed that the 5-year flood has a direct positive relationship with channel capacity. Because urbanization causes increased flood peaks, as discussed earlier, a substantial enlargement of the channels of urban streams should be expected. Apmann (1974) found that in any watershed in which the runoff volumes and flood peaks are in- creasing, the tendency will be for the stream channel to increase its cross-sectional area and length of meander. Hammer (1972) found that channel en- largement was related to the amount of impervious area, particularly impervious areas more than 4 years old. The two parameters indicating a large difference between Manor Run and the other basins were the average cross-sectional area and channel volume. Channel cross-sectional area increased 28 percent in the Manor Run basin. It increased 5 percent and decreased 6 percent in the Northwest Branch Ana- costia River and Batchellors Run basins, respec- tively. Statistical analysis indicated that the 1974 cross sections for Manor Run were significantly larger than the 1967 cross sections at the 95-per— cent confidence level. The difference between the 1967 and 1974 cross sections for the other two basins were not significant at the 95-percent confidence level. The volume of the Manor Run channel in- creased 34 percent between 1967 and 1974. Channel volume increased 12 percent in the Northwest Branch Anacostia River basin and increased 2 per- cent in the Batchellors Run basin during the same period. An examination of the individual cross sections and the streambed profiles of each channel reach in- dicated substantial differences between Manor Run and the other two streams. The increase in channel cross-sectional area in Manor Run resulted from a combination of bank erosion and degradation of the streambed. Each cross section was deeper and all except one section had greater cross-sectional areas in 1974. In straight channel reaches, in particular, cross sections were scoured to a nearly rectangular shape instead of the V- or U-shaped sections com— mon in 1967. In contrast, the cross sections of Batchellors Run and Northwest Branch Anacostia River did not exhibit any consistent pattern of bank erosion or bed degradation. The channels generally 60 EFFECTS OF URBANIZATION ON STREAMFLOW AND SEDIMENT TRANSPORT IN MARYLAND A" B APPROXIMATE SCALE 1224000 C APPROXIMATE SCALE 1224 000 V FIGURE 22,—Construction site in the Lutes Run basin, 1964 (A), 1968 (B), and 1974 (C). SEDIMENT DISCHARGE 61 FIGURE 23,—Stream channel-geometry reaches, 1974: Manor Run (A), North- west Branch Anacostia River at Nor- wood (B), and Batchellors Run (C). 62 STRAIGHT REACHES DISTANCE, IN METERS 0 5 10 10 II II MANOR~RUN Left bank SECTION 9 0 I I I I 10 I I I I I I I BATCHELLORS RUN Left bank 8 V _ .___. SECTION 8 I I l ELEVATION ABOVE ARBITRARY DATUM, IN FEET 10 I ‘ I I I NORTHWEST BRANCH ANACOSTIA RIVER Left bank SECTION 4A 0 I I I I 0 1O 20 30 DISTANCE, IN FEET FIGURE 24.—Typical cross-section changes between 1967 and 1974 in straight (sections 9, 8, and 4A) 40 50 0 CURVED REACHES DISTANCE, IN METERS 5 10 | I EFFECTS OF URBANIZATION ON STREAMFLOW AND SEDIMENT TRANSPORT IN MARYLAND 15 | l MANOR RUN Left bank SECTION 5 I I I 10 I I I I | BATCH ELLORS RUN Left ba nk SECTION 11 I I I | I ELEVATION ABOVE ARBITRARY DATUM, IN METERS I | I NORTHWEST BRANCH ANACOSTIA RIVER Left bank SECTION 9 I I | I o 10 20 30 DISTANCE, IN FEET (sections 5, 11, and 9) reaches of three typical stream channels in the study area. 40 50 and curved SEDIMENT CONTROL TABLE 11.—Summary of 63 stream-channel surveys Impervious Channel Averag e Average cross- Channel Drain e Number channel . v 1 ”g? c 13; 3 margin” Egg? v31?! sectiogtail) area ffgge (m1 ‘) sections 1966 1974 1967 1974 1967 1974 1967 1974 1967 1974 Manor Run near Norbeck ______ 1.01 12 5.6 13.5 271 269 22.6 24.9 59.8 76.8 15,600 20,900 N. W. Br. Ana- costia River at Norwood __ 2.45 15 3.4 4.4 427 444 32.6 34.5 104 109 45,200 50,500 Batchellors Run at Oakdale __- .45 8 3.6 4.0 235 243 34.7 35.9 74.8 70.4 16,800 17,200 retained their original shape within slight shifts of the meander pattern. Virtually all the increase in average cross-sectional area in the Northwest Branch Anacostia River channel was attributable to increases at three sections located on an active meander. The increase in suburban land area and imperv- ious surfaces is apparently the only factor that could have caused the channel enlargement in Manor Run. Each of the streams is carved through similar soils, and the rainfall characteristics for the three basins were similar during the period of compari- son (1967-74). The change in Manor Run channel volume over 7 years has resulted in a removal of approximately 5,300 ft3 (150 m3) of soil from the area of the stream channel. Based on an assumed density of 100 lbs/ft3 (1,600 kg/m“), 265 tons (240 t) of soil was eroded from a channel reach 270 ft (82 m) long. This erosion is equivalent to about 1 ton per foot (3 t/m) of channel over 7 years. In other words, every 7 ft (2.1 m) of channel length contrib- uted approximately 1 ton (0.9 t) of sediment to the stream annually. Based on a total channel length of 5,000 ft (1,520 m) from the outlet of storm sewers in the upstream developments to the gaging station, the contribution of stream-channel erosion to- the total sediment load measured at the gaging station would be about 700 tons/yr (640 t/yr). Suspended-sediment load would probably represent about 90 percent of the total load or about 630 tons/yr (570 t/yr). Suburban land averaged about 250 acres (101 hm?) between 1967 and 1974. If all the stream-channel erosion was attributable to in- creased runoff from impervious surfaces in this area, the sediment yield from stream channels would be about 2.5 tons/yr (2.3 t/yr) for each acre (0.4 hm?) of suburban land. This is a temporary source of sediment that would be expected to diminish to zero» when the stream channel reached equilibrium with the post-development runoff regime. SEDIMENT CONTROL In the past 10 years, sediment control has become an integral part of urban construction. Control pro- grams have been implemented by all local jurisdic- tions in the Baltimore—Washington area. The im- petus for these programs was a study begun by the Interstate Commission on the Potomac River Basin in the early 1960’s. The Commission’s study group recommended that sediment control become the stated policy of local governments, that urban development be planned carefully, and that local ordinances be adopted to require developers to employ erosion- control measures on construction sites (Guy and others, 1963). The Montgomery County sediment-control pro- gram evolved from a Sediment Control Task Force formed to study the possibility of reducing sedi- ment from developing areas in the Rock Creek watershed, which was being considered for a Public Law 566 flood-control project. The Task Force drew up a program for voluntary sediment control, and it was adopted as county policy in 1965. This action was followed by a mandatory program in 1967 that required developers to submit sediment control plans along with preliminary subdivision plans. This regu- lation provided for review of plans by the Mont- gomery County Soil Conservation District; however, there was no provision for field inspections to assure that the approved control measures were installed and properly maintained. In 1971, however, the Montgomery County Council enacted a Grading and Sediment Control ordinance, which established an enforcement unit within the county’s Department of Environmental Protection. This unit has the re- sponsibility of inspecting construction sites to insure that approved sediment-control measures are prop- erly installed and adequately maintained through- out the construction period. The urban sediment-control program adopted by Montgomery County is similar to the program used 64 EFFECTS OF URBANIZATION ON STREAMFLOW AND SEDIMENT TRANSPORT IN MARYLAND for many years in agricultural areas. The basic principles, as outlined by the Maryland-National Capital Park and Planning Commission (1967), are: 1. The smallest practical area of land should be exposed at any one time during development. 2. When land is exposed during development, the exposure should be kept to the shortest prac- tical period of time. 3. Temporary vegetation and (or) mulching should be used to protect critical areas exposed during development. 4. Sediment basins (debris basins, desilting basins, or silt traps) should be installed and main- tained to remove sediment from runoff waters from land undergoing development. 5. Provisions should be made to effectively accom- modate the increased runoff caused by changed soil and surface conditions during and after development. 6. The permanent final vegetation and structures should be installed as soon as practical in the development. 7. The development plan should be fitted to the topography and soils so as to create the least erosion potential. 8. Wherever feasible, natural vegetation should be retained and protected. FIELD APPLICATIONS The field application of the principles outlined above required a gradual training program for de— velopers, engineers, and grading contractors. In the past, developers generally did all the rough grading on the site at one time to minimize expenses for heavy grading equipment. Engineers and grading contractors designed and built drainage systems to facilitate rapid runoff from the sites. Final road surfaces, drainage channels, and vegetative cover were installed after houses were sold and ready for occupancy. Major changes in these methods of operation were required to accomplish the principles of the sediment-control program. The first attempt at sediment control was the in- stallation of sediment basins. These basins could be built at minimal cost at the same time heavy equip— ment was rough-grading street rights-of—way. They were generally located in small drainage swales at the edge of the development so they would not inter- fere with the standard construction methods. In general, they were small and underdesigned, as il- lustrated by figure 25. Many of these basins were f} S3 3M“ (a: 3.. FIGURE 25.—Typical sediment basin used in 1966. filled with sediment in one or two storms and were of no value thereafter. The net result was a gen- erally ineffective program between 1965 and 1968. The lack of effective control during the early stages of the program occurred because it was a learning period for all concerned. Representatives of the Soil Conservation District and the county were trying to develop effective standards and speci- fications for measures that would control sediment transport from construction sites, but that would not place undue restrictions on construction activi- ties. At the same time, developers and engineers were trying to control erosion without the benefit of adequate standards. As. specifications became available, the pro-gram gradually improved: Better planning limited grading to the smallest practical area to reduce the amount of land exposed to erosion at any one time; straw mulch and temporary vege- tation were used to protect exposed soils; diversion berms, level spreaders, and stabilized water-ways were used to reduce erosion on critical slopes; and properly designed sediment basins were used to trap the sediment before it could leave the construction site (U.S. Dept. of Agriculture, 1969). Between 1968 and 1971, most of the sediment- co-ntrol plans were implemented in the field; how- ever, there were still problems with contractors un- familiar with sediment control measures. Some control measures were installed improperly, and many others were not adequately maintained. There was also a lack of coordination between the different SEDIMENT CONTROL 65 contractors involved in the housing developments. Rough-grading, utility installation, and paving were not coordinated so as to reduce the length of time that construction sites were subject to erosion. Sometimes utility contractors would remove or alter controls that were installed by the developer. For example, interceptor berms on rights-of-way were frequently removed during utility installation and not replaced after the utilities were in place. In other instances, the embankments of sediment basins were breached for installation of sewer lines and were not rebuilt, resulting in the subsequent dis- charge of sediments previously trapped by the basin (fig. 26). Photograph A in figure 26 shows a well- built basin at one of the construction sites in the study area. Photographs B and C show the same basin after the embankment was breached for a sewer line. This embankment was eventually re— built, but not until much of the trapped sediments had been scoured by a series of storms. Another problem between 1968 and 1971 was a general lack of maintenance of control measures. Berms destroyed or altered by construct-ion traffic were not replaced. Temporary vegetation and mulch were not maintained. Sediment basins, the prime means of control, were not maintained. When basins became filled, the accumulated sediments were not removed. As the trap efficiency is related to the ratio between the storage capacity of basins and. the inflow, the accumulation of sediment: greatly re- duced the value of the basins. During some storms, sediment-filled basins probably had a negative trap efficiency, as runoff scoured more sediment from the basins than was deposited in them (fig. 27). After the Sediment Control Section was estab- lished in the Montgomery County Department of Environmental Protection, control measures im- proved markedly. Inspectors were on hand to see that controls were installed properly. Inspections at regular intervals throughout the period of construc- tion ensured that controls were adequately main- tained and that additional controls were installed as needed. Improvements in the sediment control pro- gram are illustrated by the summary of field evalua- tions listed: FIGURE 26.———Sediment basin before (A) and after (B and C) it was breached for installation of a sewer line. Note the scour of previously accumulated sediments. 66 Percentage of total cort- struction area with approved sediment- control measures Year 1968 _______________________________ 36 1970 _______________________________ 41 1972 _______________________________ 56 1974 _______________________________ 74 These field evaluations were made by several groups. The 1968 and 197 0 evaluations were made by mem- bers of the Montgomery County Sediment Control Task Force. The 1972 evaluation was made by the Montgomery County Sediment Control Section, and the 1974 evaluation was made by representatives of the Sediment Control Section and US. Geological Survey personnel assigned to this project. All except the 1974 evaluation were countywide. The 1974 evaluation considered only sites in the study area. EFFECTS OF SEDIMENT-CONTROL MEASURES The effectiveness of the sediment-control program was illustrated in the earlier discussion of construc- tion-site sediment yields. The quantity of suspended sediment transported by the streams decreased markedly as the percentage of construction sites with approved controls increased. In the Bel Pre Creek basin, suspended-sediment yields from con- struction areas decreased from 26.4 tons/ acre (59.2 t/hmz), when there were minimal controls, to 7.2 tons/acre (16.1 t/hmz), when about 70 percent of the construction area. was adequately controlled. Similar decreases were observed in the other study basins. Sixty to eighty percent decreases were ob— served in the Williamsburg Run, North Branch Rock Creek, and Manor Run basins. Another example of reduced sediment yields from construction sites is shown in figure 28. The curves FIGURE 27.—Small sediment basin requiring maintenance. EFFECTS OF URBANIZATION ON STREAMFLOW AND SEDIMENT TRANSPORT IN MARYLAND in figure 28 represent the average relation between storm runoff and suspended-sediment discharge in the Northwest Branch Anacostia River near Coles- ville for three periods: 1963-67 water years, when no controls were used on construction sites; 1968—71 water years, when controls were mandatory; and 1972—74 water years, when controls were mandatory and subject to inspection by the Sediment Control Section. The curves indicate a substantial reduction in sediment discharge during each succeeding period. An analysis of covariance found that the re- gression curves for the dormant-season storms were significantly different for all periods. The differences of the adjusted means of the dormant-season loads for the periods 1963—67 versus 1968—71 and 1968—71 versus 1972-74 were significant at the 95-percent confidence level. The difference betWeen the 1963—67 and 1972—74 adjusted means were significant at the 99-percent confidence level. Growing-season ad- j usted means were significantly different at the 99- percent confidence level, except for the periods 1963—67 and 1968—71, which were not significantly different at the 95- or 99-percent confidence level. Several factors could account for the reduced suspended-sediment discharge from the basin. A decrease in the amount of sediment source area would be the most obvious. Land under construc- tion averaged 2.87, 3.15, and 2.63 percent of the total basin area for 1963—67, 1968—71, and 1972—74, re- spectively. On the basis of construction area alone, a slight increase over the 1963—67 period would be expected in 1968—71 and a slight decrease expected between 1972 and 1974. However, a decrease was observed between 1968 and 1971. The large decrease in sediment discharge between 1972 and 1974 is not the likely result of the small decrease in con- struction area. Cropland decreased during the period, but the increase in new urban land was greater than the decrease in cropland. The sediment yield from these two sources probably did not change appreciably because the yield from urban land and stream channels downstream from urban land was greater than or equal to the cropland yields observed for the small study basins. Changes in soil and slope conditions on construc- tion sites and a decrease in the size or intensity of storms could also affect the sediment discharge, but these did not change significantly during the study' period. Soil conditions were generally uniform ex- cept for the greater erodibility of the subsoil in the Manor series. A substantial decrease in construc- tion on the Manor Soils was not observed. Slope varied between construction sites, but there was no SEDIMENT CONTROL 67 STORM RUNOFF, IN CUBIC HECTOMETERS 0.05 0.1 0.5 1 10'000_. . "5. - EXPLANATION 3 : 1963-67 0 ‘. 5000 . 1968—71 A ———————— -_ 500° - 1972—74 . _ w ' e Z - .- z 9 E 2 $1000 - @000 g o - m‘ 2% I 2- g 5 500 - _: 500 < r2 - - 5 o . ' Q ’— D E ' 5 E u.| a 100 g If) 100 : 2; 8 E — :- U) 1: _ -' E o - 50 I I- 50 ' _ w - 2 . U) . ‘A . 10 . ..-...... . .......h10 1o 50 100 500 1000 STORM RUNOFF, IN CUBIC FEET PER SECOND-DAYS STORM RUNOFF, IN CUBIC HECTOMETERS 0.05 0.1 0.5 1 10,000 “"‘” .J. W‘ , EXPLANATION : , 1963—67 . _: 5000 * 1968—71 A ______ __ 5000 1972—74 - __ '0 ' 1000 500 a O O STORM SEDIMENT DISCHARGE, IN TONS U1 0 STORM SEDIMENT DISCHARGE, IN TONNES 10 . . . ._._. . .. . . . . . - 10 10 50 100 500 1000 STORM RUNOFF, IN CUBIC FEET PER SECOND-DAYS FIGURE 28.—Relation between suspended sediment and storm runoff for dormant season (A) and growing season (B) storms, Northwest Branch Anacostia River near Colesville, 1963—74. trend of decreasing slope conditions during the 12- year period. Total precipitation varied during the periods, but the magnitude and intensity of storms did not decrease significantly. The use of sediment controls is the one factor that has changed appreciably during the three periods represented by the curves in figure 28. Total suspended-sediment discharge that would have oc— curred if each condition had existed throughout the 12-year period was estimated as shown: 12-1” sediment load transported by storms Degree of sediment control Growing Dormant season 3808011. (tons) (tons) None _____________________ 114,000 93,000 Mandatory, no enforcement __ 79,000 60,000 Mandatory, with enforce- ment ___________________ 52,000 52,000 The seasonal regression curves and the storm runoff for all significant storms that occurred be- tween October 1962 and September 1974, except for Hurricane Agnes, were used in the computations. The Agnes storm runoff was not used because it far exceeded the sample range of the 1963—67 and 1968- 71 periods. Suspended-sediment discharge of the storms used in the computations is 92 percent of the 12-year total sediment discharge, excluding the load transported during Hurricane Agnes The other 8 percent represents the load transported during base flow and small storms for which no samples were collected. The estimated loads indicate that a substantial reduction in sediment discharge would have oc- curred if sediment controls had been used through- out the study period. Sediment yields under manda- tory controls were estimated to be 33 percent lower than yields estimated for conditions of no control. A slightly larger reduction would have occurred during the dormant season. Enforcement of controls throughout the study period would have resulted in a total yield reduction of 103,000 tons (93,400 t), or 50 percent. The largest reduction under enforced controls occurred during the growing season. This is probably the result of an improvement in the de- signs of structures and better maintenance, which provided greater control of runoff during intense summer storms. Scheduled inspections of construc- tion sites resulted in prompt replacement and main- tenance of sediment controls damaged by normal construction traffic and a substantial reduction in sediment discharge. COST OF SEDIMENT CONTROLS Sediment-control measures implemented on urban construction sites apparently have reduced the sedi- 68 EFFECTS OF URBANIZATION ON STREAMFLOW AND SEDIMENT TRANSPORT IN MARYLAND ment loads in the receiving streams. The installation of these controls, however, represents an additional expense to developers, builders, and ultimately, the property owner. The standard control measures used in the field, the cost of the control program, and the relative costs incurred due to sediment damage downstream from construction sites are discussed in this section. After standards and specifications for soil erosion and sediment control (US. Department of Agricul- ture, 1969) were prepared, sediment-control mea- sures in the field have become fairly uniform. The conventional method used on most construction sites includes a combination of: 1. Mulch and (or) temporary vegetation to protect exposed slopes. 2. Interceptor dikes to reduce erosion on rights-of— Way by diverting storm runoff to temporary outlets where the water can be transported with minimal erosion. 3. Grassed waterways, level Spreaders, and grade stabilization structures to convey storm runoff through the construction site without erosion. 4. Diversion berms to divert storm runoff from areas with critical slopes. 5. Sediment basins to trap and store sediment eroded from construction sites before it can enter the stream system. Complete descriptions and illustrations of these and other commonly used control measures are available in reports by the U.S. Department of Agriculture (1969), Thronson (1971), and Becker and Mills (1972). The cost of these conventional controls has been evaluated in several studies. Brandt and others (1972) interviewed developers in Montgomery County, Md., to determine the average cost of indi- vidual controls. Hotes and others (1973) obtained similar information for a cost comparison between sediment controls in northern Virginia and central California. The costs for the commonly used control measures in Maryland and northern Virginia were estimated as: Surface stabilization _______ $209—$567/ acre Diversion and intercepter dikes ___________________ $1.25—$3.70/1n ft (linear foot) Sodded ditches _____________ $4.50/ln ft Grade stabilization structures _______________ $2.80—$28/ln ft Level Spreaders ____________ $1.36—$3.16/ln ft Sediment basins ___________ $1,500—$2,560/basin The range in costs reflects differences in the type of materials used for the controls, differences in sediment-control specifications in Maryland and Vir- ginia, and different time periods during which the costs were determined. An average unit cost of $1,026/acre for sediment control was estimated by Brandt and others (1972) , based on sediment-control plans submitted by seven developers in Montgomery County. This represents an average cost of $55/acre for engineering, plus $421/acre for structures and sediment basins, plus $500/acre for surface stabilization, plus $150/acre for maintenance of structures and repair of erosion damage, minus $100/acre for the residual value of surface stabilization. The significance of these costs can be shown by the total cost of sediment control in the Anacostia River basin. Between 1962 and 1974, 1,900 acres of land was developed for houses, apartments, schools, and shopping centers. If an average cost of $1,026/acre is assumed, the total cost of sediment control would have been $1.9 mil- lion. This represents an approximate cost of $19 for each ton (0.9 t) of sediment that would have been prevented from entering the stream system if mandatory controls had been enforced between 1962 and 1974. p The cost of onsite sediment control falls some- where within the range of costs incurred for sedi- ment removal from ponds, reservoirs, streets, base- ments, and sewers. These costs range from $2.50/ ton for removal of sediment from an urban pond (Becker and Mills, 1972) to $134/ton for removal from sewers (Hot-es and others, 1973). Dredging cost can range from $0.11 to $9.00/ton depending on the amount of sediment to be dredged and the distance from the dredging operation to a suitable disposal site (Hows and others, 1973). Many other costs are incurred as a result of increased sedimen- tation caused by urban construction; however, an analysis of these costs is beyond the scope of this study. These include: increased water treatment costs, increased flooding, decreased commercial fish and shellfish harvests, decreased sport fishing, de- creased esthetic values in streams and estuaries, and decreased shipping due to harbor shoaling. Most sediment damages and the related reduction in costs resulting from onsite sediment control can be calculated, but there are several damages that cannot be evaluated based on present knowledge. Several unanswered questions are: What are the toxic levels of turbidity, suspended sediment, and pollutants associated with suspended sediment for indigenous fish and other aquatic life in the local streams and ponds? What is the effect of estuarine sedimentation on marine life cycles? What is the SUMMARY 69 acceptable level of turbidity and sedimentation for recreational use of streams and lakes? Will sediment controls reduce the impact of urban sediment on these conditions? When these questions are an- swered, local officials will be better able to deter- mine whether the costs of sediment control are justified by the benefits. SUMMARY This report presents the results of a field investi- gation of the effects of urbanization and sediment- control measures on streamflow and sediment trans- port in Montgomery County, Md. The study was made between 1962 and 1974 in a 32-mi2 (83-km2) area in the headwaters of the Rock Creek and Anacostia River basins. Data from 9 streamflow-sediment sta- tions, 9 recording rain gages, numerous nonrecord- ing rain gages, 14 sets of aerial photographs, and channel surveys on 3 stream reaches were used for the study. Land use/land cover varied considerably in the study basins. Three subbasins remained virtually rural, the others underwent urban development with as much as 20 percent of one subbasin undergoing active construction at one time. Urban area ranged from 0 to 60 percent in the subbasins in 1974. Suburban land in the Anacostia River basin in- creased from 6 to 23 percent of the drainage area between 1962 and 1974. Precipitation and streamflow between October 1962 and September 1974 were essentially the same as the long-term conditions in. the study basins. Streamflow was higher than normal between 1966 and 1974, and the increase in volume was mostly due to higher base flows. Urbanization did not affect median and low flows in the study basins; however, storm runoff and peak discharges were greater in the urban basins. Suspended sediment transported from the Ana- costia River basin averaged 14,800 tons (13,400 t) per year between 1962 and 1974. The annual sus- pended-sediment discharge ranged from 5,500 tons (4,990 t) in 1974, to 31,000 tons (28,100 t) in 1972, when the runoff from Hurricane Agnes transported 12,800 tons (11,600 t) in a 2-day period. Bedload was estimated to be 6 to 13 percent of the total load. Most of the load was transported during storms, 73 percent in 2.2 percent of the time and 94 percent in 5.7 percent of the time. The temporal distribution of suspended sediment closely approxi- mated the seasonal variability of the erosion index; high loads generally occurred in the late winter, early spring and midsummer, and low loads occurred in the fall and early winter. Sediment loads were highly variable from storm to storm. Loads were most closely related to the volume of storm runoff and peak discharge. Other factors affecting the load variation for individual streams include the relative basin area under con- struction, antecedent moisture levels, total precipi- tation, and an index of the storm intensity. Useful storm-intensity indices were maximum 15- or 30- minute rainfall or the ratio of peak discharge to runoff volume. Cropland, urban land, and construction sites were the major sources of sediment. Suspended-sediment yields from land under cultivation ranged from 0.65 to 4.3 tons/acre (1.5 to 9.6 t/hmz). Yields from forest lands and grasslands were estimated as 0.03 ton/acre (0.07 t/hmz) and 0.2 ton/acre (0.45 t/hmz), respectively. Urban land yielded about 3.7 tons/acre (8.3 t/hmZ), mostly from stream-channel erosion immediately downstream from newly com- pleted residential and commercial areas. In one urban basin, channel surveys indicated that stream- channel erosion contributed 1 ton (0.9 t) of sedi- ment per foot (0.3 m) of main channel length to the total sediment yield of the basin between 1967 and 1974. Suspended-sediment yields from urban construc- tion sites ranged from 7 to 100 tons/acre (16 to 224 t/hm2) and averaged 33 tons/acre (74 t/hmz). Most of the variation in yields was attributable to the increased use of effective: sediment-control mea— sures in construction areas between 1966 and 1974. The average slope of the construction sites also was determined to be a significant factor affecting sedi- ment yields. Other factors affecting sediment yields to a lesser extent were the proximity of construc- tion to stream channels and the existence of buffer zones of undisturbed vegetation between construc- tion sites and stream channels. Construction-site sediment yields decreased as use of sediment-control measures increased. Yields were reduced from 26 tons/ acre (58 t/hmz) to 7 tons/acre (16 t/hmz) in the Bel Pre Creek basin. Sixty to eighty percent decreases were observed in the other subbasins. If controls had been used between 1963 and 1974 in the Anacostia River basin, suspended- sediment discharge would have decreased about 103,000 tons (93,400 t), or 50 percent between 1963 and 1974. Reportedly, the cost of controlling sedi— ment on the 1,900 acres developed during this period would have been $1,030/acre (0.4 hm?) or $19 for each ton (0.9 t) of sediment controlled. The 70 EFFECTS OF URBANIZATION ON STREAMFLOW AND SEDIMENT TRANSPORT IN MARYLAND data indicate that control could be achieved at less expense if construction were limited to areas with slopes less than 10 percent and if sites immediately adjacent to stream channels were avoided. If con- struction were required in these critical areas, sub- stantial control could be achieved by limiting con- struction to the period of low erosion potential be- tween September and January. These restrictions would reduce the direct cost of sediment control, but they may introduce additional costs whose evalu- ation is beyond the scope of this report. Standard cost-benefit methods can only be used to partly assess the effect of enforcing sediment-control measures or alternative restrictions on construction activities. Although costs of controlling sediment on site can be compared with the costs of removing sediment at downstream locations, there are in- tangible damages associated with allowing sediment to leave the construction site. It is difficult to assign costs of increased flood potential, deterioration of the esthetic quality of streams, and damage to the flora and fauna of streams, lakes, and estuaries, but there can be little doubt that these costs are real. SELECTED REFERENCES Anderson, D. G., 1970, Effects of urban development on floods in northern Virginia: U.S. Geol. Survey Water-Supply Paper 2001—C, 22 p. Apmann, R. P., 1974, The influence of urbanization on stream channel behavior, in National Symposium on urban rairn- fall and runoff and sediment control, July 1974: Univ. of Kentucky Bull. 106, 246 p. Becker, B. 0., and Mills, T. R., 1972, Guidelines for erosion and sediment control planning and implementation: U.S. Environmental Protection Agency Rept. EPA—R2—72—015, 228 p. Brandt, G. H., and others,’1972, An economic analysis of erosion and sediment control methods for watersheds undergoing urbanization: Dow Chemical Co. Rept. sub- mitted to U.S. Dept. of Interior (Water Resources Re- search Contract No. 14—31—0001—3392), 181 p. Brown, D. A., 1971, Stream channels and flow relations: Water Resources Research, v. 7, no. 2, p. 304—310. Carter, R. W., and Davidian, Jacob, 1968, General procedures for gaging streams: U.S. Geol. Survey Techniques Water- Rcsources Inv., book 3, chap. A6, 13 p. Guy, H. P., 1964, An analysis of some storm-period variables affecting stream sediment transport: U.S. Geol. Survey Prof. Paper 462—E, 46 p. 1965, Residential construction and sedimentation at Kensington, Maryland, in Federal Inter—Agency Sedimen- tation Conf. Proc., 1963: U.S. Dept. of Agriculture Misc. Pub. 970, p. 30—37. Guy, H. P., and Ferguson, G. E., 1962, Sediment in small reservoirs due to urbanization: Am. Soc. Civil Engr., Jour. Hydraulics Div., v. 88, no. HY2, p. 27—37. Guy, H. P., and Norman, V. W., 1970, Field methods for measurement of fluvial sediment: U.S. Geol. Survey Tech- niques Water-Resources Inv., book 3, chap. C2, 59 p. Guy, H. P., and others, 1963, A program for sediment control in the Washington metropolitan region: Interstate Comm. Potomac River Basin Tech. Bull., 1963—1, 48 p. Hammer, T. R., 1972, Stream channel enlargement due to urbanization: Water Resources Research, v. 8, no. 6, p. 1530—1540. Hotes, F. L., and others, 1973, Comparative costs of erosion and sediment control, construction activities: U.S. En- vironmental Protection Agency Rept. EPA 430/9-73— 016, 205 p. Leopold, L. B., 1968, Hydrology for urban land planning—A guidebook on the hydrologic effects of urban land use: U.S. Geol. Survey Circ. 554, 18 p. Maryland-National Capital Park and Planning Commission, 1967, Sediment control program for Montgomery County, Maryland: Maryland-National Capital Park and Plan- ning Comm., 6 p. Miller, C. R., 1951, Analysis of flow-duration, sediment-rating curve method of computing sediment yield: U.S. Dept. of Interior, Bur. of Reclamation, 55 p. Putman, A. L., 1972, Effect of urban development on floods in the Piedmont province of North Carolina: U.S. Geol. Survey open— file report, 87 p. Searcy, J. K., 1959, Flow4duration curves: U.S. Geol. Survey Water-Supply Paper 1542—A, 33 p. Sheppard, J. R., 1960, Investigation of Meyer-Peter, Muller bedload formulas: U.S. Bur. of Reclamation, 22 p. Shulits, Sam, 1935, The Schoklitsch bedload formula: Engi- neering, v. 139, p. 644—646, 687. Thronson, R. E., 1971, Control of erosion and sediment deposi- tion from construction of highways .and land develop- ment: U.S. Environmental Protection Agency, 50 p. U.S. Department of Agriculture, 1961, Soil Survey of Mont- gomery County, Maryland: U.S. Dept. of Agriculture, Soil Conservation Service, 107 p. 1969, Standards and specifications for soil erosion and sediment control in urbanizing areas: U.S. Dept. of Agri- culture, Soil Conservation Service, College Park, Md. 1970, Soil loss prediction guide for construction sites in Virginia; U.S. Dept. of Agriculture, Soil Conserva— tion Service, Virginia. Ursic, S. J., and Dendy, F. E., 1965, Sediment yields from small watersheds under various land use and forest covers, in Federal Inter-Agency Sedimentation Conf. Proc., 1963: U.S. Dept. of Agriculture Misc. Pub. 970, p. 47—52. Vice, R. B., Guy, H. P., and Ferguson, G. E., 1969, Sediment movement in an area of suburban highway construction, Scott Run basin, Fairfax County, Virginia, 1961—64; U.S. Geol. Survey Water-Supply Paper 1591—E, 41 p. Wark, J. W., and Keller, F. J., 1963, Preliminary study of sediment sources and transport in the Potomac River basn: Interstate Comm. Potomac River Basin Tech. Bull., 1963-11, 28 p. Williams, K. F., and George, J. R., 1968‘, Preliminary appraisal of stream sedimentation in the Susquehanna. River basin: U.S. Geol. Survey open-file report, 73 p. Wischmeir, W. H., and Smith, D. D., 1965, Predicting rain- fall—erosion losses from cropland east of the Rocky Moun- SELECTED REFERENCES 71 tains: U.S. Dept. of Agriculture, Agriculture Handb. 1964, Problems posed by sediment derived from con- 282, 47 p. struction activities in Maryland: Rept. to the Maryland Water Pollution Control Comm, Annapolis, Md., 125 p. Wolman, M. G., 1955, The natural channel of Brandywine Yorke, T. H., and Davis, W. J., 1972, Sediment yields of urban Creek Pennsylvania: U.S. Geo]. Survey Prof. Paper 271, construction sources, Montgomery County, Maryland: 56 p. U.S. Geol. Survey open-file report, 39 p. inUS. Government Printing Ofiice: 1978—240-961/163 [I PREPARED IN COOPERATION WITH THE MARYLAND-NATIONAL CAPITAL PARK AND PLANNING COMMISSION (MONTGOMERY COUNTY AND PRINCE GEORGES COUNTY PLANNING BOARDS), THE WASHINGTON SUBURBAN SANITARY COMMISSION, THE DISTRICT OF COLUMBIA DEPARTMENT OF ENVIRONMENTAL PROF 88 I 03 UNITED STATES DEPARTMENT OF THE INTERIOR SERVICES, THE MONTGOMERY COUNTY DEPARTMENT OF ENVIRONMENTAL , \, GEOLOGICAL SURVEY PROTECTION, AND THE MARYLAND DEPARTMENT OF NATURAL RESOURCES ,_\ M W i, IBIjATE )1 7/ 00' '” ‘ 1," 77°05“ K , {I/ 1%- ~ ,_ ., “{/Z~§£IEII.Q.I..IIIeII:: 39CIO’ ——-- . 9 Sandy Spring EXPLANATION Average land slope, in percent 39"‘05’ w 0_3 25—45 W? 8W Drainage basin boundary QM Sub-basin boundary \VI’ICOIOI'I RPQIOHAI Park SCALE 1:24 000 I L 1’? . , , . 9 1M|LE 1 .5 O 1K|LOMETER I-—-I I—I I—-—I I—I I—I . fi INTERIOR—GEOLOGICAL SURVEY, HESTUN, VA,—19777W76315 32:233.“Igf-ksfisafiizz'sifigafimm MAP OF AVERAGE SLOPE IN NORTH BRANCH ROCK CREEK AND NORTHWEST BRANCH I, ERIELIIIIIITIIIIIWIL,US. ANACOSTIA RIVER BASINS, MONTGOMERY COUNTY, MARYLAND {)3 L13 PREPARED IN COOPERATION WITH THE MARYLAND-NATIONAL CzPITAL PARK AND PLANNING COMMISSION (MONTGOMERY COUNTY AND PlINCE GEORGES COUNTY PLANNING BOARDS), THE WASHINGTON SUBURBAN SANITARY COMMISSION, THE DISTRICT OF COLUMBIA DEPARTMENT OF ENVIRONMENTAL UNITED STATES DEPARTMENT OF THE INTERIOR SERVICESWHE MONTGOMERY coumy DEPARTMENT OF ENVIRONMENTAL PROFESSIONAL PAPER 1003 GEOLOGICAL SURVEY PROTECTION, AND THE MARYLAND DEPARTMENT OF NATURE RESOURCES PLATE 3 I I x‘ c , x” \ \ > \. ‘, {\fl) MIL M A R v L A N D M ,9?» . . AI 9 I T A H . ‘ 1 \V‘», v”; Z‘ M“ . I} f 9 ”a ‘\ M MD" I I h in. 7/” “x”? \ '/ L.” \thrgb‘ E f : \ I» f Carnage Park WM f” £4 \ é; \VLV‘; \AMKASHINITVI‘ON m: 2 > \ VIRGINIA Iifi / Pg / 33/ :z Qfl§%) K I" I I ,7 {W '. Sandy Spring EXPLANATION Land use/Land cover Forest land 1‘ 3"?” ' ”WWW, J} Pasture Cropland Urban residential Urban commercial Reservoirs Construction SCALE 1 :24 000 Drainage basin boundary I 1 I6 0 1 MILE 1:! - A ' . I—fi: - . L fl. WARNS.“ mem 1 . Sub-basm boundary 5 0 J KILOMETER i I M" INTERIOR—GEOLOGICAL SURVEY, RESTUN, VA.—19777W75315 gglstivtiloemDiasrksefigoilgfisufisgfiyand Sandysprin91;24,000,1971 LAND USE/LAND COVER IN NORTH BRANCH ROCK CREEK AND NORTHWEST BRANCH ANACOSTIA RIVER BASINS, MONTGOMERY COUNTY, MARYLAND, 1974 PREPARED IN COOPERATION WITH THE MARYLAND—NATIONAL CAPITAL PARK AND PLANNING COMMISSION (MONTGOMERY COUNTY AND PRINCE GEORGES COUNTY PLANNING BOARDS), THE WASHINGTON SUBURBAN SANITARY COMMISSION, THE DISTRICT OF COLUMBIA DEPARTMENT OF ENVIRONMENTAL UNITED STATES DEPARTMENT OF THE INTERIOR SERVICES, THE MONTGOMERY COUNTY DEPARTMENT OF ENVIRONMENTAL GEOLOGICAL SURVEY PROTECTION, AND THE MARYLAND DEPARTMENT OF NATURAL RESOURCES PAPER 1003 PLATE 2 CWT: HEW AREA OF THIS REPORT I \ WAS! IINGTY ‘7 "kl , "i )N / M; m ‘3 i 4’17" r31” mr‘ééi ‘ 4 2%."? , :4 47 M M- v "Wm TI,” ,, ’I‘ H 1 (Wm 7 I, flaw/W” MN m 7’ . v, m u Hurt, M33“ ‘ w v‘w A“ u" 1 WW, ‘a:fi“r£ 4a,, W I r,“ ,, w 5 r I: J 7‘ MT ‘ WI” :4 1" T rw‘ 5-,, fly; ai'wf’f‘r ‘ ' I; m aims: ~71? if; A, a... . . . . . r w a . 4:, a * w :1: ' I . . . ., MM: N a. ”1 Amt-274 4 "I‘mzrvflilgi" may; ‘ . ‘ . Exam :7. . um . aim . I 3 . ’45:: ia .m 7m . 7w . u :3, 7 J O my a2 M ”in: SrJé‘i Raniwfx r w Earl? T 4: Ming” ,4 P «32, . WW" HER“; T337, . ’73.]: 1N?!” figfly «ngzz‘wpgflf. ‘Iltyu‘v ”a flaw”: “ 753% 5447;335:377“ 4:, 7 aw : ‘ ‘ ”$53 $55 ‘4 i" h A u , w o 7 519%”; '11:? ‘ :55 W I 'rl:;41§5¢,4‘_‘£u4{4‘r a a”? my, Aw" ,. a, 745' :57in ‘ (fl: . A” l: 717%}! ‘ib'vfi‘a: 7's” a, ‘7 ”lip ,, _ mm", ”WE n , 7i ”:vffin/“V NWT/n ‘ " W7 7:5“):‘77515'7IZ' ~ 15%;;15 ' ’72,: - ”In umm» MW" "’37? “WAE w e WW ’{u‘mhlmkpéfl Mg 3; 7 “Wuwiv‘flyp” ‘ «:iv'ww ' “"J .I‘Wf; 77'7‘ , EXPLANATION W uggfisu‘mwy. W! :meTh “Vi Land use/Land cover a ”7,0” ' Wm. Forest land Pasture Cropland I , A W1 figfizrgfiné" I «s A» “57% “M- M WW theaxon Regmrmi Park Construcnon I W Q ® W Dramage basm boundary 1 1 a >——4 I——I I. I W Sub-basm boundary 1 KILOMETRE 7.? ’05’ Base from US. Geological Survey mane,0mm,KensmgmnandSandyspmgmm,1971 LAND USE/LAND COVER IN NORTH BRANCH ROCK CREEK AND NORTHWEST BRANCH ANACOSTIA RIVER BASINS, MONTGOMERY COUNTY, MARYLAND, 1966 INTERIOR—GEOLOGICAL SURVEY, RESTON,VA.A19777W75315 V Cooling and Crystallization of Tholeiitic Basalt, 1965 Makaopuhi Lava Lake, Hawaii By THOMAS L. WRIGHT and REGINALD T. OKAMURA GEOLOGICAL SURVEY PROFESSIONAL PAPER 1004 An account of the 4-year history of cooling, crystallization, and difirerentiation of tholeiitic basalt from one of Kilauea’s lava lakes UNITED STATES GOVERNMENTCPRINTING OFFICE, WASHINGTON : 1977 UNITED STATES DEPARTMENT OF THE INTERIOR CECIL D. ANDRUS, Secretary GEOLOGICAL SURVEY V. E. McKelvey, Director Library of Congress Cataloging in Publication Data Wright, Thomas Loewelyn. Cooling and crystallization of tholeiitic basalt, 1965 Makaopuhi LaVa Lake, Hawaii. (Geological Survey Professional Paper 1004) Includes bibliographical references. Supt. of Docs. No.: I 19.16:1004 l. Basalt—Hawaii-Kilauea. I. Okamura, Reginald T., joint author. II. Title. III.Title: Makaopuhi Lava Lake, Hawaii. IV. Series: United States. (Geological Survey. Professional Paper 1004) QE462.B3W73 552’.2 76-608264 For sale by the Superintendent of Documents, US. Government Printing Office Washington, DC. 20402 Stock number 024-001-02990-1 LL K“ CONTENTS Page Page Abstract ____________________________________________________ 1 Observations—Continued Introduction ________________________________________________ 1 Oxygen fugacity measurements __________________________ 18 Acknowledgments __________________________________________ 2 Changes in surface altitude ______________________________ 24 Previous work ______________________________________________ 2 Chemical and petrographic studies ______________________ 25 The eruption of March 5—15, 1965 ____________________________ 2 Major element chemistry ____________________________ 25 Chronology ____________________________________________ 2 Petrography ________________________________________ 31 Initial conditions following formation of a permanent Distribution of olivine __________________________ 36 crust ________________________________________________ 3 Variation in grain size __________________________ 36 Definition of “crust” and ”melt” ______________________________ 4 Core density and vesicle distribution ____________ 36 Methods of study ____________________________________________ 5 Discussion __________________________________________________ 37 Measurement of surface altitude changes ________________ 5 Chemical differentiation in the lava lake __________________ 40 Core drilling ____________________________________________ 5 Gravitative settling of olivine ____________________________ 40 Sampling of melt and casing drill holes below the crust-melt Flow differentiation of 01ivine-augite-plagioclase __________ 47 interface ____________________________________________ 8 Segregation veins ______________________________________ 42 Measurement of temperature ____________________________ 8 Convective cooling in the lava lake ________________________ 44 Additional field studies __________________________________ 9 High-temperature oxidation of basalt ______________________ 45 Observations ______________________________________________ 10 Interpretation of surface altitude changes __________________ 46 Thermal history ____________-__--____________________;__ 10 Summary: Cooling and solidification history of Makaopuhi lava Temperature of the crust-melt interface ______________ 16 lake —————————————————————————————————————————————————————— 46 Cooling of the crust (T<1,070°C) ____________________ 16 References cited ____________________________________________ 47 Cooling of the melt (T>1,070°C) ______________________ 18 ILLUSTRATIONS Page FIGURE 1. Index map showing Kilauea lava lakes ______________________________________________________________________ 2 2. Photograph showing posteruption surface, Makaopuhi lava lake ________________________________________________ 3 3. Index map showing surface features of Makaopuhi lava lake __________________________________________________ 4 4. Schematic cross section, west pit of Makaopuhi crater ________________________________________________________ 4 5. Photographs of tramway used to transport equipment, 1965—67 ________________________________________________ 5 6. Photographs of tramway used to transport equipment, 1968—69 ________________________________________________ 6 7. Photograph of portable drill rig used in 1965—66 ______________________________________________________________ 7 8. Photographs of trailer-mounted drill rig used in 1968—69 ______________________________________________________ 7 9. Photographs of field studies __________________________________________________________________________________ 10 10. Diagram showing isotherms plotted as a function of square root of time and depth ______________________________ 11 11. Diagram showing log f02 plotted against the reciprocal of absolute temperature for profiles obtained in 11 drill holes 18 12. Maps showing contoured altitude changes on the surface of Makaopuhi lava lake ________________________________ 21 13. Maps showing tilting of the surface of Makaopuhi lava lake ____________________________________________________ 24 14. Diagram showing volume rate of subsidence or uplift plotted against square root of time ________________________ 25 l5. MgO variation diagrams ____________________________________________________________________________________ 30 16—23 Diagrams showing: 16. Modal data plotted as percent minerals against percent glass ______________________________________________ 34 17. Percent glass plotted against collection temperature ______________________________________________________ 35 18. Weight percent minerals plotted against temperature ______________________________________________________ '35 19. Schematic distribution of large (>1 mm diam) olivine crystals in drill holes 68—1 and 68—2 __________________ 37 20. Median volumes of augite, plagioclase, ilmenite, and vesicles plotted against depth __________________________ 39 21. Bulk density of drill core plotted against depth ____________________________________________________________ 40 22. Results of differentiation calculations: percent crystals removed plotted against depth ______________________ 41 23. Results of differentiation calculations: composition and amount of olivine, augite, and plagioclase plotted against amount of residual liquid ______________________________________________________________________________ 42 24. Photographs of drill core for 68—1 ____________________________________________________________________________ 50 25. Photographs of drill core for 68—2 ____________________________________________________________________________ 51 26. Photographs of drill core for 69—1 ____________________________________________________________________________ 52 27. Temperature profiles: uncorrected ____________________________________________________________________________ 53 28. Temperature profiles: corrected for contamination ____________________________________________________________ 57 III IV TABLE CONTENTS TABLES Page Contamination effects of thermocouple elements used in Makaopuhi lava lake __________________________________ 1 Depth to isothermal surfaces in the upper crust of Makaopuhi lava lake ________________________________________ 12 Temperature profiles used to construct isotherms in the melt of Makaopuhi lava lake ____________________________ 15 Depth to crust-melt interface determined from drilling, and temperatures measured following drilling in Makaopuhi lava lake ______________________________________________________________________________________________ 17 Melting point data for Ag (961°C), Au (1,063°C), and GeOz (1,115:4°C) ________________________________________ 17 Rainfall record at Makaopuhi crater __________________________________________________________________________ 17 Summary of observations on oxygen fugacity profiles __________________________________________________________ 21 Volume of subsidence/uplift determined from surface altitude changes, Makaopuhi lava lake ____________________ 25 Sample data for chemical analyses shown in tables 10 and 11 __________________________________________________ 26 Major element chemical analyses, Makaopuhi lava lake: samples collected in 1965—66 __________________________ 27 Major element chemical analyses, Makaopuhi lava lake: samples collected in 1968—69 __________________________ 28 Average composition of pumice (MPUMAV) erupted into Makaopuhi lava lake, March, 1965 ____________________ 31 Modal data, Makaopuhi lava lake ____________________________________________________________________________ 32 Chemical mode for MPUMAV, Makaopuhi lava lake __________________________________________________________ 34 Mineral paragenesis, Makaopuhi lava lake —————. _____________________________________________________________ 35 Temperature of melt samples estimated in two different ways__________________-h_hv,__11_§_ _________________ 36 Composition of residual glass from Makaopuhi and Alae lava lakes ____________________________________________ 37 Summary of grain-size measurements for analyzed samples from drill holes 68—1 and 68—2 ______________________ 38 Volumes of crystals and vesicles; samples from drill holes 68—1 and 68—2 ________________________________________ 38 Density of drill core from holes 1—23, 1965—66 and hole 68—2 __________________________________________________ 39 Adjusted mOdal data for differentiated melt samples from drill holes 68—1, 68—2, and 69—1 ______________________ 43 Results of mixing calculations for segregation veins __________________________________________________________ 43 Attitudes of segregation veins, Makaopuhi lava lake __________________________________________________________ 44 Core logs for drill holes 1—24, 1965—66 ________________________________________________________________________ 60 Core logs for drill holes 68—1, 68—2, and 69—1; 1968—69 ________________________________________________________ 65 Temperature profiles measured in drill holes 2—24, 68—1 _________________________________________________ _ _____ 6 6 Oxygen-fugacity profiles, Makaopuhi lava lake ________________________________________________________________ 73 Altitudes obtained by leveling the surface of Makaopuhi lava lake ____________________________________________ 74 Altitudes and differences corrected for ground tilt associated with the December 25, 1965, and October, 1968, East rift eruptions ______________________________________________________________________________________________ 78 COOLING AND CRYSTALLIZATION OF THOLEIITIC BASALT, 1965 MAKAOPUHI LAVA LAKE, HAWAII By THOMAS L. WRIGHT and REGINALD T. OKAMURA ABSTRACT A pond of lava 84 m deep and 800 m wide formed in Makaopuhi crater during an eruption of Kilauea volcano March 5—15, 1965. The tholeiitic lava (MgO about 8 percent, Si02 about 50 percent) was erupted at about 1,190°C. Foundering of the lake-surface crust fol- lowing the eruption reduced the temperature of the upper 10 m of the lake by as much as 40°C. The following studies were conducted be- tween March 1965 and February 1969, when the lake was covered by lava from a subsequent eruption of Kilauea. 1. Twenty four small-diameter (1.5 cm) and three large-diameter (6 cm) holes were drilled to trace the growth of the upper crust of the lake. The 1,070°C isothermal surface separates rigid, partly molten crust (solidus= 980°C) from melt into which probes may be pushed by hand. 2. Open holes and holes cased with stainless steel were used sub- sequently for measurement of temperature, oxygen fugacity, a single measurement of melt viscosity, and collection of samples of melt and gases exsolving from the melt. 3. Elevations of a grid of nails set in the lake surface crust were periodically determined. 4. The specific gravity of drill core was measured, and the core was used in a variety of petrographic and petrochemical studies. This report summarizes all of the quantitative studies relevant to the cooling and crystallization of ponded basaltic lava. Principal re- sults and interpretations of these aspects are as follows: 1. Thermal history: Isotherms in the upper crust migrated down- ward in the lake, first, as a linear function of square root of elapsed time (\/t); then, they were irregularly depressed to greater depths than predicted by this function, the initial change in slope being triggered by a period of heavy rainfall. Isotherms in the melt (1,070—1,130°C) were likewise initially linear with \/t-, but their slopes flattened and began to vary erratically in the period March- December 1966. This behavior is interpreted as being caused by the initiation of convection in the ponded basaltic melt. By late 1968 all . isothermal surfaces were at greater depths than predicted from the Vt—function. 2. Oxygen fugacity (f02): Drill holes showed buffered f02-T pro- files over much of the period, with f 02 values slightly higher than the quartz-fayalite-magnetite buffer. Superimposed on normal profiles were transient high values of f 02 as much as 10—2 atmospheres (atm) between 400—800°C. Core became oxidized soon after exposure to high f02, evidenced by hematitic alteration of mafic. minerals and increased Fe2 Os/FeO ratios in analyzed core. The high oxygen fugacities are tentatively ascribed to deep circulation of oxygen- saturated rain water. 3. Chemistry and petrography: Samples collected during the erup- tion and to a depth of 7.9 m have MgO contents of 7.5—8.5 percent and show olivine-controlled chemical variation. Large olivine crystals are inferred to have settled toward the bottom of the lake. Below 7.9 m, MgO decreases to 6.1 percent at 16.5 m, and samples are dif- ferentiated by removal of olivine, augite, and plagioclase. This proc- ess is interpreted as flow differentiation promoted by convection of the melt. Segregation veins have compositions explainable by low- temperature (1,030°—1,070°C) filtration of liquid from partially mol- ten crust into open-gash fractures. Grain size increases markedly to a maximum at a depth of 14 m, where median-grain values exceed 0.001 mm3 (diam=0.2 mm). Residual glass composition is that of a calc-alkaline rhyolite, and the content of residual glass increases slightly with depth as an apparent function of the amount of differ- entiation. 4. Core density: Core density reaches a maximum at 6.1 m depth (sp gr=2.7 g/cc) and then decreases to 2.5 at 15.2 m. Melt density is low (<25 g/cc) at a depth of 4.6—7.6 in (numerous vesicles) but high (2.8 g/cc) at a depth of 16.5—18.3 m. Evidently the temperature of vesiculation decreases with increasing pressure, and the resultant difference in the amount of an exsolved gas phase in the melt causes density differences that are considered to be the driving force for convection. 5. Surface altitude changes: The surface of the central part of the lake subsided at a decreasing rate (relative to Vt) until the middle of 1967 and was subsequently uplifted through the last measurement date in 1968. This variation in altitude is the resultant of thermal contraction and density change on solidification, the latter critically controlled by the temperature at which vesiculation takes place. INTRODUCTION Many eruptions of Kilauea volcano have left thick ponds of liquid lava, called lava lakes, in pit craters (fig. 1). Three historic lakes were accessible and suffi- ciently long lived to warrant study; these were formed in 1959 (Kilauea Iki), 1963 (Alae), and 1965 (Makao- puhi). Of these only Kilauea Iki lava lake is still ex- posed, the others having been covered by flows erupted in 1969 and later. In addition to the historic lava lakes, one prehistoric lava lake, exposed in the east wall of the west pit of Makaopuhi crater, was studied by Moore and Evans (1967) and Evans and Moore (1968). Data for this lake are important because they represent a complete section through a solidified lava lake whose size and chemical composition are comparable with the 1965 Makaopuhi lava lake. The Kilauea lava lakes have been natural laboratories in which to study the cooling and crystal- lization of tholeiitic basalt magma. The solidification of Makaopuhi lava lake was followed from the time of its formation in March 1965 up to a few weeks before the surface was covered by new lava in February 1969. Methods of study included core drilling of the upper crust, repeated altitude surveys of the lake surface, 1 2 COOLING AND CRYSTALLIZATION OF THOLEIITIC BASALT, 1965 MAKAOPUHI LAVA LAKE, HAWAII l [55’10' KEANA KAKOI LA VA LAKE 19 o 1 2 MILES :- \_|_|___l “a“"a 60 \A Lea >< - 0 1 2 KlLOMETERSsl-L L_l_;l CONTOUR INTERVAL 250 FEET / MAKA OPUHI LA VA LAKE I965 HAWAII l ’L FIGURE 1.—Index map of Kilauea lava lakes (modified from Peck and others, 1966). Craters, lava lakes, and roads are shown as they existed prior to 1968. Eruptions beginning in February, 1969 crossed the Chain of Craters Road in several places. The long Mauna Ulu eruption, centered between Aloi and Alae craters, erupted lava that by the summer of 1974 had completely filled Aloi crater, Alae crater, and the west pit of Makaopuhi crater. HM=Halemaumau crater. Black rectanglein inset indicates area covered by index map. and direct measurement of temperature and oxygen fugacity in holes drilled into the upper crust. Addi- tional field studies of the molten basalt underlying the upper crust included temperature profiles in holes cased with stainless steel, sampling of melt in ceramic and stainless steel tubes emplaced through open drill holes, and a single set of measurements of melt Viscos- ity. Laboratory work included studies of bulk chemical analyses of petrography of drill core and melt samples, and of bulk densities determined on drill core. Cur- rently we are modeling the thermal history of Makaopuhi lava using thermal conductivities meas- ured by Robertson and Peck (1974) and a computer- based finite element analysis. The purpose of this paper is to describe the methods of study, to summarize the data gathered, and to interpret these data in terms of the changing chemical and physical properties of the basalt during solidification. This paper was originally conceived as a progress report, as we had plans for further drilling into the lake at intervals of 2 to 5 years. However, lava erupted in February 1969 crossed the Chain of Craters Road (fig. 1), cutting off access to Makaopuhi crater and spilling into the crater to form a 3 m—deep sheet of aa over the surface of the 1965 lava lake. In May 1969, Kilauea began essentially continuous activity centered at Mauna Ulu (fig. 1), further blocking the crater from access. In early 1972, lava from Mauna Ulu began cas- cading into the west pit of Makaopuhi, which was filled by spring 1973. Thus, the present paper presents the results of an aborted, though nonetheless rather suc- cessful, study of the 1965 Makaopuhi lava lake. ACKNOWLEDGMENTS The Kilauea lava lake studies represent a combined effort of the entire staff of the Hawaiian Volcano Ob— servatory. H.A. Powers was scientist-in-charge during the study of Makaopuhi lava lake and his consistent encouragement is gratefully acknowledged. Personnel who were essential to this study were: Dallas Peck, Richard S. Fiske, Donald A. Swanson, Elliot Endo, George Kojima, Bill Francis, John Forbes, Burton Loucks, Ken Yamashita, and Jeffrey Judd. H. R. Shaw helped in the early drilling of the lake and designed experiments to measure the viscosity of the Makaopuhi melt. P. R. Brett constructed the standards for calibration of temperatures in the drill holes. Continuing collaboration with Shaw and with Rosalind Tuthill Helz have materially improved the "Discussion” section of the paper, although the conclu- sions reached are the responsibility of the authors. PREVIOUS WORK Published studies on Makaopuhi lava lake include an account of the eruption (Wright and others, 1968), preliminary mineralogic data (Wright and Weiblen, 1967; Hakli and Wright, 1967; Evans and Wright, 1972), determination of Viscosity of the Makaopuhi melt (Shaw and others, 1968), determination of oxygen fugacity of gas in the drill holes (Sato and Wright, 1966), composition of gases emitted from the drill holes (Finlaysen and others, 1968), and a study of oxidation during cooling of the lava lake (Grommé and others, 1969). Related studies of other recent Kilauea lava lakes are summarized by Wright, Peck, and Shaw (1976; see especially the annotated bibliography). THE ERUPTION OF MARCH 5—15, 1965 CHRONOLOGY The following account of the March 1965 eruption and the formation of the lava lake is condensed from Wright, Kinoshita, and Peck (1968). The eruption in Makaopuhi Crater occurred in two stages: an initial period of fountaining on March 5 lasted 8 hours and filled the crater to a depth of 50 m1; eruption resumed on March 6 after an 18-hour pause and continued until the end of the eruption on March 15. Crust formed during stage 1 and the early days of ‘All measurements were originally recorded in feet. In particular the core barrels used in drilling and the thermocouples used to measure temperature were marked in feet. In this report we observe the following conventions: I. In the text. all values are given in meters or in feet and meters if the original meas- urement was in feet. 2. In the tables. values are reported in feet. 3. In the figures, scales are given in both English and metric units. 4.f 1’4 THE ERUPTION OF MARCH 5—15, 1965 3 stage 2 piled up in a circumferential pressure ridge, and parts of this ridge persisted as “islands” of crust barely emergent from the surface during the rest of the eruption. During the last 5 days of eruption, the rate of extrusion increased, and the surface of the lake con- sequently remained hot and plastic, so that no perma- nent crust formed during this time. Following the end of eruption, the surface dropped 20 m as lava drained back down the vent, and remnants of the pressure ridge were left standing 5—10 m above the general lava surface. Following drainback, much of the solid crust on the lake was renewed repeatedly by episodes of crustal foundering, the last of which was on March 19. The process of crustal foundering is illustrated and briefly described by Wright, Kinoshita, and Peck (1968, fig. 10d, and p. 3191 ff), and by Shaw, Kistler, and Evernden (1971). The lake surface as it appeared following the erup- tion is shown in figure 2. Surface features are labeled in figure 3. Figure 4 shows the relation of lake levels during the eruption to the thickness of the upper crust at the time the lake was last drilled in 1968—69. INITIAL CONDITIONS FOLLOWING FORMATION OF A PERMANENT CRUST The mode of eruption affects the chemical composi- tion, initial temperature distribution, volatile content, and distribution of previously formed crust in the lake. These factors cannot all be quantitatively evaluated but are important in interpreting the subsequent cool- ing and solidification history. Chemical composition—Samples of pumice and lava collected during the eruption are uniform in chemical composition and in phenocryst content (Wright and FIGURE 2.—Posteruption surface, Makaopuhi lava lake. Vertical photograph. The vent is to the left and the diameter of the lake surface is about 365 m. Drainback rim surrounds the lake. The east wall of the craters is in shadow. Part of the shallower east pit is shown at the far right. Surface features are labeled in figure 3. The relatively smooth surface, marked by polygonal cracks and flow lines, formed following the last crustal foundering episode on March 19, 1965. The elongate area in the right center and some smaller areas near the vent are lower in altitude than the overturn crust surface and were the locus of the latest ooze-outs on March 20. Darker areas are islands of crust; those associated with the low area to the right are remnants of a pressure ridge that was formed during the first eruptive phase on March 5. Those at the left are pieces of the vent rafted out on the lake surface. 4 COOLING AND CRYSTALLIZATION OF THOLEIITIC BASALT, 1965 MAKAOPUHI LAVA LAKE, HAWAII 7gp: "5 a"32 rise—2 $1.22!,2a EXPLANATION Lm surface affect“ In last opium of mm! humming m , Low mu: sin 0! animals following In uplands law I mining 200 rm F—r—v—t—v—H u so MEYERS FIGURE 3.—Index map showing surface features of Makaopuhi lava lake' drawn from figure 2. Level stations are numbered and shown ~as small solid dots (0). Drill holes, labeled in larger type, are shown as open circles (0). Cross section shown in figure 4 is drawn approx- imately along the line connecting level stations 1 and 52. WEST PIT, MAKAOPUNI CRAIER, HAWAII 2600 2‘00 LAKE LEVEL, END OF ERUPTION LAKE LEVEL, AFTER DRAINBACK 2200 CRUST-MELT INTERFACE \ , DECEMBER 1968 DH ;/ 2 633—. lAKE lEVEL, END OF STAGE I ALTITUDE. IN FEEI 200° FEEDER o loo 200 FEET IBOO O 20 40 60 METERS FIGURE 4.—Schematic cross section through the west pit, Makaopuhi crater. Horizontal = vertical scale. Line of section goes through the vent and level station 1 (fig. 3.). The island near station 4 is projected onto the section; its base is conjectured. others, 1968, table 6; this report, table 10). An average composition is given in table 12. Thus we assume that any chemical heterogeneity is imposed by processes taking place after lava ponded in the crater. Temperature distribution—The best estimate of eruption temperature is 1,190°C, if we use the crystal- linity of pumice and apply it to the plot of glass versus temperature (fig. 17; see also discussion in Wright and others, 1968, p. 3198 ff). Cooling of the lava-lake sur- face by radiation augmented by foundering of solidified crust during the eruption reduced temperatures in the upper part of the lake as evidenced by the greater crys- tallinity of samples collected from the rising lake com- pared with the erupted pumice (Wright and others, 1968, table 4). Crustal foundering following the erup- tion further reduced temperatures in the upper part of the lake. Crystallinity of the upper surface of the per- manent crust (sample M— 1—1G) yields a temperature of 1,140°C. H. R. Shaw (written commun., 1975) calcu- lated the total heat loss during crustal foundering and found that it corresponded to a cooling of 50° over the upper 40 feet (12.2 m) of the lake. A temperature pro- file obtained 1 month after the eruption (table 3, fig. 10) showed temperatures of 1,128°—1,136°C between 9 and 21 feet (2.7—6.4 m). Finite-element modeling shows that these temperatures can be matched if the lake is layered with a cooler upper 40—45 feet (12.2—13.7 m) at 1,140°C and a lower part where the maximum temper- ature is 1,190°C. Thus the crystallinity of the surface crust, measured-temperature distribution, and two kinds of theoretical analysis all suggest that, when final solidification began, the upper 40 feet (12.2 m) of the lake was 1,140°C, and the lower part was 1,190°C with a transition zone of unkown but probably small thickness between the two layers. Volatile content—We have no way of quantitatively specifying the volatile content of the magma. In addi- tion to volatiles released during fountaining, signifi- cant degassing accompanied crustal foundering. Thus it is possible that Makaopuhi lava lake was relatively degassed prior to solidification. Distribution of foundered crust—The “islands” of crust left after the eruption have an unknown vertical extent. They were sufficiently rigid to resist movement during crustal foundering, yet did subside isostatically following drainback. No evidence of foundered crust was found in drilling holes 68—1 to depths of 54 feet (16.5 111). However, there is the possibility that unre- melted foundered crust was present elsewhere in the lake. Some of the irregularities in temperature dis- tribution and core density may reflect inhomogeneities traceable to foundered crust, but we have no way of treating these effects quantitatively. DEFINITION OF “CRUST” AND “MELT” The investigations of the lava lake provide a rigor- ous, if empirically derived, definition of the terms “crust” and “melt,” terms that are used throughout this report and in reports on other Kilauea lava lakes. The interface separating crust from melt is a zone across which the rigidity of the basalt changes drastically. A finite hydraulic pressure must be maintained when drilling through crust, but the drill will fall under its own weight with no additional hydraulic pressure once melt is penetrated. The crust-melt interface is found to be an essentially isothermal surface, 1,065°C in Kilauea Iki and Alae lava lakes (Richter and Moore, 1966; Peck and others, 1964, 1966); 1,070°C in Makaopuhi lava lake. The temperature of the crust- melt interface represents the “softening” temperature METHODS OF STUDY 5 of the basalt, above which a steel or ceramic probe may be pushed by hand into the melt. This property makes it possible to take samples and measure temperatures in the melt below the depth to which core-drilling is possible. When drilling within the crust, coolant water is dis- sipated as steam through fractures in the basalt in the vicinity of the drill hole and does not reappear at the top of the drill hole. In contrast, coolant water intro- duced into melt already saturated in H20 collects as a giant “bubble,” and when the drill string is discon- nected the water is expelled as a blast of superheated steam. When the upper crust is less than about 9.1 m thick, the steam returns to the top of the drill hole. At deeper levels, steam generated at the crust-melt inter- face condenses and is lost through fractures before reaching the surface. METHODS OF STUDY During and after the eruption, the surface of the lake was reached by a pig-hunter’s trail that originated about 160m east of the Makaopuhi crater overlook (Wright and others, 1968, fig. 9). Within 1 month after the end of the eruption, an aerial tramway was set up about 62 m east of the overlook and was affixed at its lower end to an island of crust near the center of the lake (fig. 5). The tramway was used to carry supplies for drilling and scientific studies. In 1968 it was re- placed by a larger tramway at the same location (fig. 6). MEASUREMENT OF SURFACE ALTITUDE CHANGES Within 72 hours after the permanent crust was formed, a grid of nails was installed to detect altitude changes of the lake surface during solidification. Later, more stations were added. The final net is shown in figure 3, and the dates when stations were added are given in table 28. Altitude changes were measured using a Zeiss self-leveling pendulum level and 12-foot (3.658 m) Invar rods (fig. 9A). The reference point as- sumed to remain unchanged in altitude was originally a nail driven into a talus block near the trail on the east end of the lake (station 44 in table 28, not shown in fig. 3). Later, station 45, at the base of the drainback was used as the reference, and eventually station 1, on the lake surface, was used. Readings were made to 0.001 foot (0.3 mm) and are precise to about 0.005 foot (1.5 mm). Most observed altitude changes exceed the precision of measurement by a factor of 10 or greater. CORE DRILLING Twenty four holes were drilled between April 19, 1965, and July 22, 1966, using 2.9 cm (SP size)2 2Hole #17 (fig. 3) was drilled with Ex bits U > — — 600 — 1: 3 ' - 200 a g 5 ° ‘ 0 g U U FIGURE 10,—Isotherms plotted as a function of square root of time Vt (days) and depth below lava-lake surface. Isotherms are dashed where interpolated. Data used to construct the figure are given in tables 2—5. Plotted below, at the same square root of time scale, is the cumulative rainfall that fell on the lake surface. The inset (lower left) shows all measurements made at least 2 weeks after drilling for three different isotherms. Single temperature profiles are reproducible within 5°C; most of the scatter represents real temperature differences between drill holes. A single temperature profile in melt on April 19, 1965, is shown by a vertical dashed line. Solid squares ( I ) represent temperatures of melt samples estimated from their crystallinity (fig. 17); these are consistent with the isotherm extrapolations. Temperature fluc- tions. For example, a hole that intersected few frac- tures might be hotter than a hole in which heat was dissipated laterally through fractures. Insofar as pos- sible, the isotherms drawn in figure 10 were con- structed from data in a single drill hole over the period of time that the maximum temperature in the drill tuations in the 1,070—1,140°C isotherms in early 1966 are real. All isotherms lower than 1,130°C can be extrapolated linearly back to a common origin at depth=0, Vtime=0.5 day, but each isotherm begins to deviate from linearity at different times between October 1965 and January 1966. Depression of low temperature isotherms reflects the effect of heavy rainfall; fluctuations of high— temperature isotherms are attributed to melt convection. Isotherms have been dashed to fit the last temperature profile ob- tained in January 1969 but the temperatures in the crust mea- sured at this time may be depressed because of the effects of dril- ling water. See text for further explanation and interpretation of the figure. hole exceeded the isotherm in question. During early cooling, the isothermal surfaces moved downward as a linear function of the square root of time (Vt (days) ), behavior expected of lava cooling mainly by conduction. The isotherms extrapolate to an intersection at Vt (days)=0.5 days, about 17 hours 12 COOLING AND CRYSTALLIZATION OF THOLEIITIC BASALT, 1965 MAKAOPUHI LAVA LAKE, HAWAII TABLE 2.—Depth to isothermal surfaces in the upper crust of Makaopuhi lava lake [Left hand column gives the date, followed, in parentheses, by the square root of time in days after March 19, 1965; the second column gives the drill hole number, followed, in paren- theses, by the number of days separating drilling of the hole from measurement of the reported temperatures; the remaining columns give the depth (in feet) to the isotherm listed in the reading. Depths enclosed by parentheses are extrapolated; others are read directly from the temperature profiles (see table 26, figures 27, 28)] Temperature 1°C) Drill Date hole No. 100 200 300 400 500 600 700 800 900 961 1,000 1,063 4/21/65 1(2) 1.9 2.3 2.8 3.3 3.8 4.4 5.1 5.9 6.7 (5.72) 4/28/65 3(1) 1.9 2.3 2.8 3.3 3.9 4.6 5.5 6.6 7.4 (6.31) 5/5/65 3(7) 2.9 3.3 3.7 4.2 4.8 5.5 6.4 7.5 (6.83) 5/17/65 3(19) 2.9 3.6 4.2 4.9 5.7 6.5 (7.67) 5/24/65 3(26) 1.7 2.3 2.7 3.3 4.1 4.9 5.6 6.6 (8.11) 4(7) 1.3 2.1 2.9 3.7 4.5 5.3 6.3 7.4 8.2 8.9 5/26/65 6(9) 1.0 1.7 2.4 3.1 3.8 4.6 5.5 6.5 7.7 8.5 9.2 (8.23) 6/7/65 4(21) 0.9 1.7 2.6 3.5 4.4 5.4 6.6 8.0 8.9 (8.93) 5(19) .9 1.7 2.5 3.5 4.5 5.6 6.8 8.3 9.3 10.0 6(21) 1.1 1.8 2.6 3.5 4.5 5.6 6.8 8.3 8(14) 1.1 1.8 2.6 3.4 4.4 5.4 6.6 8.0 8.9 9.6 9(12) 1.1 1.8 2.6 3.5 4.5 5.6 6.8 8.1 9.0 97 10(12) 1.3 2.1 2.9 3.8 4.8 5.8 6.9 8.2 9.1 9.8 6/16/65 4(30) 0.8 1.6 2.4 3.3 4.4 5.5 6.8 8.4 (9.42) 5(28) .8 1.5 2.4 3.4 4.5 5.7 7.0 8.5 9.5 6(30) 0.6 1.0 1.6 2.4 3.4 4.5 5.7 7.0 8.5 8(23) .9 1.7 2.5 3.4 4.4 5.6 6.9 8.4 9.5 9(21) .9 1.6 2.5 3.4 4.5 5.7 6.9 8.4 9.4 10(21) .9 1.7 2.6 3.5 4.6 5.8 7.1 8.5 9.6 10.3 6/26/65 6(32) 2 5 3 5 4.6 5 9 7 4 (9.93) 6/30/65 4(44) 1.0 1.8 2.7 3.6 4.7 6.0 (10.14) 8(37) 1.0 1.7 2.5 3.5 4.5 5.7 7.2 8.8 9(35) .9 1.7 2.6 3.6 4.7 6.2 7.6 9.4 10(35) 1.6 2.5 3.6 4.7 6.0 7.4 9.0 10.3 7/21/65 10(56) 2.0 2.8 3.9 5.2 6.6 8.2 10.0 (11.13) 13(40) 1.1 2.0 2.9 4.0 5.2 6.6 8.2 10.0 14(35) 2.0 2.9 4.0 5.2 6.5 8.0 9.8 11.1 17(5) 1.8 2.6 3.4 4.4 5.5 6.8 8.4 10.2 11.5 12.4 14.2 8/10/65 10(76) 1.3 2.1 2.9 4.1 5.5 7.0 8.7 (11.99) 8/17/65 8(85) 0.9 1.8 2.8 4.0 5.5 7.2 9.1 (12.28) 16(33) 1.3 2.2 3.3 4.7 6.2 7.9 9.7 11.9 13.6 20(15) 1.5 2.5 3.7 5.0 6.3 7.8 9.5 11.4 12.8 13.9 OBSERVATIONS 13 TABLE 2.—Depth to isothermal surfaces in the upper crust of Makaopuhi lava lake—Continued Temperature (”C) Drill Date hole N0. 100 200 300 400 500 600 700 800 900 961 1,000 1,063 8/25/65 6(100) 2.0 3.0 4.3 5.9 7.6 (12.60) 17(40) 0.8 1.4 2.4 3.5 4.7 6.1 7.8 9.6 9/23/65 6(129) __ 4.9 6.5 8.2 (13.70) 10(120) 1.5 2.5 3.7 5.0 6.5 8.1 9.9 9/27/65 16(74) 2.5 3.7 5.1 6.6 8.4 10.4 12.8 (13.85) 20(56) 2.7 3.9 5.1 6.6 8.3 10.2 12.4 14.0 10/24/65 17(101) 2.0 3.0 4.1 5.4 7.0 8.8 11.0 (14.82) 21(24) 3.2 4.3 5.6 7.1 8.9 10.8 13.0 14.6 16.0 10/27/65 8(156) 1.0 1.9 2.9 4.1 5.4 7.0 8.8 (14.92) 16(104) 0.6 1.8 2.9 4.1 5.5 7.1 9.0 11.1 13.6 : 11/4/65 10(162) __ 2.6 3.8 5.2 6.9 8.8 (15.16) 20(95) 0.9 1.8 3.0 4.3 5.8 7.3 9.0 11.1 13.5 11/27/65 21(57) 3.2 4.1 5.1 6.1 7.3 8.6 10.1 11.9 14.1 15.8 (15.90) 12/13/65 20(134) 3.4 3.6 4.7 6.1 7.5 9.1 10.8 12.7 (16.38) 22(34) __ 4.0 5.5 7.0 8.7 10.5 12.5 14.8 163 12/20/65 21(80) 4.0 4.6 5.5 6.5 7.8 9.3 10.8 12.7 15.2 (16.61) 1/5/66 22(57) 1.8 2.7 4.1 5.7 7.3 9.1 11 0 13.1 15 6 (17.08) 23(23) 4.8 6.5 8.2 9.9 116 13.6 159 17.5 18.7 1/19/66 23(37) 11.9 14.0 16.5 18.2 19.5 (17.48) 2/1/66 21(123) 4.0 5.5 6.9 8.5 10.3 12.3 14.5 (17.85) 22(84) 2.4 4.0 5.7 7.6 9.6 11.7 14.0 2/3/66 20(186) 1.6 3.2 4.7 6.4 8.1 10.0 12.1 (17.91) 2/18/66 23(15) 173 19.0 (20.2) (22.8) (18.33) 2/25/66 14(254) 3.6 4.5 5.9 7.3 8.9 10.6 (18.51) 16(225) 4.7 6.0 7.4 9.0 10.9 12.9 20(208) 5.3 6.8 8.5 10.5 12.6 21(147) 7.3 8.8 10.6 12.5 14.8 22(108) 6.8 8.5 10.3 12.4 14.8 23(22) (20.6) (23.2) 3/3/66 23(34) 10.9 12.8 15.0 17.7 (19.5) (20.8) (23.6) (18.31) 14 COOLING AND CRYSTALLIZATION OF THOLEIITIC BASALT, 1965 MAKAOPUHI LAVA LAKE, HAWAII TABLE 2.—Depth to isothermal surfaces in the upper crust of Makaopuhi lava lake—Continued Drill Temperature (°C) Date hole No. 100 200 300 400 500 600 700 800 900 961 1,000 1.063 3/15/66 23(41) 18.0 19.6 20.8 23.6 (18.99) 3/28/66 21(178) 9.1 11.0 13.1 15.7 (19.34) 4/4/66 23(61) 11.1 13.1 15.6 18.4 20.3 21.7 24.2 (19.52) 4/11/66 20(253) 5.1 7.0 8.9 11.0 13.4 (19.70) 21(192) 7.1 8.9 11.0 13.4 22(153) 8.4 10.5 12.9 15.6 23(68) 20.5 21.9 24.7 4/19/66 23(76) 20.9 22.2 25.1 (19.90) 4/25/66 23(82) (20.7) (22.0) (24.9) (20.05) 5/6/66 23(93) 20.9 (22.2) (25.0) (20.32) 6/1/66 22(204) 7.1 9.0 11.2 13.7 (20.95) 6/2/66 20(304) 5.3 7.3 9.3 11.5 14.0 (20.97) 24(9) 5.7 7.8 10.0 12.4 14.7 17.1 19.8 21.9 23.3 7/8/66 21(280) 11.8 14.4 17.2 (21.81) 22(241) 7.3 9.2 11.5 14.1 24(45) 2.4 3.6 5.5 7.5 9.5 11.9 14.3 17.2 20.4 22.4 8/15/66 21(318) 10.1 12.1 14.8 17.7 (22.67) 9/13/66 21(347) 10.3 12.5 15.3 18.3 (23.30) 24(112) 10.1 12.5 15.2 18.2 21.4 23.3 25.4 10/13/66 24(142) 18.4 22.0 24.4 26.0 29.2 (23.93) 11/29/66 24(189) 7.2 7.6 8.4 9.7 11.5 13.7 16.3 19.5 23.0 25.5 27.2 30.4 (24.89) 2/2/67 24(254) 7.2 7.9 9.5 11.3 13.3 15.7 18.1 21.0 24.5 27.1 28.8 (26.17) 4/7/67 24(319) 10.0 10.3 12.7 14.7 17.0 19.8 22.8 26.3 28.8 (27.36) 6/22/67 24(395) 8.8 10.8 12.7 14.8 17.0 19.5 22.2 25.0 (28.72) 12/11/68 68-1 20.8 23.9 27.0 30.1 33.2 36.3 39.4 42.4 45.0 47.0 48.2 51.8 (36.90) (21) 1/22/69 68-1 20.6 24.5 27.4 29.8 32.5 35.4 38.2 41.5 45.0 47.3 48.6 52.1 (37.47) (35) OBSERVATIONS 1 5 TABLE 3.—Temperature profiles used to construct isotherms in the melt of Makaopuhi lava lake [A complete record of measured temperature is included in figures 27,28. See text for discussion of thermocouple contamination and the corrections applied as a result of contamination. Numbers in parentheses have been corrected from measured values] \/t_(days) Drill hole Defpth Temperature Date after 3/19/65 No. ( t) (°C) Thermocouple notes‘ 4/21/65 5.72 2 9.1 1,131 5 junction Cr-Al (18 gauge wire). 12.1 1,134 (cased with 15.1 1,128 drill steel) 18.1 1,134 21.1 1,136 4/28/65 6.31 3 9.0 1,110 7/20/65 11.08 17 12.0 995 Single junction Pt-PtRhio (20 gauge (cased with 13.0 1,032 wire). ceramic) 14.0 1,064 14.6 1,076 11/9/65 15.35 22 21.0 1,085 Singlejunction Pt-PtRth (temperature (open stainless steel 21.6 1,092+ still rising when ooze came into sam- sarnpling tube) pler). 12/13/65 16.39 23 20.9 1,072 Single junction Pt-PtRhio (20.9 ft just above crust-melt interface. Tempera- ture rose to 1072°, and levelled off in 6 hours). 2/9/66 18.08 23 36.7 1,150 Single junction Pt-PtRhio. Reading ob- (cased with tained in first 10 minutes prior to stainless steel) contamination. 2/18/66 18.33 23 23.1 (1,070) Single junction Pt-PtRhm. 23.6 (1,080) 24.6 (1,090) Reproduced profile corrected by +17°C 25.9 (1,100) from measured values. 27.3 (1,110) 28.8 (1,118) 32.7 (1,130) 36.0 (1,140) 2/23/66 23 Single junction Pt-PtRth. Erratic pro- 2/28/66 files with apparently high tempera- 3/1/66 ture corrections. Not used. 3/3/66 3/8/66 18.81 23 24.2 (1,070) Single junction Pt-PtRth. Reproduced 24.9 (1,080) profile corrected by +27°C from mea- 25.9 (1,090) sured values. 26.8 (1,100) 28.0 (1,110) 29.0 (1,118) 31.0 (1,130) 33.7 (1,140) 37.0 (1,150) 3/15/66 18.99 23 24.0 1,070 New, single junction Cr—Al, 14 gauge 24.7 1,080 wire. Uncorrected profile. 25.5 1,090 26.5 1,100 28.0 1,110 29.4 1,118 32.5 1,130 37.5 1,140 3/22/66 23 Erratic profile. Not used. 3/28/66 19.33 23 24.4 (1,070) Single junction Cr-Al, 14 gauge wire. 25.0 (1,080) Reproduced profile corrected by 25.7 (1,090) +20°C from measured values. 26.6 (1,100) 27.7 (1,110) 29.0 (1,118) 31.0 (1,130) 34.2 (1,140) 4/4/66 19.52 23 24.7 1,070 New, single junction Cr-Al, 14 gauge, 253 1,080 wire. Some erratic readings. Uncor- 26-5 1,090 rected profile. 27.5 1,100 28.5 1,110 29.3 1,118 30.7 1,130 33.0 1,140 4/11/66 1970 23 25-1 1070 Single junction, Cr-Al, 14 gauge wire. 268) i838 Uncorrected profile. 27.8 1,100 29.0 1,110 30.1 1,118 32.5 1,130 35.6 1,140 16 COOLING AND CRYSTALLIZATION OF THOLEIITIC BASALT, 1965 MAKAOPUHI LAVA LAKE, HAWAII TABLE 3.—Temperature profiles used to construct isotherms in the melt of Makaopuhi lava lake—Continued V “days! Drill hole Depth Temperature Date after 3/19/65 No. (ft) (°C) Thermocouple notes‘ 4/19/66 19.90 23 25.5 1,070 New, single junction Cr-Al, 14 gauge 26.2 1,080 wire. Uncorrected profile. 27.1 1,090 28.0 1,100 290 1,1 10 30.3 1,1 18 32.6 1,130 37 1,140 (extrapolated) 4/25/66 20.05 23 25.2 (1,070) Single junction Cr-Al, 14 gauge wire. 25.9 (1,080) Reproduced profile corrected by 26.5 (1,090) +17°C from measured values. 27.3 (1,100) 28.3 (1,110) 29.2 (1,1 18) 30.6 (1,130) 32.3 (1,140) 5/6/66 20.32 23 25.5 (1,070) Single junction Cr-Al, 14 gauge wire. 26.1 (1,080) Reproduced profile corrected by 26.8 (1,090) +13°C from measured values. 27.8 (1,100) 28.8 (1,110) 29.7 (1,1 18) 30.7 (1,130) 32.5 (1,140) 9/13/66 23.28 24 28.6 1,070 New, two-junction Cr-Al, 14 gauge (cased with 29.4 1,080 wire. Uncorrected profile. stainless steel) 30.0 1,090 31.0 1,100 32.0 1,1 10 332 1,1 18 36.0 1,130 39.0 1,135 10/13/66 23.92 24 29.6 1,070 Two-junction Cr-Al, 14 gauge wire. Un- 30.3 1,080 corrected profile. 31.0 1,090 32.0 1,100 33.1 1,1 10 34.6 1,1 18 37.5 1,130 40.5 1,135 1Several ty es of thermocouples were used in the melt, single junction Pt~PtRhm, single junction Cr-Al, and multijunction Cr.A1. All thermocouples were protected by stainless steel casings and al)l made with uncontami , with repeated use, showed effects on contamination of thermocouple elements above 1,100 C, as described in the text. The data plotted in fig. 10 is taken from profiles nated thermocouples and from profiles with contaminated thermocouple elements to which a constant independent temperature measurement correction based on melting of Ag, Au, and GeOz (table 5). has been applied. The melting experiments indicate that the contamination corrections are essentially linear to at least 1,120 C. after a permanent crust formed on the lake. This 17 - hour displacement reflects the fact that the lake sur- face was hotter than ambient air temperature during the early cooling. Later deviations from linearity of the isotherms (apart from brief effects of coolant water noted immediately after drilling) are ascribed to two factors: 1. Depression of isotherms in the crust following periods of heavy rainfall; 2. Variable nonlinear behavior of isotherms because of nonconductive cooling. These factors are discussed below in connection with the crust-melt cooling regime. TEMPERATURE OF THE CRUST—MELT INTERFACE The depth to the crust—melt interface was deter- mined during drilling (table 4). The temperature of the interface was estimated from measurements made immediately after drilling and by extrapolation back- wards from later temperature profiles obtained in cased drill holes in the melt. (table 3, fig. 10). The best estimate of the temperature of the crust-melt interface is 1,070°: 5°C. The 1,063° isotherm (fig. 10), calibrated by melting of gold wire, is clearly at depths shallower than the crust-melt interface, and thus established a lower limit to the temperature of the interface. The actual points determined from drilling fall Within the envelope approximately defined by the 1,063°—1,090° isotherms (fig. 10 inset). We do not interpret these as reflecting real temperature differences at the interface but rather reflecting the uncertainty in estimating the depth at which the interface was penetrated. These un- certainties become greater at greater depths, and in the 1968 drilling the depth is only known to within 0.6 m. COOLING OF THE CRUST (T<1070°C) The position of isotherms in the crust is a function of both conductive cooling and of the amount of rainfall on the surface. During the first 7 months following the OBSERVATIONS TABLE 4.—Depth to crust- melt interface determined from drilling, and temperatures measured following drilling in Makaopuhi lava lake. \/t (days) Drill Depth to Temperature measured after hole crust-melt following drilling‘ Date 3/19/65 No. interface (it) depth (ft) temperature (°C) 4/19/65 5.57 1 6.8 o H, "Hm H. 5.57 2 6.5 4/21 5.72 2 7.2 Do ,,,,,,,,,,,,,,,, 5.72 1 6.7 1,064 4/22 5.80 1 7.0 Do ,,,,,,,,,,,,,,,,, 5.80 2 7.0 1,055 4/28 6.31 3 8.0 Do , ""1." _ W, 6.31 3 8.25 1,064" 8.0 1,056“ 5/6 6.91 7 8 8 Do ,,,,,,,,,,,,,,,,,, 6.91 7 9.0 1,017 5/7 6.98 7 9.5 1,064 5/17 7.67 6 100 Do ,,,,,,,,,,,,,,,,,,, 7.67 4 9.3 5/19 7.79 5 10.3 5/24 8.11 8 9.9 5/26 8.23 9 10.2 Do ,,,,,,,,,,,,,,,,,, 8.23 10 10.4 6/7 8.93 13 1085 6/11 9.16 13 113 Do ,,,,,,,,, , ________ 9.16 13 11.0 1051*" Do ,,,,,,,,,,,,,,,,,,, 9.16 13 11.75 1068* 6/16 9.42 14 11.25 D0 ,,,,,,,,,,,,,,,,,, 9.42 14 11.9 1,060 7/12 10.72 16 12.9 7/16 1091 16 13.9 Do ,,,,,,,,,,,,,,,,,, 10.91 17 14.2 8/2 11.65 20 14.65 8/17 12.28 21 160 Do ,,,,,,,,,,,,,,,,,, 12.28 21 18.0 1,069“ 8/30 12.81 21 15.7 9/16 13.42 21 16.9 D0 ,,,,,,,,,,,,,,,,,, 13.42 21 19.0 1076* 10/1 14.00 21 17.7 Do ,,,,,,,,,,,,,,,,,, 14.00 21 20.0 1,075* 11/9 15.33 22 197 12/13/65 16.39 23 20.9 ,,,,,,,,,,,,,,,,,, 16.39 23 20.9 1,0727" 1/19/66 17.48 23 22722.5 D0 777777777777777777 17.48 23 20.9 1,072”: 2/3/66 17.91 23 23: 0.5 5/23/66 20.74 24 26: 7/28/66 22.26 24 29.0—29.3 9/13/66 23.30 24 29.0 1,075’?’ 11/18/68 36.59 MP68—1 52:2 12/11/68 36.90 MP68—1 51.0 1,063* 12/18/68 37.00 MP68—2 53:2 1Represents a minimum value for the equilibrium temperature at the stated depth. Starred * values are believed to be within 10° of equilibrium. eruption, the isotherms moved downward as a linear function of the square root of time. This behavior is apparently independent of rainfall (table 6, fig. 10) which was great for 2 months following the eruption and much less for several months thereafter. We have no explanation for the absence of a rainfall effect, al— though it may be related to the position of the 100°C isotherm at the surface of the lake. From mid-October to early November 1965, the isothermal surfaces began to be depressed to greater depths, the lower temperatures showing the effect ear— lier. This change is correlated with heavy rains begin- ning in October and especially evident in November (table 6; base of fig. 10). Then the isotherms showed recovery toward the linear extrapolation of their initial slope between March and September 1966, a dry period. Isotherms at 961°C and higher returned to their original projected slopes. After September 1966, the isotherms plunged abruptly, probably in part due again to increased rainfall and continued to be de- pressed through the last measurement date in June 1967. Two temperature profiles (table 26) were obtained in 17 TABLE 5,—Melting pomt data for Ag (961 °C), Au (1,063°C), and Ge02 (1,115":4°C)l [Runs in Drill hole N0. 23 are 10—15 minutes. Runs for G902 in Drill hole No. 24 are )2 hour] \/t_ (days) Drill after hole Date 3/19/65 No. Sample Results 5/6/66 20.32 23 Ag Melt precisely at 20.9 ft. 5/13/66 20.49 23 Au 25.5 (melt); 25.0 ft (no melt). 5/17/66 20.59 23 Au 25.6 it (melt); 25.5 ft (no melt). Ge02 30.8 ft (completely melted). 5/23/66 20.74 23 G602 30.8 ft 30.3 ft (melt). 6/1/66 20.95 23 Ag 21.85 R (melt); 21.70 3 (no melt). Au 26.2 ft (melt); 26.0 it (no melt). Ge02 30.8 ft (melt with few crystals); 30.2 ft (crystalline). 8/15/66 22.67 24 Au 27.9—29.0 fl. (melt). Ge02 33.0 ft (melt); 32.8—32.2 ft (melt+ progressively more crystals). 8/22/66 22.80 24 Au 28.03 R (melt); 27.89 ft (no melt). Ge02 32.643185 ft (melt+crystals). 9/6/66 23.13 24 Ag 23.8 6. (melt); 23.6 ft (partly melted); 23.4 ft (no melt). Au 28.45 it (melt); 28.35 ft (no melt), GeO2 33.0 ft (melt); 32.2—32.8 it (melt+crystals) 9/13/66 23.28 24 G602 310—3025 if (some melt, mostly crystalline). 10/13/66 23.92 24 Ag 24.5 it (melt); 24.3 it (no melt). Au 29.5 ft (melt); 29.3 ft (no melt). GeO2 34.5 it (melt). 10/31/66 24.29 24 GeOz 33.6—34.6 fl (mostly crystalline). 12/11/68 36.90 MP6&1 Ag 47.5 ft (melt). Au 51.5 ft (melt); 50.5 ft (no melt). lWhen GeO2 was melted concurrently with thermocouple measurements, the best melting temperature was 1,118°C. TABLE 6.—Rainfall record at Makaopuhi crater [Rain gage installed 3/19/65; destroyed 2/22/69] Rainfall Cumulative rainfall Month Year (in) (in) March 1965 4.56 4.56 A ril 28.70 33.26 A ay 27.40 60.66 June 399 64.65 July 3.85 68.50 August 284 71.34 September 5.27 76.61 October 7.63 84.24 November 27.74 111.98 December 10.73 12271 January 1966 6.82 12953 February 953 139.06 March 5.18 144.24 A ril 3.02 147.26 ay 8.27 155.53 June 5.62 161.15 July 867 169.82 August 437 174.19 September 878 182.97 October 1136 194.33 November 27.00 221.33 December 5.95 22728 January 1967 14.22 24150 February 980 25130 March 15.25 266.55 A ril 12.65 279.20 ay 12.14 29134 June 5.42 296.76 July 902 305.78 August 1606 321.84 September 2.65 324.49 October 465 329.14 November 27.63 356.77 December 24.39 381.16 January 1968 29.49 410.65 February 10.58 421.23 March 877 430.00 A ril 20.16 450.16 ay 6.89 45705 June 8.79 465.84 July 640 47224 August 486 47710 September 3.53 48063 October 807 48870 November 10.05 498.75 December 34.09 532.84 January 1969 34.27 56711 February 9.21 576.32 drill hole MP68—1 in December 1968 and January 1969. Temperatures from the later profile are plotted in figure 10 and represent maximum depths for the labeled isotherms. An independent estimate of the depth to the 1,000° isotherm is made from microscopic 18 COOLING AND CRYSTALLIZATION OF THOLEIITIC BASALT, 1965 MAKAOPUHI LAVA LAKE, HAWAII study of samples cored from MP68—1 assuming the sol- idus is the same temperature (980°C) as found higher in the hole (Wright and Weiblen, 1967). This depth is about 1 m shallower than that indicated by the temp- erature profile. We can conclude that either (1) the temperatures are not at equilibrium because of large amounts of water introduced during drilling or (2) that the assumption regarding solidus temperature is wrong, the solidus at these depths being closer to 900° than 980°C. COOLING OF THE MELT (T>1,070°C) Limited parts of isothermal surfaces in the Makaopuhi melt are shown in figure 10 for the follow— ing isotherms: 1,100°C, 1,110°C, 1,118°C (estimated melting point of GeOz), 1,130°C, 1,140°C, and 1,150°C (extrapolated). Most of the data were obtained from profiles in cased holes 23 and 24 (fig. 10, table 3) ob- tained from February to October 1966. Data from a single profile obtained on April 21, 1965 (table 3), is also shown, as are single temperature readings ob- tained during sampling of melt, and also temperatures of melt samples estimated from their crystallinity (Ta- ble 14). Corrections were made for thermocouple con- tamination as discussed in connection with figure 28. The lines in figure 10 are our best estimate of isother- mal surfaces consistent with all of the data. Isothermal surfaces up to about 1,130°C are drawn between the origin for crustal isotherms (Vt=0.5 day at depth 0.0 ft) and the time of the first complete profile obtained in February 1966. These are consistent with single temperatures, either measured or inferred, at times in between. Subsequently the isothermal sur- faces are perturbed, and slopes are generally flatter in the period March to October 1966, when the last com- plete profile was measured. The slope of the 1,070° isotherm also flattened in the period January to Oc- tober 1966. Much later, when hole 68—1 was drilled, the 1070° isotherm was depressed relative to the position predicted from its initial slope. From February to May 1966, the melt isotherms fluctuated erratically in depth. These fluctuations can- not be attributed to errors of measurement and correc- tion for the following reasons: 1. The points defined by melting Ge02 are not co- linear, nor are isotherms defined by uncorrected pro- files alone. 2. The shape of successive temperature profiles changes, and crossovers are not uncommon (figs. 27, 28). The temperature variation in the melt is interpreted in a later section. We emphasize that the depths to isotherms show distinct departures from a linear rela- tion to Vt expected from a conductive cooling model, and over part of the time the change in slope of the melt isotherms is opposite in sign to the change in slope of the crust isotherms. OXYGEN FUGACITY MEASUREMENTS Sato and Wright (1966) reported preliminary results of measurement of oxygen fugacity in drill holes. The data upon which their work was based, as well as those obtained subsequently, are summarized in table 27 and figure 11. Two kinds of measuring devices, called oxygen pro- bes, were used. The first, described by Sato and Wright (1966) and Sato (1971), used a mixture of nickel and nickel oxide (Ni—NiO) packed in a zirconia tube. A temperature profile was determined in each drill hole on the same day that the oxygen probe was used. Emf was measured on a high-impedance electrometer and converted to log f02 using the following formula. log f02 (unknown) = longz (Ni-NiO) — emf (volts) (1) RT/4F where R is the gas constant, F is the Faraday constant, and T is in Kelvins. The reference oxygen fugacity for Ni-NiO is given by the formula (Huebner and Sato, 1970): log f02 (Ni—NiO) = w + 9.36 (2) The Ni—NiO oxygen probe was used with good repro- ducibility through September 1965. Subsequently, problems were encountered that were not fully diag- nosed and never resolved. The emf fluctuated errati- cally during the later measurements, apparently be— cause of an electrical short in the casing of the probe. The fluctuations of emf, although annoying at the time, correspond to f 02 variation of no more than one order of magnitude and do not seriously affect the interpreta- tion of the results. In December 1966, we began to use a gas reference oxygen probe which had a design similar to that de- scribed by Sato and Moore (1973, fig. 2). Oxygen was FIGURE 11,—Log f02 plotted against the reciprocal of absolute temperature for profiles obtained in 11 drill holes. A , DH6. B, DH8. C, DH9. D, DH10. E, DH11. F, DH14, 16, 17. G, DH20. H, DH21, 24. Measurements were made using an oxygen probe like that pictured in figure 9. The date of measurement is given beside each profile. Each profile is represented by a different set of symbols. Most holes show at least one buffered profile with values lying close to and parallel with the experimental buffers quartz- fayalite-magnetite (QFM) and nickel-nickel oxide (Ni-NiO). Most drill holes also show some time period when [‘02 values approached or exceeded that of the hematite-magnetite (H-M) buffer at tem- peratures from 450 to 800°C. See text for further explanation and interpretation. OBSERVATIONS TEMPERATURE (°C) TEMPERATURE (°C) 2400 500 600 700 800 900 1000 1100 400 500 600 700 300 900 1000 noo - I I I I I I I I —2 I I I I I I x’ 9/23/65 —20 - —22 15 H 13 12 11 10 9 8 7 TO‘/T A TEMPERATURE (°CI TEMPERATURE (°C) 400 500 600 700 300 900 1000 1100 400 500 600 700 300 900 1000 1100 ‘2 I I I I I I I I ‘2 I I I I I I I I 20 N N O o “- —12 — “‘~ —|2 - U) o 8’ 6/10/66 —‘ _I 12/20/65 —14 - —IA — — —16 — — —16 -— _. —IB - — —18 — _l ~20 — — _2o — _ —22 _22 | 15 14 13 12 11 10 9 8 7 15 14 13 12 11 10 9 8 7 WW 10‘/T TEMPERATURE (°C) 500 7 800 900 1000 1100 _2 “l” I “I” a” I I I l TEMPERATURE (°C) 400 500 600 700 800 900 1000 1100 —2 I I I I I I I I _4 _ ——6 _ _3 _ —10 -' _|2 _ N O u _14 ._ U) 0 __. ~16 - _13 _ _20 _ _22 DH 14,16,17 COOLING AND CRYSTALLIZATION OF THOLEIITIC BASALT, 1965 MAKAOPUHI LAVA LAKE, HAWAII 500 TEMPERATURE (°C) 600 700 800 900 10001100 l | l | 1 l 6/23/65 6/30/65 TEMPERATURE (°C) 400 500 600 700 800 900 1000 1100 I IIIIIIT, FIGURE 11.—Continued. prevailing in the open drill holes. The last set of mea- surements was obtained in February 1967. used as the reference gas, and reproducible measure- ments were obtained at the low temperatures then OBSERVATIONS The following observations can be summarized from the oxygen fugacity studies (see also Sato and Wright, 1966). 1. Most drill holes showed, at some interval of time, a f02—T profile in which log fO2 varies linearly with the reciprocal of absolute temperature. The position of such normal profiles indicates that the equilibrated basalt-gas system had oxygen fugacities greater than that of the QFM (quartz-fayalite-magnetite) buffer and within one log unit of the Ni-NiO reference buffer. The slopes of such profiles are commonly shallower than those of the buffers; that is, the conditions in the drill hole became relatively more oxidizing as temperature decreased. The position of the profiles changed with time, varying within one order of magnitude of fO2 at the same temperature. The variations are not regular, and although they probably represent real changes in the drill hole conditions, some variation could have been caused by differing probe construction, aging of the Ni-NiO, or other factors related to the techniques of measurement. 2. All drill holes showed at some time a zone of very high fO2 between extreme temperatures of 800° and 400°C. Generally, the highest [‘02 exceeds that of the hematite-magnetite (HM) buffer at the same tempera— ture, but the values of the bracketing temperatures and maximum f02 vary among drill holes and among different times in the same drill hole. Drill hole 11, near the edge of the lake, showed the zone of high fO2 on the first measurement date (June 23, 1965). Drill holes 20 and 21, near the center of the lake, did not show high f02 until 1966. Drill hole 8 showed a small f02 anomaly in September 1965 bracketed by normal profiles in August and October. Drill hole 16, within 15 In of drill hole 8, was not measured in September 1965, but showed normal profiles in August, October, and November. 3. The zone of high [‘02 moved down in the hole with the isothermal surfaces for a limited period of time, but in many holes a normal profile was reestablished. The duration of the high f02 anomalies in each hole is summarized in table 7. TABLE 7.—Summary of observations on oxygen fugacity (f02) profiles. Dates of observation Zone of high f02 Drill hole Distance from No. edge of lake Beginning End First Last observed observed T 1°C) 11 60 it 623/65 114/65 7 9/27/65 760—550 10 120 ft 7/3/65 11/4/65 9/27/65 .. 730—550 9 230 ft 6/30/65 10/27/65 9/23/65 9/23/65 790—<450 8 440 ft 71’3/65 11/27/65 9/27/65 9/27/65 680570 16 440 11 8/7/65 11/27/65 10/25/65'.’ 10/25/65? N750 14 500 ft 7/3/65 (only date measured - no high f02l 6 640 ft 6/28/65 11/27/65 8/25/65 9/23/65 >74(L<450 20 740 (1 8/17/65 6/10/66 6/10/66 7 650‘<500 21 740 fi 1/5/66 2/2/67 12’1/66 ? 700—<400 17 780 11 7/23/65 (only date measured - no high [02) 21 4. The zone of high f02 is definitely correlated with oxidation of the adjacent basalt. Core from drill hole 11 shows hematitic alteration of olivine in the same depth range that showed high [‘02. Fez 03 shows values up to 3.89 weight percent (M11—11, table 10), and (Fe203/ FeO+Fe2 03) to 0.34, in the oxidized zone compared with normal values of 1.3 and 0.11 respectively in un- oxidized core (table 10, 11). Core obtained from drill hole 6 is unoxidized, in agreement with a normal oxy- gen fugacity profile obtained during the first set of measurements (fig. 11). Subsequently, however, drill hole 6 showed a zone of high f02. Drill hole 23 was drilled next to hole 6 to see if the basalt was altered, and hematitic alteration of olivine was found at a depth corresponding to the measured high fO2 in hole 6. RAVE IN FEET/AWIDAVS) SUBSIDENCE UPLIFT < 010 fill-.020 .021- 030 .031— 040 .OAIv 050 7/26/65 In 9/9/65 0517.060 .061—080 833‘!!! ML 100 - Jul-.120 - >120 A FIGURE 12.—Contoured elevation changes on the surface of Makaopuhi lava lake. A, 7/26 to 9/8/65.B, 9/8 to 10/20/65. C, 10/20 to 12/22/65. D, 12/22/65 to 3/7/66. E, 3/7 to 5/18/66. F, 5/18 to 8/9/66. G, 8/9 to 10/31/66. H, 10/31/66 to 1/31/67. I, 1/31 to 5/31/67. J, 5/31 to 10/2/67. K, 10/2/67 to 1/29/68. L, 1/29 to 7/10/68. M, 7/10 to 12/11/ 68. Base map is taken from figure 3 and shows only the outlines of physical features on the lake. Surface contouring was done afier converting the altitude differences for each station (table 28 and 29) to a rate by dividing each difference by the difference in square root of time (days) for each leveling period. This is done to reduce the effect of the changing crustal growth rate and instead em- phasize changes between level maps that result from density con- trast between melt and the crust forming from it. The heavy solid line separates uplift from subsidence. The heavy dashed line enclosing the area within which the volume of subsid- ence (uplift) is calculated (table 8, fig. 14). The contoured data are corrected for tilts associated with East rift eruptions in the periods 12/22/65—3/7/66 and 7/10/68—12/11/68 (see fig. ‘13). The perturbed pattern shown for the period 12/22/65—3/7/66 is probably induced by the tilting; the effects are seen to die out by 8/9/66. The interpretation of the level maps is complicated and is dis- cussed in the text. 22 COOLING AND CRYSTALLIZATION OF THOLEIITIC BASALT, 1965 MAKAOPUHI LAVA LAKE, HAWAII RATE IN FEEV/AVTmAm RATE IN imm‘fiwus. SUDSIDENCE UPLIFT SUBSIDENCE UPLIFY (.010 [:1 ( olo - ammo OlI-.020 :I Oil-.030 o2I—mo [:33 cal—mo 03!,vo E] ctr—.050 ‘onfloso 05km, 9/8/65 1010/20/65 D“, 060 CI 3/7/66 ..1 my“ .oauoao m oaI—oso OSLIOO - oaI—Ioo - ,lOl-JYO B .101420 - > I20 > 120 E RAVE m rm'mfitmvsx HAVE IN fEEY/AVTmAVS) sunsmENcE umn suasmsucs upun < mo I20 > I20 RAVE IN FEEWAWDAVM ms m rssmwmus) sunsmsnc: uvun suasmmcz umn (.010 ( mo MW 0"--°1° on—ozo 1: 9217.030 02,430 03'~°‘° oar-mo E: DAV—.050 04.450 05“-°°° 12/22/65 '° 3/7’“ .0st Doc 8/9/66 vo lam/ea Dbl-.080 D - Del-050 E: OBI-,IOO OBI-loo Imano - 101. I20 > .20 m D ) no G FIGURE 12,—Continued. 1m: IN rsum Wan“) suasmencs < 010 .a11-.ozo 1:] 9214330 [2:] 0317.0“: [I] .0414)» Jul-.060 E .oeI—oao .osl— I00 101—120 >|20 an: In FEET/A v71mvs1 SUISIDENCE < 010 .ommo [:1 .o21-.o:o .0314“) .on—.oso .Us|~.060 .oonmo ,oa1—.uoo .1o1..120 >.120 RAVE IN FEEI/A VTmAvsl SUBSIDENCE < .010 .011 —,oza [2 9217.030 031- om .on 7.050 1151-.060 oolroao 3317.100 .101 ..1 20 >.120 UPLIF'I UVLIFT UPLIFY IOIJI/M to “31/67 1/3I/67 to 5/3/67 5/3l/67 to l0/2/67 OBSERVATIONS SUISIDENCE < .OIO 0 ll -.020 fill-.030 OJI- 0‘0 .04! -.050 GEM—.060 .OM mOEO OBI -. IOO 110]».110 H‘ >120 suasmchs < .010 .o11-mo [: .021— 030 031-040 .041— 050 .051 woo .06! » 090 051400 .101.,120 I >I70 suosmeucs <.olo .on-mo [:1 \ .021—030 Dill-.040 .on—mo .OSLJMO ,ooI—OBO 031—100 J 101—120 >l20 FIGURE 12.—Continued. an: IN FEEI/A VTmAVS) ms IN Fey/n VTmAvs) RATE IN FEEV/AVTmAvsJ UPLIFT umn EM“ UPun “7/2/67001/29/68 ”29/68 to 7/]0/68 7/10/6819 ”Ill/68 23 24 CHANGES IN SURFACE ALTITUDE Altitudes of each leveling station are given in table 28 along with altitude differences for each survey inter- val. Altitude differences are assumed to reflect chang- ing lake conditions, not external causes, such as uplift or subsidence of crater floor, although tilting of the crater floor was detected twice (see below). The altitude data have been converted to rate by dividing the differ- ence in altitude by the difference in Vt in days be- tween each leveling period.5 The reduced data are contoured for each leveling interval in figure 12, be- ginning July 26, 1965, when the full level net was first occupied. Before then, changes (all in a subsidence sense) were larger and more erratic than later, pre- sumably because of early, continued degassing of the molten lava. For example, some stations subsided sev- eral feet in the first survey period (March 24 to April 7, 1965). Leveling data for two periods (December 22, 1965— March 7, 1966 and July 10—December 12, 1968) have been corrected for apparent tilt of the lake surface that took place during eruptions elsewhere on the upper east rift zone. Figure 13 shows the observed elevation changes for inactive stations near the edge of the lake and the best fit tilt vector estimated from these data. Tilt corrections were calculated for each station, and the corrected elevations and differences are shown in table 29. The tilting is assumed to represent a perma- nent deformation of the lake surface, and elevation dif- ferences subsequent to tilting are compared with those of the tilted surface. Tilting associated with the eruption of December 25, 1965, induced a very complicated pattern (fig. 12), perhaps reflecting the “sloshing” of liquid lava beneath a thin crust, which died out over several months. No anomaly was associated with tilting during the Oc- 5The theoretical basis for contouring the leveling data in terms of differences in rate instead Ofelevation is as follows: The general formula for altitude changes as a function of a phase change (density change) on solidification: pC away? 1 , pm dh’/d\/t (3) Where pc :bulk density of crust (including densities of upper and lower crust) at the crust melt interface. pm = bulk density of melt at the crust-melt interface. dL : change in volume of a fixed mass of melt becoming crust per unit time (positive dW sign if the volume of crust is greater than the volume of melt). (1L 2 rate of thickening of crust (upper plus lower) per unit time. dv? We make the following simplifying assumptions: (1) the rate of growth of crust (dh’/d\/t) is constant and (2) the altitude change (dh) is ‘Ax ofthe volume change (dV), and then divide dh by the difference in Vttaken over each leveling period (dWcalculated from the second line from the top of table 28). The crust/melt density ratio should be proportional to 1 minus the rate of altitude change or pc‘ i 1 7 ilh—><(con tant) pm _ th S ‘ On these assumptions a constant ratio of crust to melt density would be reflected by a constant set of dh/d\/t values, at least for stations near the center of the lake. COOLING AND CRYSTALLIZATION OF THOLEIITIC BASALT, 1965 MAKAOPUHI LAVA LAKE, HAWAII tober 1968 eruption, perhaps because the crust had grown much thicker. In order to present more clearly the overall pattern of uplift and subsidence with time, an integrated vol- ume rate of subsidence was calculated using a compu- ter program based on Simpson’s rule (USGS computer lZ/ZZ/GS In 3/7/66 7110/6310 mil/as B FIGURE 13.—Tilting of the Makaopuhi lava lake surface. A, 12/25/65 to 3/7/66. B, 7/10/68 to 12/11/68. Altitude changes (ft) are plotted for edge stations that should have showed no change over the period. Orientation and magnitude (in microradians (ur) ) of the tilt vector are determined at right angles to the approximate azimuth of “0” change. Map base is the same as figure 12. Larger dots represent station locations whose numbers are adjacent. OBSERVATIONS program C628=“Volume of ground swelling” by Pat- rick C. Doherty). A rectangular grid of 500 points, out- lined by a short dashed line in figure 12, was superim- posed on the contour maps, and the interpolated rate data at each grid intersection were used as input to the program. Data obtained from the program are sum- marized in table 8 and figure 14. The calculated vol- umes are meant as a qualitative means of interpreting average changes of the lake surface through time whereas the contour maps show the pattern of changes related to position on the lake surface in a single time interval. The size of the grid is arbitrary and not altogether satisfactory, because in the later leveling periods the effects of cooling from the crater wall were more pro- nounced at the edges of the grid. Relative volumes are consistent for all leveling periods—the absolute values of volume depend on the grid area chosen and may be compared with the total lake volume (after drainback) of about 5.12 X 108 m3 (Wright and others, 1968, fig. 11). The overall pattern of altitude and volume changes after July 26, 1965, is as follows: net subsidence from August 1965 through October 1966, becoming less in magnitude over the next 7 months, and finally chang- ing to uplift in the center of the lake, though subsid- ence continued around the margin after May 1967. This pattern, unpredicted by any simple model of con- stant density change on solidification combined with TABLE 8.—Volume of subsidence/uplift determined from surface al~ titude changes, Makaopuhi lava lake Dates of AVUdays) AV(m3I AV(m-") Cumulative AV(m3) observation after 3’24/65 AVfldays) from July 26, 1965 7/26/65 11.14 to 71853 —1018 71853 9/8/65 12.96 9/8/65 12.96 to ‘1654 71081 73507 10/20/65 14.49 10/20/65 14.49 to 72057 71013 75564 12/22/65 16.52 12/22/65 16.52 to 71720 7808 77284 3/7/66 18.65 3/7/66 18.65 to 71533 7833 78817 5/18/66 20.49 5/18/66 20.49 to —1617 -833 710434 8/9/66 22.43 8/9/66 22.43 to +1437 +807 711871 10/31/66 24.31 10/31/66 24.31 to —934 —511 —12805 1/31/67 26.04 1/31/67 26.04 to 71021 7462 713826 5/31/67 28.25 5/31/67 28.25 to +203 +96 713623 10/2/67 30.36 10/2/67 30.36 to +642 +338 712981 1/29/68 32.26 1/29/68 32.26 to +609 +250 -12371 7/10/68 34.70 7/10/68/ 34.70 to +148 +69 712224 12/11/68 36.85 VT (DAYS) AFVER MARCH I9, 1965 In I5 20 5 :0 35 L A 1 I I I I I I I fr I I I I I lIlIlllIllT IIIIrTnIrI IIranIIIII M M J J A s ONDJ FMAMJJASOND)FMAMJJASON JFMAMJJASO J I965 Woo (957 I965 + 8 I I uPqu suasmsncz —¢oo we or suaSIoENCE on umn IN cualc MEVERS PER MT IN mvs — aoo — F— —Iooo lllllllllll (_ LARGE RAYES or SUISIDENCE _,-— _ DUE Io mumucx, DEGASSING, — ~|200 FIGURE 14.—Volume rate of subsidence or uplift of the lava surface plotted against square root of time. Volumes are calculated by computer for the area enclosed by a dashed line in figure 12. Dif- ferences in the values, for different time periods, are caused by variable crustal growth rate (fig. 10), changes in vesicularity (fig. 21) as a reflection of changes in crust/melt density ratios, and possible irregular behavior of different parts of the lava lake. Further discussion and interpretation are given in the text. thermal contraction, is discussed later in the paper. CHEMICAL AND PETROGRAPHIC STUDIES All samples collected during the eruption were studied in thin section, and many were analyzed chem- ically to define the variation in the erupted magma. Core and melt samples from drill holes 1—24 were studied in thin section, and modes were made of all partially molten samples and selected holocrystalline samples in order to determine the mineral paragenesis as a function of temperature. Selected samples were analyzed chemically. Later, a systematic chemical and petrographic study was made of core from holes 68—1 and 68—2.6 Samples were analyzed at 4-foot intervals to a depth of 22 feet and 2-foot intervals to the crust-melt interface at 54 feet. Modes were made of all samples collected at temperatures of more than 950°C, in order to define the mineral paragenesis at greater depth than could be obtained from drill holes 1—24. Melt samples and segregations were analyzed to detect differences in composition due to differentiation of the initial magma. After chemical and petrographic studies were completed, the drill core was again inspected to de- scribe megascopic changes in mineralogy and texture with depth in the lake. The results of all these studies are summarized in the next few sections. MAJOR ELEMENT CHEMISTRY Chemically analyzed samples (tables 9, 10, 11) may GDrill hole 6&1 was placed next to an island of layered crust (fig. 3) in order to study the interaction of the molten lake with layered crust. The anomalous textures and chemistry of core from this hole are the subject of a separate study, not reported here. Melt from the bottom of the drill hole is. however, considered to be part of a continuous molten zone beneath the center ofthe lake and this is described along with melt obtained from drill holes 6&1 and 68—2. 26 COOLING AND CRYSTALLIZATION OF THOLEIITIC BASALT, 1965 MAKAOPUHI LAVA LAKE, HAWAII TABLE 9.—Sample data for chemical analyses shown in tables 10 and 11 [All samples were analyzed in the US. Geological Survey rock-anal sis laboratory, Denver, Colorado, under the direction of L. C. Peck. Column 1 gives the sample number keyed as follows: M—etc, Sample collected during the eruption; M13—etc. ample collected during 1965—66 from drill hole number 13 (l—cm diameter core); MP68—1—etc, Sample collected durin 1968—69 from drill hole 68—1 (6—cm diameter core). Column 2 describes the type of material. Columns 3—5 give date, depth, and temperature of collection. Columns 6—8 give the la oratory identification. Column 9 gives additional information on the chemistry. Samples are classified as 'olivine-controlled," "contaminated," or, "differentiated," according to their chemistry compared with that of the erupted pumice (see text for further explanation)] Laboratory data Sample Type of Date of Depth of Temperature of No material collection collection (it) collection (°C) Analyst Lab. No. Report No. Comment M—l Pumice March 5, 1965 ,,,,,,,,,,,,, G.O. Riddle D100969 65DC—33 Olivine-controlled M—1A Do ,,,,,,,,,,,,,,,,,,,,, do ,,,,,,,,,,,,, E. L, Munson D101324 66D032 Do. M—20 Do H .March 14, 1965 G. O, Riddle D100972 65DC—33 Do, M—26 Do __H ,,,,,,,,, March 15, 1965 ,,,,,,,,,,,,,,,,,,,,,, do ,,,,,,,, D101012 65DC—60 D0. M1—1G Glassy skin on lava April 19, 1965 Surface E. L. Munson D101352 66DC—32 Do. lake surface M—3 Melt dip d from March 7, 1965 Surface G. O. Riddle D100970 65DC—33 Do, rising ava lake M—8 March 11, 1965 ______________ do ,,,,,,,, D101010 65DC—60 Do. M—12 March 12, 1965 E. L. Munson D101325 66D032 Do. M—18 March 14, 1965 G. O. Riddle D100971 65D033 Do. Mv22 March 15, 1965 HH do ,,,,,,,,,,,, doHHHH' E. L, Munson D101326 66D032 Do. M1—7 April 23, 1965 10,8—11.4 Approx. 1,130 HH do ,,,,,,,,,, D101351 66D032 Fe contamination and reduction of Fe203 by reaction with steel. M21—24 Melt collected in August 30, 1965 22 HH do H ,,,,,,,,,,,, do ,,,,,,,,, D101332 66D032 Do. bottom drill steel - M21—24 D0 ,,,,,,,,,,,,,,,, August 30, 1965 19 (flowed in i H. do ,,,,,,,,,,,,,, do ,,,,,,,, D101331 66D032 Do. top from 22) M21725 Ooze lefi in drill September 16, 1965 15.4—16.05 1,0601,080 HH do ______ H D101333 6613032 Pyroxene-enriched hole after collec- tion of M21724 M21~26 Melt collected in September 16, 1965 25—2675 Approx. 1,130 __,,. do ,,,,,,,, D101327 66D032 Slight Fe contamination sampler ceramic M21~26 Melt collected in HH do ,,,,,,,,,,,, 2526.75 , .H do ............... do ,,,,,,,, D101329 66D032 Do. top stainless steel M21—27 Melt collected in September 27, 1965 29—305 Approx. 1,135 HH do ,,,,,,,, D101328 66DC—32 Slight Fe contamination sampler ceramic M21727 Melt collected in HH do _____________ 29—305 HH do ,,,,,,,,,,,,,, do ,,,,,,,, D101330 66DC—32 Do, top stainless steel M23-21 Melt in bit January 19, 1966 240 Approx, 1,095 G, O. Riddle D101741 67D021 Differentiated by loss ofaugite and plagioclase during collection. M24»3 Melt in core barrel July 28, 1966 29.0—300 1,0901,100 HH do ________ D101740 67D021 D0. M1—6 Drill core April 19, 1965 6.1—7.1 1,0401,100 E. L. Munson D101334 66D032 Olivine-controlled M2r4 D0 ,,,,,,,,,,,,,,, April 19, 1965 6.1~7,1 1,040—1,100 HH do ,,,,,,,, D101335 66DC—32 Do. M13-13 D0 H June 7, 1965 9.17101 980—1025 D101337 66D032 Do. M13—14 Do H 19254080 D101336 66D032 Do. M02 Do H , Ambient D101345 66DC—32 Do. M4—1 Do , , . Ambient D101346 66D032 Do. M09 Do , . . 850—930 D101347 66DC—32 Do. M1010 D0 H . 7. 805—885 D101348 66D032 Do. M11—9 Do , . 6. 560—645 D101340 66D032 Do. M11710 Do H . 7. 645—725 D101341 66D032 Do. M11711 Do . 8. 725—795 D101342 66D032 D0. M11—12 D0 , . 9. 795865 D101343 66D032 Do. M11—14 Do H HH do ,,,,,,,,,,,,, . 1 . 920—980 D101344 66DC—32 D0. M13—11 D0 H June 7, 1965 .1481 845—915 D101339 66DC—32 D01 M1012 Do , ,,,,, do ,,,,,,,,,,,, .1—9.1 910980 D101338 66D032 Do. M21712 Do August 17, 1965 10.9 11.95 890—935 , D101349 66DC—32 Do. M22—10 D0 ,,,,,,,,,,,,,,,,, November 1, 1965 11.0—12.0 780—820 . H , do ,,,,,,,, D101350 66D032 DO. M23—19A Drill Core January 19, 1966 19.4—21.0 l,000»1,060 G. O. Riddle D101739 67D021 Difgerenitialtled; liquid segregated a ter ri ing, 68—2—100 Do ,,,,,,,,,,,,,,,, December 12, 1968 10.0 <100 HH do ......... D103502 75 LA CR 0007 Differentiated Segregation vein 60017.7 Drill core contain- December 16, 1968 177 <100 ,. H do HH, . , . D103503 75 LA CR 0007 Do. ing a small vesicle sheet 6&1v28 Drill core: segrega- November 15, 1968 28.0 320—350 E. L. Munson D102404 69 DC729 Do. tion vein 6&17445 Drill core: segrega- November 18, 1968 445 >900 G. O. Riddle D103504 75 LA CR 0007 Do. tion vein 69—1—555 Drill core: segrega- January 31, 1969 55.5 Not known HH. do ........ D103506 75 LA CR 0007 D0, tion vein 68—1 Drill Core November 6, 1968 3.7 Not accurately‘ E, Engleman D103069 72 D010 Olivine-controlled 68—1—2 D0 ,,,,,,,,,,,,,,,, November 12, 1968 8.0 72 D010 Do. 68—1-3 Do HH do ,,,,,,,,,,,,,, 12.0 D103071 72 D010 D0. 68—1—4 Do 17.5 , D103072 72 DC—10 D0. 68‘145 D0 , 22.1 ,, D103073 72 D010 D0. 68—1—6 D0 . 24.0 D103074 72 D010 D0. 68—1—7 Do . 26.0 , ,D103075 72 D010 D01 68— 1—8 Do 27.9 113103076 72 D010 Do. 68—1—9 Do 30.0 D103077 72 D010 Differentiated 6&2~10 Do _ December 16, 1968 32.0 , D103078 72 D010 Do. 601711 Do . ,, November 15, 1968 33.8 , D103079 72 D010 Do. 601712 D0 November 18, 1968 36.0 D103080 72 D010 D01 68—2—13 D0 . _ December 16, 1968 38.0 D103081 72 DC—10 D0. 68—1—14 Do ,. _ November 18, 1968 39.8 , D103082 72 D010 Do. 602—15 Do ,. December 16, 1968 42.0 D103083 72 D010 D01 68-1716 Do . December 18, 1968 4412 D103084 72 D010 Do. 601717 Do _.H , do ,,,,,,,,,,,,,, 46.1 ,D103085 72 D010 D0. 6&1v47v6 Do H 47.6 D102405 69 D029 Do. 601718 Do H _ . 48.0 D103086 69 D029 D0- 68— 1719 Do ,,,,,, do ,,,,,,,,,,,,, 49.5 D103087 69 D029 D01 68—020 D0 ,,,,,,,,,,,,,,,,, December 18, 1968 51.5 H V D103088 69 D029 D0- 601721 Melt in bit November 18, 1968 54.0 HH do ,,,,,,,, D103089 69 D029 DD- 601—57 "Black sand” November 20, 1968 57.0 G.O, Riddle D103505 75 LA CR 0007 D0. 602—59 Melt in hit December 18, 1968 59.0 E, L. Munson D102406 69 D029 Do. 69—17207 Drill corle January 28, 1969 24.7 Not known G. O. Riddle D103507 75 LA CR 0007 Olivine-controlled veszcu ar 69—1925—6 Drill core HH do ,,,,,,,,,,,,,, 25.6 HH do , _ ,. _,,,,H, .H do ,,,,,,,,, D103508 75 LA CR 0007 Differentiated masswe 69— 1—41—0 Drill cor? January 31, 1969 41.0 HH do ,,,,,,,,,,,,,, do ,H _ ..H D103509 75 LA CR 0007 Olivine1,070°C Sample M1—7 M21-24B M21—24T M21725 M21468 M21—26T M21—27S M21—27T M23—21 M24—3 50.24 49.83 49.59 50.28 50.16 50.14 50.11 50.22 50.47 50.42 13.31 13.29 13.04 13.35 13.51 13.36 13.23 13.34 13.54 13.58 0.51 1.29 1.32 1.15 1.10 1.01 1.32 0.93 1.37 1.54 11.12 10.90 11.16 10.16 10.22 10.39 10.06 10.44 10.33 10.45 7.92 7.81 8.37 8.19 8.12 8.22 8.29 8.24 7.16 6.85 10.83 10.82 10.61 10.85 10.85 10.83 10.76 10.86 10.64 10.55 2.33 2.33 2.30 2.37 2.32 2.30 2.31 2.33 2.45 2.46 0.51 0.51 0.50 0.52 0.53 0.52 0.52 0 53 0 61 0 61 0.04 0.03 0.0 0.02 0.05 0.03 0.20 0.06 0.05 0.06 2.63 2.65 2.61 2.63 2.61 2.62 2.63 2.64 2.93 3.01 0.28 0.28 0.27 0.28 0.28 0.28 0.27 0.28 0.29 0.30 0.21 0.18 0.18 0.17 0.18 0.18 0.17 0.18 0.17 0.18 0.01 0.01 0.02 0.01 0.01 0.01 0.0 0.01 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.01 0.02 F 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.04 0.05 Subtotal 99.94 99.93 99,97 99.98 99.94 99,89 99.87 100.06 100.06 100.08 less 0 0.0 0,0 0.0 0,0 0.0 0,0 0.0 0.0 0.02 0.03 Total 99.94 99.93 99.97 99.98 99.94 99.89 99.87 100.06 100.04 100 05 .9 F6203 .047 .096 096 092 088 081 106 075 106 117 .9 Fean+FeO C. 1965—66 Drill core, T <1,070°C Sample M1v6 M2—4 M3—2 M4—1 M5—9 M10»10 M11—9 M11710 M11A11 M11—12 M11—14 M13—11 M13—12 M13—13 M13—14 M21—12 M22—10 SiO2 ,,,,,,,,,,,,,, 50.44 50.51 50.20 50.19 50.40 50.42 50.40 50.34 50.27 50.36 50.29 50.41 50.39 50.43 50.45 50.40 50.44 A1203 , 13.48 13.45 13.22 13.23 13.49 13.25 13.44 13.34 13.36 13.38 13.26 13.30 13.36 13.25 13.31 13.26 13.31 F6203 , 1.30 1.31 3.50 3.57 1.33 1.20 2.70 3.01 3.89 1.91 1.31 1.35 1.24 1.30 1.29 1.31 1.26 F90 __ , 9.98 9.94 8.04 7.95 9.83 10.03 8.65 8.46 7.60 9.37 10.08 9.99 10.08 10.09 10.04 10.06 10.00 MgO __ , 7.93 7.79 8.43 8.45 8.01 8.25 7.85 7.99 8.14 8.18 8.30 8.00 8.08 8.17 8.06 8.10 8.09 CaO __ . 10.92 10.97 10.76 10.70 10.99 10.87 10.94 10.88 10.85 10.94 10.80 10.87 10.86 10.80 10.91 10.81 10.88 N820" . 2.33 2.37 2.07 2.12 2.33 2.33 2.33 2.32 2.32 2.32 2.31 2.34 2.38 2.33 2.36 2.37 2.36 K20 7. _ 0.52 0.53 0.50 0.52 0.51 0.52 0.52 0.52 0.51 0.51 0.52 0.53 0.52 0.52 0.52 0.52 0.52 EEO __ 0.01 0.03 0.11 0.12 0.06 0.04 0.03 0.03 0.02 0.03 0.02 0.03 0.04 0.01 0.02 0.06 0.04 Ti02 __ , 2.62 2.68 2.60 2.62 2.63 2.60 2.65 2.68 2.61 2.60 2.65 2.69 2.67 2.67 2.63 2.68 2.64 P205 7- 0.28 0.28 0.28 0.28 0.27 0.28 0.28 0.28 0.28 0.27 0.28 0.29 0.28 0.28 0.28 0.28 0.28 MnO ,. A 0.18 0.17 0.17 0.17 0.17 0.18 0.17 0.17 0.17 0.17 0.17 0.18 0.18 0.17 0.17 0.17 0.17 C02 0.01 0.01 0.01 0.01 0.01 0.01 0.01 0.01 0.01 0.0 0.01 0.01 0.01 0.01 0.01 0.01 0.01 Cl __ . 0,0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0,0 0,0 0,0 F W" 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 Subtotal 99.98 100.04 99.89 99.93 100.03 99.98 99.97 100.03 100.03 100.04 100.00 99.99 100.09 100.03 100.05 100.03 100.00 less 0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 Total 99.98 100.04 99.89 99.93 100.03 99.98 99.97 100,03 100.03 100.04 100.00 99.99 100,09 100.03 100.05 100.03 100.00 .9 F920; .105 .106 .282 .288 .109 .097 .219 .243 .315 155 105 108 100 104 104 105 102 .9 Fe203+FeO D. Segregation veins Sample M2ZL19A 68—2—10 68—1—177 68—1—28 68—1—44.5 69—1—5535 49.57 50.77 50.17 50.07 50.96 52.66 13.05 12.27 13.57 12.05 13.41 12.29 1.90 4.26 1.26 2.41 1.83 1.85 12.69 10.45 9.92 12.92 10.90 13.14 4.60 4.23 7.99 4.30 5.55 3.26 8.70 8.47 10.95 8.49 9.77 7.59 2.91 2.75 2.27 2.73 2.69 3.09 0.88 1.11 0.56 1.02 0.80 1.38 28 COOLING AND CRYSTALLIZATION OF THOLEIITIC BASALT, 1965 MAKAOPUHI LAVA LAKE, HAWAII TABLE 10.—Major element chemical analyses, Makaopuhi lava lake samples collected in 1965—66—Continued D. Segregation veins—Continued Sample M23—19A 6&210 68—1—17.7 68—1—28 68—17445 69—1—555 0.01 0.34 0.06 0.0 0.05 0.16 4.84 4.49 2.78 5.26 3.57 3.49 0.43 0.52 0.25 0.52 0.39 0.70 0.21 0.20 0.16 0.21 0.18 0.20 0.0 0.0 0.01 0.01 0.01 0.03 . 0.02 0.02 0.01 0.02 0.02 0.04 1‘ H. ..... 0.06 0.08 0.04 0.06 0.06 0.10 Subtotal 99.87 99.96 100.00 100.07 100.19 9998 less 0 0.03 0.04 0.02 0.03 0.03 0.05 Total 99.84 99.92 99.98 100.04 100.16 99.93 .9 F920; .112 .268 .102 .144 .131 .112 .9 Fean+FeO TABLE 11.—Major element chemical analyses, Makaopuhi lava lake: samples collected in 1968—69 68—171 68-1—2 68—1—3 6&1v4 68—1—5 68—1—6 6&1—7 6&1—8 68—1—9 68—2—10 68—1—11 68—1—12 68—2—13 50.27 50.21 50.18 50.31 50.23 50.26 50.24 50.32 50.43 50.34 50.50 50.48 50.72 13.59 13.38 13.45 13.51 13.52 13.44 13.46 13.43 13.53 13.62 13.65 13.55 13.53 1.56 1.42 1.33 1.24 1.25 1.25 1.20 1.21 1.24 1.32 1.32 1.22 1.43 9.67 9.88 9.92 9.99 9.99 9.99 10.08 10.08 10.26 10.01 10.24 10.34 10.17 7.72 8.21 8.12 7.99 8.10 8.11 8.10 8.02 7.42 7.78 6.99 7.24 6.74 10.96 10.78 10.90 10.92 10.90 10.90 10.90 10.90 10.82 10.88 10.79 10.78 10.70 2.33 2.30 2.31 2.31 2.30 2.31 2.30 2.31 2.39 2.37 2.46 2.41 2.43 0.53 0.53 0.52 0.52 0.54 0.53 0.53 0.53 0.55 0.55 0.57 0.57 0.60 0.19 0.14 0.13 0.04 0.05 0.02 0.01 0.0 0.01 0.01 0.01 0.01 0.02 2.71 2.67 2.68 2.66 2.68 2.69 2.70 2.69 2.82 2.69 2.93 2.89 3.08 0.26 0.27 0.26 0.26 0.26 0.26 0.27 0.27 0.29 0.28 0.29 0.28 0.30 0.17 0.17 0.17 0.17 0.17 0.18 0.18 0.18 0.18 0.17 0.18 0.18 0.18 0.01 0.01 0.01 0.01 0.0 0.0 0.0 0.0 0.01 0.0 0.0 0.0 0.0 0.01 0.01 0.01 0.01 0.01 0.01 0.01 0.01 0.02 0.01 0.01 0.01 0.01 0.04 0.04 0.04 0.03 0.03 0.03 0.03 0.04 0.04 0.04 0.04 0.04 0.04 100.04 100.00 100.03 99.97 100.03 99.98 100.01 99.99 100.01 100.07 99.98 100.00 99.95 0.02 0.02 0.02 0.01 0.01 0.01 0.01 0 02 0.02 0.02 0.02 0.02 0.02 Total ,,,,,,,,,,,,,,,,,, 100.02 99.98 100.01 99.96 100.02 99.97 100.00 99.97 99.99 100.05 99.96 99.98 99.93 L .127 .115 .108 .101 .102 .102 .097 .098 .098 .106 .104 .096 .113 .9 Fe203+FeO Sample 68—1—14 68—2-15 68—1—16 68-1-17 684—47 68—1—18 68—1—19 68—2—20 68—1—21 68—1—57 68—2—59 6914247 6917256 69—14410 69—1—42.0 69—1—22 50.65 50.58 50.54 50.65 50.59 50.60 50.79 50.75 50.71 50.62 50.58 50.32 50.39 49.98 50.16 50.90 13.56 13.45 13.30 13.45 13.65 13.29 13.14 13.35 13.35 13.41 13.41 13.73 13.69 13.77 13.60 12.97 1.32 1.51 1.47 1.49 1.33 1.47 1.49 1.51 1.59 1.65 1.45 2.79 2.23 4.93 1.51 1.65 10.18 10.31 10.42 10.41 10.55 10.42 10.71 10.53 10.67 10.63 10.51 8.42 9.18 6.20 9.80 11.70 6.96 6.82 6.74 6.55 6.62 6.71 6.47 6.39 6.13 6.29 6.83 7.66 7.33 7.67 8.00 5.18 10.81 10.61 10.56 10.45 10.47 10.53 10.19 10.33 10.19 10.25 10.44 11.20 10.95 11.14 10.96 9.38 2.46 2.46 2.46 2.50 2.50 2.45 2.51 2.54 2.58 2.51 2.48 2.24 2,40 2.19 2.30 2.73 0.58 0.59 0.60 0.63 0.60 0.60 0.67 0.66 0.69 0.71 0.62 0.54 0.62 0.48 0.54 0.80 0.01 0.0 0.0 0.02 0.07 0.03 0.06 0.0 0.07 0.10 0.09 0.04 0.06 0.49 0.04 0.12 3.00 3.10 3.26 3.24 3.14 3.27 3.33 3.31 3.39 3.30 3,09 2,60 2.77 2.49 2.68 3.89 0.29 0.31 0.31 0.32 0.32 0.30 0.34 0.32 0.36 0.34 0.33 0.25 0.29 0.24 0.24 0.41 0.18 0.19 0.19 0.19 0.18 0.19 0.19 0.19 0.18 0.18 0.18 0.16 0.17 0.16 0.17 0.20 0.0 0.0 0.0 0.0 0.02 0 01 0.0 0.0 0.0 0.02 0.02 0.01 0.01 0.02 0.02 0.0 0.01 0.01 0.01 0.02 0.02 0 02 0.02 0.02 0.02 0.02 0.02 0.02 0.02 0.11 0.01 0.03 0.04 0.04 0.04 0.04 0.05 0 04 0.05 0.05 0.05 0.05 0.05 0.05 0.04 0.23 0.04 0.06 100.05 99.98 99.90 99.96 100.11 99.93 99.96 99.95 99.98 100.08 100.10 100.03 100.15 100.10 100.05 100.02 0.02 0 02 0.02 0.02 0.03 0.02 0.03 0.03 0.03 0.03 0.03 0.03 0.02 0.12 0.02 0.03 Total __________ 100.03 99.96 99.88 99.94 100.08 99.91 99.93 99.92 99.95 100.05 100.07 100.00 100.13 99.98 100.03 99.99 11L .105 .117 .112 .114 .102 .112 .111 .114 .118 .123 .110 .230 .180 .417 .114 .122 .9 Fe203+FeO be subdivided into the following categories: | 3. Melt samples (T>1,070°C) collected in drill holes 1. Samples collected during the eruption 4. Segregations collected either as veins (solidified a. pumice or partially molten) found during drilling or as b. melt samples dipped from the rising lava melt that flowed into open drill holes and was lake subsequently drilled out. 2. Drill core Most samples were collected before September 1966 a. Subsolidus: T <1.000°C before drilling at depths of less than 4.6 m (core) or less than 7.6 m b. partially molten: 1,000°<0.3 mm grid) Olivine ________________ 4.4 4.0 3.7 3.1 5.3 4.9 4.4 6.8 Pyroxene ,,,,,,,,,,,,,, 15.4 16.4 13.6 6.0 6.2 7.5 12.3 4.7 Plagioclase ______________ 8.3 10.1 6.6 0.5 0.4 0.4 4.1 0.3 Fe—Ti oxide ____________ 1111 1111 1111 1111 1111 1111 ____ _1__ Glass __________________ 60.9 50.8 71.1 89.1 86.4 81.2 75.5 79.6 QuenCh ________________ 11.0 18.7 5.1 1.4 1.7 6.0 3.7 8.6 B. Melt samples collected through Drill holes (sample data in table 10) M21—21 M21724 M21726 M21727 Sample M2013 Bottom Top Bottom Middle Top M21725 Sampler Bottom Top Sampler Bottom Top Depth (ft) __________ 20—23 17—18 19—22 15.4—16.1 25—2675 29—30.5 T(°C) ______________ 1,130—1,135 1,075—1,105 1,110—1,125 1,065—1,075 ~1,130 ~1,135 Number of points 1,229 678 1,284 2,456 1,500 1,907 1,478 1,000 2,000 1,500 1,862 2,000 2,000 (0.3X0.3 mm grid) Olivine ____________ 3.0 3.7 4.5 3.0 2.9 3.4 3.4 5.0 2.8 3.3 3.3 4.4 4.5 Pyroxene __________ 16.1 9.5 8.0 12.2 7.2 8.9 62.1 17.7 13.6 11.8 13.4 15.1 11.5 Plagioclase ________ 10.7 6.7 8.6 9.1 6.2 8.8 19.8 9.0 7 9 10 0 7.1 9.9 10.2 Fe—Ti oxide ________ 1111 1111 -111 1111 1111 1.7* 15.2 1111 1111 1111 11-1 __1_ 1111 Glass 11111111111111 6.2 1111 0.2 6.2 2.8 .11. ___1 56.8 58.8 11__ 60.6 43.0 37.1 Quench 111111111111 64.0 80.4 78.7 69.3 81.0 77.2 1.1 11.5 17.0 74.9 15.6 27.6 36.7 B. Melt samples collected through drill holes—(Continued) M22—18 M2&24 Sample M22—15 Bottom Center Top M23—21 Bottom Top M2$25 M2492 M24—3 Depth (ft) ____________ 21 21—23 24 24 24 27—29 29—30 T(°C) ,,,,,,,,,,,,,,,,,, 1,095 1,095_1,110 1,095 1,090 1.06(H.085 1.085—1.100 Number of points 1,004 1,400 1,633 1,500 984 1,235 1,720 1,246 953 800 (0.3><0.3 mm grid) Olivine 1111111111111111 4.9 2.6 2.2 1.9 1.7 1.8 2.0 2.1 0 .4 Pyroxene ______________ 22.2 18.3 13.8 19.7 12.2 8.2 7.7 7.4 24.9 15.5 Plagioclase ____________ 13.3 11.9 11.7 15.4 6.3 6.9 7.3 8.7 23.6 16.0 Fe—Ti oxide 111111111111 05* 1111 1111 1111 1111 1111 1111 __1_ 1111 1111 Glass 1111111111111111 1--- 64.5 65.0 7.5 72.5 82 0 78 2 77.6 4.2 3 9 Quench ______________ 59.1 2.7 7.3 55.6 7.3 1.1 4 8 4.2 47.3 64 2 C. Drill core, partially molten 1,070 >T >980° Sample M1—6 M2r4 M5—13 M777 M913 M10—12 M1(L13 Depth (ft) ———————————————— 6.1—7.1 6.1—7.1 10.1—11.0 7.9—8.9 10—10.5 8.0-9.0 9.0—10.0 T(°C) ———————————————————— 1,040—1,100 1,04 1,100 LOGO—1,070 1,030—1,085 1,055—1,070 955—1,010 1,010—1050 Number of points 1,500 1,404 1,500 1,848 2,000 1,281 1,500 (0.3x0.3 mm grid) Olivine __________________ 1.4 2.7 1.6 4.2 2.3 1.9 1.0 Pyroxene ________________ 35.7 33.6 45.4 32.3 40.4 52.0 49.5 Plagioclase ______________ 20.1 21.6 29.4 19.6 23.5 31.4 28.3 Fe-Ti oxide ______________ 1.1 .5 5.5 1___ 3.7 11.4 6.7 Glass ____________________ 38.4 37.9 18.1 34.9 24.2 3.3 15.4 2.3 3.9 1111 9.0 5.9 1111 1111 Quench 111111111111111111 OBSERVATIONS TABLE 13.—Modal data, Makaopuhi lava lake—Continued 33 C. Drill core. Partially molten—Continued Sample M10—14 M11-15 8 MIL’16 C M1a12 M13—13 M13—14 M21_17 Depth (ft) ,,,,,,,,,, 10.0—10.6 11.0—11.9 11.9—12.9 8.1—9.1 9.1—10.1 10.1—11.1 15.9—16.9 T°C) ______________ 1,050—1,070 980—1,030 1,025—1,070 915—980 980—1,025 1,025—1,080 1,065—1,085 Number of points 4,000 1,254 2,000 2,000 1,500 1,500 2,000 2,000 1,436 (0.3><0.3 mm grid) Olivine ____________ 2.8 2.8 2.2 2.9 1.8 2. 3 2 6 2. 4 1.6 Pyroxene __________ 37.7 40.8 39. 4 49.1 46.0 51.1 50. 3 42. 3 29.7 Plagioclase ________ 25.4 24.0 25.0 28.8 29.6 30. 7 30. 7 26.3 19.6 Fe—Ti oxide ________ 3.3 3.7 3. 8 9.1 4.6 11.3 8.3 4.4 ____ Glass ______________ 30.8 28.7 29. 6 10.1 17.1 3.5 7.2 24.3 29.0 Quench ____________ ,1“ A,“ ____ 11,, 0.9 1.1 1.0 0.3 20.2 C. Drill core. Partially molten—Continued M2242 M2&15 M23—17 M2348 Sample Bottom Middle Top M22—13 Bottom Top M2:L16 Bottom Top Bottom Top Depth (ft) __________ 16.0—17.0 18.0—19.0 15.9—16.9 16.9—17.9 17.9—18.65 19.7—20.9 T(°C) ______________ 975—1,010 1,035—1,060 950—975 975—990 990—1,005 LOGO—1,085 Number of points 2,000 2,000 2,500 1,404 1,424 1,000 1,500 1,500 1,413 1,424 1,000 (0.3 X 0.3 mm grid) Olivine ____________ 1.5 2.2 2.5 1.4 0.9 1.4 1.2 0.7 1.2 1.8 1.4 Pyroxene ,,,,,,,,,, 48.1 51.8 47.8 46.1 54.3 53.8 50.0 48.6 51.5 34.4 31.5 Plagioclase ________ 31.7 30.0 34.2 25.6 28.5 31.0 27.1 32.6 31.7 24.1 26.8 Fe-Ti oxide ________ 6.0 5.5 5.2 3.7 11.9 9.3 14.3 4.4 6.7 1.4 ? Glass 11111111111111 12.2 10.6 10.1 16.0 3.5 3.3 2.1 13.7 9.9 38.3 ____ Quench 111111111111 .6 ____ .1 7.2 1.1 1.1 5.3 ____ ____ A--- 40.3 D. Drill core, subsolidus, T<980°C Sample MELI M9—2 M9—3 M9—4 M946 M9—7 M9—8 M9—10 M9—11 M2—1 M771 Depth (ft) __________ 0—0.2 0.2—0.4 0.4—0.65 0.65—0.8 1.0—2.0 2.0—3.0 3.0—4.0 5.0—6.0 6.0—7.0 0—1.0 0—1.0 T(°C) 11111111111111 <100° 100—140 140—190 190—210 240—385 385—510 510—625 720—805 805—885 0—310 30—100 Number of points 1,500 820 1,500 1,500 1,500 1,000 1,500 1,500 1,185 1,343 1,000 (0.3><0.3 mm grid) Olivine ____________ 3.0 0.7 3.0 2.4 2.1 1. 7 1.1 2. 4 1.9 3.0 4.7 Pyroxene __________ 29.4 38.8 42.0 45.1 46.1 46. 6 50.4 51.4 54.4 34.7 25.6 Plagioclase ........ 9.4 15.6 16.9 18.6 22.2 28. 3 25.1 24.0 27.8 24.1 15.4 Fe-Ti oxide ________ 1--- 1.2 2.4 5.1 8.6 9.7 12.0 13. 9 11.1 8.6 2 8 Glass ______________ ____ ____ ",1 ____ ____ 1.5 0.9 2.0 2.5 29.6 ,___ Quench ____________ 58.2 43.7 35.7 28.8 21 0 12.2 10.5 6.3 2.3 ____ 51 6 D. Drill core, subsolidus—Continued Sample M746 Mllkll M11713 Mil—14 M13—11 M22—10 M22r11 M23—6 M23—9 Depth (ft) 7777777777 4.9—5.45 7.0—8.0 9.0—10.0 10.0—11.0 7.1—8.1 11.0—12.0 14.0—15.0 4.9—5.9 8.9—9.9 T(°C) 11111111111111 725—800 885—955 865—920 920—980 845—915 780—820 900—935 360—425 600—670 Number of points 1,000 1,500 1,000 2,000 2,000 1,732 1,791 1,500 1,500 (0.3><0.3 mm grid) Olivine ____________ 1.8 2.2 2.5 2. 0 2.0 0. 7 1.1 1. 2 1.9 Pyroxene ____________ 52.5 45.9 48.5 52. 6 52.8 51.6 50.6 53. 4 50.4 Plagioclase __________ 21.1 28.4 31.4 30. 8 30.6 33. 8 32.6 24.8 29.0 Fe—Ti oxide ,,,,,,,, 15.0 12.8 11.6 10 4 9.3 7.9 9.0 13.9 14.3 Glass ______________ 1"- ____ 2.7 3.3 2.7 5.0 5.6 2.3 1.7 Quench ____________ 9.4 10.7 3.3 0.9 2.6 1.0 1.3 3.8 2.7 (sp gr=5.13 R.I—5.41)9. The results of such a construc- tion are shown in figures 16, 17, and 18. The S-shaped curves of figures 17 and 18 show that the crystalliza- tion rate is fastest at temperatures close to the crust- "R. T. Okamura, unpub. data, Hawaiian glasses. melt interface, far exceeding the rates at near-liquidus or near-solidus temperatures. The order of crystallization of minerals and their temperatures of first appearance are summarized in 34 COOLING AND CRYSTALLIZATION OF THOLEIITIC BASALT, 1965 MAKAOPUHI LAVA LAKE, HAWAII TABLE 13.—Modal data, Makaopuhi lava lake—Continued 68—1, 68—2, 69—1 Sample 69—1—22 68—2—59 6&1721 68-2520 68—1— 19 6&1418 6&1—17 68—1—16 68—2—15 68—1—14 68—1—7 Dgpth (ft) ———————————— 60—66 59 54 51.5 49.5 48.0 46.1 44.2 42.0 39.8 26.0 T( C) ———————— , ---------- 1,10a1,105 1,100 1,082 1,060 1,055 1,043 1,029 1,011 <1,000 <1,000 <1,000 Number of pomts , 1,616 956 1,996 1,673 1,816 1,828 2,000 1,376 1 901 1,859 1 934 (0.4><0.5 mm grid) ’ ’ Ohvme 7777777777777777 0 0.3 1.4 0.7 0.5 0.8 0.7 0.1 0.2 0.8 2.1 Pyrqxene —————————————— 0 7.3 9.5 15.7 20.4 25.3 31.1 34.2 32.5 36.5 43.1 Plagioclase ———————————— 0.2 5.8 9.7 21.1 24.3 35.9 41.3 43.7 48.3 43.0 41.9 Fe-T1 ox1de 444444444444 0 0 0 1.3 1.8 4.1 6.6 5.8 6.2 7.3 6.5 Glass ________________ 99.8 83.6 76.9 50.0 27.2 14.4 11.8 9.3 7.3 3.0 Quench ______________ 0 3.0 2.5 11.2 53.0 6.7 5.9 4.4 3.5 5.1 3.4 Ves1cles Volume percent ______ 0 0.1 0.2 3.7 4.9 13.2 12.8 12.8 12.3 10.5 8.1 Size(mm)(median) _- 1,2, .15><.15 .35><.35 .3><.3 S.3><.3 .5>< 1X1.5 6x.85 .5><7 .5><5 .5><.5 (max1mum)- (.7><.7) (1><.5) (1.2><1.2)(1.5><2) (1.5><2) (1x2) (1x1) (1x2) 'Glass skin, surface of lava lake near Drill hole #1. 2Smallyflow from original line of vents in Makaopuhi crater. Erupted March 5, 1965. TABLE 14.—Chemical mode for MPUMAV, Makaopuhi lava lake [Calculated by methods described in Wright and Doherty (1970). Mineral compositions are not all unique but are chosen to be reasonable for the bulk rock composition. MPUMAViaverage composition of erupted pumice, table 12] Corrected to As is 2 percent olivine‘ Olivine __________________________________ 6.5 2.0 (F070) Augite __________________________________ 33.4 (EmsFSzaWOM) 41.7 Pigeonite ________________________________ 6.4 (EIISIFSSQWOIO) Plagioclase ______________________________ 42.5 44.5 (A1157) Ilmemte __________________________________ 4.2 5.5 Titanomagnetite __________________________ 1.0 Apatite __________________________________ 0.6 6 3 Glass ____________________________________ 5.4 ‘These values of olivine, total pyroxene, pla 'oclase, total Fe«Ti oxide, and glass + apatite are used as a reference to correct optical mo es (See fig. 16 and text, p. 31). table 15, after Wright and Weiblen (1967).10 Pigeonite, magnetite, and iron-rich rims on olivine, not distin- guished optically, have been identified by the electron microprobe. Melt samples that flowed into drill casings lost crys- tals during flow and appear to have lost augite relative to plagioclase and olivine. The latter effect can be seen in table 13, comparing samples collected at the top of a casing and those at the bottom (M21—26 top and bot- tom). The evidence for total amount of crystals lost is shown in table 16. The temperature of each sample is estimated from its crystallinity (glass content) using figure 17. This is compared with the actual tempera- ture of collection derived from figure 10. Loss of crys- tals during flow into the casing results in a tempera- "The glass percentages given as volumes in Wright and Weiblen (1967) are given as weight percent here. I00 IZOO‘ __l \\ 1 I I I I I | I I I I ”00. I l80' — \ EXPLANAVION “80‘ ‘I x ”60’ 9° _ + l \ + MODAL DAYA, VOLUME VERCENY _ 90 ”50' + AVERAGE MODAL DAVA CONVERYED YO WEIGNY VERCENV ”‘0.V 80 _ [SYIMATED YRUE AVERAGE WEIGHV —- 80 — IIAO. ‘ FENCENY MODE BASED ON H‘ + CALCULATION OF CHEMICAL MODE H20'—70—_* -‘70' ”20' H00“ — ___ bD — _ 60 1100- 2 ‘ E w 6. § 8 m IOBO‘ —so —+ — so— mm It“ a In I 3 0 Min MELT u ID70‘ —‘0_ _70— I070- CRUSV CIUSY 30 ' %+- + -‘ 30 I060' — # . . _ ,0”. 70 - < 20 I040' —+ ‘F ——~ Imo' . Io» — Io I020 * ill, i . * I020- 1000‘ — M., 000- 950' f I it \ 2d; #- no- 0 _1_ I l I I 1 | 0 O 5 0 IO 20 30 ‘0 50 0 0 i0 OllVlNE PERCENY CLINOPVROXENE PERCENV FIGURE 16.—Modal data (transmitted light—table 13) plotted as per- cent minerals versus percent glass. Temperatures (from fig. 17) are interpolated on the vertical axis. Data are converted to weight percent and compared with theoretical curves that pass through the mode calculated from the bulk chemistry (table 14). In all PLAOIOCLASE PERCENV ououz PERCENI samples the pyroxene and opaque minerals are overcounted and plagioclase is undercounted due to masking effects promoted by the fine grain size (table 18). The true variation of olivine with temperature and glass is not known because of possible crystal settling. OBSERVATIONS 125° I I I I I I ‘I I I _ LIQUIDUS ”00 EXPLANATION I —+— MODAL DATA, VOLUME PERCENT GLASS H50 _ /‘ AVERAGE CURVE, WEIGHT PERCENT GLASS _. 9 I IIoo — _ 5 I; MELT g CRUST 2 1050 ~ _ .1. ._ Iooo — — ‘souous 950 — _ 900 I I I l I L I I J 0 I0 20 30 40 so 60 70 so 90 I oo PERCENY GLASS FIGURE 17.——Percent glass plotted against collection temperature. This is an updated version of figure 12 of Wright, Kinoshita, and Peck (1968). Length of bars Show uncertainties in both tempera- ture and modal composition. All samples used to define curve were collected above 8.5 m depth and were neither contaminated nor flow-differentiated. Liquidus is estimated from the MgO/FeO ratio (Tilley, and others, 1964; see also Wright and others, 1968, table 5), and solidus is drawn where residual glass content becomes ap~ proximately constant. The interface between crust and melt is de- fined from the drilling to be at a temperature of 1,070°C (fig. 10). Crystallization rate (percent crystallization per unit drop in temp- erature) is higher near the interface than at temperatures closer to either solidus or liquidus. 1250 I'— | I I I I l I I LIQUIDUS 1200 - 1150 1100 MELT CRUST 1050 - — TEMPERATURE (°C) 1 000 -' SOLIDUS I I I I l I I I 0 10 O 10 20 30 40 50 O 10 OLIVINE CLINOPYROXENE 950 20 30 40 50 0 10 FLAGIOCLASE OPAQUE FIGURE 18.—Weight-percent minerals plotted against temperature. These are the solid lines of figure 16 plotted with temperature instead of glass content as the ordinate. This figure emphasizes, as does figure 17, the increased rates of crystallization across the crust-melt interface. ture estimate that is higher than the true temperature of collection. This is confirmed for two chemically analyzed samples (M23—21 and M24—3) which show a 35 TABLE 15.—Mineral paragenesis, Makaopuhi lava lake [After Wright and Weiblen (1967)] Temperature Mineral Composition (“Ct10°) Glass (weight percent) Olivine ________ Fmoas 1,205 100 Augite ,,,,,,,, Em-1FS13W040 1,185 94 Plagioclase ____ Ans7 1,180 92 Ilmenite ______ IlmssHemu 1,070 44 Olivine ________ F055 1,050 17 Pigeonite ,,,,,, Ens1FSazWO7 1,050 17 Magnetite ____ UspssMagm 1,030 9 Apatite ,,,,,,,,,,,,,,,,,,,,,, 1,020 7 Solidus ______________________ 980 4 (residual glass) large temperature discrepancy and have differentiated bulk compositions. One sample (M20—13) shows a lower temperature (higher crystallinity) then the temperature of collec- tion. This sample was collected on the stainless-steel rotor used to measure viscosity in the melt (Shaw and others, 1968). The sample probably crystallized some- what during the several hours it took to pull out rotor and casing. Samples collected during the eruption show a range of temperatures, all of which are probably lower than the temperature of the main body of the lake because of cooling near the exposed edge of the lake before collec- tion. The higher temperature samples were collected when the eruption rate was accelerating and the lake surface was visibly hotter (Wright and others, 1968). The petrography of the core changes as a function of increasing depth in the following ways: 1. Olivine decreases both in size and amount; large (>1 mm diam) phenocrysts are absent below 10.1 m. 2. Grain size of all minerals increases with depth to a maximum in partially molten core collected at 13.4 m, then decreases as the amount of liquid in- creases in core collected at higher temperatures and greater depth. 3. Vesicle size and amount increase in the same way as grain size, reaching a maximum at about 14.0 m. 4. The amount of residual glass apparently in- creases with increasing depth. Modal counts given about 12 percent by volume at subsolidus temperatures near 12.2 m compared with 4—6 percent at depths less than 4.6 m. Compositions of two residual glasses, determined by electron microprobe by R. L. Helz are given in table 17 compared with residual glass from Alae lava lake. The Alae glass was initially analyzed by wet-chemical methods, then used as a standard for probe calibration. The two Makaopuhi glasses differ significantly from the Alae glass only in iron content. All are mimimum melting calc-alkaline rhyolite compositions having about 3 percent admixture of mafic components. From mixing calculations (Wright and Doherty, 1970), the calculated amount of residual glass should be only 36 COOLING AND CRYSTALLIZATION OF THOLEIITIC BASALT, 1965 MAKAOPUHI LAVA LAKE, HAWAII TABLE 16,-Temperature (°C) of melt samples estimated in two different ways [T lmeasurodl gives the temperature of collection read from the reconstructed temperature data of figure 10 using the date and depth ofcollection. T lmodall is the temperature Inferred from the crystallization of the collected sample using figures 16 and 17. Where T (measured) exceeds T Imodal) the sample is inferred to have crystallized to some extent during slow quenching. Where T 1modal) exceeds T (measured), the sample is inferred to have lost crystals during flow differentiation. This effect is evident for two analyzed samples IMZK'JI and 3124—3) both of which have differentiated chemical compositions] Date of Depth of T 1measuredl T lmodal) Sample no. collection collection (it) (“Cl I‘f‘r Comment M-3 March 7, 1965 surface DOt known 1,125 Sample collected from the rising lava lake on ceramic tubes pushed into the melt. Quenched in air. M—7 March 8, 1965 "u do eeeeee not known 1,120 Do. M—8 March 11, 1965 n" do ,,,,,, not known 1,135 Do. M—12 March 12, 1965 1..” do ______ not known 1,165 Do. M—18 March 14, 1965 "H do ,,,,,, 121,160 1,160 Do. quenched in water. M—22 March 15, 1965 __,, do ...... not known 1,160 Do. M1—1G April 19, 1965 0.0 21,140 1,140 Glassy skin on lava—lake surface adjacent to drill hole 1. M60—1B April 1, 1965 surface not known 1,160 Small flow from early, short-lived, vent. M20713 August 3, 1965 21—23 1,130—1,135 1,120 Melt in casing used to emplace viscometer (see Shaw and others, 1968). Probably crystallized during recovery of viscome- ter and casing. M21—21 August 17, 1965 17—18 1,095—1,105 1,135—1,140 Flowed igto drill steel. Probably differ- entiate . M21—24 August 30, 1965 19—22 1,110—1,125 1,135—1,145 Lost augite relative to plagioclase during steam-impelled flow into drill steel. M21425 September 16, 1965 154—1605 1,065—1,075 <980 Pyroxene-rich residue (see table 13) left in hole after collection of 21—24. M21—26 September 16, 1965 25—2675 1,130 1,115—1,130 Collected on ceramic and by flow into stainless steel casing. Probable loss of some crystals during flow. M21—27 September 27, 1965 29—305 1,135 1,1201,130 Do. M22—15 November 9, 1965 21.0 1,095 1,095 Collected on ceramic. M22—18 November 9, 1965 21.0—23.0 1,095y1,110 1,100—1,125 Collected in flow into stainless steel cas— ing. Probably lost some crystals during flow. M23—21 January 19, 1966 24.0 1,095 1,140 Flowed into drill bit. Differentiated com» osition (table 10) M23—24 February 3, 1966 24.0 1,090 1,145 elt in drill bit. M24—2 July 28, 1966 27.0—29.0 1,060_1,085 1,085 Flowed into core barrel, bottom sampler. M24—3 July 28, 1966 29.0—30.0 1,085—1,100 1,115 Do. Top sample Differentiated composi» tion (table 10). ‘Thermocouple reading during collection. 2Minimum temfierature inferred from profile obtained on April 21. 1965 ltable 3) is consistent with finite element modeling and heat flow calculations related to loss of heat during crustal foundering arch 1&19, 1965 (HR. Shaw. written commun. 1974). about 10—20 percent greater for differentiated samples. Thus part of the difference in modal glass content is probably not real but instead due to grain-size-related counting errors. DISTRIBUTION OF OLIVINE The distribution of large (>1 mm) phenocrysts crys- tals of olivine in drill holes 68—1 and 68—2 is shown schematically in figure 19 as determined from megas- copic examination of drill core. The largest olivine phenocryst observed measured 7 by 8 mm, but most do not exceed 5 mm in diameter. There is a good corre- spondence between the MgO content and amount of observed olivine; samples in which MgO is less than about 7.8 percent contain no large olivine phenocrysts. The distribution of olivine is not the same in the two drill holes. Relative to 68—1, the upper part of 68—2 is depleted in olivine, and the zone from 9.1 to 10.1 In has visible olivine, missing in 68—1. VARIATION IN GRAIN SIZE Maximum, minimum, and median grain diameters have been estimated from thin sections for each analyzed sample in holes 68—1 and 68—2 (table 18), and median grain volumes were computed (fig. 20, table 19). Grain size remains nearly constant to about 6.1 m, then increases irregularly, reaching a maximum in the partially molten sample 68—1—17 col- lected at 13.4 m. The grain size then decreases down- ward as the percent of melt increases. CORE DENSITY AND VESICLE DISTRIBUTION The vesicle distribution in the lake was studied by inspection of drill core from holes 68—1 and 68—2 and by measurements of abundance and maximum and me— dian diameters in thin sections (tables 13 and 19). Bulk core densities were measured on large pieces of core (table 20; fig. 2). Core densities vary irregularly with depth, and the variation is different in the two drill DISCUSSION TABLE 17.—Composition afresidual glass from Makaopuhi and Alae lava lakes Makaopuhi lava lake Alac lava lake 6&147 68—1A14 A1249 AILIOII Depth 26.0'1 Depth 39.8" Iprobel' (wet chemicall2 75.0 75.1 75.9 75.8 12.6 12.6 11.9 12.2 .76 1.64 1.1 1.6 .03 .05 .05 .2 .40 .43 .54 .6 3.25 3.44 3.37 3.4 5.82 5.64 5.99 5.5 45 33 24 .7 98.2 99 2 99 2 100.0 ‘Electron microprube analyses by R. T. Helz. 2Recalculated from analyses of {our impure separates after correcting for feldspar and apatite, Analyst: R. Meyerowitz and J. Marinenko. holes, as is evident in the photographs of the two cores (figs. 24, 25). Megascopic inspection of core from holes 68—1 and 68—2 shows that the vesicle distribution is irregular in the upper part of the lake. Down to a depth of 3.0 m, the vesicles are large (as much as 15 mm diam, median diam=4—6 mm) and abundant. Their abundance in 68—1 decreases to a minimum around 6.1 m, corre- sponding to the location of the maximum core density (fig. 21). Large single vesicles persist to a depth of 6.1 m in both drill holes; the largest observed measures 20x16 mm at 2.5 m. Below 7.6 m the vesicle distribu- tion is quite regular. Below 3.0 m, vesicle cylinders and sheets are com- mon although poorly formed in both drill holes. These concentrations of vesicles are always associated with a greater amount of glass than is in the host rock, and where they are well formed, the rock breaks along them and blades of ilmenite line the vesicle walls. The glass and ilmenite suggest that the cylinders and sheets mark zones of low pressure into which relatively late-stage liquid was beginning to segregate. Thin-section study shows that median vesicle size increases with depth in the lake similar to the changes in grain size, reaching a maximum at about 14.0 m (fig. 20). The vesicularity of partly molten samples is great- er than that of the solidified rock immediately above, although the deepest partly molten samples in drill hole 68—2 are quite dense. Melt samples collected be- tween depths of 16.5 and 18.3 m are dense, containing 1—2 vesicles per thin section. The specific gravity of sample of melt containing only traces of crystals or vesicles (69—1—22) is 2.78:0.02. This contrasts with samples of melt collected in 1965—66 from depths of 4.6—6.1 m, which had many long tubular vesicles and a low bulk density. The relation between core density and depth are dif- ferent for the small pieces of drill core collected in 37 DEPTH MP—68—1 MP—68—2 METERS FEET OLIVINE OLIVINE _ SURFACE DIST. M90 DIST. M90 0 O 2 0 O 4 o — 7.7 o 0 o 6 °°°° o 8 0:0 — 8.2 o 10 o 0 0 ° 12 :0 o — 8.1 ° 0 14 0 ° 5 0° 00° 0 16 o o o o o 18 00:: — 8.0 o o 20 ° ° ° 0 ° ° 0 o 0 ° 22 o o — 8,] 0 0° 24 ° ° — 8.1 0 o 26 3 .. ° — 8.1 °°°° ° 28 °o °° °o — 8.0 10 30 — 7.4 o o o o 32 0 o °o — 7.8 70 O 0 0 34 — . 36 — 7.2 38 — 6.7 40 — 7.0 42 — 6.8 I 44 — 6.7 5 46 — 6.6 _ 6.6 43 ' 6.7 50 " 6.5 52 — 6.4 54 — 6.1 OLIVINE CONCENTRATION WITH DEPTH, MAKAOPUHI LAVA LAKE DRILL HOLES FIGURE 19.—Schematic distribution of large (>1 mm diam) olivine crystals in drill holes 68—1 and 68—2. Note the different distribu- tion in the two drill holes and the correlation between MgO con- tent and the incidence oflarge olivine. Samples below 8.5 m depth are flow differentiated (fig. 22). 1965—66, compared with the large pieces collected in 1968 (fig. 21). For a given depth, bulk core densities are lower in the larger core reflecting the irregular dis- tribution of vesicles and the presence of some large vesicles. Drilling without a core spring reduces recov- ery and biases it toward denser pieces of rock. DISCUSSION All of the data summarized under “Observations” contribute to our understanding of the cooling and crystallization of basaltic lava in Makaopuhi lava lake. Unfortunately none of the sets of measurements forms COOLING AND CRYSTALLIZATION OF THOLEIITIC BASALT, 1965 MAKAOPUHI LAVA LAKE, HAWAII {Grain size (mm)] TABLE 18.—Summary ofgrainsize measurements for analyzed samples from drill holes 68—] and 68—2. Olivine Augite Plagioclase llmenite Sample Depth‘ (ft) T (°C)2 median min/max median min/max median minrmax median minimax 68—1—2 8.0 <100 0.15—0.3>< 0.1><0.1 0.015><0.015 <.015 0.03><0.15 <.015 0.03><.6 .15><.15 .06><.35 .07><.2 68-1—3 12.0 <100 .3><.3 .1><.1 .015—.03 <.015 .015 to .04 <.015 .015—.04 .015 1X1 ><.015—.O3 .25><.25 ><.15—.2 .07><.5 ><.15—.3 .07><.3 68—1—4 17.5 <100 .2—.35 .1 ><.1 .015—.03 <.015 .015—.04 <.015 .03>< .015 ><.2—.35 .6><.l ><.015—.03 .35><.35 ><.15—.2 .1><.4 .15—.20 .07><.35 68—1—5 22.0 >120 .2—.3 .07><.07 .015~.03 <.015 .015—.04 <.015 032.04 .015 ><.2—.3 .4><1 ><.015—.03 .15><.15 ><.15—.2 0.1><0.9 ><.2 .15><.4 68—1—6 24.0 >180 .3><.3 .1><.1 .015—.03 <.015 .015—.04 <.015 .04—.07 .015 (rare) 1x2 ><.015—.03 .15><.2O ><.5—.2 .3><.3 ><.15—.2 .15><.3 68—1—7 26.0 >250 .2><.3 .1><.1 .015—.04 <.015 .015—.04 2.015 .04—.07 .015 (rare .8><1.1 ><.015—.04 .3><.4 ><.15 .5><.5 ><.15—.3 .2><.5 68—1~8 28.0 >330 .3><.3 .1><.1 .015—.04 <.015 .015—.04 2.015 .04—.07 .015 (rare) .6><.7 ><.015—.04 .35><.35 ><.15 .15><.5 ><.15—.3 .2><.4 68—1—9 30.0 >410 .35><.35 .2><.2 .03—.06 <.015 042.07 .02 .06—.08 .02 .4><.6 ><.03—.06 .2><.65 ><.2—.4 .15><1 ><.2—.45 .15X.4 68—2—10 32.0 >480 .3><.3 .1 ><.1 .015—.04 <.015 .015—.04 2.015 .03—.06 .015 (rare) .35><1.2 ><.015—.O4 2><.2 ><.15 .5><.5 ><.2 .2><.3 68—1—11 338 >545 .3><.3 .7><.7 .03—.07 2.015 .07 .025 .07—.10 .02 (rare) ><.03—.07 .3><.3 x.%.3 .35><.5 ><.3—.45 .2><.7 68—1—12 36.0 >620 .2><.2 .75><.75 .03—.07 2.015 .07 .02 .07—.10 .03 ><.03~.07 .25><.25 ><.2—.3 .1><.7 ><.3—.45 .15><.4 68—2—13 38.0 >690 .15—.35 .8><.8 .04—.07 .025 .04—.07 .02 .07—.15 .03 ><.15—.35 ><.O4—.07 .2><.2 ><.2—.3 .15><.8 ><.3—.45 .2><1 68—1—14 39.8 >750 .2><.2 1.2><.25 .04—.07 .02 .03—.07 .03 .07—.15 .03 (rare) ><.04—.07 .25><.25 ><.2—.35 .15—.7 ><.3—.5 .2><.5 68—2—15 (41.0) 990 2><.2 .25><.25 .07—.14 .03 05—.1 .03(rare) .07—.15 .03 (rare) 42.0 ><.07—.14 .4><.4 ><.4—.7 .3><.6 ><.3-.5 .15><.7 68—1—16 44.2 1,011 none .07—.15 .04 .07—.15 .03 .07—.15 .02 (rare) ><.07—.15 .3><.45 ><.3—.45 .35><.8 ><.3—.55 .3><.6 68—1—17 46.1 1,029 .3><.3 .5><.5 .07—.15 .03 06.10 .03 .07—.15 .015 ><.07—.15 .15><.6 ><.3~.4 .4><.6 ><.25—.55 .35><.6 68—1—18 48.0 1,043 rare .06—.10 .03 .04—.07 .03 .04—.09 .02 (rare) 2><.2 .4><.4 ><.06—.1 .15><.3 ><.2—.4 .2><.7 ><.3—.55 .2><.7 6&1—19 49.5 1,055 rare 042.10 .03 .03—07 .02 cannot .2><.2 .2><1 ><.04—.1 .15><.3 ><.15—.35 .1><.6 .07><.4 tell 682220 (50.5) 1,060 2><.2 .35><.6 .03—.07 .02 .03-.06 .015 .04 .01 51.5 ><.03—.07 2><.2 ><.15—.30 25—.5 ><.25—.45 .07><.9 68—1—21 54.0 1,082 rare .03—.07 .02 .03—.06 .015 t .2x.2 .3><.5 ><.03—.07 .2 ><.4 ><.15—.20 .15><.3 “0. Present 1Samples collected in drill hole 6&2 were slightly cooler at the same depth than samples collected in drill hole 6&1. For consistency in comparing data from the two drill holes, the value in parenthesis is the estimated depth in 6&1 that fits the temperature of collection. The 2d value is the actual depth of collection. 2Temperatures are estimated from the last temperature profile measured. The crust-melt interface was assumed to be at 1,070°C, and the solidus was assumed to be at 980°C. The position of the LOOO° isotherm was estimated from the glass content of the core using figure 17. The temperature data are subject to uncertainties in these assumptions. TABLE 19.—Volumes (mm3X106) of crystals and vesicles; samples from drill holes 68—] and 68—2 Augite Plagioclase Ilmenite Vesicles Sample Depth N0. lfll Median Maximum Median Maximum Median Maximum Median Maximum 6&1—2 8 0.00005 3.4 0.13 4.2 0.13 1.9 rvglldgee 15625 1—3 12 .0013 15.6 .13 9.8 .11 3.9 20.8 5832 1—4 17.5 .0013 42.9 .13 10.0 .16 5.1 20.8 125 1—5 22 .0013 3.4 .13 45.0 .25 16.5 76.8 1953 1—6 24 .0013 4.5 .13 27.0 .53 10.1 76.8 1000 1—7 26 .022 36.0 .11 125.0 .68 35.0 144.7 236 1—8 28 .022 42.9 .11 3.0 .68 24.0 231.2 729 1—9 30 .091 26.0 .91 86.3 1.59 16.5 301.8 800 2—10 32 .022 8.0 .11 125.0 .41 15.0 74.4 512 1—11 33.8 .125 27.0 1.23 73.5 2.71 63.0 231.2 625 1—12 36.0 .125 15.6 1.23 28.0 2.71 16.5 130.0 452 2—13 38 .166 8.0 .76 57.0 4.54 120.0 480 729 1—14 39.8 .166 15.6 .69 44.6 4.84 35.0 166.6 452 2—15 41.0 1.16 64.0 3.09 81.0 4.84 44.6 195.3 875 1—16 44.2 1.33 16.2 4.53 161.0 5.14 81.0 421.9 5832 1—17 46.1 1.33 13.5 2.24 120.0 4.84 99.8 693 14976 1—18 48.0 .512 6.8 .91 63.0 1.80 63.0 421.9 5832 1—19 49.5 .343 6.8 .63 21.0 1.96 19.7 729 2—20 50.5 .125 8.0 .46 46.9 1.06 30.6 144.7 1000 1—21 54 .125 17.6 .35 10.1 ____ 11-1 64.0 125 DISCUSSION 39 SAMPLE DEW“ IN METERS TABLE 20.—Density of drill core from holes 1—23, 1965—66, and 68—1 , 5 10 15 20 68—2, 1968—Continued 10 I I I I 10,000 Depth Weight Density ' 'LMEN'TE Core 110. (ft) (grams) (g/cc) x PLAGIOCLASE M—1176 2.0.3.0 11.032 4 2.20 ’ AUG'TE M71178 4.0.5.0 20.139 230 M_11_9 5.0—6.0 14.907 226 ° VES'CLE M—11—10 607.0 13.358 2.30 M_11_12 8.0—9.0 9 794 2 38 M—11—13 90—10 0 8 268 2 57 1— —1000 M—11—14‘? 100—110 5 751 2 48 ,2 M—11—16 11912 9 27 868 2 64 o 65 .2) pr“ M—11—16 11.9.1; 9 "6.554 2 66 7 _ 2 M—13—14? 10.1-11 1 4.862 2 54 x x M—13—15 11.0.11 8 6.902 2 60 ”E ”E M—13—13 9.1-10.1 16.141 2 54 5 g M4314 10 1-11.1 25.356 2 65 q 655 W M41314 10 1—11.1 10.047 2 66 *- m ; M71&3 (H8 10.334 1.75 g 3 M—16—6 3.8648 11 808 2.47 6 5 M71689 5.92419" 6 "88 2.58 > > M7169 6.92492 13 379 2.65 w hl—l&11 8.92—9.92 7.060 2.57 g 4‘ ' — ‘00 d M71€¥14 119241292 7460 272 a: a M174 0915 14.915 2.26 .7 055 0 2 0171774 0915 14.620 2.25 2.2 2. w M—17—10 13514.5 6.498 240 n: 5 M72043 5.1-6.1 10.303 2.55 g z M—2(LQ? 121.131 3.371 2.50 g .7, M—20—10 13144.1 9.688 . z M—20—10 13.1—14.1 8.819 g 5 M—21—14 129613.95 7.245 5 3 M—22—10 ? 11012.0 4.531 g 2 M_23_1 0.1.2 20.244 M—23—2 1.2-2.2 10.848 -01 - ~10 M—23—3 2,243.2 8339 M—23—7 597.9 7.664 M4238 7.9—8.9 12.345 M~i2£L9 899.9 6.558 M42313 12913.9 16.070 M72315], 15916.9 3.638 Drill hole 6&1: Drill hole 68—2: Defpth Weight Density Degth Weight Density ( t) (grams) (g/ccl ( l lgrams) (g/cc) ‘00] I I I l I I I l I I I I I I I l I I I I I l I I I I I I I I I I I l ‘l ° 1° 2° 3° 4° 5° °° 1.0 350.9 1.95 0.5 35.2 1.94 SAMPLE DEPTH, IN FEET 140 1.91 2‘0 63-5 2-32 1.9 72.8 2.46 3.7 284.8 2.51 . . ' _ . 3.7 2.44 3.9 135.8 247 FIGURE 20.—Med1an volumes of augite, plagioclase, 1lmen1te, 675.3 3324! 3.3 322.8 3.32 and vesicles plotted against depth (note different scale for vesi- 8:8 {1:76 922 723225 5:59 cles). Data are given in table 19, calculated from grain size 32 5.22 1343 3952-9, 21%? data given in table 18. Grain size is nearly constant between I833 32; fit: 1%?3438 2.66 . . . . _. 1 1.8 m .and 6.1 m. Between 1.8 m and the surface, the core 114 257 11.7 495.5 2.;7 shows increasing amounts of devitrified glass and fine-grained 1:7,; :21) 13% 37,1515; 3422 quench intergrowths of the mineral phases. Below 6.1 m, grain I48 2.28 12.8 284.2 2.61 . . . . . 1: 1.. .1 2 _. 2. Size increases and reaches a maximum in partially molten core 131;; 184 E5 11%.?) 223 collected at 13.4— 14.0 m. Grain size decreases below 14.0 m 135-3 13% 311] Egg}? Egg correlated with decreasing crystallinity of the partially molten 188 2.7: 1715 1896 5:69 . 2.7 17.9 211.9 272 samples. 18.3 2.71 18.2 155.8 2.66 18.6 2.78 18.3 2150.0 2.67 20.0 2.78 19.5 955.0 2.68 20.6 2.74 20.0 968.5 2.66 25:8 273 221 1385.0 2.75 _ . L. . 2.72 22.9 1257.0 2.74 TABLE 20.—Denszty of drill core from holes 1—23, 1965—66, and 68—1, 24.2 2.72 23.6 2880.0 . _ 25.0 2.76 3 239.5 68 2’ 1968 25.9 2.68 2 .5 302.0 26.5 2.69 26.0 680.0 Depth Weight Density 335'; Elf/5:13 £33 3329 Core no. (ft) (gramsi (g/ccl 2935 2.67 27:9 545:3 30.7) 2.659) 30.9 263.2 _ .7 ., 31. 2. 31.4.1 289.1 1132 533 13:23:; 5:33 33-5 2-58 3145 604-2 M»3—6 1 08187 11.028 2.39 3446 2'57 33-0 58345 M_4_3 4 010 7 161 163 37.2 2.63 34.1 3010 M64 0.10 9.595 1:60 38-5 2'57 3‘10 3§4-Z M435 1 0.21 8.043 2.33 45'1 3:445? '37'0 ’1'? M54; 21—31 9.404 2.36 45'5 3'5“ 388 .1935” M‘s—7 3.1—4.1 5.042 2.51 45's 354 “9-? 108-5 M—EH) 6.1—7.1 9.979 2.45 22% 3553 :g‘g’, 11321 M—5—11 8.1-9.1 15.448 2.43 0 - - ‘ .y 0 - 7 ' M—5—11 8149.1 15.148 2 58 30’ 47‘1 0‘5“ 47‘0 80")" M—5—13 10.1—10.3 13.568 2 66 47-5 .,'46 47-1 1953-85 M—7—5 3 5—4 9 33.872 2 48 47-6 450 4950 141-5 M_9_7 5&3'0 6.840 ., 44 48.9 2.48 49.1 207.0 M98 30.40 4.141 2 37 48-0 2‘50 50-0 704-0 M99 40.50 7.942 2.42 49-0 ”‘53 5 4° 928-5 M4014 10010.6 5.367 2.61 50-7 “790 M_11.4 0.10 19.558 2 03 M71174 01.0 16.481 2 03 M—11—5 1.0.2.0 12.724 2 24 40 MP—6B—I +1965766 DRILL CORE MP—68—2 0 I I I I 0 \\'\I I II ”I L' II I I ‘ | A A I I ‘4 I I M \\f” I II | ,0_ NI' I”| ‘ ‘ ‘ '. i —10 I I0 fly II | A ‘A‘ I g\ 5 A '. I ‘ A A AA 20 — . w . q 20 5/. s : 5/ E - n C ._ 2 ‘ . .2 p so — ,4 I0 Z i A. ' 30 m u. ‘ m I . / a a ; g . I“ g . . ‘ z a 40 — / — 40 g ‘Aél‘ 15 M . so — “t - . — 5° ’EXPLANATION 6:0 — I 1965—at: SP DRILL CORE (WT. <27g) — 60 . WT. <5009 7o _ - 5009 I0009 I l l I IL I I I I 2.0 2.2 2.4 2.0 2.3 2.0 2.2 2.4 2.6 2.8 3.0 SPECIFIC GRAVITY specmc GRAVITY FIGURE 21,—Bulk density of drill core plotted against depth. The left-hand plot compares the density of 1-cm drill core collected in 1965—66 with that of 6-cm drill core collected from drill hole 68—1 in 1968. The incidence of large vesicles and vesicular zones recovered in the larger core explains the somewhat lower bulk density of this core at similar depths compared with the 1-cm core where only the denser pieces were recovered. The right-hand plot shows the density profile of drill hole 68—2 drilled only a few metres away. The scatter is much greater although both holes show a maximum density of 2.78 g/cc at depths of 5e6 m. Densities of partially molten core reach minimum values of 2.5 g/cc at a depth of 15 m before increas— ing again below the crust-melt interface (DH 68—2). Density of crystal- and vesicle—free glass that flowed into the core barrel during drilling ofhole 69—1 is 2.78 g/cc for comparison (sample 69—1422), tables 9 and 11). a complete record, because of the difficulties of meas- urement and the premature burial of the lava lake. The following sections discuss only those topics for which information is sufficient to form reasonable in- ferences about processes that took place in the lake. Final answers will come only from study of other bodies of molten basalt and from experimental studies and scale modeling of cooling and crystallization proc- esses. We hope that the following sections will stimu- late others to pursue topics of use in refining the hypotheses presented in this paper. CHEMICAL DIFFERENTIATION IN THE LAVA LAKE Three differentiation processes were observed in the lava lake. 1. Gravitative settling of large olivine crystals. 2. Removal of augite, plagioclase, and smaller olivine crystals from melt, possibly during convective flow. 3. Formation of liquid segregations by flow into open fractures in the partly molten crust. These processes are seen on a small scale in the lava COOLING AND CRYSTALLIZATION OF THOLEIITIC BASALT, 1965 MAKAOPUHI LAVA LAKE, HAWAII lake, but they are also important on a larger scale in explaining the overall variation in composition of Kilauea lavas, for they are inferred to take place in the conduits and magma reservoirs of the volcano (Wright and Fiske, 1971). GRAVITA'I‘IVE Sl‘l’l'TLIXG ()l‘ OLIVINI’. The incidence of large olivine phenocrysts decreases markedly, if erratically, with increasing depth in the lava lake (fig. 19). We infer that many of the large olivine phenocrysts settled through the melt and be- came concentrated near the bottom of the lake. Support for this inference comes from the study of the prehis- toric Makaopuhi lake, which has a zone of high olivine concentration about three quarters of the way from the top (Moore and Evans, 1967). Is the presence of large olivine phenocrysts at the observed depths in the 1965 lava lake consistent with a simple crystal settling model? To answer this question, we assume that Stokes law is obeyed and make the additional simplifying assumption that the olivine set- tles at a constant rate at temperatures of more than 1,100°C. If the settling rate exceeds the rate at which the 1,100° isotherm moves downward, olivine will be lost to the bottom part of the lake. The slope of the 1,100° isotherm, taken from figure 10, as a function of depth is as follows: Depth (m) ______________ 1.52 3.05 4.57 6.10 7.62 Slope><105 (cm/sec) ______ 6.64 3.55 2.40 1.82 1.46 Viscosities of melt at 1,100°, 1,110°, and 1,130°C were estimated from glass analyses (table 11, analysis 69—1—22; Weiblen and Wright, unpub. data) using the method described by Shaw (1972, table 2 and equation 3, p. 873). Results are as follows: Sample No ________ 69—1—22 M22—18GL M21—26G1 Temperature (°C) __ 1,100° 1,110° 1,130° Viscosity (poises) __ 1,148 1,072 832 Ifwe use the maximum Viscosity value (1,148 poises) and assume a density difference of 0.6 (g/cc) between olivine and liquid, the Stokes law equation can be solved for the radius (r) of a crystal whose settling rate would equal the rate of movement of the 1,100° isotherm. _ 2gr2Ap V 9n (5) where V = velocity of the particle Ap = difference in density between olivine and liquid 7) = viscosity of the liquid in poises g = gravitational constant DISCUSSION Solving for a velocity equal to the rate of depression of the 1,100° isotherm at a depth of 1.52 m, we get: 6.64x 10*5 = 2 x980><0.6 >1,070°C) at velocities related to their size and the changing viscos- ity of the enclosing well. Some bubbles became frozen in position as they reach the crust-melt interface (T=1,070°C). The presence of inclined vesicle sheets and incipient vesicle cylinders, as well as direct obser- vation of gas escape at the surface, indicates that some gas escapes through the crust, eventually connecting with fractures open to the surface. A profile of core density versus depth principally re- flects the relative amount of gas trapped by the crystal- lizing crust. The difference in core density from drill holes 68—1 and 68—2 (fig. 21) is an indication of the irregular nature of gas movement and entrapment. We attach significance to the fact that the volume rate of subsidence of the lake surface (fig. 14) decreased COOLING AND CRYSTALLIZATION OF THOLEIITIC BASALT, 1965 MAKAOPUHI LAVA LAKE, HAWAII abruptly about the time at which the crust attained its maximum density in late 1965 (figs. 10, 21, 14). Sub- sequently, the rate of subsidence continued to decrease in stepwise fashion, while core density decreased (ves— icularity increased) in more regular fashion. The in- creased gas trapped in and immediately below the crust may have provided the increased buoyancy to offset the increase in crust density due to crystalliza- tion and thermal contraction. Continuation of this process eventually resulted in uplift of the lake sur- face. (fig. 12, 14). We do not know why the amount of gas frozen in the crust differed from one time to another. This could be related to a thickening crust and to the decreased temperature of vesiculation postulated earlier from ob- servation of glass density. When bubbles were freely moving in the melt, they could perhaps escape laterally along the crust-melt interface before being frozen into the permanent crust, as well as escaping upward through the cooling crust. As the beginning of vesicu— lation moved closer in temperature and distance to the interface between crust and melt, a lesser percentage of gas would be able to escape laterally until eventu- ally, when the temperature of beginning of vesicula- tion was less than 1,070°, all of the exsolved gas en- tered, and most was trapped, in the growing crust. This is one way to relate the pattern of vesiculation to the changes in level of the lake surface. In conclusion, we again emphasize that our interpre- tations are based on incomplete data on a model that assumes a similar pattern of solidification for all parts of the lake. Crustal foundering could have produced local points of gas concentration and inhomogeneities in the temperature distribution that might affect the vesiculation process. Possibly even factors external to the lake, such as tilting during intrusion into the upper east rift zone, (fig. 13) could also have affected the pat- tern of vesiculation. We hope that comparison of our data with data from other lava lakes can eventually resolve some of the factors that affect the altitude changes on the lake surface. SUMMARY: COOLING AND SOLIDIFICATION HISTORY OF MAKAOPUHI LAVA LAKE Our interpretations of the cooling, solidification, and differentiation of Makaopuhi lava lake are made in terms of several interrelated processes. 1. Density change on solidification. Where the crust is less dense than the melt from which it forms, there is a tendency to uplift the surface of the lake and vice versa. 2. Temperature of vesiculation. Vesiculation at tem- peratures between 1,190° and 1,070°C complicates the REFERENCES CITED density distribution in the melt as the bubbles will rise and become larger after vesiculation begins. Vesicles formed at temperatures less than 1,070" are trapped in place in the growing crust. This factor is at least as important as the change in density during crystalliza- tion in affecting the relative densities of crust and melt. 3. Convection of the melt. This is likely to occur in dense, nonvesiculated melts and is believed responsi- ble for differentiation of the melt in which small crys- tals of iron-rich olivine, augite, and plagioclase are concentrated downward resulting in eventual cyrstal- lization of a differentiated melt. We can trace these processes by looking at the changes of chemical composition, core density, temper- ature profiles, and surface-altitude changes through time as shown in figures 10, 14, 19, 21, and 22. The earliest cooling regime extends from the time of formation of the permanent crust of March 19, 1965— January 1966 when the upper crust was 6.7 m thick. During this time we see evidence of the upper cooled layer of melt and infer conductive cooling of both melt and crust from the linear variation of isotherm depth with \/t (fig. 10). Finite—element calculations show that the initial thermal layering is largely eliminated during this period. Crustal isotherms are depressed near the close of this period because of the accumula- tive effect of rainfall on the surface. Crustal densities increase until near the end of this period when they reach a maximum value of 2.7 8g/cc, then begin to de- crease again (fig. 21). Rates of subsidence of the surface are high at first because of initial degassing, then con- stant from July to December 1965 (fig. 14). The rates of subsidence abruptly decrease near the time of reversal of core densities. The chemistry and petrography of the upper crust is uniform except for the erratic decrease in olivine content with depth (fig. 19). Melt collected in this time period is frothy and of low bulk density and, apart from contamination effects introduced during sampling, is undifferentiated relative to the crust col- lected at the same or shallower depths. A second stage of cooling extends from February through about December 1966, when the crust was 9.1 m thick. During this period the depth to isotherms in the melt (1,07 O°—1,140°C) fluctuates and the general slope of the isothermal surfaces in the melt is shal- lower than the slopes during stage 1 (fig. 10). Differ- entiation of solidified crust is observed in core collected near the end of this period (fig. 15, 22) and the density of crust shows distinct decrease with increasing depth (fig. 21). The rate of subsidence (fig. 14) decreases to less than half the rate observed during stage 1 near the end of the period. We interpret stage 2 as the time when the initial thermal layering in the lake was 47 eliminated and when vesiculation of the lava was sup- pressed at temperatures of more than 1,070°C. Ther- mal convection was initiated at high temperatures leading eventually to differentiation of the solidified crust. (From fig. 10 it can be seen, on this model, that crystal-liquid differentiation was effective, beginning in February 1966 at temperatures more than about 1v,118°C in order to cause the observed differentiated compositions below a depth of 8.5 m.) The increasing density of melt (lacking vesicles) compared with crust forming from it (fig. 21) is inferred to be responsible for the decreasing rate of subsidence (fig. 14). The last time period is from 1967 to the last tempera- ture measurements in February 1969. We know little about the thermal history except that the slope of the 1,07 0° isotherm steepens past the extrapolation of the slope during stage 1. Core density continues to de- crease and, beginning in mid-1967 the central part of the lake shows net uplift which continues to the end of the period. The crust continues to become more differ- entiated with increasing depth during this period. The combined data imply that the convective regime was still operative and that the crust forming in the center of the lake was considerably less dense than was the melt. None of the history can be exactly described in terms of any simple predictive model. Many of the changes we see and their timing may reflect inhomogeneities in either the initial temperature distribution or in terms of the presence of former foundered crust. Where we correlate the various aspects of the lake there are often timelags between one type of observation and another observation that is assumed to follow; for instance, the earliest evidence of differentiation in newly formed crust followed by several months the earliest evidence of convection in the melt. Nonetheless, we feel that there is sufficient information to present these ideas as best working hypotheses. REFERENCES CITED Bradley, W. H., 1965, Vertical density currents: Science v. 150, no. 3702, p. 1423—1428. 1969, Vertical density currents: Limnology and Oceanography v. 14, p. 1—3. Evans, B. W., and Moore, J. G., 1968, Mineralogy as a function of depth in the prehistoric Makaopuhi tholeiitic lava lake, Hawaii: Contr. Mineralology and Petrology, v. 75, no. 2, p. 85—115. Evans, Bernard W. and Wright, T. L., 1972. Composition of liquidus chromite from the 1959 (Kilauea lki) and 1965 (Makaopuhi) eruptions of Kilauea volcano, Hawaii: Am. Mineralogist, v. 57, no. 1—2, p. 217—230. Finlayson, J. B., Barnes, I. L., and Naughton, J. J., 1968, Develop- ments in volcanic gas research in Hawaii in The crust and upper mantle of the Pacific area: American Geophys. Union Geophys. Mon. 12, p. 428—438. Grommé, C. 8., Wright, T. L., and Peck, D. L., 1969, Magnetic prop- erties and oxidation of iron titanium oxide minerals in Alae and 48 Makaopuhi lava lakes, Hawaii: Jour. Geophys. Research, v. 74, no. 22, p. 5277—5293. Hakli, T. and Wright, T. L., 1967. The fractionation of nickel be- tween olivine and augite as a geothermometer. Geochim. et Cosmochim. Acta, v. 31, p. 877—884. Huebner, J. S., 1971, Buffering techniques for hydrostatic systems at elevated pressure in Research techniques for high pressure and high temperature, Gene C. Ulmer, ed.: New York, Springer Ver— lag, p. 123—177. Huebner, J. S., and Sato, M., 1970, The oxygen fugacity-temperature relationships of manganese oxide and nickel oxide buffers. Am. Minerologist, v. 55, p. 934~952. Moore, J. G., and Evans, B. W., 1967, The role of olivine in the crys« tallization of the prehistoric Makaopuhi tholeiitic lava lake, Hawaii: Contr. Mineralology and Petrology, v. 15, no. 3, p. 202—223. Murata, K. J., and Richter, D. H., 1966, Chemistry ofthe lavas ofthe 1959—60 eruption of Kilaeua volcano, Hawaii: US. Geol. Survey Prof. Paper 537—A, p. 1—26. Peck, D. L., Moore, J. G., and Kojima, George, 1964, Temperatures in the crust and melt of Alae lava lake, Hawaii, after the August 1963 eruption of Kilauea volcano—a preliminary report in Geological Survey research 1964: US. Geol. Survey Prof. Paper 501—D, p. D1—D7. Peck, D. L., Wright, T. L., and Moore, J. G., 196 , Crystallization of tholeiitic basalt in Alae lava lake, Hawaii: Bull. Volcanol. v. 29, p. 629—656. Richter, D. H., and Moore, J. G., 1966, Petrology of the Kilauea Iki lava lake, Hawaii: US. Geol. Survey Prof. Paper 537—B, 26 p. Rittman, A., 1973, Stable mineral assemblages of igneous rocks—A method of calculation, in Minerals, rocks, and inorganic mate— rials, v. 7: New York, Mendleberg, Berhn, Springer-Verlag, 262 p. Robertson, E. C. and Peck, D. L., 1974, Thermal conductivity of ve« sicular basalt from Hawaii. Jour. Geophys. Research, v. 79, p. 4875—4888. Sato, M., 1971, Electrochemical measurements and control of oxygen fugacity and other gaseous fugacities with solid electrolyte sen- sors, in Research techniques for high pressure and high temper- ature, Gene C. Ulmer, ed.: New York, Springer Verlag, 368 p. Sato, M., 1971 M., and Moore, J. G., 1973, Oxygen and sulphus fugacities of magmatic gases directly measured in active vents of Mount Etna. Royal Soc. [London] Philos. Trans. A, v. 274, p. 137—146. COOLING AND CRYSTALLIZATION OF THOLEIITIC BASALT, 1965 MAKAOPUHI LAVA LAKE, HAWAII Sato, Motoaki, and Wright, T. L., 1966, Oxygen fugacities directly measured in magmatic gases: Science, v. 153, no. 3740, p. 1103—1105. Shaw, H. R., 1972, Viscosities of magmatic silicate liquids: an empir- ical method of prediction: Am. Jour. Sci., v. 272, p. 870—893. Shaw, H. R., Wright, T. L., Peck, D. L., and Okamura, R., 1968, The Viscosity of basaltic magma: an analysis of field measurements in Makaopuhi lava lake, Hawaii: Am. Jour. Sci. v. 266, p. 225— 264. Shaw, H. R., Kistler, R. W., and Evernden, J. F., 1971, Sierra Nevada Plutonic Cycle: Part II Tidal energy and a hypothesis for orogenic-epeirogenic periodicities. Geol. Soc. America Bull. v. 82, p. 869—896. Swanson, D. A., and Fabbi, B. P., 1973, Loss of volatiles during foun- taining and flowage ofbasaltic lava at Kilauea volcano, Hawaii. US. Geol. Survey, Jour. Research, v. 1, p. 649—658. Tilley, C. E., Yoder, H. S., and Schairer, J. F., 1964, New relations on melting of basalts; Carnegie Inst. Washington Yearbook 63, 1963—64, p. 114—121. Wright, T. L., 1973, Magma mixing as illustrated by the 1959 erup- tion, Kilauea volcano Hawaii: Geol. Soc. America Bull. v.84, no. 3, 849—858. Wright, T. L., and Doherty, P. C., 1970, A linear programming and least squares computer method for solving petrologic mixing problems. Geol. Soc. America, Bull. v. 81, no. 7, 1995—2007. Wright, T. L., and Fiske, R. S., 1971, Origin of the differentiated and hybrid lavas of Kilauea volcano, Hawaii, Jour. Petrology, v. 12, no. 1, p. 1—65. Wright, T. L., and Weiblen, P. W., 1967, Mineral composition and paragenesis in tholeiitic basalt from Makaopuhi lava lake, Hawaii [abs]: Geol. Soc. America Program 1967 Annual Meet- ing, p. 242—243. Wright, T. L. Kinoshita, W. T., and Peck, D. L., 1968, March 1965 eruption of Kilauea volcano and the formation of Makaopuhi lava lake: Jour. Geophys. Research, v. 73, no. 10, p. 3181—3205. Wright, T. L., Swanson, D. A., and Duffield, W. A., 1975, Chemical compositions of Kilauea east-rift lava, 1968—1971. Jour. Petrol- ogy, v.16, p. 110—113. Wright, T. L., Peck, D. L., and Shaw, H. R., 1976, Kilauea Lava lakes: natural laboratories for study of cooling, crystallization, and differentiation of basaltic magma, in Sutton, G. H., Man- ghnani, M. H. and Moberly, Ralph, eds., The geophysics of the Pacific Ocean basin and its margin: Am. Geophys. Union Geophys. Mon. 19, p. 375—390. FIGURES 24—28; TABLES 24—29 COOLING AND CRYSTALLIZATION OF THOLEIITIC BASALT, 1965 MAKAOPUHI LAVA LAKE, HAWAII 50 ‘90; '5wa l 1 D FIGURE 24.—Photographs of drill core for 6&1. A, 0—144 ft. B, 14.4—28.8 ft. C, 28.8—54.0 ft. D, Bags containing shattered glass (quenched melt) that came out around the drill hole collar While drilling between about 44 ft and 57 ft. FIGURES 24—28; TABLES 24—29 FIGURE 25.—Ph0tographs of drill core for 68—2. A, 0—17.0 ft. B, 17.0—33.0 ft. C, 33.0—52.0 ft. 51 52 COOLING AND CRYSTALLIZATION OF THOLEIITIC BASALT, 1965 MAKAOPUHI LAVA LAKE, HAWAII TOP FEET O D FIGURE 26.—Ph0tographs of drill core for 69—1.A, 0—230 ft. B, 23.0— 41.0 ft. C, 41.0—56.0 11. D, Core barrel filled with glass (quenched melt) collected at approximately 61—66 R. FIGURES 24—28; TABLES 24—29 I o I I I I I I I I I 0 l I - I I I _2 g '5 § E 5/17/65 , 5/2A/as Z ‘ 8 ’ 5/5/05 Z ‘ :5 E I: a I - 3 w ‘“ a a Io - l 12- _ ‘ l I u - DH No.3 - 5 m I I I I I I I I I loo 200 300 400 500 600 700 800 900 1000 I I 00 I TEMPERATURE “c A 0 I I I I I I I I l 0 2 _ — I I l _ ° ‘ - 2 ‘2 I—‘ a E / E 0/ 2 Us ZI 5 ' Z ‘ f E 6/63 L‘ n. I / 0‘, - 3 w SI ”3% 91/ I: I0 - Ir q, I I2 - - A H _ DH No.4 5 I 16 I I I I A I I 1 I I 00 zoo 300 400 500 (:00 700 800 900 I 000 I I oo TEMPERATURE “c B l 0 I l I I I I I I r 0 I 7 ' - I I ‘ _ l ° ‘ -I 2 V, K ‘ E ’- IN E a 2 II 5‘ T I 5 5/24/65 . 3 9i E I0 "9" _ 6/16/05 é/7/65 I2 _ - 4 u _ DH No.5 < s ,0 I I I I I I . I I ‘ I00 200 300 400 500 600 700 500 900 I000 IIoo TEMPERATURE “C C DEPTH, IN FEET DEPTH, IN FEET DEPTH, IN fEET 53 o _ I - 2 ‘ 3 I2 - l A ,4 _ DH No.6 - s ,6 I I I I I I I I I I00 200 300 400 500 600 700 900 900 I000 I 100 TEMPERATURE °C o ”5. I I I I I I I I . o \ \n 2 . -I 4 I- é ‘ - 2 s _ - 3 0/7/65 ‘0’ 10/27/65 8/13/65 5/ /, 30/ 6/6 q, .r l2 - _ ‘ H _ DH No.8 5 ,4, I I I I I I I I I I 00 200 300 400 500 600 700 300 900 I000 I I oo TEMPERATURE ”C 0 I I I I I I I I I 0 2 E -I A . a - . 2 3 . «3 6/30/65 10 - 6/16/65 6/7/65 ”I . . u . DH No.9 - 5 M I I I I I I I I I I00 200 300 400 500 $00 700 800 900 I 000 I I00 TEMPERATURE °C DEPTH, IN METE RS DEPTH, IN METERS DEPTH, IN METE RS F FIGURE 27.—Uncorrected temperature profiles. A, DH 3. B, DH 4. C, DH 5. D, DH 6. E, DH 8. F, DH 9. G, DH 10. H, DH 11.], DH 12. J, DH 13, 14, and 16. K, DH 17.L, DH 20. M, DH 21.N, DH 22. 0, DH 23. P, DH 23 uncorrected. Q, DH 24. R, DH 24 uncorrected. S, DH 68—1. Temperature in °C, depth in feet. Each profile is represented by a simple set of symbols. COOLING AND CRYSTALLIZATION OF THOLEIITIC BASALT, 1965 MAKAOPUHI LAVA LAKE, HAWAII TEMPERATURE °C 0 I I I l I I l l I 0 2 _ .1 A - é ' - 2 2 {1" ,. ¥ : E 3 ” Z ‘ f 35 - 3 K 5" 2: D 10 _ 12 - < ‘ u _ DH No. 10 - 5 lb 4 I I I I I I I I 100 200 300 400 500 500 700 800 900 1000 I 100 TEMPERATURE °C G o I I I I I I I I I 0 _ I _ 2 3 3‘1 ._ I“ W E 2 z z E‘ E L E 3 a; 3 o - A 9, cr / ’ T; ’17 _ DH No.11 9, o / 14 , a '1: q, Ir - 5 ‘6 I I I I I I I I I 100 200 300 400 500 600 700 300 900 1000 1100 TEMPERATURE °C H 2 ' ' i l V T I 6/7/55 - I a/ao/as A , . 5 - \ _ 2 \ b/Ia/os s - . \ \. Y ‘ 3 \ Io - O u, I: " E III . w 1‘ —\ 2 z '1 - In 1 E. E \.\ _.=: L II. 1, w ° 14 » .\ :3 IL I I 16 - W .0 11] .o « 6 20 - . BASE or LAKE _ 7 22 I I I I I I I00 200 300 A00 500 600 700 800 DEPTH, IN FEET DEPTH, IN FEET DEPTH, IN FEET 0 I I I I 1 I I I I 0 ~ 1 - 2 < 3 A u. DIINII.I:I,I4,ANDI6 l6 '6 '6 10/27/55 8/l7/65 - 5 M I I I I I I I I I 100 200 300 400 500 600 700 500 900 1000 1100 TEMPERATURE 'C 0 I r I l I I I I I 0 - I — 2 - 3 10/25/65 8/25’°5 12 _ _ 4 “L ‘31 14 _ DH No 17 aw 7/21/65 - 5 M I I I I I I I I I 100 200 300 400 500 600 700 800 900 1000 TEMPERATURE °C 0 I I l I I I I I I O 92.4" . K 2 >\ ‘ l 4 - ° ' — 2 3 _ - 3 lo I. 12 - A u on NII. 20 4: < s ,6 I I I I I "II 100 200 300 A00 500 600 700 900 1000 I I 00 I 200 TEMPERATURE °C FIGURE 27.—Continued. DEPTH, IN METERS DEPTH, IN METERS DEPTH, IN METE RS DEP‘I’H, IN FEEY DEPYH, IN FEET FIGURES 24—28; TABLES 24—29 55 FIGURE 27.—Continued. ‘0 I I I I I I I I I O 16 I I I I o ,. 97” Q- Is I- . o - l 1 4 2° ’ \ - 7 o q 2 2 22 - \ Iu ._ III 2 I B Z 24 - a :5 _ 3 E D 26 - IO I3 I; III E 9 m z 25- 2 I2 . 4 1 z 3-: ". L I 3 E 14 Du NII. 2I / 30 lo 3 0. ‘3: ‘5‘ - 5 32 - Ho . I I 100 200 300 400 000 1100 M .. 34 . o I I I I I I I I | VT 3" II 2 . " _ I 3| _ DH No. 23 UNCORRECTED PROFILES - 13 40 I I I I I I I ‘00 500 600 700 800 900 IOOO 1 100 1200 6 _ _ 2 2 TEMPERAYURE °c P III ._ III 2 a - z ‘. o I I I I I I I I I 0 I - 3 E 3 IIo - 7 . 24 — I I2 . _ ‘ 4 I‘ _ on No. 22 / Ii ° « 2 / J/ D- 4 ‘II ‘ 5 a '6 I I I I I I I “- 100 200 300 400 500 e00 700 300 900 I000 Moo 3 3 TEMPERATURE °c N E E - 3 10 o I I I I I I r I I 0 12 2 - .\ _ 4 \ \ - I \\ 4 - \ H \\ DH No 2A \ 5 \ \ ,6 I I I I I I 6 - \\ . 2 g 100 200 300 400 500 coo 700 800 900 moo IIOO III \ m TEMPERATURE °C \ 2 \\ z a - \ '. ,1. 319/ ‘ 3 fi 0 IO - 12 - - A u . on No. 23 5 ,6 I I I I I I I00 200 300 400 500 500 700 900 1, 900 I000 IIOO 5’, TEMPERATURE °c a, DEPTH, IN METERS COOLING AND CRYSTALLIZATION OF THOLEIITIC BASALT, 1965 MAKAOPUHI LAVA LAKE, HAWAII DEPTH, IN FEET DEPTH, IN FEET 22— 24— 20- 28- 30- 34- 36- 38' A2 DH No.24 UNCORRECTED PROFILES 6/2/66 6/22/é7 DEPTH, IN METERS l0/lJ/66 F/IJ/M I I I I I I MX) 600 700 800 900 '000 1100 1200 TEMPERATURE “C R I I I l I I I 5 DEPTH, IN METE RS TEMPERATURE 'C S FIGURE 27.—Continued. FIGURES 24—28; TABLES 24—29 TEMPERATURE “C 750 900 350 900 950 1000 1050 1 100 1 150 ‘IB I I I I I I I - 6 I .20 - _ ’7 22 _ 24 _ — 8 26 __ .. III-I III “-1 - 9 Z 29 _ f l- . I- I“ O 30 _ . - 10 32 _ DH No. 23 2/14/66 PI-Pthm _ 11 34 ' I =MEASURED NOT REVERSED o =THEORETICAL 3a _ CORRECTION:VARIABLE PROFILE N01 USED I - I2 1 38 I l L I I I I TEMPERATURE °C ‘ 75° goo 350 900 950 1000 1050 1100 1150 I8 I I I I I I I - 6 20 - I 7 22 _ 24 _ 8 2e _ _z_ 23 _ .E‘ E ‘ 30 - - IO 32 _ DH No.23 2/18/66 PI-PIRIIm I=MEASURED PROFILE REVERSED o :THEORETICAL I I 34 _ CORRECTION=14° TO 1100' 20° ABOVE 1100“ 36 _ PROFILE USED. CORRECTION: +14° TO 1100" +18“ ABOVE 1100' - 12 38 I I 1 I I 1 I DEPTH, IN METERS DEPTH, IN METERS DEPTH, IN FEET DEPTH, IN FEET 750 TEMPERATURE °C 5 II50 22. 24- 26- 28- 30- 32- 34- 36- 38 800 850 900 950 I 000 I 050 l I 00 I I I I I I DH No. 23 2/23/66 Pv-PIRhw I =MEASURED PROFILE REVERSED O = THEORETICAL CORRECTION = + I 7° PROFILE USED I I 1 4L I I I ‘6 750 TEMPERATURE °C 800 850 900 950 I 000 1050 I I 00 1 I I150 18 20- 22_ 24.. 26 28 30- 32- 34- 36- 30 I I I I DH 23 3/I/66 Pv-PIRIIIO I=MEASURED PROFILE NOT REVERSED O = THEORETICAL CORRECTION : 3 I ° PROFILE ERRATlC-USED AS DRAWN -.6 O FIGURE 28.—Temperature profiles for DH 23 with correction for thermocouple contamination.A, 2/14/668, 2/18/66. C, 2/23/66. D, 3/1/66. E, 3/3/66. F, 3/8/66. G, 3/22/66. H, 3/28/66. 1, 4/4/66. J, 4/25/66. K, 5/6/66. Temperature in °C, depth in feet. A=difference between measured and theoretical temperature pro- I = measured temperatures. q=theoretica1 temperatures based on uncorrected profiles obtained before and after the measurement date. file. 7_ DEPTH, IN METERS DEPTH, IN METERS 58 DEPTH, IN FEET DEPTH, IN FEET COOLING AND CRYSTALLIZATION OF THOLEIITIC BASALT, 1965 MAKAOPUHI LAVA LAKE, HAWAII TEMPERATURE °C 750 800 850 900 950 1000 1050 1 I00 1 150 18 I I I I I I - 6 20 - - 7 22 _ 24 - 8 26 _ u) 8 IN I- I“ _ 9 2 28 _ Z 1‘ I- I. 3 30 _ DH 23 3/3/66 PI-PIRhlo “0 32 F I =MEASURED. PROFILE REVERSED ‘ :25, o :THEORETICAL 34 CORRECTION:31° TO 1070" ' " 25° ABOVE 1100“ 36 _ PROFILE USED WITH 31° CORRECTION - I2 38 I I I I I I I - E TEMPERATURE °C 750 800 850 900 950 Iooo 1050 I 100 I 150 18 I I I I I I I - 6 20 _ - 7 22 _ 24 _ 8 26 _ V) 8 MI I- W - 9 E 23 _ Z :I:‘ I— n. E 30 _ - 10 32 _ DH No,23 3/8/66 PI- PI Rhlo 34 I=MEASURED PROFILE REVERSED '1 ' o :THEORETICAL CORRECTION: +2e° 36 _ PROFILE USED 12 38 L I I I I I I F DEPTH, IN FEET DEPTH, IN FEET TEMPERATURE °C 750 800 850 900 950 1000 1050 I I00 I 150 18 I I \ I I I I - 6 20 _ — 7 22 _ 24 _ - 8 26 _ A :17" - 9 28 _ -17° 30 _ :17“ - 10 DH No. 23 3/22/66 Cr-AI 32 _ o :THEORETICAL A I:MEASURED SINGLE JUNCTION Ic. NOT REVERSED :I9° CORRECTION:18° PROFILE USED 11 34 _ ' o :MEASURED 5 JUNCTION Ic. NOT REVERSED CORRECTION VARIABLE 36 ,_ PROFILE NOT USED - 12 38 I 1 l I I I L TEMPERATURE °C 750 800 850 900 950 1000 1050 I 100 1150 18 I I I I I I - 6 20 - - 7 22 ,_ 24 _ _. B 26 _ - 9 25 - A215“ 30 _ - 10 32 _ DH No. 23 3/28/66 Cr-Al .2THEORETICAL O 5 JUNCTION Cr»AI I=MEASURED PROFILE REVERSED - 11 34 _ I CORRECTION: +16° TO THEORETICAL 19" TO 5 JUNCTION Cr-AI 36 _ PROFILE USED WITH 19° CORRECTION - 12 38 I 4 I I I I I FIGURE 28,—Continued. DEPTH, IN METERS DEPTH, IN METERS DEPTH, IN FEET TEMPERATURE °C FIGURES 24—28; TABLES 24—29 59 TEMPERATURE °C DEPTH, IN METERS 750 800 850 900 950 1000 1050 I 100 I I50 750 800 850 900 950 1000 1050 I 100 1150 I8 I I r T I I I - 6 18 I I I I I r I _ 6 20 .. 20 _ _ 7 _ 7 22 - 22 - ‘24 _ .. _ 8 24 _ a ‘26 _ In 26 - 8 w l- ,_ Lu “I, II“ — 9 i “ — 9 Z 23 _ Z 1 28 ~ ~ 1: a I: a. II” 3 n 30 _ 30 - - 10 - I0 DH No. 23 4/4/66 Cr-Al ‘32 _ 0 =THEORETICAL A 5 JUNCTION Cr-AI 32 _ I =MEASURED. PROFILE REVERSED DH No.23 4/25/66 Cr—AI - 1| .=THEORETICAL NOT REVERSED I] 34 _ NO CORRECTION TO 35-6 FEET OFF BY 34 _ I = MEASURED 10° AT T) IIOO“ IN REVERSED SEGMENT CORRECTION = I O—I 4° 35 _ PROFILE USED AS DRAWN. COULD BE As 36 __ moms USED 14° CORRECTION MUCH AS 10" LOWER 12 12 I 38 I l I I I I I I 38 I I I I I I I TEMPERATURE °C 750 800 850 900 950 1000 1050 1100 1150 I8 T I I I I I I - 6 20 — - 7 22 I- 24 — - 8 26 .— a, K E I“: W In M. - 9 2 % 28 - z '3': 2‘ l v- IIU t o 3 30 _ DH Na 23 5/6/66 Cr—AI A:IO° 0 — I O :THEORETICAL PROFILE 3; .— I :MEASURED PROFILE A:12° O :SILVER CALIBRATION - II 3‘ _ CORRECTION=15° FROM Aq CALIBRATION =IO—IS° FROM THEORETICAL 3o ._ PROFILE USED WITH 15" CORRECTION - I2 38 I I I 1 I 1 L FIGURE 28.—Continued. 60 COOLING AND CRYSTALLIZATION OF THOLEIITIC BASALT, 1965 MAKAOPUHI LAVA LAKE, HAWAII TABLE 24.—Core logs for drill holes 1—24, 1965—66 Drill hole Core interval f ) Temperature (°C) Core recovery Chemical No. Date weight (g) percent analysis Comment 1 4—19—65 0—0.02 M—1—1 yes Surface glass. 0—0.8 0—260° 283 70 M—1—2 Mast hole, 1—11/2 in. piece. 0—1.4 0—400 28 25 M—1—3 1.4—2.9 400—655 49 20 M—1—4 Three 1/2 in. pieces. 29—495 655—960 0 495—61 960—1,040 85 60 M—1—5 Two 1/2 in. pieces. 6.1—7.1 1,040—1,100 47 25 M—1—6 yes 4—23—65 10.8—11.4 z1,130 M—1—7 yes Melt on bottom of steel rod. 2 4—19—65 0—1 0—310 56 45 M—2—1 One 3 in. piece. 1—2 310—505 34 25 M—2-2 One 2 in. piece. 2.0—5.1 505—940 0 5.1—6.1 940—1,04O 68 50 M—2—3 6.1—7.1 1,040—:1,100 18 20 M—2—4 yes 4—21—65 18 ? M—2—5 Ooze in drill pipe probe. 3 4—22-65 0—0.02 M—3—1 Glassy surface of lake. 0.02—0.45 0—160 137 55 M—3—2 yes Two 1/2 in. piece from top. 0.45—0.7 160—240 98 5O M—3-3 0.7—0.9 240—280 93 60 M-3—4 4—27-65 0—108 0—300 96 80 M—3—5 One 1/2 in. piece from upper half of interval, 4>1/2 in. 1.08—1.87 300—455 63 55 M—3—6 1.87—2.87 455—600 62 45 M—3—7 2.87—3.8 600—720 0 3.8—4.85 720—850 9 5 M—3—8 4.85—5.83 850—940 65 ca. 30 M—3-9 All small pieces. 5.83—11 EMU—1,130: 0 _ 4—28—65 11 1,130: 90 M—3—10 Melt sample oozed 1nto base of stainless steel probe at 11 ft.; collected on old thermocouple steel at 9% ft. Water quench. 4 4—22—65 0—0.45 179 70 M—4—1 yes 3 in. piece at top Mast 0.45—0.8 162 75 M—4—2 One 1/2 in. piece- h 1 5—17—65 0—1.0 30—250 80 8O M—4—3 bottom 0 9' One 1 in. piece from upper 11/2 ft. Three 1/2 in. pieces. 1.0—2.0 250—410 96 80 M—4—4 2.0—3.0 410—540 33 20 M—4—5 3.0-4.0 540—650 19 10 M—4—6 4.0—5.0 650—750 61 40 M—4—7 One 1/2 in. piece. 5.0—6.0 750—840 47 20 M-4—8 6.0—7.0 840—920 18 10 M—4—9 7.0—8.0 920—990 26 10 M—4—10 8.0—9.5 990—1,080 0 0 Crust melt interface at 9.3 ft. 5 4—22—65 i 0—.45 <180 216 95 M—5—1 2 in. piece—top, 1 in. piece— bottom, 11/2 in. piece—middle. .45—.65 180—230 102 95 M—5—2 2 in. piece. 65—85 230—280 127 95 M_5_3 21/2 in. piece, one 1 in. piece from top. 5—19—65 0—1.0 0—240 62 50 M—5—4 1.0—2.1 240—430 122 80 M—5—5 Three 1/2 in. pieces. 2.1—3.1 430—560 96 75 M—5—6 One 1 in. piece. 3.1—4.1 560—665 65 40 M—5—7 Two 1/2 in. pieces. 4.1—5.1 665—760 0 5.1—6.1 760—850 152 80 M—5—8 One 1/2 in. piece. 6.1—7.1 850—930 96 55 M—5—9 yes One 1 in. piece, two 1/2 in. pieces. 7.1—8.1 930—990 56 35 M—5—10 8.1—9.1 990—1,040 . 126 80 M—5—11 Three 1 in. pieces. 9.1—10.1 1,040—z1,060 18 10 M—5—12 One 1 in. piece. 10.1—11.0 1,060—1,070 48 30 M—5—13 One 1 in. piece. 6 4—23—65 0—.4 <170 125 60 M—6—1 .4—6 170—220 104 80 M—6—2 1.6—.75 220—250 92 90 M—6—3 1 in. piece from bottom. 5—17—65 0—1.0 <240 109 85 M—6—4 11/2 in. piece—top of hole; 2 in. piece—just below. 1.0—1.9 240—400 58 55 M—6—5 One 3 in. piece. 1.9—2.9 400-540 98 75 M—6—6 2.9—3.9 540—645 81 50 M—6—7 One 1/2 in. piece. FIGURES 24—28; TABLES 24—29 TABLE 24.—Core logs for drill holes 1-24, 1965—66—Continued 61 Drill hole No. Core interval Temperature (“C Core recovery Sample No. Chemical Date (ft) ) weight (g) F percent analysis Comment 6—Con. 3.9—4.9 645—740 96 80 M—6—8 One 1/2 in. piece. 4.9—5.9 740—830 109 60 M—6—9 5.9—6.9 830—915 21 25 M—6—10 6.9—7.9 915—980 88 7O M—6—11 7.9—8.9 980—1035 30 2O M—6—12 7 5—5—65 0—1 30-100 24 25 M—7—1 Much core left in hole; one 3A in. lece. 1—1.85 100—120 69 60 M—7—2 Olfe 1/2 in, piece, 1.85—2.85 120—275 60 50 M—7—3 One 3A; in. piece; top marked. 285—35 275—460 35 50 M—7—4 One 1/2 in. piece; top marked. 5—6—65 3.5—4.9 460—725 160 65 M—7—5 One 21/2 in. piece; top marked; one 1 in. piece; top marked. 4.9—5.45 725—800 30 50 M—7—6 All small pieces. 5.45-7.9 800—1,030 0 7.9—8.9 1,030—1,0851- 11 5 M—7—7 One good glassy piece. 8.9—15 1,085:1,130 O 5—7—65 15—20 1,130—1,140 M—7—8 Melt oozed into steel probe for ball experiment. Collected on push rod. 8 5—24—65 0—.45 <180 47 25 M-8—1 .45—.7 180—195 47 20 M—8—2 .7—.8 195—210 58 50 M—8—3 0—1 <240 58 50 M—8—4 One 1 in. piece; one 1/2 in. piece. 1—2 240—385 28 30 M—8—5 2—3 385—510 11 10 M—8—6 3—4 510—625 11 5 M—8—7 4—5 625—720 9 5 M—8—8 5—6 720—805 0 6—7 805—885 27 10 M—8—9 7—8 885—955 49 25 M—8—10 8—9 955—1,025 26 5 M—8—11 9 5—24—65 0—.2 <100 62 7O M—9—1 One 11/2 in. piece. .2—.4 100—140 93 75 M-9—2 One 2 in. split piece. .4—.65 140—190 156 100 M—9—3 One 2 in. piece. .65—.8 190—210 69 80 M—9—4 One 1 in. piece. 0—1 <240 56 50 M—9—5 Three % in. pieces. 1—2 240—385 47 40 M—9—6 One 1/2 in. piece. 2—3 385—510 39 30 M—9—7 Two 1/2 in. pieces. 3—4 510—625 43 30 M—9—8 One 1/2 in. piece. 4-5 625—720 121 80 M—9—9 One 1/2 in. piece. 5—6 720—805 125 75 M—9—1O 5—26—65 6—7 805—885 46 30 M—9—11 7—9 885—1,015 0 9—10 1,015—1,055 43 25 M—9—12 10—10 5 1,055—1,070 46 60 M—9—13 10 5—26—65 0—.35 <130 134 75 M-lO—l 2 in. piece. 35—55 130—165 98 85 M-10—2 2 in. piece Mast hole. .55—.7 165—195 91 80 M—10—3 2 in. piece. 0—1.1 <26O 42 50 M—10-4 One 1/2 in. piece. 1.1—2.0 260—385 34 30 M—10—5 2.0—3.0 385—510 8 10 M—10—6 3.0—4.0 510—625 12 5 M—10—7 4.0—5.0 625—720 51 50 M—10—8 5.0—6.0 720—805 30 2O M—10—9 6.0—7.0 805—885 43 10 M—lO—lO yes 7.0—8.0 885—955 45 40 M—10—11 8.0—9.0 955—1,010 13 5 M—10—12 9.0—10.0 1,010—1,050 65 50 M—10—13 10.0—10.6 1,050—1,070 63 7O M—10—14 Three 1/2 in. pieces. 11 5—26—65 0—.4 <80 130 90 M—11—1 Two 1% in. pieces. .4—.6 80—100 85 75 M—11—2 2 in. piece l Ma“ 11016- .6—.8 100—130 90 80 M—11—3 0—1.0 <150 119 90 M—11—4 Numbered from top: two 11/2 in. pieces; four 1 in. pieces; two 1/2 in. pieces. 1.0—2.0 150—255 84 60 M—11—5 Four 1 in. pieces. 2.0—3.0 255—360 43 30 M—11—v6 One 1 in. piece. 3.0—4.0 360—465 57 35 M—11—7 One 1/2 in. piece. 62 COOLING AND CRYSTALLIZATION OF THOLEIITIC BASALT, 1965 MAKAOPUHI LAVA LAKE, HAWAII TABLE 24.——Core logs for drill holes 1—24, 1965—66—Continued Core recovery Drill hole Core interval Temperature Sample Chemical No. Date lft) (°Cl Weight lg! percent N0. analysis Comment 11—Con. 4.0—5.0 465—560 84 5O M—11—8 One 1V2 in. piece. 5.0—6.0 560—645 101 50 M—11—9 yes Three 1 in. pieces. 6.0—7.0 645—725 90 50 M—11—10 yes One 1 in. piece. 7.0—8.0 725—795 71 5O M—11—11 yes 8.0—9.0 795—865 95 40 M 11—12 yes One V2 in. piece. 9.0—10.0 865—920 80 50 M—11—13 Two V2 in. pieces. 10.0—11.0 920—980 44 25 M—11—14 yes One V2 in. piece. 110—11 9 980—1,030 27 15 M—11—15 One V2 in. piece. 6—1—65 119—12 9 1,025—1,070 203 100 M—11—16 Three 2 in. pieces,two V2 in.pieces. 12 6—1—65 0—0.2 1 22 30 M—12—1 .2—.5 26 30 M—12—2 0—.5 81 100 M—12—3 One 5V2 in. piece. 5—105 0 1.05—2.3 82 50 M—12—4 Five V2 in. pieces. 2.3—3.3 28 25 M—12—5 One V2 in. piece. 3.3-4.3 0 4.3—6.3 225 75 M—12—6 One 1V2 in. piece. 6.3—8.3 109 40 M—12—7 One V2 in. piece. 8.3—8.75 41 50 M—12—8 One V2 in. piece. 6—4—65 875—97 67 20 M—12—9 Four V2 in. pieces. 97—12. 222 50 M—12—10 Seven 1 in. pieces, many V2 in. pieces. 12.3—14.0 97 20 M—12—11 Two 1 in. pieces, many V2 in. pieces. 14.0—15.9 0 15.9—17.0 65 50 M—12—12 Four V2 in. pieces. 17.0—19.0 54 20 M—12—13 One 1 in, piece, one V2 in. piece. 19.0—21.0 0 21.0—24.0 8 <5 M—12—14 One V2 in. piece and fragments (lava or talus?). 13 6—7—65 0—0.4 <120 71 40 M—13—1 0.4—0.65 120—170 99 75 M—13—2 } Mast hole. 0.65— .85 170—210 110 95 M—13—3 2 in. piece. 0—0.9 <220 92 80 M—13—4 Two 1 in. pieces, two V2 in. pieces. 0.9—2.15 220—380 101 70 M—13—5 One 1 in. piece, One V2 in. piece, 2.15—3.1 380—500 74 50 M—13—6 3.1—4.1 500—600 75 50 M—13—7 4.1—5.1 600—690 132 85 M—13—8 5.1—6.1 690—775 133 80 M—13—9 6.1—7.1 775—845 104 70 M—13—10 7.1—8.1 845—915 90 60 M—13—11 yes 8.1—9.1 915—980 48 30 M—13—12 yes 9.1—10.1 980—1,025 128 75 M—13—13 yes One in. piece, two V2 in. pieces. 10.1—11.1 1,025—1,080 133 75 M—13—14 yes One 2 in. piece, three 1 in. pleces. 11.0—118 1,080—1,100 51 4O M—13—15 Two V2 in. pieces from top of interval. 14 6—11—65 0—0.5 <160 92 40 M—14—1 0.5—0.8 160—200 94 50 M—14—2 1‘ Ma“ 11013 6—16—65 0—0.85 <200 47 60 M—14—3 Two 1 in. pieces,three V2 in.pieces. 0.85—20 200—350 42 35 M—14—4 Two V2 in. pieces. 2.0—3.0 350—460 36 40 M—14—5 3.0—4.0 460—565 30 30 M—14—6 4.0—5.0 565—660 95 80 M—14—7 5.0—6.0 660—765 18 2O M—14—8 One V2 in. piece. 6.0—7.0 765—815 72 60 M—14—9 , . _ . 7.0—8.0 815—885 82 70 M—14—10 One plelce With large 011v1ne 8.0—9.0 885—950 58 50 M—14—11 crysm ' 9.0—10.0 950—1,010 92 60 M—14—12 10.0—11.0 1,010—1,060 26 20 M—14—13 11.0—11.9 LOGO—1,070 17 10 M—14—14 15 6—11—65 0—025 <100 62 60 M—15—1 Broken 0.2e0.55 100—130 167 100 M—15—2 Irregular 3 in. pieces MaSt 0.55m 7 130.170 86 100 M—15—3 Good 2 in. pieces hole- 6—16—65 0—1.0 <210 35 30 M—15—4 One V2 in. piece. 1.0—2.0 210—350 12 10 M—15—5 2.0—2.9 350—450 29 20 M—15—6 2.9—4.0 450—565 48 50 M—15—7 One V2 in. piece. 3.85—4.85 550—645 92 70 M—15—8 One V2 in. piece. 4.85—5.95 645—740 31 30 M—15—9 FIGURES 24—28; TABLES 24—29 TABLE 24.—Core logs for drill holes 1—24, 1965—66—Continued 63 Drill hole Core interval Temperature (°C) Core recovery Sample No. Chemical No. Date (ft) weight (g) percent analysis Comment 16 6—30—65 0—0.4 <100 135 70 M—16—1 Mast hole. 0.4—0.7 100—140 125 80 M—16—2 Mast hole. 0—1.8 <300 75 45 M—16—3 2 in. piece, two 1 in. pieces. 1.8—3.0 300—450 18 10 M—16—4 3.0—3.85 450—520 8 5 M—16—5 3.85—4.8 520—600 115 85 M—16—6 1 in. piece. 7—12—65 48—592 600—670 78 55 M—16—7 1/2 in. piece. 5.92—6.92 670—740 155 60 M—16—8 Two 1/2 in. pieces. 6.92—7.92 740—800 76 60 M—16—9 1 1/2 in. piece. 7.92—8.92 800—860 84 6O M—16—10 8.92—9.92 860—910 128 80 M—16—11 Two 1/2 in. pieces. 992—1092 910—965 9 10 M-16—12 10.92—11.92 965—1,010 30 10 M—16—13 11.92—12.92 1,010—1,055 84 60 M—16—14 One 1 in. piece, one 1/2 in. piece. 7—15—65 1285—139 1,055—21,100 43 20 M—16—15 17 7—12—65 0—0.48 <100 146 85 M—17—1 0.48—0.75 100—150 133 85 M—17—2 7—15—65 0—0.9 <190 106 95 M—17—3 One 1 in. piece, three 14 in. pieces, three 1/2 in. pieces. 0.9—1.5 190—270 104 90 M—17—4 Three 1% in. pieces, one 2 in. piece. 1.5—2.5 270.370 49 25 M—17—5 One 1 in. piece. 2.5—3.5 370—460 49 25 M—17—6 Two 1/2 in. pieces. 3.5—4.5 460—560 13 10 M—17—7 4.5—5.5 560—640 50 20 M—17—8 Two 1/2 in. pieces. 5.5—12.5 640—1,030 0 12.5-13.5 1,030—1,070 58 20 M—17—9 Three 1/2 in. pieces. 13.5—14.5 1,070—1,100 23 5 7—16—65 14.5—16.1 0 17A 7—28—65 0—1.0 <190 26 20 M—17A—1 1.0—2.0 190—300 94 75 M—17A—2 One 1/2 in. piece. 2.0—3.0 300—400 38 30 M—17A—3 Two 2 in. pieces. 3.0—4.0 400—500 0 4.0—5.0 500—580 90 70 M—17A—4 5.0—6.0 580—650 113 80 M—17A—5 Three pieces>1/2 in. 6.0—7.0 650—710 98 70 M—17A—6 One piece>1/2 in. 7.0—8.0 710—770 88 70 M—17A—7 80—885 770—820 54 40 M—17A—8 885—100 820—880 10 10 M—17A—9 18 7—12—65 0—0.5 113 60 M—18—1 - 050.57 39 90 M_1&2 } Mast hole. Slte abandoned. 19 7—12—65 0—0.4 169 88 M—19—1 . 0.4—0.8 226 88 M—19—2 }Mast hole. Slte abandoned. 20 8-2—65 0—0.90 <190 17 15 M—20—1 0.9—1.9 190—300 35 30 M—20—2 1.9—3.1 300—410 10 10 M—20—3 3.1—4.1 410—500 24 10 M—20—4 4.1—5.1 500—570 102 70 M—20—5 One 1/2 in. piece. 5.1—6.1 570—645 76 60 M—20—6 One 1 in. piece. 6.1—7.1 645—700 0 7.1—8.1 700—770 45 40 M—20—7 8.1—9.1 770—830 0 9.1—10.1 830—880 81 50 M—20—8 10.1—12.1 880—970 0 12.1—13.1 970—1,010 82 60 M—20—9 . . 13.1—14.1 1,010—1,045 104 70 M—20—10 One 1 m. plece, four l/2 in. pieces. 14.1—21 0 21-23 1,130—1,135 M—20—11 Melt on pusher rod above paddle. 21—23 1,130—1,135 M—20—12 Melt on paddle wheel. 21—23 M—20—13 Melt in stainless steel casing. 21 8—17—65 0—0.95 <175 112 95 M—21—1 Four 1/2 in. pieces, three 1 in. pieces, one 2 in. piece. 0.95—1.95 175—280 113 90 M—21—2 Three 1/2 in. pieces, two 1 in. pleces. 1.95—2.95 280—380 70 60 M—21—3 Five pieces>1/2 in. 2.95—3.95 380—465 33 40 M—21—4 3.95—4.95 465—550 114 I 80 M—21—5 64 COOLING AND CRYSTALLIZATION OF THOLEIITIC BASALT, 1965 MAKAOPUHI LAVA LAKE, HAWAII TABLE 24.—Core logs for drill holes 1—24, 1965—66—Continued Core recovery Drill hole Core interval Temperature Sample Chemical No. Date (ft) (”C) weight (g) percent No. analysis Comment 21—Con. 4.95—5.95 550—620 104 80 'M—21—6 - 5.95—6.95 620—685 94 80 M—21—7 Three 1/2 in. pieces. 6.95—7.95 685—750 55 40 M—21—8 7.95—8.95 750—800 0 8.95—9.95 800—845 79 50 M—21—10 995—1095 845—890 52 40 M—21—11 10.95—11.95 890—935 59 4O M—21—12 yes 11.95—12.95 935—980 101 80 M—21—13 One 1/2 in. piece. 12.95—13.95 980—1,010 98 80 M—21—14 Three 1/2 in. pieces. 13.95—14.95 1,010—1,040 5 3 M—21—15 14.95—15.95 1,040—1,064 0 15.95—16.95 1,065—1,085 8 3 M—21—17 Base of crust 160 ft. 16.95—18.95 21,085—z1,130 0 18.95—19.6 =1,130—~1,140 38 20 M—21—20 17—18 1,095—1,105 194 20 M—21-21 Ooze from drill steel, three 3-in. pieces, one 2 in. piece, one 1 in. piece. 18 1,105 26 M—21—22 Ooze-end of stainless steel rod. 8—30—65 20—22 1,115—1,125 M—21—23 Melt from outside of drill steel. 19—22 1,110—1,125 M—21—24 yes 3 ft of melt in drill steel. 9—16—65 154—1605 55 M—21—25 yes One 2 in. piece, good ooze sam- p e. 9—27—65 25—2675 M—21—26 yes Dense glassy melt surrounding long tubular voids—collected in ceramic tube and stainless , steel casing. Four pieces >% 1n. 29—305 M—21—27 yes ‘Melt in ceramic and in steel cas- mg. 22 11—1—65 0—0.9 39 45 M—22—1 Two pieces>‘/2 in. 0.9—1.9 71 55 M—22—2 Three pieces>1é in. 1.9—3.0 0 l 3.0—4.0 26 20 M—22—3 4.0—5.0 95 80 M—22—4 5.0—6.0 465—530 69 60 M—22—5 l 6.0—7.0 530—580 45 5O M—22—6 ‘ 7.0—8.0 580—635 53 50 M—22—7 , 8.0—9.0 635—690 0 90—100 690—740 85 7O M—22—8 10.0—11.0 740—780 27 3O M—22—9 11.0—12.0 780—820 78 60 M—22—10 yes One piece>1/2 in. 12.0—14.0 820—900 0 14.0—15.0 900—935 132 80 M—22—11 Six pieces>‘/2 in. 15.0—16.0 990—1,020 11—9—65 16.0-17.0 975—1,010 100 M—22—12 17.0—18.0 1,010—1,035 0 18.0—19.0 1,035—1,060 50 M—22—13 19.0—20.0 LOGO—1,075 20 M—22—14 Base of crust 19.7 ft. 20.0—21.0 1,075—1,095 O 21.0 1,095 M—22—15 End of sampler. 20.5—21.6 ~1,095 M—22—16 Ooze 0n thermocouple. 21.0 1,095 M—22—17 Plug at 21.0 ft (ooze). 21—23 1,095—1,110 M—22—18 Melt in 2 ft stainless. 23 12—1—65 0—1.2 <100 85 80 M—23—1 Four pieces>1é in. 1.2—2.2 <100 60 6O M—23—2 Two pieces>1/2 in. 2.2—3.2 <100—150 83 65 M—23—3 Two pieces>1/2 in. 3.2—3.6 150—250 28 40 M—23—4 12—6—65 3.5—5.0 250—360 130 80 M—23—5 One piece>1/2 in. 4.9—5.9 360—425 95 80 M—23—6 One piece>‘/2 in. 5.9—7.9 425—540 195 75 M—23—7 Four pieces>1/2 in. 7.9—8.9 540—600 110 80 M—23—8 Several>1/2 in. 8.9—9.9 600—670 85 80 M—23—9 Two pieces>1/z in. 9.9—10.9 670—730 23 10 M—23—10 One piece>1/2 in. 10.9—11.9 730—780 31 20 M—23—11 11.9—12. 780—820 64 50 M—23—12 12.9—13.9 820—870 75 50 M—23—13 One large piece. 13.9—15.9 870—950 41 20 M—23—14 15.9—16.9 950—975 80 60 M—23—15 Two pieces>1/2 in. 12—13—65 16.9—17.9 975—990 60 60 M—23—16 FIGURES 24—28; TABLES 24—29 TABLE 24.—-Core logs for drill holes 1—24, 1965—66—Continued 65 Core recovery Drill hole Core interval Temperature Sample Chemical ‘ No. Date (ft) PC! weight Igl percent No. analysis Comment 23—Con. 17.9—18.65 990—1,005 50 40 M—23—17 One piece>1/2 in. 19.7—20.9 LOGO—1,085 35 25 M—23—18 1—19—66 19.4—21.0 LOGO—1,060 80 M—23— 19 yes Ooze drilled out as four separate pieces: a, b, c, d. May have come in from different hori- zons. 24.0+ 1,095+ M—23—21 yes Small piece lodged in bit. 23.5 1,090 M—23—22 Melfain bit and on stainless steel ro . 2—3—66 210—22 0 1,030—1,060 M—23—23 Small piece in bit. 24.0+ 1,085+ M—23—24 Pushed 3—4 ft into melt which was collected in bit. 24 5—2—66 0—4.9 <290 400 51 5—23—66 4.9—7.0 280-390 90 38 7.0—9.0 390—480 150 80 9.0—11.0 480—580 140 80 11.0—13.0 580—670 60 40 130—15 0 670—740 170 60 150—17 0 740—815 145 50 2M—24—16.9 17.0—19 0 815—870 0 0 190—21 0 870—950 105 60 M—24—19.1 M—24—20.4 21.0—23.0 950—1,005 180 75 M—24—21.1 M—24—22.0 M—24—22.9 23.0—25.0 1,005—1,045 245 90 M—24—23.1 M—24—24.0 M—24—25.0 25.0—27.0 1,045—1,085 0 0 7—28—66 22.0—25.0 940—1,015 6 5 M—24-1 25.0—27.0 LOIS—1,060 0 O 27.0—29.0 LOGO—1,090 5 M-24—2 yes Melt (in core barrel). 29.0—30.0 1,090—1,100 6 5 M—24—3 Do. 29.0—31.0 1,090—1,110 Hole cased to ~ 40 11;. 11—29—66 30.0—30.2 1,053:3 8 M—24—4 Film of ooze on oxidized ther- mocouple sheath in cased hole. ‘Hole near edge of lake. No way to extrapolate temperature back to the time of drilling. 2Analyzed or thin-sectioned samples from drilling dates 5/2/66 and 5/23/66 coded as follows: Mv24—(depth in feetl, for example. core collected at 21 feet is labeled M—24—21. Core fron redrilling in 7/28/66 numbered consecutively. TABLE 25.——Core logs for drill holes 68-], 68—2 and 69—1; 1968—69 Core recovery 1 Sample Nos. Drill hole Core interval Temperature‘ intervals of for chemical No. Date (ft) (c°C) percent no core (it) analysis Comment 68—1 11/6/68 0—1.0 <100 100 1.0—1.9 <100 100 Sublimates in vesicles. 1.9—3.7 <100 5O 1.9—2.8 68—1—1 (3.7 ft) 3.7—7.3 <100 81 4.7—4.9 5.9—6.4 11/12/68 7.3—10.0 <100 89 8.5—8.8 68-1—2 (8.0 ft) 10.0—14.8 <100 100 68—1—3 (12.0 ft) Thin section (14.8 ft). 14.8—19.8 ~100 100 68—1—4 (17.5 ft) 11/15/68 19.8—23.0 100—200 100 68—1—5 (22.1 ft) 23.0—28.0 200—320 100 68—1—6 (24.0 ft) Thin section (27.4 ft). 68—1—7 (26.0 it) 68—1—8 (27.9 ft) 28.0—28.8 320—350 100 Thin section (28.3 ft). 28.8—29.35 350—390 100 29.35—35.1 390—590 71 scattered 68-1—9 (30.0 ft) Thin section (33.5 ft). 68—1—11 (33.8 it) 11/18/68 35.1—40.0 590—770 57 scattered 68—1—12 (36.0 fl) Core lost in drilling. 68—1—14 (39.8 ft) 40.0—44.2 770—900 0 40.0—44.2 Probably drilled along vertical crack. 44.2—49.5 900—1,030 100 68—1—16 (44.2 ft) 68—1—17 (46.1 ft) 68—1—47.5 (47.5 R) 68—1—18 (48.0 ft) 68—1—19 (49.5 it) 66 COOLING AND CRYSTALLIZATION OF THOLEIITIC BASALT, 1965 MAKAOPUHI LAVA LAKE, HAWAII TABLE 25.—-Core logs for drill holes 68—1, 68—2 and 69—1; 1968—69—Continued Core recovery Sample Nos. Drill hole Core interval Temperature‘ intervals of for chemical No. Date (ft) (c°CI percent no core (ft! analysis Comment 68—1 11/18/68 49.5—54.0 1,030—1,075 18 rob. 68—1—21 (54.0 ft) Thin section (53.5 ft, 54.0 ft). 49.5—53.2 11/20/68 47.9—57.0 -1 ",1 Black sand erupted at top of drill hole throughout entire drilling interval. Collected in bags. 68—2 12/12/68 0—1.0 <100 100 1.0—2.0 <100 100 2.0—3.0 <100 80 2.8—3.0 3.0—8.0 <100 74 scattered 5.9—8.0 ft: vertical fracture in core. in 5.9—8.0 8.0—13.0 <100 100 12/16/68 13.0—17.9 <100 84 scattered in 15.0—17.9 17.9—22.9 <100—200 100 22.9—27.9 200—320 100 27.9—33.0 320—500 84 scattered 68—2—10 (32.0 ft) in 28.3—30.3 ’ 33.0—38.0 500—720 88 36.5—37.1 68—2—13 (38.0 It) 38.0—43.0 720.870 62 39.5—41.4 68—2—15 (42.0 ft) 12/18/68 43.0—44 3 870—900 100 443—49 1 900—1010 0 Core spring left out of core barrel. 443—49 5 900—1,020 56 not known Redrilled—void to 46.0 ft. 49.5—50.0 1,020—1,040 100 50.0—55.2 1.040—1,080 27 scattered 68—2—20 (51.5 ft) Probably crossed crust-melt interface. 52.0—55.2 55.2—59.0 1,080—1,100 low not known 68—2—59 (59 ft) Melt in bit, no core. 69—1 1/22/69 0—1.3 38 0.1—0.4, 0.7—0.8, 0.9—1.3 1.3—2.3 80 scattered 2.3—4.0 59 scattered 4.0—7.7 100 7.7—12.3 81 scattered Thin sections (8.3.ft, 8.4 ft, 9.6 ft, 9.7 ft, 11.6 ft 12.3—15.0 52 scattered Thin section (14.9 ft). 15.0—19.2 52 scattered 19.2—20.0 38 scattered 20.0—24.7 55 scattered Thin section (20.0—22.7 ft, 24.6 ft). 1/28/69 24.7—27.1 83 scattered 69—1—24.7 (24.7 ft) Thin section (24.7 ft, 25.6 ft). 69—1-25.6 (25.6 ft) 27.1—30.0 100 Thin section (28.0 ft, 28.3 ft, 28.7 ft, 29.0 ft, 29.5 ft, 29.8 ft). 30.0—35.0 90 scattered Thin section (31.2 ft, 31.7 ft). 35.0—38.5 100 Thin section (37.5 ft, 37.9 ft). 38.5—41.0 80 scattered Thin section (41.0 ft). 1/31/69 41.0—46.0 30 not known 69—1—41.0 (41.0 ft) Thin section (exact depth not known). 69—1—42.0 (42.0 ft) 46.0—50.9 90 scattered Thin section (47.3 ft, 47.6 it, 48.3 ft, 49.5 ft). 50.9-56.0 76 scattered 69—1—555 (55.5 ft) Thin section (51,0 ft, 52,3 ft, 53.1 ft, 55.1 ft—segregation vein, 55.0 ft, 55.7 ft). 56.0—61.0 0 In melt. 61.0—66.0 0 In melt. Melt flowed into core barrel which was then recovered (fig. 26). ‘Temperatures are not accurately known for drill holes 68—1 and 6&2 because the only temperature profiles showed effects of thermal depression from water introduced during drilling. No temperature data are obtained for 6&1. Where temperatures are given they are probably minimum values. TABLE 26.—Temperature profiles (°C) measured in drill holes 2—24, 68—] [Data for each drill hole given separately, with subheadings giving date (and hour for the earliest temperature profiles) followed in parenthesis by the square root of time in days since formation of the permanent crust on March 19, 1965. Example: 5/24/65 6/7/65 6/16/65 6/30/65] (8.11! (8.93! (9.42} (10.14! Drill hole No. 2 Depth 4/19/65 (5.58) 4/21/65 (572! (ft! 09:55—10:45 14:09 14:45 15:13 15:32 1.0 137: 2.0 334.7 3.0 549.3 FIGURES 24—28; TABLES 24—29 TABLE 26.—Temperature profiles (°C) measured in drill holes 2—24, 68—1—Continued Drill hole No. 2—Continued De th 4/19/65 (558) 4/21/65 (572) ( t) 09:55—10:45 14:09 14:45 15:13 15:32 3.1 441 449 4.0 734.9 5.0 892.8 6.1 1,016.6 954 1,031 963 9.1 1,131 1,109 1,088 1,116 12.1 1,134 1,110 1,088 1,115 15.1 1,128 1,103 1,083 18.1 1,134 1,106 21.1 1,136 Drill hole No. 3 Depth 4/28/65 (6.31) 5/5/65 5/7/65 5/1 7/65 5/24/65 (ft) 10:4?»1126 15:1&15:49 (6.83) (6.98) (7.67) (8.11) 1.0 163.0 102.5 102.0 100.0 2.0 332.3 188.5 129.0 145.3 103.5 3.0 550.3 329.3 314.5 368.5 3.5 95 4.0 715.3 685.1 561.1 470.0 496.8 5.0 849.8 738.0 484.4 615.3 618.0 6.0 955.5 899.0 860.8 741.8 748.3 6.5 787.8 6.7 1,011.5 7.0 993.8 964.3 7.2 863.0 830.3 7.6 1,012.5 8.0 1,056.3 961.0 9.0 1,110.0 1,063.5 9.5 Drill hole No. 4 Depth 5/24/65 6/7/65 6/16/65 6/30/65 (ft) (8.11) (8.93) (9.42) (10.14) 168.3 213.3 228.8 203.0 327.8 351.1 322.0 410.0 451.3 468.0 435.5 555.3 565.3 620.0 611.8 581.0 661.1 645.0 651.3 709.1 696.0 657.8 792.5 775.3 73313 865.0 86715 845.0 933.0 907.8 $°@99@?'WF%PN? ommomuomoooo 1,00318 Drill hole No. 5 Depth 5/24/65 6/7/65 6/16/65 (ft) (8.11) (8.93) (9.42) 1.0 166.5 215.3 232.3 2.0 338.7 352.5 3.0 394.0 454.0 463.6 4.0 554 8 560.9 5.0 608.4 63818 642.0 6.1 739.0 733.0 7.1 810.0 818.5 809.8 8.1 891.5 879.3 9.1 975.5 953.8 938.0 10.1 1005.8 987.8 Drill hole No. 6 Depth 5/26/65 6/7/65 6/16/65 6/26/65 7/25/65 9/23/65 (ft) (8.23) (8.93) (9.42) (9.93) (12.60) (13.70) 1.0 93.1 182.0 203.8 146.0 2.0 249.3 325.8 349.5 338.9 293.0 3.0 387.8 443.8 460.8 450.2 326.5 4.0 515.0 549.1 560.0 542.9 466.0 407.1 4.9 635.5 637.3 626.3 540.0 5.0 641.6 637.3 644.5 503.1 5.9 736.0 720.7 716.5 699.6 594.2 565.5 6.9 835 0 802 8 792.3 769 3 652.5 627.8 7.9 918.2 874.5 859.5 828.7 705.8 686.0 8.9 982.5 936.8 91913 883.6 761.0 737.3 Drill hole No. 7—No measurements made Drill hole No. 8 Depth 6/7/65 6/16/65 6/30/65 Depth 8/17/65 Depth 10/2 7/65 (ft) (8193) (9.42) (10.14) (ft) (12.28) ((1.) (14.89) 170.8 0.5 91.5 0 0 265.0 .5 156.8 0 371.6 2.5 258.5 .0 184.8 212.3 202.5 1 .0 320.8 335.3 337.8 2. .0 449.3 456.3 453.8 3. “NH 68 COOLING AND CRYSTALLIZATION OF THOLEIITIC BASALT, 1965 MAKAOPUHI LAVA LAKE, HAWAII TABLE 26.—Temperature profiles (°C) measured in drill holes 2—24, 68-1—Continued Drill hole No. S—Continued De th 6/7/65 6/16/65 6/30/65 De th 8/17/65 Depth 10/27/65 (t) (8.93) (9.42) (10.14) (t) (1228) (ft) (1489) 4 0 559.0 559.3 546 0 4 0 436.3 3.5 351.8 5 0 649.5 643.5 5 0 536.0 4.5 437.1 5 7 724.0 710.8 697 8 5 6 591.8 5.5 505.8 6 7 805.8 787.8 769 3 6 6 660.7 6.5 573.8 7 7 881.8 858.0 835 0 7 6 720.0 7.5 629.8 8 7 948.3 920.3 891 8 8 6 776.8 8.5 686.8 9 7 1,004.8 974.0 941 8 9 6 824.0 9.5 734.0 Drill hole No. 9 Depth 6/7/65 6/16/65 6/30/65 (ft) (8.93) (9.42) (10.14) 1.0 184.3 219.3 205.3 2.0 320.8 344.0 332 8 3.0 448 0 462.0 446 0 4.0 554 5 559.3 538 2 5.0 643 8 638.8 620 1 6.3 760 5 749.3 707 3 7.3 840 7 822.8 776 0 8.3 915.0 892.3 842.0 9.3 977.4 952.3 895.3 10.3 1,031.8 1,004.0 Drill hole No. 10 Depth 6/7/65 6/16/65 6/30/65 Depth 7/21/65 Depth 8/10/65 8/13/65 Depth 9/27/65 11/4/65 (ft) (8.93) (9.42) (10.14) (ft) (11.13) (ft) (1199) (12.11) ((1) (13.85) (15.16) 1.0 156 0 206.0 222 5 1 0 182.3 1.0 1.1 173.8 187.0 2.0 283 8 332.5 345 3 2 0 307.3 2.0 2.1 260.8 254.3 3.0 411 1 446.9 451 8 3 0 409.8 3.3 436.0 438.8 3.1 351.1 343 9 4.0 522 7 544.9 543 1 4 0 506.0 4.3 518.0 522 9 4.1 430.8 428 4 5.0 616 0 627.8 618 0 5 0 583.1 5.3 588.0 593 5 5.1 512.5 494 7 6.4 752 4 744.0 729 8 6 3 680.2 6.3 652.0 656 7 6.1 578.2 555 8 7.4 832 5 818.0 798 0 7 3 746.0 7.3 723.0 7151 7.1 651.6 611 2 8.4 910 5 888.0 863 0 8 3 805.5 8.3 778.2 772 0 8.1 707.0 664 9 9.4 976.3 948.5 919.8 9 3 863.0 9.3 836.0 827 5 9.1 759.3 714 0 10.4 1,031.0 1,001.0 968.0 10 3 909.8 10.3 882.4 875 0 10.1 803.5 761 3 Drill hole No. 11 Depth 6/7/65 (8.93) 6/16/65 6/30/65 Depth 8/13/65 Depth 9/23/65 10:40—11:45 (9.42) (9.93) (ft) (1211) (ft) (13.70) E 1.0 92.7 131.3 166.0 2.0 180.3 235.8 250.3 3.0 285 0 331.8 340.8 4.0 376 8 410.6 413.5 4.0 309.3 5.0 466 0 487.0 486.8 5.0 465.3 5.0 373.0 6.0 552 3 573.3 557.1 6.0 525.5 6.0 433.1 7.0 640 4 652.2 624.8 7.0 590.5 7.0 512.0 8.0 724 2 730.3 692.6 8.0 654.3 8.0 586.2 9.0 804 8 801.5 762.5 9.0 713.5 8.8 648.8 10.0 877 0 864.5 822.0 9.9 767.3 9.8 711.0 11.0 947 3 931.0 878.0 10.9 815.5 10.8 762.0 12.0 1,006.5 984.0 929.4 11.9 862.3 11.8 811.0 13.0 1,052.3 1,027.0 976.9 12.9 909.3 12.8 854.0 Drill hole No. 12 6/7/65 (8.93) Depth 6/16/65 Depth 6/30/65 (ft) 10:53 (ft) 12:00 (ft) 12:25 (ft) 13:30 (ft) (9.42) (ft) (9.93) 9.3 503.1 8.0 462.0 7.0 404.5 2 0 249.6 7.8 480.9 3.0 222.0 12.3 599.0 11.0 576.8 10.0 546.5 5.0 355.1 9.3 531.1 6.0 439.8 15.3 639.3 14.0 632.2 13.0 612.9 8.0 497.3 11.0 579.6 8.0 488.4 18.3 648.3 17.0 649.8 16.0 643.8 11.0 563.8 12.3 609.1 9.0 506.7 21.3 609.6 20.0 644.3 19.0 645.3 14.0 630.0 13.8 633.1 9.4 525.5 15.3 646.4 11.0 569.8 16.8 650.7 12.0 587.3 18.3 649.8 12.4 585.8 19.8 642.5 14.0 613.3 21.3 615.5 15.0 621.6 15.4 623.8 17 0 632 4 18 4 628 0 20 0 627.3 21 4 599.5 Drill hole No. 13 Depth 7/21/65 (m (11.13) 1.0 181.3 2.0 304.5 3.0 409.3 4.0 503.3 FIGURES 24—28; TABLES 24—29 TABLE 26.—Temperature profiles (°C) measured in drill holes 2—24, 68—1—Continued Drill hole No. 13—Continued Depth 7/21/65 (ft) (1113) 582.9 687.0 752.7 812.0 870.3 .0 .4 .4 .4 .4 .4 922.0 owooqaaol Umntncncn 1 Drill hole No. 14 Depth 6/16/65(9.42) Depth 7/21/65 Depth 2/25/66 (ft) 12:30 13:33 (ft) (11.13) (ft) (18.51) 7.9 724.7 799.3 1.5 241.5 2.0 95 8.9 791.8 867.3 2.5 360.9 2.9 95 9.9 881.0 944.5 3.5 457.5 4.0 167.0 10.9 1,004.5 1,025.5 4.5 548.4 4.9 228.5 11.9 1,059.4 1,070.5 5.5 624.9 6.0 310.5 7.05 736.5 6.9 377.0 8.05 799.3 8.0 442.7 9.05 857.1 8.9 502.0 10.05 914.0 10.0 563.3 11.05 960.5 10.9 617.6 Drill hole No. 15—No measurements made Drill hole No. 16 Depth 8/17/65 Depth 9/27/65 Depth 10/27/65 2/25/66 (ft) (12.28) (ft) (1385) (R) (14.89) (18.51) 1.0 161.8 2.0 257.1 0.7 122.8 2.0 277.5 4.0 423.0 1.7 199.5 3.0 375.6 6.0 560.2 2.7 282.3 4.0 450.4 8.0 677.2 3.7 371.1 5.0 523.3 10.0 779.1 4.7 445.0 199.0 6.0 591.3 12.0 866.8 5.7 515.8 275.7 7.0 651.4 13.8 931.7 6.7 574.5 347 5 8.0 703.0 7.7 631.4 414 5 9.0 758.0 8.7 684.9 480 2 9.7 798.8 9.7 734.6 539 1 10.0 807.8 10.7 781.8 589 1 10.7 848.3 11.7 824.0 641 4 11.7 892.5 12.7 865.8 687 3 12.7 929.0 13.7 901.8 735 1 13.7 966.5 Drill hole No. 17 Depth 7/20/65 Depth 7/21/65 Depth 8/25/65 Depth 10/25/65 (m (11.08) (ft) (1113) (m (12.60) (m (14.82) Pt-PtRhw 4.8 546.0 0.5 90.3 1.0 159.5 2.0 203.3 6.0 637.1 1.5 160.8 2.0 259.8 3.15 310.8 7.0 714.3 2.5 287.1 3.0 351.0 4.0 389.1 8.0 783.0 3.5 408.4 4.0 447.8 5.15 480.4 9.0 842.1 4.5 514.3 5.0 524.3 6.0 541.3 10.0 897.3 5.5 599.8 6.0 588.0 7.15 621.6 11.0 949.1 6.5 675.3 7.0 646.7 8.0 656.4 12.0 995.2 7.5 743.8 8.15 722.8 9.15 716.3 13.0 1,032.4 8.5 806.4 9.15 776.5 10.0 752.3 14.0 1,064.3 9.5 856.9 10.15 822.5 11.15 807.8 14.6 1,076.2 10.5 911.0 11.15 868.5 11.5 961.3 12.5 1,006.0 13.5 1,043.0 14.5 1,064.0 Drill hole No. 20 Depth 8/17/65 Depth 9/27/65 Depth 11/4/65 Depth 12/13/65 Depth 2/3/66 Depth 2/25/66 Depth 4/11/66 Depth 6/2/66 (ft) (1228) (if) (13.85) (15.16) (ft) (1638) (ft) (17.91) (R) (18.51) (Pt) (19.70) (ft) (2097) E 1.0 152.3 2.0 240. 1.0 116.3 1 0 94.5 1.0 92.4 5.0 282.0 4.85 284.5 4.8 274.0 2.0 252.8 4.0 413.3 2.0 237.0 2 0 98.7 2.0 110.0 5.95 340.5 5.85 340.3 5.8 327.5 3.0 342.3 6.0 563.6 3.0 295.1 3 0 101.0 3.0 188.3 7.0 408.5 6.85 392.5 6.8 378.0 4.0 424.0 8.0 684.8 4.0 399.3 4 0 245.0 4.0 254.8 7.95 462.7 7.85 446.4 7.8 428.0 5.0 004.5 10.0 792.3 5.0 454.3 5 0 292.0 5.0 315.1 9.0 526.2 8.85 497.0 8.8 478.0 6.0 586.4 12.0 886.7 6.0 535.8 6 0 403.3 6.0 381.1 9.95 575.2 9.85 548.2 9.8 524.5 7.0 654.5 14.5 980.4 7.0 586.0 7 0 459.0 7.0 429.8 11.0 619.5 10.85 590.0 10.8 570.5 8.0 714.4 8.0 653.8 8 0 532.5 8.0 489.8 13.95 758 0 11.85 638 0 11.8 614.5 9.0 770.0 9.0 698.7 9 0 599.6 9.0 536.8 12.85 677 3 12.8 655.0 10.0 826.0 10.5 768.8 10 0 653.0 10.0 592.0 13.85 719 3 13.8 693.0 10.5 849.3 11.0 797.1 11 5 703.0 11.0 635.4 70 COOLING AND CRYSTALLIZATION OF THOLEHTIC BASALT, 1965 MAKAOPUHI LAVA LAKE, HAWAII TABLE 26.—Temperature profiles (°C) measured in drill holes 2—24, 68—1—Continued Drill hole No. 20—Continued Depth 8/17/65‘ Depth 9/27/65 Depth 11/4/65 Depth 12/13/65 Defpth 2/3/66 Depth 2/25/66 Degth 4/11/66 De‘pth 6/2/66 (t) ( ) ( (ft) (12.28)) (ft) (13.85) (ft) (15,16) (ft) (1638) (17.91) (ft) (18.51) (19.70) ) (20.97) 11.5 904.0 12.5 858.0 12.1 768.0 12.0 692.0 12.5 948.0 13.0 886.0 13.5 814.0 13.0 719.5 13.5 983.5 14.5 937.5 14.1 861.3 14.0 777.0 14.5 1,022.0 Drill hole No. 21 Depth 10/25/65 Depth 11/27/65 Depth 12/20/65 Depth 2/1/66 Depth 2/25/66 Depth 3/28/66 Depth 4/11/66 Depth 7/8/66 Depth 8115/66 Depth 9/13/66 Depth 10/13/66 (ft) (14.82) (ft) (15.90) (ft) (16.61) (R) (17.85) (ft) (18.51) ((1) (19.34) (ft) (19.70) (ft) (2181) (ft) (22.67) (ft) (23.28) (ft) (23.93) 1.0 97.0 1.85 99.6 1.0 98.0 1.0 91.5 6.65- 353.8 7.8 427.0 6.6 375.8 10.8 556.0 8.6 412.0 10.65 520.0 10.5 498.0 2.0 165.3 2.85 101.0 2.0 98.0 2.0 98.7 7.65 420.3 9.8 538.5 7.6 430.8 12.8 640.4 9.6 460.0 12.65 607.0 12.5 588.0 3.0 275.8 3.85 166.3 3.0 98.0 3.0 114.8 8.65 489.3 11.8 641.0 8.6 484.9 14.8 715.3 10.6 532.5 14.65 682.0 14.5 661.0 4.0 373.3 4.85 288.5 4.0 104.0 4.0 200.8 9.65 547.8 13.8 727.0 9.6 536.0 16.8 784.5 11.6 573.0 16.65 745.5 16.5 726.0 5.0 459.8 5.85 375.3 5.0 255.0 5.0 274010.65 605.7 15.8 805.0 10.6 586.4 18.8 848.0 12.6 620.0 18.65 812.0 18.5 789.0 6.0 527.1 6.85 469.3 6.0 349.5 6.0 354.2 11.65 659.0 11.6 634.5 13.6 653.0 7.0 597.3 7.85 548.7 6.85 433.1 7.0 408.3 12.65 706.5 12.6 675.1 14.6 695.0 8.0 650.0 8.85 618.5 7.0 432.0 7.8 455.6 13.65 751.8 13.6 718.3 15.6 724.0 9.0 711.8 9.85 684.7 7.85 509.3 8.0 476.9 14.65 796.0 14.6 756.0 16.6 763.0 10.0 758.3 10.85 743.8 8.0 509.5 9.0 524.5 15.65 834.0 15.6 794.8 17.6 790.0 11.0 814.8 11.85 799.8 8.85 583.8 9.8 573.0 18.6 826.0 12.0 856.0 12.85 844.5 9.0 573.0 10.0 588.7 13.0 901.3 13.85 888.5 9.85 641.3 11.0 635.0 14.0 936.3 14.85 929.8 10.0 637.4 11.8 680.0 15.0 975.8 15.85 964.0 10.85 701.0 13.0 731.0 16.0 1,000.3 11.85 754.0 13.8 771.1 12.85 803.0 15.0 816.0 13.85 846.1 15.8 849.3 14.85 890.5 15.85 922.8 Drill hole No. 22 Depth 12/13/65 1/5/66 2/1/66 2/25/66 Depth 4/11/66 6/1/66 Depth 7/8/66 (ft) (1638) (17.08) (17.85) (18.51) (ft) (1970) (20.95) (ft) (21.81) 1.0 144.8 157.0 167.0 2.0 173.0 145.0 172.0 3.0 214.5 221.0 260.8 4.0 301.3 295.4 308.0 5.0 371.6 353.0 361.1 6.0 439.0 415.5 414.0 7.0 496.3 464.5 462.2 413.0 6.9 426.0 384.5 7.0 383.5 8.0 557.3 520.8 510.2 466.7 7.9 471.3 438.0 8.0 438.0 9.0 614.0 582.0 564.8 527.3 8.9 523.0 491.0 9.0 492.3 10.0 672.8 634.5 610.0 579.8 9.9 568.4 540.0 10.0 537.8 11.0 726.5 692.3 662.2 633.8 10.9 617.1 586.0 11.0 582.2 12.0 780.3 741.5 708.8 682.0 11.9 660.7 628.0 12.0 620.4 13.0 822.8 788.0 758.7 725.1 12.9 700.3 670.0 13.0 658.8 14.0 868.8 829.8 798.0 771.3 13.9 740.4 708.0 14.0 694.2 15.0 907.0 873.0 840.2 809.0 14.9 777.8 746.0 15.0 730.4 16.0 945.3 906.5 872.5 847.0 15.9 813.0 781.0 16.0 763.0 Drill hole No. 23 Depth 1/5/66 Depth 1/19/66 Depth 2/9/66 Depth 2111/66 Depth 2/14/66 Depth 2/18/66 Depth 2/18/66 Depth 2/18/66 Depth ‘2/23/66 Depth 2/23/66 (ft) (17.08) ((1) (17.48) (ft) (18.08) (11) (18.13) (ft) (18.21) (ft) (1832) (ft) (18.32) (11) (18.32) (ft) (18.46) (ft) (18.46) Pt-PtRhm Pt-PtRhm Pt-PtRhm Pt-PtRhm Cr»A1 Pt-PtRhm Pt-PtRhw 2.05 135.7 10.4 624.0 36.7 1,149.8 24.65 1,099 24.65 1,099 20.65 1,000 31.65 1,109 11.5 642.3 “0.65 982 32.65 1,110 10.0 604.1 12.4 729.5 25.65 1,106 25.65 1,106 21.65 1,025 30.65 1,104 10.6 597.3 21.65 1,016 31.65 1,111 12.0 720.3 13.4 770.2 26.65 1,111 26.65 1,111 22.65 1,047 29.65 1,102 22.65 1,038 Cr-Al 14.0 817.4 14.4 825.5 27.65 1,118 27.65 1,118 23.65 1,064 28.65 1,098 23.65 1,058 19.5 970.2 15.0 861.2 15.4 859.8 28.65 1,119 28.65 1,119 24.65 1,075 27.65 1,092 24.65 1,071 18.5 937.3 16.0 903.5 16.4 903.5 29.65 1,120 29.65 1,120 25.65 1,083 26.65 1,086 25.65 1,079 17.5 901 16.5 923.4 17.4 932.8 20.6 1,004 26.65 1,089 25.65 1,079 26.65 1,085 16.5 865 17.0 941.7 18.4 973.3 21.6 1,037 27.65 1,094 24.65 1,071 27.65 1,092 15.5 825.8 17.5 959.8 19.4 998.3 22.6 1,056 28.65 1,101 23.65 1,062 28.65 1,101 14.5 786 18.0 978.3 23.6 1.075 29.65 1,109 22.65 1,043 29.65 1,108 13.5 737.3 18.5 996.4 30.65 1,110 21.65 1,022 30.65 1,112 12.5 693 19.0 1,011.0 24.6 1,081, 31.65 1,108? 20.65 996 31.65 1,115 11.5 637.3 19.4 1,017.0 26.6 1,086 33.65 1,113 19.65 963 32.65 1,118 10.5 585.7 29 6 1,098 34 65 1,117 18 65 932 33 65 1,121 32 6 1,109 35 65 1,117 Cr-Al 34 65 1,124 35 6 1,112— 37 65 1,122— 19.5 976 7 35 65 1,125 382 1,118 1,126 18.6 947 5 36 65 1,126 34 6 1124 37 65 1,129 17.5 904 3 37 65 1,130 28 6 1,127 38 0 1,130 16.6 871 8 38 3 1,130 27 6 1,122 36 65 1,125 15.5 828 7 37 65 1,128 25 6 1,106 35 65 1,122 14.6 792 8 36 65 1,128 1,100 34 65 1,120 13.5 704 4 35 65 1,122 1,087 33 65 1,115 12.6 7010 34 65 1,119 FIGURES 24—28; TABLES 24-29 71 TABLE 26.—Temperature profiles (°C) measured in drill holes 2—24, 68—1—Continued 1 Drill hole No. 23—Continuecl Depth 2 '28/66 Depth 3/1/66 Depth 3/3/66 Depth 3/8/66 Depth 53/8/66 Depth 3/15/66 Depth 3/2 2/66 Depth 3/2 2/66 Depth 3/2 8/66 (1ft) (18.59) (ft) (1862) (ft) (1867) (ft) (18.81) (ft) (1881) (ft) (1899) (ft) (19.18) (ft) (19.18) (ft) (19.34) ) Pt—Ptthu Pt-PtRhm Pt-PtRhiu Pt-PtRhm Cr-A1 Cr-Al 21.65 1,004 19.65 938 20.65 967 21.65 988 19.0 945.3 8901 30.0 1,087 35.65 1,123 1,0361 2 .65 1,034 20.65 972 21.65 997 1,0161 18.0 909.8 17.65 890 29.0 1,083 37.65 1,128 23.6 1,028 23.65 1,057 21.65 999 22.65 1,020 22.65 1,016 17.0 876.5 18.65 929 28.0 1,079 37.75 1,127 ,0561 24.65 1,074 22.65 1,022 23.65 1,040 23.65 1,036 16.0 838.5 19.65 9591 27.0 1,076 36.65 1,125 24.6 1 050 25.65 1,082 23.65 1,040 24.65 11,0551 1,0501 15.0 800.8 962 26.0 1,062 34.65 1,120 25.6 1,062 28.65 1,099 24.65 1,054 1,055 24.65 1,049 14.0 758.2 20.65 998 25.0 1,053 30.65 1,110 1,098.51 . 25.65 1,060 25.65 1,064 25.65 1,060 13.0 709.8 21.65 1,0181 24.0 1,043 28.65 1,098 26.6 1 073 1 26.65 1,065 26.65 1,0751 1,0711 12.0 662.5 ,022 23.0 1,025 26.65 1,081 27.6 1,082 ‘ 27.65 1,075 ,074 26.65 1,068 11.0 609.1 22.65 1,046 22.0 1,010 24.65 1,047 1,095.51 28.65 1,084 27.65 1,085 27.65 1,079 10.0 556.7 23.65 1,0611 21.0 986 24.15 1,049 28.6 1 091 29.65 1,090 28.65 1,0921 1,0881 ,068 20.0 960 24.35 1,051 29.6 1 098 , 30.65 1,094 ,092 28.65 1,086 24.65 1,084 19.0 930 24.55 1,054 1121 1 31.65 1,099 29.65 1,099 29.65 1,093 26.65 1,102 18.0 903 24.75 1,058 31.6 1,109.4 3265 1,099 30.65 1,1021 1,1011 28.65 1,115 17.0 867 24.95 1,062 1181 33.65 1,102 1,103 30.65 1,101 30.65 1,124 16.0 829 25.15 1,064 33.6 1,114.5 34.65 1,101 31.65 1,108 1,1061 32.65 1,127 15.0 788 25.35 1,066 ,1221 35.65 1,109 32.65 1,1091 31.65 1,102 34.65 1,131 14.0 747 25.55 1,070 35.6 1,118.5 1 36.65 1,118 1,110 1,1171 36.65 1,137 13.0 700 37.4 1 126 33.65 1,114 32.65 1,109 37.80 1,141 12.0 650 36.6 1,123.5 34.65 1,1181 1,1141 35.65 1,135 11.0 599 34.6 1 120 1,118 33.65 1,109 33.65 1,131 25.65 1,067 32.6 1 114 35.65 1,1191 1,1181 31.65 1,124 27.65 1,088 30.6 1,107 1 1,120 34.65 1,116 29.65 1,116 28.65 1,102 25.1 1 062 36.65 1,1221 1,1201 27.65 1,104 31.65 1,111 24.1 1,047.5 1,121 35.65 1,118 25.65 1,088 33.65 1,115 24.35 1,051.3 37.65 1,1241 1,1211 22.6 1,012 1,112 36.65 1,119 21.6 988 1 37.9 1,123? 37.65 1,123 19.6 929.5 1 37.95 1,124 18.6 896 20.65 964 18.65 902 1 Drill hole No. 23—Continued Depth 419) 66 Depth 4 19/66 Depth 4/25/66 Depth 5/6/66 Depth 3/28/65 Depth 4’4/66 Depth 4/4/66 Depth 4/1 1/66 Depth 4/11/66 1(ft) (19.90) (ft) (1990) (ft) (20.05) (ft) (20.32) (ft) (1934) (ft) (19.52) (ft) (19.52) (ft) (19.70) (ft) (19.70) 19.6 921 33.2 1,129 19.95 910 19.65 906.8 17.6 861.5 15.65 796 37.55 1,139 31.65 1,1241 14.9 777.8 20.6 955 34.2 1,134 20.65 942 20.65 937.8 15.6 784 17.65 871.8 18.65 907 , 19 13.9 740.4 31.6 985 35.2 1,136 21.65 972 21.65 968.3 19.4 943 19.65 9401 16.65 837.8 1,1321 12.9 700.3 2.6 1,011 36.6 1,139 22.65 1,000 22.65 996.5 18.4 909.5 940 14.65 762.8 33.65 1,129 11.9 660.7 23.6 1,037 23.65 1,025 23.65 1,022 17.4 876 21.65 1,003 12.65 677.8 34.65 1,141 10.9 617.1 24.6 1,056 24.65 1,045 24.65 1,043 16.4 840 22.65 1,024.3 1065 581 1,1481 9.9 568.4 25.6 1,071 25.65 1,059 25.65 1,058 15.4 805 1,0491 35.65 1,140 8.9 546.4 6.6 1,081 26.65 1,074 26.65 1,075 14.4 764 23.65 1,049.5 1965 932 7.9 471.3 7.6 1,094 27.65 1,085 27.65 1,085 13.4 719.5 24.15 1,062.3 20.65 963 6.9 426 8.6 1,105 28.65 1,097 28.65 1,095 12.4 675.5 2465 1,069.5 9931 29.6 1,112 29.65 1,106 29.65 1,105 11.4 625 2515 1,076.8 2165 993 30.6 1,119 30.65 1,113 30.65 1,115 10.4 576.5 25.65 1,0815 1,0201 1.6 1,124 31.65 1,116 31.65 1,121 2615 1,086.5 22.65 1,019 2.6 1,129 32.65 1,125 32.65 1,124 26.65 1,091 1,1431 33.6 1,131 33.65 1,128 33.65 1,132 27.15 1,097.81 36.65 1,140 34.6 1,134 34.65 1,132 34.65 1,131 1,0963 1,1381 35.6 1,136 35.65 1,134 35.15 1,134 27.65 1,101 34.65 1,135 35.6 1,137 36.3 1,136 36.15 1,135 28.65 1,110.5 1,1381 $0.2 937 2965 1,120.9 33.65 1,132 2.2 997 3065 1,128.9 1,1321 24.2 1,045 1,124.91 32 65 1,129 26.2 1,070 3165 1,1355 1,1231 28.2 1,099 3265 1,139.5 30.65 1,121 30.2 1,115 33.65 1,140 29.65 1,114 31.2 1,120 3465 1,142.8 28.65 1,107 32.2 1,126 35.65 1,133.11 27.65 1,097 1,143.3 26.65 1,088 3665 1,140.3 2565 1,075 1 24.65 1,062 1 23.65 1,043 15.9 813 1 Drill hole No. 24 Depth 7’8/66 Depth 9/13/66 Depth 10/13’66 Depth 11 29 66 Depth 11 2966 Depth 2’2/67 Depth 2/2/67 Depth 4/7/67 Depth 6.22/67 (ft) (21.81) ((1) (23.30) (ft) (2393) (ft) (24.89) ((1) (24.89) (ft) (26.17) (ft) (26.17) (ft) (27.36) (ft) (28.72) 22.0 946 40.5 1,137 40.5 1.134+ 30.2 1.056 12.0 527 26.0 937 23.0 851 Cr-AI 26.4 842 21.0 918 39.0 1,1357 39.0 1,133 29.0 1.036 11.0 473 25.0 910.5 22.0 822.5 26.0 892.5 ‘ .5 816 20.0 890 38.0 1.131 + 38.0 1,130+ 28.0 1.016 10.0 416 24.0 884 21.0 793.5 24.0 832 24.4 776.5 19.0 858.5 37.0 1,131+ 37.0 1,129 27.0 993.5 9.0 346 23.0 837 20.0 762.5 22.0 775 23.5 749 18.0 828 35.0 1.125, 36.0 1.121? 26.0 971 8.0 265 22.0 829.5 20.0 707 22.4 709 17.0 794 34.5 1,123.5 35.0 ’ 2' 0 949 7.0 98 21.0 798.5 18.0 625 21.5 681 16.0 761.5 38.0 1.117 34.5 24.2 931 20.0 768 16.0 546.5 20.4 637.0 15.0 725.5 32.0 1,109+ 33.0 23.0 902 19.0 733 14.0 460 19.5 607 14.0 687.5 31.0 1.101 32.0 1,101 22.0 875.5 18.0 695 12.0 344 18.5 560 13.0 648.5 30.0 1,086+ 31.0 1,090* 21.0 846 17.0 656 10.0 100+ 17.5 517 12.0 605 29.0 1,075+ 30.0 1.077 20.0 817 16.0 613.5 16.5 476 lWhere two temperatures are given opposite one depth, the one without an arrow is taken moving downward in the hole and the one with the arrow (1) is taken moving up the hole. 72 COOLING AND CRYSTALLIZATION OF THOLEIITIC BASALT, 1965 MAKAOPUHI LAVA LAKE, HAWAII TABLE 26.—Temperature profiles (°C) measured in drill holes 2—24, 68—1—Continued Drill hole No. 24——Continued Depth 6/2/66 Defpth 7/8/66 Depth 9/13/66 Depth 10/13/66 Depth 11/29/66 De th 11/29/66 Depth 2/2/67 Depth 2/2/67 Depth 4/7/67 Depth 6/22/67 (ft) (2097) (t) (21.81) (ft) (2330) (ft) (23.93) (ft) (24.89) (1%) (24.89) (ft) (2617) (ft) (26.17) (ft) (2736) (ft) (28.72) 13.3 640.5 11.0 563.6 28.0 1.054 29.0 1.057 19.0 786.5 15.0 570.5 Pt-PtRhm 15 5 431 12.3 596 10.0 520 26.0 1.013 28.0 1.037 26.12 974 14.0 532 15.0 405 11.3 554 9.0 474.5 24.0 970+ 26.0 996 25.0 950 13.0 482 30.0 980 14.5 389 10.3 509 8.0 428 22.0 917+ 24.0 949 24.0 925.5 12.0 437 29.25 970 14.0 355 9.3 465+ 7.0 378.5 20.0 858.5 220 900 211.0 900 11.0 380 28.5 950 13.5 341 8.3 418.5 6.0 328 18.0 793 20.0 845 22.0 872 10.0 327 27.5 917.5 13.0 314 7.3 374 5.0 271.7 16.0 727 18.0 784+ 21.0 845.5 9.0 265 26.5 902.5 12.5 297 6.3 3‘ + 4.0 21 1' 12.0 583 16.0 719 20.0 816 8.0 205 25.5 870 12.0 260 5.3 ' 3.0 156.5 10.0 496 19.0 788 7.0 97 24.5 840.5 11.5 245 18.0 755 30.0 1,019 11.0 212 17.0 722.5 29.0 998 10.0 158 16.0 685 28.0 974.5 9.0 106 15.0 650 27.0 953 8.0 100+ 14.0 612 26.0 931 13.0 565 24.0 880 Drill hole No. 68—1 Depth 121168 Depth 1/22/69 Depth 1/22/69 (11) (36.90) (ft) (3747) (11) (37.47) 1.085 1.079 1,067 1,053 1,035 1.015 991 963 935 908 879 854 826 798 765 734 305 266 231 199 165 134 107.5 100 96 MN‘NNNNNNIV. 90~NF%?@fl békfidkbo‘kb-fik HI§XQICNNNNCJQJWMWJK>$£~>>§>A>{>919qu ‘ ©~W¢<1015 Btu, and that for longer-term con- sumption (for the year 2000) is 162X1015 Btu. The val- ues in table 2 show the domestic supply for each major energy source in selected years, along with the total energy consumption for those same years, and by dif- ference the energy shortfall. The latter values represent the amounts that will have to be satisfied by imported fuels or by some method of rationing if energy conser- vation measures are not adopted. OIL SAVINGS BY MEANS OF ENERGY STORAGE At the present time, most utilities use gas turbines as the major source Of peaking power. These devices are low in capital cost, but they are expensive to operate; TABLE 1.—Um'ted States consumption of energy resources by major sources1 10“" Btu 1950 1960 1970 Petroleum ______________ 13.489 20.067 29.614 Natural gas ____________ 6.150 12.699 22.029 Coal ____________________ 12.913 10.140 12.922 Hydro __________________ 1.440 1.657 2.650 Nuclear ________________ __ .006 .229 Total primary energy ____ 33.992 44.569 67.444 ‘Dupree and West, 1972. LITHIUM RESOURCES AND REQUIREMENTS BY THE YEAR 2000 3 TABLE 2,—Um'ted States energy shortfall1 TABLE 3,—Energy storage research and development program. 10'thu 1971 1975 1980 1985 2000 Domestic supply: Natural gas ________ 21.810 22.640 22.960 22.510 22.850 Petroleum __________ 22.569 22.130 23.770 23.600 21.220 Coal 12.560 13.825 16.140 21.470 31.360 Hydro... 2.833 3.570 3.990 4.320 5.950 Nuclear ____________ .391 2.560 6.720 11.750 49230 Total ____________ 60.163 64.725 73.580 83.650 130.610 Domestic consumption __________ 68.728 80.265 90.075 105.300 162.450 Shortfall to be satisfied by im orts ______________ 8.656 15.540 16.495 21.650 31.840 (1 a bbl oil) __________ (1,546) (2,775) (2,946) (3,866) (5,686) ‘Dupree and West, 1972; Hoffman, Beller and Doernberg, 1975. moreover, they produce significant quantities of nitro- gen oxides and are noisy. Development of alternative electrical—storage technology would reduce the need by the utilities for the high-grade fossil fuels and natural gas that are required to operate peaking turbines. With the increasing introduction of nuclear central electric generation, electric energy storage becomes an increas: ingly important consideration because safety and eco- nomics dictate that reactor systems operate at base load. The transportation sector of the energy economy is presently almost totally dependent on the use of gasoline and distillate oils. Thus, the development of energy storage for transportation systems by means of batteries or of the use of hydrogen as an automotive fuel is potentially of great value. The energy supplied to charge batteries or to produce hydrogen would be off- peak power and thus would serve to level the load, in a manner analogous to load leveling at a utility substation or central station.. Petroleum and natural gas savings would also result if economical systems can be devised to utilize replenish- able energy sources such as solar energy and wind energy. Because the availability of these sources is in- termittent, economical storage devices are needed. For heating homes with solar energy, thermal storage ap- pears to be the most appropriate. If solar or wind energy are to be used to generate electricity, however, the use of storage batteries is appropriate. The re- quirements for such batteries would be similar to those for utility storage batteries except that units of smaller capacity would be needed. A summary of the goals of the energy storage pro— grams, expected oil savings, and the research costs are shown in table 3. It is estimated that these programs will save 24 percent of the imported oil projected for the year 2000; this is approximately 13 percent of the total oil demand in the year 2000. The electric utility industry has estimated that 15 to 25 percent of the daytime generating capacity of future Projected annual savings of Total imported oil, research, 10" bbl/yr costs, Program Program goals 1985 2000 $106 1. Energy storage for Reduce oil consumption in 39b 681 300 electric utilities. gas turbines by 80 per- cent in the year 2000,21 reduce requirements for new transmission facili- ties by 10 rcent. ‘ 2. Advanced energy stor- Introduce 22.5 million elec- 3 120 45‘ age for automotive tric and hybrid vehicles applications. into the transportation sector by 2000. 3. Industrial Reduce fossil fuel (oil and 3 212 40C applications. gasuonsumption in in~ ustry by 10 percent in the year 2000. Reduce fossil fuel (gas and nil 123 8‘5 oil) consumption in space heat applications by 12 rcent m the year 2000. Al ow practical implemen- 26 215 30c tation of solar energy to save 2 rcent of the total oil in 11:: year 2000. 4. Commercial and residential. 5. Applications to solar energy. 71d 1,351e Total 2‘ Assumed 6 percent of electricity generated by the utilities was produced by gas turbines with an efficiency of 40 rcent. b Assumed that all oft e new gas turbine capacity installed between 1977 and 1985 and 12 percent of the capacity existing in 1977 were replaced by energy storage devices. 6 It is assumed that the costs required for the other storage programs are incremental cost to the utility stora e program. The utility need is current and these energy storage systems can be implemente in the near term. d Ap roximately 2 ercent savings of projected oil imports in year 1985; about 1 percent of total oilpdemand in 19.85. e About 24 percent savin s of projected oil imports in year 2000; about 13 percent of total oil demand in the year 2000. utility systems could be supplied by batteries that are charged during off-peak periods. Unfortunately, a suit- able battery does not yet exist for this application. The introduction of bulk energy storage batteries on utility networks would eliminate the need for gas turbines which are presently used to generate power during peak periods. If 80 percent of the gas turbine units could be replaced by battery units, the oil savings would be ap- proximately 450 million bbl/year by the year 2000. Plants employing these batteries are expected to be compact, efficient, quiet, and nonpolluting. Con— sequently, siting problems should be minimal, and bat— tery plants could be located near load centers to achieve maximum savings in transmission costs. Also, the bat- teries could be supplied in modular form, which would allow the storage capacity to be easily altered as the elec- tric utility requirements change. The advanced battery systems currently under de- velopment for use in utility networks after 1985 include sodium/sulfur, lithium/metal sulfide, and zinc/chlorine, as well as several other systems which are presently in an earlier stage of development. The sodium/sulfur, lithium/metal sulfide, and zinc/chlorine systems are ap- proaching the battery hardware stage, but will require additional research and development to improve elec- trical performance and lifetime and to reduce costs. None of these battery systems have advanced to the point where successful development has become cer- tain; however, it appears very likely that at least one of these advanced batteries will be successfully developed, 4 LITHIUM RESOURCES AND REQUIREMENTS BY THE YEAR 2000 thereby resulting in the major petroleum savings de- scribed above. Present plans call for the sodium/sulfur, lithium/metal sulfide, and zinc/chlorine batteries to be- come operational in 1980, 1981, and 1983, respectively. (The other electrochemical systems appear to require major technical advances before the hardware stage is reached.) The ERDA program provides for research and development on the most promising alternative bat— tery systems. If major technical advances are made in one or more of these systems, the level of funding will be increased, and the more promising systems will be brought to the demonstration and commercial de- velopment stage, if warranted. The total cost of the ERDA program for research and development of ad- vanced batteries is projected at $200 million through the year 2000. It is anticipated that industry will participate on a cost sharing basis and that some of the demonstra— tion projects may be funded as loans. A Battery Energy Storage Test (BEST) Facility will be constructed for the testing of advanced batteries and power-conditioning equipment. The electrical perform- ance and lifetime capabilities of the batteries will be de- termined under standardized conditions to allow direct comparisons to be made among the various systems. The test results will be used to determine what im- provements are necessary for each system and to aid in determining which systems should receive additional support. The BEST Facility will become operational in 1979 and will be used first for testing advanced lead-acid batteries. Large-scale testing of lithium-aluminum/metal sulfide, sodium/sulfur, and zinc/chlorine batteries should be under way in the BEST Facility in the period from 1980 to 1985. BATTERIES FOR ELECTRIC VEHICLES Existing electric vehicles for highway use are severely limited in their range and performance because of the low specific energy and low specific power of the bat- teries that are available. Such vehicles normally have a range of 20 to 50 miles and a top speed of 30 to 50 mph, depending on the driving conditions. This lack of per— formance not only discourages consumer acceptance, but also creates safety problems. Vehicles of this type may be suitable for certain interim, limited applications; however for the long range, electric vehicles of accepta- ble performance will require new types of batteries of high specific energy (120 to 150 W-hr/kg) and high peak specific power 150 to 200 W/kg). It appears that these long-range requirements can be met only by new ad— vanced battery systems—very likely some of the systems are currently under development. For near-term application, the battery systems being investigated include lead-acid, nickel/iron, and nickel/ zinc. Improvements in these systems will make it possi- ble to attain interim goals that will aid in assessing the long-range viability of electric vehicles. Among the sys- tems that are under consideration for the longer range application are sodium/sulfur, lithium-aluminum/Fe82, zinc/chlorine, zinc/bromine, zinc/air, iron/air, lithium/ air, and aluminum/air. The sodium/sulfur and lithium- aluminum/iron sulfide systems are both well advanced into the cell hardware stage and will soon be tested in battery configurations. Both systems have achieved cycle lives of several hundred charge-discharge cycles and lifetimes of several thousand hours. (The minimum re- quirements for an electric vehicle battery are about 1000 cycles and 25,000 to 50,000 hr.) Both systems also show promise for meeting the long-term specific energy and specific power goals. A zinc/chlorine battery has been demonstrated in a test vehicle and a rechargeable zinc/ chlorine battery has attained about 100 cycles in the lab- oratory. The zinc/air and iron/air systems are under de- velopment by industry and are now capable of about 150 cycles. However, both these systems have low specific powers (20 to 40 W/kg) and low overall energy efficiencies of only about 30 percent. The zinc/bromine, lithium/air and aluminum/air systems are under consid- eration for development, but little or no work is cur- rently in progress. Only a few thousand electric vehicles are on the road today in the US. However, the development of high- performance batteries is expected to increase their ac- ceptability to the public, and the number of electric ve— hicles on the road is expected to increase from 10,000 in 1978 to 18 million by the year 2000. The corresponding annual production of electric vehicle batteries would in- crease from 7,000 in 1978 to 10 million in the year 2000. On the basis of a battery capacity of 35 kW-hr per vehi- cle, the cumulative battery capacity produced over this period would be about 2 million MW-hr. The total value of all of the batteries produced during this period at a price of $30/kW-hr would amount to approximately 60 billion dollars. On this same basis, the annual value of batteries produced would increase from 8.5 million dol- lars in 1978 to about 12 billion dollars in 2000. The widespread use of electric vehicles will affect sev- eral areas of our society. The overall environmental im- pact of electric vehicles should be favorable. Although the level of pollution will not be decreased, the pollution problem will be transferred from the highways to the electric generating plant, where it can be coped with more easily. The use of electric vehicles will also have a significant economic impact on the automotive, petro— leum, battery manufacturing, and electric utility indus- tries. However, the changeover from gasoline-powered to electric-powered vehicles is expected to be gradual, and no major disruptions are anticipated. LITHIUM RESOURCES AND REQUIREMENTS BY THE YEAR 2000 5 MATERIAL SUPPLIER FOR ADVANCED BATTERIES The manufacture of batteries for storing energy on electric utilities and for propulsion of electric vehicles will be a multibillion-dollar business by the year 2000, if ‘ any of several battery programs now under develop- ment are successful. The manufacturers of these bat- teries will require huge quantities of strategic materials. Of particular interest here is that one of the most prom- ising systems—the lithium/metal sulfide system—utilizes lithium, not only as one of the active materials, but also as a constituent of the LiCl-KCl molten salt electrolyte. If the performance goals for the lithium/metal sulfide battery are achieved, the demand for these batteries will, in turn, create a greatly increased demand for lithium. Accordingly, serious consideration should be given by the lithium industry to possible problems related to the availability of lithium resources and to greatly increased production rates. However, the problem of lithium availability may not continue for the long term if other molten—salt systems can be developed that have per— formance capabilities similar to that of the lithium/iron sulfide systems. Work on molten—salt systems not con- taining lithium is in the early development stage at Ar- gonne National Laboratory. ‘ REFERENCES CITED Dupree, W. G., Jr., and West, ]. A., 1972, United States energy through the year 2000'. US. Department of the Interior report, 53 p. and 27 tables. Hoffman, K. C., Beller, M., and Doernberg, A. B., 1975, Current BNL reference energy system projections: “Base Case (Sept. 19, 1975)”: Brookhaven National Laboratory. N BATTERY SYSTEMS FOR LOAD-LEVELING AND ELECTRIC-VEHICLE APPLICATION, NEAR-TERM AND ADVANCED TECHNOLOGY1 By N. P. YAO and W. J. WALSH, ARGONNE NATIONAL LABORATORY, ARGONNE, IL ABSTRACT Intensive efforts are underway in the United States and elsewhere to develop secondary batteries for utility load-leveling systems and elec- tric vehicles. The lead-acid battery is the only candidate system that can meet the performance requirements using presently available technology, but it has the drawbacks of higher-than-desired battery cost and marginal energy-storage capacity for the vehicle application. The Zn/Ni system appears to have a good chance of capturing the near-term electric-vehicle market, provided that advances can be made in reducing costs and increasing battery lifetime. Advanced sec- ondary battery systems that may come into commercial use by 1985 include Li-Al—FeSx, Na/S, and Zn/Clg. These advanced systems are expected to possess superior performance characteristics and economic prospects compared to lead-acid and Zn/Ni batteries. 1Work supported by the Conservation Division of the Energy Research and Development Administration N LITHIUM REQUIREMENTS FOR HIGH-ENERGY LITHIUM-ALUMINUM/IRON-SULFIDE BATTERIES FOR LOAD-LEVELING AND ELECTRIC-VEHICLE APPLICATIONS1 By A. A. CHILENSKAS, G. J. BERNSTEIN, and R. O. IVINs, ARGONNE NATIONAL LABORATORY, ARGONNE, IL ABSTRACT Lithium-a1uminum/iron-sulfide batteries are being developed at Argonne National Laboratory for use as energy-storage devices on electric utilities and as power sources for electric vehicles. These bat- teries are expected to come into commercial use around 1985 and to achieve rapid market penetration thereafter. By the end of the year 1Work performed under the auspices of the Assistant Administrator for Conservation, Energy Research and Development Administration. i 6 LITHIUM RESOURCES AND REQUIREMENTS BY THE YEAR 2000 2000, as much as 3 percent (3X108 kW-hr) of the projected total US. energy consumption could be supplied by lithium-aluminum/iron- sulfide off-peak energy-storage batteries. At the same time 18 million electric vehicles or more may be powered by lithium-aluminum/iron- sulfide batteries. Projections have been made of the quantity of lithium required for lithium-aluminum/iron-sulfide batteries produced from 1985 to 2000. The projected lithium requirement in the year 2000 represents a sub- stantial fraction of the present estimate of world lithium resources and is nearly equal to the estimated economically recoverable lithium re- sources in the US. In addition, the potential market for primary bat- teries using lithium will add to these requirements. Expansion of explo- ration for new sources and increases in production capacity will be needed to meet the projected lithium requirements for the introduc- tion of these batteries in sufficient quantities to produce the desired impact upon petroleum-conservation goals in the US transportation sector and electrical-power generation. INTRODUCTION The Chemical Engineering Division of Argonne Na- tional Laboratory is engaged in a research and de- velopment program on high-performance lithium- aluminum/iron-sulfide batteries suitable for powering electric automobiles and for storing off-peak energy to level the loads on generating equipment in electric util- ity systems. At the present time, commercially available batteries are limited in their ability to meet the stringent performance, cost, and lifetime requirements for these applications. The successful development of more economic, high-specific—energy batteries for either ap- plication would provide a means of using energy pro— duced from nonpetroleum sources (for example, nucle- ar, coal) and reduce our dependence on foreign oil sources. The current effort in the battery program consists of about 65 engineers, chemists, metallurgists, and other technical personnel, including temporary industrial par- ticipants, postdoctoral appointees, students, and faculty members, and is funded chiefly through the Division of Conservation Research and Technology in the Energy Research and Development Administration (ERDA) (Nelson and Webster, 1974; Nelson and others, 1974). The batteries consist of cells containing a lithium- aluminum alloy in the negative electrode and a metal sulfide such as FeS or F652 in the positive electrode. The electrolyte is the LiCl—KCI eutectic, which has a melting point of 352°C. For this reason the cells are operated at temperatures between 380° and 450°C. A drawing show- ing a full-scale cell designed for a utility load-leveling battery is shown in figure 1. To provide an opportunity for industry to develop manufacturing techniques and expertise in the fabrica- tion of lithium—aluminum/metal-sulfide cells, commer- cial development contracts have been let with Gould, Inc.; Eagle-Picher Industries, Inc.; and Catalyst Re- search Corp. to develop, fabricate, and deliver cells to Argonne for testing and evaluation. Argonne has also 3|.lcm STEEL HOUSING (NEGATIVE TERMINAL) 7 POSITIVE TERMINAL /»I\ 2.5 cm \ POSITIVE ELECTRODE (AND CURRENT COLLECTOR 3““, .y» 4615 " NEGATIVE ELECTRODES NEGATIVE ELECTRODES (LITHIUM-ALUMINUM POSITIVE ELECTRODE PARTICLES) (FERROUS SULFIDE PARTICLES) ELECTROLYTE'. LITHIUM CHLORIDE- POTASSIUM CHLORIDE EUTECTIC IRON CURRENT COLLECTOR BORON NITRIDE CLOTH OR PAPER SEPARATOR FIGURE l.—Design of prismatic cell for Argonne National Laboratory lithium/aluminum/iron sulfide battery. contracted with Atomics International for a supporting development program; at the present time their ap- proach incorporates a lithium—silicon-alloy negative elec- trode of their own development. Test cells and electrodes have been fabricated by both Gould and Eagle-Picher and are now being tested. Dur- ing 1976 the industrial-cell-fabrication development will be extended to the fabrication of battery systems. Cur- rent plans include large-scale battery tests both in the laboratory and in a test automobile early in 1978. The cells currently being tested are square in shape, 13 cm on a side, and 2 to 3.8 cm thick. Both F68 and FeS2 positive electrodes are being evaluated. The former appear attractive for electric-utility-storage ap- plication owing to potentially lower cost, whereas the latter appear more suitable for automobiles owing to their higher power capability. Cells containing FeS elec- trodes have achieved a capacity of about 90 A-hr at 1.2 V, and cells with Fe82 electrodes have demonstrated a capacity of about 120 A-hr at 1.5 V. These values, which correspond to specific-energy values of about 100 W-hr/kg for FeS cells and 150 W-hr/kg for Fesg cells, are about three to five times that obtainable from a lead-acid battery. With further improvements in cell design, the specific energy is expected to increase by at least 25 per- cent and possibly as high as 50 percent. LITHIUM RESOURCES AND REQUIREMENTS BY THE YEAR 2000 7 LITHIUM REQUIREMENTS Because of the potential demand for large quantities of lithium for fabrication of lithium-aluminum/iron- sulfide batteries, we have attempted to forecast the need for lithium in the time period 1985—2000. In this re- gard, not only the total amount of lithium contained in the batteries, but the production rate necessary for meeting annual needs is of importance. The projected lithium requirement for off—peak energy-storage batteries is based upon a total US. electrical—energy consumption of 4X 109 MW-hr in 1985, 6X109 in 1990, and 10X10‘9 in the year 2000 (Electric Research Council, 1971). The fraction of this energy supplied from lithium-aluminum/iron-sulfide batteries was estimated to be minimal in 1985, about 1 percent in 1990, and about 3 percent in 2000. Based upon these assumptions, the projected energy supplied from battery storage plants in the year 2000 is 3X108 MW-hr. Laboratory test cells have achieved 150 W- hr/kg. These cells have a lithium content of 6.5 weight percent. About 4 percent of this is in the electrodes, and the remaining amount is contained as lithium chloride in the molten salt electrolyte. Assuming 250 charge- discharge cycles per year and the above values, the lithium content in battery systems on utility networks would be about 5><108 kg in the year 2000. In table 4, we estimate the annual requirement for lithium based upon our projected rate of growth of util- ity storage facilities and assuming 10-yr life-time for the batteries. We have also indicated the production rate of new lithium if the lithium in used batteries can be recy- cled with a 90 percent recovery. Recycling of lithium has the advantage of reducing the amount of new lithium that has to be extracted from available lithium re- sources. Because of the relatively small use of these bat- teries prior to 1990, recycling of utility storage batteries does not become important until after 1995. A projection was also made of the number of electric vehicles in operation in the US. until the year 2000. These vehicles are expected to be introduced in signifi- cant numbers around 1985, with annual production ris- TABLE 4,—Projeeted requirements for lithium in lithium-aluminum/iron- sulfide batteries for utility energy storage 1985 1990 2000 Total US. electrical-energy consumption, MW-hr ______________ 4X10” 6><109 10x 109 Electrical energy supplied by batteries, MW-hr __________________ 5 ><105 6X107 3 X 10a Lithium in utility batteries, kg ________ 1X10“ 1><10s 5.2X10“ Annual new lithium production; no lithium recycle, kg ______________ 1><106 2.8X10" 8X 107 Annual new lithium production; 90 percent lithium recycle, kg ______ 1X106 2.8><107 5.3x107 ing to about 2.7 million in the year 2000. Based upon an expected average lifetime of these vehicles of at least 10 years, about 18 million electric vehicles would be on the road in the year 2000. Figure 2 shows the growth rate of electric vehicles starting with modest production in 1982. A typical vehicle would be equipped with a 42 kW—hr lithium-aluminum/iron-sulfide battery weighing about 350 kg and would be capable of driving the vehicle more than 240 km per charge. The electric-vehicle cells would contain about 7.2 weight percent elemental lithium as metal and salt. Thus, each vehicle battery would hold about 15 kg of lithium; this value represents a total of 2.7x 108 kg in vehicle batteries in the year 2000. During the period between 1985 and 2000, about 20X 106 vehicles would have been built and about 2 x 106 would have been scrapped, based upon an average lifetime of 10 years. Consideration was also given to the effect upon annual demand for new lithium if the vehi- cle batteries had lifetimes of 5 years instead of 10 years and if the lithium in used batteries could be recycled with 90 percent recovery. (see fig. 3.) Table 5 shows the annual demand for new lithium for 5— and lO—year bat- tery lifetimes and under the conditions noted above. Table 6 summarizes the total annual requirements for lithium and the total lithium in battery use for both l l | l l l l 107 _— : — Electric-vehicle population ~ — Annual production rate ‘ _ / / _ _ / _ m / w / .J g 106 — — I I 2 Lu > _ _. u T . _ o — Annual scrapping role - a: — —1 Lu a: _ _ z D 2 _ _ I05 _— __ — -l .04 1 l 1 l 1 I I980 1985 2005 mm l990 l995 YEARS 2000 FIGURE 2.—Electric-vehicle production. 8 LITHIUM RESOURCES AND REQUIREMENTS BY THE YEAR 2000 | I l l — 5—year battery, no recycle — lO-year battery, no recycle I06 — 5—year battery, 90% recycle — : lO-year battery, 90% recycle : O x S- ._ _ 2 I ': _I I07 — —: I06 1 1 1 1 1 1 |985 1990 1995 2000 2005 2010 20|5 YEARS FIGURE 3.—-Annual lithium requirements for electric-vehicle batteries. TABLE 5.—Projeeted requirements for lithium in lithium-aluminum/ iron-sulfide batteries for electric-vehicle propultiorz 1985 1990 2000 Annual electric vehicle production ,,,,,,,, 8.0X10‘ 7.0X105 2.7)(106 Total electric vehicles in operation cccccccc 1.:':><105 2.1 ><106 1.8><107 Lithium in operating electric vehicles, k __________________________ 2.3X10" 3.2><107 2.7)(10‘ Annual new ithium production: 5-yr batter life, no recycle, kg ________ 1.2x10“ 1.2X107 7.7><107 Annual new ithium production: 5-yr battery life, 90 percent recycle, kg ______ 1.2)(10a 1.1 X107 4.4x10’ Annual new lithium production: 10-yr battery life, no recycle, kg ______________ 1.2x10‘ 1.1)(101 4.1X10" Annual new lithium production: 10—yr battery life, 90 percent recycle, kg ______ 1.2X10“ 1.1 X 10" 3.1X10" TABLE 6.—Pr0jected requirements for lithium in lithium-aluminum/ iron-sulfide batteries for utility energy storage and electric vehicles 1985 1990 2000 Lithium in utility facilities, kg ____________ 1.0><10'5 1.0X10“ 5.2x10‘ Lithium in electric vehicles, kg; 2.3X10° 3.2X10" 2.7><10a Total lithium in use, k __________________ 8.3X10“ 1.3X10" 7.9x10“ Total annual new lithium production: lO-yr battery life, no recycle, kg ________ 2.2X10“ 3.9x107 1.2x 10" Total annual new lithium production: 10-yr battery life, 90 percent recycle, kg ____________________________ 2.2x10‘ 3.9X10’ 8.4x 101 Cumulative produc n of no recycle, kg ______________________ 3.3x10a 1.3X10“ 9.3x10“ Cumulative production of lithium: 90 percent recycle, kg ________________ 33x10s 1.3><10a 8.1 X 10" electric vehicles and energy storage under the condi- tions noted above. The requirements shown in this table must be compared with estimated lithium resources and the required rates of production. In addition, some al— lowance must be made for lithium requirements for other battery applications, such as primary batteries for military and civilian use. Considerable uncertainty surrounds estimates of lithium resources. World lithium resources have been estimated as high as 2X1010 kg (H. R. Grady, written commun. 1974). Other estimates suggest that econom- ically recoverable U.S. resources might be less than 1><10g kg (I. D. Vine, written commun. 1975). Table16 shows that the cumulative production of lithium by the year 2000 could equal the presently identified, econom- ically recoverable U.S. resources. Consideration must be given to lithium production rates. The US. Bureau of Mines (I. D. Vine, written commun., 1975) estimates that lithium production was about 2.6)(106 kg in 1968. On the basis of an estimated annual growth rate of 10 percent, the production in 1974 would have been about 4.5X105 kg. Ifa 10 percent annual growth rate is maintained until 1985, annual production would be 1.2x 107 kg. This would be about 10 times the projected requirements for lithium- aluminum/iron-sulfide batteries in that year. If the same growth rate continued until 1990, annual production would reach 2><107 kg. This is less than the annual battery—production requirements under the conditions of 10-yr battery life and 90 percent recycle (3.9 X 107 kg). The requirements for the year 2000 (8.4x 107 kg) would also exceed projected capacity at 10 percent growth (5.2x 107 kg). In addition to expansion of capacity for producing new lithium from ores, it will be necessary to establish a lithium-recovery technology to permit recy— cling of the increasing amounts of lithium in used bat- teries. It should be noted that the chemical form of lithium that will be used in commercial cells will very likely be L128, as suggested by recent results of the cell- fabrication development effort. The commercial produc- tion of Li2S will be based upon the chemical conversion of lithium ores to Li2S rather than the current practice of producing the sulfide from lithium metal. Theravailability of lithium clearly constitutes a major supply problem for the period 1985—2000, and acceler- ated efforts at discovery of new resources and expansion of production facilities will be needed if lithium- aluminum/iron-sulfide batteries are to be introduced into the economy at rates sufficient to produce the de- sired impact upon petroleum conservation goals in the US. transportation sector and in electrical power gen— eration. REFERENCES Electric Research Council, 1971, Electric utility industry R 8c D goals through the year 2000: Publication No. 1—71, Appendix F. Nelson, P.A., and Webster, D.S., 1974, High-energy battery program at Argonne National Laboratory: Argonne Natl. Laboratory REpt. 8064. LITHIUM RESOURCES AND REQUIREMENTS BY THE YEAR 2000 9 Nelson, P.A., Chilenskas, A. A., and Steunenberg, R. K., 1974, The need for development of high energy batteries for electric au- tomobiles: Argonne Natl. Laboratory Rept. 8075, 24 p. available from Natl. Tech. Inf. Service, US. Dept. of Commerce, 5285 Port Royal Road, Springfield, VA 22161. N LITHIUM REQUIREMENTS FOR ELECTRIC VEHICLES USING LITHIUM-WATER-AIR BATTERIES1 By]. F. COOPER, I. Y. BORG, L. G. O’CONNELL, E. BEHRIN, B. RUBIN, and H.]. WIESNER, LAWRENCE LIVERMORE LABORATORY, UNIVERSITY of CALIFORNIA, LIVERMORE, CA ABSTRACT The lithium-water-air battery is a new primary battery of such ex- ceptional power and energy that it is a candidate to provide propulsion for electric automobiles of the future. In the electrochemical reaction involved, lithium, oxygen, and C02 are combined, leaving Li2C03 as a by-product to be removed from the battery and recycled. A subcom- pact car weighing 910 kg would transform 7.2 kg of lithium in travel- ing 320 km at 97 km/hr. At least an equal amount of lithium per car would be unavailable because of the need to recycle the by-product (Li2C03). Thus, a minimum of 14.4 kg Of lithium per car is required to support a transportation system based upon this power source. Assum— ing that in the year 2000, 12 to 16 percent of all vehicles are powered by lithium-water-air batteries, we will need 234,000 to 425,000 metric tons of lithium. This amount is somewhat less than the total known U.S. lithium reserves, but if the current rate of consumption Oflithium for other purposes continues, the supply of lithium will have to be increased. INTRODUCTION The 1973 petroleum embargo by major Arab petro- leum exporters and the subsequent price increases by OPEC (Organization of Petroleum Exporting Coun- tries) focused attention On Oil-consuming technologies and industries. Because the passenger automobile con- SIImes about 4.8 million bbl/day (approximately 28 per- cent of the nation 5 total Oil consumption), the automO-V tive industry has been pressured to modify existing vehicles so that gasoline economy is increased. There is an urgent need tO develop vehicles that use alternate fuels or propulsion systems. One alternative to the gasoline automobile is an all- electric vehicle. However, the most advanced batteries available today have power and energy capacities per unit weight that are'too low to provide performance and range comparable tO conventional automobiles. The lithium-water-air battery, which is closely related to the lithium-water battery developed for marine use by the Lockheed Missile 8c Space Company (Halberstadt, 1973), offers apparent high-performance potential. Further development could provide the basis for an all— IThis work was performed under the auspices of the US. Energy Research and Develop- ment Administration under contract No. W—7405—Eng—48. electric vehicle with the range, speed, acceleration, and rapid refueling capabilities Of current automobiles. CHARACTERISTICS OF THE LITHIUM-WATER-AIR BATTERY The projected performance Of the lithium-water and lithium-water—air batteries approaches that of the inter- nal combustion engine. We have compared the per— formance Of various power sources on a graph of specific power against specific energy in the manner Of Ragone (1968) (see fig. 4). For fixed battery and vehicle weights, top speed and acceleration are determined by specific power, whereas the vehicle range for a given speed is proportional to specific energy. On the basis of demonstrated cell performance data, a lithium-water battery weighing 230 kg could be built to deliver an energy density of 225 to 300 Wh/kg at a specific power of over 100 W/kg. This would suffice to power a vehicle weighing about one tonne for 300 tO 400 km at about 100 km/h. Modification Of the lithium-water battery to include an air cathode has been shown to increase cell 1000 IIITI III 'Li -H20—air yInternai - combustion 7' "‘1“ - x . engine \‘ Molten saits wwrojected): \ ‘ANL \ _ _ \ L1 -S \ 10 _ LI \ . _ : . Nonaqueous : Lead ac1d ' Iectroiytes _ Zn—air _ 1 I I I I I I I I I I I I I l 10 100 1000 Specific energy (Wh/kg) Iooh Specific power (W/kg) (D FIGURE 4.—Performance characteristics of vehicle power sources. 10 LITHIUM RESOURCES AND REQUIREMENTS BY THE YEAR 2000 voltage. A specific energy of 350 to 400 Wh/kg at 110 W/kg has been projected, based on specific cell data and a preliminary battery design. PRINCIPLES OF OPERATION OF BATTERIES OF THE LITHIUM-WATER TYPE » The lithium—water battery utilizes the electrochemical reaction of lithium and water: Li=Li++le— H20+le'=OH‘+1/2H2 Li+ H2O=LiOH(aq)+ 1/2H2 (anode) E8=—3.05V; (cathode) E8: —0.83V; (net) AE8 =2.22V. In a simple version of the cell, an iron wire mesh cathode current collector is pressed against the lithium anode and the leads are connected through a load. When this electrode assembly is immersed in an aqueous solution of lithium hydroxide, a thin film of reaction products forms on the lithium surface and separates the two electrodes. The film inhibits corrosion of the lithium metal and prevents internal shorting of the cell. Prac- tical cell voltages of 1.6 V are obtainable at current den- sities of 103 A/m2. The lithium-water-air battery is considered to be the most energetic of the batteries of the lithium—water type, although other cathodes and electrolytes are under in- vestigation (O’Connell and others, 1974). In the lithium-water—air battery, the iron mesh is placed in con- tact with a catalyzed, porous graphite electrode doped with an appropriate catalyst such as platinum (fig. 5). The battery utilizes atmospheric oxygen with the net cell reaction, Li+1/1O2+1/2 H2O = LiOH(aq)AE8=3.45 V. Iron—wi re-mesh Porous, cathode current catalyzed collector graph1te air electrode Air“. we \\\\\k\\\\\\\\\\ *Ioad l—4M L1'0H electrolyte Case Oxidic film Lithium metal FIGURE 5.—A simple lithium-water-air cell. Protective properties of the anode surface film are re- tained, and cell voltages of 2.6 to 2.7 V at 103 A/m2 are practical, depending on the choice of air cathode. Stable cell discharge requires maintenance of electro— lyte temperature and concentration within certain bounds. Precipitation of lithium as LiOH-Hgo upon electrolyte cooling is one means of concentration con— trol. Concentration might also be controlled by precipi- tation of lithium hydroxide with carbon dioxide. The overall reaction for the lithium-water-air battery would then become Li+ 1402+ l/2COQ(g) = l/ZLIQCOg The energy losses and hazards associated with hydro- gen evolution in the lithium—water-air battery are thus avoided, and a 50 percent increase in cell power is ob- tained. ANODE REPLACEMENT AND LITHIUM RECYCLING The lithium-water-air battery is a primary battery. Recharging is accomplished by removing the reaction product (a slurry of lithium carbonate) and replacing the battery’s supply of lithium and carbon dioxide. For a 910 kg vehicle, 7.2 kg of lithium and 23 kg of carbon dioxide suffice for a 320-km range at 97 km/h. For this range, the present distribution of gasoline service sta— tions is adequate. One attractive option, depicted in figure 6, is to remove the lithium carbonate and replace the carbon dioxide every 320 km, while replenishing the supply of lithium only at intervals of 1,600 km. This would raise the necessary initial inventory of lithium from 7 to 36 kg, making the average amount carried 21 kg. Such an option would reduce the frequency of the most difficult phase of refueling and reduce the number FIGURE 6.—The lithium—water-air battery concept for electric vehicles. LITHIUM RESOURCES AND REQUIREMENTS BY THE YEAR 2000 11 of service stations required for handling lithium metal. The mechanics of refueling the battery pose no serious problems, and details will be determined by considera— tion of economics and convenience. The lithium carbonate reaction product would be transported to a recycling plant for reduction to lithium metal and carbon dioxide, as Shown schematically in figure 7. Such a distribution system involving reproc- essing and service centers does not currently exist. The cost of reprocessing lithium carbonate to lithium metal in large quantities is largely conjectural, but Foote Mineral Company (written commun., March 1975) stated that it may be in the range of $3.30/kg. If the cost of recycling lithium could be reduced to $1.30/kg, the fuel costs for the vehicle would be roughly $0.05/mile. This rate corresponds to a gasoline cost of $1.50/gal for an internal combustion engine vehicle achieving 30 miles/gal. HYBRID BATTERY SYSTEMS The lithium—water-air battery can be employed in sev- eral different configurations. Thus far, it has only been considered as a primary battery system. However, dual battery hybrid systems could allow more economical op— eration without a sacrifice of performance. A secondary battery (for example, Ni-Zn) would provide propulsion energy for all short trips (that is, under 32 km), which constitutes 55 percent of total vehicle travel (Motor Ve- hicle Manufacturers Association, 1973—74, p. 37). The primary battery would be activated when needed for greater range or rapid refueling. Hybrid systems would thus serve to reduce operating costs and make the vehi— cle competitive with internal combustion engines. A dual-battery hybrid system is an attractive means of combining the high performance of the lithium-water— air primary battery with the economical operation of the secondary (table 7). ELECTRICAL CARS IN THE FUTURE We do not know with certainty what percentage of Lithium anodes Regional reprocessing plant C02 C02 2LiCl + 2Li + “2 Electricity Local _ ' H33; auto Ll2C03 + Clz + ‘0‘ " service , —"“> center 2L1Cl + co2 + 1/2 02 Oxygen Li2c03 +2Li + CO2 + l/2 02 l l )Lizcoa/ FIGURE 7.—An automotive system involves recycling of Li2C03 (O’Connell and others, 1974). automobiles used in the U.S. will have electrical propul- sion systems- Table 8 includes estimates from several sources (Kalhammer, 1974; Nelson and others, 1974; Harvey and Menchen, 1974). These suggest that the penetration of the conventional automobile market by such vehicles will be gradual. By the year 2000, electric vehicles are expected to comprise 12 to 16 percent of all automobiles. Electric vehicles require considerable design study in addition to development of an adequate battery. Lower cost electric motors are necessary, lighter weight vehicles are desirable, and reliable and inexpensive electronic controls must be devised. Much new servicing equip- ment must be introduced also. These problems are not major ones, however. Vehicle design will vary in detail and complexity depending on the type of battery. TABLE 8.——-Projeeted effect of electric vehicles on the automotive industry 1975 1985 1990 2000 Auto population (mil ions) ____________________________ 92 126 134 151 Estimated electric vehicle population (millions) __________________ 0 0.2—1 1.8—5 18—25 Percent of electric vehicles ______________________________ 0 0.1—,8 1.3—3.7 12—16 TABLE 7,—Performanee specifications for dual battevy hybr'uis [Assumptions: Propulsion energy and power requirements for 97 km/h cruise: 169 Wh/km and 40 kW for 0—97 km/h acceleration .in 25 sec. Propulsion requirements based on the repre- sentative vehicle: curb weight, 91.0 kg; battery weight. one-quarter of vehicle weight; aerodynamic drag coefficient. 0.4; rolling resistance (constant), 137 nt; frontal area, 1.86m’; drive train efficiency, 75 percent] S cific Battery Specific giver Power Ran e— . weight ener y (average) (peak) at 97 m/h Mileage us: Vehicle power (kg) (Wh/ g) (W/kg) (kW) (km) (percent) Hybrid (secondary with 32-km ranie): Primary atteryl ________________ 146 370 l 10 16 320 45 Secondary batteryl ______________ 81 67 200 24 32 55 Effective total __________________ 227 262 70 40 350 100 Hybrid (secondary with 80-km range) " Primary batteryI ________________ 146 370 110 16 320 25 Secondary battery‘ ______________ 81 167 200 24 80 75 Effective total __________________ 227 297 70 40 400 100 'Alone. 12 LITHIUM RESOURCES AND REQUIREMENTS BY THE YEAR 2000 LITHIUM REQUIREMENTS OF ELECTRICAL VEHICLES POWERED WITH LITHIUM-WATER-AIR BATTERIES A total of 7.2 kg Of lithium is required for a subcom- pact vehicle with a range of 320 km. For every battery in use, an equivalent amount Of lithium will be in process. Thus, for each vehicle in use, about 14.4 kg of lithium are needed. A better estimate of the requiremet would be 13 to 17 kg per vehicle, because the actual designs may cause the requirements tO change. Taking all vehi- cles into account and using the projections from table 8, we can arrive at the following estimate: Number of electric vehicles Lithium required Year (million) (tanner) 1985 ________________ 0.2—1 2,600- 17,000 2000 ________________ 18 —25 234,000—425,000 It is estimated that by the year 2000, 36,000 tonnes of lithium will be produced annually. Estimates of antici- pated cumulative production of lithium in the US. to the year 2000 are on the order of 400,000 tonnes (Vine, written commun. Feb. 1975; Borg and O’Connell, 1976). This supply is adequate to meet the demand for lithium in glass, ceramics, aluminum, fluxes, greases, etc., but clearly, any additional demand, such as exten- sive use of lithium in automotive batteries or in a fledg- ling controlled thermonuclear reactor industry, will tax existing reserves and resources which are on the order of 320,000 and 830,000 tonnes respectively (Vine, writ- ten commun., July 1, 1975). The supply of lithium would have to be increased to accommodate these newly emerging technologies. FEASIBILITY OF THE LITHIUM-WATER-AIR BATTERY At this point, the lithium-water-air battery, as part of an automobile propulsion system, is in a developmental stage. The exact nature of the air cathode is still open, and the required utilization of lithium might be reduced to about 5 kg per 320 km. Alternate electrolytes may prove more advantageous. Alternative ways of control- ling lithium hydroxide concentration may be feasible. The behavior of lithium anodes in the aqueous electro- lyte batteries must be explored. In addition, vehicle de- sign options must be investigated so as best to accom— modate the specific features of the battery, and the required servicing and fuel distribution system must be investigated. The economical reprocessing of the lithium carbonate is also crucial to successful employ- ment of the battery. The total cost per vehicular kilometre must be assessed carefully and compared to the systems involving other battery types or types of fuel. Lastly, the availability of lithium must be critically examined. We do not want to create in the year 2000 an analogue to the present oil crisis. Much remains to be done, but it is Clear that the lithium-water-air battery shows great promise as part of the transportation system Of the future. \_/-\ FUSION POWER AND THE POTENTIAL LITHIUM REQUIREMENT By S. LOCKE BOGART. U.S. ENERGY RESEARCH and DEVELOPMENT ADMINISTRATION, WASHINGTON, DC INTRODUCTION Thermonuclear fusion is regarded by most people as one of the attractive long-range solutions for the Na— tion’s and the world’s energy supply problem because of its major advantages: 1. Effectively infinite fuel supply at low cost (< 7. Flexibility to site plants near load centers, possibly even in urban areas Fusion is the process by which the nuclei of light ele- ments are forced to combine and form new light ele- ments with a net yield of significant quantities of energy per reaction. Fusion compares with fission in the sense that the fission process is the conversion of heavy nuclei to nuclei of intermediate weights with comparable re- leases of energy. Thus, the fundamental distinctions be- tween fusion and fission are the nature of the fuels, the nature of the reaction products, and the mechanisms by which the two processes are made to occur. In this paper, I will restrict the scope to the fusion process and one of the basic fuels for first generation reactors. LITHIUM RESOURCES AND REQUIREMENTS BY THE YEAR 2000 13 The first generation fusion reactors will be based on the deuterium—tritium (D—T) reaction, within the deuterium-lithium fuel cycle. This choice results simply from the fact that the D—T reaction will be easiest to achieve given what is felt as achievable physics and en- gineering in the next three to four decades. The deuterium resource is affectively infinite, being found in one part in 6,500 in water. Commercial lithium is ob- tained from pegmatite ore and specific brines in large but not unlimited quantities. Lithium also is found in seawater, but present availability of this element has not required utilization of this source. The lithium requirement for different fusion reactor concepts is strongly design dependent. When lithium is used both as a breeder and a coolant in either the metal- lic or the salt form, the inventories are very high. In alternative solid breeder concepts the lithium require— ments is much smaller, but this lithium must be enriched in the isotope lithium-6, and a neutron multiplying ele- ment also must be used. Because of the high burnup rate in solid breeder concepts, periodic refueling must be performed which will affect the annual net lithium requirement. This is not the case in liquid lithium breeder/coolant concepts which have such a high initial inventory of lithium-6 that the small fractional burnup per year does not mandate periodic refueling. In the balance of this paper, these concepts will be developed more fully. The basic properties of the fusion process will be described, and the requirements for this process will be related to the potential lithium demand. An energy supply scenario will be developed which in- cludes fusion as part of the mix. The lithium require- ment for a developing fusion reactor economy will be presented. THE CONTROLLED THERMONUCLEAR FUSION PROCESS The fusion of light nuclei (deuterium and tritium) with the subsequent release of energy results from forc- ing the nuclei together, thus overcoming the coulombic barrier, to form a compound nucleus which spontane- ously decays to stable products. The energy released is simply the energy equivalence of the difference between the mass of the fuels and the mass of the products, a fundamental result of the expressionE =mc2. To achieve significant reaction rates, the temperature of the reac- tants must be very high (about 100 million degrees Kel- vin) and they must be confined away from surfaces which would cool them. The last restriction basically de— fines the two types of confinement methods presently being investigated. The Division of Controlled Thermonuclear Research focuses on magnetic confinement. In this method, the reactants which are highly ionized are confined by magnetic fields which are kept away from the walls of the reactor vessel. This satisfies two conditions: reducing the energy flow from the fuels to the wall and confining the fuels for a sufficiently long time to allow a significant number of energy releasing reactions to occur. The technical definition of this condition is called the Lawson criterion, which prescribes the plasma requirements for net energy release from the reaction. Figure 8 illustrates the Lawson operating regime. An alternative to fusion by magnetic confinement is fusion by inertial confine- ment, which is a process being sponsored by the Division of Military Applications in the US. Energy and Re- search Development Administration. In this concept, a small particle of fusionable fuel is imploded by the ir- radiation of either laser light or relativistic electron or proton beams. The Lawson criterion also must be obeyed in this concept, but energy loss to the reactor vessel is not as important. Fundamental to either approach using deuterium and tritium as the basic fuel is the necessity to breed the tritium in a lithium blanket surrounding the reactor ves- sel. First, however, some background information of the basic fuel cycle must be presented. The fusion of a deuterium and tritium atom results in a helium nucleus and a neutron as the products (see fig. 9). The energy release from this reaction is 17.6 MeV with 3.5 MeV being carried away by the helium nucleus and 14.1 MeV being carried away by the neutron. The helium nucleus, being charged, is confined by the magnetic field and subsequently heats incoming fuel to keep the reaction going. The neutron, however, is not charged, and it rapidly escapes the magnetic field and enters the blanket region. Inelastic scattering with blanket material con- verts the kinetic energy of the neutron to thermal 1000 TWO COMPONENT TORUS BREAKVEVEN a 100- _, uwsou D-T g 2 x II B (1975} K BREAKEVEN z r-= ion MILuou 05an 2 to —_' _____ E. ; nousm m (1978) g zx MIRROR \ . . _ d L“ m usual an 1197s) a 1- SCYLlA IV | _. mm PINCH E ALCATOR (1975) l E >- 1 — I3 I a . E St TUKAMAKS | n 740” E ORMAK I g 01— | I l .001 — I l l l l l l l l 109 10‘" 10“ m12 10'3 10" 10‘5 1015 DENSITY-CUNFINEMENT TIME PRODUCT FIGURE 8.—“Break even” plasma conditions for fusion power. l4 LITHIUM RESOURCES AND REQUIREMENTS BY THE YEAR 2000 DEWWUM LI'I'HIIM FROM WATER usuinnu ‘— ______ -i I ENERGY D + T + "E‘zfl‘filw D “‘0 + "9* (17.6 MeVl “ l usuum —————— .1 r-— _ vv STABLE HELIUM av mnucr ENERGY GAINv~ZflM—1w FIGURE 9.—Deuterium-tritium reaction. energy which is removed by the coolant for the conver— sion to electricity. Now comes the part critical to the deuterium-tritium fusion reaction concept. To regenerate sufficient amounts of tritium to maintain the fuel cycle, the neu- tron must be absorbed by either a lithium-6 or a lithium-7 atom. By proper blanket design, the sum of these reactions will result in the net production of more than one tritium atom per fusion event. As it turns out, the lithium-6 reaction predominates in all concepts, and it is this reaction that provides the bulk of the tritium fuel. However, the lithium-7 reaction is required to achieve a breeding ratio of greater than 1 providing other neutron multipliers such as beryllium are not used. To avoid this increase in blanket design complex- ity, very thick liquid lithium blankets have been pro- posed (fig. 10). Unfortunately, large quantities of natu- ral lithium must be provided for this concept. If one is willing to accept the complexities introduced by neutron multipliers in the blanket, then lithium in the enriched form of lithium-6 can be used which significantly re- duces the quantity of lithium required (fig. 11). The choice of either of these two concepts cannot be made at this time simply because the basis for such a decision is HEAT EXCHANGER MAGNETIC CDtL m LIOU|D . LIIHIUM .' /' ....... ' -1. VACUUM ALTERNAT lNG CURRENT YRITIUM SEPARATOR GENERATOR NEUTRON DEUIERIUM TRmuM J CONDENSER {Am m lRlIIUM DEUIENIUM INJECIOR FIGURE lO.—Liquid lithium blanket design for a fusion reactor. not yet developed. However as we shall see, the choice of either blanket concept strongly effects the quantities of lithium required in given fusion reactor design. FUSION POWER DEVELOPMENT PLAN Let us look at the four basic approaches to fusion power reactors. The Division of Controlled Thermonu- clear Research presently studies the theta-pinch, the mirror, and the tokamak concepts (fig. 12). Both the theta-pinch and the mirror approaches are regarded as backup to the tokamak approach, so I will focus on the latter as being typical of what might be a commercial fusion power reactor after the turn of the century. Fig- ure 13 is an artist’s rendering of such a plant. Since the inertial confinement concept is not part of my program I will not discuss it at all, but mention that the laser fusion reactor is equally as probable as a tokamak fusion reactor. The development of tokamak fusion reactors is planned such that a demonstration commercial fusion power plant will be built before the turn of the century. Figure 14 presents the current plan to achieve this goal. The first major fusion facility designed to burn deuterium and tritium will be the Tokamak Fusion Test Reactor presently planned to operate in the 1980—19811 time frame. This is a plasma physics experiment which will release significant quantities of thermonuclear power but will be incapable of producing any net electrical power. The subsequent experiment is called the first Experimental Power Reactor. This machine will operate in the 1986—1988 interval and probably will be capable of producing net electric power from the fusion process. It probably will not be a complete fusion power reactor .in the sense that it is not likely that tritium will be bred in sufficient quantities to maintain the fuel cycle equilib- rium. The provision for fueleycle equilibrium may add unnecessary complexity which could compromise the design goal for the machine. However, the second ex- perimental power reactor, presently planned for opera- tion in the 1990—1991 time frame, would be a miniature fusion power electric plant. This machine would dem— onstrate the technologies of all future fusion power reactors but would not provide the basis from which vendors and utilities could assess the economic viability of the concept. The last experimental power reactor, the Demonstration Fusion .Power Plant (DEMO), would op- erate before the end of the century and would provide the basis for an economic assessment of the viability of fusion power as an alternative to other energy supply options that may exist at that time. The DEMO will be a small scale commercial fusion power plant. Of course the development schedule presented above depends upon the solution to a number of significant technical problems that have been identified at this time. LITHIUM RESOURCES AND REQUIREMENTS BY THE YEAR 2000 15 HEUUM SUPPLY HELIUM RETURN HEADERS HEADERS 3m- / ////// [/ /,//./i,///://y////// ////i’i////\ fl: , : GRAPHITE I 5.5. CLAD ~ ~ VERTICAL . STRUCTURE i 48cm . SECTION ! i ~ TRITIUM as STAINLESS : COLLECTION cm STEEL MANIFOLDS CANS I A / I A . x A . A: A —— U N 0. SEAL ATTACHMW ! BETWEEN POINT J— _ 4 CELLS T is ism Be filflcm LL015ml SEAL BETWEEN seem" A" 27”" SECTION ‘ L. AI O, REMOVABLE FRONT WALL _ A SECTION COOLANT A 23”“ CHANNELS PLASMA FIGURE 11.—Blanket design using enriched lithium-6 and neutron multipliers. Perhaps foremost among these is the development of materials which are highly resistant to the energetic reaction products that result from the fusion process. These materials will be exposed to neutron and charged particle flux and will experience the potential for dam- age not yet observed in reactor engineering. This prob— lem was early recognized and an intensive program is underway to test selected materials in a simulated fusion radiation environment and to develop new materials which are resistant to fusion radiation. The prognosis is optimistic because of the experience in fission technol- ogy and the lead time intrinsic to the program. How- ever, the effect of materials problems on economics has yet to be demonstrated. The second major problem (characteristic) of the fu- sion process is recirculating power and the need for energy storage and transfer. With all known controlled fusion concepts, a significant amount of power will be required to bring the fuels to reaction conditions and temperatures. For example, although the tokamak seems least demanding in this respect, calculations indi- cate that perhaps 500 MW(e) over 10 seconds may be required for reactor start-up. In principal, the energy produced by the fusion reactions will be significantly larger than the invested energy providing that the burn time is reasonably long. However, for shorter burn times or lower duty cycles, there may be strong economic incentives to recover the invested energy at a high efficiency at the end of the burn. Answers to this problem are now significantly inconclusive because it is not yet clear what the duty cycle of a tokamak reactor may be. -A third problem facing fusion by magnetic confine- ment is the development and construction of supercon- ducting magnets which are more than several times larger than the present state-of—the-art of this technol- ogy. Superconducting magnets are necessary because the energy balance of the facility cannot tolerate the use of normal conducting copper or aluminum magnets. Large superconducting magnets have been built for physics facilities such as the Brookhaven National Labo- ratory and the Argonne National Laboratory bubble 16 LITHIUM RESOURCES AND REQUIREMENTS BY THE YEAR 2000 MIRROR MACHINE FIGURE 12.—Comparison of four basic designs for fusion power reactors. chambers, but these are of a very simple solenoidal de- sign. Further, these are by a factor of two or more small- er than those which will be required in a tokamak fusion power reactor. But size is not the most important crite- rion; for toroidal systems, the superconducting magnet set must be built in the shape of a doughnut. Is is not clear how the complex magnetic field and force struc- ture will affect the stability of such magnet sets and especially how these forces may prescribe the structural material necessary to make the device rigid. A major part of the Development and Technology Program of the Division of Controlled Thermonuclear Research is the design and construction of large toroidal supercon- ducting magnet sets. Another potential problem area arises from the fun- damental properties of one of the basic fuels, tritium. Since tritium is mildly radioactive, there are prescribed certain maximum allowable releases either in normal operation or in accident situations. Tritium is an isotope of hydrogen and, consequently, has similar diffusive properties of that element through hot metallic sur— faces. The diffusivity of tritium affects the normal re- lease characteristics of a fusion power plant whereas the inventory of tritium in a state which possibly could be- come volatile affects the accidental release characteris- tics. In either case it is felt that adequate tritium con- finement will be achieved, but with a nontrivial effect on the economic parameters of the fusion power plant. It is evident that the fusion power development schedule can be affected strongly by the degree of diffi- culty of these and other problems. Also becoming evi- dent as inhibiting the development of pure fusion LITHIUM RESOURCES AND REQUIREMENTS BY THE YEAR 2000 I7 MAJOR COMION'ENTS Reactor Buildi . Vacuum Chamber 13. Hot Helium Line . Flihe Blanket 14. Helium Return Line Shield Blanket 15. Steam Generators . Divertors 16. Steam Headers . Vacuum Pumps 17. Feedwater Header . 0H Magnet Coils 18. Reheat Steam Return Headers . Toroidal Magnet Coils 19. Helium Circulators . Liquid Helium Tank 20. Cryogenic System . Magnet Support Cylinder 21. Flibe Sparging Units . Helium, Flibe 5. Water Lines 22. Flibe Storage Tanks . Fuel Injector . 23. Reactor Maintemnce Hat Cell Biological Shield 24. Polar Crane Turbine Hall . Very High Pressure Turbine 29. 1800 RPM Generator . High Pressure Turbine 30. 3600 RPM Generator . Intermediate Pressure Turbine 31. Feeduater Heaters low Pressure Turbines 32. Feedwater Treatment Equipment FIGURE l3.—Artist’s conception of a tokamak fusion power reactor plant. power is the inadequate commitment of both public and private sector funds necessary for the prerequisite re— search to enable commercial fusion to be demonstrated in this century. This is an institutional problem arising from the reluctance to commit near term funds for long term benefits and seems to be endemic to all long term energy supply options. For the sake of argument, the balance of the analysis presented in this paper will as- sume that adequate funding is provided for the fusion power research and development program. ASSUMPTIONS AND ANALYTICAL TECHNIQUE Fundamental to this analysis as well as any other re- source demand projection is how the ultimate demand is expected to grow in future years. For pure fusion reac- tors, the basic product will be electric power for which there now exist a number of projections, many of which have electric power growing at the near historical or historical rate. However, recent estimates are not as generous with electric power production growth, and I will use these as they represent a fairly realistic scenario. Two cases are extrapolated from data in the Project In- dependence Report which was published little more than one year ago (fig. 15). The first scenario is defined as the Base Case which is the continuation of present electric energy growth trends with a smaller growth rate after the year 1990 as a result of conservation. The Base Case assumes no significant change in energy use pat- terns with the growth of electrical energy demand prin- cipally based upon population expansion. The second case is called the Massive Shift which assumes essentially the same total energy use as the base case but that many energy uses now satisfied by conventional fuels will be satisfied by electricity. This projection could come about if nuclear energy thrives and is competitive in historical nonelectric final use categories or if coal cannot be con- verted competitively to nonelectric energy products. The selected implementation schedule for fusion power reactors as presented in figure 16 follows trends exhibited by other high technology programs (for 18 LITHIUM RESOURCES AND REQUIREMENTS BY THE YEAR 2000 FUSION POWER DEVELOPMENT 1970 1975 1980 OONCEPTUAL DESIGN,- SYSTEMS ANALYSIS, MATERIALS RESEARCH AND DEVELOPMENT, TRITIUM DEVELOPMENT HANDLING, ETC. AND TECHNOLOGY SUPERCONDUCTING MAGNETS, NEUTRAL BEAMS, ENERGY STORAGE, ETC ////////////////////// ////////////////////// ;;{4////////////////// ................ / /;SMALLSCALESTUDES / /, OF EQUILIBRIUM, . NEAR-REACTOR -.. :~ LEVEL -=. CONDITIONS IN HYDROGEN PLAsM s TEsT ;; STABILITY, AND xx HEATING // ,,,,,,,,,,,,,,,,, I / ////////////////////// ////////////////////// ////////////////////// ////////////////////// ////////////////////// ////////////////////// ////////////////////// ////////////////////// //// // / / ” CONFINEMENT SYSTEMS RESEARCH FUSION REACTORS : To lem 1985 2000 COMPONENT DEVELOPMENT, SUBSYSTEM INTEGRATION, SYSTEMS DEVELOPMENT AND OPTIMIZATION, ETC. :EXPERIMENTA EXPERIMENTAL ; POWER ' REACTOR No. 2 ; :nO-IO>I"|:U ZO—HhN-IV’ZOZMU >100 MINIel CONFINEMENT THEORY, COMPUTER SIMULATION OF CONFINED PLASMAS, RELEVANT FUNDAMENTAL PLASMA STUDIES FIGURE l4.—Fusion power development plan. example, nuclear fission, space exploration, and so forth). The point in time at which implementation be- gins is chosen by the following arguments. Commercial fusion power reactors will be built after utilities and vendors are satisfied that fusion power is an energy source competitive with alternative technologies. Purchasing decisions will depend on the total cost of fusion power as compared with that of competing technologies. Disadvantageous characteristics of all energy sources are assumed to be included in costs. This accounting procedure is not uniformly practiced today but certainly will be before the end of the century. If the DEMO is operating in 1998, then it is expected that the first order for a commercial reactor will be placed several years later, say in the year 2000. Sub— sequent orders would be placed in a fashion typical of most new technologies. Assuming that the time from order to operation is seven years, achievable for nuclear plants today, then the operation of the first commercial reactor would not take place until the year 2007. We will take this point in time as the implementation date of commercial fusion power with the implementation rate following the order rate but lagging by seven years. Consequently, we assume for the sake of this assessment that lithium will not be required for fusion power reac- tors in commercial quantities until the year 2007. Note that fusion electric power does not take over the market completely. This occurs because there usually will be cases where unique circumstances make an alter- native more attractive (economically, environmentally, socially). The saturation level here is taken to be about 80 percent of new orders, but there is no quantitative basis for this choice. The effect, however, is not too important. This reactor type we will consider is the tokamak as this concept seems the most promising at this time. Since the breeding characteristics of fusion reactors are rela- tively design independent, this choice will not bias the analysis. Two types of blankets will be considered which represent the extreme of the lithium demand range. Table 9 presents the inventory and annual replacement lithium requirement for both concepts. The first type is the liquid lithium blanket which performs the simul- taneous functions of tritium breeding and reactor cool- ing. This blanket type has a large lithium inventory but has such a small fractional burnup rate of lithium-6 that periodic refueling is normally not required. The second blanket type is the gas cooled solid lithium breeder LITHIUM RESOURCES AND REQUIREMENTS BY THE YEAR 2000 50F 40>— 20!— INSTALLED ELECTRICAL GENERATING CAPACITV (IN HUNDREDS OF GIGAWA‘I’TS) lllllllllllIlILJ 1970 1980 1990 2000 2010 2020 2030 YEAR 1950 FIGURE 15.—-—Installed electrical generating capacity as a function of time for Project Independence Base Case and Major Shift to Elec- tricity scenarios with the FPC 7 percent growth rate for comparison. BC, Base Case, MS, Massive Shift, and FPC, Federal Power Commis- snon. 100 ASVMPTOTE _ .82 PERCENT OF NEW ADDITIONS 0 2 4 6 8 10 12 14 16 18 20 YEARS SINCE COMMERCIALIZATION FIGURE 16.—Projected penetration of fusion into electrical generating market. 19 TABLE 9,—Blanket lithium requirements Liquid‘ Solidz breeder/coolant breeder (MT/MWe)’ (MT/MWe) High tails Low tails Natural (25 percent) (0.3 percent) Inventor ______________________________ 1.15 .0558 .03918 0 .00338 .00239 Annual replacement‘ ____________________ *At a thermal conversion efficiency of 40 rcent and a plant factor of 80 percent. lUWMAK-I University of Wisconsin Blan et. 2UWMAK-I Brookhaven National Laboratory Blanket. JMT, Megatonnes. MWe, Megawatts electric power. which is enriched 90 percent in the isotope lithium-6. Very small lithium inventories are required but there is need for a neutron multiplier such as beryllium to achieve a tritium breeding ratio of greater than unity. Furthermore, the fractional burnup rate of lithium-6 is relatively high and frequent refueling is required (perhaps as often as two years). A notable feature of the gas cooled solid breeder con- cept is that the lithium material “tails” resulting from the enrichment process (depleted in lithium-6) become available for non-nuclear applications. Consequently, the lithium demands computed for this particular blan- ket concept are not absolute in the sense that all the lithium is removed from the marketplace. The natural lithium demand in the ith year for fusion power reactors, being commercially implemented in the year 2007, is defined to obey the following expression: D,- = X1i+Y(Ai_1+1i/2), X E Natural lithium inventory demand in Kg/MW(e) of installed fusion reactor capac1ty, 11- E Incremental fusion reactor capacity in MW(e) installed in the ith year, Natural lithium required to replace burn-up in Kg/MW(e) of installed fu- sion reactor capacity with an 80 per- cent duty factor and a conversion effi- ciency of 40 percent, Cumulative installed fusion reactor capacity in MW(e) at the end of the (i—1)th year. The accumulated natural lithium demand at the end of the yearj simply is Y E s m J D = 21Di=D1+D2+...+D/._]+Dj. 1: For the case of the solid breeder enriched in the isotope lithium-6, the natural lithium demand as a function of enrichment is calculated from the following expression (0-6) X (or Y) = Z [(.0742—b) ’ 20 where X orY E Natural lithium feed per MW(e) of in- stalled fusion reactor capacity, Z E Enriched lithium feed per MW(e) of in- stalled fusion reactor capacity, a E The percentage of the Li-6 in the fusion reactor fuel, I) E The percentage of the Li-6 in the lithium enrichment plant tails, .0742 E The Li-6 fraction of natural lithium. The natural lithium demand associated with the solid breeder concept will be calculated for two tails fractions. The first case is 2.5 percent lithium-6 remaining in the tails which possibly could represent the most economical process. The second tails fraction is selected to be 0.3 percent which presently is characteristic of the U-235 fraction remaining in depleted uranium. One may wish to have such low tails fraction if lithium were indeed to become a very dear commodity after the turn of the century. In this case, one would attempt to extract as much lithium-6 as possible out of the natural lithium. RESULTS Tables 10 and 11 present the annual and cumulative natural lithium demand posed by a growing fusion reac- tor energy economy for both the Massive Shift and the Base Case. Table 10 suggests what might be required of the lithium industry while table 1 1 presents the potential impact on the resource base itself. It is clear that the choice of the basic breeder concept strongly affects both the annual and the cumulative demand and is more LITHIUM RESOURCES AND REQUIREMENTS BY THE YEAR 2000 1000- 800 - 700 - 600 - 500— 400 _ LIQUID MASSIVE SHIFT BASE CASE — SOLID LINE — DASHED LINE 300 — 100— 90' 30" 70- 50 — SOLID 40 _ HIGH TAILS SOLID 30 ' LOW TAILS NATURAL LITHIUM CUMULATIVE DEMAND (THOUSANDS OF METRIC TONS) 20- 10 l I 1 I I J 2000 2005 2010 2015 YEAR FIGURE l7.—Cumu1ative lithium demand curve for different reactors. Table lufiAnnual natural lithium demand Massive Shift Base Case Liquid Solid Solid Liquid Solid Solid Electric Fusion breeder/ breeder, breeder, Electric Fusion breeder/ breeder, breeder, ca acity ca acity coolant high tails low tails ca acity acity coolant low tails high tails Year (8W8) ( We) (MT) (MT) (MT) ( WC) ( We) (MT) (MT) (MT) 2000 __________________________ 1600 DEMO ~0 ~0 ~O 1400 DEMO ~0 ~0 ~0 DEMO ~0 ~0 NO 1500 DEMO ~0 ~0 ~0 19.8 8,700 470 335 1600 8.6 4,140 220 160 92.6 24,800 1,470 1,040 1700 41.8 10,120 615 435 239.6 38,900 2,620 1,860 1800 101.8 15,870 1,085 765 425.8 42,800 3,440 2,430 1900 180.4 20,490 1,565 1,110 614.4 43,400 5,000 2,900 2010 270.6 20,750 1,880 1.330 TABLE ll.—Cumulative natural lithium demand Massive Shift Base Case Liquid Solid Solid Liquid Solid Solid Electric Fusion breeder/ breeder. breeder. Electric Fusion breeder/ breeder, breeder, czzpacity ' city coolant high}; (ails low tails ca city cagacily coolant low tails hi h tails Year ( We) ( We) (MT) ( T) (MT) ( We) ( We) (MT) (MT) EMT) 2000 __________________________ 1600 DEMO ~0 ~0 ~0 1400 DEMO ~0 ~0 NO DEMO ~0 ~0 ~0 1500 DEMO ~0 ~0 ~0 19.8 22,780 1,200 850 1600 8.6 9,890 520 365 92.6 106,500 6,080 4,300 1700 41.8 48,070 2,745 1,945 239.6 275,600 16,940 11,990 1800 101.8 71 17,070 7,245 5.130 425.8 489,700 32,830 23,250 1900 180.4 207,481 13,950 9,880 614.4 706,600 52,050 36,860 2010 270.6 811,210 22,750 16,110 LITHIUM RESOURCES AND REQUIREMENTS BY THE YEAR 2000 21 than a factor of 10 larger for the liquid lithium breeder/coolant concept (fig. 17). For the liquid breeder/coolant concept, the cumula- tive inventory lithium requirement reaches 7X 108 kg by the year 2030 in the Massive Shift case. This quantity corresponds to currently reported estimates of between 5X 108 and 109 kg of economically recoverable domestic reserves. For the more conservative Base Case, the de— mand could exceed the reported reserves by the year 2040, even before the first plant, operating in 2007, was retired. In both cases, the potential cumulative natural lithium requirement for a fusion reactor economy is compared against what presently is regarded as the eco- nomically recoverable reserves. Experience shows that as one is willing to spend more for feedstock, more feedstock becomes available. Consequently, the demand posed by liquid lithium coolant/blanket concepts may only serve to increase the lithium supply at a higher, but yet to be quantified, cost. The solid breeder concepts for both the high and low tails cases are far less demanding on the lithium supply. Furthermore, there does not seem to be a significant advantage to operate at the very low tails fraction of 0.3 percent. This result is placed in even greater perspective by the fact that most of the lithium is returned to the marketplace after the lithium—6 has been removed. For example, with a tails fraction of 2.5 percent Li-6, 94.38 percent of the calculated lithium demand is returned to the marketplace. This compares with only a modest de- crease to 92.06 percent if the tails fraction was reduced by a factor more than eight to 0.3 percent. It is evident that only if lithium separative work were to be very cheap or if lithium were to be very expensive would there be an incentive to maintaining a low tails fraction. CONCLUSIONS Fusion power reactors may place a significant demand on what presently is thought to be the economically re- coverable reserves of lithium if lithium is used for reac- tor cooling in addition to tritium breeding. The en- riched solid breeder concept is far less demanding on the domestic lithium resource since much less feed is required and most of the feed is returned to the mar- ketplace after the Li-6 isotope is removed. However, the solid breeder concept has a neutron deficit which must be made up by using neutron multipliers such as beryl- lium. Other multipliers exist but all, including beryl- lium, have potential technological or resource limita- tions. It is premature to select either the breeder/coolant or solid breeder concept as the most likely candidate for fusion reactors. Furthermore, there may exist attractive alternatives which lie within the lithium demand ex- tremes represented by those two concepts. The lithium supply should be more than adequate to meet the long term needs of fusion power. Even if the cost oflithium rises to many times its present level, utili- zation of lower grades of lithium bearing materials, in- cluding seawater, appears economic. A paper to be pre- sented later in this symposium will indicate that the cost of lithium extracted from the oceans is only three to four times the present market price. This mild rise in cost is well within allowable limits for fusion power reac- tors. Since the question of fuel resources is so fundamental to all energy supply options, we will continue to study the implications of blanket choices on lithium demand. However, with respect to fusion, it is premature to em- bark on programs to increase the supply or to stockpile material. Fusion power reactors are too far in the future to warrant the expenditure of present funds for these purposes given the current domestic lithium resource estimates. Our real needs at this time lie in two areas. First, is an assessment of the total domestic and world supply of lithium as a function of both its cost and availability. Second, and perhaps more important, is a coordinated study of the different potential demands for lithium and how they will be met. REFERENCES Borg, I. Y., and O’Connell, L. G., 1976, Lithium’s role in supply energy in the future energy sources: Energy Sources, v. 2, no. 4 (in press). Halberstadt, H. j., 1973, The Lockheed Power Cell: Proceedings of Intersociety Engineering and Energy Conference, Detroit, Michi- gan, Soc. Automotive Engineers, Paper No. 739008, p. 63—66. Harvey, D. E., and Menchen, W. R., 1974, The Automobile—Energy and the Environment: Columbia, Md., Hittman Associates Inc., 159 p. Kalhammer, F., 1974, Energy Storage—Incentives and prospect for its development: Electric Power Research Institute, Palo Alto, Calif., Report presented to Am. Chem. Soc. (Atlantic City, Sept. 12, 1974), 19 p. Motor Vehicle Manufacturers Association, 1973—74, Automotive Facts and Figures: Detroit, Michigan, p. 37. Nelson, P. A., Chilenskas, A. A., and Steunenberg, R. K., 1974, The need for development of high energy batteries for electric au- tomobiles: Argonne National Laboratory, Report ANL—8075, 24 p. O’Connell, L. G., Rubin, B., Behrin, E., Borg, I. Y., Cooper,]. F., and Wiesner, H. j., 1974, The lithium-water-air battery—A new con- cept for automotive propulsion: Lawrence Livermore Laboratory Report UCRL—5181], 36 p. Ragone, D. V., 1968, Review of battery systems for electrically pow- ered vehicles: Detroit, Michigan, Soc. Automotive Engineers, Paper No. 680453, 8 p. K'\_/ 22 LITHIUM RESOURCES AND REQUIREMENTS BY THE YEAR 2000 US. LITHIUM SUPPLY AND DEMAND AND THE PROBLEMS INVOLVED IN COMPILING STATISTICS By HIRAM B. WOOD, U.S. BUREAU OF MINES, WASHINGTON, DC ABSTRACT The United States has the reserves and the capability to more than satisfy all current domestic demands for lithium products. For the 1970—75 period, US exports Of lithium averaged about 21 percent of US. production. The US. is the largest exporter of lithium chemicals, exceeding Russia for the past 3 years. Currently, only three US com- panies are producers of primary lithium products. The total 1974 US. published annual production capacity was about 4,500 metric tonnes (5,000 tons) of contained lithium. For the 1970-74 period, U.S. pro- duction ranged from 65—70 percent of world estimated production. Predicting the future demand for lithium products requires predict- ing the future growth of the many industries that use lithium prod- ucts. For the 1970—74 period, the industry grew at the rate of 6 per- cent per year. Because of the current worldwide industrial slump, the 1975 growth was much lower, and a conservative growth rate of 5 percent annually is predicted through 1980. But a breakthrough, such as in nuclear fusion or lithium batteries, could by 1980 double the 1975 demand. The compilation of dependable statistics has been hampered by the lack of standard terminology, by difficulties with surveys, and by some reluctance on the part of private companies to reveal their production. The United States and the world in general have the lithium reserves and the capabilities to supply demands . at the current growth rate at least through this century. The 1975 world production and/or demand was about 6,300 tonnes (6,900 tons) of contained lithium. During the 1970—74 period, the United States exported about 20 percent of its production, and US exports equaled about 14 percent of world production. For the same period, the US. accounted for 65 to 70 percent of world production. Throughout the world, in the 1972—74 period, de- mand for lithium carbonate by the aluminum and ceramic industries increased notably. Demand for lithium hydroxide also continued strong. Demand de- creased during the last quarter of 1975 making 1975 production about the same as 1974. Details on world trade of lithium compounds are available from trade tables published by the major industrial countries. Most of these countries, including West Germany, Italy, and France, import both raw mineral concentrate and lithium compounds and then export a variety of lithium chemicals and lithium metal. Only the United States and Russia are known to produce lithium mineral concen- trate and to support significant quantities of lithium compounds and lithium metal. Until 1965, the largest producer and exporter of min- eral concentrate, mostly petalite, was Rhodesia. At that time the United Nations applied trade restrictions to that country, and since then export figures for Rhodesia have been obscured. On the Bureau of Mines world production tables, it was assumed that Rhodesia had continued to mine at the same rate of about 860 metric tonnes (950 tons) of contained lithium in mineral con- centrates as in 1965. Based on verbal information, how- ever, their production decreased greatly in 1974 and 1975. There are only three US. producers of lithium, and two of these have been responsible for 98 percent of domestic output; consequently, the US. Bureau of Mines cannot publish domestic production figures re- ceived from these companies. However, the Bureau can use production capacity data which have been published either in company reports or in trade magazines, and it can publish US. Census Bureau export and import figures. From these data, we have prepared table 12. All figures in table 12 are shown as tons of contained lithium. Three classes of marketable material are pre- sented. As previously explained, lithium mineral production figures for the United States, which were reported to the US. Bureau of Mines as spodumene flotation concentrate containing 5 to 6.8 percent lithium oxide, are not shown. They have been small and erratic. Imports, which are mostly lepidolite, have been small. The production of lithium compounds and metal represent published figures. These figures of estimated production capacity, consumption, and exports were compiled from different sources. The United States production estimates published by Lithium Corporation of America (LCA) and by the Canadian Department of Energy indicate an 8 percent annual growth rate. The estimate by Roskill Information Service for only a 3-year period indicates only a 4 percent growth rate. These estimates appear to be realistic, considering U.S. indus- try demands and exports, but the production tonnage estimates for 1972 through 1974 are notably different. From these figures, it is apparent that the United States production for 1974 and 1975 was between 4,200 and 4,600 tonnes (4,600 and 5,100 tons) of lithium. The sale by the US. General Services Administration (GSA) of lithium hydroxide monohydrate from excess stocks is listed to show how much of this material has been made available to industry. These were published sales with the material going to the highest bidder whether a US. or a foreign country. All of this lithium hydroxide monohydrate came from USAEC (now US. Energy Research and Development Administration, LITHIUM RESOURCES AND REQUIREMENTS BY THE YEAR 2000 23 TABLE 12.—U.S. salient lithium statistics [Lithium content in short tons (tonnes in parentheses)] 1970 1971 1972 1973 1974 ”1975 Lithium mineral concentrate: Shipments: Chemical and ceramic grade ____________________ W W W W W W Exports _______________________ _ _________________ 66 81 38 (50) (73) (34) Imports1 _________________________________________ 63 128 36 170 88 NA (57) (116) (33) (154) (80) Lithium compounds and metal marketed: Production est. Lithcoa2 ________________________________ 2,444 2,820 3,008 3,290 NA NA (2,217) (2,558) (2,729) (2,985) Production est. Canadiana ____________________________ 3,200 3,200 3,600 4,280 ”4,700 5,170 (2,903) (2,903) (3,266) (3,883) (4,264) (4,690) Production est. Roskill“ ________________________________ NA NA ,120 5,450 5,820 NA (4,645) (4,944) (5,280) Exports of metal and compounds1 __________________________ 747 516 503 805 894 5850 (678) (468) (456) (730) (811) (771) Lithium hydroxide (LiOH-H2O): Exports‘ ______________________________________________ 113 39 91 86 99 NA (103) (35) (83) (78) (90) General Services Admin. Excess sales ______________________________________ None None None 157 433 61 (142) (393) (55) Price per pound ___________________________________________________ $0.65 $0.87 $0.95 Inventory, Dec. 31 ________________________________ 1,071 1,071 1,071 914 482 420 1 (972) (972) (972) (829) (437) (3811) Yearend prices: Spodumene mineral concentrate, per ton ________________ NA NA $80 $88 $107 $130 L1thium carbonate, per pound ............. ___ $0.52 $0.54 $0.55 $0.58 $0.78 $0.80 Lithium metal, per pound ______________________________ $8.17 $8.17 $8.43 $8.53 $9.38 $11.30 eEstimate; W, withheld; NA, not ap licable. lU.S. Census Bureau and Roskill In ormation Service. ’Source: Roskill Information Service, p. 30 ’Source: Roskill Information Service. Production by Canadian Department of Energy and Resources, p. 30, 1974—75 projections by USBM; Exports of Metal and Compounds, p. 9: Production estimate by Roskill, p. 32. ERDA) stockpiles and was devoid of the isotope, lithium-6. A recent published report stated that ERDA has in stock about 36 million kg (80 million pounds) of lithium hydroxide monohydrate. For those who are not familiar with the prices of the major lithium products, table 12 lists the approximate yeare‘nd price for spodumene mineral concentrate, lithium carbonate, and lithium metal from the Chemical Marketing Reporter. The price of lithium mineral con— centrate increased more than 60 percent in the past 4 years, and lithium carbonate increased more than 45 percent. The estimated world capacity, by country, and total estimated world production from 1970 through 1975 are shown on table 13. The tonnage estimates compiled by the U.S. Bureau of Mines may disagree notably from the world consumption estimates by Luckenback and from the production estimates by the Canadian De- partment of Energy and Mines as shown on the bottom two lines. The totals do not include GSA stockpile sales, but they do include tonnages of similar lithium hydrox- ide monohydrate, also devoid of isotope lithium-6, which were exported from the USSR. The U.S. Bureau of Mines world production figures indicate, for the 1970—74 period, a 6 percent annual growth. The esti- mated world production capacity figures for the same period indicate a 9.4 percent annual growth, and Luc- kenbach’s figures for the same period indicate a 10.5 percent annual growth. Table 13 shows that the United States, Russia, Rhodesia, Brazil, and South Africa have been and still are the major mineral producing countries. Although Luere does not appear to be much difference between the estimated production and the estimated production capacity, the important implication shown is that, if necessary, world production can in l or 2 years be in- creased 25 to 30 percent. Beyond that, new mines would have to be opened and new beneficiation and chemical plants built. The most significant production increases can most readily come from the United States, Canada, Rhodesia, Russia, Southern Africa, and Brazil. In time and with proper demand, new production may be realized in Chile and Zaire. The Zairetain mining com- pany recently announced plans to build a $4 million spodumene concentrator at Manono. Table 14 is included to show world trade for lithium. The exports from the United States, Russia, West Ger- many, and some other countries were derived from pub- lished trade statistics and show shipments mostly to western European countries and to japan. This table shows West Germany as both a major importer and ex- porter, but they are not a primary ore producer. Trade data on the United Kingdom and Canada should be included in this table, but 1974 trade books from these countries had not arrived. This table does not show all world trade shipments, but it is probably 85 to 90 per- cent complete. In general, current market conditions and future predictions are favorable. Although the 1970—74 growth rate averaged 6 per- cent per year, it dropped to about 2 percent in 1975, and unless industry rebounds notably in 1976, the growth this year will not be much better. However, any good growth in any one of the major consuming indus— 24 LITHIUM RESOURCES AND REQUIREMENTS BY THE YEAR 2000 TABLE 13.— Estimated maximum world productwn cabacity of lithium contained in mineral concentrate or brine solutiam [Short tons (tonnes in parentheses)] 1970 1971 1972 1973 1974 21975 United States ____________________________________________ 3,600 4,000 4,500 5,000 5,200 5,500 (3,266) (3,629) (4,082) (4,536) (4,717) (4,990) Argentina ________________________________________________ 5 5 5 10 10 10 (4.5) (4.5,) (4.5) (9) (9) (9) Australia ________________________________________________ 20 40 40 40 30 30 (18) (36) (35) (36) (27) (27) Brazil ____________________________________________________ 60 120 150 150 150 150 (54) (109) (136) (136) (136) (136) Canada __________________________________________________ 40 40 40 90 90 100 (36) (36) (36) (82) (82) (91) Mozambique ______________________________________________ 10 10 20 20 20 10 . (9) (9) (13) (18) (18) (9) Portugal __________________________________________________ 10 20 25 40 40 10 (9) (18) (23) (36) (36) (9) South Rhodesia ________ __-. ________________________________ 950 950 950 950 950 950 (862) (862) (862) (862) (862) (862) South Africa, Republic of, and Southwest Africa ____________ 100 250 250 250 300 300 (91) (227) (227) (227) (272) (272) U.S.SR ___________________________________________________ 700 1,500 1,500 1,800 1,800 2,000 (635) (1,361) (1,361) (1.633) (1,633) (1,814) Total world Capacity 5,495 5,935 7,480 8,350 8,590 9,060 (4,985) (5,384) (6,786) (7,575) (7,793) (8.219) Total world production‘ ____________________________ 5,100 5,300 5,900 6,500 6,800 6,900 (4,627) (4,808) (5,352) (5,897) (6,169) (6,260) Total world consumption W. F. Luckenbach’ __________ 3,055 2,886 3,478 4.653 5.020 NA (2,771) (2,618) (3,155) (4,221) (4,554) World lithium production3 __________________________ 5,060 5,290 5,842 6,532 NA NA (4,590) (4,799) (5,300) (5,926) eEstimated; NA, not applicable. IU.S. Bureau of Mines. 2Source: Mining journal, Mining Annual Review (1975, p. 112). 8Canadian Department of Energy Mines and Resources. TABLE 14.——World trade [Imports of lithium compounds and lithium metal as short tons of contained lithium (tonnes in parentheses” Exporting countries Other U.S.A. U.S.S.R West Germany countries Total 1 t' ' £1,321? 1973 1974 1973 1974 1973 1974 1973 1974 1973 1974 Belgium and Luxemborg ___ 2 13 33 5h 12 ———— 60 56 (1.8) (11.8) (30) (51) (10.8) (54) (511) France ________________ .___ 75 77 59 35 91 70 .__- 29 225 212 (68) (69) (54) (33) (83) (64) (26) (204) (192) Italy _____________________ ll __-. "-9 68 55 47 19 126 77 (10) (2.7) (62) (50) (43) (17) (114) (70) Japan ................... 301 375 305 302 .-__ .___ ___. 606 684 (273) (340) (277) (274) (6.4) (550) (621) Netherlands _____________ 3 11 ___. 4 24 26 ____ 1 27 42 (2.7) (10) (3.6) (22) (24) (0.9) (24) (38) West Germany ___________ 322 327 264 104 ..-- ___- 72 41 658 472 (292) (297) (239) (94) (65) (37) (597) (428) Other countries3 _________ 111 115 NA NA 119 78 NA NA 230 193 (101) (104) (108) (71) (209) (175) Total ____________ 825 908 641 446 335 285 131 97 1,932 1,736 (748) (824) (582) (405) (304) (259) (“9) (88) (1,753) (1,575) e750 6550 (680) (499) eestimated; NA, not applicable. IData from trade tables of respective importing countries, exclusive of United Kingdom. 2Data from West German export tables. “US, exports of lithium hydroxide to other countries were estimated, based on partial data. tries will certainly reflect a proportional increase in lithium demand. When one tries to predict the future demand for lithium and its seemingly unlimited number of compounds, one also has to predict the future growth for such industries as the aluminum, ceramic, glass, rubber, refrigeration, nonferrous alloys, industrial machinery, and organic chemicals, and also the unpre- dictable potential demand for automobile batteries, for nuclear fusion reactors, and the switching to absorption refrigeration to avoid use of fluorocarbons. A 5 percent per annum growth over the next 5 years is a conserva- tive but practical estimate. Increased use of lithium car- LITHIUM RESOURCES AND REQUIREMENTS BY THE YEAR 2000 bonate by aluminum companies who currently do not use lithium carbonate to improve the thermal stability and electrical conductivity of their potlines will account for most of this growth. Undoubtedly, this is why Foote Mineral Company is building a new carbonate plant at Kings Mountain and why LCA, in their 1974 annual report, discussed plans to increase their output 50 per- cent. Without a scientific breakthrough, the conservative 5 percent annual growth would easily require by 1980 at least an additional 1,180 tonnes (1,300 tons) of lithium or 6.35 million kg (14 million pounds) of lithium carbo- nate. A higher growth rate for the European nations and japan is anticipated because more severe energy shortages could force them to use more lithium carbo- nate in their aluminum smelting potlines. In view of the present embryo state of the lithium industry, I think a US. demand of about 5,700 tonnes (6,300 tons) and a world demand of 9,070 tonnes (10,000 tons) oflithium by 1980 is reasonable. However, a breakthrough in any of the potential fields just men- tioned could triple the growth rate to~15“percent annu- ally. If the growth should increase 15 percent annually by 1980, US. and world production would have to dou— ble to 9,100 and 12,700 tonnes (10,000 and 14,000 tons) annually. A warning on overenthusiasm in predicting future growth should be given because even if there is a break- through in 1976 or 1977, the consuming industry could not tool up and make the changes necessary to demand this much more lithium by 1980. Terminology is one of the worst problems involved in 25 compiling US. and world lithium statistics because production may be reported as lithium ore, lithium mineral concentrate, or lithium compounds. Often the lithium or lithium oxide content of the ore or concen- trate is not given. Ifthe kind oflithium mineral concen- trate, such as spodumene, petalite, lepidolite, or amblygonite is listed, a fair estimate of the lithium con- tent may be made. But in the international trade books, which list foreign shipments, lithium assays are seldom recorded. For years, the US. Bureau of Mines, in the producers canvass survey, has asked for shipments of lithium minerals, but seldom did the producers list the lithium oxide assays. In 1975, a new survey form was designed to be more specific. Also, in the US. Census tabulations, only the lithium hydroxide monohydrate is listed separately. All of the other lithium compounds are grouped under a blanket category which may include lithium amide, borate, bromide, chloride, carbonate, fluoride, and so forth. Also, the producing companies, in their annual reports or in trade magazines, report capacity and sometimes U.S. production as lithium car- bonate equivalent. Usually, the lithium carbonate equiv- alent includes only the lithium compounds, some 30 or more that are produced commercially, and lithium metal. These estimates may or may not include the lithium mineral concentrate produced. Only the lithium companies can standardize their terminology and per- mit their output figures to be published; however, with only two significant, competitive companies in the in- dustry, there are understandable reasons for their reluctance. /'\/ THE LITHIUM INDUSTRY By GERALD ]. ORAZEM, LITHIUM CORPORATION OF AMERICA, GASTONIA, NC ABSTRACT The image of the lithium industry in the past has been clouded by lack of information available to the industrial world. The industry consists of several ore producers and marketers, but only three major basic producers of lithium chemicals; Lithium Corporation of America, Foote Mineral Company, and the USSR. The USSR exports to the Free World, but their supply is unpredictable in any one year. Their exports consist mostly of chemicals depleted in the lithium-6 isotope, a byproduct of their atomic energy program. Foote Mineral extracts lithium from brines at Silver Peak, Nev., and Lithium Corp. from spodumene concentrates in North Carolina. A minor producer is Kerr-McGee Chemical Company at Trona, Calif. Besides the basic producers, there are several converters, the most important of which are Metallgesellschaft in West Germany and Honjo Chemical injapan. The major lithium chemicals are carbonate for use in ceramics and the aluminum industry, hydroxide in the grease industry, bromide and chloride in air conditioning, hypochlorite in sanitation, metal in pharmaceuticals, and butyllithium in synthetic rubber. The Free World demand for lithium chemicals, based on carbonate equivalents, was about 21 million kg in 1974 and 17 million kg in 1975. The United States capacity for lithium chemicals was 20 million kg in 1975. The largest potential in the next 5 years for lithium demand is the aluminum industry. The lithium battery is not expected to have a major impact on the lithium industry until the early 1980’s. N 26 LITHIUM RESOURCES AND REQUIREMENTS BY THE YEAR 2000 LITHIUM RESOURCES—PROSPECTS FOR THE FUTURE By IHOR A. KUNAsz, FOOTE MINERAL COMPANY, Exr0N, PA ABSTRACT The future of the lithium industry will be controlled by the demands for lithium ores, chemicals, and metal. The ability of the lithium indus- try to meet any large future commitment must be appraised not only in terms of reserves, which are based on economic and legal extracta- bility at the present time, but also in terms of identified resources, which represent sources economically and legally extractable at some time in the future. The point in time at which they will become re- serves will be determined by the rate of lithium demand. Consequently, all important identified resources must be considered if an accurate representation of the potential of the lithium industry is to be made. These resources are: Tin-Spodumene Belt, North Carolina, USA. Clayton Valley, Nevada, USA. Bikita Tinfields, Rhodesia Preissac-Lacorne District, Quebec, Canada Bernic Lake District, Manitoba, Canada Great Salt Lake, Utah, USA. Salar de Atacama, Chile Manono-Kitotolo District, Zaire These represent producing districts, districts with past or planned production, and districts with large identified resources. An evaluation of these deposits and the probability of discovery of additional re- sources in existing districts suggest that lithium resources will support the needs of the lithium industry for several decades. INTRODUCTION In reviewing the recent activity in the field of lithium, it appears that a great deal of attention has been devoted to it. In addition to last year’s edition of Industrial Min- erals and Rocks, lithium has been reviewed by Norton in US. Geologic Survey Professional Paper 820 in 1973, by Industrial Minerals of London in 1974, by Roskill In- formation Services of London in 1975, as well as by members of the newly created Lithium Exploration Group of the US. Geologic Survey. To these one must add the regular annual reviews prepared by the US. Bureau of Mines, the Engineering and MiningJournal, Mining Engineering, and Miningjournal. This surge in interest can be attributed to the anticipated potential of lithium in battery and possibly in thermonuclear power development. The continued interest in this fascinating element was amply demonstrated by the impressive at- tendance at this symposium. RESERVES AND RESOURCES The contents of the publications reveal two opposing viewpoints. Some maintain that resources are suffi- ciently large to satisfy future demands. Others suggest a serious forthcoming shortage by the end of the decade because of the large quantities of lithium predicted for battery and thermonuclear power development. This disparity of opinion is not unusual or surprising. It is the result of a grave concern over the dependence of our economy on the imports of a significant number of strategic mineral commodities. This has led to a reevaluation of our mineral industry on the basis of a new reserve concept (McKelvey, 1973). This concept has been used by the US. Geological Survey in the recent evaluation of the lithium industry (Vine, 1975). In addi- tion to a reserve evaluation, attempts to “estimate” probable lithium yields to be expected from the industry by the year 2000 have also been made. The “estimates” constitute the basis for a shortage prediction by some people. While recognizing the validity of the reserve concept as it is proposed and applied to the evaluation of the lithium industry at the present time, one must neverthe~ less exercise great caution in using reserves (and even more, probable yields) to predict the ability of the lithium industry to fulfill future commitments, such as the battery or thermonuclear power. Such commitments are based on tenuous data. A contradiction seems ap- parent. Since reserves are defined on the basis of pre- sent steady state technological and economic conditions, they can only validly predict a steady state future. De- mands resulting from battery and thermonuclear power development will create unsteady state conditions which may alter existing technological and economic condi- tions. For this reason, the future of the lithium industry cannot be precisely, nor validly, predicted on the basis of today’s reserves. This point can be illustrated by lithium’s past history (fig. 18). Following World War II, the lithium industry experienced a period of slow growth consistent with commercial demand. In the few years which preceded the Atomic Energy Commission (AEC) program, the in- dustry did not have the capacity necessary to meet the AEC requirements. And yet, within the amazingly short period of 24 months, Foote Mineral Company, Lithium Corporationn of America, and American Potash and Chemicals Corp. were able to meet this excessive de- mand. Clearly, had the lithium industry been evaluated on the basis of reserves and of “probable yields," the results could not have justified the initiation of an AEC program. In fact, economic parameters had changed drastically, and resources became reserves practically overnight. While the industry was able to fulfill its commitment by increasing its capacity, the latter caused its downfall as well. From 1955 to 1960, commercial demand for LITHIUM RESOURCES AND REQUIREMENTS BY THE YEAR 2000 27 50-— 30— I 20— I Li2C03 EQUIVALENT, IN MILLION POUNDS IO — I ”I A ’z’ 0 I I I | I I930 I940 I950 I960 I970 I980 Y E A R 3 FIGURE 18,—History of lithium production in the United States. A, Use (during World War II) of large quantities of lithium hydride to inflate rescue apparatus, and in the development of multipurpose lithium greases. B, Lithium purchase program of the U.S. Atomic Energy Commission. lithium had only doubled, while the industry was faced with a 500 percent overcapacity. Plants operated at about 20 percent of their installed capacity, and the lithium industry experienced a period of economic de- pression. In order to determine the ability of the lithium indus- try to meet future requirements, a comprehensive study of the lithium industry must be made. It is proposed that a valid evaluation can only be made on the basis of identified resources and not on the basis of reserves. IDENTIFIED RESOURCES An evaluation of identified resources reveals that in addition to existing producing districts, there are several areas around the world which must be included in a proper evaluation of the lithium industry. These areas contain identified resources (conditional resources) which could justify chemical processing centers in the future. The following paragraphs review existing and potential districts in the United States, Canada, Africa and South America. TIN-SPODUMENE BELT, NORTH CAROLINA, U.S.A. As stated by Kunasz (1975), the tin-spodumene belt of North Carolina constitutes the largest developed reserve of spodumene in the world. Following a short period of mining activity during World War II, Foote Mineral Company began the mining of spodumene in 1951. In the 1960’s, Lithium Corporation of America opened up its own mine some 5 miles northwest of its Bessemer City, North Carolina plant. Although Foote Mineral Company also produces significant quantities of lithium carbonate from its Silver Peak brine operation in Clayton Valley, Nev., the tin—spodumene district consti— tutes the most important source of lithium in the world today. In 1959, when Foote Mineral Company completed its drilling program over the main portion of its ore body, a total resource of about 36,000,000 tons of pegmatite was reported (Kesler, 1960). In the following 16 years, sev- eral million tons of ore have been mined out. The pegmatite belt is known to extend through the remaining portion of the Foote property—its resource potential essentially unknown. As a result of a recent modest drilling program, a resource of 38,000,000 tons has been identified over approximately one-half of the Foote property. Resources on the remaining half of the property are still unknown but will be identified in the next few years. Drilling increased the pegmatite re- sources by a full 20 percent, not only replacing the mined out ore, but adding several million tons of a new resource. Lithium Corporation of America has more than dou— bled its reserves from 13.6 million tons grading 1.4 per- cent Li20 to 30.5 million tons grading 1.5 percent LigO. A preliminary mining plan indicates that 27.5 million tons are recoverable by open pit mining methods. The outlined pegmatite remains open downdip (Evans, 1976). Under an extensive exploration program which would take place under conditions of increased lithium demand, the probability of discovery of considerable additional reserves and resources is considered high. CLAYTON VALLEY, NEVADA, U.S.A. A benchmark in the history of lithium has been the discovery of this element in the brine of Clayton Valley, Nev. The uniqueness of this brine deposit lies in the fact that lithium is a primary, and not a co- or by—product. The resources of Clayton Valley have been a subject of controversy and need to be clarified. As pointed out by Norton (1973), the large resources previously re— ported also included lithium contained in Tertiary and Quaternary lacustrine sediments present in the valley. In the last edition of AIME’s Industrial Minerals and Rocks, an updated resource figure was reported (Kunasz, 1975). The new figure of 775,000 tons as Li includes only that amount calculated to be contained in the brine body. In addition, a yield of 44,500 tons of Li as recoverable product was estimated under present economic and technological conditions. This is only a 28 LITHIUM RESOURCES AND REQUIREMENTS BY THE YEAR 2000 small fraction of the resource present in Clayton Valley. Favorable technological and economic conditions will increase the yield since weaker and deeper hypothetical brine resources are not being exploited or included in published data. The yield estimate for Clayton Valley has been re- cently misinterpreted by Vine (1975) in a paper entitled “Are Lithium Resources Adequate for Energy Self- Sufficiency?” Using the published yield of 44,500 tons Li and assuming a 50 percent recovery factor, Vine calcu- lated and reported lithium resources at 90,000 tons, in- stead of the 775,000 tons reported by Kunasz (1975). In a similar evaluation of the tin-spodumene belt, a proba- ble yield of 680,000 tons as Li is reported. In this case, however, even hypothetical resources have been in- cluded in the estimation of the yield. Clearly, the yield from a hypothetical source can only be zero. An artificial set of numbers has thus been generated and used to predict the potential of the lithium industry. The valid- ity of such an evaluation is questionable and, more im- portantly perhaps, is misleading. GREAT SALT LAKE, UTAH, U.S.A. The Great Salt Lake constitutes a considerable re- source oflithium. The importance and the reality of this source was well demonstrated when the Gulf Resources Corp. announced production plans. However, since the production of lithium is tied to the production of other salts such as potassium sulfate, sodium sulfate, mag— nesium chloride, contractual difficulties resulted in a postponement of the project. Total resources in the Great Salt Lake have been re- estimated at 526,000 tons as Li (Whelan, 1976). The quantity of lithium which can be extracted from this body of brine is restricted by the production levels of primary salts and by the efficiency of the extractive technology. An agreement with the State requires Gulf Resources Corp. to return all waste salt streams back to the lake. This eliminates waste disposal problems and results in a more subdued alteration of the brine salinity of the lake. CANADA In Canada, there are two significant sources of lithium raw materials. One is located in the Preissac-Lacorne area of Quebec and the other in the Bernic Lake area of Manitoba. In the Preissac-Lacorne region, several pegmatites occur in an area approximately 8,000 feet long and 2,000 feet wide. The identified resources have been cal- culated at 15,000,000 tons, carrying 1.2 percent Li2O (Mulligan, 1965). In 1955, the Quebec Lithium Division of the Sullivan Mining Group estimated their resources to last 45 years at current production rates. The prop- erty is developed with an underground mine, a mill, and a chemical plant. Its operational capacity is 1,000 tons of ore per day with a rated capacity of 2,000 tons per day. Although inactive at the present time, the company did operate for a number of years when it supplied the Lithium Corporation of America’s North Carolina chemical plant. Following completion of the AEC con- tract, Quebec Lithium operated at reduced levels until 1965. The overcapacity in the lithium industry created by the 5 year noncommercial demand by the AEC, forced the closure of the operation. In Manitoba, Tanco began the exploitation of a com- plex zoned pegmatite which probably constitutes the world’s richest tantalum and cesium mine. The pegma- tite contains an uncommonly rich spodumene zone which averages almost 3 percent Li20. A previously un- known spodumene zone has been intercepted beneath the main spodumene ore body by exploratory drilling (Pearse, 1973). The resources of this new body are un- known. Resources for the main spodumene ore body are reported at 5,000,000 tons. In 1972, Tanco undertook a pilot plant study to evaluate the possibility of spodumene production. No significant commercial production has yet been achieved, as this company has been plagued by a number of financial problems. In 1974, Kawecki- Berylco Industries bought a 24.9 percent interest. In addition to their cesium and rubidium interest, Kawecki-Berylco plans the production of about 10,000,000 lb of lithium carbonate and about 15,000 tons of spodumene concentrates per year (Pearse, 1973). Production is scheduled to begin in 1977. BIKITA TINFIELDS, RHODESIA The Bikita ore body became important in the 1950’s as a result of the AEC program. In order to partially satisfy the needs of the program, American Potash and Chemicals processed Bikita lepidolite in Texas. Although United Nations sanctions have been im- posed on Rhodesian trade, eliminating an important source of petalite for the U.S. ceramic industry, Rhode- sian ores are still mined and sold. Resources have been reported at about 6,000,000 tons (Symons, 1961) with an average composite grade of 2.9 percent Li20. The pegmatites contain petalite, spodumene, lepidolite, and eucryptite. This appears to be a deposit similar to Bernic Lake in Manitoba. A chem- ical processing plant is, therefore. not inconceivable. ZAIRE The pegmatites of Manono and Kitotolo were dis- covered in 1910. Exploitation for cassiterite and tanta- lite began in 1929 when the Belgian company, Geomines, undertook the project. LITHIUM RESOURCES AND REQUIREMENTS BY THE YEAR 2000 A report by Barzin, Managing Director of Geomines, states that the pegmatites are both 5.5 km long and average 400 metres wide (Barzin, 1952). The pegmatite sills have been proven to a depth of 125 metres by drill- ing. A calculation using these dimensions shows that re— sources are enormous. Barzin reports, however, that up to 50 metres are kaolinized and altered. On the other hand, exploratory drilling ended in pegmatite, indicat- ing the presence of additional resources. Over the years, low cost mining of cassiterite and tantalite has been achieved from the upper decomposed layer. The peg- matites of Manono and Kitotolo probably represent the largest spodumene resource in the world. However, they have never constituted reserves because of their remote location in south-central Africa. The tin and columbium production is exported via the Angolan port of Lobito, located some 2,000 km away. Nevertheless, this deposit must be included in an appraisal of the fu- ture of lithium. In fact, the government of Zaire must think so, because plans for the production of lithium carbonate have been officially announced. The docu- ment states that Zaire will become the third producer of , lithium carbonate in the world. Zairetain (50 percent government controlled) is planning a production of 10,000,000 lb of lithium carbonate per year beginning in 1980. SALAR DE ATACAMA, CHILE The latest development in the field of lithium has been the announcement by the Chilean government of the discovery of a large brine field containing unusually high lithium values. The statement indicated concentra- tions of 2,000 ppm Li and resources on the order of 1,200,000 tons of lithium, presumably as Li. In january 1975, Foote Mineral Company entered into a contract with the Chilean government and has since undertaken a program of geological, hydrological, and chemical research in order to assess the feasibility of producing lithium chemicals from the brines of the Salar. The Chilean government is also interested in re- covering potash in order to develop its fertilizer industry. The geology of the Salar de Atacama and of the sur- rounding areas has been described in several publica- tions (Dingman, 1967; Chong, 1971; Moraga and others, 1974; Stoertz and Ericksen, 1974). The Salar has been interpreted as a graben (Frutos, 1972), bounded on the east by the Andean Cordillera and on the western side by the Cordillera de la Sal, a structural block which consists of evaporites including halite. The Salar lies at an elevation of 7,500 feet. The south- ern portion consists of a halite nucleus which occupies an area of 1,400 km? The surface of the nucleus is very rugged, and the halite occurs as razor-sharp jagged ridges. The area is extremely arid. Several streams drain 29 into the Salar. The largest of these is the Rio San Pedro which enters the basin from the north. Although the resources reported by the government have not yet been verified: the work Foote has done to date supports their earlier Claim. CONCLUSIONS The deposits which have been described do not con- stitute hypothetical resources of lithium. These are all reserves or conditional resources. Today, lithium chemicals are produced in North Carolina (Kings Mountain and Bessemer City); in Vir— ginia (Sunbright); Silver Peak, Nevada; Exton, Pennsyl— vania; New johsonville, Tennessee; and Searles Lake, California. Under favorable economic conditions, one may expect production from Quebec and from new cen- ters such as Manitoba, Utah, Rhodesia, Chile, and Zaire. An evaluation of the resources held in all these de- posits yields a total of 4,100,000 tons as Li. If we assume as little as 15 percent recovery from these resources, there is 617,000 tons of Li available (table 15). If only existing and actually announced capacity is considered, an annual production rate of 90,000,000 lb lithium carbonate may be reached by 1980. This repre- sents 9,000 tons as Li. If we assume a consumption of 200 kg of Li per one thousand MW(e) (Mills, 1975), a 500,000 MW(e) capacity would require 100 tons of met— al. This corresponds to the total 1970 U.S. electric power consumption. In addition, approximately 100,000 tons of metal will be needed in the reactor blan- ket. The projected industrial capacity could supply this quantity in 11 years. Obviously, the lithium industry will not devote 100 percent of its capacity for this exclusive use. However, since such a thermonuclear power capac- ity is predicted for the year 2020, the industry would have ample time to expand if and when this use ap- peared likely. The projected capacity suggests that commercial as well as thermonuclear power and battery requirement can be met by the lithium industry. As a final comment, one must again ask the question of the fate of the lithium industry after the peak de- TABLE 15.—Major identified resources of the world Short Ions Li Production Reference Tin-spodumene Belt. North Carolina, U.S.A ______ 473,000 Active ______________ Kunasz (1975): Evans (1976). C121 ton Valley, evada. U.S.A ______________ 775,000 _do ____________ Kunasz (1975). Great Salt Lake, Utah, U.S.A ________________ 526,000 Planned ____________ Whelan (1976). Preissac-Lacorne, Queber Canada ____________ 84,000 Temporarily Pearse (1973). inactive. Bernie Lake. Manitoba, Canada __________ 69.000 Announced ________ Pearse (1973). Bikita, Rhodesia ______________ 81,000 Active _________ __ Symons (1961). Atacama, Chile __________ ___ 1,200.000+ Possible ....... __ ______________ Manono, Zaire ________________ 1,000,000 + Announced ______________________ Total __________________ 4,208,000+ 30 mand for batteries and thermonuclear power has been met. A hasty increase in capacity installed to meet a projected demand based on tenuous data may have harmful consequences. Existing data suggests that re- sources are sufficiently large to satisfy projected higher lithium demands in the future. The existing and an- nounced capacity of the industry are likely to satisfy the demands for a considerable number of years. REFERENCES Barzin, H., 1952, Geomines—A major open-pit tin producer in the Belgian Congo: Eng. and Mining jour. v.153, no. 11, p. 86—89. Chong, G. D., 1971, Depositos Salinos en el Norte de Chile y el Salar de Atacama: Geochile, no. 3, p. 13—27. Dingman, R. j., 1967, Geology and ground-water resources of the northern part of the Salar de Atacama, Antofagasta Province, Chile: U.S. Geol. Survey Bull. 1219, 49 p. Evans, R. K., 1976, Lithium Corporation of America—ore reserves: Company news release. Frutos, j., 1972, Ciclos tectonicos sucesivos y direcciones estructurales sobreimpuestas en Los Andes del Norte Grande de Chile, in Sim- posio sobre los resultados de investigaciones del manto superior con enfasis en America Latina: Buenos Aires, p. 473—483. LITHIUM RESOURCES AND REQUIREMENTS BY THE YEAR 2000 Kesler, T. L., 1960, Lithium raw materials: Industrial Minerals and Rocks, [3d ed.], p. 521—531. Kunasz, I. A., 1975, Lithium raw materials: Industrial Minerals and Rocks, [4th ed.], p. 791—803. McKelvey, V. E., 1973, Mineral resource estimates and public policy, in United States mineral resources: U.S. Geol. Survey Prof. Paper 820, p. 9—19. Mills, R. G., 1975, Problems and promises of controlled fusion power: Mechanical Eng, no. 9, p. 20—25. Mulligan, R., 1965, The Geology of Canadian Lithium Pegmatites: Geol. Survey of Canada Econ. Re’pt” no. 21, 131 p. Moraga, B. A., Chong, D. G., Forttz, M. A., and Henriquez, A. H., 1974, Estudio geologico del Salar de Atacama, Provincia de An- tofagasta: Inst. Inv. Geol. Bol., no. 29, 56 p. Norton, ]. j., 1973, Lithium, cesium, and rubidium, the rare alkali metals, in United States mineral resources: U.S. Geol. Survey Prof. Paper 820, p. 365—378. Pearse, G. H. K., 1973, Lithium: Canadian Minerals Yearbook 1973. Stoertz, G. E., and Ericksen, G. E., 1974, Geology of Salars in North- ern Chile: U.S. Geol. Survey Prof. Paper 811, 65 p. Symons, Ralph, 1961, Operation at Bikita Minerals Ltd., Southern Rhodesia: Inst. Mining and Met. Bull., no. 661, p. 129—172. Vine, j. D., 1975, Are lithium resources adequate for energy self- sufficiency?: Am. Nuclear Soc. Trans, v. 22, p. 55—56. Whelan, ]. A., 1976, this report. m LITHIUM ORES By R. KEITH EVANS, GULF RESOURCES 8c CHEMICAL CORPORATION, HOUSTON, TX ABSTRACT Chemical differences and commercial advantages of the principal lithium ore minerals—spodumene, petalite, lepidolite, eucryptite, and amblygonite—are described. Ores and concentrates of these minerals are used directly in specialty glasses, glass ceramics, porcelain enamels, fritted glasses, raw glazes, ceramics (whitewares), and refractories. These lithium minerals can also be used in the production of lithium chemicals and lithium metal. The quality requirements for these uses are not as stringent as they are for “direct usage" ores. An estimate of the demand for “direct usage” lithium ores under conditions of normal trading suggests a probable demand between 3.2 and 4.5 million kg (7—10 million pounds) oflithium carbonate equiva- lent. The inherent uncertainties in making this estimate are discussed. Potential consumption of “direct usage" lithium ores would constitute a significant percentage of the total Free World lithium production if such Ores were more readily available. LITHIUM MINERALS Although lithium occurs in a variety of minerals, those of principal economic interest are spodumene, petalite, lepidolite and, to a much lesser extent, amblygonite and eucryptite. In the future, lithium- containing clays could assume some economic signifi- cance, but neither clays nor lithium-containing brines fall within the scope of this short paper. In terms of volume usage, spodumene is much the most important mineral; it constitutes the feed to the western world’s current and proposed major lithium chemical plants. In addition, it meets a high percentage of the current demand for “direct use” ore, or concen— trate for the manufacture of glasses and similar prod- ucts. The current bulk of sales for this purpose is as low-iron (about 0.1 percent Fe203) concentrates grading between 6 and 7 percent LiQO; historically significant tonnages of a quartz-spodumene intergrowth similar in chemical composition to petalite and grading about 4.5 percent Li20 have been sold, both for direct usage and chemical manufacture. Prior to the imposition of the United Nations sanc- tions against Rhodesia, petalite grading approximately 4.3 percent LI2O and 0.03 percent Fe203 was the prin- cipal lithium mineral used in the glass and related industries. New deposits have been brought into pro- duction, stimulated by the virtual absence of Rhodesian ore, but these have either been smaller and of poorer quality than the source they have sought to replace or have lacked the consistency of the grade of Rhodesian material. LITHIUM RESOURCES AND REQUIREMENTS BY THE YEAR 2000 31 In most applications petalite and spodumene of com- parable iron'content can be substituted for one another. The silica/alumina ratios are different, but in the case of glass manufacture, for example, these can be balanced by the volume of other additives. This interchangeabil- ity, however, is only true in cases of total melting; peta- lite behaves in a more refractory manner than spod- umene with no volume change accompanying its con- version from an alpha to beta phase and is the preferred lithium mineral in strictly ceramic applications. The more refractory nature of petalite makes it a marginally less attractive feed for chemical production. Following phase conversion, petalite retains a strong resistance to the fine grinding which is a normal prerequisite for effi- cient acid attack, whereas the phase change in spod- umene coincides with decrepitation. In the bulk of nonchemical uses the choice between the use of spodumene and petalite is dictated by the price of each per lithia unit and by the relative concen- trations of deleterious constituents. The latter vary with end use, but the key ones are iron, soda, and potash. Lepidolite differs from the two other major minerals in that it contains significant percentages (2—4 percent) of both fluorine and rubidia. The added fluxing powers of these make the ore an attractive one for certain glass and porcelain enamel compositions. Until recently lepidolite was the preferred source of lithia for a high percentage of the world’s monochromatic television tube glass. Although known high grade sources are rare, ade- quate flotation capacity exists to produce upgraded acceptable concentrates (3.8—4.3 percent Li2O) to meet a declining demand due to the increased concerns regard— ing fluorine emissions. Lepidolite was the ore feed for large scale lithium hydroxide production in the United States in the 1950‘s. Bikita, Rhodesia, is the only known source of eco- nomic tonnages of eucryptite where it is a co—product from the mineral assemblage there. The mineral in its alpha form has no current unique commercial applica— tion. Amblygonite rarely occurs in large tonnages but because of its high lithia content (8 percent Li20) is gen- erally saleable for use, particularly, in certain porcelain enamel formulations. Amblygonite is the feed for small scale chemical production in Brazil. DIRECT USAGE ORES Apart from the production of lithium chemicals and metal, the principal current uses of lithium ores are in glasses, glass ceramics, porcelain enamels, fritted glazes, raw glazes, ceramics (whitewares), and refractories. Glasses are formed when mixtures of inorganic sub- stances are melted and subsequently cooled in a manner which usually prevents crystallization. The main ingre- dient is normally silica sand but, to reduce the melting temperature to an economically acceptable level, an al- kali flux is added. This flux is normally in the form of soda ash; lime is added as a stabilizer. These soda-lime- silica glasses constitute the bulk of the world’s glass output. Small amounts of alumina, magnesia, boric oxide, and other chemicals are added to impart specific properties in response to market demands. Each glass is a com- promise with soda being the principal source of alkali because of its cost, inspite of the fact that its use intro- duces many undesirable qualities. Although lithium results in the lowest increase in un- desirable qualities, the difference in price and the toler- ance of these in normal applications generally restricts its use to specialized glasses. Specific advantages of lithia additions include the fol- lowing: Viscosity reductions; reduced melting tempera- tures; the reduction in the thermal expansion of the glass; increased density resulting in good electrical prOperties and high chemical durability; increased sur- face hardness; and other improvements which cannot be commented upon with brevity. The use of lithia in container glasses is limited and generally restricted to those few low-tonnage types that require a high thermal shock resistance. Glass lenses subject to rapid temperature fluctuations, such as many sealed beam headlight glasses and borosilicate opal glasses, contain lithia. High-barium monochromatic television tube produc- tion results in a major demand for lithium ores, princi- pally as a melting aid. Similarly in much foam glass production, lithium is utilized as a replacement for other potentially polluting additives. There is increasing usage of lithium in the glass fiber industry. Lithium is added to the glass batch for the production of large telescopic mirrors to help the glass withstand a pro- longed annealing process. It is also added to a wide variety of other low tonnage glasses, where one or more of the unique properties of lithium are necessary. Glass ceramics are nonporous crystalline materials de- rived from noncrystalline glass containing nucleating agents. When reheated under very careful control, the nuclei act as centers of crystal growth. Many formula- tions have been devised, but the most commercially im- portant are those for end products requiring a high thermal shock resistance. These are produced in Europe, the United States, and japan and the resulting demand for lithium is substantial. Originally the de- mand was almost exclusively for high quality ores. The shortage of good grade ores coupled with increasing quality demands, especially in respect to low potash and soda contents, has resulted in a partial replacement of ore by lithium chemicals, particularly lithium carbonate. 32 Vitreous or porcelain enamels are essentially fritted glasses that are fused to metals for decoration and cor— rosion resistance. Properties attributable to the use of lithia in glasses apply equally to enamels; lithia is a common constituent in a wide range of formulae. A high percentage of domestic United States’ lithium demand for these applications is as lithium carbonate. This is because the large volume formulae are low in alumina, and the addition of lithia as ore minerals re- sults in a significant and intolerable alumina addition. Elsewhere, low-alumina enamels are less common; pro- viding the ore is of acceptable purity, the choice of the lithia source is based on the relative costs of ore and carbonate. Glazes in general vary in type from “raw” to fully fritted, with composition differences from enamels which generally allow longer firing times. With increas- ing pressure against the use of lead oxides in glazes, there is a continuing increase in lithium usage because of its low fusion point and low viscosity, to provide bril- liance, luster, and smoothness. In terms of tonnage, lithium ore usage is greatest in glazes on low expansion ceramic bodies. The benefits of adding lithia to ceramic (whiteware) bodies to allow rapid firing have been well publicized. Apart from some low volume specialized ceramics, usage is small, except in Japan where large tonnages of petalite are used in inexpensive ovenware. There appears to be a great potential for lithium ore usage in the ceramics industry for domestic and indus- trial use. In the domestic field, technical problems exist in producing glazes with coefficients of thermal expan- sion to match the very low thermal expansion ceramic bodies possible with the use of petalite. Finally, a moderate amount of lithium ore is used in some refractories and kiln furniture. ORES AS CHEMICAL PLANT FEED All of the principal lithium minerals are, or have been, used as feedstock for lithium chemical production in the United States (spodumene, petalite, and lepido- lite), Canada (spodumene), Brazil (amblygonite), Eng— land (petalite and eucryptite), West Germany (petalite and spodumene principally), Italy (spodumene and spodumene—quartz intergrowth), and Japan (petalite). The remaining operations in other countries are small, while others have been abandoned completely as uneconomic. The closures, in part reflect the cyclical nature of the industry in its earlier years, but also dem- onstrate the generally inherent uncompetitiveness of LITHIUM RESOURCES AND REQUIREMENTS BY THE YEAR 2000 lithium chemical plants located away from their ore sources. Even the theoretical maxima of lithia contents in the common minerals are low. Pure spodumene grades a theoretical 8 percent LigO; 7 percent material is produc- ible if one is prepared to sacrifice recovery, but optimisa- tion of recovery and grade generally results in the pro- duction of chemical feed concentrates grading between 5 percent and 6.5 percent Li2O. Petalite and lepidolite normally grade between 3.5 percent and 4.5 percent LiQO, and this cannot be improved upon. Quality requirements for chemical feed are not as stringent as for direct usage ores. Iron content is not a major concern, but other elements such as soluble alumina, phosphates, fluorine, and particularly mag— nesium result in complicated processing requirements and higher costs. DEMAND Approximate lithium ore requirements for lithium chemical production can be translated back from pub- lished chemical production statistics by making reason— able assumptions regarding recoveries in milling and processing. Direct usage ore demand is more difficult to estimate because nearly 90 percent of the Free World’s supply originated from a source that is no longer open to the major consuming nations. There are uncertainties concerning the decline in ore demand in certain key consumption areas such as (1) monochromatic television tubes and enamels, (2) the problem of more stringent specifications for certain high technology products resulting in an increasing de- mand for lithia as a chemical rather than an ore, (3) problems concerning fluorine emissions resulting from lepidolite usage, and (4) others which may be offset by increased ceramic demand. Finally there is the question of the relative prices per lithia unit of ores and lithium chemicals. My personal opinion is that in normal trading condi- tions current lithium ore demand for uses other than chemical production would amount to between 8,000 and 12,000 tonnes/year for lepidolite (on a 4 percent LigO basis) and between 40,000 and 50,000 tonnes/year for petalite or spodumene if reduced to a common 4 percent Li2O basis. These are approximately equivalent to lithium carbonate demands of between 3.2 and 4.5 million kg (7.0—10.0 million pounds) a year and, thus, represent a significant percentage of total current lithium demand. \_/-\ LITHIUM RESOURCES AND REQUIREMENTS BY THE YEAR 2000 33 LITHIUM PRODUCTION FROM SEARLES VALLEY By LARRY E. RYKKEN, KERR-MCGEE CHEMICAL CORPORATION, TRONA, CA ABSTRACT Lithium occurs in Searles Lake brine, which is processed primarily for other minerals. A portion of this lithium is recovered as a crude concentrate in a two-stage foam-flotation process. This concentrate (dilithium sodium phosphate) is then converted to a relatively pure lithium carbonate by digestion using sulfuric acid and conversion using sodium carbonate. INTRODUCTION Lithium in the form of lithium carbonate is produced as a byproduct of the Trona, Calif, Plant main-plant- cycle process. In this process, a dry—lake brine, pumped from permeable salt beds, is essentially evaporated to dryness in triple-effect evaporator units. Potash (KCI), borax (Na2B4O7'10H20), salt cake (NaQSO4), and soda ash (N32C03) are recovered by a series of fractional crystallizations in a cyclic process. Lithium in the form of dilithium sodium phosphate (Li2NaPO4) is separated from a process stream by foam flotation. This separation is part of the refinement process for the production of salt cake and soda ash in the soda products section of the main plant cycle. The lithium values in the dilithium sodium phosphate are then converted to a refined lithium carbonate (Li2C03) product, while the phos- phate values are converted to a 78—percent-crude or-‘ thophosphoric acid (H3P04). Annual production is ap— proximately 153,000 kg as lithium (900 tons/year of lithium carbonate), which represents recovery of ap- proximately one-third of the lithium contained in the brine fed to the main plant cycle. DEVELOPMENT The first attempts to remove part of the lithium- bearing slimes from the soda-products liquors were prompted by the problems they presented in processing (foaming of liquors, poor filtration, product contamina- tion) rather than by their potential sales value. The re- covery of lithium concentrates (licons in plant terminol- ogy) was started in 1936 by processing the foam from a process vessel on a campaign basis. Liquor was drained from the tank, water added, and the slurry circulated several hours through a line in which steam was injected to remove entrained airsThe wet licons were removed on a Sweetland filter press, washed with water, partially air-dried, scraped from the press leaves, and hauled to drying plates, where the material was sun-dried for a week or two and them sacked for shipment. Only minor improvements in this process occurred until May 1942, when the Federal War Production Board issued an urgent request to triple the lithium tonnage. An intensified research and development pro- gram was inaugurated, which resulted in the design and construction of a commercial-scale flotation plant in 1943. The original plant has undergone considerable {revision since then, although the air—flotation principle “as been retained. The dried licons were sold until late 1951 when the lithium carbonate plant in Trona began operation. This plant was plagued by heavy tarlike gums, which formed when the licons were digested with sulfuric acid. This problem resulted in the addition, in 1954, of a Her- reshoff furnace for burning Off the organic material from the licons. The last major addition was a second- stage flotation unit in 1956; only minor refinements to the process have been made since then. FLOTATION-PROCESS DESCRIPTION The Searles Lake brine contains from 0.007 percent to 0.008 percent lithium (~0.015 percent Li20). Most of the lithium-bearing salt is precipitated in the main- plant-cycle evaporators and accompanies the burkeites through the classification and dewatering steps. The burkeites are double salts of sodium sulfate and sodium carbonate, which constitute the feed to the soda prod— ucts section. After further processing a burkeite filter cake is dissolved in water, leaving the relatively insoluble lithium-bearing solids in suspension. The exact molecu— lar structure of the lithium salts is in question, but the constituents closely conform to the formula LigNaPO4 or dilithium sodium phosphate. It is relatively insoluble in neutral or basic water solutions, but very soluble in even slightly acidic solutions. The lithium-plant flotation process uses soap as a flo— tation agent. The soaps are derived from coconut-oil fatty acids, which are added to the alkaline liquors at the main-plant-cycle evaporators. The fatty acids serve an additional function as foam-control agents in the evaporators. The sodium soaps of the fatty acids are believed to coat the near-colloid licon particles, making them relatively insoluble. Residence time is minimized between the solution of the lithium-containing burkeite and the flotation because the amount of dissolved lithium increases with time. The best recovery is ob- tained from a near—saturated solution of burkeite at a 34 temperature just above the crystallization temperature of glauber salt (~809F). Water addition at the burkeite dissolver is controlled by the density of the liquor leav- ing the dissolver. The dissolver liquor is cooled by pass— ing it through two induced-draft spray-cooling towers. Flotation is carried out in two stages, with air added at both feed-pump suctions. The air is compressed by a pressure-controlled surge chamber at the discharge of each feed pump. Release of pressure in the flotation units results in the formation of a multitude of tiny air bubbles, which collect the soap-coated licon particles and float them to the surface. The first-stage flotation takes place in four 10,000- gallon (37,850—l)—capacity tanks. The tanks contain an inner cone-bottomed tank. The liquor is released tangentially along the bottom of the outer tank, rises through the annular space surrounding the inner tank, and then sinks downward in the inner tank. The lithium froth overflows the lip of the outer tank into a foam trough, and the liquor is discharged from the bottom of the inner cone-bottomed tank. Liquor levels are au- tomatically controlled and maintained within 2~3 inches (5—7.6 cm) of the top. The second stage consists of an industrial pneumatic foam separator. The aerated feed liquor (discharge from first stage) is released inside a hooded skirt in the center of the unit: it spreads to the outer wall, then down and under the wall and up through risers to dis- charge into a circular launder. Lithium froth forms on the surface and is removed by a sweep. Additional foam is obtained from a sweep on the surface of one of the soda-products process tanks, which is kept at a constant liquor level. The foam is collected in tanks where it is agitated and heated to break down the foam. The lithium salt is fil- tered from the foam—storage tanks in a batch process using two Sweetland filter presses. The cake is washed, dried with an air-stream mixture, and removed into hoppers. Conveyor belts carry it to the licons roaster, where the organics are removed by roasting and bleach- ing with sodium nitrate in a six-hearth, gas—fired Her- reshoff furnace. LITHIUM-CARBONATE PROCESS In early 1945, work was started on the production ofa comparatively pure salt of lithium in place of the crude lithium concentrates. Several possibilities were investi— gated, but a process for the production of lithium car- bonate was finally decided upon. The availability of sodium carbonate as an in-house product for the con- version and the marketability of the carbonate form of LITHIUM RESOURCES AND REQUIREMENTS BY THE YEAR 2000 lithium were factors in this decision. Actual production did not begin until 1951. In the lithium-carbonate process, the licons are first digested with concentrated (93 percent) sulfuric acid. The digestion reaction may be represented as: 2 L12N3PO4+3H2SO4_‘)2 L12 SO‘4+N32SO4+2H3PO4. This reaction is accomplished in a 6,000-gallon (22,710- l) tank equipped with a top-entering agitator and a double-walled inner tank or pachuca. The double- walled pachuca serves as the heat-exchange surface for both heating during the digestion of licons and cooling the batch preparatory to the separation of the precipi- tated solids. The lithium and sodium sulfates (called mixed sulfates) are centrifuged from the digester tank, washed, and dissolved in water. The filtrate is a 45—50 percent orthophosphoric acid solution, which is evapo- rated down to a 78 percent phosphoric acid solution. The evaporation reduces the lithium in solution from about 3 percent as Li20 to less than 1 percent, thus increasing recovery. After evaporation, the solids (Li2504 and Na2804) are settled out and recycled as a sludge back to the digester. The dissolved, mixed-sulfate solids are introduced into a saturated solution of sodium carbonate at a tem- perature of 200°F. The lithium carbonate has a solubil- ity of about 0.6 percent as L120 versus about 4 percent for the lithium sulfate. The lithium carbonate precipi- tates out according to the reaction 1.5304(1) + Nagcogapmgcogt + Na2SO4(l). The lithium-carbonate solids are centrifuged out, washed, dried in a steam-heated rotary-drum air dryer, and packaged in 50-pound (22.7-kg) sacks and 325- pound (l47-kg) fiber drums for shipment. Lithium re— covery is increased by treating the end liquor (lithium carbonate centrifugal filtrate) using dilute phosphoric acid and evaporating to near —Na2SO4 saturation. The phosphate ions cause precipitation of Li3PO4 as a result of the decreased solubility of Li3P04 (<. 15 percent Li20 versus 0.6 percent Li20 for Li2C03). The reaction is 3 Li2C03(l)+2 H3PO4(l)—>2 Li3P041+3 co2T+3 H20. The Li3P04 solids are settled out and recycled to the digester. Lithium is also recovered from spillages by pumping sump liquors back to the end-liquor evaporator. “Theoretical” lithium losses are reduced to those dis- solved inconcentrated phosphoric acid and treated end liquor. Theoretical recovery is 92—94 percent, whereas actual recovery has averaged about 88 percent. \/\ LITHIUM RESOURCES AND REQUIREMENTS BY THE YEAR 2000 35 THE LITHIUM-RESOURCE ENIGMA By JAMES D. VINE, LLS.GEOUMHCALSURVEY,DENVER,CC) ABSTRACT The identified lithium resources of the United States that may be recoverable by the year 2000 will probably yield less than one million tonnes. About a third of this will be used in conventional ways, leaving about two-thirds for new energy-related uses, including use in bat- teries for electric vehicles. Exhaustion of our recoverable lithium re- sources for these uses would limit the development of thermonuclear power in the early decades of the next century. Depletion of a com— modity essential to future energy independence can best be prevented by identifying new resources oflithium. A program of exploration for new deposits well in advance of the demand would help to assure a stable price and a steady growth and would be of benefit to the Ameri- can public as well as the lithium industry. The best chance of finding new deposits of lithium is in association with nonmarine salt bodies. Undiscovered resources may exist in other geologic environments that have not as yet been evaluated. A modest U.S. Geological Survey effort to search for such deposits shows promise. THE NEED FOR RESOURCE DATA Previous estimates of the availability of lithium tend to be misleading to those who plan to use this rare metal in alternative energy applications. Reasons for an overly optimistic outlook are twofold: (1) Geologic estimates of resources (for example, Norton, 1973) summarize data 'on amounts of lithium in the ground without regard to recoverability and (2) all recently published estimates are biased by an unrealistic preliminary forecast of ,lithium available from one large brine field in Nevada. Although brine fields account for a large part of the lithium resource base, they will represent a much small— er fraction of the identified lithium recoverable by the year 2000. That is because, in practice, a large part of the resource base is inaccessible as a result of technical or other problems. Many of these other problems, in- cluding economic, political, and environmental restric- tions, can be as difficult to overcome as any technical problem. An accurate forecast of the amount of lithium avail- able to the market place rather than a measure of the lithium contained in various inaccessible parts of the earth is desirable, in order to plan for the effective de— velopment of lithium batteries for automobiles and for storage of off-peak power for utility nets. Furthermore, if the scarcity of the Lib- isotope, which is required in the production of tritium, is a limiting factor in the de- velopment of controlled thermonuclear power, as sug- gested by Hubbert (1969, p. 230—233), consideration should be given to stockpiling this isotope prior to using the rest of the lithium for batteries and other conven- tional uses. Any mineral industry, such as the lithium industry, is concerned with the quantity of ore that can be profitably extracted from the ground.1 This is the category of re- sources that we call reserves, and it may represent only a small fraction of the total resource base. Most companies will attempt to block out an ore reserve adequate to supply 5—10 years of mill product, or whatever is re- quired to amortize the capital investment in mining and milling facilities. Under the pressure of competition they may tie up additional mineral leases, but most companies cannot afford the investment required to hold leases that they do not anticipate using for a few years. Even if available to the government, the-proprietary data on company-held reserves would not be adequate for long-range government planning. Thus, from the broader perspective of the long-term economic health of the nation, various agencies of government may need data on the total resources of a commodity, including company-held reserves, identified but subeconomic re- sources, and the undiscovered resources that may be predictable on the basis of geologic probability. Rea- sonably accurate forecasts are required for many differ- ent uses—from the ranking of national priorities to military and diplomatic strategy. REVISED LITHIUM RESOURCES A classification of lithium resources that includes an estimate for three main categories of resources, revised to include information up to February 1976, is shown in table 16. The categories include (1) identified reserves, which are economic at today’s prices and technology; (2) identified low-grade resources, which are not economic at today’s prices and technology but may become economic at a higher price or improved technology, and (3) hypothetical and speculative resources, which will have to be discovered and developed in order to bring the cumulative yield by the year 2000 up to 1 million metric tonnes. All classes show both the “in place” re- sources, which refers to the amount of lithium that is in the ground, and the probable yield, which refers to the 'The latest definitions for various categories of reserves and resources can be found in U.S. Geological Survey Bulletin I450—A. 36 TABLE 16.—Cla33ificati0n of U .S. lithium resources in millions of tonnes of contained lithium Estimated percent In recoverable Probable Sources place by 20001 yield Reserves (economic) Kings Mtn., N.C.2 ______________________ 0.510 35 0185 Clayton Valley, Nevi‘fie .080 50 .040 Searles Lake, Calit" _-_ .009 50 .005 Total ,,,,,,,,,,,,,,,,,,,,,,,,,,,, 0.599 0.230 Low-grade resources (subeconomic) Kings Mtn., N.C.2 ,,,,,,,,,,, 0.840 25 0200 Clayton Valley, Nev“, .625 10 .060 Searles Lake, Calif‘ _ .027 10 .003 lm rial Valley, Califs , 1.000 1 .010 Oil ield brines ccccccccccccccccccc 1.000 1 0.10 V.» Great Salt Lake, Utah’_-c ________ .2762), 5 .010 Borate mine waste” _______________ .060 7 .004 Black Hills, SD.‘ “A ,,,,,,,, .010 30 .003 Total ____________________________ 3 825 0.300 Hypothetical and speculative resources” Clays __________________________________ 0.800 25 0.200 Brines ____________ _-_. 1.000 20 .200 Pegmatites .___ u» .350 20 .070 Total ____________________________ 2.150 0.470 ‘ various sources, including hearsay. 2“In place" reserves estimated for two known ore bodies (Keith Evans and Ihor Kunasz, oral commun.,]an. 1976). Total resources for the district from Norton (1973, p. 372). ICalculated from the estimate of recoverable reserves and total resources listed by Kunasz (1975, p. 796). ‘Calculated from Norton (1973, p. 372). 5Extrapolated from two analyses and the geometry of the brine field suggested by White (1968, p. 312—313). “Based on analytical data of Collins (1974, p. 16—20). No data on brine volume. 7Estimate by j. D. Whelan (oral commun., Jan. 1976) incorporating recent downward revision of the volume of brine in the lake by Carol A. Peterson. 8Calculated from data presented by Robert B. Kistler (oral commun.,]an. 1976). 9This amount, in addition to reserves and low-grade resources, must be discovered and developed to produce a cumulative yield of 1 million tonnes in the years 1976 through 2000. amount oflithium one might expect to be recovered and made available to the market place. While the in—place resources are a measured volume of rock times the grade of ore, the accuracy of which may be good (:10 percent) or poor (: 100 percent), the factor of recovera- bility is not a measurable quantity. Recoverability in— volves unpredictable economic and technical factors that can be determined only in practice. Although the accu- racy of the resulting data may be questioned, the totals probably represent a reasonable value. For example, a marked price increase relative to other products and services could increase the total yield by a factor of 10 or 20 percent, to more than a million tonnes oflithium, but additional restrictions on the use of land for mining could reduce the total by an equal or greater amount. Although the existence of hypothetical and speculative resources can be predicted on the basis of geologic probability, the quantities listed here are uncertain and will be a function of the imagination applied by geologists and the incentives available for exploration. It seems reasonable to forecast that the United States will produce somewhere between 750,000 and one million tonnes of lithium by the year 2000, but such production LITHIUM RESOURCES AND REQUIREMENTS BY THE YEAR 2000 will require an economic climate that provides incentive for considerable exploration and development of lithium in previously unknown deposits. COMPARISON OF RESOURCES AND REQUIREMENTS In the years 1971 through 1974, conventional uses for lithium grew at an average rate of 8.8 percent, accord— ing to proprietary data. If this figure is projected into the future, conventional uses may consume about 3 ><108 kg of lithium by the year 2000. Requirements for new energy—related uses for lithium, including batteries for electric automobiles, batteries for the off-peak storage of power for utility nets, absorption—type refrigeration and air-conditioning units, and controlled thermonu- clear power plants, are much more difficult to predict. Requirements for these uses depend on the success of current technologic research and its acceptance by the American public. Estimates of these requirements, as detailed in a previous report (Vine, 1975b) are from 3 to 15><108 kg for battery applications alone. Thermonu- clear power plants are not expected to require much lithium before the year 2000 but may need as much as 5><108 kg by the year 2020. Thus, about one-third of our presently identified re- coverable resources will be consumed by conventional uses by the year 2000. The remaining two-thirds could easily be spent in battery applications before the year 2000, leaving nothing for thermonuclear power, unless we start now to search for and develop new deposits. Failure to identify new resources of lithium may be re- garded as irresponsible by those of the next generation. UNDISCOVERED RESOURCES An appraisal of lithium resources would be incom- plete without discussion of undiscovered resources, which include extensions of identified resources in known districts plus identification of new districts previ— ously unknown. Whereas the rate of discovery of new oilfields can be extrapolated from the past history of exploration and discovery for petroleum, no such rec- ord of exploration exists for the lithium industry. Moreover, the number of known districts is too small to constitute a statistical basis for projection. Even pros- pecting guides to aid in the exploration for lithium are minimal. Small lithium-bearing zoned pegmatites, such as those in the Black Hills of South Dakota, are generally characterized by large crystals, as much as 1—2 metres in length; their size makes them easy to recognize and un- likely to be overlooked when exposed at the surface, but their irregular distribution makes prediction of covered pegmatites nearly impossible. Unzoned pegmatites, such as those in the Kings Mountain tin-spodumene belt LITHIUM RESOURCES AND REQUIREMENTS BY THE YEAR 2000 37 of North Carolina, are large features that may extend for many kilometres; hence they are unlikely to be over- looked, except in areas of deep weathering or jungle cover. Thus, the chances of discovering a major new pegmatite district within the conterminous United States seem rather poor. The chances of discovering a new brine field, on the other hand, are relatively good, because relatively little effort has been devoted to searching for brines. Furthermore, lithium-rich rocks or minerals may be associated with nonmarine salt bodies, volcanic rocks, and vein deposits, none of which have received more than casual attention with respect to lithium (Vine, 1975a). . Whereas, by definition, reserves are identifiable, mea- sured amounts ofa commodity, the concept of resources is abstract in that the values can change with changes in economics, technology, and geologic understanding. For example, the possibility of discovering previously unrecognized lithium resources is very good because previous exploration was minimal and limited to a few geologic environments. Hence, a program of explora- tion for lithium in previously untested geologic envi- ronments could greatly improve the outlook for lithium resources. If the future growth of the lithium industry is de- pendent on presently identified resources, a considerar ble price increase in lithium can be predicted, which will bring today’s subeconomic resources into the category of economic reserves. However, any marked price in- crease would seriously limit the economically feasible uses of lithium, especially in direct consumer products such as electric automobiles. On the other hand, an ex- ploration program that appraises previously untested geologic-Environments may result in a more optimistic forecast of resources because of the probability of dis- covering new deposits. Thus, a program to search for such deposits could make the difference between rea— sonably priced electric automobiles and some other form of personal transportation that might be too ex- pensive for the average consumer. Although private industry is capable of conducting a program of exploration for lithium, little incentive to do so exists at the present time, because industry has suffi- cient reserves for the short term and no assurance that major investments in exploration would ever be re- turned. Meanwhile, large amounts of money are being spent to develop new energy-related uses for lithium, partly based on the assumption that lithium will be available at a reasonable price when the manufacturers are ready to start up production lines. However, if new resources are not identified and developed well in ad- vance of this need, any sudden increase in demand will be met by a spectacular rise in price which will justify having to mine and mill lower grade lithium ore. A crisis in demand does not permit the time required to develop new resources. Moreover, a spectacular price rise would be self-defeating because less desirable but less expen— sive substitutes could be found to replace lithium in bat- teries. A stable price for lithium would be beneficial to both the industry and the American public. The best way to assure both a stable price and a steady growth in the demand for lithium would be to conduct a program for the exploration of new deposits well in advance of the actual increase in demand. A modest program along this line has already been initiated by the US. Geological Survey and has begun to produce encouraging results. Those of us involved in this program solicit both the confidence and encouragement of the lithium- producing industry as well as the support of the scien- tists and engineers from national and private labora- tories who are concerned with the development of new energy-related uses for lithium. REFERENCES CITED Collins, A. G., 1974, Geochemistry of liquids, gases, and rocks from the Smackover Formation: U.S. Bur. Mines Rept. Inv. 7897, 84 p. Hubbert, M. K., 1969, Energy resources, [Chap] 8, in Resources and Man, a study and recommendations by the committee on Re- sources and Man of the Division of Earth Sciences, National Academy of . Sciences—National Research Council: San Francisco, W. H. Freeman and Company, p. 157—242. Kunasz, I. A., 1975, Lithium raw materials, in Lefond, S. j., ed., In- dustrial minerals and rocks [4th ed.]: New York, Am. Inst. Min- ing, Metall, Petroleum Engineers, p. 791—803. Norton, 1. _]., 1973, Lithium, cesium, and rubidium—the rare alkali metals, in Brobst, D. A., and Pratt, W. P., eds., Unified States Mineral Resources: US. Geol. Survey Prof. Paper 820, p. 365— 378. Vine, j. D., 1975a, Lithium in sediments and brines—How, why, and where to search: U.S. Geol. Survey jour. Research, v. 3, no. 4, p. 479—485. Vine, J. D., 1975b, Are lithium resources adequate for energy self- sufficiency: U.S. Geol. Survey open-file rept. 75—682. White, D. E., 1968, Environments of generation of some base-metal ore deposits: Econ. Geology, v. 63, no. 4, p. 301—335. N 38 LITHIUM RESOURCES AND REQUIREMENTS BY THE YEAR 2000 LITHIUM RESOURCE ESTIMATES—WHAT DO THEY MEAN? By JAMES ]. NORTON, U.S. GEOLOGICAL SURVEY, DENVER, CO ABSTRACT Lithium resources have been poorly known, mainly because there has been little need for information. The supply of lithium in known deposits, minable under prevailing conditions, has been far more than adequate to meet demand. Potential resources are very much larger. Most of the major mines are in deposits that had been discovered long before they were developed or were found more or less by chance. Systematic exploration for new deposits, especially new kinds of de- posits, has been meager, and thus experience with exploration techniques is slight. Anticipated large demand, now centering on lithium for storage batteries, greatly changes the outlook but can probably be met, though to justify the costs of exploration and development requires assurance that the demand will in fact materialize. Mutual understanding among those who will participate in enlarg- ing the supply of lithium and those who develop new uses is a neces- sity. Confidence arising from large resource estimates can be seriously misplaced; the resources must first be developed and made ready for market. The most reliable category of resources consists of the lithium proved to be recoverable under current conditions from known de- posits. Almost equally important in an expanding market is lithium in known deposits that is almost but not quite minable now. The large expected demand, however, can be met only if new deposits are found, as they probably can be. BACKGROUND The decision to publish the articles in this volume was triggered by misunderstandings of lithium resource es- timates, by shortcomings in published statements, by a tendency to misuse the estimates because Of inadequate comprehension of what underlies them, and by need for better communication among those who plan future uses of lithium, those who do the geologic work, and those who mine or will mine lithium. An article of mine (Norton, 1973) is at the center of some of the problems. Many of the users of that article are not mining engineers or geologists. They have tended to quote the resource estimates without studying the text of the article to understand the data on which the estimates are based. But I too have been at fault, in that article and in personal communications, by accepting premature company announcements of lithium reserve estimates for Clayton Valley, Nev.,‘which later proved to be tOO large and which seriously distorted the total estimated reserves for the United States. One awkward aspect of compiling mineral resource estimates is that companies ordinarily do not reveal the data on which their calcula- tions are based, and the shortcomings in their an- nouncements may not be visible even to themselves. In a generally overlooked paragraph (Norton, 1973, p. 375), I protested against a tendency to call my lithium resource estimates conservative in disregard of the fact that they have been the largest that available data would allow, but added that any truly large new demand would cause a search for new deposits and probably still larger estimates of resources. The conclusion of the paragraph emphasized the necessity of finding deposits and deter- mining their lithium content before any large use actu- ally begins. Such uses now appear tO be forthcoming. Hindsight shows that my 1973 article was defective in not being written so as to be more useful to persons not connected with mining or geology. A particular shortcoming was the absence of discussion of the outlook for lithium min- ing under various sets of possible circumstances. Lithium almost certainly can become in ample supply to meet even the largest needs now foreseen for the remainder of this century, but the end of the century is not far away. The problem of how to bring demand and supply together at acceptable costs must be understood by all persons involved or likely to become involved with lithium. MINERAL RESOURCE CLASSIFICATION AND NOMENCLATURE The word “reserves” is widely misinterpreted as in- cluding all material expected ever to come to market. Idiosyncratic usage of the term is not unusual among mining engineers and geologists, and the published record contains definitions that are incompatible with one another. More rigorously defined, “reserves” include only material that can be profitably mined, or at least is ex- pected to be profitable, under present economic and technologic conditions, and for which considerable in- formation has been obtained. The drilling, geologic in- vestigation, chemical analyses, and other work necessary to outline reserves are ordinarily time consuming and costly: Reserves may be said to be “bought” rather than “found,” though at some single point in the process there is, of course, a discovery and at the same time or a later time there is a decision that profitable mining is possible. LITHIUM RESOURCES AND REQUIREMENTS BY THE YEAR 2000 39 “Resources” is the term used to include all material, discovered and undiscovered, profitable and now un- profitable, which seems likely to become a source of a mineral commodity. The nature of the various categories of mineral re- sources has been greatly clarified by what has come to be known as the McKelvey diagram, a simplified version of which is shown as figure 18A. This is the basis for the current US. Department of Interior nomenclature and classification of mineral resources, most recently de- scribed in comprehensive form by the US. Geological Survey (1975). Among the many other publications dis— cussing the diagram are McKelvey (1972), Brobst and Pratt (1973), and Schanz (1975). Resources may, as a first step, be divided into four categories: (1) those classified as reserves because they have already been discovered and can now be worked at a profit, (2) those yet to be found that would also be profitable, (3) discovered material of too low quality to be used, and (4) material that is both undiscovered and unprofitable. Reserves is the most important category for the next few years. The undiscovered but profitable category generally takes second rank, especially when the time span is some decades. Yet under special circumstances the known low-quality deposits can take over the second or even the first position, as during times of great in- crease in demand or price, or after a major technologi- cal change, or when conventional sources are cut off by war or embargoes. The necessity for speed in mining may then be too great to allow time to explore. An example for lithium was the intensive mining in the 1950’s of deposits in granitic pegmatites that previously had been only moderately attractive. The possibility of finding nonpegmatitic deposits was already suspected, but there was no time to look for them. If demand again Discovered Measuredl Indicated ] Inferred Undiscovered In known districts ] In undiscovered districts 1 // , at a profit -._. at aprofii ] Paramargi’nal ‘LI Not currently Submarginal INCREASING ECONOMIC FEASIBILITY 0F MlNlNG—§ l l l | DECREASING GEOLOGIC INFORMATION—> FIGURE ISA—Classification of mineral resources. Modified from McKelvey (1972, fig. 3). increases greatly and if exploration of brines, clays, and so on is insufficiently intense or is unsuccessful, peg- matitic deposits will again assume their earlier impor- tance, despite their probably higher costs of mining. Industry’s interest is generally in or near the category represented by the rectangle labeled reserves on figure 18A. Most industry reserve estimates are for single mines or small groups of mines, chiefly for the purpose of acquiring the information needed to plan operations and to keep financial arrangements on a firm footing. Government has broader and longer term respon- sibilities, and thus tends to be concerned in considerable degree with other segments represented by the dia- gram. Reserves are customarily divided into smaller categories on the basis of how much is known about the position, size, shape, grade (ordinarily expressed as the percentage of a metal in the deposit), amenability to processing, and other factors necessary to appraise a deposit. In figure 18A, reserves are divided into classes called measured, indicated, and inferred, following the prac- tice of the US. Department of Interior. The mining industry prefers the terms “proved,” “probable,” and “possible” ore (Banfield and Havard, 1975). Measured and proved are generally agreed to have nearly the same meaning. They refer to ore that is so well known as to be ready for development and extraction. Neither the words “probable” and “indicated” nor “possible” and “inferred” seem to be equivalent, but the issue does not need to be discussed here. A formerly unprofitable deposit is promoted to the reserve category usually because of higher prices, technological advances leading to lower costs, or an in- crease in demand permitting larger scale and hence more efficient operation. A deposit can also cease to be profitable and lose its standing as a reserve, as by a de- crease in price caused by lowered demand. Many other circumstances can cause a body of ore to move into or out of the reserve category or even to be a reserve for one company and not another. Construction of public highways or railroadsvcan make mining more feasible. A deposit can become of value through an op- portunity for cheap transport of the product in ships; loss of such an opportunity can return the deposit to worthlessness. After the gold mine closing order of World War II, many operators of small gold mines were unable to keep their facilities in working order and never reopened the mines; their unmined ore ceased to be reserves. Other governmental actions, such as start— ing or ending a price support program, can move de- posits from one category to another. The opening of better mines can cause other mines to become unprofit- able, as has happened twice with lithium: (1) in the 40 LITHIUM RESOURCES AND REQUIREMENTS BY THE YEAR 2000 1950’s, when greatly increased demand led to the open- ing of large mechanized mines and ultimately to the closing of the old small handsorting mines and (2) when the Clayton Valley, Nev., brine operation caused the Quebec Lithium mine, near Barraute, Quebec, to be no longer competitive. Some mineral deposits may be prof- itably worked by a company that already has a large plant nearby but not by a company that must build new facilities. Other possibilities can be conceived by anyone who gives a little thought to the subject. What constitutes a profit introduces questions. A min- ing company expects a rate of return commensurate with its risks. Certainly the prospective return must be greater than the yieldl of ‘high-grade bonds to be attrac— tive. Yet once a mine has been developed and a plant built, mining will continue as long as cash income ex- ceeds cash outgo, even if part or all of the capital cost must be written off. Individuals making reserve calculations may have dif- ferent motives that can lead to different results from essentially the same data. Congressman Harold Runnels (US. Congress, 1975, p. 28) has expressed this neatly by saying that a person borrowing money to develop a newly discovered oilfield will want the largest possible reserve estimate, but when evaluating the same oilfield for an estate he would want the smallest possible esti— mate. Mining company reserve calculations made dur- ing the planning process that precedes the large finan- cial outlays necessary to start mining have a tendency to be low: Caution can do no harm to the company except to lead to missed opportunities, but recklessness can do the company very serious financial damage. If the reserve concept is so complex and the calcula- tions are hard to evaluate, it is not suprising that those who estimate undiscovered resources open themselves to disputes. The test of the precision of a calculation is the reproducibility of results obtained by different per- sons using different methods. Oil is currently the most instructive example, partly because the urgency of its problems has given it wide attention but mainly because an enormous quantity of information is available. The disparities in the estimates of undiscovered oil in the United States seem extraordinary to many persons, but actually the spread is equivalent merely to several dec- ades of US. consumption at its current rate. Though this spread is important in planning the future of US. energy usage, its size is insignificant compared to what would result from similar examination of most other mineral commodities. A mineral commodity for which reserves are much larger than required by the market presents great un- certainties. Lithium is a good example. Most of the major known lithium deposits were easily found in sur— face exposures, and the others were found virtually by chance. Their lithium content is far greater than has ever been needed to meet the demands for lithium. The circumstances create little reason to search for more de- posits, thus offering slight experience or information to use as a guide in estimating the magnitude of undiscov- ered resources. My recent estimates of undiscovered re- sources (Norton, 1973, table 73, p. 372) were little more than informed guesses, as the text of the report makes clear, but they have been widely credited with more ac- curacy than they possess. . Resource estimates for coproducts and byproducts have problems which are neglected in figure 18A and which rarely, if ever, are examined in the literature, though the mining industry is well acquainted with them. The lithium produced as one of the minor prod- ucts, along with several major products, from brine at Searles Lake, Calif, is a useful example. A change in the processing plant to add to the overall efficiency of the operation could have the side effect of making the lithium unrecoverable. The reserve (if it ever should have been called a reserve, for the lithium content is only about 70 ppm) then no longer exists. Other lithium brines have marketable components that must be pro- duced to make mining economically viable. A lithium company can thus be put into another business in which it has little expertise and little influence, hence com- plicating the appraisal ofits outlook for lithium produc- tion. In short, resources of various categories can change through time for a great variety of reasons. Estimates of their magnitude range from careful, elaborate, and ex— pensive measurements to mere guesses. These estimates are furnished by persons who cover the whole spectrum from those of indisputable expertise and objectivity to those of questionable reliability. Persons who provide estimates should explain how their results were obtained so that others can make independent interpretations. With lithium the chief shortcoming in existing resource estimates is lack of information caused by the little past need for exploration. METHODS OF EXPRESSING RESOURCE ESTIMATES Metal mining companies ordinarily state reserves in terms of the tonnage of ore and its grade. Grade is gen- erally expressed as the percentage of a metal in the ore. Not all of the metal is recovered during mining and processing, but few companies issue estimates of the re— coverability. In making broad appraisals of metals re- serves, the lack of such estimates is rarely important because recoverability of metals tends to be near enough to 100 percent so as not to introduce errors that are much if any greater than the errors made in the normal reserve estimating process. Some mines actually recover LITHIUM RESOURCES AND REQUIREMENTS BY THE YEAR 2000 more than 100 percent of the measured reserves, for unanticipated ore may be found during mining or low- grade ore not originally calculated in the reserves may, inadvertently or by necessity, be extracted. The practices in stating reserves of deposits of indus- trial minerals and rocks (commonly called nonmetals) are too varied to be discussed briefly. The products may be rocks, minerals, chemical compounds, elements, or manufactured materials. A deposit’s grade (in the sense applied to this word in the metals industry) may be no more important than many of its other properties. Recoverability is critical in the oil industry’s examina- tion of its resources, for only about a third of the oil in the ground can be obtained profitably. The oil industry emphasizes what it calls “proved reserves,” which in- cludes only the amount that is reasonably certain to be recovered under existing conditions. Of course the oil terminology also provides for estimates of the total in the ground and for methods of expressing other esti— mates serving various purposes. For lithium the practice has been to state tonnage and grade, but recoverability is also very important. Of all the lithium ever extracted from pegmatites, probably at least one-third and possibly nearly half has been lost during mining and the elaborate processing needed to make the many kinds of lithium products on the market. In the brine at Silver Peak, Nev., the most recent esti- mate by Kunasz (this report) indicates a known content of 775,000 tons of lithium but a recoverable amount of only 44,500 tons. Applying the word “reserves” to the 775,000 tons, as has been done by Kunasz (1975, p. 796), is akin to the practice of the metal mining indus- try; applying it to the 44,500 tons follows the methods of the oil industry. This difference illustrates a vital prob- lem in expressing lithium reserve estimates. RELATION OF RESERVES TO PRODUCTION Often since the beginning of this century, it has been noticed that the reserves of many mineral commodities were adequate to last only a few decades. With the pas— sage of time, reserves in most of the mineral industry stayed ahead of production by about the same amount. Some have attributed this experience to transitory good luck by prospectors, to the incompetence of geologists, and to machinations of mining companies. The princi- pal reason is none of these. It is merely that the size of reserves depends largely on the magnitude of produc- tion, which in turn reflects demand. In the phraseology of mathematics, reserves are a function of production. They are also a function of the geologic availability of deposits, accessibility, political environment, and many other influences. A hypothetical example will clarify this point. Sup- pose all the world’s copper deposits were identical in 41 every way that affects how the copper is mined, proc- essed, and sold. Suppose also that the deposits contain enough copper to last thousands of years, and that price, rate of consumption, and other economic factors will remain unchanged. The effect of these assumptions is to reduce all the variables to a constant except reserves and production. Now suppose the lifetime of every deposit, at the most efficient rate of mining, is 100 years. Each year some deposits will be abandoned as mined out, and an equal number of deposits will begin to be mined. The reserves in operating mines will at all times amount to 50 years of supply, despite the existence of very large resources. No authority would disagree that the recoverable re— serve of lithium in the United States is at least 350,000 tons. At the 1975 production rate of probably between 4,600 and 5,100 tons (estimated from Wood, this re- port), this reserve would last about 70 years. Such a large reserve-production ratio is unusual, and seems in conflict with the belief that reserves tend to be kept at the minimum necessary level. Actually reserves are so large relative to demand not because there has been any need for large reserves but merely because some lithium deposits are of such great size that dividing a small number (production) into a big one (reserves) yields a large result. As a consequence, reserves are not only adequate to meet the assured market but also provide leeway to prepare to meet larger, less certain needs. Other articles in this volume suggest that annual de- mand for lithium will greatly increase by the end of this century. If demand thereafter is expected to continue at a high level, then reserves must be correspondingly in- creased to maintain an adequate backlog for subsequent years. HOW RAPIDLY CAN PRODUCTION OF LITHIUM BE INCREASED? An astonishing assumption underlying many discus- sions of future uses of lithium and whether supply will meet demand is that if resource estimates are large enough, there need be no worry about supply. Actually, the chief concern should be whether lithium can be brought to market in suitable amounts when needed. Theoretically one should also be concerned about whether the inadequately explored or undiscov- ered resources exist and can eventually be converted to reserves, but on this score I remain an optimist. A way of visualizing the circumstances is to look at the McKelvey diagram (fig. 18A) as if it were three dimen- sional and the third dimension is time. In this third dimension, the reserve box would be continuously de- pleted by mining and continuously replenished by dis- covery of additional deposits and by raising low-grade material to a profitable level. How fast this can be done 42 LITHIUM RESOURCES AND REQUIREMENTS BY THE YEAR 2000 is a critical and difficult question, especially with a com— modity for which there has been very little exploration. Figure 18B presents the problem in a different light. Wood (this report) suggests a 5-percent growth rate in U.S. production for the next 5 years. If this growth rate were to continue, the production in the year 2000 would be about 15,000 tons of lithium. At the other extreme, Chilenskas and others (this report) suggest a consump- tion for storage batteries alone in the year 2000 at somewhere between 90,000 and 170,000 tons. The maximum likely figure for U.S. annual demand at the end of the century seems to be about 180,000 tons. The disparity between 15,000 and 180,000 tons raises obvious enough uncertainty to those who must find de- posits, drill them out, develop mines, and build plants. A further problem lies in the rate of change in demand. Curve 1 of figure 18B shows the most probable form of increasing annual consumption if lithium batteries do come into common use—a modest expansion in the next few years and a great upswing at the end of the century. A curve of this shape cannot, of course, continue up- ward forever. It must eventually turn downward be- cause of a shortage of lithium, because of something else affecting lithium usage, or through other ancillary cir- a i h 2 9 9 .— 3 f o / / g 3 2 /J O. / e 4 //f //// g d 2 Z / < c / / b O A 1 1 . 1 1 I 1975 1980 1935 1990 1995 2000 YEARS FIGURE 18B.——Schematic graph of how an increase in lithium produc- tion may proceed in the years 1975 to 2000. Lines 1, 2, and 3 are possible generalized curves showing annual production (the actual curve will, of course, be far more irregular). Point A is 1975 produc- tion, which in the United States was probably somewhat less than 5,000 tons of contained lithium. Point B is U.S. demand in the year 2000, which may be as small as 15,000 tons or as great as 180,000 tons. Lines 21 through i signify production levels at which new mines and plants must start operation. - cumstances. Curve 2 represents merely a straight-line increase. Curve 3 has the shape to be expected if lithium resources minable at acceptable costs turn out to be hard to find or if the demand for lithium has a dimishing rate of increase at the end of the century. Along any curve representing increasing production, there are points at which new mines and plants must start operating. These are represented schematically by the horizontal lines labeled a through i. The lines indi- cate considerable uncertainty about when new mines will be needed. This uncertainty is important for any— thing that requires as long a lead time as a mine, especially a mine in a new kind of deposit requiring experimentation with processing methods. Mining companies in lithium and the other so-called rare ele- ments have heard about many anticipated large new uses that did not materialize, and they have learned the importance of caution and skepticism. The outlook beyond the end of the century can also be influential, though its importance may be little more than theoretical now. Nevertheless it can affect the rate at which capital costs must be amortized and the pru— dence necessary in developing new mines. The history of other mineral products that have gone through large increases in demand can give hints about how large an increase is practicable in lithium. Let us assume that the production capacity desired for the United States by the year 2000 is 150,000 tons of lithium. Starting from a current capacity of not much more than 5,000 tons (Wood, this report), an approxi- mately 30-fold expansion in 25 years would be needed. Increases of this amount are rare in segments of the mineral industry controlled mainly by private markets. For oil, which has an outstanding reputation as a growth industry, U.S. output increased 58-fold in the 75 years from 1896 through 1970, which was the peak year. The increase from 1921 through 1970 was merely 7.5 times. Copper increased only eight-fold from 1896 through 1970, also the peak year for it. To find increases that were proportionately much greater and took place over a shorter time, it is best to search among commodities which had small production until recently and for which there was assistance from the Federal Government. Uranium is the greatest suc- cess story. U.S. uranium production increased from vir- tually nothing in the late 1940’s to a high peak in 1960. Meanwhile world production, in the non-Communist group of nations, increased 22-fold. Niobium (columbium) is especially pertinent because in 1950 its situation was similar to that oflithium today. The main difference between the two is that notice of an increase in niobium demand, chiefly for use in jet en- gines, came very suddenly. The geologic literature available in 1950 indicated, as with lithium now, that LITHIUM RESOURCES AND REQUIREMENTS BY THE YEAR 2000 major deposits of niobium probably could be found, and notes in tradejournals implied that some were ready for development. A US. Government purchase program began that exceeded the wildest expectations, and after its end in 1959 world production increased even more rapidly. Niobium output of the world, outside the Communist segment, was about 14,565 tons in 1975 (US. Bureau Mines, 1976, p. 45), in contrast to about 500 tons in 1950 (when production data were, for sev- eral reasons, imprecise). The increase was about 30—fold. With such experiences in the background, and if one can assume the availability of adequate lithium re- sources, there seems little reason for pessimism about increasing lithium production enough to meet the maximum demands anticipated for the end of the cen- tury. In lithium, unlike niobium in 1950, the United States is already the world’s largest producer and has the world’s largest reserves as well as favorable geologic en— vironments for the discovery of more deposits. For the expansion to be orderly and not too costly, new deposits must be found and the demand must be assured long enough in advance for the construction of producing facilities. PROBLEMS IN RECENT ESTIMATES OF LITHIUM RESOURCES Several articles in this volume treat various aspects of existing information about lithium resources. The prin- cipal assessments of resources are in the articles by Vine and by Kunasz. Their tables of resource estimates may, to the casual reader, seem greatly different, but actually the main difference is in their approach to the available data. Vine emphasizes the very proper view that the critical figures are for lithium that almost certainly can be mined by the end of the century, for this is the only lithium that users can be sure to obtain. Kunasz takes the equally proper view that in an expanding market the known resources that are less adequately known or are oflower quality will take on increasing importance. Both of them would undoubtedly agree that search for new deposits is necessary if US. demand is indeed going to increase greatly by the year 2000. If a proper reserve base should amount to at least a 30-year supply, and if the annual production at the end of the century is to be 180,000 tons of lithium, reserves at that time should be about 5,400,000 tons, which is larger than either of their estimates of known resources and far larger than any estimate of reserves. My own recently published appraisal of resources (Norton, 1973, p. 372—375) is not fundamentally in- compatible with those of Vine and Kunasz. The most obvious difference is that I estimated or (in more correct words) made guesses of undiscovered world resources; Kunasz makes no such guesses and Vine does so only for 43 the United States, arriving at a total (if his Kings Moun- tain, N.C., low—grade resources are included) similar to mine but Via a different route. Both Vine and Kunasz use Great Salt Lake brines, which I chose to disregard. Vine also uses oil field brines, for which I had virtually no information. Kunasz promotes the Zaire and Chilean deposits to identified resources, chiefly on the basis of recent information. A source of some contention has been the estimates for Clayton Valley, Nev. The original claim was 2.5 to 5 million tons of lithium “reserves” (obviously meaning “resources"). My examination of the published data (Norton, 1973, p. 374) led to assigning 500,000 tons to reserves and relegating the rest to more dubious categories of resources. Kunasz (this report), using newer information, now estimates the identified re— sources at 775,000 tons oflithium, of which only 44,500 tons is currently recoverable. He adds that the total re- sources in the valley, apparently in clay as well as brine, are undoubtedly very much larger, though when and how' these resources can be mined seems questionable. Vine (1975 and this report) stresses the 44,500-ton figure, because one can rely on obtaining this amount. Yet to say that no more than this quantity oflithium will ultimately be extracted from Clayton Valley may be too conservative, though some such event as the develop- ment of better deposits elsewhere, making Clayton Val- ley not competitive, could prevent enlargement of its supply of recoverable lithium for a long time. OUTLOOK FOR INCREASING LITHIUM RESERVES Questions about the future success of lithium explora- tion are difficult to answer, and the answers are impos— sible to quantify. The small amount of exploration and geologic study has yielded a correspondingly small amount of information about where and how to search for new deposits. The traditional source of lithium has been pegmatite. Most lithium pegmatite has about 0.7 percent lithium, apparently for geologic reasons summarized recently (Norton, 1973, p. 368—369). The main economic differ- ence between a good and a not—so-good lithium pegma- tite is size: A large one can be mined more cheaply than smaller ones. Another difference is accessibility. The Zaire deposits may be the best in the world except for their remote location and the political environment. Pegmatites are the source of several unusual mineral products and for this reason they have been prospected carefully in much of the world. Not many large lithium pegmatites exposed at the surface seem likely to have remained undiscovered. The stage of exploration prob- ably is near maturity. Nevertheless, the great disparity between the amount of known lithium pegmatite in 44 North America and in the rest of the world suggests some hope for the potential of other continents. Exploration is decidedly immature for lithium peg- matites concealed beneath the surface and for small but possibly useful pegmatites. Kesler in this report makes suggestions about how to find additional large deposits in the Kings Mountain region, North Carolina. Several clay deposits are known to contain about 0.2 percent lithium, but none have been mined for lithium. For these deposits, little laboratory work on extraction processes has been attempted, the geology is poorly known, and all the discoveries were little more than ac- cidental. A grade of 0.2 percent by weight seems very low, but opportunities for low cost mining and process- ing may offset this defect. Furthermore, weight percent is somewhat misleading with elements of low atomic weight: for example, spodumene (LiAlSiQO6), the prin- cipal lithium mineral of pegmatites, contains 10 percent lithium in terms of atoms but only 3.7 percent by weight. The likelihood of finding lithium-rich brines was sus- pected long before the more or less fortuitous discovery of the Clayton Valley deposit, which is now probably the second most productive deposit in the world. The de- posit in the Salar de Atacama, Chile, reputedly also found by chance, appears to be richer than Clayton Val- ley and probably also large. Whether even better de- posits are awaiting discovery is unknown and will be beyond even guessing until the geology and geochemis- try of lithium brines have been investigated much more thoroughly. In short, exploration for lithium brines and clays is at a very immature stage. No other kinds of lithium deposits are known, but they may exist. The source of the lithium in lithium-rich pegmatites is uncertain (Norton, 1973, p. 368—369). The lithium may have been obtained by partial melting of lithium-bearing metamorphic rocks. If so, some of the lithium pegmatite regions of the world may also contain nonpegmatitic deposits. Because nearly all lithium peg- matite localities were first prospected for tin, the associa— tion between tin and lithium may be worth examining. Many publications indicate that rocks allied with the Cornwall, England, tin deposits have lithian mica, and one wonders if they have other lithium minerals. Ocean water has often been cited as a potential source of lithium. Its content of only about 0.2 ppm lithium multiplied by the large tonnage of ocean water is im- pressive, but actually does nothing but reaffirm the eye—catching effect of such arithmetic. With so many better possible sources of lithium to test, one cannot regard ocean water seriously except perhaps for by- product lithium. In processing ocean water, other com- modities would probably be more remunerative, and the lithium output would depend on their market. This is LITHIUM RESOURCES AND REQUIREMENTS BY THE YEAR 2000 not to say that the work of Steinberg and Vi-Duong Dang (this report) is not a useful contribution to the chemical engineering of extracting lithium from seawa- ter. The subject has been so much discussed that its technology must be examined. HOW TO ADAPT RESOURCE NOMENCLATURE TO LITHIUM Three aspects of lithium resources have caused most of the misunderstandings and disputes: (l) the differ— ence between the lithium contained in the deposits and the amount recoverable, (2) the existence of deposits that are not now minable at a profit but are equal in quality to deposits mined in the past and in some in- stances have actually been mined, and (3) the potential for finding new deposits. Uncertainty about undiscovered deposits can be al- leviated only by geologic work of many kinds coupled with exploration to teach us where, how, and at what rate new resources can be found. If uses in the next 25 years are to become as large as now anticipated, a sizable body of geologic data must be acquired soon. The problems with recoverability and with marginal deposits can be clarified by dividing the resource esti- mates into the smallest feasible categories. Resources in operating mines, closed mines, and undeveloped but known deposits should be separated from one another. Resources in different kinds of deposits should also be segregated, because future technological advances may have different effects on the various categories. Separa- tion of deposits by size, grade, mineralogy, and other traits would also be worthwhile. As Singer (1975) shows, “disaggregated” information of this kind will facilitate the process of predicting the supply outlook for the var- ious possible combinations of future economic and technologic circumstances. Reserves should include only lithium that is currently recoverable, and should exclude lithium now lost in mining and milling. In other words, reserve estimates should encompass only material reasonably certain to enter market channels. The unrecoverable lithium is not a reserve in this sense, but because the recoverability may increase or the lithium in mine dumps and mill tailings may someday become obtainable, it is of poten- tial value. Companies may be reluctant to reveal re- coverability, but probably in most instances they can hide the data they prefer not to release. Certainly frank statements of recoverable reserves would eliminate a major source of confusion that may endanger the future of the lithium business. The importance of marginal resources is illustrated by the Quebec Lithium Corp. deposit at Barraute, Quebec, which could quickly assume its former status as one of the world’s best and most productive deposits. The deci- LITHIUM RESOURCES AND REQUIREMENTS BY THE YEAR 2000 sion to stop operation because of inability to compete with Clayton Valley brine made it no longer a reserve but merely a paramarginal resource. However accurate this term may be, it implies that the deposit’s standing is lower than it really is. Other pegmatite deposits have a similar problem. They have often been called reserves, and some of them are in fact either reserves or so near to being reserves that misunderstanding is created by treating them slightingly. Lithium seems unready for a well-organized, system- atic, and exact procedure for stating resources, but a step in that direction would be well taken. One can more easily be optimistic that knowledge of the lithium poten- tial of the United States can and probably will be greatly increased in the next decade. REFERENCES CITED Banfield, A. F., and Havard, J. F., 1975, Let’s define our terms in mineral valuation: Mining Eng, v. 28, no. 7, p. 74—78. Brobst, D. A., and Pratt, W. P., 1973, Introduction, in Brobst, D. A., and Pratt, W. P., eds., United States mineral resources: U.S. Geol. Survey Prof. Paper 820, p. 1~8. Kunasz, I. A., 1975, Lithium raw materials, in Lefond, S. J., ed., In- 45 dustrial minerals and rocks [4th ed.]: Am. Inst. Mining, Metall., and Petroleum Engineers, p. 791—803. McKelvey, V. E., 1972, Mineral resource estimates and public policy: Am. Scientist, v. 60, no. 1, p. 32—40. Repr. in Brobst, D. A., and Pratt, W. P., eds., 1973, United States mineral resources: U.S. Geol. Survey Prof. Paper 820, p. 9—19. Norton, J. J., 1973, Lithium, cesium, and rubidium—the rare alkali metals, in Brobst, D. A., and Pratt, W. P., eds., United States mineral resources: US. Geol. Survey Prof. Paper 820, p. 365—378. Schanz, J. J.,Jr., 1975, Resource terminology—an examination of con- cepts and terms and recommendations for improvement: Wash— ington, D.C., Resources for the Future, 116 p. [Prepared for Elec- tric Power Research Inst.]. Singer, D. A., 1975, Mineral resource models and the Alaskan mineral resource assessment program, in Vogely, W. A., ed., Mineral materials modeling: Washington, DC, Resources for the Future, p. 370—382. US. Bureau of Mines, 1976, Commodity data summaries 1976: US. Bur. Mines, 196 p. US. Congress, 1975, Oversight—Bureau of Mines, Geological Survey, Ocean Mining Administration: US. Cong, 94th, lst sess., House of Representatives Comm. Interior and Insular Affairs, Comm. Print, 75 p. US Geological Survey, 1975, Mineral resource perspectives 1975: US. Geol. Survey Prof. Paper 940, 24 p. Vine, J. D., 1975, Lithium in sediments and brines—how, why, and where to search: U.S. Geol. Survey Jour. Research, v. 3, no. 4, p. 479—485. N OCCURRENCE, DEVELOPMENT, AND LONG-RANGE OUTLOOK OF LITHIUM-PEGMATITE ORE IN THE CAROLINAS By THOMAS L. KESLER (retired), HENDERSONVILLE, NC ABSTRACT Unzoned lithium pegmatites occur in a narrow area 50 km long in the Carolina Piedmont. The pegmatites are composed of approxi- mately 20 percent spodumene that contains about 7.5 percent LigO. The ore bodies are as much as 90 m thick and 1,000 m long. Tens of millions of tonnes of ore are minable by open pit. Development began gradually in the 1930’s, and two large mines are now in operation. The area containing the pegmatites occurs mostly within and parallelto the strike of metamorphosed shallow-water sediments of the Kings Moun- tain belt and bounds the east side of the Permian(?) to Mississippian(?) Cherryville Quartz Monzonite of the Inner Piedmont. Pegmatites elsewhere in and around the monzonite do not contain lithium miner- als. The quartz monzonite provided the main components of the lithium pegmatites, but the lithium was probably derived from con- tiguous evaporites in the metasedimentary rocks. Deep pegmatites of long-range interest for underground mining are probably west of those now known, because the metasedimentary rocks dip westerly and should dip less steeply at depth. GEOLOGIC SETTING The western world’s largest reserves of lithium are in the pegmatites of an area of metamorphic rocks in the Carolinas. Geologic maps of much of the area (Sterrett, unpublished map, 1912; Keith and Sterrett, 1931; Kesler, 1942, 1944, 1955; Griffitts, 1958; Espenshade and Potter, 1960) have been used in the preparation of figure 19, which shows features related to this discus- sion. Most of the area is in the western part of the Kings Mountain belt of the Piedmont. The country rocks are mostly metasedimentary and of lower metamorphic rank than those to the east and west. They include quartzite, conglomerate, fine-grained schist with traces of pyroclastic texture, chloritoid schist, biotite schist and gneiss, thin-layered amphibolite, crystalline limestone, and phyllitic metashale that is interbedded with the other types of rocks. The more quartzose of these rocks underlie a series of ridges, of which Kings Mountain is the most prominent (figs. 19 and 20). The less quartzose rocks, which strati- LITHIUM RESOURCES AND REQUIREMENTS BY THE YEAR 2000 46 8I°I5' ~ n erson MAJOR INTRUSIVES \\\\\\'. .‘Mtn \ Mostly Permian Yorkville \\ Quartz Monzonite \\\\\\ /’ V4 r< Permiani?) to Mississippiani? ) \/ Cherryville Quartz Monzoniie \/ pggksPoleozom metamorphic 35° _ "50‘ O 5 MILES 0 5 KILOMETRES /\ \x ‘u /4 V7” V " 7a /4 CU “ Q7 ——'_" > V‘fi-VL—hi/ OCherryville 4 v ‘/ ;v v :17 Q‘ /1V?// P/ Q 0/ 4’ ‘u oGaffney Lithium "X‘ ‘0 Corp of.. I) \\ k America \\\ ' ) \ A /\\\VM / Q /\\ /\ O \\ O \\ Bessemer City \1 e i Q- 35°_ Kings 3 l _ I5' Mountain 5 ' q Q Foote - \\ O Crowders ‘ I {0 L) $ $ EXPLANATION \ Grover Area containing ‘§— - _ .7 V\' NORTH CAROLINA 3 lithium pegmatites / \\\\\\ SOUTH CARO_|_lNA__— ’\ \\Stratigraphic zone of Whittakerf \\\\\\ discontinuous crystalline gm -\\'\\\\\ Q) Henry. limestone beds I \ .‘ /"\O\\ $ Knob/ ~---- . Quartzose7 ridge- / \\\\\\ BIGCKSburg \ / forming rocks \\ 4* / \\\\ / Boundary between A? \\\‘\ Piedmont belts Gaffney 9’ \\\\\ / O \\\ oKIngs Creek . . . . /\\\\ / ‘X‘ Active lithium-ore mine /\\\\\ / ('3 X 1'3 IO MILES I l I ‘l I i I 0 5 IO KILOMETRES / ‘ I / I FIGURE 19.—Areal features of the Carolina lithium region. graphically overlie the quartzose rocks, crop out west of identify the strongly compressed core of the structure ‘ I the ridges with uniform westerly clip except at Gaffney, where the beds are overturned. These west-dipping rocks identify the western limb of an arch original— ly more than 15 km wide, and vertical isoclinal folds mapped in the ridges by Espenshade and Potter (1960) In the northern part of the belt (section A—A ) the eastern limb of the arch is missing, having been dis- placed by a major intrusive of the adjoining Charlotte belt. This is the biotitic and porphyritic Yorkville Quartz Monzonite of Permian age (Overstreet and Bell, 1965). LITHIUM RESOURCES AND REQUIREMENTS BY THE YEAR 2000 47 NT INNER PIEDMO A Lithium Pegmatites H—L—fi IIII III :I’I/I/ll’ I’ll/NI l/II / / /WI/// Ill/.Al/llllllllll [Ill/WI] It’ll/l LII/III |e———— KINGS MOUNTAIN BELT———>I ll/l/jl/fl CHARLOTTE L BELT / I/I III/’1] I I, I’I I/ I I/’/II/I/,/’I’I,’I//I I I/I’II I//,I III/[Ill lll-l MIN/Ill, Will/l 01—0 D: UJ Z Z PIEDMONT 7F Lithium Pegmatites r’H KINGS MOUNTAIN BELT EMILES 4| I 2 KILOMETRES \ AL CHARLOTTE BELT EXPLANATION PALEOZOIC METAMORPHIC ROCKS A PALEOZOIC IGNEOUS ROCKS (—A—fi \ ’\I /_\ _\II‘,\I ’ ., A <>_, VAA --@mm Permian Permian(?) to Thin—layered Meta—argiIIa- Crystalline Schistose Schists and Yorkville MississippianI?) amphibolite ceous rocks limestone pyrocIastic gneisses of Quartz Cherryville with quartz- rock diverse origin Monzonite Quartz Monzonite ose beds FIGURE 20.—Cross sections showing geologic relations of the pegmatites. Location of sections shown in figure 19. The forceful intrusion of the Yorkville left discordant contacts and contact-metamorphic effects (Espenshade and Potter, 1960; Overstreet and Bell, 1965), but this intrusive shows no relation to the lithium pegmatites. Along the western boundary of the Kings Mountain belt there is a transition involving four features: (1) a gradual decrease in dip from that shown in figure 20; (2) an increase in metamorphic rank; (3) a change from the northeast structural grain of the Kings Mountain belt to a complex of diversely oriented, broad, open folds characteristic of the Inner Piedmont; and (4) a profusion of large and small bodies of a muscovite- biotite quartz monzonite of the Cherryville, many of which occur as concordant sheets. Limited data on the Cherryville yield radiometric ages ranging from 375 to 260 million years, a spread regarded by Overstreet and Bell (1965) as inconclusive. Whether this results from inadequate sampling or from a long and relatively pas- sive intrusive history is unknown. However, the Cher- ryville differs from the Yorkville in that it does not trun- cate major country—rock structures The Cherryville contains nonlithium pegmatite, and similar pegmatite also occurs in the surrounding coun- try rock; however, it is only along the east side of the intrusive that the outermost of the pegmatites are lithium bearing. Of particular significance (Kesler, 1961) is a compound pegmatite 30 m wide within am- phibolite and 60 m from the quartz monzonite to the west and lithium pegmatite to the east. In composition, this body grades eastward across its strike from simple quartz-monzonite pegmatite to spodumene-bearing pegmatite. The Cherryville, therefore, shows a genetic relation to the lithium pegmatites; this relationship is considered further in discussing future exploration. 4 8 THE PEGMATITES The lithium pegmatites occur at close intervals in an area about 50 km long and less than 3 km in maximum width (fig. 19). Sporadic spodumene crystals also occur in pegmatite near Gaffney, 19 km farther south. The bodies are as much as 90 m thick and almost 1,000 m long. To date drilling has reached a maximum depth more than 200 m below outcrop without encountering any change in the geologic characteristics of pegmatites or wall rocks. In general, pegmatite was intruded parallel to the foliation of mica schist. Thin-layered amphibolite, where not interlayered with schist, was ruptured irregu-_ larly with respect to dip, but parallel to strike. Thin pegmatites were intruded intojoints with diverse trends in massive hornblende gneiss. Five minerals make up about 99 percent of the lithium pegmatites. As previously reported (Kesler, 1961), long-term control on the milling of crude ore showed an original content of 20 percent spodumene, 32 percent quartz, 27 percent albite, 14 percent micro- cline, 6 percent muscovite, and 1 percent trace minerals including cassiterite in which there has been recurrent mining interest since the 1880’s. Only near the collapsed shaft of the old Ross tin mine near Gaffney have pegma- tite cores been recovered in which spodumene is very scarce. Elsewhere it is common and averages about 7.5 percent L120 (theoretical maximum 8.03). Thus, crude ore with 20 percent spodumene should grade about 1.5 percent L120, or 7,000 parts per million Li. The pegma- tites, with rare and small-scale exceptions, are not zoned, and the grade of crude ore is fairly uniform as shown in published‘assay sections (Kesler, 1961). This feature has also stimulated the recovery of feldspar and mica byproducts, which are liberated in grinding the ore. Late features include contact-alteration selvages be- tween pegmatite and amphibolite, rarely as much as 0.5 m thick, consisting of rare-alkali biotite, holmquistite, calcite, and apatite. Near some of the large pegmatites, dynamically brecciated amphibolite containing frac- tured black tourmaline has been altered to coarse chlo- rite rock containing much pyrrhotite and a little holmquistite and chalcopyrite. Extreme weathering of such breccia leaves only a conspicuous float of fragmen- tal black tourmaline. A few nearly vertical, northwest- trending faults of very small displacement cut some of the pegmatites. The walls are separated by a few cen- timetres to nearly a metre of amphibolite breccia and gouge, partly altered to a biotitic rock that resembles minette. DEVELOPMENT In 1938, our government’s concern with supplies of LITHIUM RESOURCES AND REQUIREMENTS BY THE YEAR 2000 strategic minerals caused a renewed interest in the cas- siterite content of the pegmatites. The US. Geological Survey assigned me and other personnel to an investiga- tion that included two seasons of detailed mapping, and the results (Kesler, 1942) indicated a large potential for developing minable reserves of lithium ore. Others had foreseen this possibility. Limited prospect- ing was started in the mid-thirties by the Tennessee Mineral Products Company, by Phillip Hoyt, and by L. M. Williams, who was still active when our survey began. On leases of Williams, the Solvay Process Company operated a mine and flotation plant in the early forties and did considerable diamond drilling. This effort ended in 1945, and uncertain market conditions follow- ing World War II prevented further work until 1951, when Foote Mineral Company reactivated the mine and mill and did more extensive drilling. An account of all exploration on this property through 1959 has been published (Kesler, 1961). Shortly after Foote started its operation, the Lithium Corporation of America began construction of a chemical plant at Bessemer City and did some preliminary mining at two localities 8 km south of Lincolnton. Their present mine is near Tryon School, 6.5 km northwest of their plant, where much explorat- ory drilling has also been done. The large open—pit mines of Foote and Lithium Cor- poration. are expanding, and during their early years purchases oflithium by the Atomic Energy Commission cansed a flurry of prospecting in the Carolina area as well as in Canada, its nearest potential competitor. In fact, until early in 1956, published articles had listed 286 lithium interests in 16 areas in 3 Canadian provinces plus the Great Slave Lake region. The Canadian activity had a strong influence on developments in Carolina and finally matured into temporary competition. The domestic industry should not forget. MINING The mines in the Carolinas must penetrate the zone of weathering before encountering the fresh ore that gives best recoveries in milling. The base of the zone of weathering is irregular, mostly from 10 to 20 m below the surface in pegmatite, but generally deeper in the country rocks. Little of the pegmatite in the zone of , weathering is decomposed, but substances deposited from ground water stain the cleavage surfaces of the minerals. Flotation reagents therefore cannot reach clean surfaces, and only part of the stains are acid solu- ble at practicable concentration. Separation by heavy media requires crushing that preferentially shatters the spodumene, which has better cleavage than the other minerals. Power requirements for abrasive scrubbing of total crude ore to remove the stains would become in- creasingly uneconomic because of rising fuel costs, and LITHIUM RESOURCES AND REQUIREMENTS BY THE YEAR 2000 this disadvantage would also apply to any effort to bypass the concentration of spodumene by roasting the crude ore for treatment. Whatever may be the solution to this problem, it must be found to eliminate a con- tinual drain on ore reserves. LONG-RANGE OUTLOOK It is generally agreed (Kesler, 1961; Norton, 1973) that reserves of minable lithium ore in the Carolina area amount to tens of millions of tonnes. The rate at which they will be consumed is unknown, but their location gives them a competitive advantage over other pegma— tite ores. The advantage will continue until mounting costs of open-pit stripping and initial stages of under- ground mining bring competition from foreign sources. As reserves in the Carolina area are depleted and costs increase, the lithium industry will then enter a sec— ond stage in which the domestic market may be divided between domestic and foreign suppliers. A third stage will eventually be reached in which foreign sources will also experience rising costs, bringing the location of ore reserves back into major importance. At that time the domestic industry, if it prepares well in advance, can be in position to recapture lost segments of the world lithium market. The preparation will require exploration beyond any- thing yet attempted, based on the certainty that all ore of sufficient bulk for open-pit mining will have been discovered and that new targets will be much deeper. Where and how these efforts should be directed re- quires consideration of the origin of the Carolina lith- ium pegmatites. Scarcely half of the Cherryville Quartz Monzonite as illustrated in figure 19 actually consists of monzonite. The remainder consists of Inner Piedmont metamor- phic rocks with widely divergent structural trends, a fea— ture also characteristic of Inner Piedmont rocks margi- nal to the Cherryville. The quartz monzonite is accepted as the source of all pegmatites in and around its area of outcrop, but only those containing lithium occur in an area having a preferred and essentially straight trend. That trend is along the west side of the Kings Mountain belt and parallel to its structural grain as reflected by the zone of limestone shown in figure 19. Thus, the dis- tribution of the lithium, considered apart from the na- ture of its host, seems to be more closely related to the Kings Mountain belt than to the Inner Piedmont. In an effort to account for high concentrations of pegmatitic lithium without an obvious source, Norton (1973) has suggested that deep melting, or anatexis, of lithium-rich metasedimentary rocks could form the pegmatite magma. This concept, though not entirely compatible with known conditions in the Carolina area, provides a new approach. 49 The low metamorphic rank and well-preserved sedimentary structures of the rocks at Kings Mountain preclude their having reached conditions usually specified for anatexis, but their lithology suggests an original shallow-water environment that could have in- cluded evaporites. All of the rocks are thin bedded, and the bedding planes are sharp and continuous along the trend of the belt without appreciable distortion. The limestone is lenticular and was originally silty; the quartzite is relatively pure and reflects good sorting; the conglomerate is mostly fine textured, but contains peb- bles as much as 7.5 cm in diameter; and the meta- argillaceous rocks are nonuniform in composition, par- ticularly their micas. These features show that the origi- nal sediments at Kings Mountain were deposited under conditions of variable currents, in water too shallow for massive, poorly sorted deposits to be formed. They were apparently deposited in a basin bordering an area that would later become the Charlotte belt (fig. 19) on whose margin Espenshade and Potter (1960) have found evi- dence of volcanism, and therefore a possible source of thermal waters. The final stage in the filling of such a basin could have included the formation of evaporite deposits whose po- sition would have been in the present downdip section beneath the eastern edge of the Inner Piedmont. Dur- ing the slow invasion of the Cherryville Quartz Monzo- nite, volatiles expelled into the rocks of the Kings Moun- tain belt would have served as collectors of loosely bonded elements, including lithium if present; such v01- atiles are essential parts of pegmatite magma. This hypothetical framework implies that rocks of the Kings Mountain belt should persist in depth as the wall- rocks of the lithium pegmatites. Their normal dip is westward beneath the rocks of the Inner Piedmont (sec- tionA—A ’) and adjacent to the Cherryville Quartz Mon- zonite. In the entire extent of the Cherryville and the metamorphic rocks (fig. 19), the many bodies of the monzonite are so uniformly concordant with the Inner Piedmont metamorphic rocks that abrupt departure from this habit is unlikely. The eastern limit of the Cherryville Quartz Monzonite and its bordering fringe of lithium pegmatites can therefore be expected to shift westward at depth, and deeper pegmatites of interest for future exploration should lie west of the axis of present outcrops. Outside of this area, the apparently barren gap be- tween Grover and Gaffney could reflect an original di— vide between evaporate deposits, but the presence of spodumene near Gaffney requires reconsideration of that area. Closer study is also needed beyond the pres- ently known limit of the pegmatites to the northeast, where conditions are not clear (Broadhurst, 1956) but where the continuity of typical metasedimentary rocks 50 in the Kings Mountain belt may indicate continuity of lithium source beds. There is no known physical reason why underground mining of the strategically located Carolina ore bodies cannot reach depths many times the present maximum; it is also an economic probability for the future, not only because mineral values rise as raw materials are de— pleted, but also because essentially the entire mineral content of this ore has large-scale industrial use. REFERENCES CITED Broadhurst, S. D., 1956, Lithium resources of North Carolina: North Carolina Div. Mineral Resources Inf. Circ. 15, 37 p. Espenshade, G. H., and Potter, D. B., 1960, Kyanite, sillimanite, and andalusite deposits of the southeastern states: U.S. Geol. Survey Prof. Paper 336, 121 p. Griffitts, W. R., 1958, Pegmatite geology of the Shelby district, North LITHIUM RESOURCES AND REQUIREMENTS BY THE YEAR 2000 Carolina: US Geol. Survey open-file rept., 123 p. Keith, Arthur, and Sterrett, D. B., 1931, Description of the Gaffney and Kings Mountain quadrangles: U.S. Geol. Survey Geol. Atlas, Folio 222, 13 p., maps. Kesler, T. L., 1942, The tin—spodumene belt of the Carolinas: U.S. Geol. Survey Bull. 936—], p. 245-269. 1944, Correlation of some metamorphic rocks in the central Carolina Piedmont: Geol. Soc. America Bull., v. 55, p. 755—782. 1955, The Kings Mountain area, in Russell, R. J., ed., Guides to southeastern geology: New York, Geol. Soc. America, p. 374—387. 1961, Exploration of the Kings Mountain pegmatites; Mining Eng, v. 13, p. 1062—1068. Norton, J. J., 1973, Lithium, cesium, and rubidium—the rare alkali metals, in Brobst, D. A., and Pratt, W. P., eds., United States mineral resources: U.S. Geol. Survey Prof. Paper 820, p. 365—378. Overstreet, W. C., and Bell, Henry, III, 1965, The crystalline rocks of South Carolina: US. Geol. Survey Bull. 1183, 126 p. Sterrett, D. B., 1912, Geology of the Lincolnton quadrangle, North Carolina: US Geol. Survey, unpublished map. N A COMPARISON OF THREE MAJOR LITHIUM PEGMATITES: VARUTRASK, BIKITA, AND BERNIC LAKE ByEflWM.HEmRmH UNIVERSITY or MICHIGAN, ANN ARBOR, MI ABSTRACT A reassessment of the mineralogical and geological characteristics of three major lithium pegmatites, Varutr'ask, Sweden; Bernic Lake, Manitoba; and Bikita, Rhodesia, shows that Bernie Lake and Bikita are mineralogically and structurally very similar. Both also contain sig- nificant reserves of lithium, especially in very low iron secondary spodumene which is especially suitable for utilization as glass-ceramic raw material. INTRODUCTION Three of the world’s most highly localized major Li-Rb concentrations are the structurally and mineralog— ically complex pegmatites of Varutrask, Sweden; Bikita, Rhodesia; and Bernie Lake, Manitoba, Canada. This paper compares these deposits and assesses their past importance and future potential as Li-Rb sources. The writer has examined all three: Varutr'ask first some 25 years ago; Bikita, in 1966 and 1967; and Bernic Lake several times beginning in 1972. PEGMATITIC LITHIUM MINERALOGY There are some 26 different “independent” lithium minerals, of which nine silicates and eight phosphates occur in granitic pegmatites. Five other lithium minerals occur almost exclusively in highly alkalic subsilicic peg- matites (nepheline-syenitic) and.related deposits. Of the many subordinate elements that characterize granitoid pegmatites, lithium is undoubtedly one of the most versatile and, indeed, is so in several ways. It not only forms its own pegmatitic species but can also ap- pear vicariously in other minerals, including such di— verse groups as pyroxenes, amphiboles, micas, and tourmalines. Its granitic pegmatitic minerals are equally varied: the pyroxene, spodumene; the amphibole, holmquistite; the micas, lepidolite, zinnwaldite; and the zeolite, bikitaite. Both a variety of silicates and of phos- phates (for example, amblygonite, triphylite) are repre- sented. In alkalic granitic pegmatites lithium may even appear as the fluoride, cryolithionite. In granitic pegmatites lithium minerals appear in all three structural types of pegmatites: unzoned, zoned, and structurally complex. In addition, holmquistite oc— curs as an exogenic wallrock constituent (Heinrich, 1965). In zoned pegmatites, spodumene and amblygonite-montebrasite are not uncommonly mul- tigenerational, with differences among generations in form, crystal size, and composition (Heinrich, 1953). Lepidolites of complex pegmatites are normally mul- tigenerational with conspicuous variations among gen- erations in color, grain size and form, and, of course, in composition (Heinrich and Levinson, 1955; Heinrich, 1967). The lithium minerals range, paragenetically, from such earliest species as phenocrysts of iron-rich LITHIUM RESOURCES AND REQUIREMENTS BY THE YEAR 2000 51 spodumene, through the intermediate magmatic stages (spodumene, petalite, amblygonite, triphylite), through the post magmatic replacement period (lepidolites), into intermediate to low-temperature hydrothermal altera- tion stages (eucryptite, bikitaite, cookeite, and secondary phosphates). PARAGENETIC TYPES OF SPODUMENE Recent studies (Heinrich, 1975; 1976) have identified three distinct paragenetic types of spodumene in peg- matites: 1. Phenocrystic spodumene in essentially unzoned pegmatites of the “Kings Mountain type.” These laths, usually less than a foot long, commonly are greenish and contain essential Fe3+ in substitution for Al3+, usually in the range of Fe203=0.6—0.9 percent. 2. Zonal spodumene, commonly as large laths that occur in intermediate zones and cores of well- zoned pegmatites. This spodumene is low in iron (Fe203=0.01—0.03 percent). It may contain sig- nificant manganese (X.0—0.X percent), and whereas most is white, some is pink or lilac (kun- zite). 3. Secondary spodumene. Produced by the isochemical decomposition of petalite: petalite —> spodumene + quartz LiAlSi4010 —>LiAlSi205 + 28102 This spodumene is relatively fine grained; com- monly the aggregate retains the crystal form and basal cleavage of the parent petalite. It is white and has very low iron (F6203=0.007-—0.03 percent), because of the low—iron content of the parent peta- lite. The substitution of even small amounts of Fe3+ for Al?‘+ in spodumene markedly increases its thermal sta- bility field (Appleman and Stewart, 1968). Iron spodu— mene was synthesized in air at 1,100°C and is stable up to at least 800°C at 2Kb H20 pressure compared to ~565°C at 2Kb for essentially iron-free spodumene. Geothermometric data obtained by Sheshulin (1963) from fluid-inclusion studies, indicate that early (his Generation 1) spodumenes crystallized between 600° and 640°C, whereas younger (his Generation 11) spod- umenes crystallized at 330°—390°C. Thus phenocrystic spodumene, which is iron rich as compared to paragenetically younger intermediate zone and core spodumene, represents a higher temperature spodumene. After secondary spodumene has been formed from petalite, it may be recrystallized and texturally re- arranged, becoming coarser, unoriented, and very dif- ficult to distinguish from primary white inner zone spodumene. Some such spodumene occurs at Bernic Lake where Cerny and Ferguson (1972) distinguish three types of spodumene: Type A, tabular aggregates of fibrous spodumene intergrown with quartz, clearly pseudomorphous after petalite which constitute 90 per- cent of all spodumene in the pegmatite. Type B, laths up to 1.5 m long, in the intermediate core-margin zone, not associated with petalite, and identified as primary zonal spodumene. Type C, fibrous to columnar spodumene in quartz, but not tabular (may be secon— dary). VARUTRASK, SWEDEN The Varutr'ask pegmatite, described in 37 articles by Percy Quensel and his students (see Quensel, 1952, 1956) is “C”-shaped in plan with the “C” open to the north, the western, thinner limb which trends north- westward dips inward to the northeast at a low angle, whereas the eastern thicker limb which trends north- eastward dips 30° NW. Thus the pegmatite is synclinal with the trough plunging to the north at a shallow angle. The internal structural units are as follows: 1. Border zone, 1—3 cm, fine-grained, quartz and mus— covite. 2. Wall zone, up to 5 m, coarse-grained, quartz, musco— vite, schorl, and locally beryl. 3. Outer intermediate zone, microcline-perthite, quartz, muscov1te. 4. Inner intermediate zone, microcline-perthite, quartz, muscovite, amblygonite, spodumene-quartz pseu- domorphous after petalite, petalite remnants. Units 3 and 4 together are 40—75 m thick. 5. Pollucite masses (Quensel’s “Caesium replacement unit”), as much as 5 in across, which in the eastern limb are marginal to the core. Core, quartz, 12 m across. . Lepidolite replacement unit. This is centrally located in the western limb and replaces mainly rock of the inner intermediate zone. In the eastern limb it also replaces parts of this zone but in addition trans- gresses into the outer intermediate zone and the wall zone. _ 8. Cleavelandite replacement bodies, irregularly de- veloped in all older units. “F” BERNIC LAKE, MANITOBA The geology of the Tanco pegmatite at Bernic Lake, Manitoba, has recently been described in detail in a series of papers by Cerny and Ferguson (1972). The east-west—striking pegmatite is an almost horizontal sheet that crops out only in the bed of Bernic Lake from whence it dips at a low angle to the north and plunges slightly both toward the east and west. As in most fiat- lying pegmatites, the zonal configuration is markedly 52 asymmetric with outer zones 1, 2, 4, and 5 (see below) forming a series of shells that enclose the lensoid to tabular inner units 3, 6, 7, 8, and 9 (see below). Of these inner tabular units the albitic fine- to medium—grained assemblages are concentrated in the lower interior, whereas lithium- and silica-rich assemblages with giant- crystal dimensions predominate in the upper interior. The sequencevof units in probable order of formation is (Crouse and Cerny, 1972): 1. Border zone, albite, quartz, locally tourmaline. 2. Wall zone, albite, quartz, microcline—perthite, locally tourmaline. 3. Sheet and lenses of aplitic (“sugary”) albitite. 4. Lower intermediate zone, microcline-perthite, albite, quartz, spodumene + quartz, (replacement of petalite), locally amblygonite. 5. Upper intermediate zone, spodumene (laths), spodumene + quartz (replacement of petalite), quartz, locally amblygonite and petalite. 6. Central intermediate zone, microcline-perthite, al- bite, quartz, locally beryl and wodginite. 7. Core units, lenses and sheets of quartz in several dif- ferent positions. 8. Lens-shaped bodies of pollucite in zone 5 and be- tween zone 6 and hanging-wall zones 1 and 2. 9. Sheets and pods of lithian muscovite rock in zone 6. BIKITA, RHODESIA The Bikita pegmatite is an irregular sheet that strikes north-northeast and dips 14—45° eastward.1 It ranges in thickness from 95 to 210 ft and has been explored over a strike length of 5,100 ft. Again the internal zonal struc- ture is strongly asymmetric, and there are major changes in the sequences of units along the strike. Our first knowledge of the internal structure stemmed from Symons (1961) and Cooper (1964) who both unfortu- nately use a variety of nonequivalent terms (mineralogi- cal, structural, textural, and grade) to designate the units. Studies by Gallagher (1962) and by the writer in 1966 and 1967 (Heinrich, unpublished company report, 1966) have supplied additional information on the un- its. Table 17 is an attempt to decipher and interpret the interrelationships. COMPARISON The three pegmatites are compared in table 18. The reassessment of the nature of the finer grained units in the Bikita pegmatite demonstrates that Bikita and Ber- nic Lake are structurally very similar, both having aplitic albitites (some wavy banded) and lepidolite replacement ‘Actually the “main Bikita pegmatite" may be but one limb (the easternmost) of a much larger undulatory sheet. The eastern limb passes north of the Al Hayat sector into an anticli- nal unit, and thus connects southwestward into a synclinal phase to the east of which lies a third main unit, an east—dipping synclinal limb (see Gallagher, 1962). LITHIUM RESOURCES AND REQUIREMENTS BY THE YEAR 2000 units. The latter is centrally located in both but is much more strongly developed, both in size and grade, at Bikita than at Bernic Lake. At Bikita the lithium mica is chiefly lepidolite, whereas at Bernic Lake it is predom- inantly lithian muscovite. As at Bernic Lake, the aplitic albitites at Bikita are concentrated in the lower part of the pegmatite. One of the main differences between Bikita and Ber- nic Lake is in the extent to which the petalite has been transformed to spodumene + quartz. At Bikita about 50 percent of the petalite has been converted; at Bernic Lake probably about 90 percent of the petalite has been transformed. RESERVES VARUTRASK There has been no mining at Varutrask since early 1950, but in late 1960 “some stockpiles of lithium min- erals” still were available (Erland Grip, written com- mun., 1967). Doubtless these are small, and the total amount of lithium in all reserves at Varutrask (petalite, spodumene, lepidolite, amblygonite) is small, possibly no more than about 100 tons of ore. BERNIC LAKE Lithium reserves for the Bernie Lake pegmatite were calculated by Howe and Roundtree (1967) who esti- mated lepidolite reserves at 107,700 tons containing an average of 2.24 percent LigO and spodumene reserves at 4,727,263 tons containing an average of 2.01 percent L120, excluding ore that must be left in a pillar unless the overlying lake is drained. These data indicate a total of 45,000 tons of lithium (Norton, 1973). Since then drilling has considerably extended the dimensions of the spodumene ore body, and the total lithium reserves are at least 50 percent greater. BIKITA Symons (1961) calculated reserves of 6,700,000 tons of all lithium-mineral ores (petalite, spodumene, lepido- lite, amblygonite, eucryptite) averaging 2.90 percent L120 or 90,000 tons of lithium (Norton, 1973). Calcula- tions on measured reserves of all types of ores known at the end of 1965 available to the writer show that 76,000 tons of lithium were in these reserves. This total does not include the unmeasured resources of the northern extension and the two western “arms.” REFERENCES CITED Appleman, D. E., and Stewart, D. B., 1968, Crystal chemistry of spodumene-type pyroxenes [abs]: Geol. Soc. America Spec. Paper 101, p. 5-6. Cemy, Petr, and Ferguson, R. B., 1972, The Tanco pegmatite at Ber- LITHIUM RESOURCES AND REQUIREMENTS BY THE YEAR 2000 53 TABLE l7.-——Intemal structure of the Bikita, Rhodesia, pegmatite from hanging wall downward to footwall Symon‘s (1961) and Cooper's (1964) designations Al Hayat Sector Bikita Sector Mineralogy Interpretation (this paper) Border zone ____________________________ Wall zone—mica band ___ Hanging-wall feldspar zone ______________ Intermediate zone: petalite-feldspar zone __________________ Intermediate zone: spodumene zone ______________________ Intermediate zone: feldspathic lepidolite zone ______________ Intermediate zone: pollucite zone ________________________ Intermediate zone: “all mix" zone ________________________ Footwall feldspar zone __________________ Border zone Wall zone—mica band ...... - Hanging—wall feldspar zone ____________ Intermediate zone: petalite-feldspar zone ________________ Intermediate zone: spodumene zone ____________________ Intermediate zone: pollucrte zone ________________________ Intermediate zone: feldspar-quartz ______________________ Intermediate zone: “all mix" zone ________________________ Core zone: “High grade" ,1 lepidolite __________ “Near solid“ 1, lepidolite __ - lepidolite—quartz shell ______________ Lower intermediate zone “cobble zone.“ Lower intermediate zone feldspathic 1e idolite. Not recognized?) Rhythmically banded beryl zone Muscovite band ________________ ”Spotted dog" zone ____________________ y Cooper ______________ f. gr. alb., qtz., musco ____________________ c. gr. musco., qtz _____ micro.-perth., qtz.. musco ________________ pet., micro.-perth., alb., musc., qtz ________ spod.—qtz.,_ in matrix alb., musco., qtz., eucryptite, blkltaite. f. gr. lepid., alb., qtz ____________________ pollucite, lepid __________________________ micro.-perth., qtz ________________________ micro.-perth., qtz., pet, spod.-qtz., beryl + cleav., lepid. lepid __________________________________ lepid., qtz ________ lepid., gtz” amblyg. lepit., a b., qtz __________________________ lepid., alb., qtz __________________________ alb., qtz ________________________________ lepid., alb., qtz., beryl, tant. musco., qtz alb. + blebs of musco.—qtz- 1) Hanging-wall border zone. 2) Hanging-wall wall zone. 3) Intermediate zone. 4) Intermediate zone. The spodumene—quartz aggregates represent former petalite crystals. 5) Aplitic lepiodolite albitite. 6) Pollucite zone. 7) Core-margin intermediate zone. 8) Spodumene-quartz re lace petalite. Clevelandite and lepiiiolite are sec- ondary. Part of core-margin inter- mediate zone. 9) Lepidolite replacement of a quartz core. Quartz increases downward. 10) Remnants of quartz-amblygonite core. Lepidolite “bands“, lenses, veins in banded a litic albitite. Lepidolite isseminated in aplitic al- bite—quartz rock. Aplitic albitite. Banded aplitic lepidolite albitite. I2) Footwall wall zone. Footwall border zone. Abbreviations: alb. = albite; amblyg.‘= ambly onite; cleav. = cleavelanditic albite; lepid. = lepidolite; micr0.-perth. = microcline-perthite; musco. = muscovite; pet. = petalite; qtz. = quartz; spod. = spodumene, tam. = tantalite; f. = me; c. = coarse; gr. = grained; l~= sequence from center core outward. TABLE 18.—Comparzlron of Varutréisk, Bemic Lake, and Bikita pegmatite: Characteristic Varutrask (1.7 b.y.) Bernic Lake (2.6 b.y.) Bikita (2.65 b.y.) Shape and attitude __________ Size (strike and dip) __________ Thickness Wall rock ____________________ Wallrock alteration Internal structure ____________ Major accessory elements Minor accessory elements ____ Lithium mineralogy __________ Types of spodumene ________ Economic minerals Flat-troughed sheet __________ 1,050 ft(+?) __________________ 10-100 ft ____ Amphibolite __ Albite, tourmaline ____________ Relatively symmetric; lepidolite re lacement. Li, Nb, Ta, Ti, Sn, U, Mn Be, As, Sb. Petalite, spodumene, amblygo- nite, lepidolite, Li-muscovtte, cookelite, triph lite -lithi- ophilite, ferri-siczlerite. Secondary (major); recrystal~ lized secondary (minor). Petalite, spodumene-quartz, p01- lucite, lepidolite. s, Rb, B, P, F ____________ ' Nearly horizontal sheet ________ 4,000 by 1,500 (+) ____________ 60—280 ft ______________ __ Amphibolite __________________ Biotite, holmquistite, tourmaline, apatite. Strongly asymmetric; aplitic albi- tites; lepidolite re lacement. Li, Cs, Rb, Ta, Ti, Nb, Sn Mn, B, P, Mo, Bi ______ Petalite, spodumene, amblygo- nite, lepidolite, Li-muscovrte, eucryptite, cookeite, triphyllite- lithiophilite, lithiophosphate. Zonal (very minor); secondary (major) (minor). Spodumene-quartz, Ta minerals, pollucite. Irregular to tabular gently dip- pm . 5,100 (+) 95—210 ft Amphibolite. Biotite, tourmaline (P). Strongly asymmetric aplitic albi- tites; lepidolite replacement. Li, Cs, Rb, Be, P. Nb, Ta, Ti, Sn B, Cu. Petalite, s odumene amblygo- nite, 1e idolite, Li-muscovrte, eucryptite, bikitaite. Secondary; recrystallized sec- dondary (very minor). Petalite, spodumene-quartz, lepidolite, amblygonite, eu- cryptite, pollucite, beryl, micro— cline, Ta minerals, cassiterite. nic Lake, Manitoba; IV, Petalite and spodumene relations: Cana~ dian Mineralogist, v. 11, p. 660—678. Cooper, D. G., 1964, The geology of the Bikita pegmatite, in The geology of some ore deposits of Southern Africa, v. II: p. 441— 461. Crouse, R. A., and Cerny, P., 1972, The Tanco Pegmatite at Bernic Lake, Manitoba; I, Geology and paragenesis: Canadian Mineralogist, v. 11, p. 591—608. Gallagher, M. J., 1962, Mineralogy of the Bikita pegmatites, Southern Rhodesia with special reference to beryl deposits in Bikita Main pegmatite: Geol. Survey Great Britain, Atomic Energy Div., Rept. 242, p. 1—20. Heinrich, E. Wm., 1953, Chemical differentiation of multi-generation pegmatite minerals [abs.]: Am. Mineralogist, v. 38, p. 343. 1965, Holmquistite and pegmatitic lithium exomorphism: In- dian Mineralogist, v. 6, p. 1—13. —1967, Micas of the Brown Derby pegmatites, Gunnison County, Colorado: Am. Mineralogist, v. 52, p. 1110—1121. ——1975, Economic geology and\ mineralogy of petalite and spodumene pegmatites: Indian Jour. Earth Sci., v. 2, p. 18—29. —1976, Mineralogy and structure of lithium pegmatites: journal Mineralogia, Djalma Guimaraea Vol. (in press). Heinrich, E. Wm., and Levinson, A. A., 1955, Geological, mineralogi- cal, and crystallographic factors in the exploitation of lepidolite deposits [abs.]: Am. Inst. Mining Engineers Program, Ann. Mtg, 1955, p. 29—30. Howe, A. C. A., and Roundtree, J. C., 1967, Geology and economic significance of the Bernic Lake pegmatite: Canadian Inst. Mining 54 Metallurgy Bull, Feb. 1967, p. 207-212. Norton, J. J., 1973, Lithium, cesium, and rubidium—the rare alkali elements, in United States mineral resources: U.S. Geol. Survey Prof. Paper 820, p. 365—378. Quensel, Percy, 1952, The paragenesis of the Varutrask pegmatite: Geol. Mag. [Great Britain], v. 89, p. 49—60. 1956, The paragenesis of the Varutrask pegmatite including a review ofits mineral assemblage: Arkiv. Mineral. Geol., v. 2, no. 2, LITHIUM RESOURCES AND REQUIREMENTS BY THE YEAR 2000 p. 9—125. Sheshulin, G. I., 1963, Composition of gas-liquid inclusions in the minerals of spodumene pegmatites, in New data on rare element mineralogy, Ginzburg, A. 1., ed.: New York, Consultants Bur., p. 47—55. Symons, Ralph, 1961, Operation at Bikita Minerals (Private), Ltd., Southern Rhodesia: Inst. Mining and Metallurgy Bull. 661, p. 129—172. \_/—\ NONPEGMATITE LITHIUM RESOURCE POTENTIAL ByJAMEs D. VINE U.S. GEOLOGICAL SURVEY, DENVER, CO ABSTRACT The lithium content of most sedimentary rocks ranges from about 0.0005 to 0.01 percent, but rocks deposited in basins that contain evaporite minerals may contain 0.01-0.1 percent lithium, and clay minerals related to hectorite containing 0.3—0.5 percent lithium have been reported from several localities in the western United States. Although marine evaporite sequences that include potassium salts may suggest conditions favorable for the entrapment of lithium-rich brines, nonmarine evaporite basins probably represent the most favorable trap for concentrated lithium brines. Areas known to con— tain such exotic salts as borates and nitrates in nonmarine basins of Tertiary age also contain anomalous concentrations of lithium in clas- tic rocks, volcanic ashes, clays, and magnesium-rich carbonate rocks. Once dissolved, lithium tends to remain in solution in residual brines, even after evaporative concentration and precipitation of the salts of sodium, potassium, and magnesium. Brines associated with these evaporite minerals may contain as much as 1,000 times more lithium than seawater contains. INTRODUCTION The distribution of lithium is less well understood in sedimentary rocks than in igneous rocks where the con- centration of lithium in the late stages of crystallization and in mineralizing fluids expelled from the magma chamber is fairly well documented (Heier and Billings, 1972). All the commercially important lithium minerals occur in pegmatites or greisen associated with igneous rocks. Recent studies (Horstman, 1957; Ronov and others, 1970; Tardy and others, 1972) suggest that most of the lithium content of sedimentary rocks is associated with the clay minerals. Because clays are so widely dis— persed in sedimentary rocks and the manner of lithium fixation on clays is generally weak, lithium is only rarely concentrated in sedimentary rocks in amounts of more than a few tens of grams per tonne. The exceptions to this broad dispersal of lithium in sedimentary rocks are of special interest in the search for potential new sources of lithium. NORMAL MARINE SEDIMENTS Seawater has an average lithium content of slightly less than 0.2 mg/l (Goldberg, 1963), an amount that is equal to the influx from river waters in about 6.4:x106 years (Wedepohl, 1968, p. 1010). Hence, the residence time for lithium in the sea is a relatively short interval of geologic time, equal to about one—tenth the interval of time since the end of the Cretaceous Period. Thus, an amount of lithium equal to the total quantity in the ocean must be removed every 6.4 m.y. Marine clays, which contain an average of about 60 g/t lithium, pro- vide the reservoir for this excess lithium. However, this process of adsorption of lithium on clay particles is so uniform throughout the ocean that it is unlikely that any significant enrichment of lithium will occur in the sedi— ments deposited in normal marine environments. MARINE EVAPORITES Local marine basins with restricted circulation may develop hypersaline environments, and in warm dry re- gions these waters will become concentrated enough to precipitate evaporite minerals, including gypsum (CaSO4- 2H20), halite (NaCl), and various complex salts of potassium and magnesium. Such evaporite deposits have been estimated to compose no more than 1 percent of the total mass of sedimentary rocks (Wedepohl, 1969, p. 265), and most of these are of marine or marginal marine origin. However, lithium is so soluble it will not precipitate by evaporative concentration until after most of the sodium, potassium, and magnesium salts have been precipitated. Lithium precipitation is so rare a situation that natural lithium salts have never been ob- served. In order for a restricted basin to accumulate a large quantity of salt, it must have a regular source of fresh LITHIUM RESOURCES AND REQUIREMENTS BY THE YEAR 2000 seawater by transport across a shallow bar or by periodic flooding during storms. The heavier, more concen- trated brines may be preserved by entrapment in the pore spaces of the already precipitated salts, but some brine may also be mixed and diluted with the fresh sea- water. Some of the diluted brine will generally escape to the open sea by ebbtide or through gravity return across a permeable bar. Hence, in a basin of marine evaporites, if we sum the sodium and lithium in both the precipi- tated salts and the entrapped brine, the ratio Li:Na may be less than, but cannot exceed, the ratio (0.0001) for normal seawater. With continued deposition and burial, the body of permeable salt and brine will recrystallize into a solid dry mass causing the brine that once filled the pores to be expelled into adjacent clastic rocks such as sandstone. Most of the lithium will be carried away from its original environment of deposition into these adjacent sedi- ments. Because only those salts deposited from the most con- centrated marine brines contain potassium minerals, the occurrence of potassium may be an indication of the former existence of a lithium-rich brine. Lithium does not significantly substitute for sodium or potassium in the common salt minerals, but the salts deposited from the most concentrated brines may contain anomalous amounts of bromine. This relation provides a way to recognize areas favorable for potassium exploration, and it may help to guide the search for lithium brines or for neoformed lithium minerals that are a result of the interaction between a brine and the rock in which it was contained. NONMARINE EVAPORITES Nonmarine evaporites are the product of evaporative concentration of water in undrained basins within the continental landmass. Most rivers drain into the sea, but an undrained basin may occur where the crust of the earth is deformed so rapidly—or the rate of evaporation is so high—that the streams cannot maintain their course to the sea. Streams may also be blocked by vol— canic activity, landslides, and glacial activity, but the larger basins such as Death Valley in California and the Dead Sea trough in Israel are the result of tectonic de- formation of the earth’s crust in an arid climate. The Great Basin area in the southwestern United States (fig. 21) includes a relatively large area of un- drained basins, both large and small. Some major rivers flow into this area and contribute to the load of dissolved mineral matter that is carried into a terminal lake or playa. The Owens River of California drains the eastern Sierra Nevada and terminates at Owens Lake, Calif. The Walker River also drains an area of eastern California and empties into Walker Lake, near Hawthorne, Nev. The Truckee River forms the outlet for Lake Tahoe in eastern California and flows into Pyramid Lake, north- 55 east of Reno, Nev. The Bear River drains part of south- eastern Idaho and southwestern Wyoming, and flows into the Great Salt Lake, near Ogden, Utah. The Sevier River, which drains central Utah, terminates at Sevier Playa, south of Delta, Utah. The Humboldt and Reese Rivers flow through a large part of central and northern Nevada and drain into Carson Sink, a large playa, near Fallon, Nev. The structural framework responsible for these basins originated in mid-Tertiary time—some 20—25 m.y. ago—when faulting produced the system of basins and ranges extending from the front of the Wasatch Range, near Salt Lake City, Utah, across western Utah and Nevada to the eastern front of the Sierra Nevada, near Bishop, Calif. Since the initial breakup, some parts of this region have been integrated into the major drainage of the Colorado River. However, the present course of the Colorado River is geologically quite recent, and so some basins that contain upper Tertiary evaporites have only-recently been exposed by erosion. One such body of salt is exposed along the Overton Arm of Lake Mead, Nev. Salt deposits of nonmarine origin differ somewhat from those of marine origin in not being laminated and in being very low in bromine. Moreover, the ratio Li:Na may be much greater in nonmarine evaporites than in marine evaporites. Chemical differences between evaporite basins are pronounced. Great Salt Lake, Utah, and the Great Salt Lake Desert to the west are characterized by high con- centrations of sodium chloride. Others, such as Railroad Valley, southwest of Ely, Nev., contain a large body of salt that includes a body of gaylussite (Nanga(C03)2 '5H20). Bristol Lake in California is characterized by an unusually high concentration of calcium chloride. Other basins in the Mojave Desert area in California are characterized by high concentrations of boron. Clayton Valley, southwest of Tonopah, Nev., is characterized by a high lithium-to-sodium ratio in a chloride brine. At- tempts to explain these differences in terms of the chemistry of the immediate drainage basin are only partly satisfactory. This is because the history of the re- gion is complicated. Many basins received overflow from other basins during previous wet climate, periods‘, and some receive drainage by ground water flow from adja- cent basins at the present time. Significant ground water flow has been suggested as the reason that some basins such as Clayton Valley have a wet surface while adjacent slightly higher basins have a hard dry surface. The origin of most of the major constituents of non- marine evaporites—including the sodium, calcium, .magnesium, as well as the sulfate, chloride, and carbonate—can be explained either by weathering of rocks, including marine evaporites, within the drainage basin or by the airborne salt carried into the area by the prevailing westerly winds off the Pacific Ocean. How- 56 LITHIUM RESOURCES AND REQUIREMENTS BY THE YEAR 2000 ‘0 (VJ/ASHINGTON MONTANA WYCJMI NG COI.ORI\DO Area of Huefior drainage 200 400 MILES l I I I I I 200 400 KILOMETRES o——o FIGURE 21.—Undrained basins in western United States. LITHIUM RESOURCES AND REQUIREMENTS BY THE YEAR 2000 ever, neither explanation seems to account for the un— usually high ratio of lithium to chloride that distin— guishes some of these brines, such as that of Great Salt Lake, from sea-water. Deep ground water circulation with subsurface heat and volcanic activity seems to be the best alternative explanation for the source of the lithium. HYDROTHERMAL WATER Although freshwaters generally contain only a few tens of micrograms per litre of lithium, many hot springs and geothermal waters contain 1—10 mg/l lithium, and a few contain more. The high ratio of lithium to sodium is so characteristic of geothermal wat- ers that a ratio approaching 1:100 may be useful in rec- ognizing a geothermal source, even where the water has been cooled and diluted by surface waters (Ellis, 1975, p. 516). The elevated temperature of geothermal water generally indicates deep circulation of ground water in the vicinity of a buried heat source, such as a magma chamber or volcano. Pore water in the rocks adjacent to such a heat source will tend to circulate as in a giant convection cell. Taylor (1974) has shown that the quan- tity of water pumped to the surface in this manner in mineralized districts such as Tonopah, Nev., is very large. Hot water is more effective than cold water in leaching lithium from rocks; hence, the amount of lithium carried to the surface in hot springs is probably very large. In Yellowstone National Park, and in similar geothermal areas, most of the lithium is lost to streams that carry it directly to the sea. However, the lithium mica lepidolite has been reported from a drill core in the Lower Geyser Basin in Yellowstone National Park (Bar- gar and others, 1973), and lithium—rich clays, from the Pliocene Teewinot Formation, have been reported south of the park near jackson, Wyo. (I. D. Love, writ- ten commun., 1973). Geothermal systems not only circulate lithium, but may in some areas represent the ore-forming fluid re- sponsible for deposition of base and precious metals. A lithium asbestos, known as eckermannite, is associated with lead and other metallic ores at the Camp Albion district, Boulder County, Colo. (Wahlstrom, 1940). Another example is the Spor Mountain beryllium de- posit, Juab County, Utah, where lithium occurs with be- ryllium, fluorite, manganese, and uranium. The possi- bility of finding commercially significant quantities of lithium minerals associated with hydrothermal mineral deposits should not be overlooked. BURIAL METAMORPHISM When sedimentary rocks are deeply buried, many changes take place that we associate with compaction, diagenesis, and burial metamorphism (Noble, 1963). Hydrous minerals that are stable during weathering at the surface of the earth are, when deeply buried, trans- 57 formed to less hydrous or anhydrous species (Coombs and others, 1959; Jolly, 1974). Thus, at depth, hydrous clay minerals such as montmorillonite are transformed to anhydrous illite or mica, zeolites may be altered to feldspars, hydrous iron oxides related to limonite are transformed to hematite or may be reduced to pyrite, and gypsum is transformed to anhydrite. All these changes involve the loss of water, which can be quantita— tively significant. Moreover, the lithium, which is lightly adsorbed by many of these hydrous minerals, may be released to solution. The water, released at‘depth, can- not readily escape to the surface but will be forced into fractures where it may be responsible for the deposition of various vein minerals, such as cookeite. SUMMARY Because pegmatites and brines have supplied suffi- cient lithium to date, there has been no pressing economic reason to look for new sources of lithium. The predicted future demand for lithium suggests that new sources of lithium will be needed. To find new re— sources, new prospecting techniques based on increased knowledge of the behavior of lithium in different geologic environments have to be developed. Hydro- thermal sources of lithium may result in lithium deposits similar to the beryllium deposit at Spor Mountain, Utah. Concentrations of lithium in brines or sediments may also be associated with nonmarine evaporite depos— its or with extremely concentrated marine evaporites. Cookeite deposits may be found in low-grade metamor- phic terranes, and other lithium deposits associated with hydrothermal activity may also be found. REFERENCES CITED Bargar, K. E., Beeson, M. H., Fournier, R. 0., and Muffler, L._]., 1973, Present-day deposition of lepidolite from thermal waters in Yel- lowstone National Park: Am. Mineralogist, v. 58, no. 9—10, p. 901-904. . Coombs, D. S., Ellis, A. j., Fyfe, W. S., and Taylor, A. M., 1959, The zeolite facies, with comments on the interpretation of hydrother- mal syntheses: Geochim. et Cosmochim. Acta, v. 17, p. 53—107. Ellis, A. 1., 1975, Geothermal systems and power development: Am. Scientist, v. 63, no. 5, p. 510—521. Goldberg, E. D., 1963, Composition of sea water, in Hill, M. N., ed., The composition of sea—water [and] comparative and descriptive oceanography, v. 2 of The sea—Ideas and observations on prog- ress in the study of the seas: New York, Interscience Publishers, p. 3—25. Heier, K. S., and Billings, G. K., 1972, Lithium, [Chap.] 3, Pt. B, in Wedepohl, K. H., ed., Handbook of geochemistry, Volume 2—1: Berlin, Springer-Verlag, looseleaf. Horstman, E. L., 1957, The distribution of lithium, rubidium, and cesium in igneous and sedimentary rocks: Geochim. et Cos- mochim. Acta, v. 12, nos. 1—2, p. 1—28. Jolly, W. T., 1974, Behavior of Cu, Zn, and Ni during prehnite- pumpellyite rank metamorphism of the Keweenawan Basalts, Northern Michigan: Econ. Geology, v. 69, no. 7, p. 118—1125. Noble, E. A., 1963, Formation of ore deposits by water of compaction: Econ. Geology, v. 58, no. 7, p. 1145—-1156. 58 Ronov, A. B., Migdisov, A. A., Voskresenskaya, N. T., and Korzina, G. A., 1970, Geochemistry of lithium in the sedimentary cycle: Geochemistry Internat, v. 7, no. 1, p. 75—102. Tardy, Yves, Krempp, Gerard, and Trauth, Norbert, 1972, Le lithium dans les mineraux argileux des sediments et des sols [Lithium in sediment and soil clay minerals]: Geochim. et Cosmochim. Acta, v. 36, no. 4, p. 397—412. Taylor, H. P., jr., 1974, The application of oxygen and hydrogen isotope studies to problems of hydrothermal alteration and ore LITHIUM RESOURCES AND REQUIREMENTS BY THE YEMOO deposition: Econ. Geology, v. 69, no. 6, p. 843—883. Wahlstrom, E. E., 1940, Ore deposits at Camp Albion, Boulder County, Colorado: Econ. Geology, v. 35, no. 4, p. 477—500. Wedepohl, K. H., 1968, Chemical fractionation in the sedimentary environment, in Ahrens, L. H., ed., Origin and distribution of the elements: Oxford, Pergamon Press, p. 999—1016. 1969, Composition and abundance of common sedimentary rocks, Chap. 8, in Wedepohl, K. H., ed., Handbook of geochemis- try, Volume 1: Berlin, Springer-Verlag, p. 250—271. m LITHIUM CONTENTS OF THERMAL AND MINERAL WATERS By DONALD E. WHITE, J. M. THOMPSON. and R. O. FOURNIER U.S. GEOLOGICAL SURVEY, MENLO PARK, CA Reported lithium contents of natural waters range from less than 0.01 ppm to more than 500 ppm (table 19). Although absolute contents are of major interest in commercial recovery, the ratios of Li to Other major constituents are more significant in evaluating origin and diverse effects of evaporation, dilution, and water- rock interaction. The most useful single ratio is Li/Cl, which is emphasized in this text, but Li/Na has nearly equal usefulness. Lithium/chlorine (by weight) is lowest in ocean water (0.00001, table 19) and becomes progressively higher in Thermal and mineral waters of high salinity tend to have intermediate Li/Cl ratios, generally ranging from about 0.0001 to 0.001 (table 19). The median ratio for oil-field waters is probably near 0.0003, but a few ratios attain 0.001 and even higher; one notable example is brine from the Jurassic Smackover Formation of Texas and Arkansas, which averages 0.001. Individual analyses are greater than 0.002 and Li contents are as TABLE 20.—Lithium contents of average rocks, in ppm [Data fromjKrauskopf (1967)] major North American rivers (0.00004) and in small Li M £35“: c1 hit/xiii 1 streams and meteoric springs (variable and unreliable . i 20 24,000 0.0008 130 0.15 values because so dilute, but ranglng from 0.001 to 0.1). 30 28,000 .00” 200 .15 For comparison, the ratios in major crustal types are (133 12.233 28335 123 g; granite, 0.15; basalt, 0.17; and shale, 0.38 (table 20). TABLE 19).—Lithium amtents qf various waters, in ppm Li/Na Li/Cl Li Na ratio Cl ratio Source Water Ocean ___________________________________________________ 0.17 10,500 0.00002 19,000 0.00001 Krauskopf (1967). North American rivers ________________________ - .003 ~9 .0003 8 .0004 Livingston (1963). Meteoric warm springs, ,(median of 5, ~50°) ____ _ .04 ~30 .001 5 .008 Mariner and others (1975). Salton Sea brine, Calif. (~350°C) ___________________________ 215 50,400 .043 155,000 .0014 Whlte (1968). Oilfield brines jurassic Smackover Formation brines of Gulf Coast (average) 174 67.000 0.0026 172,000 0.0010 Collins (1974). Alberta, Canada _____ 81 42,000 .0019 113,000 .0007 White (1965). Kern County, Calif . 9.9 6,700 .0015 15,300 .0006 White (1965). Evaporate brines Clayton Valley (connate?), Nev 1-- 380 66,200 0.0057 95,200 0.004 Kunasz (1970). Searles (connate?), Calif 81 110,000 .0007 121,000 .0007 G. 1. Smith (oral commun., 1975). Permian Salado Formation (conna M 16 14,300 .0011 200,000 .00008 White and others (1963). Permian Salado Formation (solution of salts), N. Mex ....... 2.8 95,000 00003 157.000 00002 White and others (1963). Pennsylvanian Paradox Member; Hermosa Formation (connareP), Utah _______________________________________ 66 6.000 .011 241,000 0003 White (1965). LITHIUM RESOURCES AND REQUIREMENTS BY THE YEAR 2000 much as 500 ppm (Collins, 1974). The interstitial waters of some evaporites, marine and nonmarine, are notably high in Li (table 19): 81 ppm in the upper salt brine of Searles Lake (Li/C1 ratio 0.007) and 300 ppm in the commercial Clayton Valley brines (Li/Cl ratio 0.004). Nearly all saline waters of high temperature geother- mal systems are much enriched in Li. The Salton Sea brine (350°C and 26 percent salinity) has 215 ppm of Li (Li/Cl ratio of 0.0014 and an estimated 1.0 million ton- nes of Li in 5 km3 of brine; White, 1968). Moderately saline geothermal waters exceeding 200°C and closely 59 associated with young silicic volcanism range from about 1 to 50 ppm of Li and have the highest Li/Cl ratios of any major group of thermal and mineral waters, rang- ing from less than 0.001 to more than 0.02; the median ratio is about 0.004 (White and others, 1963). The rela- tively dilute thermal waters of Yellowstone National Park are especially notable, having an average Li con- tent of about 3 ppm and an Li/Cl ratio of about 0.01. The calculated discharge of Li from Yellowstone Park is about 480 tonnes per year, or 48 million tonnes over the probable minimum activity of 100,000 years (table 21). TABLE 21,—Lithium contents of waters of Yellowstone National Park, Wyoming, in ppm [From Thompson and others (1975)] Li/Na Li/Cl Tem erature Location pH Li Na ratio Cl ratio C‘) Near-neutral chloride waters U r Ge ser Basin: pp(Ear “y _________________________________ 9.2 5.2 335 0.016 '417 0.012 95 Giantess ______________________________ 9.2 5.7 365 .016 439 .013 95 Punch Bowl __________________________ 8.3 4.0 425 .009 295 .014 95 Sapphire ______________________________ 9.0 2.3 450 .005 308 .007 95 Interchange __________________________ 8.6 3.2 300 .011 226 .014 74 S ring near Y—8 ______________________ 8.9 2.9 429 .007 293 .010 79 yriad near C—l ______________________ 9.9 4.0 310 .013 404 .010 94 Midway-Lower Basins: Rabbit Creek (Y—5) ____________________ 8.4 3.3 375 .009 271 .012 95 Fountain P.P. ‘near Clepsydra __________ 9.4 2.8 380 .007 325 .009 92 Porcupine Hills near Y—I3 ______________ 9.2 4.6 341 .013 312. .015 93 Ojo Caliente (Y—3) ____________________ 8.2 4.2 335 .013 331 .013 95 River Group __________________________ 9.4 3.4 339 .010 323 .011 93 Fairy Meadows ________________________ 8.7 3.4 390 .007 319 2011 87 Imperial Group ________________________ ~9.l 2.4 290 .008 222 .011 93 Sentinel ______________________________ ~8.5 1.7 309 .006 300 .006 94 S lvan ____________________________________ 5.7 4.8 410 .012 531 .009 93 orris: . Cistern ________________________________ 7.1 4.3 316 .014 476 .009 93 Fenner drillhole ______________________ 4.6 7.9 440 .018 720 .01 1 92 Porcelain Terrace, base ________________ 8.4 6.8 400 .017 713 .010 93 Pork Cho ____________________________ 8.0 6.3 470 .013 780 .008 91 Yellowstone Canyon (7 mile) ________________ 8.9 3.3 347 .010 376 .009 92 Vermillion Group __________________________ 8.6 .35 430 .0008 244 .0014 67 Rainbow Grou ____________________________ 8.4 .55 354 .0016 304 .0018 92 Heart Lake, Spike Group __________________ 9.5 6.7 400 .017 363 .018 93 Lewis Lake ________________________________ 8.0 .6 140 .004 79 .008 Hot Shoshone Basin, Bronze Group ____________ 8.8 1.5 315 .005 167 .009 93 Lone Star (Y—6) ____________________________ 8.2 1.7 329 .005 417 .004 68 Travertine-depositing springs Hillside __________________________________ 8.6 0.84 151 0.006 73 0.012 83 Firehole Lake, Steady ______________________ 8.4 .40 85 .005 44 .009 93 Terrace Springs __________________________ 7.3 .75 305 .002 65 .012 59 Mammoth ________________________________ 7.3 1.6 130 .012 166 .010 73.5 Thermal, acid SO4—Cl waters Norris Basin: . Little Whirligig ________________________ 3.2 4.6 345 0.013 582 0.008 93 Realgarz ______________________________ 3.2 3.6 334 .011 492 .007 88.5 Horseshoe2 ____________________________ 2.5 2.8 235 .012 341 ' .008 86 Echinus2 ______________________________ 3.2 .93 160 .006 103 .009 85 Emerald2 ______________________________ 3.05 3.6 308 .012 460 .008 89 Unnamed ____________________________ 3.3 3.7 290 .013 468 .008 54 Green Dragon ________________________ ~2.7 2.3 145 .016 216 .011 ~93 Green Dragon ________________________ ~2.7 4.0 273 .015 447 .009 92 Northwest of Fenner drill hole __________ acid .70 63 .011 88 .008 63 Shoshone Basin ____________________________ 2.4 .16 60 .003 60 .003 93 60 LITHIUM RESOURCES AND REQUIREMENTS BY THE YEAR 2000 TABLE 2l.——Lithium contents of waters of Yellowstone National Park, Wyoming, in ppm—Continued . . l.i/Na Li/CI Temperature Location pH Li Na ratio C1 ratio ( C‘) Thermal, acid SO4 waters Summit Lake2 ____________________________ 2.8 0.02 30 0.0007 5 0.004 94 Washburn ________________________________ 4.5 .02 14.8 .001 .1 .2(?) ~90 Hot Sprin s Basin ________________________ 1.8 .01 11.0 .0009 .l .1(?) 76 East of Su phur Cauldron __________________ 2.0 .01 25.4 .0004 2.1 .005 88 Heart Lake2 ______________________________ 2.8 <.01 2.2 <.005 3 <.003 87.5 joseph Coats .............................. 1.8 .01 7.0 .001 .1 .1(?) 86 Dilute neutral waters Cold Streams Thumb Creek ________________________ 7 5 0.04 27 0.0015 1.4 0.03 5 Little Thumb __________________________ 7.4 .01 4.4 .002 1.3 .008 7 Elephant Back ________________________ 7.3 .02 9.2 .002 .8 .03 Cold Cold Sprin s: West 0 Black Sand ____________________ 7 5 .02 6.7 .003 3.8 005 14 Midway Picnic area ____________________ 7.3 .12 32 .004 17 007 21 Sylvan area ____________________________ 7.80 .02 3.0 .007 .2 1(?) 5 Thermal meteoric waters: Fairy Meadows ________________________ not reported .09 16.2 .006 l .06 26 Southwest of Gibbon Hill ______________ 8.6 .12 63 .002 1.6 .08 68 North of Gibbon Falls __________________ 6.0 .1 24 .004 .8 12 31 Specimen in Geyser Creek area ________ 7.9 .05 65 .0008 5 01 92 Hillside ______________________________ 7.4 .04 10.4 .0004 4.0 01 32 North of Biscuit Basin __________________ 6 9 .01 10.0 .001 .8 01 48 ‘Boiling at this elevation is 93°C. Springs above 93°C are superheated. 2Unpublished data of Thompson. The most probable explanation for Li—enriched geoth- ermal waters is selective leaching of part of the Li from rocks at temperatures above 200°C; 300° to 500°C may be especially favorable. The evidence for direct in- volvement of magmatic fluid high in Li and other con- stituents is a real but still controversial possibility. Lithium is evidently removed from river and ocean waters and is fixed by low-temperature reactions in clay minerals. In simple evaporational processes, most brines require source waters that are higher in Li/Cl than is the brine. The Clayton Valley brine (table 19), which has an Li/Cl ratio of 0.004, probably requires a main supply from hot springs containing an Li/Cl ratio of 0.01 or higher and a localized evaporation basin with little inflow of other C1. The upper Tertiary rhyolitic vol- canism and hydrothermal ore deposits west of Clayton Valley may have been essential in supplying Li. Diagenetic, metamorphic, and igneous-related water-rock interactions at moderate to high tempera— tures are probably essential in accounting for the rela- tively high Li contents of most thermal and mineral waters. REFERENCES CITED Collins, A. G., 1974, Geochemistry ofliquids, gases and rocks from the Smackover Formation: U.S. Bur. Mines Rept. Inv. 7897, 84 p. Krauskopf, K. B., 1967, Introduction to geochemistry: New York, McGraw-Hill Book Co., Internat. Ser. Earth and Planetary Sci., 721 p. Kunasz, I. A., 1970, Geology and geochemistry of the lithium deposit in Clayton Valley, Esmeralda County, Nevada: Pennsylvania State Univ. Ph.D. thesis, 128 p. Livingston, D. A., 1963, Chemical composition of rivers and lakes: U.S. Geol. Survey Prof. Paper 440—G, 64 p. Mariner, R. H., Presser, T. S., Rapp, J. B., and Willey, L. M., 1975, The minor and trace elements, gas, and isotope compositions of the principal hot springs of Nevada and Oregon: Menlo Park, Calif, US. Geol. Survey 0pen~file rept., 27 p. Thompson,j. M., Presser, T. S., Barnes, R. B., and Bird, D. B., 1975, Chemical analysis of the waters of Yellowstone National Park, Wyoming, from 1965 to 1973: US. Geol. Survey open-file rept. 75—25, 53 p. _ White, D. E., 1965, Saline waters of sedimentary rocks, in Young, A., and Galley, ]. E., eds., Fluids in subsurface environments—a sym- posium: Am. Assoc. Petroleum Geologists Mem. 4, p. 342—366. 1968, Environments of generation of some base-metal ore de- posits: Econ. Geology, v. 63, no. 4, p. 301-335. White, D. E., Hem, j. D., and Waring, G. A., 1963, Chemical composi- tion of subsurface waters, in Data of geochemistry [6th ed.]: U.S. Geol. Survey Prof. Paper 440—F, 67 p. m LITHIUM RESOURCES AND REQUIREMENTS BY THE YEAR 2000 61 LITHIUM RECOVERY FROM GEOTHERMAL FLUIDS1 By C. E. BERTHOLD, HAZEN RESEARCH, INc., GOLDEN, CO, and D. H. BAKER, JR.,U.S. BUREAU or MINES, BOULDER CITY, NV ABSTRACT The lithium resource of the Imperial Valley, Salton Sea KGRA (Known Geothermal Resource Area), in southern California, appears to be of significant size, comparable to the Clayton Valley, Nevada, lithium reserve in quantity of contained lithium. Lithium occurs, presumably as lithium chloride, in amounts of about 200 ppm (as Li) in the geothermal fluids obtained from depths of 5,000 feet or more (as typified by Sinclair No. 4 fluid). Recovery of lithium from this resource can be achieved by chemical concentration (precipitation) techniques and ion-exchange methods, both liquid and solid, among others. The paper will discuss in general terms the studies undertaken and preliminary results obtained in attempting to recover lithium from this resource. INTRODUCTION Geothermal fluids from various sources in the West- ern United States contain quantities of lithium ranging from the low parts per million level to concentrations of hundreds of parts per million (ppm). Some of these geothermal resources can be considered to be signifi- cant, if not major, resources or reserves of lithium. In some studies on recovery of mineral values from geothermal fluids for the US. Bureau of Mines, 3 sam- pling survey was made of several hot springs and geo- thermal sites in the Western United States. Table 22 presents a summary of the lithium content found in some of these geothermal fluids and hot springs waters. As can be seen, the geothermal fluids of the Salton Sea Known Geothermal Resource Area (KGRA), Imperial TABLE 22.—Lithimn contents of selected geothermal fluids and hot spring waters1 Lithium concentration Location (ppm Li) Mesa L—6—1, Imperial County, Calil'., pre-flash, 8,000-ft depth ______________________________________________ 55 Magmamax No. 1, Imperial County, Calif, pre-flash, 2,400-ft de th ______________________________________________ 143 Sinclair No. 4, ImperialPCounty, Calif., post-flash, 5,400-ft de th ____________________________________________ 238 Coso Hot Spring), Inyo ounty, Calif, average of pu lished analyses ________________________________________ Trace to 0.1 Darrou h Hot Springs, Nye County, Nev. post- ash ____________________________________________________________ .06 Golconda Hot Springs, Humboldt County, Nev.: No. 1 ____________________________________________________________ .45 1.0 post-flash ____________________________________________________________ 3.3 ‘From work performed by C. E. Berthold and others,]une 1973 for the US. Bureau of Mines under contract No. SO 133084, Process Technology for recovering geothermal brine minerals. ‘The data and results presented herein were obtained during the performance of work for the US. Bureau of Mines under contracts 50 133084 and H0 144104. The views and conclusions presented herein should not be interpreted as necessarily representing the official policies or recommendations of the Interior Department, Bureau of Mines, or of the US. Government. County, Calif., contain significant concentrations of lithium. Lithium reserves within this KGRA site have been compared in quantity with those of the Clayton Valley, Nevada, area, that is, approximately 40,000 tonnes of contained lithium (Kunasz, 1975). The lithium content of the Salton Sea geothermal fluids appears to vary, both with respect to location within the reservoir and depth of the producing zone. As an example, Magmamax No. 1 well produces brine from the 2400-foot level and has a lithium content of about 150 ppm with a total salinity of approximately 180,000 ppm. Contrasted to this geothermal fluid are those of Sinclair No. 4 brine (well located one mile south of Magmamax No. 1) at 5400-foot depth, with over 200 ppm lithium and a total salinity of nearly 300,000 ppm. Figure 22 shows the locations of these geothermal wells in the Imperial Valley area. Lithium occurs in the Salton Sea KGRA fluids, pre- sumably as lithium chloride, since the predominant anion is chloride, with small amounts of bromide, fluoride, and sulfate also present. The major cations present in these geothermal fluids are sodium, calcium, potassium, and ammonia, with minor amounts of mag- nesium, rubidium, and cesium also present. Table 23 presents a partial analysis for Mesa, Sinclair, and Mag— mamax geothermal fluids, together with a comparison of lithium-sodium and lithium-calcium ratios for these geothermal fluids with other naturally occurring saline brines. It appears that brine from Mesa L—6—1 well at the southern fringe of the Salton Sea KGRA is considerably higher in Li/Ca ratio as compared to other geothermal fluids found in this KGRA. Whether this implies a dif- ferent source for this fluid or a different subsurface rock matrix that supplies the lithium for this fluid is not known. In any event, recovery of lithium from Mesa L—6—1 fluid would appear to be somewhat easier than from either Sinclair No. 4 or Magmamax No. 1, owing to the relatively higher lithium level with respect to the other salines present. Recovery of lithium from such naturally occurring geothermal fluids is complicated by several factors: (1) relatively low concentration in the fluid, (2) high temp- erature of the fluid, and (3) presence of relatively large amounts of other potentially valuable minerals such as iron, manganese, silica, zinc, lead, barium, strontium, and magnesium. In general, geothermal fluids from the 62 LITHIUM RESOURCES AND REQUIREMENTS BY THE YEAR 2000 8/1 L 7'0/V SEA Mogmomax No. zuMogmamax No. 3 @ Magmamax No. |..Woolsey NO-l Sinclair No. 4' , Sinclair NO. 3 ‘Cah'pafria v \l 1% ‘ N Z i 9—; O .Westmoreland @ o 5 IO MILES i . . .1 . i ' l 1 o 5 IOKILOMETRES Brawley [(A V. ( «We m 2. <5 \ Imperial $1“ C) Holtville El Centro a U.SB.R. Mesa 6-! ... '3) .i/\i Hol’rz No. I Holt\z No. 2. Nowlin No 2 Nowlin NO. I /\ cALlFot‘g‘IIfi __- -—— Calexico ______ "B‘EfiTAI—LIFORNIA s T A T S”.— - “'— N T EB.._————--—' . . ___—_.._—--——--"M I C Mexucah FIGURE 22.—Imperial Valley geothermal area showing location of geothermal wells, highways, and cities. Salton Sea KGRA require clean-up, or removal and perhaps recovery, of the silica, iron, manganese, lead, zinc, and magnesium therefrom prior to recovery of the lithium values. RECOVERY METHOD Recovery methods for lithium contained in solution can be broadly categorized into two general schemes: (1) LITHIUM RESOURCES AND REQUIREMENTS BY THE YEAR 2000 63 TABLE 23.—Analyses, in ppm, of geothermalfluids Sample Br Cl Na Ca K Mg NH4 Li F 504 Li/Na Li/Ca Mesa L—6—1 35 18,000 11,000 1,370 1,430 22 39 55 1.5 16 0.005 0.04 Sinclair No. 4 - 162 186,000 71,000 35,000 18.000 152 611 238 5.8 42 .003 .007 Magmamax N0 109 102,000 44,000 21,000 8.300 100 504 143 4.6 50 .003 .007 Clayton Valley, Nev. .1 .0045 .6 Searles Lake, Calif ____________________ .007 No Ca present direct recovery from the brine via liquid (Baniel and Blumberg, 1963; Fletcher and Wilson, 1961; Grinstead and Davis, 1970; Morris and Short, 1963; Morrison and Freiser, 1962; various patents2 and written communica- tions3) or solid ion exchange (Kennedy, 1961) and (2) removal of the lithium by precipitation as an insoluble compound or double salt. In discussing methods of lithium recovery from geothermal fluids, it should be kept in mind that such recovery processes are secondary to the original intent behind the development of geothermal resources, namely power generation and production of freshwater. Whatever method or methods are used to recover lithium from such resources must be adaptable to the constraints imposed upon geothermal wells. These are: l. The almost universally accepted belief that waste geothermal fluids must be reinjected back into the producing structure to maintain hydraulic balance and prevent ground surface subsidence. Alterna- tively, fresh or brackish water, if available, could be used as a substitute fluid, thus allowing some con— centration of the geothermal fluid, if this would aid in recovering some of the constituents from the geothermal fluid. 2. Whatever methods or means used for lithium recov- ery must be compatible with the need for geother- mal fluid reinjection. That is, no material should be present that might precipitate out in the subsur- face structure, thereby leading to plugging-off of the aquifer(s). The presence of organic com— pounds, such as might result from liquid ion- exchange methods of lithium recovery, constitutes yet another unknown if reinjected back into the producing zone. Keeping these constraints in mind, a series of labora- tory studies on lithium recovery was started, using Sinclair No. 4 geothermal fluid as the raw material. Post-flash Sinclair No. 4 geothermal fluid required a “clean up” treatment to remove soluble silica, iron, ’[srael patent 16017, and U.S. patents numbered 3,307,922; 2,964,381; 3,306,700, U.S. Department of Commerce, Patent and Trademark Office, Washington, D.C. C’C. Hanson, 1970 written commun., “Extraction of magnesium chloride from brines using mixed ionic extractants,” University of Bradford, Bradford, York. W. C. Keder, 1970, written commun., from Kedder and others, “Separation of alkali metals by solvent extraction with mixtures of organic-soluble acids and phenols," Battelle Pacific Northwest Laboratory, Rich- land, Washington. R. L. Focht and others, 1961, oral commun., “Separation of lithium alumi- nate" paper presented at the Pittsburgh Conference on analytical chemistry, Pittsburgh, Penn., 1961. manganese, zinc, and lead prior to lithium removal. This was achieved by adjusting pH to a nominal value of 7.5 to 8.0 and removing the precipitated hydroxides, together with adsorbed zinc and lead values. At this point, the purified Sinclair No. 4 fluid can be used di- rectly for lithium recovery or else concentrated (using steam evaporation or solar evaporation), if such con- centration is deemed necessary and/or desirable. Our studies utilized both types of Sinclair No. 4 geothermal fluid, post-flash and concentrated. Figure 22A illustrates the behavior of lithium in post-flash Sinclair No. 4 fluid, during the course of concentration by simulated solar evaporation. Using magnesium as a tracer element to follow the course of evaporation, it appears that no lithium is lost from solution up to a concentration ratio in excess of 5. The limiting factor in degree of solution concentra- tion becomes saturation with respect to the calcium chloride content of the geothermal fluid. Further con- centration of Sinclair No. 4 fluid results in crystallization of large quantities of calcium chloride hexahydrate (CaClg-GHQO) which entrains significant quantities of the valuable, concentrated, lithium-bearing fluid. While much has been published in the literature and in the patent files regarding recovery of lithium values from solution, our studies have indicated that precipita- tion of lithium values as the aluminate complex to be the most effective. Organic solvents such as butyl and amyl alcohols, have been used to selectively dissolve away lithium chloride from mixtures of alkali metal chlorides (Lindal, 1970; Morrison and Freiser, 1962). Alcohol-ketone mixtures have been used to extract lithium from brines, but urea or ammonia additions are required to prevent calcium interference“. Lithium Corporation of America has patented a process5 where lithium is extracted as a com- plex lithium tetrachloroferrate, but this scheme requires pretreatment of the lithium-containing fluid with both hydrochloric acid and ferric chloride to establish proper operating conditions. In addition, the application of liquid or solid ion- exchange techniques to such hot saline fluids (T2110°C) V‘U.S. patent, 3,807,922, U.S. Department of Commerce, Patent and Trademark Office, Washington, D.C. sU.S. patent, 3,537,813, U.S. Department of Commerce, Patent and Trademark Office, Washington, D.C. 64 3400 — 3200 — N 0| 0 O O I SINCLAIR NO. 4 BRINE N a) O O I 2600 — N N m 4: o o o o I 2000 - X IN PARTS PER MILLIO “IBOO — / (O 5 K) E 5 o o o o o o o o | / Li CD 0 o I 03 O O _ Ba ELEMENT CONCENTRATION I I O I 2 3 LITHIUM RESOURCES AND REQUIREMENTS BY THE YEAR 2000 l I l 4 5 6 CONCENTRATION RATIO, Mg TRACER FIGURE 22A.—Behavior of lithium and other elements during concentration by evaporation. leads to problems regarding solvent degradation and losses due to evaporation of the carrier solvent, while ion-exchange resins are subject to dehydration and shrink-swell fracturing with consequent destruction of the resin beads. The preferred method for lithium recovery from Sinclair No. 4 geothermal fluid is the precipitation as a complex lithium aluminate using either freshly precipi- tated aluminum hydroxide or in-situ precipitation of aluminum hydroxide via aluminum chloride with pH control. Ratios of concentration in excess of 150 have been obtained by this method with lithium recoveries in ex- cess of 98 percent. Figure 23 presents data on lithium recovery versus mole ratio of aluminum to lithium used to effect the precipitation. Apparently the optimum Al/Li ratio for 98+ percent lithium recovery is about 2.75/1. Precipitation temperature for these results was in the 75°C—85°C range. Control of pH is critical to achieve maximum lithium recovery, with a range of 5.5 to 8.2 appearing to be optimum for Sinclair No. 4 brine as shown in figure 24. This pH range is somewhat different than that called for in the literature and is probably dependent to some degree on the composition of the fluid used. For exam- ple, R. L. Focht (1961, oral commun.) recommends a pH range of 10 to 12.5 as being optimum for maximum lithium extraction. DISSOLVED Li, IN PERCENT OF FEED RATIO: Al203/Li2 0 av WEIGHT T 1 6. . '2 0 T I I 2 MOLE RATIO Al/Li FIGURE 23.—Effect of AlzLi ratios on lithium recovery. LITHIUM RESOURCES AND REQUIREMENTS BY THE YEAR 2000 65 I00 — 90— 80— 70-— 60— 50— 4o— 30— 20— LITHIUM PRECIPITATED, IN PERCENT OF FEED pH FIGURE 24.-—Effect of pH on precipitation of lithium with aluminum hydroxide. A major difficulty associated with the aluminate method of lithium recovery is the gelatinous nature of the precipitate formed. This is probably due to the high mole ratio of aluminum to lithium used in previous work. In addition, pH of precipitation exerts a pro- found effect on the settling characteristics of the pre- cipitate as shown in figure 25. Optimum pH of precipitation for maximum settling rate appears to be about 8.3, using an Al/Li ratio of 4/1. However, a pH of 8.3 is just outside the optimum pH range for maximum lithium recovery. Using an aluminum to lithium ratio of 3/1 and a pre- cipitation pH of 7.5 resulted in both maximum lithium recovery and a reasonably rapid settling precipitate as shown in figure 26. In conclusion, it should be noted that the conditions outlined for maximum lithium recovery from Sinclair No. 4 geothermal fluid may not be optimum for other SOLIDS VOLUME IN PERCENT OF INITIAL SLURRY VOLUME o 30 so 90 I20 150 I30 TIME. IN MINUTES FIGURE 25.—Settling data for aluminum (lithium) hydroxides. I00 - 80 AIILi=32| SOLIDS VOLUME IN PERCENT OF INITIAL SLURRY VOLUME o 20 40 so so IOO I20 TIME. IN MINUTES ,— FIGURE 26.——Settling curve—aluminum (lithium) hydroxide at pH=7.5. geothermal fluids.‘ A properly planned experimental program should enable one to determine optimum lithium recovery conditions for each type of lithium- containing fluid. REFERENCES CITED Baniel, A., and Blumberg, R., 1963, Concentration of aqueous solu- tions by liquid-liquid extraction: London Indust. Chem., v. 39, p. 460. Fletcher, A. W., and Wilson, J. C., 1960—61, Naphthenic acid as a liquid-liquid extraction reagent for metals: Inst. Mining and Metallurgy Trans., v. 70, p. 355. Grinstead, R. R., and Davis,J. C., 1970, Extraction by phase separation with mixed ionic solvents: Indust. Eng. Chem., v. 9, no. 1, p. 66. Kennedy, A. M., 1961, The recovery of lithium and other minerals 66 from geothermal water at Wairakei: U.N. Proc. 1961, Art. 6/56, p. 502. Kunasz, I. A., 1975, Lithium raw materials, in Industrial minerals and, rocks (1975 revision): Am. Inst. Mining, Metall., and Petroleum Engineers, p. 791—803. Linda], B., 1970, The production of chemicals from brine and sea water using geothermal energy: U.N. Symposium on the De- LITHIUM RESOURCES AND REQUIREMENTS BY THE YEAR 2000 velopment and Utilization of Geothermal Resources, v. 2, pt. 1, p. 910. Morris, D. F. C., and Short, E. L., 1963, Extraction of lithium chloride by tri-butyl phosphate: jour. Inorganic Nuclear Chemistry, v. 25, p. 291. Morrison, G. H., and Freiser, H., 1962, Solvent extraction in analytical chemistry: New York, John Wiley and Sons, 269 p. m LITHIUM RESOURCES OF SALARS IN THE CENTRAL ANDES By GEORGE E. ERICKSEN, U.S. GEOLOGICAL SURVEY, RESTON, VA; GUILLERMO CHONG D., UNIVERSIDAD DEL NORTE, ANTOFAGASTA, CHILE; and TOMAS VILA G., INSTITUTO DE INVESTIGACIONES GEOLOGICAS, SANTIAGO, CHILE. ABSTRACT Lithium occurs in potentially commercial concentrations in salars of the central Andean region of northern Chile, western Bolivia, and northwestern Argentina, and in the nitrate deposits of northern Chile. Preliminary exploration of the salars indicates that Salar de Atacama, one of the largest in the region, contains large amounts of brine that average between about 2,000 and 4,000 mg/l lithium, so that this one salar may contain several million tonnes of recoverable lithium. Brines from several smaller salars in the Chilean Andes were found to contain as much as 200 mg/l lithium. Information about the lithium content Of brines in the many other salars of the Andean Highlands is lacking, although some, such as Salar de Uyuni in Bolivia, appear favorable for accumulation of large amounts of lithium-rich brine. Lithium is con- centrated locally in brines and saline crusts of salars in the coastal desert of northern Chile in amounts of several hundred to a few thousand parts per million, but potential resources of material averag- ing more than 100 mg/l lithium are small. Nitrate ore generally con- tains 20—50 ppm lithium, which is recoverable as a byproduct. INTRODUCTION Saline-encrusted playas, locally known as salars, and assayciated brines1 that have accumulated in closed ba- sins in the central Andean region contain huge amounts, of saline materials. Recent exploration has shown that some of the salars contain relatively high concentrations of lithium, and it seems likely that this region has poten- tially large lithium resources. Associated saline re- sources include sodium chloride, sodium sulfate, potas- sium, boron, and magnesium. Some lithium-rich brines contain small but potentially recoverable amounts of cesium and rubidium. The salars are in a region of internal drainage in northern Chile, western Bolivia, and northwestern Ar- gentina, an area of more than 1 million km2 (fig. 27). The region has more than 100 closed basins in which are found at least 75 salars ranging in size from about 1 km2 Iln this report, a brine is defined as a body of water containing more than 35,000 mg/l dissolved solids—the approximate concentration of sea water—and saline water is one con- taining 10.000-35,000 mg/l. to several thousand square kilometres. Salar Uyuni in Bolivia, the largest of the salars, has an area of about 9,000 kmz’. Saline lakes oCcur in several of the basins, and lakes formerly existed in many others that are now occupied by salars. Lake Titicaca, the largest freshwater lake in South America, is at the border between Bolivia and Peru and in the northern part of the region of interior drainage. The region has an arid climate, and precipitation is now more or less balanced by evaporation so that lake levels and ground—water levels are relatively stable, though during the last several thousand years evapora- tion has tended to exceed precipitation in the region as a whole. Precipitation is in part dependent upon altitude, ranging from less than 10 mm per year in the Atacama Desert of coastal Chile, at altitudes below 2,000 m, to a maximum of about 300 mm in the Andean Highlands of Chile and nearby parts of Bolivia and Argentina. Most of the region ofinternal drainage in the Andean Highlands is underlain by volcanic rocks of Pliocene and Quaternary age, which are considered to be a major source of salines in the region (Ericksen, 1963). Thick rhyolitic ash-flow tuffs, chiefly of Pliocene age, are pres- ent throughout the area, representing one of the greatest known accumulations of rocks of this type. Superimposed on this rhyolitic terrane are several hun- dred volcanoes, most of which are of andesitic compo- sition but a few of which are rhyolitic. Most of these volcanoes are now dead; only a few show fumerolic ac- tivity. Ash—flow tuffs from this part of the Andes have been found to contain about 500—3,500 ppm of water- soluble salines (Ericksen, 1961, 1963). Saline thermal springs are found at many places today and probably were widespread during former times of intense vol- canic activity. Several spring deposits of the borate min- eral ulexite have been found in northwestern Argen-. 26° LITHIUM RESOURCES AND REQUIREMENTS BY THE YEAR 2000 67 68° o 'l STBA.J GB .u Alemaniu o 45/? I 63 Voldivio ,' CD Siurra Gorda 28 M ”as o o 3130““? anuedono >‘ g ’32 ‘ Antofagasta 2&0: a ‘ 33 / 21) 2‘5 ‘ lo :34 35/ so 22 1‘ ¢ ’37. : - I Aqua: .1 ‘ V. Blanca: 2466 39 ‘ 4/0‘ t ‘ 64 Copiapg 51 . 52 433) 57 58’ :r’i 55 18° 20° 22° 24° 70° 68° SALARS AND LAKES l—Laguna Blanco 2—Solar de Surire 3—Solor de Coipasa 4—Salar de Obispo 5—Salar de Uyuni 6—Salor de Empexo 7—Solar del Guasco 8—Solor de Coposa 9—Solar de Pintados lO—Solor de Bella Vista I I—Salar Grande lZ-Solor de Sur Vieio l3-Solar do Lagunas l4—Solar do Llamora IS—Cerro Soledod (mounloin) lb—Salar do Ollagiie l7—Salar de San Martin lS—Salar do Chiguana 19—Solar do Ascolon 20—Solor del Carmen 21—Solor de Navidod 22—Salar Mor Muerto 23—Pampo Elvira 24—Salor de lmilac 25—Salar de Punlo Negro 26-Llano de lo Paciencia 27-Solor de Atacama 28-Solor de Toro 29—Pampa de Taro 30—Solor de Aguas Calienles 31—Solor de Puiso 32—Solor de Quisquiro 33-Salar do Aguas Calienles 34—Lagqi’o Miscanli 35-Solar de Loco 36—Solar de Talar 37—Loguna de Tuyailo 38—Solars de Aguos Calienles and Purisunchi 39—Solor do Pular 40—Salar de lncaguasi Al -Solor do Aguas Colienles 42—Laguno de la Azufrero 43—Solor de Paionoles 44—Salar de Gorbeo 45—Solor de Agua Amorga 46—Salar de lo lsla 47—Salar de Ios Parinas 48—Solar de Aguilar 49—Solor do lnfieles 50—Solor Grande 5i—Salar de Pedernales 52—Solar de Piedra Poroda 53—Lagunas Bravos 54—Loguno Wheelright 55—Laguna Escondido fi-Salar Wheelrighf 57-Loguna Verde 58—Solar de Maricunga 59—Laguna del Negro Francisco 60—Solor de Couchari 61—Solar de Arizaro 62—Solor del Hombre Mueno 63-Salar de Anfofollo 64—Salar de Pocilos 65—Solar de Pampo Blanca 66-Nnme unknown 0 50 KILOMETRES ;.1 Figure 27.——Index map of the central Andean region showing location of salars and saline lakes. (Modified from Stoertz and Ericksen, 1974). 68 tina, some of which are associated with present-day thermal springs (Muessig, 1966, p. 153—154). In contrast to the Andean Highlands, the Atacama Desert to the west is underlain chiefly by sedimentary, volcanic, and plutonic rocks of Jurassic to early Tertiary age. Except for the salines in continental sedimentary rocks ofjurassic to Tertiary age in the region of Salar de Atacama, the rocks of the Atacama Desert lack salt- bearing strata and have contributed a relatively minor portion of the widespread and abundant salines here. A considerable portion of the salines of the Atacama Des- ert is probably of oceanic origin, having been released from the ocean surface as sea spray and bubble spray and transported inland by wind and fog. GEOLOGY OF THE SALARS Three types of salars can be recognized as follows: (1) zoned salars of the Andean Highlands and front ranges, (2) dry silty salars of the Atacama Desert, and (3) Salar Grande, which is a mass of recrystallized halite. Zoned salars show lateral changes in crustal type and composi— tion (fig. 28); they generally have a relatively thick, hard, dry interior crust of halite or gypsum and a thin, soft, moist outer crust consisting of mixtures of these miner- als with mirabilite, thenardite, and ulexite. The gypsum crust commonly contains layers or abundant nodules of ulexite. The saline minerals in crusts of zoned salars crystallized chiefly from surface and near-surface saline waters and brines. Ponds and lakes, ranging from a few tens of metres to about a kilometre in diameter, now exist perenially at the margins and in interior low areas of many zoned salars, and the low areas are more exten- sively flooded during the rainy season. The silty-salar crusts in the coastal region of Chile formed by crystalli- zation of saline minerals, chiefly halite, which served as near-surface cementing material in silty playas in the capillary fringe of the ground-water table. The resulting crust generally is a few tens of centimetres to a metre in thickness and consists of a somewhat loosely cemented mass of nodular to platy fragments of silty salt. The fragmental character of the crust is due to the combined processes of desiccation, heaving, and breaking of old crust as new crust forms beneath it and to leaching by infrequent rains. Salar Grande, near the coast in north- ern Chile (fig. 27), is unique. It is a huge deposit of coarsely crystalline pure halite nearly 50 km long, as much as 8 km wide, and as shown by drill cores, it has a maximum thickness of at least 160 m. The salt is dry, as is the alluvium beneath it, except for abundant fluid inclusions; former brines evidently drained away along recent, faults that cut the salar. As the result of faulting and tilting during their for- mation, many of the Andean salars are crowded toward one side of their basin and show asymmetrical zoning LITHIUM RESOURCES AND REQUIREMENTS BY THE YEAR 2000 (fig. 29). Salar surfaces also may be gently inclined. Many of the basins were tilted towards the northwest, as is well shown by three of the four salars in figure 29. Others were tilted towards the west or southwest; one example is Salar del Guasco, also shown in figure 29. In some cases old salar crusts may have been covered with marginal alluvium deposited during and after tilting. Lacustrine and playa sediments beneath the salar crusts are known only from shallow pits and short drill holes. Layers of saline minerals have not been encoun- tered within these sediments and probably do not exist beneath most of the salar crusts, which may have formed as single-stage drying ofa saline lake or by capil- lary evaporation of near-surface ground water. On the other hand, some of the large salars such as Atacama, which have had a more complex history, may have addi- tional saline layers at depth. Such layers might contain significant quantities of “fossil” lithium-rich brines. The lacustrine sediments of the salars penetrated by drill holes consist of poorly sorted to well-sorted sand, silt, and clay with varying amounts of calcium carbonate-rich mud and sporadic diatom— and ostracod-rich layers. Several drill holes in marginal areas of Salar de Atacama show abundant calcium carbonate mud and ostracodes to depths as great as 43 m (Moraga and others, 1974, p. 47—51), and thin, relatively pure diatomite layers have been found interbedded with elastic sediments at several salars. CHEMICAL COMPOSITION Chemical analyses of saline waters of northern Chile show them to be chloride waters (Clark, 1924, p. 202) in which sodium is the dominant cation, as indicated in table 24. The similarity of salar crusts elsewhere in the central Andes suggests that essentially all of them formed from chloride waters. Concentrations of saline components range from several tens of g/l in saline ground water or brine within gypseous salar crusts, to nearly 400 g/l in brine within halite-rich salar crusts. The saline components in brine tend to show relative concentrations as follows: Na+>K+>Mg++ >Ca++ >pLi+ Cl->SO4=>B03= > HCOS‘. Spring waters, which generally contain less than 10 g/l dissolved solids, also are chloride waters but have vari- able relative amounts of K, Ca, and Mg; HCOS— is much more abundant than BOf‘ in spring waters and may occur in amounts equal to or greater than SO4=. As indicated by the analyses in table 24 and by other analyses of saline waters and brines associated with sa- lars in northern Chile, lithium and potassium both tend to show an increase in concentration as total salinity increases. Potassium is generally 10—15 times more LITHIUM RESOURCES AND REQUIREMENTS BY THE YEAR 2000 69 89° 07‘ Solar dc Ayuas Colin/es 0 , 2KILOMETRES LLI Solar do Pu/ar L lano oh la Paciencia / Solar: dd , Fur/synchi EXPLANATION 32 m 3:193:20025, wet to dry, Lakes or areas frequently . flooded .. o 2 KM Oulsqu/ro l_l._l EKM |8,606 Mountaln peak showing A elevation, in feet Pa/onales Solar or old Iakebed , (I000 Feet = 305 Metres) --------- . I . ' undlfferantiated; mosture ‘ shown In detail on inset maps Area of Iakebeds and shore- -— DIVIdOS 0‘ closed WW“ — lines of former lakes 25 Nuihber of solar and basin keyed to figure 27 25° 0 5 IO KILOMETRES Hard salIne crusts of salars L_I_J i FIGURE 28.—Salars and basins of interior drainage in the eastern central part of Antofagasta Province, Chile. (Modified from Stoertz and Ericksen, 1974). abundant than lithium, but this relationship is not have been tested for lithium. Figures 30 and 31 Show the necessarily linear, nor does it apply to all brines that relation between concentrations of lithium and potas- 70 LITHIUM RESOURCES AND REQUIREMENTS BY THE YEAR 2000 W: P—26'15 I ‘Salar de Pedernales \” m: Salav de Punta Negro 68‘45’ IO KILOMETRES / 5.1., d.l Guasco IO KILOMETRES . \5 ‘U’ 0 5 l0 KILOMETRES l_l__l EXPLANATION Drainage basin border Slightly elevated marginal zone of solar, characterized by shallow stream channels Hfi'd °V°P°fil° "V3“: predominantly and shovelines of former lakes; generally massive to silty rock salt (chloride contains abundant gypsum zone) § Soft, moist saline soil, subiect to Former salar or lakebed, apparently above &\ flooding; chiefly sulfate zone, con- the level of present flooding or ground-water taining gypsum and\_some halite evaporation; includes deltas, shorelines, and beaches of former lakes - Lakes or areas frequently flooded FIGURE 29.—Asymmetrical zoning of salars in basins ofinterior drainage in northern Chile, attributed in part to faulting and tilting of the basins during late Pleistocene and Holocene time. (Modified from Stoertz and Ericksen, 1974). LITHIUM RESOURCES AND REQUIREMENTS BY THE YEAR 2000 71 TABLE 24.—Chemical concentrations, in milligrams per liter, in selected lithium-rich brinesfrom northern Chile Dissolved Locality solids‘ Na K Mg Ca Li Cl SO, HCO, B Salar de Atacama2 ____________________ 370,000 91,100 23,600 9,650 450 1,570 189,500 15,900 230 440 Do __ 310,000 85,800 13,000 6,350 1,100 940 163,900 8,540 280 360 Do _- 190,000 45,100 9,000 5,330 900 520 83,780 18,170 240 360 Do __ 73,000 18,220 4,220 1,810 360 290 36,750 3,430 320 100 Do __ 62,000 14,840 2,900 1,930 1,080 190 27,500 7,900 100 88 Do _______ 40,100 10,280 1,690 750 1,160 130 20,300 2,160 92 61 Salar de Ascoton . 47,022 13,870 1,670 827 1,195 82 24,000 4,693 0 595 Salar de Pujsa5 ___ ____ 89,298 28,500 1,295 653 375 137 27,660 28,110 0 675 Salar de Aguasa Calientes 1 81.436 25,460 1,183 1,361 2,538 152 46,690 3,154 0 474 Salar de San Martin‘ ______ 4 102,138 28,160 2,614 6,252 1,566 187 60,050 2,490 625 426 Salar de Bellavista‘ Pimados 170,300 50,000 5,403 3,665 5,935 85 100,600 2,720 178 225 Salar de Lagunas‘ _____________________ 390,000 126,800 14,280 3,630 1 10 412 176,600 47,770 406 979 ‘By evaporation. “From Moraga and others (1974, table 4). 317mm Vila (1975, table 6). ‘Institulo de Investigaciones Geologicas (unpubi data, 1975). 1.400 '_ ‘ o O O 0 1,200 — o O O . 0 O 0 I .000 " O m . ' o t: O ._ O :1 \ O U) 800 — o E: o n: 0 ° (9 :1- 0 o O .1 ° ' . o z 600 — Z — O O O .. O O ._I O O 400 — .0 0 ° I 0 ~ . O Q ' 200 — o . o O o 1 1 1 41 0 5,000 10,000 15,000 20,000 K, W MlLLlGRAMS/LITRE FIGURE 30.—Lithium-potassium ratios (dots) in saline ground water and brine, Salar de Atacama. Average LizK ratio as indicated by line is 1:15. Analytical data from Moraga and others (1974, table 4). 72 LITHIUM RESOURCES AND REQUIREMENTS BY THE YEAR 2000 L400 " o o o 0 1,200 - o o o o . 0 Low — ' - o m ' ' o E ° ' j o a 800 ~ ’ - 3 o a: o 0. g 0 0'0 .1 ' , 0 g 600 - z ’ '. - . '3 ,' ..° . 400 — .. . . o .0 . . ' . . o .0. a. :0 o 200 - . .0 o .. . 0 o o - a. o l I l l O IO0,000 200.000 300.000 _ 400,000 DISSOLVED SOLIDS. IN MlLLlGRAMS/LITRE FIGURE 3l.—Lithium-salinity ratios (dots) in saline ground water and brine, Salar de Atacama. Average Lizdissolved solids ratio as indicated by line is approximately 1:300. Analytical data from Moraga and others (1974, table 4). sium and between lithium and total salinity for saline waters and brines from Salar de Atacama. DISTRIBUTION OF LITHIUM IN SALINE DEPOSITS OF THE CENTRAL ANDES Lithium is concentrated in brine in salars and saline lakes, in salar crusts, and in the nitrate deposits of the coastal region of northern Chile. Brines associated with salars contain by far the greatest amounts oflithium and offer the greatest potential for future production. Some salar crusts may prove to be exploitable, but the amount of recoverable lithium is relatively small. The nitrate deposits contain lithium that is recoverable as a by- product. The few saline lakes tested for lithium show concentrations of 20—40 mg/l (Vila, 1975). Other high- salinity lakes, which have not yet been tested for lithium, may have still greater concentrations. Of the more than 75 salars in the central Andean region, only four—Atacama, Lagunas, Bellavista, and Pintados—have been-extensively tested for lithium and potash. More than 200 analyses of saline water and brine from Salar de Atacama have been published (Moraga and others, 1974), and it is evident that this salar contains potentially large lithium resources. Most of these analyses are from shallow pits along the south- ern and eastern margin of the salar, and relatively few are from the area within the salar that is covered by a thick halite crust. Nine drill holes as much as 43 m deep also were sampled. Moraga and others (1974) show that amounts of both lithium and potassium in brine increase inwards from the salar margin. They reported that the LITHIUM RESOURCES AND REQUIREMENTS BY THE YEAR 2000 lithium content of brine in a marginal zone (referred to as the zone of efflorescence) averages between 200 and 300 mg/l, whereas an intermediate or transition zone has BOO—1,600 mg/l and the marginal interior halite zone has 1,5 10—6,4OO mg/l. Guillermo Chong, co-author of this report, found that brines sampled along a 20-km line within the halite zone in the southern part of the salar averaged 4,100 mg/l lithium and 12,000 mg/l potassium. Analyses of saline water and brine from another 13 salars in the Andean Highlands of Chile show lithium contents ranging from 3 to 200 mg/l, and potassium contents ranging from 154 to 3,000 mg/l (Vila, 1975, G. E. Erickson, unpub. data, 1975). Total dissolved sol- ids range from 17,000 to 400,000 mg/l. Correlations be- tween lithium, potassium, and salinity in these salars are Only fair in comparison to those of Salar de Atacama. Brines containing more than 100 mg/l lithium were found in the salars of Aguas Calientes I, Laco, Pujsa, and San Martin (fig. 27). Brines of Salar de Pedernales, which are known to be high in potassium, may contain more than 100 mg/l lithium. Because of the lack of chemical analyses from salars in the Andean Highlands of Bolivia and Argentina, it is not possible to estimate their lithium potential. Nevertheless, it can be suggested that the Salars of Coipasa and Uyuni in Bolivia are among the most favor- able for accumulation of large quantities of lithium-rich brine.2 These salars consist chiefly of thick halite crusts that are subject to partial flooding during the annual rainy season. Together with the nearby salars of Empexa and Chiguana (Fig. 27) and Lago Poopo, a large saline lake northeasteast of Salar de Coipasa, they are remnants of the former Lago Minchin, an upper Pleistocene lake that extended over an area of about 40,000 km2 of the present-day Bolivian Altiplano (Ahlfeld and Branisa, 1960, p. 158—163). Salars Coipasa and Uyuni contain most of the saline materials that ac- cumulated in Lago Minchin and in an earlier Pleistocene lake that existed in the same area. Chemical compositions of brines and crusts of salars in the Atacama Desert of coastal Chile were reported by Ericksen (1963) and are recorded in unpublished analyses of the Coporacion de Fomento de la Produc- cién of Chile (CORFO), the Instituto de Investigaciones Geologicas, and the former nitrate company Cia, Salitr- era Tarapaca y Antofagasta. The crust and brine of Salar de Lagunas are unique in that they contain rela- tively high concentrations of nitrate, perchlorate, and iodate, as well as abundant lithium and potassium. Lithium content of the brine and crust of this salar is as much as 500 ppm. The presence of unusual saline com- 2The firstrknown lithium analyses of brine from Saler de Uyuni were made by Shirley L. Rettig, U.S. Geological Survey, on two samples collected in April 1976 by William D. Carter, U.S. Geological Survey, and Paul Ballon. Servicio Geolég‘ico de Bolivia, at the request of the senior author of this report. These analyses show 490 and 1,510 ppm lithium, respectively. 73 ponents in Salar de Lagunas indicates that the brines were enriched, at least in part, by rainwater leaching of nearby nitrate deposits. Local concentrations also may be due to contamination by leach solutions from waste dumps at three nitrate beneficiation plants that once operated here. Salar de Bellavista and Salar de Pintados at the west- ern side of the Central Valley in northern Chile have silty saline crusts relatively rich in potassium and perhaps also in lithium. Ericksen (1963, p. 128—132) re- ported that 1,700,000 tonnes of salar crust averaging 2.87 percent K was mined from the southern part of Salar de Bellavista. A pit was opened in this area to test the rate of regeneration of new potassium-bearing salt crust, and two samples of new cr‘ust which formed in the pit were found to contain 350 ppm and 3,100 ppm lithium, and 3 percent and 10 percent potassium. Brine from this same pit was found to contain 2,130 mg/l potassium, but lithium was not determined. The nitrate deposits of Chile contain anomalous amounts of lithium in the water-soluble salines from which the nitrate is recovered. The amount of lithium present in this form is generally in the range of 20—50 ppm but locally is as much as 250 ppm, as indicated by about 50 analyses we performed on nitrate ore samples. This lithium becomes concentrated in the recycled brines utilized to extract the nitrate from the ore. The former owner of the two largest nitrate beneficiation plants, the Anglo-Lautaro Nitrate Corp., determined that the lithium could be recovered by a solar evapora- tion process integrated with the plants. The calculated annual yield was estimated to be 1,610 tonnes oflithium sulfate (209 tonnes of lithium) from 17,000,000 tonnes of nitrate ore, the annual capacity of the two plants. This is based on a concentration of about 60 ppm lithium in the water soluble portion of the nitrate ore, which is about a third of the ore. SOURCES AND FACTORS FAVORING LITHIUM CONCENTRATION IN ANDEAN SALARS Information about chemical composition of salars and asSociated saline waters is too scanty to allow a precise evaluation of the sources of lithium and ofthe causes for its occurrence in relatively high concentrations in cer- tain salars and not in others. However, a volcanic source of lithium is suggested, because Pliocene and Quater- nary volcanic rocks and their erosion products cover large areas in the Andean Highlands where salars are most numerous and are virtually the only rocks exposed in many of the closed basins. Furthermore, it seems likely that thermal spring waters associated with these volcanic rocks were major sources of the lithium. White, Thompson, and Fournier (this report) point out that high-temperature geothermal systems are generally en- riched in lithium and that moderately saline waters 74 above 200°C associated with silicic volcanic rocks contain 1—50 ppm Li. They conclude that these high- temperature waters selectively leach lithium from the volcanic rocks. Ellis (1975) reported that high- temperature thermal water from the El Tatio geyser field, which is in a drainage basin to the north of the Salar de Atacama basin, contains 47.5 ppm lithium. Conditions that would favor concentration of lithium in the salars are as follows: (1) the relatively old age of basins—some may have been in existence since middle or even early Pleistocene; (2) the large basin/salar ratio, which might contribute to the high degree of concentra- tion of saline waters; (3) the high degree of basin closure so that water loss by either surface or underground flow was minimal; and (4) abundant saline-rich rhyolite ash- flow tuff subject to leaching and widespread fumarolic activity and saline thermal springs associated with vol- canic activity. Salar de Atacama, which has the greatest lithium concentrations yet known in the region, has three of these attributes. It has a relatively small basin/ salar ratio as is indicated in figure 27. On the other hand, this basin is unique in havingjurassic, Cretaceous, and Tertiary saline continental sedimentary rocks (Dingman, 1967) that furnished a considerable amount of saline material to the salar. These rocks may have been a major source of the lithium. POTENTIAL LITHIUM RESOURCES Salars in the central Andean region have large poten- tial resources of lithium chiefly in brines and to a lesser degree in saline crusts. Of the few salars that have been tested for lithium—all in northern Chile—Salar de Atacama appears to have the greatest quantity of lithium—rich brine. Moraga and others (1974) estimate that the brine in a 400-km2 area of Salar de Atacama contains 40,000 tonnes of lithium per metre depth. (Their calculation was based on an average lithium con- tent of 2,000 mg/l and a brine recovery of 100 1/m3; this gave an estimate of 80,000 tonnes, which the authors reduced by half to be conservative.) This estimate was made for an area within the central part of the salar where several drill holes show the halite crust ranges from 50 to 335 m in thickness and where brines sampled LITHIUM RESOURCES AND REQUIREMENTS BY THE YEAR 2000 along a 20-km line (see paragraph 2, “Distribution of Lithium in Saline Deposits of the Central Andes”) were found to average 4,100 mg/l lithium. Consequently, total lithium resources may be very large, perhaps amounting to several million tonnes. The potential lithium resources in brines of other salars in northern Chile would be far less. Lithium contents generally are not greater than 200 ppm, and the amounts of brines are such that a comparable-sized area in most salars would yield less than 1,000 tonnes of lithium per metre depth. The lithium potential of salars in the Andean High- lands of Bolivia and Argentina cannot be estimated be- cause of the lack of lithium analyses. However, Salars Coipasa and Uyuni in Bolivia appear to be most favora- ble for accumulation of large amounts of lithium—rich brine. REFERENCES CITED Ahlfeld, Federico, and Branisa, Leonardo, 1960, Geolog’ia de Bolivia: La Paz, Inst. Boliviano Petroleo, 245 p. Ahlfeld, Federico, and Schneider-Scherbina, Alejandro, 1964, Los yacimientos minerales y de hidrocarburos de Bolivia: Dept. Nac. de Geologia B01. 5, 388 p. Clark. F. W., 1924, The data of geochemistry [5th ed.]: U.S. Geol. Survey Bull. 770, 841 p. Dingman, R. j., 1967, Geology and ground-water resources of the northern part of the Salar de Atacama: U.S. Geol. Survey Bull. 1219, 49 p. Ellis, A. j., 1975, Geothermal systems and power development: Am. Scientist, v. 63, no. 5, p. 510—521. Ericksen, G. E., 1961, Rhyolite tuff, a source of the salts of northern Chile, in Geol. Survey research 1961: US. Geol. Survey Prof. Paper 424, p. C224—C225. 1963, Geology of the salt deposits and the salt industry of northern Chile: U.S. Geol. Survey open-file rept., 164 p. Moraga, B. A., Chong D., Guillermo, Fortt Z., M. A., and Henriquez A., H., 1974, Estudio geologico del Salar de Atacama, Provincia de Antofagasta: Instituto de Investigaciones Geologicas Chile, Bull. 29, 56 p. Muessig, Siegfried, 1966, Recent South American borate deposits, in Rau, J. L., ed., Second symposium on salt, Volume 1: Northern Ohio Geol. Soc., p. 151—159. Stoertz, G. E., and Ericksen, G. E., 1974, Geology of salars in northern Chile: U.S. Geol. Survey Prof. Paper 811, 65 p. Vila G., Tomas, 1975, Geologia de los depositos salinos Andinos, Pro- vincia de Antofagasta, Chile: Revista Geological de Chile no. 2, p. 41—55. m LITHIUM RESOURCES AND REQUIREMENTS BY THE YEAR 2000 75 LITHIUM RESOURCES OF UTAH1 By]. A. WHELAN, UNIVERSITY OF UTAH, SALT LAKE CITY, UT and CAROL A. PETERSEN, UTAH GEOLOGICAL and MINERAL SURVEY, SALT LAKE CITY, UT ABSTRACT Known lithium resources of Utah consist of Great Salt Lake brines, subsurface brines of the Great Salt Lake Desert and Sevier Lake, and Spor Mountain beryllium ores. The total metric tonnage of lithium in Great Salt Lake is 526,000, of which perhaps half is recoverable. Extraction of lithium from Great Salt Lake will be as a byproduct of magnesium-metal or salt recovery. The Bonneville-Salt-Flats part of the Great Salt Lake Desert has a recoverable resource of from 1,300 to 2,000 tonnes. This amount could be increased from 1,700 to 2,700 tonnes by recharging a shallow brine aquifer now being utilized to produce potassium and mag- nesium salts with brines from existing deep wells. Sevier Lake contains a shallow subsurface brine containing 450 to 540 tonnes of recoverable lithium. The beryllium Ores contain 1,400 to 18,000 tonnes of lithium. The average lithium content of brines from 58 oil wells in Utah is 7.1 mg per liter. Brines from three oil wells contain over 39 mg/l lithium. ' The mean lithium content of 35 hot springs in Utah was 2.33 ppm. The total lithium resources of Utah are between 200,000 and 240,000 tonnes. No lithium is produced in Utah at present. Plants to recover lithium as a byproduct from existing brine-processing or beryllium-extraction plants could recover from 160 to 600 tonnes of lithium per year. INTRODUCTION Only nonpegmatite sources need to be considered in evaluating the lithium resources of Utah, because neither spodumene nor other lithium minerals are re- ported from the few pegmatites of Utah (Bullock, 1967). Evaluating nonpegmatite lithium resources is a new en- deavor for most geologists, and some of the various types of occurrences are discussed by Vine (1975). Lithium in the following occurrences will be considered in this paper: Great Salt Lake brines, subsurface brines, volcanic rocks, and miscellaneous lithium-bearing min- eral localities. Location Of the occurrences are shown in figure 32. GREAT SALT LAKE BRINES The brines of Great Salt Lake constitute the largest lithium resource of Utah. The lake contains three dis— tinct brine types: (1) a nearly saturated brine (320 g/l total dissolved solids) in the area north of a semiperme- able railroad causeway, (2) a relatively fresh (117 g/l) brine in the south part of the lake at depths less than about 7 m, and (3) a brine of intermediate concentration (222 g/l) in the south part at depths greater than 7 m. The three brine types have formed because Great Salt Lake is cut in two by an east-west-trending railroad ‘Permission to publish granted by Director, Utah Geological and Mineral Survey. causeway built of gravel armored with rip-rap. Over 90 percent of the surface inflow enters the lake south of the causeway. Flow through the causeway into the north arm is restricted, and high evaporation there forms a concentrated brine. The shallow south-arm brine is be- coming fresher each year. The deep south-arm brine represents return flow from the north arm. The north— arm brine contains 64 ppm lithium, the shallow south- arm brine contains 22 ppm lithium, and the deep south-arm brine contains 43 ppm lithium. This amounts to tonnages of 214,000 tonnes, 203,000 tonnes, and 109,000 tonnes of lithium, respectively, in each of the brines; the total tonnage is 526,000 tonnes. Because lithium follows magnesium during process- ing of the brines, and because of the low lithium con- tent of the brines, economic recovery of lithium from ”4° ° a 42° "3 “2 GREAT SALT LAKE 4v Great Salt "'° “I°° . Luke Desert Clay Basin /‘7 Field JBonneville Sol! Flat Red Wash Field . 40° Tinfic Mining _ .Spor Mountain 0'5"”, ‘p .Honeycomb Hills °C‘ ms 39° F» _ S /Sew'er Lake U .Rooseveli KGRA Seven Mile Canyon 0 .Marysvole 39° _. Rabbi] Ears 37" l | I Fieldl . 0 50 I00 MILES o 50 I00 KILOMETRES FIGURE 32.—Locations of lithium resources and mineral occur- rences in Utah. 76 Great Salt Lake depends on a viable magnesium indus- try. NL Industries has a magnesium plant at Rowley, on the west side of the lake, that uses shallow south-arm brines. Design capacity is 41,000 tonnes of magnesium per year; but, to date, the plant has had start-up difficul- ties and has only produced at a maximum rate of about 14,000 tonnes per year. About 5 kg of lithium can be produced per tonne of magnesium, so the NL plant has the potential to produce 70—200 tonnes of lithium per year. Great Salt Lake Mineral and Chemical Corporation, located west of Ogden uses north-arm brines. It has a design capacity of 220,000 tonnes of potassium sulfate annually. Although production figures are not available, the plant has been operating successfully. When the Great Salt Lake Mineral and Chemical Corporation project was started, Dow Chemical Company contracted to buy 90,000 tonnes per year of bischofite (MgC12'6H20). In 1970, with the magnesium plant nearly complete, Dow cancelled the contract. The par- tially completed plant is now in standby status. Ninety thousand tonnes per year of bischofite would represent about 8,020 tonnes of magnesium and a potential lithium production of 41 tonnes. Thus, if all the magnesium-production facilities on Great Salt Lake should become operative, the lake could provide feed for lithium plants capable of producing 241 tonnes oflithium per year. Of the lithium resources in Utah, Great Salt Lake brines probably could be most easily developed. SUBSURFACE BRINES Subsurface brines considered to be possible sources of lithium in Utah include (1) subsurface brines of Bon— neville Salt Flats in the Great Salt Lake Desert, (2) sub- surface brines of Sevier Lake, (3) oil—field brines, and (4) brine that may be produced incidental to geothermal development. SUBSURFACE BRINES OF THE BONNEVILLE SALT FLATS The high-potash, shallow subsurface brines in the western part of the Great Salt Lake Desert were first discussed by Nolan (1927). Later studies include those of Turk (1973) and Lindenberg (1974). Turk (1973) discussed the hydrology of the Bon- neville Salt Flats, the part of the desert now being exploited for potassium and magnesium salts by Kaiser Chemicals. He described a shallow brine aquifer, which provides brine for potassium and magnesium chloride production, and a deep brine aquifer at depths greater than about 325 metres. Lindenberg (1974) provided additional geochemical data on the shallow brines, which he sampled over a considerably larger area than that described by Turk (1973). Typical analyses of the LITHIUM RESOURCES AND REQUIREMENTS BY THE YEAR 2000 shallow brines are given in table 25. Turk (1973, p. 17) predicted that the current produc- tion rate from the shallow brine aquifer can be main- tained for 25—40 years before the brine quality deterior- ates from slightly less than 1 percent KCl to 0.5 percent KCl. Assuming the same proportional decline, lithium would decrease from about 0.03 g/l to 0.015 g/l. If the decrease were linear, the average lithium content dur- ing this time period would be 0.0225 g/l. The assump- tion of linear decline in brine quality is not valid, as Turk (1973, p. 17) noted, but was made by us to simplify calculations. Turk assumed a discharge of 700 acre-feet (8.6X105 m3) per year. This represents some 19 tonnes of lithium per year. On the. basis of a field life of 25—40 years, a potential production of 475—760 tonnes of lithium would be reasonable. Lindenberg’s study (1974) increased the known area of shallow brine by 2.8 times, so the estimated potential production of lithium should be increased to between 1,300 and 2,000 tonnes. The deep brine aquifer is penetrated by 13 wells. Typ- ical analyses of the brine from two wells are given in table 26. Turk noted that by using the deep brine for TABLE 25.—Analyses of shallow aquifer brines, Great Salt Lake Desert, in grams per liter [N.D., not determined] Nolan Turk Lindenberg lon (1927, p. 39) (1973)1 (1974)2 Overalls . 139.63 122.37 119.38 . .038 N.D. N.D. . \‘uD. N.D. N.D. . 4.24 5.75 5.69 . .00 N.D. N.D. . .023 N.D. N.D. . 76.7 74.71 69.57 .. 4.43 3.49 3.62 . .003 .003 .003 .' L25 1.39 1.38 . .057 N.D. N.D. 1.9 1.73 1.84 1.83 N.D. .0087 N.D. N.D. N.D. .041 N.D. N.D. N.D. .0018 N.D. N.D. N.D. .0013 N.D. N.D. N.D. .0048 N.D. N.D. Total dissolv8d solids ...... 164.7 228.28 209.55 N.D. @180°C 5 N D S cilic ravit ,,,,,,,,,,,, 1.11 41.150 ‘1.140 '. . PC g y @ 25°C 61.145 IH.136 pH ________________________ N.D. 7.1 N.D. N.D. @ 25°C not detected: Fe, Mn, As, P04 1Average of 15 analyses given in Turk (1973, p. 13). 1Average of65‘analyses given in Lindenberg (1974, p. 22—28). aArithmetic mean of first three columns. ‘Extrapolated from table, specific gravity versus total dissolved solids, given in Levorsen (1958, p. 663). 5Interpolated from table, specific gravity versus total dissolved solids, given in Levorsen (1958, p. 663). 6From graph. total dissolved solids versus specific gravity. given in Whelan (1973, p. 19). TABLE 26.—Chemical analyses of brine from deep brine aquifer in parts per million [Analyses by Kaiser Chemicals, San Leandro, California; from Turk, 1973, p. 9] Well number Ca Mg Na Li K S‘O4 Cl DBW8 ............ 1,600 1,400 41,400 16 1,800 6,000 70,000 DBWIS ____________ 1,500 1,400 46,000 17 2,000 6,200 72,800 LITHIUM RESOURCES AND REQUIREMENTS BY THE YEAR 2000 recharge, the life of the project could be extended some 30—40 percent, indicating a total potential lithium prod- uction from the Great Salt Lake Desert of 1,700—2,700 tonnes. The deep brines are also a possible geothermal resource (Whelan and Petersen, 1975). SEVIER LAKE Lithium-bearing brines are found at shallow depths in the sediments of Sevier Lake, a playa lake in southwest— ern Utah (Whelan, 1969). The mean of 10 brine analyses, all samples from sites in the southern half of the playa, is given in table 27. Assuming that the hydrologic characteristics of Sevier Lake sediments are similar to those of the Great Salt Lake Desert sediments and calculating the possible production for relative sampled areas and brine con— centrations, one would obtain a potential of about 450—— 540 tonnes lithium. Because all of the sampled points are in the south half of the lake, the resource may be double the value given. At present, there is no develop— ment of saline industries on Sevier Lake. OILFIELD BRIN ES Another possible source of lithium in Utah is oilfield brines. The mean of 58 available lithium analyses of Utah oilfield brines is 7.1 mg/l lithium, with a standard deviation of 14.1 mg/l. Oil production in Utah during 1974 was about 39,400,000 bbl, with a water production of about 107,500,000 bbl. Thus, some 17 tonnes of lithium are brought to the surface each year in oilfield brines. Because of the low lithium content of the brines and the fact that brines are produced in varying quan- tities from scattered fields. Production of lithium from most oilfield brines in the near future is doubtful. Certain oilfield brines, however, are of interest and should possibly be investigated further. A sodium chloride—type brine from a well (SEIANWV1 sec. 22, T. 3 N., R. 24 E., Salt Lake Basin and Meridian) in the Clay Basin field contained 85 mg/l lithium and 50,200 mg/l total dissolved solids. This field is reported to have no water production. TABLE 27.—Mean analysis of 10 shallow sabsurface brines, southern half of Sevier Lake, Utah Value in Ion grams per liter .035 Total dissolved solids (g/l) ______________________ 185.14 Specific gravity ________________________________ 1.126 77 A well (SEIANEl/i sec. 19, T. 43 S., R. 22 E.) in the Rabbit Ears field produced a sodium chloride-type brine containing 72,830 mg/l total dissolved solids and 58 mg/l lithium. This is a one-well field which, in 1974, pro- duced 3,413 bbl of oil and 60,390 bbl of water. One deep well (SEMISEI/i sec. 21, T. 7 S., R. 24 E.) in the Red Wash field of Uintah County produced a sodium-calcium chloride brine containing 144,104 mg/l total dissolved solids and 39 mg/l lithium. This was a dry hole, but it indicates that potential for lithium-bearing brines may exist at depth in this field. GEOTHERMAL BRINES Brine produced from hot-water geothermal systems has been considered as a potential lithium resource. Thirty-five analyses of hot-spring water in Utah (Mun- dorff, 1970, p. 12—19) give a mean lithium content of 2.33 ppm lithium. The highest reported lithium value for a Utah hot spring is 2.7 ppm from Roosevelt Hot Springs, near Milford, Utah (White and others, 1963). White (oral commun., 1976) noted that this same lithium analysis was incorrectly reported as 0.27 ppm by Mundorff (1970, p. 17). Most of the hot-spring analyses indicate little potential for lithium production as a byproduct of future geo- thermal development. However, using an estimated po- tential energy production of 33 MWe—centuries from the Roosevelt Known Geothermal Resource Area (Nathenson and Muffler, 1975, p. 110) and assuming that all the water produced contains 27 ppm lithium, about 40,000 tonnes of lithium could be brought to the surface in this area. The possibility that lithium-bearing clays exist in the hydrothermally altered rocks associated with geother- mal resources should not be overlooked. VOLCANIC ROCKS Mineralized rhyolitic crystal-lithic tuffs are mined for beryllium in the Spor Mountain district northwest of Delta, Utah. The ore-grade tuffs contain anomalous amounts of lithium. Lindsey, Ganow, and Mountjoy (1973) discussed the area and reported 75 ppm lithium in unmineralized vitric tuff and 55 ppm lithium in un— mineralized zeolitic tuffs. They described hydrother- mally altered argillic tuffs containing 105 ppm beryl- lium and 230 ppm lithium. Hydrothermally altered feldspathic tuffs contained an average of 14 ppm beryl- lium and 315 ppm lithium. j. A. Whelan (written com- mun., 1962) reported that a sample from the Hogsback ore body contained 1.25 percent beryllium oxide and 0.10 percent Li20 (420 ppm lithium). A section of the Rainbow open pit that averaged 0.19 percent BeO also averaged 4,231 ppm lithium (Park, 1968, p. 93—94). Williams (1963, p. 58) reported 3,175,000 tonnes of 78 blocked-out beryllium ore, with an anticipated tonnage greatly in excess of 4,500,000 tonnes. Thus, the lithium resource of ores of the Spor Mountain beryllium district is probably between 1,400 and 18,000 tonnes oflithium. At present, about 90,000 tonnes of beryllium ore are processed yearly, carrying between 35 and 350 tonnes of lithium. Similar rhyolitic tuffaceous beryllium-bearing rocks occur in the Honeycomb Hills, west of the Spor Moun- tain beryllium deposits. 'An average beryllium value of three analyses (Park, 1968, p. 72) from this area is 0.24 percent beryllium oxide. Unfortunately, the tuffs were not analyzed for lithium. Bullock (1967, p. 167) re- ported spodumene from these tuffs. The beryllium occurrences of Utah and other silicic volcanic rocks should be prospected for lithium. Other beryllium occurrences besides Spor Mountain and the Honeycomb Hills that might contain anamalous amounts oflithium are found in the Deep Creek Moun- tains (pegmatitic), Granite Peak (pegmatitic), Sheeprock Mountains (beryl disseminated in granite), and Mineral Range (skarns). MISCELLANEOUS LITHIUM MINERAL OCCURRENCES IN UTAH Lithiophorite, a wad containing about 1 percent L120, has been reported by Bullock (1967, p. 137) from the following localities: Seven Mile Canyon and Shinarump Nos. 1—3 claims, Grand County; Marysvale district, Buddy, Bullion Monarch, Freedom No. 1, and Prospec- tor No. 1 and 3 mines, Piute County. He also reported (p. 163) that the lithium iron manganese phosphate, sicklerite, occurs in the Tintic district, Juab County. Dyni (1973) reported a trioctahedral smectite from the Roan Cliffs, Uinta Basin, that might be similar to the lithium-bearing montmorillonite, hectorite. These reported occurrences of lithium minerals have yet to be evaluated. CONCLUSIONS The known lithium resources of Utah include Great Salt Lake brines, subsurface brines of the Great Salt Lake Desert and Sevier (dry) Lake, and the beryllium ores of Spor Mountain. Probable recoverable tonnages of the resources are given in table 28. Production could be started with minimum delay by adding lithium- recovery plants to the brine-processing plants on Great Salt Lake and the Great Salt Lake Desert and to the beryllium mill of Brush-Wellman, near Delta, Utah. Probable annual outputs from additions to these plants are given in table 28. LITHIUM RESOURCES AND REQUIREMENTS BY THE YEAR 2000 TABLE 28.—Estimated recoverable lithium resources of Utah and estimated annual recovery possible by adding lithium-recovery plants to existing facilities [Leaders ( ______ ) indicate no existing facilities] Probable annual production obtainable using additions to existing plants (tonnes) Probable recoverable resource Source (tonnes) Great Salt Lake brines ______________________ 260,000 110—240 Great Salt Lake Desert brines ____________ LOGO—3,000 10 Sevier Lake brines... 500 ________ Geothermal brines..- 0—40,000 ________ Beryllium ores _____ _ LOCO—20,000 35—350 Approximate total ________________ 262,500—283,500 158—600 REFERENCES CITED Bullock, K. C., 1967, Minerals of Utah: Utah Geol. and Mineral. Sur- vey Bull. 76, 237 p. Dyni, J. R., 1973, Trioctahedral smectite in the Green River Forma- tion, Duchesne County, Utah: US. Geol. Survey open-file rept., 25 p. Levorsen, A. I., 1958, Geology of petroleum: San Francisco, W. H. Freeman and Company, 703 p. Lindenberg, G. J., 1974, Factors contributing to the variance in the brines of the Great Salt Lake Desert and the Great Salt Lake: Utah Univ., M.S. thesis, 70 p. Lindsey, D. A., Ganow, H., and Mountjoy, W., 1973, Hydrothermal alteration associated with beryllium deposits at Spor Mountain, Utah: US. Geol. Survey Prof. Paper 818—A, 20 p. Mundorff, J. C., 1970, Major thermal springs of Utah: Utah Geol. and Mineral. Survey Water-Resources Bull. 13, 60 p. Nathenson, Manuel, and Muffler, L.J. P., 1975, Geothermal resources in hydrothermal convection systems and conduction-dominated areas, in White, D. F., and Williams, D. L., Assessment of geo- thermal resources of the United States—1975: U.S. Geol. Survey Circ. 726, p. 104-121. Nolan, T. B., 1927, Potash brines in the Great Salt Lake Desert, Utah: US. Geol. Survey Bull. 795—B, p. 29—44. Park, G. M., 1968, Some geochemical and geochronologic studies of the beryllium deposits in western Utah: Utah Univ. M.S. thesis, 105 p. Turk, L. J., 1973, Hydrogeology of the Bonneville Salt Flats, Utah: Utah Geol. and Mineral Survey Water-Resources Bull. 19, 81 p. Vine, J. D., 1975, Lithium in sediments and brines—~how, why, and where to search: U.S. Geol. Survey Jour. Research, v. 3, no. 4, p. 479—485. Whelan, J. A., 1969, Subsurface brines and soluble salts of subsurface sediments, Sevier Lake, Millard County, Utah: Utah Geol. and Mineral. Survey Spec. Studies 30, 13 p. 1973, Great Salt Lake, Utah—Chemical and physical variations of the brines, 1966—1972: Utah Geol. and Mineral Survey Water-Resources Bull. 17, 24 p. Whelan, J. A., and Petersen, C. A., 1975, Great Salt Lake, Utah— Chemical and physical variations of the brine, water-year 1973: Utah Geol. and Mineral Survey Water-Resources Bull. 20, 29 p. White, D. E., Hem,J. D., and Waring, G. A., 1963, Chemical composi- tion of subsurface waters: U.S. Geol. Survey Prof. Paper 440—F, 67 p. Williams, Norman C., 1963, Beryllium deposits, Spor Mountain, Utah, in Guidebook to the geology of Utah, No. 17, Beryllium and uranium mineralization in western Juab County, Utah: Utah Geol. Soc., p. 36—59. \_/—\ LITHIUM RESOURCES AND REQUIREMENTS BY THE YEAR 2000 79 PRELIMINARY DESIGN AND ANALYSIS OF A PROCESS FOR THE EXTRACTION OF LITHIUM FROM SEAWATER1 By MEYER STEINBERG and VI-DUONG DANG, DEPARTMENT OF APPLIED SCIENCE, BROOKHAVEN NATIONAL LABORATORY, UPTON, NY ABSTRACT The U.S. demand for lithium by the industrial sector is estimated to increase from 2.7><106 kg (3,000 short tons) in 1968 to 1.2X107 kg (13,200 tons) in the year 2000. Projections of the demand for lithium by a fusion power economy are subject to uncertainties in the future power demands of the United States and the wide range Of require- ments for various designs of controlled thermonuclear reactors (CTR). For a one million MWe CTR (D-T fuel cycle) economy, grow- ing into the beginning of the next century (the years 2000 to 2030), the cumulative demand for lithium is estimated to range from (0.55~4.7)>< 107 to 1.0X 109 kg ((0.6~5.l)><104 to 106 tons). The pre- sent estimates of the available U.S. supply are 6.9><108 kg (7.6)(105 tons) of lithium from mineral resources and 4.0><109 kg (4.4)(106 tons) of lithium from concentrated natural brines. With this kind of demand approaching supply, there is concern for lithium availability in a growing CTR (D-T fuel cycle) economy. There is, however, a vast supply of lithium in seawater. Although the concentration of lithium in seawater is dilute (170 ppb), the total quantity available is large— estimated to be 2.5x1014 kg (2.8X1011 tons)—thus insuring an unlim- ited resource for a long-term fusion economy. These estimates pro- vide the incentive for devising an economical process for the extrac- tion of lithium from the sea. A preliminary process design for the extraction Of lithium from seawater is presented on the basis of the literature data. The essential features of the process are that seawater is first evaporated by solar energy to increase the concentration of lithium and to decrease the concentration Of other cations in the bittern which then passes into a Dowex-502 ion exchange bed for cation adsorption. Lithium ions are then eluted with dilute hydrochloric acid, forming an aqueous lithium chloride which is subsequently concentrated. Lithium metal is then formed by electrolysis Of lithium chloride. The energy requirement for lithium extraction varies between 0.08 and 2.46 KWh(e)/g for a range of production rates varying between 104 and 108 kg/yr. This energy requirement is relatively small when compared to the energy produced from the use of lithium in a CTR having a value of 3,400 kWh(e)/g Li. Production cost of the process is estimated to be in the range of 2.2 to 3.2 cents/g Li. To Obtain a more definitive basis for the process design, it is recom- mended that a thorough phase equilibria study of the solid-liquid crystallization processes of seawater be conducted. Uncertainties exist in the Operation Oflarge solar ponds for concentrating large quantities of seawater. A continued thorough search for a highly selective ad- sorbent or extractant for Li from low concentration aqueous solutions should be made. Investigation Of other physical separation processes such as the use of membranes should be investigated. INTRODUCTION The lithium industry was a small industry before ‘This work was performed under the auspices of the Controlled Thermonuclear Reactor Division (CTR) of the U.S. Energy Research and Development Administration. ’Any use of trade names is for descriptive purposes only and does not constitute endorse- ment by the U.S. Geological Survey. World War 11. During the war, it was necessary to de- velop a hydrogen storage means for the warships to lift equipment by balloon in case of emergency. Lithium hydride was developed for that purpose. A small amount Of lithium hydride can react with seawater to release a large quantity of hydrogen (Hader and Others, 1951). After the war, lithium requirement for nuclear purposes stimulated the production of lithium for some years during 1954—1960. Excess lithium produced in those years required more commercial end use after the need for lithium by the Atomic Energy Commission (AEC) decreased in 1960. Commerical uses of lithium and its compounds are in the following categories (Cummings, 1968; Shreve, 1967). Lithium itself is mainly used for ceramics, glass, and multipurpose greases. Other commercial uses of lithium and its compounds are for automotive, metal- lurgy, air conditioning, batteries, and so forth. Besides the commercial use of lithium, it is anticipated that lithium can be used as blanket material in the fusion reactor. Lithium has been proposed for the blanket material of the first wall of a fusion reactor (Controlled Thermonuclear Research Division, 1973) because its nuclear properties are ideal, its thermal conductivity is high, it is a desirable coolant, and above all, it is tritium breeding material. A major difficulty with lithium as a controlled thermonuclear reactor (CTR) coolant is its high electrical conductivity which leads to excessive pressure drop and pumping power in circulating it as coolant in the vicinity of the strong magnetic field Of the plasma. To overcome this difficulty, Sze and Stewart (1974) proposed a method Of electrical insulation. Other alternative methods (El-Wakil, 1971) are the use of lithium compounds which may be one of the following four materials: (1) molten salts such as LiF-BeF2 eutectic called flibe, (2) lithium nitrates along with other nitrates, (3) mixtures of liquid metals and molten salts such as lithium and flibe, and (4) LiAl compound (Lazareth and Others, 1975). Use of lithium for blanket or coolant material can be viewed from an energy point of view (Holdren, 1971). The basic deuterium-tritium reaction will release a neu- tron, an alpha particle, and an energy of 17.6 MeV car- ried by the neutron and alpha particle. Deuterium is available in seawater but natural tritium is scarce. Bom- 80 barded by a neutron, lithium can be used as a breeder for tritium in the fusion reactor. Because of this charac- teristic, lithium becomes one of the unique elements in the fusion reactions. The energy value of one gram of lithium can be increased by increasing the breeding ratio (tritons produced/tritons consumed) or increasing the amount of energy per fusion. Lee (1969) estimated that the effective energy content of natural Li ranges from about 8.5X103 kWh (normal) per gram to 2.7x 104 kWh (thermal) per gram of natural lithium. Natural lithium consists of 7.5 percent Li6 and 92.5 percent Li7. SUPPLY AND DEMAND OF LITHIUM Production of lithium for various commercial end uses has been described previously. Fusion power should be available at the turn of the century. Natural lithium demand for fusion reactors varies depending on the reactor type as is shown in table 29. For a solid lithium blanket, the quantity required is about 5.5~46.8 kg/MWe of natural lithium or 0.46~3.9 kg/MWe of 90 percent enriched in Li6 depending on the first wall load and thermal conversion efficiency. For a liquid lithium blanket, the quantity required may be as high as about 1,000 kg/MWe. It should be noted that the minimum activity solid blanket is 90 percent enriched in Li6, while the liquid coolant blanket is natural lithium (7.5 percent Li5). In the following, all the quantities of lithium ex- pressed are in terms of the amount of natural lithium. The total lithium demand consists of both use in the blanket inventory and burning on a continuing basis as to form tritium. Tritium is thus part of the fuel cycle. With the present state of knowledge of the technology, an order of magnitude estimate rather than exact values can only be made at this time. Figure 33 shows the production and demand of lithium as a function of time. This chart indicates the slow growth of the industry prior to World War II and shows the influence of war- time demands, especially in the peak production year of 1944. Projection of future production for conventional use is obtained by a linear regression based on the pro- duction rate of previous years. Production of lithium in 1968 was about 3,000 short tons (2.7X106 kg, Cum— mings, 1968). With an annual increase in the rate of about 5 percent, the lithium production in 2000 is esti- mated to be 1.2 X107 kg/year. When the requirement of TABLE 29.—Lithium requirement per megawatt for various kinds effusion reactors inventory BNL Reactor ORNLa LLL D'l'a minimum type __________ ORMAKa Tokamaka 9Pincha laser mirror activil) Lithium required per me await (kg/M ) ______ 804 454 377 325 182 5.5~46.8 3Controlled Thermonuclear Research Division (1973'). bPowelI, ]. R. (1975). LITHIUM RESOURCES AND REQUIREMENTS BY THE YEAR 2000 9 '0 z I I | I I : _ RANGE OF — CUMULATIVE I08: DEMAND FOR : a : CTR REACTOR: x _T_ I o” : : ~ 3: MINIMUM — 5 7 ANNUAL _ . /DEMAND _ E '0 E ./// FOR CTR 5 o ; , REACTOR : z _ /// _ 9 _ // _ '— _ _. Q g 6 PROJECTIONSZ g '0 :— — LITHIUM DEMAND FOR THE E O. : CONVENTIONAL INDUSTRY : 2 I I 2 _ CUMULATIVE LITHIUM DEMAND _ I _ FOR THE CONVENTIONAL INDUSTRY _ '5 PLUS FUSION REACTORS I05: — MINIMUM ANNUAL LITHIUM DEMAND: 3 FOR FUSION REACTORS PLUS 5 — CONVENTIONAL INDUSTRY : '04 I I I I I I |930 I950 I970 I990 20IO 2030 2050 YEARS FIGURE 33.—Lithium production rate and demand projections. fusion power in the United States reaches about 106 MWe beyond 2030, the quantity of natural lithium in- ventory required ranges from about 0.55X 107~4.7X 107 kg (Powell, 1975) for enriched solid CTR blankets to as high as 109 kg for liquid lithium coolant blankets. Superimposing these quantities on the previous one for commercial use, the total quantity of lithium required lies between 4.3><107 and 109 kg beyond the year 2030 (the Energy Advisory Panel, 1973). Figure 33 gives the general trend of lithium production and demand for various time periods. Future projections of these curves may change with respect to the time scale, but the shape of the curve should remain the same. Conventional lithium sources are minerals and brines. In the United States most of the lithium minerals come from Kings Mountain, North Carolina. Mainly spodumene and some amblygonite are mined in North Carolina. South Dakota also has some reserves which consist of spodumene, amblygonite, and lepidolite. Elemental composition of these minerals is shown in table 30. Spodumene has the highest lithium composi- tion as compared to the other minerals listed in the ta— ble, and more than half of the Free World’s output of lithium oxide came from this source in 1963. Estimation LITHIUM RESOURCES AND REQUIREMENTS BY THE YEAR 2000 of the total quantity of resources from these materials is about 6.89X 108 kg, as can be seen in table 30. In order to produce more lithium and to reduce min— ing operations which rely on relatively expensive mining and conventional hydrometallurgical processing, brine has been investigated for an alternative source for lithium. Lithium has been found in brines in several places such as Silver Peak, Nevada; Great Salt Lake, Utah; Searles Lake, California; and Bonneville, Utah. Of these locations, lithium concentration is highest in Silver Peak, Nevada, as can be seen in table 30 (Bren- nan, 1966). Commercial production has been carried out in Nevada, Utah, and California. Estimates of the content of two of the richest lithium brines are 3.45 X 109 kg in Silver Peak, Nevada, and 5.91 X 108 kg in the Great Salt Lake in Utah, as is shown in table 30. The total estimated supply of lithium from brine and mineral reserves is thus about 4.73X109 kg, which comes close to the maximum cummulative quantity de- manded for fusion reactors after 2030, estimated to be as high as 109 kg. If there is any additional large de- mand for lithium, such as for lithium sulfur batteries and for Li7 as water conditioner in fission reactors, the reserves will not be sufficient to sustain a long term fu- sion economy. To remedy this situation, a large new lithium source of supply should be explored in conjunc— tion with a reduction of fusion reactor lithium demand. In this report, seawater is examined as an alternate large source for Li. The concentration of lithium in sea- water has been determined to be 0.17 ppm (Riley and Tongudai, 1964; Riley and Skirrow, 1965; Chow and Goldberg, 1962). Although the lithium concentration in the sea is very dilute, its quantity is tremendous— estimated to be about 2.5x 1014 kg. If annual require- ment for lithium is a maximum of 4.32X106 kg/yr after 2030 (Powell, 1975) for solid blanket or even about twenty times larger than the above quantity for liquid blanket, the supply of lithium from the sea can last a million years or more. Thus the presence of an unlim- 81 ited fuel resource as promised for a deuterium economy becomes possible. Obviously this is possible only if the technology for extraction of lithium from seawater be- comes economically feasible. The objective of the present paper is to outline a process for the extraction of lithium from seawater and to determine in a preliminary manner the mass, energy balance, and economics of such a process. PROCESS CONCEPT FOR EXTRACTION LITHIUM FROM SEAWATER OF Regardless of which process scheme is considered, the minimum theoretical energy requirement for extraction of lithium from seawater can be determined from the equation of minimum free energy change, AF =RT1n ('yN), where AF is computed to be 0.07 kWh (thermal)/ gmLi; AF is the difference in free energy, R is the ther- modynamic gas constant, I is absolute temperature in degrees Kelvin, and 7N is the activity of the dilute species. Several possible routes for lithium extraction were considered mainly based on a study of the literature. The various processes considered for the concentration of dilute species included adsorption, solvent extraction, ion exchange, evaporation, precipitation, and elec- trochemical separation. Ideally, it would be most advan- tageous to have a specific selective adsorbent for lithium. Although this is technically possible our literature sur- vey did not reveal such an agent. As a result, prelimi- nary calculations indicated that an extraction process involving concentration by solar evaporation followed by ion exchange and finally electrolytic separation may be a reasonable process approach. This process has been determined to be a more economical one than a process using solar energy alone for salt separation. The plant can be located either by_ the seashore or on an offshore island. The ion exchange beds may be built in the strait so that power required to pump seawater to the beds may be eliminated. ' TABLE 30.—Mineml contents and sources of lithium [weight percent] Silver Peak, Great Salt Lake, cl" (1* * (1* Mineral Nevada“ Utaha Sea waterb Spodumenec‘ Petalite ‘3 LepidoliteC-d Amblygoniteic' Sodium ________________________________________ 7.5 7.0 1.08 0.36 0.68 0.50 0.88 Magnesium ____ _________________________ .06 .8 .13 .60 ___- -___ ___- Calcium --- .05 .03 .04 .34 ___- ---_ ___- Lithium .04 .006 .000017 2.80 1.58 2.26 .85 Potassmm 1.0 .4 04 .60 .17 10.18 ___- Silicon__ .0. .0004 0003 29.7 36.2 23.63 ___- Aluminum- .-._ ______________ 13.9 8.81 13.30 17.87 Iron _______________ -__- ______________ .95 .14 .29 _--- Sulfate .75 1.5 .27 ---- _--- ____ ___- Chloride--. 11.7 14.0 1.94 ____ ____ --__ __-- Bromide__- ________________ .0 .0 .0004 ___- -___ ____ ___- Total lithium estimated (kg) ________________ 3.45x10' 5.91 x 108 2.5x 1014 689x10“ ~All the metals are in oxide forms. Another major element in amblygonite is phosphorus. a Brennan (1966). Riley and Skirrow (1965. p. 164). C Cummings (1968). d Me110r( 946. p. 425). 82 Extraction of sodium chloride and magnesium from seawater by solar evaporation has been practiced on a commercial scale for many years (Shrier, 1952; Schambra, 1945). Extraction of uranium from seawater by an ion exchange method was proposed by Davies and others (1964). The concentration of lithium in seawater lies between that of sodium and uranium. The applica- tion of solar evaporation alone can hardly produce pure lithium without subsequent chemical treatment, because of its dilute concentration. Application of ion exchange beds alone will require a large quantity of ion exchange resins and therefore become uneconomical. Hence, a method of combination of solar evaporation and ion exchange is proposed for extraction oflithium from the sea. The separation and quantitative determination of lithium in seawater have been carried out by several investigators (Chow and Goldberg, 1962; Riley and Tongudai, 1964; Leont’eva and Vol’khin, 1973) using ion exchange methods. Table 31 lists a survey of these methods. Strelow and others (1974) separated lithium from sodium, beryllium, and many other elements by eluting lithium with 1 molar nitric acid in 80 percent methanol from a column of sulphonated polystyrene cation—exchange resin. This method is similar to that of Riley and Tongudai (1964) and may be used for separa- tion of lithium from seawater. The proposed large-scale engineering process design is based on the laboratory results of Riley and Tongudai (1964). A schematic flow diagram of the process is shown in figure 34. Seawater flows in the solar pond by tidal waves. Ponds can be built according to the topography of the seashore. The concentration ponds should be built at a higher elevation and the crystallization pond will be built at a lower elevation so that concentrated brine can flow from the concentration pond to crystalli- zation pond by gravity. In the concentration areas, water LITHIUM RESOURCES AND REQUIREMENTS BY THE YEAR 2000 is evaporated successively to different concentrations of lithium in ponds of various sizes. In the process of frac- tional crystallization of seawater, sodium chloride, cal- cium sulfate, and magnesium chloride will precipitate first because of their much higher concentration and thus will reach their solubility limit first. Assuming the salts precipitated out in the order with calcium sulfate first, followed by sodium chloride and magnesium chloride, it is found that when the seawater is concen- trated to 105‘ parts of original seawater by evaporation, then lithium chloride will begin to precipitate. The solar pond systems can be designed to be big enough so that two parts in 106 parts of water is left after solar evapora- tion. In this manner, the size and capital investment of the ion exchange bed can be greatly reduced. The concentrated lithium brine is passed into the pre- cleaned ion exchange bed of Dowex 50~X16 (polymer beads of polystyrene crosslinked with 16 percent polyvinylbenzene). The characteristic of the Dowex resin is selective for exchange of its hydrogen ion with the cations in the seawater in the order of potassium, sodium, and lithium followed by magnesium. After lithium ions are adsorbed, the remaining seawater will be concentrated in a smaller solar pond to concentrate the chloride ions in solution to a stronger hydrochloric acid. One can design the size of the bed to be big enough so that it can be operated 24 hours before the bed is saturated. In order to maintain continuous operations, a second bed is needed so that while one bed is function- ing for the adsorption mode, the other bed can operate in the elution mode. Lithium ions are first eluted from the pregnant bed with 0.2—0.5 normal hydrochloric acid. The eluted lithium chloride solution from the bed can flow in the evaporator to obtain concentrated lithium chloride or recycle back to the bed in order to elute more lithium ions. The lithium chloride then flows TABLE 31.—Literature on separation and determination of lithium from seawater Li concentration Method of Author Ocean or sea (lag/l) Method of separation determination (1) Marchand (1899) English Channel __ 200 ________________________________________ Gravimetric as LiJPO‘. (2) Thomas and Thompson (1933)-- North Sea ________ 100 Precipitate Ca and Mg as carbonates, NaCl Visual flame spectro- recipitated with alcohol or HCl. Residual photometry. Rig precipitated. Filtrate evaporated to dryness and residue extracted wrth alcohol. (3) Goldschmidt and others (1933) __________ North Sea __________________ 72 ________________________________________ Spectrographic on sea salts. (4) Strock (1936) ____________________________ North Sea __________________ 140 ________________________________________ Spectrographic on dry salts. (5) Bardet and others (1937) ________________ North Atlantic ______________ 200 Li separated from NaCl by amyl-alcohol ex- Gravimetric as Li2804. traction. Other elements removed by re~ peated preci itation and evaporation. (6) Ishibashi and Kurata (1939) ______________ Pacific (Ilapanese 170 Concentrated y eva oration and elements Gravimetric. coasta waters). removed by precipitation. (7) Chow and Goldbert (1962) ________________ Pacific ______________________ 173 Sample enriched wrth 'Li. Li se rated by Isotope dilution. ion exchange and eluted with EC]. (8) Kappanna and others (1962) ____________ Indian (coastal waters) ______ 160 Li + residual Mgdprecipitated as phosphate Gravimetric. } and weighed. g determined utrimetric- i ally with EDTA. (9) Riley and Tongudai (1964) ______________ All oceans __________________ 183 ample enriched with Li. Li separated by Flame photometry. 1 ion exchange and then eluted with HCl. (10) Leont‘eva and .Vol‘khin (1973) ____________ High mineral solution ______ 11,000 (lsM-l ion-sieve cation exchange. Do. LITHIUM RESOURCES AND REQUIREMENTS BY THE YEAR 2000 83 ION A EXCHANGE: l BED SEA I SOLAR WATER POND PUMP 2.74xm" g/m I If I POND FOR HCI CONCENTRATION V ”j E aN HCI ‘NaCI,KCI.MqCI, T0 SEA 0.2-0.5 NHCI OW. CONDENSER PROCESS STEAM o.2-o.5u NCI A ‘ v Ii BUENER , 7 . //// CONDENSER APORATOR A A A II, CI. ‘ LiCI ma" ———— ‘: / 4.I9XIo’q/nr I v A : LITHIUM > METAL I PRODUCT | PUMP | I I I | I GENERATOR CTR : TURBINE FIGURE 34,—Schematic flow diagram for extraction of lithium from seawater. into an electrolyzer where lithium is electrolyzed. Hy— drogen and chlorine from the electrolyzer react in a burner to form hydrogen chloride which is then dis— solved in water from the evaporator to form dilute hy- drochloric acid. The dilute hydrochloric acid is recycled back to the ion exchange bed to elute lithium ions. Other cations such as sodium, potassium ions, and so forth will be eluted by concentrated hydrochloric acid (8 N) from the ion exchange bed after lithium ions are eluted out by dilute hydrochloric acid and before fresh seawater is passed in the bed. The eluted effluent is then passed back to the sea. Evaporators used are triple-effect backward-feed evaporators. The assumed steam economy (total water evaporated/total steam supplied) is three. Boiling point elevation in each evaporator is neglected. Electrolyzer design follows closely the one described by Hader and others (1951) for electrolysis of lithium chloride. PROCESS DESIGN AND ECONOMIC ANALYSIS3 In the present design calculations, W gm/hr of lithium is taken as the production unit throughout and all the design quantities are based on this variable. Once W is 3The following discussion of the engineering design for a method of extracting lithium from seawater employs English units as the fundamental data for discussion. It involves land area in acres and square miles, equipment size and capacity in standard American sizes, and costs in dollars and cents. Conversion to metric units is not practical. Editor. defined, all other quantities may be computed. The first major unit to design is the solar pond system. The rate of evaporation of water by solar energy is an important issue here. From a simple energy balance, one can take the heat required for evaporation to be supplied by solar radiation according to the following equation: \ MA=BAt, (1) where A is latent heat of water taken to be 1,000 Btu/lb; B is the daily average solar energy received on the ground that varies from a few hundred to 2,500 Btu/ft2 (Lof, 1966; Pancharatnam, 1972); t is the time interval during which the seawater receives solar energy for evaporation;A is the pond area in ft2; and M is the mass of water evaporated in pounds. Validity of this equation can be seen from data of Silver Peak (Brennan, 1966) which shows that 107 lb/yr of lithium carbonate is pro- duced in a pond of 910 acres. Daily solar energy radia- tion received on the ground is taken to be 1,500 Btu/ftg. In order to obtain the number of active days of evapora- tion from about May to September each year (the period where abundant sunshine is received on ground in the West Coast), a correction factor of 0.5 is needed to vali- date the results of Silver Peak (Brennan, 1966) in using equation (1) which is a simple and good approximation for large complex solar pond systems. Hence, one can 84 then apply equation (1) for solar pond design for extrac- tion of W gm/hr of lithium from seawater having an initial lithium concentration of 170 ppb. If we select a plant location on the Southwest coast of the country with hourly solar energy intensity of 187.5 Btu/hr ft2 and time of evaporation to be 8,000 hr/yr, we can calculate the pond area to be 0.288X105W ft2 in order to evapo- rate 2.36X1010 W g/yr of seawater. For a production rate of lithium of 106 kg/yr which is sufficient for one 1,000 MWe fusion reactor of ORMAK or one hundred 1,000 MWe fusion reactors of Brookhaven National Laboratory (BNL) minimum activity blanket, the area of solar pond required is about 155 square miles. Utiliza- tion of waste heat of about 3X 109 Btu/hr from a 1,000 MWe fusion reactor for part of the heat supply for solar evaporation can also be considered. With the present design, it is found that using the waste heat for solar evaporation can only be reduced by about 0.6 square miles, or about 0.3 percent of the land calculated with- out using waste heat. The concentration of cations initially in the seawater which will precipitate in a salt form such as sodium chloride, calcium sulfate, magnesium carbonate, potas- sium chloride, and so forth, is about 0.0351 g/g seawa- ter. When two parts of the initial 105parts of seawater is left from evaporation (lithium chloride will begin to precipitate when 1 part of the initial 106 parts of sea- water is left), 12.6 W g/hr of bittern flows into the ion exchange bed for cation adsorptions. More than 99.9 percent of CaSO4, NaCl, MgCl2, and KCl is precipitated in the crystallization ponds. The flow rate of bitterns that contain cations is 0.568 W g-equivalent/hr. The capacity of D0wex-50 is 4.32 meg/g and the design of the ion exchange bed is for a 24-hour operation period before switching over to another bed. Then the total amount of resin required is 6,320 W g. The density of the resin is 0.802 g/cms, so the volume of each bed is 4.14X103 (W/N) cm3, where N is the number of beds. The ratio of height to diameter of the bed is taken to be 15 (Dow Chemical Company, 1958; Chow and Goldberg, 1962; Riley and Tongudai, 1964), so height and diameter of each bed are 93.5 (W/N)‘/3cm and 6.23 (W/N)V3 cm, respectively. According to Riley and Tonguadi (1964), 500 ml of 0.5 normal HCL is used to elute 1.17 mg of lithium. Theoretically one equivalent of hydrogen ion can dis- place one equivalent of lithium so the hydrochloric acid can recycle 1,176 times to elute lithium in the bed before it passes into the evaporator‘to drive off water from aqueous LiCl. The circulation load of hydrochloric acid is 1.5 W gal/min. Riley and Tonguadi reported that one liter of 4 N HCl is used to elute other cations such as sodium, potassium, and so forth from the bed. In the present operation, the concentrated hydrochloric acid LITHIUM RESOURCES AND REQUIREMENTS BY THE YEAR 2000 required is less than that used by Riley and Tonguadi because more than 99.9 percent of the cations are pre- cipitated. With this reduction of cations in the bittern, 25.6 W1 g/hr of 4 N hydrochloric acid is needed. When all the hydrogen ions in 0.2N hydrochloric acid are displaced by lithium ions in the ion exchange bed, LiCl ofi 8.48 g/l flows into the evaporator at a rate of 746 W cm3/hr. To obtain a solid lithium chloride in the evaporator, 99.152lpercent of the water has to be evaporated. The evaporator type is a triple effect backward feed with a steam economy of three. Process steam at 140 psi and 353°F is used to supply heat of evaporation of water. Using a heat transfer coefficient of 500 Btu/hr ft2°F, one can estimate the heat transfer area of the evaporator to be 4.62X10_3 W square feet. The electrolyzer used is similar to the one described by Hader and others (1951) except larger in scale. A cell with dimensions of 14 ft by 21 ft by 10.5 ft is used and its capacity is 161 lb/hr. Fourteen graphite anodes, 1.25 ft in diameter and 21 ft long, are supported from above the cell and extend downward into the electrolyte bath. The cathode is made of steel. Operating conditions would be the same as in a conventional cell (Hader and others, 1951). Table 32 shows a summary of the dimen- sions and number of major units. The area of the solar pond is a significant factor in the present process. Table 33 shoWs the areas of ponds required for various pro- duction rates of lithium based on equation (1). An important aspect of the process is the energy re- quirement. Mechanical energy is required to transport bittern and hydrochloric acid between the ion exchange bed and other pieces of equipment. The mechanical energy equation for fluid transport in the present sys- tem is _g_ h A 142 gt where the first term is for potential energy, the second term is for kinetic energy, the third term is frictional losses in the pipe line, with Fanning friction factorf, length of pipe L, diameter of pipeD and the fourth term is the pumping power required with the pumping effi- ciency 80 percent. Writing the velocity term in terms of the mass flow rate of liquid and the diameter of pipe, one can rewrite equation (2) as follows: gAh+ 862 g gt}; 2 2D4 + + Ws=0, (2) 2g: ch 32L G2 gop? 2D5 By substituting the mass flow rate for bittern (3.5 X 10—3 W lb/hr), 0.5 N HCl (6.3 W lb/hr), 4N HCl (5.64x10-2 W lb/hr), height of ion exchange bed and length of transporting line which is taken 7.1 times of the bed height, one can obtain the pumping power required in +Ws=0. (3) LITHIUM RESOURCES AND REQUIREMENTS BY THE YEAR 2000 85 TABLE 32.—Extraction of lithium from seawater—summary of dimensions and numbers of major pieces of equipment Equipment [on exchange bed Pumps Evaporator Electrolyzer Solar pond 2 Dimensions Height (h)=3.07 (W/N)“3; 7,000 gal/min Total area=4.62X10‘3W ft2 21 ft by 14 ft by 10.5 ft Area= 1:2" diameter (D)=h/15. Number of units (N) N=(153.5/h)3 N=2.14X10“W N=3 N= J— 155 mi2 or 454““ 12.4 mix12.4 mi Values are given for N=17 when 27 645 ft2 each 2 W=1.25x105g/hr. h=60 ft D: 4 ft N =5 when h =90 ft D: 6 ft TABLE 33.—Extraetion of lithium from seawater. Solar pond area as a func- tion of lithium production rate Lithium production Solar pond area rate (kg/yr) (square miles) 104 _______________________________________ 1.55 105 _______________________________________ 15.5 106 _______________________________________ 155 Note: for higher production rate, one can have plants located in several sites so that the requirement for solar pond area will not be concentrated in one location. terms of kWh(e)/lb Li as _._0-35X10‘9 W+l.35x10- 4h+0.61><10—11—@. D4 D5 The thermal energy required in the evaporation is 73.1 kwh(thermal)/lb Li. With 80 percent cell efficiency, energy required in the electrolyzer is 8.15 kWh(e)/lb Li; so, total energy required for extraction of lithium from seawater by the present process in terms of kWh(e)/lb Li is then 9%9—9 W+1.35><10-4h+0.61x0.61><10‘“ ”12” —+37.39. D5 Table 34 shows a summary of energy requirements calculated from the above expression for various lithium production rates, pipe sizes for liquid transport, and heights of ion exchange bed. When the lithium produc- tion rate is less than 105 kg/yr, the energy requirement for the present process is mainly for thermal energy in the evaporation and electrical energy for the elec- trolyzer. At a higher production rate, pumping energy becomes more important. When the lithium production rate is less than 106 kg/yr, energy requirement for lithium extraction is less than 0.2 kWh(e)/g Li. Energy requirement increases with increase in lithium produc- tion rate and the height of the ion exchange bed mainly due to increase in pumping power for liquid transport in the pipe line. However, increased pipe size will de- TABLE 34.—Summary of energy requirements for lithium extraction from seawater at various lithium production rates, pipe size for liquid transport, and height of ion exchange bed Lithium ctio H ' f' Ener re uirement for Eariiflu n Diameter of pipe exilfiliig: 6:3 Eil’rgduction“ (kg/yr) (0(a) ) (Mft) ) kWh0+.05w"0-2 8.18 2.69 2.34 2.18 100.0 100.0 100.0 --._ a arrin on and others (1966, p. 37:). biHKunin $958, p. 396). LITHIUM PRODUCTION COST (cents/gm) LB 1 I I I 10‘ 105 10‘ LITHIUM pnooucnou RATE (Kc/yr.) FIGURE 35.—Lithium production cost versus lithium production rate. nation of lithium in seawater: Jour. Marine Research, v. 20, no. 3, p. 163—67. Cummings, A. M., 1968, Lithium, in Mineral Facts and Problems [1970 ed.]: Staff of Bureau of Mines, p. 1073—81. Controlled Thermonuclear Research Division, 1973, Fusion power; an assessment of ultimate potential: WASH—1239, 44 p. plus appen- dix. Davies, R. V., Kennedy,]., McIlroy, R. W., Spence, R., and Hill, K. M., 1964, Extraction of uranium from seawater: Nature, v. 203, no. 4950, p. 1110—15. Dow Chemical Company, 1958, Dowex; ion exchange: Dow Chemical Company. El-Wakil, M. M., 1971, Nuclear energy conversion: Intext Educational Publishers, p. 545. Energy Advisory Panel, 1973, An assessment of new options, in Energy Research and Development: AET—Q November, p. 217. Goldschmidt, V. M., Berman, H., Hanptman, H., and Peters, U., 1933, Zur geochemie des alkalimetalle: Nach. Ges. Wiss. Got— tingen, Maths. Phys. K1 IV (N.S.) 2, p. 235—38. Hader, R. N., Nielsen, R. L., and Herre, M. G., 1951, Lithium and its compounds: Indus. and Eng. Chemistry, v. 43, no. 12, p. 263—46. Harrington, F. E., and Salmon, R., Unger, W. E., Brown, K. B., Cole- man, C. F., and Crouse, D. _I., 1966, Cost commentary on a pro- posed method for recovery of uranium from seawater: ORNL Central Files Number 66—2—57. Holdren, _I. P., 1971, Adequacy of lithium supplies as a fusion energy source: Lawrence Livermore Laboratory, UCID-15953. Ishibashi, M., and Kurata, K., 1939, Determination of lithium in sea- water and bittern: Jour. Chem. Soc. japan, v. 60, p. 1109—1 1. Kappanna, A. N., Gadre, G. T., Bhavnagary, H. M., and joshi, J. M., 1962, Minor constituents of seawater: Current Science (India), v. 31, p. 273. 88 Kunin, R., 1958, Ion exchange resins [2d ed.]: John Wiley and Sons. Lazareth, O. W., Powell, J. R., Benenati, R., Fillo, J., Majeski, S., Sheehan, T. V., 1975, Minimum activity aluminum reference de- sign: Trans. Am. Nuc. Soc., v. 21, 1975, p. 51—3. Lee, J. D., 1969, Tritium breeding and energy generation in liquid lithium blankets: Proc. BNES Conf. on Nuclear Fusion Reactors, p. 471. Leont’eva, G. V., Vol’khin, V. V., 1973, Determination of the lithium content of solutions with high mineral contents by means of ISM-1 cation exchange: Zhur. Anal. Kimii, v. 28, no. 6, p. 1209— 11. . Lof, G. O. G., 1966, Solar distillation, in Spregler, K. 8., ed., Principles of Desalination: Academic Press, p. 152—55. Marchand, E., 1899, Des eaux potables en general: Mem. Acad. Med. (Paris), v. 19, p. 121—318. Mellor, J. W., 1946, A comprehensive treatise on inorganic and theoretical chemistry: v. II, p. 425. Page, J. S., 1963, Estimators manual for equipment and installation cost. Pancharatnam, S., 1972, Transient behavior of a solar pond and pre- diction of evaporation rates: Indus. and Eng. Chem. Process Des. Develop. v. 11, no. 2, p. 287—92 and v. 11, no. 4, p. 626—30. Peters, M. S., and Timmerhaus, K. D., 1968, Plant design and econom- ics for chemical engineers: [2d Ed]: McGraw-Hill Book Co. LITHIUM RESOURCES AND REQUIREMENTS BY THE YEAR 2000 Popper, H., 1970, Modern cost-engineering techniques: McGraw-Hill Book Co. Powell, J. R., 1975, Beryllium and lithium resource requirements for solid blanket designs for fusion reactor: BNL 20299. Riley, J. P., and Skirrow, G., 1965, Chemical oceanography: London Academic Press, p. 164. Riley, J. P., and Tongudai, M., 1964, The lithium content of seawater: Deep-Sea Research, v. 11, p. 563—68. Schambra, W. P., 1945, The Dow magnesium process at Freeport, Texas: Trans. AIChE 41, p. 35—51. Shreve, R. N., 1967, Chemical process industries: New York, McGraw—Hill, 358 p. Shrier, E., 1952, Passing the salt: Chem. Eng., Oct. 1952, p. 138. Strelow, F. W. E., Weinert, C. H. S. W., and Van Der Walt, T. N., 1974, Separation of lithium from sodium, beryllium, and other elements by cation-exchange chromatography in nitric acid- methanol: Anal. Chem. Acta, v. 71, p. 123—32. Strock, L. W., 1936, Zur geochemie des lithiu: Nach. Ges. Wiss. Got- tingen. Math. Phys. Kl. IV (NS) 1, p. 171—204. Sze, D., and Stewart, W. E., 1974, Use Ofelectrical insulation in lithium cooled fusion reactor blankets: Proceedings of the First Topical Meeting on the Technology of Controlled Nuclear Fusion, CONF—74—4—2—Pl, p. 652. Thomas, B. D., and Thompson, T. G., 1933, Lithium in seawater: Science 77, p. 547—48. m LITHIUM BRINES ASSOCIATED WITH NONMARINE EVAPORITES By RICHARD K. GLANZMAN and A. L. MEIER, U.S. GEOLOGICAL SURVEY, DENVER, CO Lithium is enriched in brines resulting from the evap— oration of ground and surface water in the arid regions in the western United States. Most of the world’s total supply of lithium is in the continental crust (Ronov and others, 1970). Chemical and mass weathering constantly redistribute lithium on the continental surfaces where runoff in the form of both ground and surface water from the continental highlands transports lithium to val- leys from which there is no surface drainage (closed ba- sins), and continuous evaportation of the water results in brines and the precipitation of the less soluble salts. Lithium salts are very soluble; the ions of the salts tend to remain in the brine and are among the last salts to precipitate. This process was in operation during late Tertiary and Quaternary time in the semi-arid to arid regions of western Utah. The geologic histories of Lake Bonneville in the Pleistocene and the Great Salt Lake in more recent times illustrate this process. The drainage basin of Lake Bonneville, therefore, was investigated to determine the areal distribution and enrichment of lithium resulting from the evaporation process. The location, areal extent, and major physiographic subdivisions of the Lake Bonneville drainage basin are shown in figure 36. Lake Bonneville was the largest of the Pleistocene Great Basin Lakes with approximately 140,000 km2 (54,000 mi2) of drainage basin. Paleozoic marine limestones and dolomites in many of the present mountain ranges formed islands in Lake Bonneville. Weathering of these limestones and dolomites undoubt- edly contributed much of the carbonate deposited in the remnant terraces and benches of the lake as the lake water evaporated. Evaporation first exposed Sevier Lake playa, then the Great Salt Lake Desert; finally, the once great lake dwindled to the remnant brine of the Great Salt Lake of today. The tectonic setting of the eastern Basin and Range province in Utah is shown in figure 37. The north-south fault zone known as the Wasatch line nearly bisects the state. The highlands immediately east of this fault zone form the east boundary of Lake Bonneville and the an- cestral drainage basins. The east-west trending mineral belts that were created by the Oligocene-Miocene- LITHIUM RESOURCES AND REQUIREMENTS BY THE YEAR 2000 89 ”3° H2° l l | 42° -~':. Causeway "- SALT 4|° GREAT SALT LAKE DESERT 40° 39° 38° — O 50 MILES O 50 KILOMETRES FIGURE 36.——The location, areal extent, and major physiographic sub- divisions of the Lake Bonneville drainage basin. Pliocene and earlier tectonic activity form not only a structural setting but also contain elements normally as- sociated with lithium. The southern mineral belt (Ca1- laghan, 1973) includes the Marysvale mining district in the area where the belt crosses the Wasatch line. Lithiophorite, a lithium-manganese mineral, is reported to occur in this district (Bullock, 1967). The northern belt is the Nevada-Utah beryllium belt (Shawe, 1966) containing both beryllium and tungsten mineral de— posits. Lithium is generally associated with both ele- ll4° II2° 42° 41° 40° 39° Maximum extent at W Lake Bonnevllle 7" Northern mineral belt N Southern mlneral belt 35° 37° | 1 l 1 FIGURE 87.—The tectonic setting of the eastern Basin and Range province in Utah. ments. It has been known for many years that lithium is enriched in the Spor Mountain beryllium district (Shawe and others, 1964; Starkey and Mountjoy, 1973; and Lindsey and others, 1973). Lake Bonneville (outlined in fig. 36) was not the first large lake in Utah. Large lakes have occurred in Utah at least as early as Eocene time with the Green River Lake (Hunt, 1956) in the Uinta Basin (fig. 38a). The Lake Bonneville drain— age basin was a land mass with streams draining across WASATCH LINE (Hunt, I956) EOCENE (Heylmun, |965) MIOCENE- PLIOCENE b EXPLANATION Residual lake in late Eocene time FIGURE 38.——The pre-Bonneville lake and lake-playa distribu- tion in Utah. . 90 LITHIUM RESOURCES AND REQUIREMENTS BY THE YEAR 2000 the Wasatch line into the Green River Lake. These en- vironmental conditions should have effectively leached any near-surface evaporites along the Wasatch line into the Green River Lake. The lake evaporated to the re- sidual lake in late Eocene time and to dryness in the Oligocene. Tectonic activity along the Wasatch line dur— ing the Oligocene and early Miocene resulted in a lower- ing of the western part of Utah and the formation of a progressively larger lake or series of lakes and playas from the northwest corner of Utah to the extent shown in figure 38b (Heylmun, 1965). Identification of any evaporites formed in these western lake sediments will depend on drilling because of burial by Lake Bonneville sediments. According to Stokes (1973) and Axelrod (1957), during Oligocene-Miocene time the eastern Basin and Range physiographic province had a humid climate which caused deep weathering and supported deciduous hardwood forests. The Pliocene to early Pleistocene climate, however, was one of increasing arid- lt . yWater associated with volcanic rocks of Oligocene through Pliocene age within the mineral belts trans- ported lithium into Tertiary drainage basins. Because lithium is a very soluble element, water and brine are the most indicative sampling media for lithium exploration. A sampling program was conducted following a litera- ture search for all available lithium data in the eastern Basin and Range physiographic province (essentially Mundorff, 1970; Young and Carpenter, 1965; Nolan, 1927; Turk, 1973; Lindenberg, 1974; and White and others, 1963). The combination of tectonism with the formation of closed basins and of a climatic sequence of increasing aridity before Lake Bonneville gives favorable condi- tions for the concentration of lithium. The specific areas of enrichment depend on a structural basin to form the trap and a supply of lithium to be transported into the trap. Two areas within the drainage basin appear to satisfy both conditions—the Sevier—Black Rock desert area and the northern Great Salt Lake Desert area. The results of the sampling program in combination with other available data for the area between the north and south mineral belts are shown in figure 39. In gen— eral, a sample containing more than 1 mg/l (milligram per litre) lithium may indicate an area suitable for further study; more than 10 mg/l in a potable or brack- ish water certainly is suitable. The lithium concentration in samples along the upper Sevier River valley from Kingston tojust below Salina ranges from 2 to 8 mg/l. Marine evaporites near Redmond increase the total dis- solved solids but contribute little or no lithium. The Sevier River drains into the Sevier Desert. At the pre- sent time, the course of the Sevier River would carry water through the Sevier Desert to Sevier Lake Playa. LITHIUM CONCENTRATION o <| (mg/l) O I—IO (mg/I) l l0-4O (mg/I) 30-50 (mg/I) V _ /5o |00(mg/I) 0 40 MI 0 40 KM Provo \ ./I - . _ | If I IDEEP .' .- , I CREEK HONEY— SPOR : MTNS COMB MTN . I | HILLS ' I: <. . 2' I ' O > < u I— Z 3 . ' _. 39. o Redmond R [— SPELVAIYEA .Meodow o 50”ch O o i Hutton '. O | . I Roosevelt ' . . I Mar svale i ' Milford 0 y | I Thermo‘ 0 8 I 36°F _ ESCALANTE I I DESERT I] .0 . FIGURE 39.—Lithium concentration in water and brine in the Lake Bonneville drainage basin. .Surface water, however, seldom reaches the playa. Near surface brine (within upper 6 ft, approximately 2 m) contains 10 to 40 mg/l lithium (Whelan, 1969). The lithium concentration with depth is unknown. The northern mineral belt formed a highland from which water containing lithium drained both to the south into the Sevier Desert and to the north into the Great Salt Lake Desert. The mineral deposits within this belt probably contributed as much if not more lithium to the Sevier Desert than those in the southern belt. Rock and sediment samples from the area of beryllium pros- pects in the Sheeprock Mountains contain 50—300 ppm (parts per million) lithium; beryllium deposits at Spor Mountain 200—1500 ppm; a beryllium prospect in the Honeycomb Hills 100—1000 ppm; and the tungsten- beryllium deposits in the Deep Creek Range 50—500 ppm. Water, in leaching lithium from these sources and LITHIUM RESOURCES AND REQUIREMENTS BY THE YEAR 2000 91 the aforementioned sources, deposited at least 1,800 m (6,000 ft) of Tertiary and Quaternary sediments in the Sevier Desert because Gulf Oil No. 1 Gronning Well in sec. 24, T. 16 S., R. 8 W., 13 km (approx. 8 mi) north— west of Delta, Utah, showed this thickness (Heylmun, 1965). Lithium-bearing clay and (or) brine enriched in lithium in the upper Tertiary rocks of the Sevier Desert would appear to be an excellent exploration drilling target. . A line of sample localities with lithium concentrations greater than 1 mg/l crosses the area diagonally from the southwest corner in the Escalante Desert to the north- east corner in Utah Valley marking the Wasatch line. Thermal springs are the major sources of these higher lithium concentrations. Water from Roosevelt thermal spring contained the highest lithium concentration of any spring sampled with 27 mg/ 1. Water from Roosevelt as well as Thermo hot springs (1.4 mg/l), hot springs near Meadow and Hatton (3.7 mg/l) all ultimately dis- charge to the north into the Sevier Desert. These sources and the upper Sevier River valley, however, are not the only sources of lithium draining into the Sevier Desert. During the time of Lake Bonneville and for an unde- termined part of the time of ancestral Lake Bonneville, water crossed the northern mineral belt into the Great Salt Lake Desert through the “Old River Bed” described by Gilbert (1890). Anomalous lithium concentrations are quite evident north of Spor Mountain, and they ex— tend into the Great Salt Lake Desert (fig. 39). The ther- mal spring at the defunct Wilson Resort, which is ap- proximately 1.61 km (1 mi) from the nearest outcrop, Fish Springs Range, produces water that contains 2.1 mg/l lithium. Near surface brine in the Great Salt Lake Desert containing 30—50 mg/l lithium surrounds the Bonneville Salt Flats near Wendover, Utah. The Bon- neville Salt Flats area and a smaller area in Pilot Creek Valley contain near-surface brine with 50—100 mg/l lithium. Whelan and Petersen (1974) have shown that an abnormally high geothermal gradient may exist in the Bonneville Salt Flats area. The lithium concentra- tion in both areas compares very favorably with the Great Salt Lake brine which ranges in concentration from 20 to 65 mg/l. A comparison of near—surface brines in the Lake Bonneville drainage basin and the Great Salt Lake illus- trates the process of enriching lithium through evapora- tion. Brine in Sevier Lake playa, first isolated by the evaporation of Lake Bonneville, contains 20—40 mg/l lithium; Bonneville Salt Flats and Pilot Creek Valley, 50—100 mg/l; and Great Salt Lake, even with fresh water inflows, contains 20—65 mg/l. Great Salt Lake is a good example of how little time is required for evaporative processes. An approximate 35 percent increase in lithium concentration in brine in the north arm of the lake has occurred since the construction of the causeway in 1959. Eardley (1970), in his discussion of the salt bal- ance in Great Salt Lake, suggested the existence of ear- lier saline lakes in the Lake Bonneville drainage basin. He further suggested that somewhere in the basin there is a buried deposit enriched in lithium, boron, and potassium. Such a deposit may be a result of late Ter- tiary volcanic events along the Wasatch line and in the two mineral belts with the formation of lithium enriched brine and clay in several parts of the drainage basin. Gravity profiles in the northern Great Salt Lake Des- ert (Cook and others, 1964) indicate a horst and graben structure beneath the playa (fig. 40). Grabens 1.22 km (4,000 ft) deep extend beneath the Bonneville Salt Flats, extend beneath Pilot Creek Valley north to the Raft River Mountains, and extend beneath the playa west of and parallel to the Newfoundland Mountains. Southern Pacific No. 1 Lemay Well, in sec. 29, T. 7 N., R. 14 W., at Lemay siding and north of the graben structure near the Newfoundland Mountains drilled through 93 m (305 ft) of brown limestone at a depth of 422 m (1,385 ft) that Heylmun (1965) assigned to an upper Miocene- lower Pliocene sequence. Paleozoic rocks were reached at a depth of 643 m (2,108 ft). The possibility that the ”4° ”2° l 4I° WASATCH LINE 40° 39° \ \ _’ ) 38° _// EXPLANATION —' / NFG Newfoundland Groben ..........m "Old River Bed"of Gilbert (I890) —' — — Mineral belts 37° 1 l FIGURE 40.—Location of the Sevier Desert, Black Rock Desert, and the gravity-defined grabens in the northern Great Salt Lake Desert. 92 upper Tertiary section formed on a shoreline or in an ancestral Lake Bonneville that occupied these graben structures is speculative because the subsurface data currently available is meager. The existence, however, of such a lake or lakes appears to be very possible. These graben structures, like the Sevier Desert, appear to be good drilling prospects for lithium. Summary The enrichment of lithium by evaporation during the formation of nonmarine brines in Utah probably began with upper Tertiary lakes and playas, continued with Pleistocene Lake Bonneville and is continuing in the Great Salt Lake Desert and Great Salt Lake. The exis- tence of the older lakes and playas is based on late Ter- tiary geologic history of flora and sedimentation, on the interpretation of near-surface geochemical data, on the interpretation of gravity profiles, on salt balance studies, and on data from a dozen oil wells drilled several years ago. Drilling of oil wells in the near future in the Great Salt Lake will add much more information on which to base the existence, and possibly the locations, of these pre-Lake Bonneville lakes. Lithium-enriched brine and clay may be encountered by drilling in the Sevier Desert, Bonneville Salt Flats, Pilot Creek Valley, and the gravity-defined graben west Of the Newfoundland Mountains formed as a result of the evaporation process in late Tertiary and early Pleistocene time. References Cited Axelrod, D. I., 1957, Late Tertiary floras and the Sierra Nevadan uplift [Calif.-Nev.]: Geol. Soc. America Bull., v. 68, no. 1, p. 19—45. Bullock, K. C., 1967, Minerals of Utah: Utah Geol. and Mineral Sur- vey Bull. 76, 237 p. Callaghan, Eugene, 1973, Mineral resource potential of Piute County, Utah and adjoining area: Utah Geol. and Mineral Survey Bull. 102, 135 p. Cook, K. L., Halverson, M. 0., Stepp,j. L., and Berg,j. W.,Jr., 1964, Regional gravity survey of the northern Great Salt Lake Desert and adjacent areas in Utah, Nevada, and Idaho: Geol. Soc. America Bull., v, 75, no. 8, p. 715—740. LITHIUM RESOURCES AND REQUIREMENTS BY THE YEAR 2000 Eardley, A. j., 1970, Salt economy of Great Salt Lake, in Rau,j. L., and Dellwig, L. G., eds., Third symposium on salt, Volume 1: North- ern Ohio Geol. Soc., p. 78—105. Gilbert, G. K., 1890, Lake Bonneville: U.S. Geol. Survey Monograph No. 1, 438 p. Heylmun, E. B., 1965, Reconnaissance of the Tertiary sedimentary rocks in western Utah: Utah Geol. and Mineral Survey Bull. 75, 38 p. Hunt, C. B., 1956, Cenozoic geology of the Colorado Plateau: U.S. Geol. Survey Prof. Paper 279, 99 p. Lindenberg, G. j., 1974, Factors contributing to the variance in the brines of the Great Salt Lake Desert and the Great Salt Lake: University of Utah, M.S. thesis, 70 p. .Lindsey, D. A., Ganow, Harold, and Mountjoy, Wayne, 1973, Hydro- thermal alteration associated with beryllium deposits at Spor Mountain, Utah: U.S. Geol. Survey Prof. Paper 818—A, 20 p. Mundorff,j. C., 1970, Major thermal springs of Utah: Utah Geol. and Mineral Survey Water-Resources Bull. 13, 60 p. Nolan, T. B., 1927, Potash brines of the Great Salt Lake Desert, Utah: U.S. Geol. Survey Bull. 795—B, p. 25—44. Ronov, A. B., Migdisov, A. A., Vosresenskaya, N. T., and Korzina, G. A., 1970, Geochemistry of lithium in the sedimentary cycle: Geochemistry Internat., v. 7, no. 1, p. 75—102. Shawe, D. R., 1966, Arizona-New Mexico and Nevada—Utah beryl- lium belts: U.S. Geol. Survey Prof. Paper 550—C, p. C206—C2l3. Shawe, D. R., Mountjoy, Wayne, and Duke, Walter, 1964, Lithium associated with beryllium in rhyolitic tuff at Spor Mountain, west- ernjuab County, in Geological Survey Research 1964: U.S. Geol. Survey Survey Prof. Paper 501—C, p. C86—C87. Starkey, H. C., and Mountjoy, Wayne, 1973, Identification of a lithium-bearing smectite from Spor Mountain, Utah: U.S. Geol. Survey, Jour. Research, v. 1, no. 4, p. 415—419. Stokes, W. L., 1973, Geologic and climatic inferences from surficial deposits of the Colorado Plateau and Great Basin: Brigham Young Univ. Geology Studies, v. 20, pt. 1, p. 11—26. Turk, L. J., 1973, Hydrogeology of the Bonneville Salt Flats, Utah: Utah Geol. and Mineral Survey Water-Resources Bull. 19, 81 p. Whelan,j. A., 1969, Subsurface brines and soluble salts of subsurface sediments, Sevier Lake, Millard County, Utah: Utah Geol. and Mineral Survey, Spec. Studies 30, 13 p. Whelan, j. A., and Petersen, C. A., 1974, Bonneville Salt Flats—a possible geothermal area?: Utah Geology, v. 1, no. 1, p. 71—82. White, D. E., Hem,]. D., and Waring, G. A., 1963, Chemical composi- tion of subsurface waters, in Data of geochemistry : U.S. Geol. Survey Prof. Paper 440—F, 67 p. Young, R. A., and Carpenter, C. H., 1965, Ground-water conditions and storage in the central Sevier Valley, Utah: U.S. Geol. Survey Water-Supply Paper 1787, 95 p. \_/-\ ORIGIN OF LITHIUM AND OTHER COMPONENTS IN THE SEARLES LAKE EVAPORITES, CALIFORNIA By GEORGE 1. SMITH, U.S. GEOLOGICAL SURVEY, MENLO PARK, CA ABSTRACT The lithium and many of the other valuable components in the salt layers of Searles Lake came from thermal springs in the Long Valley area. However, about two-thirds of the sodium, chlorine, and bromine were apparently derived from atmospherically transported sea salts LITHIUM RESOURCES AND REQUIREMENTS BY THE YEAR 2000 93 and by solution of older salt deposits. About one-quarter of the boron may have been derived from thermal springs in the C050 Range, and about 15 percent of the sulfate may have come from solution of gyp- sum in older lake beds. The components were transported by the Owens River and its pluvial-stage downstream extension via Owens and China Lakes t0 Searles Lake. Most of the potassium, sulfate, and boron that entered the Owens River reached the salt layers in Searles Lake, but about 95 percent ofthe lithium was removed prior to crystal- lization of the salts, and 50 to more than 99 percent of certain other components are not accounted for. The time required to account for the valuable salts in Searles Lake that came from springs in Long Valley is only about 10 percent of the time that the thermal springs are likely to have been active. Either one or more large deposits remains undiscovered (unlikely), or the springs have had their present chemical characteristics for only the last 30,000 to 40,000 years. INTRODUCTION Searles Lake lies near the middle of Searles Valley in the southwest corner of the Great Basin, an area where the surface drainage of all basins flows toward a central lake or playa rather than to the sea. At present, Searles Valley and most of the other valleys within the Great Basin contain playa (dry) lakes or salt flats in their cen- ters; a few valleys contain permanent lakes but most of them are near the east and west edges of the province where they receive drainage from the bordering high mountain ranges. During the pluvial periods of the Pleistocene, how- ever, most valleys in the Great Basin contained large bodies of water as a result of increased precipitation and runoff and decreased evaporation. Several of these lakes were interconnected to form chains; the evapora- tion that occurred in each lake caused progressive in- creases in the downstream concentrations of dissolved solids. Searles Lake was a lake in one of these chains. It was the third (and sometimes fourth) in a chain that included as many as six lakes during the most intensely pluvial periods of the Pleistocene (fig. 41). Owens River, which drains the east side of the southern Sierra Nevada, was the source of most of the water that flowed into this chain of lakes; although, during some periods, Mono Lake and tributaries to that basin added their waters to those of Owens River. Owens River flowed into Owens Lake, spilled out of that lake into China Lake, then into Searles Lake; during the peaks of most pluvial episodes, Searles Lake overflowed into Panamint Lake, and, at times, that lake overflowed into Death Valley. This sequence is shown diagrammatically in cross sec- tion in figure 42. The number of lakes in the chain and the depth of water in the last lake were a function of the amount of runoff versus the cumulative evaporation from their combined surfaces. Using probable evaporation rates and measured lake areas, calculations show that when ”9° EXPLANATION Lake Present playas \\ and lakes 330_ k " Pleistocene lakes Present rivers ———— Pleistocene rivers Long Va/[ey Reserve/I ‘ (La/(e Crowley) \ \ 37o .— 36° .— O 50 MILES O 50 KILOMETRES FIGURE 4l.—Index map showing relations between present lakes and rivers, and between those features during pluvial maxima of the Pleistocene. Locations of the Long Valley caldera (from Bailey and others, 1976), Long Valley Reservoir (Lake Crowley), Coso Hot Springs, and Airport Lake also shown. Searles Lake received water from upstream sources, Owens River had to flow at a rate approximately four times more than that of the present. For Searles Lake to overflow, Owens River and other sources had to contri- bute about seven times more than the present river flow. Water overflowing from Mono Lake was included in these volumes part of the time, but present runoff pat— terns suggest that less than 20 percent of the total flow FIGURE 42.—Diagram showing relation between Pleistocene lakes in the chain that includes Searles Lake; Mono Lake is omitted. Eleva- tions of the present valley floors (and spillways) are shown in parentheses. From Gale (1914). 94 was added by that basin. When the pluvial periods of the Pleistocene ended, major concentrations of salts that were trapped in the basins crystallized as layers on the valley floor. Drilling has shown that no layers of salts accumulated in the upper several hundred metres of Owens Lake and China Lake, that only halite and gyp- sum accumulated in Panamint Lake, and that major quantities of rare elements and salts accumulated only in Searles Lake (Smith and Pratt, 1957). Searles Lake now is a nearly dry salt flat with a surface area of approximately 100 kmg. Two distinct salt bodies lie within 50 m of the surface; these are informally named the Upper Salt and the Lower Salt. The volume of the Upper Salt body is approximately 1 km3 and the volume of the Lower Salt is about halfthat amount. Two chemical plants are operated by the Kerr—McGee Chem- ical Corp. on the west side of Searles Lake, one at West- end and one at Trona. One or both of them extract sodium carbonate and sulfate, potassium chloride and sulfate, lithium carbonate, sodium borate, phosphoric acid, and bromine from the interstitial brines pumped from the Upper Salt and Lower Salt. The brine which is pumped from the interstitial fluids in the Upper Salt body contains approximately 80 ppm Li, and this is pro- cessed by the plant at Trona. Brine from the peripheral parts of the body contain 10—70 ppm Li. Lithium is pre- cipitated from the brine as LigNaPO4 and converted to Li2C03. The value of all components produced annually by the two plants is in excess of $30 million. Total pro- duction from the deposit since 1926 exceeds $1 billion. GEOLOGIC HISTORY OF SEARLES LAKE The sequence of muds and salts now found beneath the surface of Searles Lake is shown in figure 43. The mud layers are mostly composed of calcium, mag- nesium, and sodium carbonate minerals, but elastic and organic components are also present. The salt layers, which represent the economically valuable part of the deposit, are composed of a large number of minerals, which are given in table 38. No minerals in the salt layers contain lithium, and all production of lithium (as well as of bromine and phosphate) comes from components that are present only in dissolved form. The history of the fluctuations of Searles Lake during the last 150,000 years is shown in figure 44. Most of the time is represented by mud layers, and only small parts of the total time are represented by salts. The climatic history represented by this curve is representative of other parts of the Great Basin, and most of the hundred or more closed basins of this area can be considered to have had enlarged lakes for about 100,000 of the last 150,000 years. During the intervening dry periods, saline layers were deposited in some basins. LITHIUM RESOURCES AND REQUIREMENTS BY THE YEAR 2000 UNIT LITHOLOGY GEOLOGIC C-I4 AGE AGE ' Overburden Mud Interbedded mud and Late Holocene salines 6000 1'— Upper Salt Solines Early Holocene 20 IO 5 — Partlng Mud Mud Late Wisconsin ' 00 30 l t b dd d I‘ 24,000— Lower Salt " 9’ Endemuff'm Middle Wisconsin 40: 33,000— 50 Mud thin beds of . Bottom Mud ’ sallnes Early Wisconsin 60 70: —I30,0oot Interbedded salines Mixed Layer and mud, grading Sunfifinmgggfilfnd down to mud ' ' 275 FIGURE 43.—Generalized subsurface stratigraphy of upper Quater- nary units in Searles Lake. Patterns represent marl (“mud") layers; clear areas represent salts. The names of the units are informal terms used locally; the indicated geologic ages are based on correla- tions with glacial deposits in the central parts of North America. TABLE 38,—List of minerals found in the saline layers (yr Searles Lake and of their compositions Aphthitalite ________________________ K3Na(SO4)2 Borax ____________________________ Nag B4O7- IOHZO Burkeite __________________________ 2Na2504 - Na2C03 Halite ____________________________ NaCl Hanksite __________________________ 9NaZSO4 - 2Na2C03 ' KCl Mirabilite __________________________ Na2804' 10H20 Nahcolite __________________________ NaHCOar Northu ite ________________________ Na2C03 - MgCOs - NaCl Sulfoha ite ________________________ 2Na2804~ NaCl- NaF Teepleite __________________________ Na2B204 - 2NaCl'4H20 Thenardite ________________________ NaZSO4 Tincalconite ________________________ Na2B4O7 ' 51-120 Trona ____________________________ Na2C03 - NaHCOa ‘ 2H20 Tychite ____________________________ 2Na2CO3 . 2MgC03 ' Na2804 Level of lake during overflow FT ?.? METRES ABOVE SEA LEVEL J J . I . O 20 4O 60 BO IOO |50 METRES ABOVE (BELOW) VALLEY FLOOR YEARS (“03) BEFORE PRESENT FIGURE 44.—Diagram showing the fluctuations in Searles Lake during the last 150,000 years. Since the original deposition of the salts and brines in Searles Lake, dilution by surface water has reduced the concentration of the original brine. The total volume of brine has apparently remained the same, so some of the original components that were concentrated in the brine—like lithium—were lost. The displaced brine pre- sumably migrated toward Panamint and Death LITHIUM RESOURCES AND REQUIREMENTS BY THE YEAR 2000 Valleys—both of which have floors at lower elevations—but this has not been confirmed. Evidence of dilution is found in three forms: (1) Large crystal cavities are numerous in certain horizons of the salt layers, and such features are rarely (if ever) formed dur- ing natural crystallization of salt layers; cavities of these sizes virtually require the removal of some original components (Smith and Haines, 1964, fig. 8). (2) The phase chemistry applicable to these saline bodies shows that the present brine compositions are those produced by partial solution of existing solids and not those pro- duced as bitterns at the close of normal crystallization (G. 1. Smith, unpub. data, 1976). (3) Data from the deuterium—hydrogen ratios in the salts and brines show that about half of the water in the present brines has been derived from local rainwater and runoff ( G. I. Smith, Sadao Matsuo, and Irving Friedman, unpub. data, 1976). The hydrogen-deuterium ratio in the par- ent brine is represented by the ratio of these isotopes in the primary mineral borax (Matsuo and others, 1972), but the present brines in Searles Lake have values that are intermediate between the values indicated by borax (approximately 0 permil relative to SMOW, Standard Mean Ocean Water) and the values found in present- day rains and runoff in the valley, which average about -80 permil SMOW. The replacement of original brine with local surface waters has resulted in a proportional loss and lower concentrations of components found only in the original brines, but components found in both minerals and brines (potassium, borates, chlorides, car— bonates, sulfates, etc.) have concentrations that are only slightly lower, because the lost quantities were mostly replaced by partial solution of the crystal body. SOURCES OF COMPONENTS IN THE BRINES The components in the salts and brines of the Searles Lake deposit were apparently derived from at least three sources: the atmosphere, rocks, and thermal springs. Much or most of the carbonate probably came from atmospheric C02 which was dissolved in the river 95 and lake waters. About 20 percent of the sodium chloride and other seawater components were probably carried into the Owens Valley by air masses moved in- land from the sea (Holser, 1970, table 3). Some of the sodium, potassium, and magnesium, and the balance of the carbonate were derived from the solution of those components in rocks that crop out within the drainage area of the extended chain of lakes; and evidence de- scribed later shows that halite and gypsum were also dissolved and contributed Cl and 804 to the total. The substances that make Searles Lake a valuable resource and virtually unique, however, were mostly derived from hot springs that lie at the head of the Owens River drainage. Lithium was among these components. Figure 41 shows the location of the Long Valley cal- dera, which is now an area containing geysers, thermal springs, and saline marshes. Analyses of the waters issu- ing from these areas are cited by the California Depart- ment of Water Resources (1967), Mariner and Willey (1976), Eccles (1976), and from an unpublished ad- ministrative report by L. V. Wilcox (1946) that is based on work done cooperatively by the US. Salinity Labora- tory (US. Department of Agriculture) and the Los Angeles Department of Water and Power. This report is published here with the written permission of both agencies. The first line of evidence that this source con- tributed many of the uncommon elements now found in Searles Lake is provided by the abundance of these ele— ments in such waters. Table 39 lists the concentration of selected ions, and table 40 lists the concentrations and monthly tonnages of chloride and boron contained in the measured discharge above and below a major ther- mal spring in the area, as well as in the Owens River. Table 39 also lists the ratios between certain elements that will be related to the dissolved components in the Owens River and Searles Lake. These data make it clear that the quantities of the unusual components are large enough, given geologic units of time, for major deposits to form. A second line of evidence that the thermal springs in the Long Valley area contributed most of the uncom- TABLE 39.—Concentrations of selected components in waters from Long Valley area and ratios of selected ions [All values in mg/l; leaders (___-) indicate no data] Station Concentrations Ratios (listed in order of d . . ”mm“ P°s“‘°"’ Na K so. c1 r B As L. Na/K cvn cvso. KIB B/Li F/Li :3 Casa Diablo hot spring‘ ______________________________ 390 45 130 280 12 15 2.2 2.8 8.7 18.7 2.2 3.0 5.4 4.3 Mammoth Creek, below Casa Diablo; average of three analyses” __________________________ 6.6 1.3 6.8 . .1 .05 .01 ,___ 5.1 10.0 .07 26 ____ --__ Hot Creek, in gorge at hot springs‘ _.__ - 400 24 100 225 9.6 10.5 ____ 2.3 16.7 21.4 2.2 2.3 4.6 4.2 Little Hot Creek1 _____________________ - 410 30 96 200 8.4 10.6 .74 2.8 13.7 18.9 2.1 2.8 3.8 3.0 Hot Creek, below most springs; average of six analyses ______ _ 83 7.0 26 43 1.8 2.0 .20 ___- 11.9 21.5 1.6 3.5 -___ ____ Owens River, Benton Crossing2 _ 24 2.8 11 ll 6 .47 .05 ____ 8.6 23.4 1.0 6.0 ___- ____ Lake Crowley, outlet3 ________________________________ 34 4 13 18 6 .66 .03 .13 8.5 27.3 1.4 6.1 5.1 4.6 fMariner and Willey, 1976, table 1. 2Eccles, 1976. table 9. 3California Department of Water Resources, 1967, appendices F. and F; Li values obtained by spectrographic analysis, average of six analyses, 1957—64. 96 LITHIUM RESOURCES AND REQUIREMENTS BY THE YEAR 2000 TABLE 40.—Monthly boron and chlorine contents (f Hat Creek and Owens River [Data from Wilcox. 1946, tables 1 and 8. Data expressed as parts per million (ppm) and tonnes per month (t/mo)] ‘Station 2 2Station 4 Owens River at Boron Chlorine Boron Chlorine Aqueduct Intake3 Discharge Dischar e (t/mo) (l/moX 10‘) ‘ ‘ — (t/moXl 6) — ‘ _ Date 0 ppm t/mo ppm t/mo ppm t/mo ppm t/mo 1933: August __________________________________ 2.41 0.37 0.89 0.33 0.80 3.15 2.34 7.37 1.52 4.79 17.54 September 198 .42 .83 .32 .63 2.65 2.57 6.81 1.72 4.56 982 October -_ 175 .39 68 38 .67 2.51 2.86 7.18 1.81 4.54 5.17 November - ______ 1 58 .40 63 29 .46 2.37 2.98 7.06 l.90 4.50 36.34 9Igetsember - ____________ 1.54 .39 60 33 .51 2.42 2.88 6.97 1.81 4.38 27.12 1 4: March _____________ 2.13 .34 72 29 .62 3.03 2.47 7.48 1.67 5.06 25.05 april --- 196 .34 66 33 .65 2.83 2.53 7.16 1.77 5.01 34.86 ay _____ 2 35 .49 1 2 28 .66 3.17 2.35 7.45 1.46 4.63 2487 June _____ 2 37 .39 92 29 .69 2.85 2.31 6.58 1.55 4.42 2592 September- 1 24 .43 53 29 .36 2.06 3.38 6.96 2.11 4.35 11.11 Mean (:7) ____________ 1 93 .40 77 31 .60 2.70 2.66 7.10 1.73 4.62 21.78 Standard deviation (:)-- .39 .04 .20 .03 .13 .37 .34 .29 .19 .25 10.54 5/37 ________________________________________ .20 .10 .26 .l0 .21 .14 .l3 .04 .ll .05 .48 lHot Creek, in canyon, approximately 200 m upstream from thermal springs (lat 37"39.8’N., long ll7°49.5’W.). 2Hot Creek. about 200 m downstream from thermal springs. aAt Aberdeen (lat 36°58.5'N., long l 18°12.G' W), about 60 km north-northwest of Owens Lake; waters include all sources ofboron in northern Owens Valley. mon elements comes from the apparent balance be- tween the tonnages of certain elements contributed by the springs of this area to the Owens River each year, and the tonnages of those components now found in the Upper Salt of Searles Lake, which represents a known number of years. Discrepancies arise in the balance be- tween several other elements, but they are in expected directions. The balance between the source and the salt body requires two assumptions. One is that the present Owens River carries approximately the same tonnages of these components each year as it did during pluvial periods of the Pleistocene, even though the volumes of water increased greatly during those periods; data sup- porting this proposal are given in a later section. The second assumption is that virtually all the dissolved sol- ids carried by the Owens River during the last 50,000 years are now found as salts in either Owens or Searles Lake. Dissolved solids carried during that period must be in one or more of the downstream basins, yet China Lake contains no salt layers and Owens Lake contains salts only in the 1—3-m surface layer formed during this century. Searles Lake overflowed into Panamint Valley part of this time (fig. 44); and the dissolved solids that reached Searles Lake must have been in waters that overflowed into Panamint Valley. However, no known sodium borate or sodium carbonate minerals are found in that valley, and the brines apparently do not contain abnormal amounts of potassium, bromine, lithium, or tungsten (Smith and Pratt, 1957). The lack of uncom- mon elements in Panamint Valley suggests that the quantities of salts lost during those periods of overflow were small. It appears, therefore, that Searles Lake con- tains the vast bulk of the materials that were contributed by the Owens River during this time. A period extend- ing from 24,000 years ago to the present is chosen be- cause that represents the period during which the Part- ing Mud, Upper Salt, and Overburden Mud (fig. 43) accumulated; the salts and brines of the Upper Salt con- tain most of the uncommon elements. Analyses of the present Owens River water are listed in table 41. The data in the first column summarize, over a 17-year period (1929—45), a large number of analyses for boron and chlorine and a smaller number of analyses for other components. These analyses were accompanied by discharge data, thus allowing calcula- tion of absolute quantities of those substances over the same period of time. The first 11 years of that period also predate the completion of the Long Valley Dam and the Mono Basin Tunnel; those structures make in- terpretations of seasonal variations in more recent chemical data more difficult. The data in the second column mostly represent one sample of water collected from the inlet to Haiwee Reservoir, but the values for TABLE 41.—Dissolved solids in waters of the Owens River or its equivalent [Values in mg/l; leaders (----) indicate no data] Owens River at aqueduct intake. Los Angeles Los Angeles range includes aqueduct at aqueduct ears from Haiwee avera e Component 1 28 to 1946‘ Reservoir2 for 19 5: Na ........................ 23—100 26 29 K _- 7—13 3.4 2.9 Ca 21—42 23 M -_- 5—33 .-_- 5.1 C5, _-- ---- 10 105 HCO, -- 130 $0, ___ ______ 15—72 17 23 Cl-.- 11—93 14 12 B - ------ 3-2.4 -_-- .33 F _ 5—1.0 53 Li-_- ______ --_- .11, .08 _-_- P0, _ _.-_ . 11 As-_- ------ _-_- ---- .02 Br--- _--_ ---- .02 l ___ _-__ ---- .04 W--- ---- _--_ ---- .07 p_H --------------- 7.6—8.4 8.4 8.18 otal dissolved solids __________ 187—691 221 181 Bishop; Li and crosses the Owens River. sample point where aqueduct empt xWilcox, 1946, table 1; sample point is at Aberd ”Friedman, Smith, and Hardcastle water sampled ingune 1965 at P int where aqueduct e 4 determine (1976). All com een, about 45 km southeast of Bishop. ponents except Li and P04 found in nters Reservmr, about 135 km south of . from water collected at point where California Highway 136 5Los Angeles Department of Water and Power, Sanitary Engineering Division records; ies into Van Norman Reservoir, San ernando, Calif. LITHIUM RESOURCES AND REQUIREMENTS BY THE YEAR 2000 lithium and phosphate are for two water samples col- lected in 1969 at the point where the Owens River en- tered Owens Lake during a year when Owens River runoff exceeded the capacity of the aqueduct. Two val- ues for lithium are listed—0.1 1 and 0.08 mg/l suggesting an average of approximately 0.1 mg/l at the time those samples were collected. The data in the third column represent the average composition of the water in the Owens Valley aqueduct, during 1975, as it flowed into Van Norman reservoir in Los Angeles. Waters carried by the Los Angeles aqueduct are presumed to be the compositional equivalent of the Owens River. Table 42 gives a sample calculation of the amounts of seven components, whose concentrations and abun— dances are known for Owens River, Owens Lake, and Searles Lake. The amounts of potassium, sulfate, and boron in the two salt lakes are quite close to the amounts calculated to have been carried by Owens River in 24,000 years. The amounts of sodium and chlorine in the two lakes are too large, and the amounts of carbo- nate and lithium are too small. Similar calculations can be made for bromine, phosphate, arsenic, iodine, tungsten, and fluorine by assuming that the quantities of those elements now trapped in Owens Lake form a minor part of the total. With the exception of fluorine, these elements are found in Searles Lake only in the brine phase, which makes calculations of their abun- dance more reliable. The balance between the quantities of these compo— nents carried by the Ownes River over a period of 24,000 years and the amounts now present in the Upper Salt of Searles Lake is plotted in figure 45. The values from table 42 are used for those seven components; the values for other components in the Owens River are calculated from table 41, and in Searles Lake, from data cited by Smith (1973, tables 14 and 15). The amounts of components now found in the Lower Salt of Searles Lake (fig. 43), compared to the amounts that would be 97 IOOO-— CleO6 NchIO6 o—o ‘9 ET: 3 N to6 0X .3 x 5°4"'°6 c MO6 I- o 03 x—xs x 2 cmo E I; loo—— 6 . SO xIO u 3:on4 x ‘l i 0 .1 KatIO6 g co :09 g .o 3" 3 5 IO-— 4 >- lelO I: po4uo5 o I; 0 < ““04 wuo‘ a 0 o 4 QIXIO FxIO‘ 0 I 1 I I I | I 1 IO I00 IOOO QUANTITY CONTRIBUTED BY OWENS RIVER (METRIC TONNES) FIGURE 45.—Diagram showing the relations between tonnages of selected components carried by the present Owens River in 24,000 years versus tonnages in the Upper Salt of Searles Lake (circles), and the tonnages carried by the river in 8,000 years versus tonnages in the Lower Salt (X). Plotted on three-cycle log scale; quantities ex- pressed as tonnes multiplied by indicated exponent. Diagonal line indicates 1:1 relation. Multiple points connected by lines represent independent determination. contributed by 8,000 years of flow by the present Owens River, are also plotted; these are discussed later. Points that lie on the diagonal line represent perfect agreement between the predicted and observed amounts. Points above the line represent a greater amount of that com- ponent in Searles Lake than predicted, and points below the line represent a deficiency or loss. The two points for chlorine represent the two values derived for the Owens River in table 42; the four points for boron rep- TABLE 42.—Comparison of amount of selected components carried by Owens River in 24,000 years with amount now in Owens Lake and Upper Salt of Searles Lake [Quantities expressed as parts per million (ppm) and tonnes (t)] Present concentration Amount carried Amount carried Amount now in Amount in Total amount now in Owens RiverI annually2 by by Owens River in U per Salt, Owens Lake in found in Searles and Component (ppm) Owens River (X 10’ 1) 24,000 years (X 10“ t) Searles flake (x 10" t)3 1912 (X 10‘ t)‘ Owens Lakes (X 10‘ t) 26 10.9 262 462 54 516 3.4 1.4 34 28 3 31 137 57.5 1,380 163 30 193 17 7.1 170 192 15 207 14 5.9 142 448 38 486 ___- ‘7.8 187 ____ ____ ____ “.77 .32 7.7 7.9 .8 8.7 .s.. 5.23 A 5.5 ’64 _ 7.2 Li ______________________ .11 .046 1.1 5.06 ‘.005 .06+ ‘Friedman. Smith, and Hardcastle (1976) except value for B. ‘Average annual flow of river, 4.2X10“ g/yr (calculated for period before irrigation in Owens Valley, 3Values from Smith (1973, table 15). except as indicated. ‘Recalculated from Gale (1914. p. 259); he incorrectly calculated quantities in this report. but he corrected them in later reports. 5Calculation assumes average va ues of 60 ppm Li in brine and 20 m in salts. IData from Wilcox (1946); value for concentration, from his table 1) , is mean for the years 1929—45 at the inlet to the aqueduct near Aberdeen.Values of amounts of CI and B carried annually compiled by methods described in footnote 3, this chapter. "D. V. Haines. unpub. data, 1956. using data of Gale, 1914, p. 254—261). 98 resent the two values for that substance in the river and two independently determined values for the lake (Smith, 1973, p. 112), plotted in the four possible com- binations. The spread among them probably indicates about the level of uncertainty in the data. As shown in figure 45, the predicted and observed quantities of boron, sulfate, and potassium in the Upper Salt are close to the diagonal line that indicates equality. The quantity of potassium in Searles and Owens Lakes is about 90 percent of that predicted. This close balance results in part from compensating errors, because some potassium must have been added downstream, where additional drainage from bedrock sources joined the system, and some potassium is normally lost from natural solutions owing to adsorption by clay. The quan- tity of boron now found in the two lakes may be a few percent less or 60 percent more than predicted, depend- ing on which combination of the estimates listed in table 42 is used; the “best” values (5.5 versus 7.2 X 106 tonnes) suggest that the lake contains about 30 percent more boron than predicted. An even greater discrepancy may exist because additional quantities of boron are con- tained in ulexite that occurs in small quantities in lake sediments exposed around the edges of Searles Valley, and boron also may be adsorbed by clays. Some increase in the amount of boron flowing from the Long Valley area may have occurred during pluvial periods, and some boron was probably contributed to China Lake and downstream basins by the Quaternary volcanic rocks in the C050 Range (fig. 41) and the C050 Hot Springs,1 which lie along the north side of China Lake. The observed quantity of sulfate in the Upper Salt of Searles Lake is about 20 percent greater than predicted, suggesting that greater quantities of sulfate were con- tributed during pluvial periods by drainage from upper Cenozoic lake beds in Owens Valley and from the gyp- siferous sediments and fault gouge zones of the Slate Range that forms the east edge of Searles Valley. In the upper salt, only about 15 percent of the pre- dicted amount of carbonate and 5 percent of the lithium are accounted for. Large quantities of carbonate are known to have been precipitated in Owens and China Lakes as calcite in calcareous silt and clay (Smith and Pratt, 1957), and the large quantities of calcite, arago- nite, dolomite, gaylussite, and pirssonite in the mud layers in Searles Lake account for major additional xDuring pluvial periods, the C050 Hot Springs (lat 36°2.8’N., long ll7°46.2’W.) probably flowed with the waters draining south via Coso Wash to Airport (dry) Lake. Those waters probably contained large amounts of boron (Austin and others, 1971, table 6). Airport Lake is now separated from China Lake by a gravel divide about 15 m above the lake surface, but shorelines showed that it overflowed during pluvial periods into China Lake (Duffield and Bacon, 1976). That uncommon amounts of boron reached Airport Lake is shown by observa- tions made in 1955 by R. E. von Huene and D. V. Haines of the sediments exposed in a trench in the lake floor (lat 35°54.4’N., long ll7°43.7’.W,). They noted concentrations of ulexite in three layers that had thicknesses ranging from 0.1 to 0.7 m and were at depths of approxi- mately 1, 2, and 4 111. Four analyses (by H. Kramer) of bulk samples from the ulexite-rich horizons averaged 0.5 percent B and 75 ppm Li. LITHIUM RESOURCES AND REQUIREMENTS BY THE YEAR 2000 quantities. Some of the carbonate lost during transit was probably restored as the lakes and streams dissolved at- mospheric CO2, but because of the slowness of this process and the rising pH of the lake waters, the original quantities were never fully replaced. The amount of lithium now found in Searles Lake indicates even more loss. The calculations assumed the following: (1) an av- erage of 30 ppm Li in the salts of the Upper Salt (the average of 54 unpublished analyses of water-soluble lithium in cores GS—1, GS—4, and GS—lO, obtained and described by the Geological Survey (Haines, 1959)); (2) an average of 60 ppm Li in the brine (intermediate be— tween the richest (80 ppm) brines, from the center of the deposit, and the brines present around the edge (10—40 ppm)); (3) a brine:salt ratio of 40:60; (4) salt and brine densities of 2.1 and 1.3; and (5) the volume of the Upper Salt to be 1.05 X 109 m3. Reasonable variations in these assumptions are possible, but none alter the fact that much of the lithium in the Owens River disap— peared before the salts in Searles Lake were crystallized. The amounts of sodium and chlorine now in Searles Lake are two to three times greater than the present Owens River seems likely to have supplied. The excess of sodium in the Upper Salt is somewhat less, because major quantities reacted after burial to form gaylussite and pirssonite in the interbedded mud layers. The ex- cess of chlorine was noted by Holser (1970). It tends to confirm his suggestion, based on both the absolute amounts of chlorine and the Cl/Br ratios in the Upper Salt of Searles Lake, that Searles received chlorine and bromine from erosion of older lake beds as well as from springs and atmospheric sources. The erosion of older lake beds also accounts for excess sodium. Lake beds are included in the Waucoba (of Hopper, 1947) and C050 (of Schultz, 1937) Formations, of late Cenozoic age. These areas drain into Owens Valley, and although little surface runoff from these areas occurs under present climatic conditions, substantially greater amounts prob- ably occurred during pluvial maxima. According to the data plotted in figure 45, the Cl/Br ratio of the Owens River is 710 (or 935), and the ratio in the Upper Salt of Searles Lake is 975; Holser (1970), using slightly differ- ent data and assumptions, estimated these ratios as 520 (his table 4) and 860 (his table 2). All estimates of Cl/Br ratios for the Owens River and Searles Lake lie between that of sea water (about 200) and that of primary marine evaporites (greater than 1,000). As noted by Holser (p. 308—9, table 4), neither thermal springs nor atmospheric salts normally contribute these elements in this ratio, and contributions from recycled salts (ultimately of marine origin) seem required. The other components plotted in figure 45, like lithium, suggest loss during transport between the Owens River and Searles Lake. Losses range from one- LITHIUM RESOURCES AND REQUIREMENTS BY THE YEAR 2000 half order to two orders of magnitude. The post- depositional dilution and loss of brine from Searles basin might account for the observed deficiency of phosphate, but it is inadequate to account for deficien- cies of the others. Nothing is known about the concen- trations of phosphate, arsenic, iodine, tungsten, and fluorine in the clastic sediments of Owens, China, or Searles Lakes, but those deposits seem most likely to contain the missing elements if they were adsorbed by clays or fixed by biological processes during transport. The losses of fluorine are most notable. The quantity plotted in figure 45 for the Owens River is calculated from the lower value (0.5 ppm) listed in table 41, and use of the larger value would have further increased the apparent loss. Table 43 lists several analyses for fluorine and other selected components in waters from the Owens River and Hot Creek (Wilcox, 1946). It also lists the ratio of the chemical activity product (AP) to the equilibrium constant (KT) for the mineral fluorite in so— lutions having the indicated compositions and estimated temperatures. These data were provided by E. A. Jenne, who used the computer program described by Nordstrom and jenne (1976). An AP/KT ratio of 1 indicates fluorite saturation under equilibrium conditions. Although the ratios of waters at the Aqueduct Intake are only 3—5 percent of saturation, evaporative concentration of 20—30 times (assuming no other changes) would theoretically pre- cipitate fluorite. Owens River water (ca. 200 ppm), con- centrated by a factor of 30 times, would have a salinity of about 0.6 percent, more than an order of magnitude below levels that produce salts. If the lost fluorine was precipitated as fluorite during transport, it would be upstream from any highly saline lakes. Therefore, large quantities of disseminated fluorite may be present in the 99 upper 50 to 100 m of muds beneath the surfaces of Owens and China Lakes. Figure 45 also shows the balance between the quan— tities of major components in the Lower Salt of Searles Lake (Smith, 1973, table 11) and the amounts contrib- uted by the present Owens River over a period of 8,000 years. That period was chosen because it produced the best balance between the boron, potassium, and sulfate, which were apparently transported during the final 24,000 years with the least gain or loss. The pattern of Lower Salt components is strikingly similar to the pattern determined for the Upper Salt.2 Potassium and sulfate are again very close to the diagonal line, indicative of equality; boron is a little above it; sodium and chloride are substantially above it, with the excess of chloride being greater than that of sodium; and carbonate is substantially below it (al- though less so than for the Upper Salt). The detailed similarity of this Lower Salt pattern to that of the Upper Salt adds substantial weight to the likelihood that the source of the dissolved solids was the same for both units, the Owens River. A major problem, however, is that the quantity of time required to introduce the components in these two salt units—about 32,000 years—is substantially less than 2A very similar pattern also emerges when the amounts of components carried by the Owens River in 2,000 years are plotted against the amounts now found in Owens Lake (table 42, columns 3 and 6). The amounts of sulfate and potassium are almost on the diagonal line, boron is slightly above it, sodium and chlorine are substantially above it (suggesting that outcroppirig halite beds were being dissolved), and carbonate is well below it. Inasmuch as no deeper beds of salts exist in the upper part of Owens Lake, the exposed salts in Owens Lake appear to be the residue of about 2,000 years of evaporative concentration, and the salts contributed prior to that time were apparently carried downstream. Gale (1914, p. 264) and Antevs (1936. 1938) calculated the quantities of sodium and chlorine in Owens Lake and, from those data, concluded that the time since the last overflow (for saline deposition) was nearer 4,000 years. The evidence in the present paper indicates that sulfate and potassium are more reliable components for this purpose and suggests that the age of 2,000 years is more nearly correct. TABLE 43,—Artolyses for calcium, fluorine, and other elements in waters from Owens River and Hot Creek, and the ratio of the chemical activity prodmt (AP) of Ca++ and F ‘ in these waters to the equilibrium constant for fluorite (K T) [Analytical data from Wilcox, 1946, tables 1 and 8, values in parts per million;,AP/KT ratios computed by E. A._]enne using procedures described by Nordstrom and Jenne (1976). 1.eaders(---_) indicate no data] Chemical analyses AP/KT . (PPm) (fluorite, Can) Sampling (1 =saturation ‘ dischar e C° under equilibrium Station Year Date (g/d X 1 “') pH Ca Mg Na K 504 C1 B F (est: 10°) conditions) Owens Riverl ________________ 1935 10/29 80.4 7.9 21.2 6.2 47.59 ____ 18.73 25.53 0.80 0.6 20 0.035 1945 10/4 71 7.9 33.3 28.1 79.08 11.34 29.30 58.50 1.48 .6 20 .045 _-.. _--_ ___. ____ ____ _.__ ____ ____ ___- ____ 25 .038 Station 2 Hot Creek2 __________ 1933 10/3 5.6 7.7 12.2 7.9 34.48 4.30 12.97 13.48 .39 .4 10 .01 1933 11/6 5.3 7.7 14.0 10.7 25.75 11.73 15.85 10.28 .40 .5 8 .025 1934 3/16 6.9 8.1 13.4 9.7 29.43 ___, 19.69 10.28 .34 .2 7 .004 1934 4/9 6.5 7.9 13.0 7.1 27.82 6.65 11.05 11.70 .34 .2 7 .004 1934 5/7 7.6 7.8 13.6 8.5 16.55 _-._ 9.61 9.93 .49 .4 9 .016 1934 9/17 4.1 7.7 11.6 7.7 32.41 __._ 9.61 10.28 .43 .2 15 .0027 .___ __,_ ____ ____ ___. ___. .___ "-1 .___ ———— ———— 20 .0023 Station 4 Hot Creek3 __________ 1933 10/3 881 677 13.2 7.3 125.06 12.12 27.38 64.18 2.86 2.6 10 .52 1933 11/6 7.9 6.9 16.2 7.7 118.16 12.12 33.14 67.38 2.98 3.0 8 .90 1934 3/16 9.8 7 2 14.2 9.7 100.92 ____ 40.83 59.22 2.47 2.2 7 .44 _.._ .___ -_.. ____ ____ ____ -_-_ ____ __- .___ 17 .31 1934 5/7 10.0 18.2 .9 93.10 ____ 15.85 51.77 2.35 2.4 7 .71 1934 9/17 6.9 7 4 12.2 5 3 132.41 _.__ 35.06 74.82 3.38 3.2 9 .75 --—— ———— —-—— ———— --—— ———- -——— ——-- ———— -——— —-—- 19 .54 ‘At aqueduct intake. ’Station 2 'is upstream “Station 4 is downstream from major thermal springs. from the major thermal springs in canyon, but downstream from Casa Diablo hot spring and other thermal areas. 100 the age of the volcanic rocks involved in the Long Valley caldera. Minimum reasonable estimates of the time at which the associated thermal springs are likely to have been initiated are in excess of 100,000 years, and the best estimates are nearer 300,000 (Bailey and others, 1976). The 32,000 years of spring discharge that can be accounted for by the deposits in Searles Lake leaves about eight times that amount unaccounted for. It seems unlikely that existing subsurface data could have overlooked one or more deposits of this size, although obviously it is not impossible. Alternative explanations that would provide for a smaller mineral discharge from the caldera area during its first 270,000 years of cooling and much more intense discharge thereafter seem worth exploring. _ The third line of evidence indicating that streams dis- charging from the Long Valley area carried most of the valuable components now found in Searles Lake comes from the similarity in the ratios of spring-derived ele- ments in various parts of the Owens River (table 44). Changes occur, but they are mostly small or occur for reasons that can be demonstrated or are plausible. Be— tween Hot Creek and Lake Crowley, as a result of mix- ing with waters from the upper Owens River and Mono Basin, the Na/K and Cl/SO4 ratios are reduced by almost half because of the low ratios (5.9 and 0.33) in those waters (Eccles, 1976). The K/B ratio increases because of mixing with waters having ratios near 8. The Cl/B ratio increases for reasons that are unclear because the added water has a ratio near 17; possibly some Cl is added by drainage from other tributaries. Ratios for B/Li and F/Li increase slightly, possibly because some Li is adsorbed by the muds of Lake Crowley. Between Lake Crowley and Haiwee Reservoir (10 km south of Owens Lake), the Na/K ratio stays nearly the same while the Cl/SO4 ratio decreases, possibly because gypsum is being dissolved from the upper Cenozoic lake deposits that lie east of this segment of the river. The constant Na/K ratio, combined with small increases in the Cl/B and K/B ratios, suggests that proportional amounts of Na, Cl, and K are added without propor- tional amounts of B. This, combined with the fact that TABLE 44.—Comparison of ratios of selected ions in waters from Long Valley area, Owens River, and Searles Lake Owens River Hot Creek and/(or) Searles Lake Ratio gorge‘ Lake Crowley‘ L.A. aqueduct2 Upper Salt3 16.7 8.5 7.6, 10.0 l6.5 2.2 1.4 .8, .5 2.3 21.4 27.3 36.4 56.7 2.3 6.1 8.8 3.5 4.6 5.1 3.0 132 4.2 4.6 4.8 .24 ‘Data from table 39. 'Data from table 41, columns 2 and 3; both ratios listed where possible. 5Data from table 42, and Smith (1973, and unpub. data, 1976). LITHIUM RESOURCES AND REQUIREMENTS BY THE YEAR 2000 the F/Li ratio stays about the same while the B/Li ratio decreases, suggests the enroute addition of small amounts of Li and F from springs. Data plotted in figure 45 suggests that most of the boron, sulfate, and potassium in the lower Owens River reached Searles Lake. Changes in the ratios that include these elements are explained using this assumption. In- creases between the Owens River and Searles Lake in the Na/K, Cl/SO4, and Cl/B ratios thus suggest that addi- tional sources of NaCl were draining into the river dur— ing pluvial periods; this explanation is consistent with the surplus of Na, Cl, and Br indicated by figure 45 and discussed previously. The decrease in the K/B ratio is consistent with the earlier conclusion that a little K was lost and that some B was added, but that the changes in relative total amounts were not large. The notably large increase in the B/Li ratio indicates a major loss of Li. The change in ratios suggests that about 98 percent of the amount introduced into the first lake of the once-connected chain was lost during transit. (The data in table 42 andfig. 45 suggest that 95 percent was lost.) The decrease in the F/Li shows that even more F than Li was lost and that only about 0.1 percent of the F originally contributed by the Owens River reached Searles Lake. (Fig. 45 suggests that about 0.3 percent of the F reached the salt lake.) PRESENT RELATION BETWEEN RUNOFF AND BORON AND CHLORINE CONTENTS The conclusions of this paper depend in part on the likelihood that the annual tonnages of components de— rived from thermal springs remained nearly constant during times when flow in the Owens River exceeded that of the present by as much as an order of magnitude (Smith, 1976). The evidence of this constancy comes from the low to negative correlation between discharge of both the Owens River and Hot Creek and from the relative long-term constancy in the tonnages of boron and chlorine contained by them. Table 45 lists the annual discharge of the Owens River and the tonnages of boron and chlorine dissolved in it during the period 1929—45. During the 17 years of sam- pling (about 350 samples), the daily discharge varied by a factor of about 32; the concentration of boron in water samples varied by a factor of 7, and the estimated boron tonnage per day varied by a factor of almost 40. The tonnages of boron and chlorine carried during a full year, however, varied much less. During the 17-year period, the quantity of boron3 varied from 150 t/yr to 1These values for annual quantities were derived by plotting each of the samples listed by Wilcox (1946, table 1) at the appropriate point on the horizontal axis of 365=19.76). In this case lithium is considered to display a pseudolog- normal distribution. Although Link and Koch (1975) have discussed some of the consequences of making the lognormal transformation on pseudolognormal popula- tions, the analytical values were transformed into natural logarithms in order to facilitate graphic display of the data and to present some of the statistical tests discussed. The reader should be cognizant of the possi- ble bias introduced by this transformation. Figure 47 illustrates the frequency distribution of the 156 playa samples, and figure 48 is a similar diagram of the 41 playa mean values. The range in lithium values for the 156 playa samples is 10460 ppm; the arithmetic mean is 99 ppm; and the mean of the natural logarithms is 74 ppm (table 47). These estimates of the population mean as well as estimates of the standard deviation are frequently used in geochemical exploration to define anomalous values; there is 97.7 percent probability that a random sample from a normally distributed popula- tion will be less than p.+2o: Table 48 lists several esti- mates of this value. Caution should be exercised in using this type of geo- chemical reconnaissance as an exploration tool. In addi— tion to the bias introduced by the departure from a normal distribution, a large degree of variation was probably introduced by the lack of control over the loca- tion of the sample site and the mineralogical composi- tion of the individual samples. Furthermore, there is as yet no clear understanding of the relationship between the lithium content in playa sediments and the concen- tration of lithium in subsurface brines. However, the facility with which sediment samples can be collected from playa surfaces relative to the difficulty of sampling 30F I I I ~ : '5 25» : U I ° : 5 l E; 20— : I >' ' n=156 "z’ i '5' 15— I o ”.5 E «:1 t ,0_ ”1:1; In 1 > I I: l 3 5- : m I 0‘ I 1 I l 'I l I I I 2.4 3.2 4.0 4.8 5.6 6.4 LN LITHIUM (PPM) FIGURE 47.—Frequency histogram of the lithium concentration of 156 playa samples. LITHIUM RESOURCES AND REQUIREMENTS BY THE YEAR 2000 35— 30— 25— n=4| 20-— RELATIVE FREQUENCY (PER CENT) 3‘42- 2.4 3.2 4.0 LN LITHIUM (PPM) I l I 4.8 5.6 FIGURE 48.—Frequency histogram of the mean values of the lithium samples. TABLE 47,—P0pulati0n statistics of lithium analyses X , S , _ S In, arithmetic arithmetic X In, standard Sample mean standard mean of natural deviation of populailon (ppm) deviation logarithms natural logarithms (ppm) 156 playa samples ______ 99 86 4.31 0.75 (74 ppm) 41 playa means ________ 90 76 4.24 .73 (69 ppm) TABLE 48.—Stati.ttieal estimates of a lithium anomaly Sample _ _ population X +2; X1" +2811" 156 playa samples ____________ 271 ppm 5.81 (334 ppm) 41 p aya mean values __________ 242 ppm 5.70 (299 ppm) interstitial brines somewhat compensates for the disad- vantages of the method. AREAL FACTORS INFLUENCING THE DISTRIBUTION OF LITHIUM The suite of sediment samples collected in the recon- naissance was examined statistically in order to better understand the distribution of lithium in closed basin sediments. If the size of the drainage basin is an impor- tant factor in determining the amount of lithium intro- duced into the playa by weathering of the bedrock, there may be some relationship between drainage basin area and the lithium concentration in the playa sedi- ments and interstitial waters. To assess this relationship, the natural logarithm of the area of each basin was plot- 107 ted against the natural logarithm of the mean lithium value for that playa (fig. 49), and Pearson’s product- moment coefficient of linear correlation was calculated. The value r=—0.1788 indicates no significant correla- tion at the 95 percent confidence level (Till, 1974, p. 83—87). The lack of a significant correlation in the relationship between lithium and drainage area does not necessarily preclude the importance of areal factors in the distribu- tion of lithium. Many of the playa basins studied had different drainage areas during Pleistocene pluvial periods; some basins had pluvial lakes which spilled into adjacent basins. The expanded drainage area of basins which received pluvial overflow could be a significant factor in the accumulation of the unusual concentra- tions of soluble salts found in some of these basins. When the drainage basin areas were modified to allow for pluvial spillover (Snyder and others, 1964) and plot- ted against the mean lithium values (fig. 50), the correla— tion coefficient, r=0.3493, indicates significant correla- tion at the 95 percent confidence level; however r2=0.12 indicates that only 12 percent of the total variance ob- served in the lithium distribution is explained by the 6- O — O O O . O O 5—. .0 . A O z _ . ' 2 CL 2 .0... ... .0 D 4_ O. O O E o I: . - . " _ . '. r=-O.|788 z 0 O -' . n=4| 3.. V O _ O 2 | I | l | | 6 7 8 9 IO ll l2 LN BASIN AREA (KMZ) FIGURE 49.—Plot of lithium values versus drainage basin area. 108 6r— 0 O E O. 5 2 2 I t -' 0. n=24 3 1" “0.05 3— O _ O 2 l I I I I I 6 7 8 9 IO II I2 LN BASIN AREA (KM2) FIGURE 50.—Lithium versus drainage basin area. Area simulates plu- vial spillover. linear relationship between lithium concentration and area. In expanding the concept of areal influence on lithium distribution, groundwater movement can also be considered. Some data are available which indicate prob- able hydrologic underflow in many of the basins studied in this reconnaissance (Hunt and others, 1966; Walker and Eakin, 1963; Rush and others, 1971). When the drainage basin areas were modified to simulate suggested hydrologic underflow and plotted against lithium concentration, the correlation coefficient in- creased to r=0.48 (fig. 51). The coefficient 72:02?) for this improved correlation indicates that 23 percent of the variance in lithium concentration can be accounted for by the linear relationship between area and lithium concentration. This rudimentary study was based upon the assump- tion that drainage basin size is the only factor influenc- ing the weathering of bedrock and the subsequent transportation of ions to the playa. The significant cor- relation between drainage basin area and lithium con- centration in the playa suggests that weathering of bedrock does indeed contribute lithium to the playa sed- LITHIUM RESOURCES AND REQUIREMENTS BY THE YEAR 2000 6_ LN LITHIUM (PPM) 2 I l | I I I 6 7 8 9 IO II l2 LN BASIN AREA (KM2) FIGURE 51.—Lithium versus drainage basin area. Area simulates hy- drologic underflow. iments. The importance of bedrock weathering as a lithium source might be clarified by considering other factors which could influence the amount of lithium reaching the playa; the chemical composition of the bedrock, as well as factors which influence weathering such as relief, slope, climate and vegetation, and factors influencing surface flow such as flow frequency, runoff, infiltration, and evaporation may significantly affect the amount of lithium reaching the playa. When the role of these factors are more thoroughly understood, it will be possible to evaluate more accurately the importance of regional sources of lithium which are not related to drainage basin morphology or process. REFERENCES CITED Hunt, C. B., Robinson, T. W., Bowles, W. A., and Washburn, A. L., 1966, Hydrologic basin, Death Valley, California: US. Geol. Sur- vey Prof. Paper 494—B, 138 p. Link, R. F., and Koch, G. S.,Jr., 1975, Some consequences of applying lognormal theory to pseudolognormal distributions: jour. Inter- nat. Assoc. Math. Geology, v. 7, no. 2, p. 117—128. Rush. F. E., Scott, B. R., Van Denburgh, A. 8., and Vasey, B.J., 1971, Water resources and inter-basin flows, Map 8—12 in Water for LITHIUM RESOURCES AND REQUIREMENTS BY THE YEAR 2000 Nevada—Hydrologic Atlas: Carson City, Nevada, Nevada Div. Water Resources. Snyder, C. T., Hardman, George, and Zdenek, F. F., 1964, Pleistocene lakes in the Great Basin: U.S. Geol. Survey Misc. Geol. Inv. Map 109 Till, Roger, 1974, Statistical methods for the earth scientist: New York, John Wiley 8c Sons, 154 p. Walker, G. E., and Eakin, T. E., 1963, Geology and ground water of Amargosa Desert, Nevada-California: Nevada Dept. Conserv. and I—416. Nat. Resources, Ground-Water Resources—Reconn. Ser. Rept. 14, 58 p. N THE TECTONIC AND SED MENTOLOGIC ENVIRONMENT OF LITHIUM OCCU RENCES IN THE MUDDY MOUNTAINS, CLARK COUNTY, NEVADA By ROBERT G. BOHANNON, US. GEOLOGICAL SURVEY, DENVER, CO _l_— ABSTRACT In the vicinity of the Muddy Mountains of southern Nevada, ihe Horse Spring Formation of Miocene age and the Muddy Creek For- mation of Miocene(?) and Pliocene age contain various saline depoisits that are rich in lithium as well as other alkali metals and alkal ne earths. The rocks of the Muddy Mountains have been involved in several episodes of deformation. Thrust faulting occurred during Cie- taceous time and strike-slip faulting during Tertiary time. However, normal faulting in the Tertiary and Quaternary that is characteristic‘ of basin and range deformation appears to have controlled the depd tion of the Cenozoic rocks. Lithium concentrations greater than 1,000 ppm occur in rocks ini at least three geographic areas: West End Wash, White Basin, and Over- ton Ridge. High concentrations of lithium are associated with cla s, borate mineralization, and magnesite deposition. In West End W sh and White Basin, lithologies and sedimentary structures suggest fiat the Horse Spring Formation and the borate minerals within it accu u- lated in a saline lake or hot spring environment. At Overton Ridge the magnesite member of the Horse Spring Formation probably formled in a similar environment, and associated coarse conglomerate within the Horse Spring indicates a nearby source of elastic material. There is evidence that the Horse Spring Formation was depositedi in a single, closed basin that may have covered much of the Muddy Mountains and the southern part of the Virgin Mountains. The de 0- sition of the Muddy Creek Formation appears to have been confin d to valleys presently in existence in southern Nevada. ‘ si- INTRODUCTION . A varied section Of upper Tertiary and Quaternaiy sedimentary rocks rims the west and south sides of the Muddy Mountains in southern Nevada (fig. 52); it is rich in lithium, magnesium, sodium, boron and manganese. This section, which contains abundant clay—rich clastiic material, magnesite, dolomite, limestone, gypsum, hali e and volcanogenic sedimentary rocks probably origi- nated in a large, closed bolson because many of the d - posits are characteristic of alluvial fans, playa lakes and river flood plains. Longwell (1928, 1949) and Longwell, Pampeyan, Bowyer, and Roberts (1965) subdivided the Tertiary rocks in some places, into the Miocene Horse Spring and Miocene(?) and Pliocene Muddy Creek Formations, but they included equivalents of these units and some Older rocks in the Cretaceous(?) and Tertiary " Salt oGreat Salt L... ”\5‘ Lari; EXPLANATION I Desert _ » Thrust fault | L\ 40°— .l—Jl Basin and Range I margin l —-—— Fault. zone Mountain I I <9 I / a, l , ¢ | v Q N E V A D A | U T A H Nevada 7 _ — Test Site __—__ ________ 310‘ A R | Z O N A A 1. 7 u \ _ “22,9 . «“2 35 ~ Lyles Lithium Deposit CALIFORNIA (Hectorite Locallty) .Fhoenix Plcacho Basin — — - d — — 1 0 I50 Kllometres 0 I50 Milu FIGURE 52.—-—Index map of the major structural features in the vicinity of the Muddy Mountains, Nev. Square in left center of the diagram outlines the Muddy and Virgin Mountains and Frenchman Moun- tain. DV, Detrital Valley; RL, Red Lake. 110 Gale Hills and Thumb Formations in other places. Al- though geologic mapping is incomplete, much of the enrichment in light metals and boron and most of the exotic lithologies appear to be part of the Horse Spring Formation; the more uniform, clastic Muddy Creek Formation contains local pockets of manganese (McKel- vey and others, 1949). Because of the enrichment in lithium and other light elements within the Tertiary sedimentary section in the vicinity of the Muddy Mountains, a detailed study of these rocks is underway. These rocks are well suited for a study of lithium and its behavior in a nonmarine sedimentary environment because they are well ex- posed. Hence a good three-dimensional view of the rocks is possible, but complex geologic structures must be understood first. Once the details of the structural geology are ironed out, geochemical sampling of all the different sedimentary facies and environments can yield information relating to the localization of lithium, the chemical and mineralogical associations of lithium, the most likely sedimentary environments for lithium ac- cumulation, the possible sources of lithium, and any necessary tectonic or volcanologic controls on lithium accumulation. This paper marks the beginning of such a long—range study; and, although the work isjust in the stage of geologic mapping, the major structures and Tertiary sedimentary environments of the area can be described and preliminary geochemical data discussed. ACKNOWLEDGMENTS This paper has benefited from many discussions with geologists associated with the US. Geological Survey’s lithium program, and most of the chemical data refer- red to in the report were supplied by Allen Meier. Chemical analyses for lithium were performed in the field using an atomic-absorption apparatus. Powdered solid samples were dissolved in hydrofluoric acid and nitric acid and the solution taken to dryness. Hy- drochloric acid was used to redissolve the remaining res- idue; distilled water was employed to dilute that solution. TECTONIC SETTING The Muddy Mountains occur where several major tectonic elements of differing ages come together (fig. 52). A late Mesozoic zone of thrust faults and folds, the Sevier Orogenic belt (Armstrong, 1968), extends from northern Utah to southern California and includes most of the Muddy Mountains. The Glendale thrust and the Muddy Mountain thrust (Longwell, 1949) are the two principal low-angle faults resulting from this deforma- tion. Also, the Walker Lane (Locke and others, 1940), which trends northwest through most of western and southern Nevada, terminates in the vicinity of the LITHIUM RESOURCES AND REQUIREMENTS BY THE YEAR 2000 Muddy Mountains. Albers (1967) proposed that the right-slip Las Vegas Shear Zone (Longwell, 1960) is part of the Walker Lane—which he claimed is as old as juras- sic but was active through the Miocene—and that the Muddy Mountains are a giant right-lateral “oroflex.” Anderson (1973) however, has documented several left strike-slip faults north of Lake Mead within the right- lateral “orofiex”, as proposed by Albers (1967), and these left—slip faults appear to be of about the same magnitude as the Las Vegas Shear.Zone (Stewart and others, 1968). Also, Fleck (1970) concluded that the Las Vegas Shear Zone is no older than 10.7 million years, or about the same age as Anderson’s (1973) left-slip faults. More research is needed in order to resolve some of these conflicting details of the structural setting of the Muddy Mountains. Several other major tectonic features also occur in the vicinity of the Muddy Mountains, and these may have a direct bearing on lithium accumulation in the area. The Wasatch Line, a term commonly applied to the zone that divides the thick “geosynclinal” Paleozoic sedimentary sections in the western Cordillera from the thinner cra— tonal sections in the eastern Cordillera and on the Col- orado Plateau, passes through the Muddy and Virgin Mountains. In the Muddy Mountains, this transition takes the form of a major thrust fault—the Muddy Mountain thrust, which juxtaposes the thick Paleozoic carbonate rocks characteristic of south—central Nevada on top of the thinner carbonate and red bed sections characteristic of the Colorado Plateau. In central Utah, the Wasatch Line roughly corresponds to the tectonic transition from the Basin and Range Province into the Colorado Plateau, but in southern Nevada and north- western Arizona, this transition is about 70 km east of the faulted sedimentary facies change in the Muddy Mountains. The Muddy Mountains, then, occur within the Basin and Range Province but are near the margin of the Colorado Plateau. Other areas of anomalous lithium accumulations and thick evaporite deposits, mentioned below, share the above characteristics. LITHIUM AND THE TERTIARY ROCKS Geologic mapping is incomplete in the Tertiary rocks of the Muddy Mountains; thus, subdivisions within these rocks are poorly defined and understood. Several rock units have been named (Longwell, 1949; Longwell and others, 1965; Anderson, 1973); these include the Horse Spring (Miocene) and Muddy Creek (Miocene? and Pliocene) Formations, Gale Hills and Thumb For- mations (Cretaceous? and Tertiary) and the Overton Fanglomerate (Tertiary). The Baseline Sandstone and Willow Tank Formation have been referred to as Cre- taceous (Longwell, 1949), but may be, in part, Tertiary as well. Lithium occurs in abundance throughout all the LITHIUM RESOURCES AND REQUIREMENTS BY THE YEAR 2000 above-named rocks, with the possible exceptions of the Willow Tank and Baseline; but it is most concentrated in three areas: the magnesite near Overton, the White Basin borate area, and the West End Wash borate area (fig. 53). All three areas are thought to be part of the Horse Spring Formation. The two borate areas appear to be similar in their overall lithology. In addition to the presence of borate minerals, both areas contain sections of well—bedded, predominantly clastic rocks that have abundant sedimentary structures suchtas ripple marks and small- scale crossbedding. In both areas, carbonate beds, gyp- sum veins, and claystone are common as well. Although no detailed geologic mapping or sedimentological study has been conducted in either area, the lithologies, bed- ding, mineralogy, and sedimentary structures that occur in both areas suggest a saline lake and (or) spring envi- ronment for their deposition. Reconnaissance surface sampling was conducted at “4°45 ”41230' EXIPLANATION J"??— D \3 .. $- 2 Cenozoic rocks ORTH \SL ? z \\\\\\\\\ “Mn \ ’ MOUN TA I N s Overfono Tertiary volcanic \ rocks .1 36°_ ‘ Area of 30' m figure 54 AGNE Mesozoic and LOCALI Paleozoic rocks 36°_ is' M io MILES o 5 IO KILOMETRES FIGURE 53.—Map showing the distribution of three lithium-rich areas in the Muddy Mountains area. Geology modified from Longwell Pampeyan, Bowyer, and Roberts (1965). 111 both the West End Wash and White Basin localities, and some drill cuttings were sampled from the White Basin area. The results of lithium determinations on the sur- face samples are summarized in tables 49 and 50; the average lithium concentration from 194 samples taken from five drill holes in the White Basin is 474 ppm. The highest lithium value from the drill cuttings is 1,600 ppm and the lowest is 22 ppm. Most of these values, TABLE 49.——Lithium analyses of the West End Wash surface samples in the Muddy Mountains Sample Nor Description Li(ppm) MAI 161 ______ Carbonate rock ______________________ 230 162 ______ Clay ________________________________ 590 163 ______ Borate ore __________________________ 600 164 __________ do ________________________________ 560 165 ______ Channel near ore zone ________________ 660 166 ______ Channel 1.8 m above ore ______________ 730 167 ______ Ma nesite(?) __________________________ 360 168 __________ o ________________________________ 420 169 ______ Colemanite(?) ________________________ 1 10 170 ______ Clay debris __________________________ 470 171 ______ Clay and magnesite __________________ 670 172 ______ Chert ______________________________ 67 173 ______ Shaley zone __________________________ 700 174 ______ Clay ________________________________ 540 175 ______ unknown ____________________________ 410 176 ______ Volcanic ash __________________________ 130 177 __________ do ________________________________ 72 178 __________ do ________________________________ 130 179 ______ Clay with borates ____________________ 750 180 ______ Cla ________________________________ 470 181 ______ Chert(?) from biotite ash ______________ 45 182 ______ Clay ________________________________ 830 183 __________ do ________________________________ 770 184 ______ Carbonate ____________________________ 1 l 185 __________ do ________________________________ 33 TABLE 50.—Lithium analyses (f the White Basin surface samples in the Muddy Mountains Sample No. Description Li(ppm) MAC 617 ______ Clay ________________________________ 300 618 __________ do ________________________________ 800 619 ______ G psiferous clay ______________________ 220 620 ______ Clay ________________________________ 1 10 621 ______ Sandstone ____________________________ 130 622 ______ Cla ________________________________ 760 623 __________ dlo ________________________________ 670 624 __________ do ________________________________ 680 625 ______ Siltstone ____________________________ 100 626 ______ Clay ________________________________ 330 627 ______ Clay and borate ______________________ 420 628 ______ Cla stone ____________________________ 330 629 __________ o ________________________________ 1,250 630 ______ Siltstone ____________________________ 1,000 631 ______ Claystone ____________________________ 400 632 __________ do ________________________________ 750 633 ______ Siltstone ____________________________ 540 634 ______ Clay gouge in sandstone ______________ 250 635 ______ Concretion __________________________ 130 636 ______ Borate concretion ____________________ 120 637 ______ Tailings ______________________________ 540 638 ______ Volcanic ash __________________________ 550 639 ______ Fault-zone clay in borates ______________ 1,000 640 ______ Tailings ______________________________ 920 641 ______ Gy sum and clay ____________________ 540 642 ______ Ca careous sandstone ________________ 160 643 ______ Tailings ______________________________ 450 112 which are anomalously high with respect to crustal aver- ages, are based on the value of 20 ppm reported for the upper crust by Heier and Billings (1970), and many val- ues seem to be much higher than average for most nonmarine basins (I. D. Vine, oral commun., 1975). Values above 1,000 ppm are among the highest re- ported in sedimentary rocks (pure hectorite runs about 3,000 ppm Li). On the average, fine-grained clastic rocks such as claystone and siltstone contain the highest concentrations of lithium, but the borate zones them- selves are also rich in lithium. There does not seem to be a noticeable trend of lithium values with depth in the drill holes. Recent geologic mapping in the vicinity of the Over- ton magnesite deposits shows several changes from the original map of Longwell (1949) (fig. 54). The older conglomerate stratigraphically below and west of the conglomerate on Overton Ridge is not a laterally con— tinuous unit; but rather the conglomerate interfingers to the south with the quartz arenite of the Baseline Sandstone (fig. 54) and is the same age as the Baseline Sandstone. In this report, the above-mentioned unit is retermed conglomerate and is assigned to the Baseline Sandstone as the Overton Conglomerate Member. Also, the Horse Spring Formation has been divided into four informal members: the conglomerate member, the dolomite member, the magnesite member, and the sandstone and claystone member. The younger con- glomerate that makes up Overton Ridge is conformable with and is mapped as the basal member of the Horse Spring Formation (fig. 54). This same unit was formerly included with the conglomerate and quartz arenite below it as part of the Overton Fanglomerate (Longwell, 1949), but it rests above these older rocks with a slight angular discordance, and there is evidence of channel— ing at its base. The Muddy Creek Formation has been divided into two members: The conglomerate member, which only occurs near Overton Ridge; and the sandstone member, which occurs east of the conglomer- ate member. The general stratigraphic relationships found in Magnesite Wash are depicted in figure 55. Data on the age of the rocks has been supplied by An- derson, Longwell, Armstrong, and Marvin (1972) and Longwell, Pampeyan, Bowyer and Roberts (1965). Figure 55 also shows the averages of analyses for lithium in several of the units in Magnesite Wash. Lithium values in the dolomite member of the Horse Spring Formation are averaged from 62 samples with a high value of 580 ppm and a low value of 24 ppm. In the magnesite member of the Horse Spring, they are aver- aged from 118 samples with a high of 1,400 ppm and a low of 90 ppm, and in the sandstone and claystone member of the Horse Spring, the average is from 26. samples with a high of 850 ppm and a low of 90 ppm. The average lithium value in the Muddy Creek Forma- LITHIUM RESOURCES AND REQUIREMENTS BY THE YEAR 2000 tion is from seven samples that range in value from 75 to 190 ppm. The highest concentrations of lithium occur in the magnesite member of the Horse Spring Formation, where maximum values above 1,000 ppm occur in sev- eral beds. This member appears to have accumulated in place as a sedimentary deposit. The lines of evidence supporting this conclusion are the well-preserved bed- ding and other sedimentary structures such as ripple marks; the fact that the member occurs as a discrete sedimentary body; the fact that the relative proportion of magnesite to other minerals is controlled by bedding; and the fact that beds of tuff occur interbedded in the member and are unaltered. Lithium also appears to have accumulated with the sediments, because lithium concentrations show strong bedding control. The dolo— mite member of the Horse Spring Formation has slightly lower lithium values and appears to be, in part, a clastic carbonate rock. X-ray analysis shows that both dolomite and calcite occur in this member and thin— section analysis indicates the presence of angular, sand- sized grains of carbonate in a micritic carbonate matrix. It is likely that the clastic grains are calcite; the matrix is dolomite. Lithium concentrations are probably low owing to “diluting” of the material that formed in place by lithium-poor clastic grains. This idea, however, is not documented. Clay occurs in all the members of the Horse Spring and Muddy Creek Formations. In the magnesite member of the Horse Spring Formation the clay is a trioctahedral smectite (Harry Starkey, written commun., 1976) that contains as much as 3,300 ppm lithium. The coarse conglomerate interbedded with the fine- grained rocks of the Horse Spring Formation suggests that a nearby source of clastic material existed for the Horse Spring Formation in this area and that the magnesite may have formed near the edge of the Horse Spring basin. Although no detailed sedimentologic work has been attempted on this conglomerate, derivation from a terrane of carbonate rocks can be assumed on the basis of the abundance of carbonate clasts; the most logical nearby source is a carbonate terrane to the west in the north Muddy Mountains. That the Horse Spring basin east of this probable source was a closed basin is suggested by the widespread occurrence of gypsum in the Horse Spring Formation and by rock salt, which probably also originated in the Horse Spring Formation (Mannion, 1963). CONCLUSIONS Rocks called the Horse Spring Formation have been mapped in many areas in southern Nevada. These rocks are also described in the Virgin Mountains, east of the Muddy Mountains (Morgan, 1964) and in the Nevada LITHIUM RESOURCES AND REQUIREMENTS BY THE YEAR 2000 II4°29‘ 113 36f 3| 0 o——o l KILOMETRE EXPLANATION CORRELATION 0F MAP UNITS Holocene UNCONFORMITY UNCONFORMITY Pliocene and Miocene (?) } Pleistocene Miocene ., > QUATERNA RY r TERTIARY } CRETACEOUS LIST OF MAP UNITS Alluvium (Holocene) Terrace deposits(Pleistocene) FIGURE 54.~Geologic map of the area west of Overton, Nev. Geology from this report. Test Site, 100 km northwest of the Muddy Mountains (Sargent and others, 1970; Sargent and Stewart, 1971). In the Nevada Test Site, however, Sargent, McKay, and Burchfiel (1970) reported that the Horse Spring Forma- Muddy Creek Formatioanliocene and Miocene ? ) Conglomerate member Sandstone member Horse Spring Formation(Miocene) Conglomerate member Sandstone member Dolomite member Baseline Sandstone (Cretaceous) oaoucoo ohm” Overton Conglomerate Member Sandstone member Contact--Dashed where approximately located; dotted where concealed Fault—-Dashed where approximately located; dotted where concealed i Strike and dip of bedding tion is about 29 m.y. old which is older than the date of 21 m.y. reported for it in the Muddy Mountains (Ander- son and others, 1972). Also, Will Carr (oral commun., 1974) suspects that the Horse Spring Formation in the 114 LITHIUM RESOURCES AND REQUIREMENTS BY THE YEAR 2000 ,. m z E >- g E (I) Q m 8 34 a: La (n < Lu l: _ '5: Lu _l < >2 a) m a: E 2 o a: a. 2 DESCRIPTION >. uJ x I MEG. m (D a: o I— >3 ""' o — — <— 2 u. I _l I I— t .J a) C I: .2 8 *5 Sandstone Sandstone (Iithic—arkose), claystone, and siltstone .9 E with parallel bedding 2-30 cm thick. — \_ o g L S x (r: |25 a) a 1’ A v' o 2 w 5‘ 8 ‘3 E 2 Conglomerate Lenticularly bedded conglomerate with sandy matrix. )- CC < Conglomerate with poorly defined lenticular beds. Clasts — o are subround to subangular, are clast supported, and l— ‘E Conglomerate composed of carbonate rocks. The largest clast is l5 cm c A in diameter. The matrix is composed of sand and a: '9 contains crossbedding. L” ‘6 l— E 8 LL 2 8 or Sandstone Sandstone and shale. Sandstone is most abundant near the o .E (:3 307 and base of the unit and contains parallel bedding 5mm to E 5- Claystone 25 cm thick. (I) 3 a Fine-grained and parallel-bedded magnesite and dolomite. I 3 372 Magnesite Magnesite and dolomite are interbedded. Ripple marks occur on the tops of some beds. 3 I32 Dolomite lnterbedded dolomite and limestone. g Conglomerate Conglomerate With poorly defined, lenticular bedding. Carbonate clasts are set In a carbonate matrix. a) . U) 8 w B 2 8 Quartz arenite and subarkose with some conglomerate 2 gray-brown argillaceous limestone .8. ‘6 o a, g Gray-black hard shale above gray-brown dense, 32 ~ to fine crystalline sometimes oolitic and g ‘5’, - pelletoidal limestone with subordinate SE - gray limy sandstone -_|. I I I C 5 ,9 ‘lvvljll Fine white sucrosic anhydrite with subordinate 5 ‘6 twin. tan and gray, dense, sandy, and often oolitic g E 1.1521; limestone interbedded with dolomite. '3 1?. “1.51,, Massive bedded anhydrite base task“ ul’wl' \ls’v/v I I r I I I, I 5 f—_’ Gray-brown dolomite, good oolimoldic framework. 8 15 I I Dense oolite, large oolimolds often graded to '25 0 I I fine crystalline drab dolomite. Dork gray-brown E g I I limestone with interbeds of black calcareous shale. V’ E, I I Silty limestone with admixture of siltstone and o I I dark carbonate Fine, silty, argillaceous, orange sandstones Nor phlet 1" Salt _"—Formation_' Formation /// c /// I: // . a g /// Clear crystalline hallte p /// ..J /// //// /// +++-L ‘- g + 4- 4- 0r— + 4_ C ‘- . . . 5-, g ”Ct—tat White to brown masswe anhydnte ; 5 + +++ -l- “- -+ -l-+ I. + 4— FIGURE 58,—Generalized section of jurassic and Jurassic(?) rocks in northeast Texas showing types of lithologies in each unit. The units are not to scale. LITHIUM RESOURCES AND REQUIREMENTS BY THE YEAR 2000 originally deposited with the sediments include ion ex- change, infiltrating waters, sediment leaching, mineral formation, sulfate reduction, and ultrafiltration through clay-shale membranes (Collins, 1974). Most of the Smackover brines are defiicient in mag— nesium relative to an evaporite-formed brine. Table 52 illustrates the approximate amounts of calcium, mag- nesium, bromide, and sulfate that could exist in a water before and after precipitation of gypsum. Assuming that the residual sulfate 1,644 mg/l was re- moved by the dolomitization reaction, MgCT2+2CaC03—>.Ga€12-+CaMg (C0302 (1) QaCf2+MgSO4—>CaSO4+Mg€12 (2) MgSO4+2CaC03—>CaSO4+CaMg (C0502 (3) then the Mg/Br ratio would be about 883/65=13.6, as illustrated by the data in table 52. However, if the residual sulfate was removed by bacterial reduction, CnHm+Na2SO4 —> NagC03+H2S+ C02‘+H20, the Mg/Br ratio would be about 1,300/65=20. Magnesium will react with CaCOg (calcite) to form dolomite; thus, the concentration of calcium is increased in the brine, and the Smackover brines are enriched in calcium. However, the total calcium plus magnesium in the brine should remain constant; this can be calculated 24.31 as 40 08 (4) xmg/l calcium+ mg/l magnesium=total equiva— lent magnesium or Mg’. The ratio Mg’/Mg will vary, depending upon the availability of calcite, and the ratio should be indicative of the degree of dolomitization. For example, brines in equilibrium with sandstones should have a relatively low Mg’/Mg ratio; those in equilibrium with dolomite should have higher ratios; and those in equilibrium with limestone should have the highest ratios. The average Mg’/Mg ratio for the Smackover brines studied is 7; this ratio indicates that the brines were in equilibrium with limestone and dolomite. Previous studies have shown that the average ratio in sandstones is about 2.5 and in limestones about 9.5 (Collins, 1975). The formation of chlorite from montmorillonite re- quires about 9.2 moles of MgO per mole of chlorite, as follows: 1.7A1203, 0.9 MgO, 88102, 2H20+9.2 MgO+6H20 —> 10.1 MgO, 1.7 A1203, 6.4 5102, 8H20+1.6 5102 (5) Such a reaction can remove large amounts of mag- nesium from waters. Table 53 presents data obtained by comparing the average composition of the analyzed Smackover brines 119 TABLE 52.-—1Approximate composition of seawater changed by mineral forma- tion and bacterial reduction Bacterial Gypsum precipitation Dolomitization reduction Ion Before (in mg/l) After (in mg/l) After (mg/l) After (mg/l) Calcium __________ 390 0 0 0 Magnesium ______ 1,800 1,300 883 1,300 Bromide __________ 65 65 65 65 Sulfate __________ 2,580 1,644 0 0 TABLE 53.——Concentration ratios and excess factor ratios for some constituents in Smackover brines Average composition, mg/l Number of Smackover Concentration Excess Smackover Constituent Seawater brines ratio‘ factor“ samples Lithium __________ 0.2 174 870 18.1 71 Sodium __________ 10,600 66,973 6 .1 283 Potassium ________ 380 2,841 8 .2 82 Calcium __________ 400 34,534 86 1.8 284 Magnesium ______ 1,300 3,465 3 .1 280 Strontium _ 8 1,924 241 5 85 Barium _ .03 23 767 16 73 Boron-.. 4.8 134 28 .6 71 Copper _ __ .003 1.1 359 7.5 64 Iron ____________ .01 41 4,069 84.2 90 Manganese ______ .00 0 14,957 311 69 Chloride __________ 19,000 171,686 9 .2 284 Bromidec 6 3,126 48 1 74 Iodide- .05 25 501 10.4 73 2,690 446 .2 .003 271 1,543 24,362 16 .3 284 ‘Amount in brine/amount in seawater. 2Concentration ratio of a given constituent/concentration of bromide. 3Mg’=(24.31/40.08)> Waco 8 \l I O \ /‘ 5’ LIME ,/ ‘ _,\ STONE >’ \ K I \ \ s ‘/( o r ( LE N ‘U , 2\ , TEXAS \. LOUISIANA S v 8 \ -—.I ' 3|° ‘ '2’ A: r/ ‘ lu _ \ .. < \ 119/, l I I \v / I ' 0 I00 zoo KILOMETRES L FIGURE 61.—Map showing the relation of high lithium concentrations ,1 J (in mg/l) in Smackover Formation waters in Arkansas and Texas to the Mexia—Talco fault zone and to an arcuate contour line parallel to the truncated subcrop of the formation. effective porosity of 5 percent, and if the enclosed brine has a specific gravity of 1.2 and contains an average of 100 mg/l of lithium, then about 0.75X106 tonnes of lithium would be contained in 7.5 km3 of brine. While the total area underlain by the Smackover Formation may be as much as 10 times the area assumed in the above calculations, the other assumptions may be over- estimated. The amount of brine that would have to be pumped from depths ranging from about 1,800 to 4,800 m and evaporated nearly to dryness makes pri- mary product recovery of this lithium economically un- satisfactory with the technology available today. If an extraction technique could be developed to recover lithium as a byproduct of the recovery of petroleum, bromine, potassium, or other chemical products, the economic feasibility would be improved significantly. REFERENCES Baillie, A. D., 1953, Devonian names and correlation in the Williston Basin area: Am. Assoc. Petroleum Geologists Bull, v. 37, no. 2, p. 444—447. Bishop, W. F., 1969, Environmental control of porosity in the upper Smackover Limestone North Haynesville Field, Claiborne Parish, Louisiana, in Geology of the American Mediterranean: Gulf Coast Assoc. of Geol. Socs. Trans., v. 19, p. 155—169. Collins, A. G., 1969, Chemistry of some Anadarko basin brines con- taining high concentrations of iodide, in Geochemistry of subsur- face brines: Chem. Geology, v. 4, no. 1—2, p. 169—187. 1970, Geochemistry of some petroleum-associated water from Louisiana: US. Bur. Mines Rept. Inv. 7326, 31 p. 1974, Geochemistry of liquids, gases, and rocks from the Smackover Formation: U.S. Bur. Mines Rept. Inv. 7897, 84 p. 1975, Geochemistry of oilfield waters: New York, Elsevier, 496 p. Cummings, A. M., 1970, Lithium, in Mineral facts and problems: U.S. Bur. Mines Bull. 650, p. 1073—1081. LITHIUM RESOURCES AND REQUIREMENTS BY THE YEAR 2000 Dickinson, K. A., 1968, Upper Jurassic stratigraphy of some adjacent parts of Texas, Louisiana, and Arkansas: US Geol. Survey Prof. Paper 594—E, 25 p. Imlay, R. W., 1940, Lower Cretaceous and Jurassic formations of southern Arkansas and their oil and gas possibilities: Arkansas Geol. Survey Inf. Circ. 12, 64 p. 123 1945, Jurassic fossils from the southern states, Number 2: Jour. Paleontology, v. 19, no. 3, p. 253—276. Miodrag, S., 1975, Should we consider geochemistry as [an] important exploratory technique?: Oil and Gas Jour., v. 73, no. 31, p. 106— 110. N LITHIUM IN THE GILA CONGLOMERATE, SOUTHWESTERN NEW MEXICO By ELIZABETH B. TOURTELOT and ALLEN L. MEIER, U.S. GEOLOGICAL SURVEY, DENVER, CO ABSTRACT Reconnaissance sampling of the Gila Conglomerate of Pliocene and Pleistocene age in southwestern New Mexico revealed anomalously high lithium values in lake beds within the Gila Conglomerate. The lithium is probably derived from the volcanic rocks of the Mogollon- Datil volcanic province by postvolcanic hydrothermal activity as- sociated with hot springs. The alkaline lake environment in which the lake beds were deposited was a sink for the elements added to surface waters by hot-spring activity. Although no deposits of economic poten- tial were found, the geochemistry of the area suggests that beds of hectorite could occur within the lake bed sediments of the Gila Con- glomerate. INTRODUCTION This is a preliminary report based on reconnaissance sampling and field determinations of lithium content during a reconnaissance survey of the Rocky Mountain region for lithium in 1974 and 1975. Anomalous lithium values were found in some lake—bed deposits within the Gila Conglomerate of southwestern New Mexico. Lithium was determined by atomic absorption analysis using field-exploration methods (see Meier, this volume, for discussion of analytical techniques). Fluorine and chlorine analyses were made using specific ion elec- trodes. Semiquantitative spectrographic analysis was used to determine the other elements discussed in this report. In general, we have regarded any rock or sediment sample containing more than 100 ppm (parts per mil- lion) lithium as having an anomalous concentration. This value corresponds well with other studies such as that by J. R. Davis (this report), which show an average value of less than 100 ppm lithium for most common rocks and sediments. ACKNOWLEDGMENTS Discussions with colleagues have contributed to this work. We also wish to thank Wolfgang Elston of the University of New Mexico and James C. Ratté and Richard A. Sheppard of the US. Geological Survey for valuable information on the geology of southwestern New Mexico. The assistance of Laura L. Wray both in the field and in the laboratory contributed immeasura- bly to this project. Robert Brown performed the X-ray diffraction analysis and Katherine Van Weelden helped in the identification of minerals on the X-ray traces. J. C. Hamilton did the six-step semiquantitative spectro— graphic analyses. GILA CONGLOMERATE The Gila Conglomerate, named by Gilbert (1875), is a stratigraphic term that includes most of the basin fill in southwestern New Mexico (fig. 62). The Mimbres River or the Continental Divide generally is considered to be the eastern extent of the Gila Conglomerate; similar sed- iments east of this area are called Santa Fe Group. The Gila Conglomerate is not precisely dated—its age gener- ally is given as ranging from Pliocene to Pleistocene (Heindl, 1952) and its age probably varies between ba- sins. Fossils of late Pliocene age were described by Knechtel (1936) from southeastern Arizona. Basaltic andesite at or near the base of the Gila Conglomerate along Sapillo Creek (fig. 62) has been radiometrically dated at 20.6:05 my (early Miocene) and’ basalt near the top in the Mimbres River valley was dated at 6.3 my (late Miocene) (Elston and others, 1973). Heindl (1952, 1954, 1963) has discussed the problems of the name, “Gila Conglomerate” in detail. The Gila Conglomerate generally correlates with the Santa Fe Group to the east and with the Muddy Creek Formation in Nevada and Utah—all these being general names for continental rocks of late Tertiary to Quaternary age. Most of the Gila Conglomerate is made up of IOB“ . DIVIDE ') ARIZONA EW MEXICO , ' \ I’- BLACK RANGE BN Fa wood - Ho/Spr/ngs‘r . kL U N A H Deming 32° 0 go 40 MILES O 20 40 KILOMETRES FIGURE 62,—Distribution of the Gila Conglomerate (patterned areas) in southwestern New Mexico, generalized from the New Mexico geologic map (Dane and Bachman, 1965). Heavy lines indicate faults. salmon-colored conglomerate and sandstone which are coarser grained at the margins of basins. Towards the centers of basins, the rocks are finer grained, grading into siltstone and claystone. Locally, near the centers of the basins, gypsum deposits, tuffaceous beds, diatomaceous beds, and thin units of freshwater lime- stone (Heindl, 1952; Gillerman, 1964) occur. Evidence of tectonic and volcanic activity during deposition can be found in many places—lava flows and ash beds are not uncommon in the Gila. The Gila Conglomerate is locally folded and faulted (Ballmann, 1960). In some areas, such as south of Redrock, N. Mex., tufa beds, probably from old hot springs, occur in the Gila Conglomerate (Elston, 1960). BUCKHORN AREA Lacustrine clay, silt, and diatomite in the Gila Con- glomerate crop out along US. Highway 180 in the valley of Duck Creek near Buckhorn, N. Mex. Although never studied in detail, these lake beds have been described by Elston (1960), Gillerman (1964), and Ratté and others (1972). The diatomite has been prospected (Gillerman, 1964) and on the southeastern margin of the old lakebed a zeolite deposit is being prospected (R. A. Sheppard, 1975, oral commun.). Two miles (3.2 km) northwest of Buckhorn a roadcut in the lake beds exposes about 25 feet (8 m) of the LITHIUM RESOURCES AND REQUIREMENTS BY THE YEAR 2000 TABLE 55.—Mean content of minor elements in I 4 samples of lake beds in the Gila Conglomerate, southwestern New Mexico [From six-step semiquantitative Spectrographic analyses by j. C. Hamilton. Lithium analyses by atomic absorption by A. L. Meier. Values are in parts per million except for Fe, Mg, Ca, and Ti, which are in percent] Standard Element Mean deviation Average in shale‘ Li (ppm) ____ _ 75 43 66 Fe (percent)__ - 2 .76 4.7 Mg (percent) _ 1 .41 1.5 Ca (percent)__ _ 3 2.4 2.2 Ti (percent) _____ c 17 .05 .46 Mn (ppm) _______ . 200 70 850 B (ppm) __ _ 50 20 100 Ba (PPm) -- _ 500 180 580 Be (ppm) __ _ 16 .63 3 Co (ppm) _. _ 8 3.7 19 Cr (ppm) __ _ 70 27 90 Cu (ppm) ____ __________ 20 4.8 45 La ( pm) ______________ 30 8.9 92 Nb 83pm)" 7.8 4.6 11 Ni (ppm) .. 17 7 68 Pb (ppm) -- 16 2.l 20 Sc (ppm) -- 8 2.6 13 Sr (ppm) .c 380 183 300 V (ppm) ...... __ 280 130 130 Y (ppm) ______ __ 16 5 26 Zr (ppm) __________________________ 86 18 160 ‘Turekian and Wedepohl (196]). diatomaceous section. The lithium values in 14 samples from this section ranged from 14 to 170 ppm—the mean value was 75 ppm, the standard deviation 43. X-ray diffraction analyses showed that besides diatoms the rock contains a montmorillonitic clay, quartz, potas- sium feldspar, plagioclase, biotite, clinoptilolite, calcite in some samples, and minor amphibole. Looking at hand samples through a binocular microscope revealed a rock made of diatom fragments, very small detrital grains and flakes (silt size) of quartz, feldspars, biotite, and amphiboles. In some samples pieces of reed re- placed by potassium feldspar are abundant. Spectrographic analyses of these samples show that they have a fairly uniform composition. The lithology and geochemistry of the sediments suggests that in this part of the section the lake never evaporated to dryness, but the presence of authigenic potassium feldspar and clinoptilolite suggests that it was highly alkaline. Of the elements analyzed for, only lithium and vanadium occur in anomalous amounts (table 55). However, lithium and vanadium show no correlation with each other whereas lithium does show a weak correlation with magnesium and iron. Owing to the high (170 ppm) lithium value found in the lake beds, we decided to collect some more samples. The first sample site was a zeolite prospect pit 1.4 miles (2.25 km) southeast of the Post Office at Buckhorn, N. Mex., in NW% sec. 10, T. 15 S., R. 18 W. The beds in this sequence range from about 6 inches (15 cm) to 2 feet (60 cm) thick and consist of brown to gray to white zeolitic tuff, claystone, siltstone, freshwater limestone, and red, green, and gray chert. Above the excavated part of this section, limestone and chert increase. X-ray diffraction analyses showed that the claystone and siltstone consist chiefly of feldspar, quartz, clinoptilolite LITHIUM RESOURCES AND REQUIREMENTS BY THE YEAR 2000 125 and clay. The lithium content of samples from this sec- tion ranged from 44 ppm in green chert to 200 ppm in the claystone. The lithium contents of the samples above the excavated part were lower (maximum 150 ppm), but this may be because the samples are more weathered. About 11/; miles (2 km) west of the zeolite pit another prominent section of more than 200 feet (60 m) of the lake beds is exposed in a draw. The top of this section is about 400 feet (120 m) topographically above the top of the section exposed in and near the prospect pit. A west-northwest-trending fault occurs on the north side of the draw and basalt is juxtaposed to the lake beds. According to]. C. Ratté (oral commun., 1975) the basalt is older than the lake beds. The lake beds appear to be flat lying in both areas, but detailed mapping would be necessary to determine the exact stratigraphic relation- ships between the two sections. The section exposed in the cliff on the southwest side of the draw is similar to that exposed at the zeolite pit except that some prominent white beds of zeolitic tuff also occur. At the base of this section, red-gray to red- brown rocks become more prominent. Although weath- ered, the sediments in this section still show some anomalous lithium contents. The lithium contents range from 17 ppm in green chert to 240 ppm in a white ash bed near the base of the outcrop. , As mentioned, chert is abundant in these sections, and some of it closely resembles Magadi-type chert as de- scribed by Eugster (1967, 1969), Hay (1968, 1970), Sur- dam, Eugster, and Mariner (1972), and Sheppard and Gude (1974). Magadi-type chert was first reported from the Buckhorn area by D. M. Dennis (oral commun. in Sheppard and Gude, 1974). According to Surdam, Eugster, and Mariner (1972), the presence of Magadi- type chert indicates a depositional environment having a pH of at least 9.5. The lithium content of the chert sampled ranged from 17 to 84 ppm; the mean lithium content in six samples was 41 ppm and the standard deviation 23. The mean fluorine content was 1,125 ppm and the standard deviation 724. In the chert, lithium and fluorine have a correlation coefficient (r) of 0.83, indicating a strong relationship (fig. 63). We analyzed the samples from the Buckhorn area for fluorine as well as lithium. As mentioned, the strongest correlation between lithium and fluorine was found in the chert samples. In 28 claystone and siltstone samples containing less than 1 percent fluorine the correlation coefficient between lithium and fluorine (r=0.7) indi- cates a fairly strong relation. However, in the samples containing more than 1 percent fluorine, there is almost no correlation between lithium and fluorine (r=0.26 for five samples). In the samples containing less than 1 per- cent fluorine, lithium increases as fluorine content in- creases; in the samples containing more than 1 percent fluorine, lithium content tends to decrease as fluorine content increases (fig. 63). 2.5— _ EXPLANATION L . o Fine-grained sediments — or ash ._ _ x Cheri E 2.0— o m .— m n_ _ E m“ E a: O :> _| u. I.O— 0 o_9|lllI11ll|llIIllllI 6000 O 5500 5000 4500 4000 3500 3000 2500 2000 FLUORINE, IN PARTS PER MILLION |500 lIlllllllllII|llllllllllllllllllllllllllllllllllll IOOO : K,/\xy=25.6x + 75.5 _/ 500 jX *IlllIIIIIIIILIIIIIIIIIILJ O 0 50 I00 I50 200 250 LITHIUM,|N PARTS PER MILLION FIGURE 63.—Lithium-fluorine correlations in samples of lake-bed sediments in the Gila Conglomerate near Buckhorn, N. Mex. Correlations based on least squares linear regression equation, y =mx H). x, lithium value; y, fluorine value, r, computed correlation coefficient. 126 HYDROTHERMAL ALTERATION OF THE GILA CONGLOMERATE Mineralizing episodes have occurred repeatedly throughout the geologic history of southwestern New Mexico, and in several places the Gila Conglomerate is mineralized. Chrysocolla coats boulders and forms vein- lets in the Gila Conglomerate in the northern Big Burro Mountains (Hewitt, 1959, p. 92; Gillerman, 1964, p. 59—60). Fluorite veins occur in the Gila (Elston, 1960, 1970), and south of Redrock, N. Mex., manganese was mined from veins in the Gila Conglomerate (Elston, 1965, p. 212). North of Silver City, N. Mex., veins of meerschaum (Sterrett, 1908; Ratté and others, 1972) occur in the Gila. Areas of hot-spring alteration are evi— dent between Silver City and the Gila Cliff Dwellings National Monument, and hot springs are common in the area. In the Caprock district, south of Redrock, N. Mex. (fig. 62), manganese oxide veins in the lower part of the Gila Conglomerate are associated with travertine de- posits and have other features suggesting that the viens were deposited from hot springs. The lower part of the Gila in this area consists of very coarse bedded conglom- erate and sandstone; cobbles as large as 15 cm in diame- ter occur near the manganese veins. The veins are steeply dipping to vertical and fill faults which are alined parallel to major regional structures (Gillerman, 1964, p. 168). Along with psilomelane and pyrolusite, calcite, white fluorite, and quartz occur in the veins, The quartz ranges from opaline to coarsely crystalline and generally occurs near the center of the veins. The manganese oxides replace the matrix of the conglomerate near the veins, and calcite locally )cements the conglomerate. Elston and others (1973) reported that the manganese mineralization is younger than the basaltic andesites in the area dated at 2061-16 m.y. To check the possibility of lithium occurring in the mineralizing solutions, we took some grab samples of the manganese ore, vein materials, and country rock. Lithium contents of the eight samples taken ranged from 10 to 40 ppm, so we have to conclude that either the mineralizing solutions did not contain much lithium or that the lithium was not precipitated with the man- ganese and fluorine. The evidence we have now, which is mostly negative, suggests that the lithium in the sys- tem was carried in the hot springs to surface drainage. Another area of mineralized Gila Conglomerate oc- curs approximately 1 mile (1.6 km) north of New Mexico Highway 25 where meerschaum was mined (fig. 62). Meerschaum is a lightweight, tough clay that can be carved and is used for pipes (for smoking tobacco), as insulation for pipes, and for radio insulators. The oc— currence was described by Sterrett (1908) and Ratté and others (1972). The meerschaum is in steeply dipping LITHIUM RESOURCES'AND REQUIREMENTS BY THE YEAR 2000 veins and veinlets in Gila Conglomerate and the under- lying basaltic andesite (Ratté and others, 1972). Meerschaum is usually defined as sepiolite (American Geological Institute, 1957), but X-ray diffraction analysis shows the vein material from the meerschaum deposit to be palygorskite, (Mg,Al)2Si4010(OH)-4H20. The material also contains calcite crystals, clinoptilolite, a small amount of quartz, and and undetermined clay mineral having a sharp peak at 15 A. The altered con— glomerate near the veins consists of quartz grains in a matrix of clay, clinoptilolite, palygorskite, and calcite. The lithium contents of samples from the area were low: 10 ppm in the vein material and 25 ppm in the altered conglomerate. In contrast, a sample of coarse sandstone from several miles north of the meerschaum deposit contained 105 ppm lithium. A sample of altered ande- site underlying the Gila Conglomerate about 3 miles (5 km) northwest of the meerschaum deposit contained 6 ppm lithium which is half the average content (12 ppm) reported for andesites by Heier and Adams (1964). These few preliminary data suggest that lithium may have been leached from the country rock and carried out by the same solutions that deposited the palygorskite, and that similar hot-spring systems have leached the andesite underlying the Gila Conglomerate. Systematic sampling in altered and unaltered areas of Gila Conglomerate is needed to confirm or refute the leaching hypothesis. HOT SPRINGS Many hot springs occur in and near the Mogollon Mountains. The average lithium content of hot springs in this region, from our data and from published data (table 56), is approximately 0.25 mg/l. Chlorine averages 35 mg/l and conductivity averages 713 micromhos for eight springs. The lithium/chlorine ratios plot along a curve which has a slope normal for hot springs (Smith, this report). However, even though dilute, these springs do supply lithium. SUMMARY Reconnaissance sampling in southwestern New Mexico suggests an opportunity for detailed study of a geochemical cycle of lithium. We have preliminary evi- dence of leached rocks (a source), of mineralizing fluids (transport), and lake beds (deposition). Lakes, similar to those represented by the rocks in the valley of Duck Creek, can be a sink for the elements leached by hot springs. Although reconnaissance did not disclose any- thing more than anomalous lithium contents in the lake beds in the Gila Conglomerate in the Valley of Duck Creek, higher concentrations of lithium may occur within the lake bed sequence, possibly in beds of hecto- rlte. LITHIUM RESOURCES AND REQUIREMENTS BY THE YEAR 2000 127 TABLE 56.—Chemist1y of hot springs in southwestern New Mexico [Leaders(_--_) indicate no data] Temperature Li Cl Conductivity Specific Location Name (°F) (°C) (mg/l) (ppm) (micromhos) gravity NEVgtsec. 5, T. 13 5., R. 13W. Gila 148 64 0.21 110 600 1.0006 SEV4 sec. 24, T 12 S., R. 14 W. Unnamed 152 66 .34 115 700 1.0005 NEV; sec. 10, T. 13 S., R. 13 W. ‘1065 41 ‘.20-0.25 ' 98—103 1568—581 __- NEV4 sec. 51, T. 12 S., R. 13 W. ‘149.6 65 ‘.37-0.40 l105—110 ‘720—771 ___ NWV. sec. 20, T. 13 S. R. 13 W. ——— -—— ‘.25 _-_ ___ ___ NWVt sec. 20, T. 205., R. 11 W. ’129.2 54 .18 ’ 17 ‘610 ‘1.000 NEV. sec. 12, T. 18 S. R. 10 W. 1129.2 54 .12 2 19 ’440 ’1.000 NEV4 sec. 14, T. 12 S., R. 20 W. San Francisco ’117 47 ,44 2473 ’1,300 ’1.001 Average _________________________________________________ 133 56 .25 135 713 _-- ‘Summers (1972). ”Summers (1965). REFERENCES CITED American Geological Institute, 1957, Glossary of geology and related sciences: Natl. Research Council Pub. 501, 325 p. Ballmann, D. L., County, New Mexico: New Mexico Bur. Mines and Mineral Re- sources Bull. 70, 39 p. Dane, C. H., and Bachman, G. 0., 1965, Geologic map of New Mexico: Washington, DC, U.S. Geol. Survey. Elston, W. E., 1960, Reconnaissance geologic map of Virden thirty- minute quadrangle: New Mexico Bur. Mines and Mineral Re- sources Geol. Map 15. I 1965, Mining districts of Hidalgo County, New Mexico [summ.], in Guidebook of Southwestern New Mexico II—New Mexico Geol. Soc., 16th Field Conf. 1965: Socorro, New Mexico, New Mexico Bur. Mines and Mineral Resources, p. 210—214. 1970, Volcano—tectonic control of ore deposits, southwestern New Mexico, in Guidebook of the Tyrone-Big Hatchet Mountains-Florida Mountains region—New Mexico Geol. Soc. Field Conf., 2lst, 1970: Socorro, New Mexico, New Mexico Bur. Mines and Mineral Resources, p. 147—153. Elston, W. E., Damon, P. E., Coney, P. J., Rhodes, R. C., Smith, E. I. and Bikerman, Michael, 1973, Tertiary volcanic rocks, Mogollon-Datil province, New Mexico, and surrounding region; K-Ar dates, patterns of eruption, and periods of mineralization: Geol. Soc. America Bull., v. 84, no. 7, p. 2259-2273. Eugster, H. P., 1967, Hydrous sodium silicates from Lake Magadi, Kenya—precursors of bedded chert: Science, v. 157, p. 1177— 1180. 1969, Inorganic bedded cherts from the Magadi area, Contr. Mineralogy and Petrology, v. 22, p. 1—31. Gilbert, G. K., 1875, Report on the geology of portions of New Mexico and Arizona, in Report upon U.S. geographical and geological explorations and surveys west of the one hundredth meridian, in charge of First Lieut. G. M. Wheeler, Volume 3: Washington, DC, U.S. Govt. Printing Office, p. 503—567. Gillerman, Elliot, 1964, Mineral deposits of western Grant County, New Mexico: New Mexico Bur. Mines and Mineral Resources Bull. 83, 213 p. Hay, R. L. 1968, Chert and its sodium- silicate precursors in sodium- carbonate lakes of East Africa: Contr. Mineralogy and Petrology, v. 17, p. 255-274. Kenya: 1960, Geology of the Knight Peak area, Grant, 1970, Silicate reactions in three lithofacies of a semi-arid basin, Oldavai Gorge, Tanzania, in B. A. Morgan, ed., Fiftieth Anniver- sary Symposia—Mineralogy and petrology of the Upper Mantle, sulfides, [and] mineralogy and geochemistry of non-marine evaporites: Mineralog. Soc. America Spec. Paper 3, p. 237—255. Heier, K. S., and Adams, ]. A. S., 1964, The geochemistry of the alkali metals, in Physics and chemistry of the earth, Volume 5: New York, Macmillan Co., p. 253—381. Heindl, L. A., 1952, Gila Conglomerate, in Arizona Geol. Soc., Guidebook for field trip excursions on southern Arizona: p. 113—1 16. 1954, Cenozoic alluvial deposits in the upper Gila River drain- age basin, Arizona and New Mexico [abs]: Geol. Soc. America Bull., v. 65, no. 12, pt. 2, p. 1262. 1963, Cenozoic geology in the Mammoth area, Pinal County, Arizona: U.S. Geol. Survey Bull. 114l—E 41 p. Hewitt, C. H., 1959, Geology and mineral deposits of the northern Big Burro Mountains-Redrock area, Grant County, New Mexico: New Mexico Bur. Mines and Mineral Resources Bull. 60, 151 p. Knechtel, M. M., 1936, Geologic relations of the Gila Conglomerate in southeastern Arizona: Am. Jour. Sci., 5th ser., v. 31, no. 182, p. 81—92. Ratte,j. C., Gaskill, D. L., Eaton, G. P., Peterson, D. L., Stotelmeyer, R. B., and Meeves, H. C., 1972, Mineral resources of the Gila Primi- tive Area and Gila Wilderness, Catron and Grant Counties, New Mexico: U.S. Geol. Survey open-file rept. 428 p. Sheppard, R. A., and Gude, A. J., 3d, 1974, Chert derived from magadiite in a lacustrine deposit near Rome, Malheur County, Oregon: U.S. Geol. Surveyjour. Research, v. 2, no. 5, p. 625—630. Sterrett, D. B., 1908, Meerschaum in New Mexico, in Contributions to economic geology, 1907: U.S. Geol. Survey Bull. 340, p. 466—473. Summers, W. K., 1965, Chemical characteristics of New Mexico’s thermal waters—a critique: New Mexico Bur. Mines and Mineral Resources Circ. 83, 27 p. ' 1972, Factors affecting the validity of chemical analyses of natural water: Ground Water, v. 10, no. 2, p. 12—17. Surdam, R. C., Eugster, H. P., and Mariner, R. H., 1972, Magadi-type chert in juraSsic and Eocene to Pleistocene rocks, Wyoming: Geol. Soc. America Bull., v. 83, no. 8, p. 2261—2265. Turekian, K. K., and Wedepohl, K. H., 1961, Distribution of the ele- ments in some major units of the Earth’s crust: Geol. Soc. America Bull., v. 72, no. 2, p. 175-191. N 128 LITHIUM RESOURCES AND REQUIREMENTS BY THE YEAR 2000 LITHIUM IN CLAYEY ROCKS OF PENNSYLVANIA AGE, WESTERN PENNSYLVANIA By HARRY A. TOURTELOT and ALLEN L. MEIER, U.S. GEOLOGICAL SURVEY, DENVER, CO ABSTRACT Clayey rocks of Pennsylvanian age in western Pennsylvania have a wide range of lithium concentrations. The first of three data sets con- sists of analyses of 162 samples of clay and shale from the Greater Pittsburgh region which were selected for testing because of possible ceramic usefulness. Lithium concentrations for shale range from less than 33 ppm to 165 ppm and for underclay from less than 33 ppm to 270 ppm; 25 samples contain 100 ppm lithium or more. Formation name, geometric mean, geometric deviation, and number of samples are as follows: Pittsburgh Formation, 15 ppm, 3.25, 27; Casselman Formation, 12 ppm, 4.13, 34; Glenshaw Formation, 47 ppm, 2.39, 22; Freeport Formation, 93 ppm, 1.97, 8; Kittanning Formation, 23 ppm, 4.19, 29, and Clarion Formation, 56 ppm, 2.83, 16. The relatively ,large geometric deviations indicate that the stratigraphic variation among the formation means should be viewed judiciously. The second data set consists of analyses of 38 samples of shale, underclay, and siderite from Beaver County, Pa. The range of lithium concentrations is similar to that in the first data set, but one shale sample contains 280 ppm lithium and an underclay sample contains 330 ppm. Siderite contains 15-40 ppm lithium. Within two strati- graphic sections, lithium concentrations greater than 100 ppm may represent marine shale and concentrations less than 100 ppm may represent nonmarine shale. The third data set consists of a subset of analyses of 13 samples of shale and underclay from Armstrong and Butler Counties and another subset of analyses of about 20 samples of flint clay from Clear- field, Jefferson, and Somerset Counties. The shale and underclay have smaller lithium concentrations than similar rocks in Beaver County, the highest value being 160 ppm in a shale. Perhaps these rocks were deposited in an area in which marine influences were very slight. Flint clay tends to contain more lithium than any other clayey rocks analyzed: 17 of 20 flint clay samples contain 300 ppm or more and 10 of these 17 contain more than 500 ppm. In addition 6 of the 20 values cluster at about 300 ppm and another 6 cluster at about 900 ppm. Although a diaspore clay contains only 40 ppm, no differences in mineral composition that might be related to lithium content can be detected by whole-sample X-ray diffraction analyses. Flint clay is made up almost entirely of well-crystallized kaolinite. Quartz was detected in only two samples. Illite and boehmite each are present in a few sam— ples but there is no evident relation between mineral composition and lithium content. In one set of samples through a clay bed, lithium concentrations are largest in the upper part of the bed. Structural substitution of magnesium and lithium for aluminum may account for the occurrence of lithium in polymineralic shale and underclay that contain considerable illite as well as kaolinite if lithium concentrations are less than perhaps 300 ppm. Some lithium may be held in cation-exchange positions, but the amounts should be small because other cations should replace lithium in clay minerals in near— surface rocks in a humid climate. The small magnesium content of flint clays makes unlikely magnesium-lithium substitution in the kaolinite in flint clay, and cation exchange seems even more improba- ble because of the small cation-exchange capacity of well-crystallized kaolinite. The lithium content of flint clays may be due to the presence of small amounts ofa lithium mineral that has not been recognized. If a mineral contains 1.5 percent lithium, 1 percent of the mineral in a sample would give the sample a lithium concentration of 150 ppm. Underclays are thought to be the result of relatively pure accumula— tion of clay in low-energy coal-swamp environments or the result of severe leaching following coal-bed deposition. Flint clay and more aluminous minerals are generally interpreted as the result of extreme leaching. Another view is that flint clay, and perhaps other occur- rences of kaolinite as well, are the result of silication of preexisting aluminum-hydroxide minerals. Under the conditions of either ex- treme leaching or silication, it is possible that lithium-bearing minerals could be formed. Although lithium aluminosilicates are known so far only from igneous rocks, the wide range of authigenic silicate minerals in the Eocene Green River Formation makes it clear that the mineral- forming potential of unusual sedimentary environm'ents is not com- pletely understood. INTRODUCTION Clay minerals are widely believed to be the primary residence sites for trace constituents, including lithium, in sedimentary rocks. Clayey rocks contain more lithium than other kinds of sedimentary rocks, the lithium con- tents of which are grossly proportional to their clay con— tents. Horstman (1957, p. 3) estimated the average lithium content of worldwide shales to be 66 ppm, an estimate accepted by Turekian and Wedepohl (1961, table 2) and Heier and Billings (1972, p. 3—k—1); other recent estimates range from 60 ppm (Ohrdorf, 1968, p. 207) to 80 ppm (Ronov and others, 1970, p. 86). Keith and Degens (1959, p. 42) reported that marine shales of Pennsylvanian age in western Pennsylvania contain an average of 159 ppm lithium, whereas as- sociated nonmarine shales contain 92 ppm. The geochemical interest in lithium in shales of Pennsylva- nian age was reemphasized by publication of plate- reader spectrographic analyses of shale and clay in the six-county Greater Pittsburgh region as part of the US. Geological Survey’s program of environmental studies there (O’Neill, 1973). Of the 162 samples of shale and clay selected for their possible ceramic uses, 25 have a lithium content of 100 ppm or more, and the maximum value was about 270 ppm. Brief reconnaissance sampl- ing therefore was undertaken in 1974 and 1975 in con— junction with other work in the Greater Pittsburgh re- gion. This paper deals with three sets of data: (1) the plate- reader semiquantitative spectrographic analyses for lithium of 162 samples of shale and clay from the Grea- ter Pittsburgh region; (2) atomic absorption analyses for lithium in 38 samples from Beaver County, Pa. (table 57); and (3) two data subsets, first, atomic absorption analyses for lithium of 13 samples of shale and clay from LITHIUM RESOURCES AND REQUIREMENTS BY THE YEAR 2000 Armstrong and Butler Counties, Pa. (table 57), and sec- ond, 21 samples of flint clay and related materials from Clearfield, Jefferson, and Somerset Counties, gener- ously supplied for lithium analysis (table 57) by the Gen- eral Refractories Company, Pittsburgh, Pa. SHALE AND CLAY FROM DIFFERENT AREAS GREATER PITTSBURGH REGION SHALE AND CLAY Shale and clay resources in the Greater Pittsburgh region (Allegheny, Armstrong, Beaver, Butler, Washington, and Westmoreland Counties, figure 64) are being investigated by the Pennsylvania Bureau of Topographic and Geologic Survey in cooperation with the US. Geological Survey and the US. Bureau of Mines (O’Neill, 1973). The lithologic, physical, chemical, mineralogic, and use data obtained will contribute to broadening the natural resource base of the region. Re- ginald P. Briggs of the US. Geological Survey supplied the stratigraphic positions of most of the samples (writ- ten commun., 1974). The 162 samples collected by the Pennsylvania Bureau of Topographic and Geologic Survey were 80° 79° 78° I I l l 0 so KILOMETRES I L PENNSYLVANIA NEW YORK _—__—'—_FTE'NITsY_tm—-— 42° — 40° MARYLAND WEST VIRGINIA I l | I FIGURE 64.——Distribution of sample localities and rocks of Pennsylva- nian and Permian ages in western Pennsylvania Pennsylvanian and Permian rocks patterned. Stipple-outlined six counties constitute the Greater Pittsburgh region. Italic capital letters indicate counties: AL, Allegheny; AR, Armstrong, B, Beaver; BU, Butler, CL, Clearfield; ], Jefferson; S, Somerset; WA, Washington; WE, Westmoreland. Roman capital letters, sample localities in table 57. 129 selected primarily for the possible ceramic usefulness of the rock units sampled. Shale and clay of possible ceramic use are not particularly more prevalent in marine than in nonmarine strata, nor do these shale and clay units have other characteristics that would exclude them from the general class of fine-grained rocks called shale. The wide range of values found for lithium, from less than 33 ppm to 270 ppm, is thought probably to represent a major part of the potential spectrum of lithium contents in such rocks. Lithium concentrations in 96 of the 162 samples were reported as qualified val— ues, that is, as “less than 33 ppm;” this value suggests that current estimates of the abundance of lithium in shale may be too large. The lithium contents of the samples from the forma- tions of the Allegheny, Conemaugh, and Monongahela Groups of Pennsylvanian age are summarized in figure 65. The frequency distributions are positively skewed, thus the geometric mean is the best estimate of the most common value. For some formations, more than half the samples were reported 'as containing less then 33 ppm lithium. Logarithmic transformations and the Cohen method of dealing with qualified values in I0 :3 Pittsburgh Fm. g I) :: GM = I5 g a 5 3: GD = 3.25 <3; 3 n = 27 EC? 3: 2 1' I I | A Casselmon Fm] GM = 12 GO = 4J3 n = 34 ' I I I Glenshuw Fm. GM .= 4? GD = 2.39 . n = 22 o. coo-go [I | l Conemauah Group A, Freeport Fm. I GM: 93 GD = I.97 n = 8 ‘# SAMPLES l"°’ | | l Kittanning Fm. GM = 23 GD = 4.l9 . n = 29 [A 00.0.00... Allegheny Group 41 Clarion Fm. GM = 56 GD = 2.83 n = |6 oooolAAAI AA [A | 4‘ IOO I50 200 250 300 LITHIUM, IN PARTS PER MILLION O 50 FIGURE 65.—Frequency distribution of lithium concentrations by geologic formation in Greater Pittsburgh region. Solid triangle, underclay; solid circle, shale and clay; GM, geometric mean; GD, geometric deviation; n, number of samples. 130 geochemistry as discussed by Miesch (1967) were used in analysis of the data. The Freeport and Clarion Formations of the Al- legheny Group and the Glenshaw Formation of the Conemaugh Group have geometric means for lithium of 93, 56, and 47 ppm, respectively. These figures are markedly different from those of the Pittsburgh Forma- tion (15 ppm), the Casselman Formation (12 ppm), and the Kittanning Formation (23 ppm). To the extent that sample selection has not introduced large biases, these figures clearly point to the Freeport, Clarion, and Glen- shaw Formations as the most interesting for future studies on lithium in shale in the Greater Pittsburgh region. Among the samples from all formations that contain 100 ppm lithium or more, shales and underclays1 are about equally divided with 13 and 12, respectively. It is obvious, however, that all the samples containing more than 100 ppm lithium in the Kittanning and Clarion Formations are underclays, and also these formations are the only formations in which underclays make up a significant proportion of the samples that contain less than 33 ppm lithium. This data may partly be a result of sampling, inasmuch as underclays seem to be thicker and more suitable for ceramic use in these formations than in the others and would receive more attention in sampling for ceramic testing. With respect to lithium, however, the data imply that the lithium contents of underclays are independent of the special composition and position in the sedimentational sequence of the un- derclays compared to the shales that make up the bulk of the coal-bearing sequences. The reason for the varia— bility in lithium is not evident to us. BEAVER COUNTY, PENNSYLVANIA The sampling in Beaver County, Pa., mostly within the Kittanning Formation, was intended primarily to in- vestigate localities reported by O’Neill (1973) that had yielded samples containing more than 100 ppm lithium, although the exact beds yielding those samples could not be identified in the field. The correspondence be- tween the localities shown in figure 64 and those in O'Neill (1973) is shown below (capital letters are localities from figure 64 and numbers are from O’Neill): C (Cain mine) ____________________ 58 (approximately) E (Ellsman mine) __________________ 58—60 C (Glasgow) ______________________ 62 M (Midland) ______________________ 63 PR (Peggs Run) ____________________ 64 V (Vanport) ______________________ 76—79 (approximately) “‘Underclay. A layer of fine-grained detrital material, usually clay, lying immediately be- neath a coal bed or forming the floor of a coal seam" (Gary and others, 1972, p. 766). Mineralogically, most underclays consist chiefly of mixtures of kaolinite and illite in contrast to associated shales that consist of illite, kaolinite, and Chlorite with admixed quartz and com- monly carbonate minerals. For example, see table 57. LITHIUM RESOURCES AND REQUIREMENTS BY THE YEAR 2000 Shale, underclay, and siderite nodules were sampled to get information on the relations between lithium con- tent and rock type and to estimate variation within stratigraphic sections. Siderite nodules were sampled because S. Landergren reported 100 ppm lithium in marine siderite iron ores (Rankama and Sahama, 1950, p. 428). The method of atomic absorption analysis is described by Meier (this report). The analyses in table 57 confirm the general magnitude of the plate-reader spectrographic analyses reported by O‘Neill (1973). The frequency distribution oflithium in shale, under- clay, and siderite nodules in shale is shown in figure 66. The distribution is very similar to those derived from the spectrographic data in figure 65. The lithium con- tents of both shale and underclay vary widely. Most of the concentrations in shale less than 100 ppm come from the Glasgow and Vanport localities (table 57). The siderite nodules contain less than 50 ppm lithium. The shales range in clay content from 50 to 70 per— cent (table 57) with quartz making up the remainder except for minor amounts of plagioclase feldspar and siderite. The clay fraction is made up chiefly ofillite, but kaolinite predominates in some of the underclays and almost equals the amount of illite in many samples. Chlorite occurs in the shale in amounts of 5—10 percent, but it is generally absent from the underclays. No rela- tion is evident between lithium content and the mineral composition of the samples. The stratigraphic and lithologic distributions of lithium are shown in figure 67. Stratigraphic assign- ments are based on the geologic map of Patterson (1963), but uncertainty remains in our interpretation of the stratigraphic position of the Glasgow section. If an underclay exists beneath the Middle Kittanning coal, it is very thin compared to the underclay of the Middle Kittanning at the Ellsman mine. In addition, we did not find the Middle Kittanning coal at the Midland section, although it may be present above the sandstone at the top of the section in figure 67, at a much greater interval than separates the Middle and Lower Kittanning coals at the Ellsman mine. Few regularities in distribution of lithium content by lithology and stratigraphic position are evident in figure 67. The largest lithium concentration is 330 ppm in a sample of the underclay of the Middle Kittanning coal at the Ellsman mine but the next largest (280 ppm) is a siderite-rich sample of black shale at the Midland local- ity. Samples from the Lower Kittanning underclay con- tain about 150 ppm lithium at both the Ellsman mine and the Midland locality. The stratigraphic distribution of lithium in the Ellsman mine and at the Midland sections seems to be partly interpretable in terms of the difference in lithium content of marine and nonmarine shales reported by LITHIUM RESOURCES AND REQUIREMENTS BY THE YEAR 2000 131 TABLE 57,—Lithium contents and X-ray mineralogical analyses of clayey rocks of Pennsylvanian age, western Pennsylvania [Lithium analyses by A. L Meier; X-ray diffraction analyses by H. A. Tourtelot using the method of Schultz (1964). Mineral percents in parentheses are result of different kind of X-ray analysis that yields order-of-magnitude’figures only; x, mineral detected; n, mineral not detected; leaders indicate no data] Clay minerals Mineral composition (Percent of total clay) (Pcrcent)‘ Clayey rocks Sam le Li Kaolinile‘ Chlorite [Hits and Total Quartz Other loca ity (ppm) expanding clay clay minerals Armstrong County Heilman mine Varicolored clay ____________________________ HMl 77 21 (n) ’79 70 13 ________________ Gray clay __________________________________ HM2 80 22 (n) ’78 40 21 Calcite 5, West Valley mine Dolomite 2. Lower Freeport coal Dark-gray shale ____________________________ WV] 160 17(2) 23 60 45 15 Siderite 5 Gra WV2 72 15 (n) 285 40 30 Coa WV3 64 ——————————————————————————————————————————————————————————————————————————————————————————— Beaver County Cain mine Underclay (Middle Kitlanning coal) __________ C1 85 (60)(2) (n) “(40) (60) (40) Shale ______________________________________ C2 81 (50) (10) 2(40) (55) (45) Ellsman mine Underclay (Lower Kitianning coal)____ E1 150 (80)(2) (n) 2(20) (70) (30) Dark-gra shale E2 100 (45)(2) (10) ’(45) (50) (50) Gray sha e _____ E3 110 (40) (10) '(50) (60) (40) Siderite from ab E35 21 (x) (n) (x) (x) (x) E4 75 (40x2) (10) (50) (60) (40) E45 40 (x) (n) (x) (x) (x) E5 330 (60)(2) (n) 2(40) (75) (25) E6 110 (40) (10) 2(50) (50) (50) __ ......... ~ ______ E7 110 (x) * (x) (x) (30) (20) Siderite(50) Gray shale _____________ E8 100 (40)(2) (10) ’(50) (50) (50) Plagioclase(x) Gray shale _______________ E9 100 (35)(2) (15) (50) (50) (50) ________________ Siderite from above _ ______ E95 18 (x) (n) (x) (x) (x) Siderite(90) GlGray shale __________________________________ E10 90 (25)(2) (15) (60) (50) (50) Siderite(x) asgow Gray shale __________________________________ Cl 67 (40)(2) (10) ’(50) (55) (45) P1agioclase(x) Siderite from above ___ Gls 14 (x) (n) (x) (x) (x) Siderite(90), Pyrite(x) Undercla (Middle Kittanning coal) __________ G2 36 (35)(2) (n) 2(65) (60) (40) ____ Black sha e GS 76 (40) (15) (45) (55) (45) __.. ........ Siderite from above G35 21 (x) (n) (x) (x) (x) Siderite(90) Midland Gray shale __________________________________ M1 48 (30) (5) (65) (50) (50) Plagioclase(x) Silly siderite from above _ ........ M15 25 (x) (x) (x) (x) (x) Siderite(75) Gray shale ___________________ M2 74 (35)(2) (n) 2(65 (60) ‘ (40) Siderite from above -_- __ M2s 40 x) (n) (n) (x) (x) Undercla (Lower Kittanning coal) MS 160 (40)(2) (5) “(55) (60) (40) Black sha e _ _ M5 280 (45)(2) (15) 2(40) (60) (40) Gray shale -._ ______ _ M6+2 130 (40) (10) (50) (55) (45) Pla Gray shale _____________ __ M6+8 150 (35) (10) ‘(55) (55) (45) Plagioclase(x) P Sandlgtone __________________________________ M6+9 I4 (75)(2) (x) (25) (35) (65) Plagioclase(x) e 5 un gay shale __________________________________ PR1 80 (35) (10) (55) (50) (50) Plagioclase(x) Vanport Gray shale __________________________________ V1 88 (35x2) (5) (60) (60) (40) Plagioclase(x) Siderite from above ________________________ Vls 24 (x) (n) (x) (x) (x) Siderite(90), Calcite(x) Gray shale __________________________________ V2 100 (40)(2) (x) ’(60) (60) (40) _.__ Gray shale __________________________ V3 64 (40)(2) (10) (50) (55) (45) __ _______ Calcite concretion from above ________________ V35 16 (x) (n) (x) (x) (x) Calcrte(70), _ Siderite(20). Gra shale __________________________________ V4 140 (40)(2) (10) (50) (55) (45) Plagioclase(x) Un erclay (Lower Freeport? coal) __ V5 160 (35) (10) ’(55) (55) (45) ________________ Gray shale __________________________________ V6 130 (40) (10) ’(50) (50) (50) Siderite(x) Butler County McCall mine Black shale ________________________________ McCl 120 12(1) 24 64 40 I5 ---------------- Underclay _._ McC2 13 50(2) n ‘50 60 15 Siderite”; Underclay _-_ MCCS 78 27(2) 2 ’71 75 22 -———5——' ————————— Hard cla _._ MCC4 110 31(2) n 69 45 18 Goethite? Gray sha e ___ McC5 28 26(2) n 275 40 21 ’____ _ Gra shale.-- McC6 80 ___________________ Coa ___________ McC7 57 nu __ Clearfield County ' Altoona Hint clay _ ______________ L2 950 100(1) (n) (n) 100 ' (n) ________________ Freeport flint clay _________________________ (1L1 320 100(1) ____________________________ 85 15 ________________ Galbraith Property. Merc r clay Flint clay __________________________________ GA] 880 100(1) n n 95 n ________________ Do 6A2 900 100(1) n x 95 ____________________________ 6A4 900 ________________________ __ GA5 840 ‘ 100(1) n n ’95 7 n ________________ 6A6 550 100(1) n x 95 n ................ GA7 590 __ GAS 330 100(1) n n 95 n 0A9 320 .-_ __ ._ ______ GAIO 340 100(1) n n 95 n ________________ GA11 330 _________l __ ‘ Do ___._-._.-_.__.-__.__, ________________ GA12 450 100(1) 11 x 95 n ___.__L _____ Z-.- Curwensville-Anderson Creek area, Prospect Hole 1, Mercer clay . Dark-gray flint clay __________________________ KA] 1400 100(1) n n 95 n Boehmite x Light-gray flint clay _____ KA2 850 100(1) n x 95 n Boehmile x jeffries mines, Decatur Twsp., Mercer clay Dark-gray flint cla __________________________ L3 290 100(1) n x 95 n ________________ jeffries mines, Wooclward Twsp., Mercer clay Dias re flint clay __________________________ L4 41 10(1) n n 10 n Diaspore 90 Philips urg Gray silty shale___--_' ________________________ PH] 63 22(2) 15 63 45 26 ________________ jefferson County Brookville flint clay ____________________________ BRI 700 100(1) n n 95 n Boehmite x 132 LITHIUM RESOURCES AND REQUIREMENTS BY THE YEAR 2000 TABLE 57.———Lithium contents and X -my mineralogical analyses of clayey rocks of Pennsylvanian age, western Pennsylvania—Continued Clay minerals Mineral composition (Percent of total clay) (Percent)2 Clayey rocks Sam le Li KaoliniteI Chlorite Illite and Total Quartz Other loca ity (ppm) expanding clay clay minerals Somerset county Mercer flint clay ______________________________ L1 210 100(1) n n 95 n ________________ $01 150 78(1) n 22 70 23 ________________ Star Mining Co. plastic clay ‘Cr stallinit shown b fi ures in arentheses if determinable: (1) well or stallized; (2) oorl cr stallized. Y Y 7 g P Y P Y 1’ zIllite contains much expandable material. 6 SAMPLES u Ill-I II '0'. a... on.- O on 0-. o u I b o A + ‘ _l ' I ‘ fit 00 I50 200 250 300 350 LITHIUM, IN PARTS PER MILLION no Al, | 0 FIGURE 66.—Frequency distribution of lithium concentrations in underclay (solid triangles), shale and clay (solid circles), and siderite (solid squares) from Beaver County, Pa. ELLSMAN MINE 90 ppm Li 95 GLASGOW c I8 ROADCUT ‘9 > ___ METRES E _ _ Lower Freeport __ — g (300‘ _-Upper Kittanningml no“ coaI ' t I00 MIDLAND 3‘ :3, 3mm Li ROADCUT 2 q) 76 u) E / ' - .. ,/ Illlllll 55 X / C< / 2 IOO/ ‘57 s c __— = .9 33° _MiddIe < 3 —40 Kittanning E 75 coal Lg 2I o1 ”5 .E g IOO E _______________ Q -:|50 Lower Kittanning coal l'——|4 KM—l—-6 KM—I EXPLANATION Sandstone E CoaI x 1 I s 5’ 5 X35 Undercluy siderite 5“” a concretion g 3”“ 5"“ FIGURE 67.—Ellsman mine, Glasgow, and Midland stratigraphic sections, Beaver County, Pa., showing lithium concentrations by rock types. Keith and Degans (1959) and the cyclic sequences of environments of coal deposition conceived by Edmunds (1968, p. 25—37). Although such an interpretation is highly speculative, it seems worthwhile in order to point out a potential relation between sedimentation and geochemistry of an unusual type that needs further study. Edmunds’ ideal sequence (fig. 68) begins with sedi- ments deposited landwards of a coal swamp, which are transgressed by the swamp itself. The resulting under— Bruckish or marine open-water deposits Nonmarine quviatiIe I back—swamp deposits (Obscure boundary) ’ CoaI Brackish or marine open-water deposits A Nonmarine quviaIile back-swamp deposits TRANSGRESSION REGRESSION TRANSGRESSION DIRECTION OF TRANSGRESSION % FIGURE 68.—-Idealized cyclic sequence of depositional environments for coal-bearing rocks in west-central Pennsylvania. Modified from Edmunds (1968). clay and coal may be regarded as generally a single unit. As transgression continues, the coal is overlaid by sedi— ments deposited in brackish or marine open water, that transgresses the swamp. The beginning of a regression may be marked by a coal, usually without much of an underclay, but the relation between a regressive unit and the underlying transgressive unit may be obscure (Edmunds, 1968, p. 34—35). The nonmarine fluviatile deposits making up the regressive unit are complex in lithology and arrangement and are characterized by concurrent erosion-and-deposition features. Such cycles of deposition theoretically would result in nonmarine fluviatile shale containing a relatively small amount of lithium overlain by coal and then brackish or marine shale containing relatively large amounts of lithium. If the above models hold, the beds containing less than 100 ppm lithium beneath the Lower Kittanning coal at Midland would be nonmarine, whereas the beds containing more than 100 ppm lithium above the coals at Midland would be marine. The sandstone at the top LITHIUM RESOURCES AND REQUIREMENTS BY THE YEAR 2000 of the local section probably is a fluviatile sandstone of the regressive unit and the base of the sandstone a dis- conformity. Similarly, at the Ellsman mine the section between the Lower and Middle Kittanning coals could consist of transgressive marine shale containing 100 or more ppm of lithium; it is separated obscurely from an overlying regressive nonmarine shale containing less than 100 ppm lithium. The situation for the section be— tween the Middle Kittanning and Lower Freeport coals is not clear because the sample for lithium analysis was a composite of the entire unit and the possible variation in lithium concentration from botttom to top is not deter- mined. The Glasgow and Vanport sections are not interpret— able in these simplistic terms, perhaps partly because neither field observations nor samples were closely enough spaced to reveal a complex depositional se- quence. Keith and Degens (1959, p. 43—44) did not explain the mineralogical basis for the difference in lithium con- tents between marine and nonmarine shales other than to suggest that lithium and some other trace elements would be more abundant in the marine environment and might be incorporated in clay mineral structures. It was also suggested that lithium can proxy for mag- nesium in illite and chlorite. Nicholls and Loring (1962, p. 207—208) deduced that a substantial amount of the lithium in their rocks was an original constituent of de- trital illite. At concentrations of 150 ppm or so, it seems unlikely that the mode of occurrence of lithium in polymineralic shales can be determined. ARMSTRONG AND BUTLER COUNTIES, PENNSYLVANIA A few samples were collected at the Heilman and West Valley coal strip mines in Armstrong County (HM and WV in figure 64) and the McCall mine in Butler County (McC in fig. 64). Very wet ground made it im- possible to observe the rocks well or to collect samples systematically. The stratigraphic position of the coals is uncertain, but they are within the Allegheny Group. The lithium concentrations are much lower than in Beaver County, the highest value being 160 ppm in a dark-gray shale at the West Valley mine. Only two other samples contain 100 ppm or more. One underclay (McCall mine, table 57) contains only 13 and 78 ppm lithium in two samples. These rocks in Armstrong and Butler Counties may have been deposited in an area of little marine influence even though transgressive and regressive cycles are present (Edmunds, 1968, p. 30). On the other hand, if only a limited amount of lithium were available in the environments of deposition, then only small amounts of lithium could be incorporated in the rocks regardless of their marine or nonmarine origin. 133 FLINT CLAY Clear-field, Jefferson, and Somerset Counties Samples from these counties were provided by the General Refractories Company from a collection of bulk samples maintained at the Company’s field office at West Decatur, Pa. The collection had accumulated over a period of years, and most if not all the samples repre— sent mined-out deposits. Most of the samples are flint clay2 that commonly occurs as an underclay. The flint clay in National Bureau of Standards standard sample 97 was known to be a Mercer flint clay from Clearfield County and to contain 0.23 percent Li20 (1,070 ppm Li) (Bureau of Standards, 1931, 1955). Consequently, the opportunity to investigate the lithium content of mater- ial of this kind was a welcome one even though field observations could not be made. Most samples of flint clay contain more lithium than do any other clayey rocks analyzed, 17 of 20 samples containing 300 ppm or more and 10 of these 17 contain- ing more than 500 ppm (fig. 69). The range of concen- tration is 41 to 1,400 ppm. The lowest lithium concent- ration is found in a diaspore clay, which is included here with flint clay for convenience. The next lowest values, 150 and 200 ppm, come from Somerset County and may indicate regional variation in lithium content of flint clays in this part of Pennsylvania. Although most of the samples come from Clearfield County, the single sample fromjefferson County to the northwest contains 700 ppm lithium (table 57), which is in the upper range of values for all flint clay samples. According to X-ray diffraction analysis (table 57), the flint clays are made up almost entirely of well-crys- tallized kaolinite. Illite is a significant component of the clay fraction only in sample 801, which probably should be classified as a semi-flint clay (Hosterman, 1972, p. 4). However, illite is also present in detectable amounts in several other samples. Quartz was detected in only two samples. Boehmite is present in three samples in detect- able amounts that might be as large as 5—10 percent. Only diaspore was detected in the diaspore sample. 0 5 SAMPLES lomspone 3 g ‘ é CA.A+ +A|A.A++ |‘ Al 1| 0 200 400 I000 1200 I400 600 600 LITHIUM, 1N PARTS PER MILLION FIGURE 69.—Frequency distribution of lithium concentrations in flint clays and one diaspore clay, from Clearfield, jefferson, and Somer— set Counties, Pa. "‘Flint clay is defined as a sedimentary, microcrystalline clay (rock) composed dominantly of kaolin, which breaks with a pronounced conchoidal fracture and resists slacking in water" (Keller, 1968, p. 113). 134 None of these mineralogic variations, except diaspore, have any evident relation to the lithium contents of the samples. A section through the Mercer flint clay bed is shown in figure 70. Each sample represents about 15 cm (6 in.) of the bed (G. P. Jones, General Refractories Co., writ- ten commun., 1975); the bed is thus about 180 cm (6 ft) thick. The upper 75 cm of the bed contains from 840 to 900 ppm lithium in contrast to the lower 105 cm of the bed which contains from 320 to 600 ppm lithium. The bottom sample contains 450 ppm compared to the sam- ple just above which contains 330 ppm. This marked pattern of variation within the bed does not correlate with any megascopic features of the samples and con- trasts with the uniform mineral composition of the sam- ples based on X-ray diffraction. The pattern of variation is similar, however, to that for germanium and several trace elements in coal, which is interpreted as indicating post-burial addition of elements to the coal from diagenetic, or later, circulating solutions (Nicholls, 1968, p. 289—290). Data are needed on the behavior of other elements besides lithium in sections such as this before any explanation can be offered. 0— m _ E |_ 45— LLI E _ .— Z _ LL] 0 E 90— U). _ U) LL] 2 _ X Blas— I '— nso I I I I I I I 300 400 500 600 700 800 900 I000 LITHIUM, IN PARTS PER MILLION FIGURE 70.—Variations in lithium concentrations within the Mercer clay bed, Galbraith Property, Clearfield County, Pa. LITHIUM RESOURCES AND REQUIREMENTS BY THE YEAR 2000 OCCURRENCE OF LITHIUM IN CLAY MINERALS The mode of occurrence of lithium in most clay min- erals is not definitely known. In hectorite and stevensite, however, lithium seems to occur with magnesium in the octahedral layer (Deer and others, 1962, p. 230; Weaver and Pollard, 1973, p. 79). This mode of occurrence of lithium has led to the interpretation that lithium and magnesium are similarly paired in other clay minerals such as illite (Nicholls and Loring, 1962, p. 217—218; Tardy and others, 1972, p. 407). The lithium contents of perhaps 300 ppm or so in some of the shales and illite-rich underclays may be explainable in this way. Some lithium also may be held in ion-exchange posi- tions, but the amounts should be very small in clay min- erals in near-surface rocks, in moist climates such as in Pennsylvania, because any Li+ likely would be displaced by more abundant cations. Lithium is the most easily replaced of the common cations in ion-exchange proces— ses (Grim, 1968, p. 212). Neither magnesium-lithium pairing nor cation ex- change seems to explain the relatively large amounts of lithium found in these flint clays. Nearly pure kaolinites contain very little magnesium. The 14 readily recogniz- able sedimentary kaolinites for which Weaver and P01- lard (1973, p. 132, 134) give chemical analyses contain only 0.05 to 0.46 percent magnesium oxide (0.03—0.28 percent magnesium). Most of the alkali and alkaline earth elements in kaolinite are attributable to im- purities; Deer, Howie, and Zussman (1962, p. 200) give 0.15 percent magnesium oxide (0.09 percent mag- nesium) as the average magnesium content of a number of analyses of kaolinite with few impurities. This amount of magnesium would be equivalent to about 250 ppm lithium if Horstman’s theoretical “most effective magnesium-lithium ratio” of 3.5 is applicable, which is unlikely. The magnesium-lithium ratio ranges from 6 to 490 and the median value is 30 (Horstman, 1957, p. 7, 25). Foose (1944, p. 568) gives an analysis of a typical flint clay from Clearfield County that contains 0.28 per- cent magnesium oxide (0.17 percent magnesium), which would correspond to about 500 ppm lithium according to Horstman’s most effective ratio if all the magnesium is assigned to kaolinite, but only about 100 ppm lithium according to the median ratio of 30. Although mag- nesium analyses are not available for the flint clays dis- cussed here, the published data suggest that mag- nesium-lithium pairing may have little to do with the lithium content of flint clays. The cation-exchange capacity of kaolinite is the small- est of all the clay minerals and is chiefly an effect of broken crystal edges (Grim, 1968, p. 189, 193). The cation—exchange capacity of sample Kal (table 57) is about 9.5 milliequivalents per 100 grams ( (H. C. Star- key, U.S. Geological Survey, written commun., 1976). LITHIUM RESOURCES AND REQUIREMENTS BY THE YEAR 2000 The leachate contained lithium equivalent to removing 22 ppm of the 1,400 ppm that the sample contains. Lithium in this sample thus is very tightly held. If kaoli- nite is the mineral in these samples in which lithium is strongly bound in structural positions, substitution to the extent of 1,000 ppm or so has not been documented. Alternatively, the lithium in these flint clay samples may not be associated at all with kaolinite but may be occurring in a specific lithium-bearing mineral. The amounts of such a mineral could be so small as to go undetected in whole-sample X-ray analysis. Only 1 per- cent of a mineral containing 1.5 percent lithium (3.2 percent lithium oxide) would need to be present in the sample for the sample to contain 150 ppm lithium. If the lithium content of a sample is based on such a min- eral, a sample containing 1,000 ppm lithium would con- tain only about 7 percent of the mineral. Such a mineral could be derived from the same source minerals or rocks from which the kaolinite was derived. This seems a logical origin for the few to several hundred ppm lithium found in the kaolinites of Great Britain and other kaolinites derived by alteration or weathering of igneous rocks (Horstman, 1957, p. 25). The Pennsylvania kaolinites, however, and the detrital rocks with which they are associated were derived pre- dominantly from sedimentary rocks of pre-Penn- sylvanian age (Glass, 1972, p. 88) in which lithium- bearing minerals are not likely to occur. In addition, lithium-bearing flint clays in Missouri (National Bureau of Standards, 1969) seem to have been derived ulti- mately by weathering of a terrane made up chiefly of carbonate rocks of pre-Pennsylvanian age, in which it also seems improbable that lithium-bearing minerals would occur. This line of thought suggests that the lithium-bearing mineral in flint clays, if such a mineral exists, is formed by the same processes by which the flint clays themselves are formed. ORIGIN OF UNDERCLAYS AND FLINT CLAYS The origin of underclays, of which flint clays are a special type, still is much debated despite persuasive ar- guments for each of two principal viewpoints. The first view is that underclays, in which kaolinite is more abun- dant than in the underlying shales, formed by the sedimentational concentration of clay in the quiet waters that preceded the development of most coal swamps. The beginning of the coal swamp, which may be repre- sented only by roots and disseminated plant fragments in the underclay, may have had a preferential effect by trapping the very fine sediment. This was roughly the View of Schultz (1958) which he based on underclays in the Eastern Interior Coal Region, and of Wilson (1965) which he based on underclays in Great Britain. They 135 observed that highly kaolinitic underclays are usually associated with highly kaolinitic shales, from which they concluded that most of the kaolinite must have de- veloped before the sediment entered the coal swamp. They believed that only minerals highly susceptible to leaching, such as chlorite, were systematically leached in most coal swamps; and more resistant minerals, such as illite or quartz, were not generally much altered except locally under extreme leaching conditions under which flint clays developed. The second View is that underclays formed by diagenetic alteration of normal polymineralic shales beneath the coal bed. Organic—rich solutions are considered to be powerful leaching agents. This was the view of Huddle and Patterson (1961) which they based on the published record and much experience with un- derclays, and of Patterson and Hosterman (1962), which they based on studies in Kentucky. They observed all gradations between highly kaolinitic plastic underlays and flint clays, from which they concluded that both types of rock resulted from the same process— authigenic leaching. Implicit in the concept is that flint clay and aluminum hydroxides are the result of inten— sification or long continuance of the process that formed kaolinite from other clay minerals. These two views have been combined to some extent by Keller, Westcott, and Bledsoe (1954), Keller (1968), and Williams, Bergenback, Falla, and Udagawa (1968), all of whom were concerned specifically with flint clays and more aluminous materials. Great emphasis is placed on the accumulation of colloidal material in back swamps and other special environments, the colloidal material then being leached and undergoing dialysis to remove cations. The fabric of flint clays consists of tightly interlocking crystals of kaolinite, which probably accounts for the flint-like character of typical flint clay (Keller, 1968). The fabric is similar to that of igneous rocks that have crystallized from a melt, but it seemingly results from reorganization and crystallization of col- loids. Curtis and Spears (1971) suggested that flint clays, and probably other kinds of kaolinite accumulations as well, form by Silication of aluminum hydroxides such as gibbsite, and aluminum oxyhydroxides such as boehmite and diaspore. Silication of gibbsite has been demonstrated in the Arkansas bauxite deposits (Goldman and Tracey, 1946;'Goldman, 1955) and was a widespread process there. Other occurrences of mixed gibbsite and kaolinite, such as the Eufaula bauxite de— posits in Alabama (Warren and Clark, 1965), may have had a similar origin. A flint clay in Australia (Loughnan, 1970) is unusual in that it occurs in a red-bed sequence instead of a coal- bearing sequence, and it contains odd minerals such as gorceixite (a complex phosphate) and gyrolite (a calcium 136 silicate cement mineral). The minerals seem to be au- thigenic. This flint clay may represent an original sedimentary deposit of aluminum hydroxide and oxyhydroxide minerals that were silicated to kaolinite after deposition. The hypothesis that silication of aluminum-hydroxide and oxyhydroxide minerals is a principal process in the development of flint clay has not, however, been tested by an integrated field and laboratory investigation. The contrasting ideas on the genesis of flint clay also identify two possible chemical regimens under which small amounts of lithium-bearing minerals could perhaps form: (1) an extreme leaching regimen that forms kaolinite or aluminum-hydroxide minerals from cation-rich aluminum-silicate minerals, whether this leaching takes place in the source area prior to the dep- osition of an underclay or whether it takes place after- wards; (2) a silication regimen, whether it takes place in a residual bauxite deposit or an underclay, or whether it takes place during or after the erosion and sedimenta- tion of aluminum-hydroxide minerals. The leaching hypothesis implies a relatively closed and short-lived chemical system, in that the process is confined to the environments of coal-forming sedimen- tation. The silication hypothesis implies a relatively open and long-lived chemical system in that silica would be introduced into the system from sources extraneous to the coal-forming environment and after, perhaps long after, the underclay is formed. Intuitively, the regimen of the silication hypothesis seems the more likely regi- men for the formation of small amounts of lithium- bearing minerals simply because an open, long-lived sys- tem allows for a wider range of geochemical conditions, even though it may be very difficult to learn about them. A possible lithium-bearing mineral in flint clays might be a lithium aluminosilicate, with or without other ca- tions, of a type ordinarily thought of as occurring only in igneous rocks; or such a mineral might be a lithium aluminate. Considering the startling range of authigenic silicate minerals found in the oil shale of the Green River Formation (Milton and Eugster, 1959; Milton and others, 1960) and the unusual minerals in the Australian flint clay (Loughnan, 1970), the possibility of lithium— bearing minerals in flint clay should not be dismissed peremptorily. CONCLUSIONS Further investigation of lithium in flint clay and re— lated materials is needed to determine the applicability of the three alternative modes of occurrence of the lithium that now seem possible: (1) in the well— crystallized kaolinite of the flint clays by structural sub- stitution in the kaolinite, perhaps in association with magnesium; (2) in detrital minerals derived from the LITHIUM RESOURCES AND REQUIREMENTS BY THE YEAR 2000 source areas of the kaolinite; (3) in a previously unrec- ognized authigenic mineral that formed under the same conditions and by much the same processes as the kaolinite was formed. Such an investigation has significance not only for clarification of the origin of flint clays and related aluminous materials, but also for the possible discovery of sedimentary rocks that contain enough lithium to be ultimately of economic interest. REFERENCES Curtis, C. D., and Spears, D. A., 1971, Diagenetic development of kaolinite: Clays and Clay Minerals, v. 19, no. 4, p. 219—227. Deer, W. A., Howie, R. A., and Zussman, J., 1962, Rock-forming minerals—Volume 3, Sheet silicates: New York, John Wiley and Sons, 270 p. Edmunds, W. E., 1968, Geology and mineral resources of the north- ern half of the Houtzdale 15-minute quadrangle, Pennsylvania: Pennsylvania Geol. Survey, 4th ser., Atlas A85ab, 150 p. Foose, R. M., 1944, High—alumina clays of Pennsylvania: Econ. Geol- ogy, v. 39, no. 8, p. 557-577. Gary, Margaret, McAfee, Robert, Jr., and Wolf Carol, eds., 1972, Glossary of geology: Washington, D.C., Am. Geol. 1nst., 858 p. Glass, G. B., 1972, Geology and mineral resources of the Philipsburg 71/2-minute quadrangle, Pennsylvania: Pennsylvania Bur. Topo. and Geol. Survey, Atlas 953, 241 p [1974]. Goldman, M. 1., 1955, Petrology of bauxite surrounding a core of kaolinized nepheline syenite in Arkansas: Econ. Geology, v. 50, no. 6, p. 586—609. Goldman, M. I., and Tracey, J. I.,Jr., 1946, Relations of bauxite and kaolinite in the Arkansas bauxite deposits: Econ. Geology, v. 41, no. 6, p. 567—575. ' Grim, R. E., 1968, Clay mineralogy [2d ed.]: New York, McGraw-Hill, 596 p. Heier, K. S., and Billings, G. K., 1972, Lithium [Chap.] 3, Part B, in Wedepohl, K. H., ed., Handbook of Geochemistry, Volume 2—1: Berlin, Springer-Verlag, looseleaf. Horstman, E. L., 1957, The distribution of lithium rubidium, and cesium in igneous and sedimentary rocks: Geochim. et Cos- mochim. Acta, v. 12, nos. 1—2, p. 1—28. Hosterman, J. W., 1972, Underclay deposits of Somerset and eastern Fayette Counties, Pennsylvania: U.S. Geol. Survey Bull. 1363, 17 p. Huddle, J. W., and Patterson, S. H., 1961, Origin of Pennsylvanian underclay and related seat rocks: Geol. Soc. America Bull., v. 72, no. 11, p. 1643—1660. Keith, M. L., and Degens, E. T., 1959, Geochemical indicators of marine and fresh-water sediments, in Abelson, P. H., ed., Re- searches in geochemistry: New York, John Wiley and Sons, p.. 38—61. Keller, W. D., 1968, Flint clay and a Hint-clay facies: Clays and Clay Minerals, v. 16, n0. 2, p. 113—128. Keller, W. D., Westcott, J. F., and Bledsoe, A. 0., 1954, The origin of Missouri fire clays, in Swineford and Plummer, eds., Clays and Clay Minerals, Natl. Research Council Pub. No. 327, p. 7—46. Loughnan, F. C., 1970, Flint clay in the coal-barren Triassic of the Sydney Basin, Australia: Jour. Sed. Petrology, v. 40, no. 3, p. 788—854. Miesch, A. T., 1967, Theory of error in geochemical data: U.S. Geol. Survey Prof. Paper 574—A, 17 p. Milton, Charles, Chao, E. C. T., Fahey,J.J., and Mrose, M. E., 1960, Silicate mineralogy of the Green River formation of Wyoming, LITHIUM RESOURCES AND REQUIREMENTS BY THE YEAR 2000 Utah, and Colorado: Internat. Geol. Cong., 2lst, Copenhagen 1960, Rept., pt. 21, p. 171—184. Milton, Charles, and Eugster, H. P., 1959, Mineral assemblages of the Green River formation [Colo.-Utah-Wyo.], in Abelson, P. H., ed., Researches in Geochemistry: New York, john Wiley and Sons, p. 118—150. National Bureau of Standards, 1931, Certificate of analyses, Standard Sample No. 97: Flint clay, 1 p. 1955, Revised values for alkalies in NBS Standard Samples 76, 77, 97, 98, 102 and 104: l p. —)——1969, Certificate of analysis, Standard Reference Material 97a: Flint clay, I p. icholls, G. D., 1968, The geochemistry of coal-bearing strata, in Mur- chison, D. G., and Westoll, T. S., eds., Coal and coal-bearing 73 strata: New York, Elsevier, p. 269-307. icholls, G. D., and Loring, D. H., 1962, The geochemistry of some British carboniferous sediments: Geochim. et Cosmochim. Acta, v. 26, no. 2, p. 181—223. Ohrdorf, Renate, 1968, Ein Beitrag zur Geochimie des Lithiums in Sedimentgesteinen: Geochim. et Cosmochim. Acta, v. 32, no. 2, p. 191—208. O'Neill, B. J., jr., 1973, Clay and shale resources in the Greater Pittsburgh region of Pennsylvania; Phase III, Clay-shale sample test data—text and tables: U.S. Geol. Survey open-file rept., 294 p., 1 table. Iatterson, E. D., 1963, Coal resources of Beaver County, Pennsyl- vania: U.S. Geol. Survey Bull. 1143—A, 33 p. Patterson, S. H., and Hosterman, j. W., 1962, Geology and refractory clay deposits of the Haldeman and Wrigley quadrangles, Ken- ABSTRACT Three lithium minerals are found in Arkansas, all in the Ouachita ountains. Cookeite, a lithium chlorite, is associated with many small ydrothermal milky-quartz veins in the jackfork Sandstone of Missis- ippian and Pennsylvanian age, in an area extending some 77 km (48 i) from Little Rock, in Pulaski County, westward into Perry County. ookeite has been tentatively identified in association with antimony, ead, and zinc in quartz veins in Sevier County, some 240 km (150 mi) outhwest of Little Rock. Taeniolite, a rare lithia mica, occurs in moky-quartz veins and in recrystallized novaculite adjacent to alkalic yenites at Magnet Cove, Hot Spring County, and in a milky-quartz- fluorite vein in the “V” alkalic intrusive, Garland County, some 16 km (10 mi) west of Magnet Cove. Lithiophorite, a lithium-cobalt- aluminum manganese oxide, has been reported in the manganese ore deposits in Polk and Montgomery Counties; and it also occurs with quartz veins in many other Arkansas localities. The bromine-rich brines in the Upper jurassic Smackover Forma- tion in Columbia County, southwestern Arkansas, contain as much as 445 parts per million Li; smaller quantities are reported from various water wells and springs, hot and cold, in the State. The widespread distribution of lithium minerals, for the most part 137 tucky, with a section an Coal resources, by John W. Huddle: U.S. Geol. Survey Bull. 1122—F, 113 p. [1963]. Rankama, K. K., and Sahama, T. G., 1950, Geochemistry: Chicago, Chicago Univ. Press, 912 p. Ronov, A. B., Migdisov, A. A., Voskresenskaya, N. T., and Korzina, G. A., 1970, Geochemistry of lithium in the sedimentary cycle: Geochemistry Internat. v. 7, no. 1, p. 75—102. Schultz, L. G., 1958, Petrology of underclays: Geol. Soc. America Bull., v. 69, no. 4, p. 363—402. 1964, Quantitative interpretation of mineralogical composition from X—ray and chemical data for the Pierre Shale: U.S. Geol. Survey Prof. Paper 391—C, 31 p. Tardy, Yves, Krempp, Gerard, and Trauth, Norbert, 1972, Le lithium dans les mineraux argileux des sediments et des sols: Geochim. et Cosmochim. Acta, v. 36, no. 4, p. 397—412. Turekian, K. K., and Wedepohl, K. H., 1961, Distribution of the ele- ments in some major units of the Earth’s crust: Geol. Soc. America Bull., v. 72, no. 2, p. 175—191. Warren, W. C., and Clark, L. D., 1965, Bauxite deposits of the Eufaula district, Alabama: US. Geol. Survey Bull. 1199—E, 31 p. Weaver, C. E., and Pollard, L. D., 1973, The chemistry of clay miner- als: New York, Elsevier, 213 p. Williams, E. G., Bergenback, R. E., Falla, W. S., and Udagawa, S., 1968, Origin of some Pennsylvanian underclays in western Pennsylvania: Jour. Sed. Petrology, v. 38, no. 4, p. 1179—1193. Wilson, M. j., 1965, The origin and geological significance of the South Wales underclays: _]0ur. Sed. Petrology, v. 35, no. 1, p. 91—99. \/_\ LITHIUM MINERALIZATION IN ARKANSAS By CHARLES G. STONE and CHARLES MILTON, Arkansas Geological Commission, Little Rock, AR; The George Washington University, Washington, DC; US. Geological Survey, Reston, VA unrecognized until recently, warrants further search for economic concentrations oflithium in Arkansas. Furthermore, because the three known lithium minerals appear very similar to common non-lithium species, their discovery may require somewhat sophisticated methods of identification, such as emission spectroscopy, and X-ray diffraction. LOCATION OF LITHIUM MINERALIZATION IN ARKANSAS The lithium minerals cookeite, taeniolite, and lithiophorite occur in Arkansas only in the Ouachita Mountains, in the west-central part of the State (table 58; fig. 71). In this region, cookeite and taeniolite each TABLE 58.fiArkan.sas lithium minerals I r Lizo Mineral Formula (in percent) Reference Taeniolite ............ KLnggSLOWF, 4 Miser and Stevens, 1938 Lithiophorite ________ LiAl,Mn3093H 0 0—8 Fine and Frommer, 1956 Cookeite ______________ LiAISSigOMOPf a 3 Miser and Milton, 1964 138 MISSOURI REGION ARKANSAS VALLEY For. SmiIh OUACHITA MOUNTAIN L L LREGIONT OKLAHOMA MISSISSIPPI 0 50 KILOMETHES 0 50 MILES B .El 007060 33-— -33. LOUISIANA 912' 0) |94- I FIGURE 71.—Major areas of lithium mineralization in Arkansas. C, cookeite; T, taeniolite; L, lithiophorite; B, bromine-rich brines of Smackover Formation. LITHIUM RESOURCES AND REQUIREMENTS BY THE YEAR 2000 occur in a limited area, but lithiophorite is widespread. Figure 72 shows the distributions of the lithium miner- als in more detail. The bromine-rich brines of the Smackover Formation, which contain lithium, are in the southwest corner of the State, beneath the Coastal Plain sediments. LITHIUM CONTENT OF ARKANSAS MINERALS Figure 73 is a histogram showing spectrographic de- terminations of lithium content of many Arkansas min- erals. It is evident that only the three minerals discussed in this paper have significant lithium content—the “seri- cite” and “fault gouge” are probably altered cookeite. Table 59 also lists these minerals, with the serial number of the X-ray powder pattern by which the mineral was identified. TAENIOLITE The first lithium mineral reported in Arkansas was taeniolite (KLng28i4010F2) at Magnet Cove in Hot Spring County (Miser and Stevens, 1938). It contains 4 BRECC/A /// /// /// /// /// /// /// /// l// 7/77/F/77—/_/7777 // //// ////////////////// ASYNCLINO ///////////// ///////////// /////////// o ////////////,, ///////////// //////////////:. ///////////// ////////// 93606— PERRY COUNTY —C?[ :_—' \FT——_l PULASKI COUNTY —|—FRONTAL ‘ OUACHITAS ,, C (I, LCC I A I / _ _I_____ [ —34°45' \ 7M LI ' SALINE COUNTY GARLAND [ — COUNTY .____, BENTON - BROKEN BOW UPLIFT l E : 541 [NE IO KILOMETRES Io MILES FIGURE 72.—Spatial distribution of lithium minerals in central Arkansas, showing occurrences of cookeite (C), lithiophorite (L), and taeniolite (T). Dark shading represents Cretaceous alkalic intrusives as shown by Hollingsworth (1967). Geology is from Haley (1976). 3.0 \\ - § I.o—E_’ : :5 \§ 3 E. 2\\\§ 2 3 §\\ €§§<§§§ wéjg§§§§ MINERAL OR SAMPLE TYPE FIGURE 73.—Precentage of lithium in Arkansas minerals and fault gouge. Numbers in parentheses indicate the number of determina- minerals: adularia (1); albite with quartz (2); K-Na feldspar (1); K feldspar with quartz (1); anthraxolite (1); Chlorite (2); dickite (5); halloysite-kaolinite (6); hornblende (1); lithiophorite (3); K-Fe mica (1); miserite (l); muscovite (1); montmorillonite (2); narsarsukite (1); natrolite (1); pyrophyllite (2); narsarsukite (l); rutile (1); sphal- erite (1); tetrahedrite (1); thuringite (1); wollastonite (1). percent L120. Previously, taeniolite had been discovered in Narsarsuk, Greenland, and again found in the Kola Peninsula, U.S.S.R. (Later it was found elsewhere in the Soviet Union.) In Arkansas, Milton has now found a second occurrence of Taeniolite at the “V” intrusive in Garland County, some 16 km (10 mi) west of the first. The two Arkansas localities are shown in figures 71 and 72. In Arkansas, as elsewhere, taeniolite is associated with LITHIUM RESOURCES AND REQUIREMENTS BY THE YEAR 2000 139 TABLE 59.fiArkama5 minerals on which lithium determinations have been made [N umbers are X-ray films identifying the minerals. A spectrographic analysis was made of each. Only those italicized contain significant lithium} Adularia 348 Albite+quartz 126, 639 Na-K feldspar 645 K-Na feldspar+quartz 125, 16976, 16980 Anthraxolite Chlorite 16662, 16669 Cookeite 124, 1275, 1280 Cry tomelane 1261, 1267, 1213 Dic ite 589, 17028, 16873, 16875, 17016 Hallo site-Kaolinite 16874, 16875, 16983 (5?), 16877, 16902, 16868 Horn lende 7/1/65 Union Carbide Lithiophorite 573, 590, 597, 598, 1268 K—Fe mica 588 Miserite Muscovite 16989, 1281, 1257 Sericite 1253, 1257 Montmorillonite 16664, (0.001%), 16666 Narsarsukite 16671 Natrolite 16675 Pyrophyllite 17017, 17018 Quartz+muscovite+cookeite(?) 1247, 1248 Rectorite 16666, 16674, 16677, 16992 Rutile 341 Sphalerite Stevensite 1285 Taeniolite 17037 Tetrahedrite 130 Thuringite 16899, 16977, 1282 Wollastonite alkalic syenitic intrusives. The Arkansas taeniolite is re- lated to early Late Cretaceous volcanism (Zartman and others, 1967). At Magnet Cove, taeniolite occurs in small smoky- quartz-brookite veins and in recrystallized Arkansas Novaculite of Devonian-Mississippian age; it is as— sociated with dickite, brookite, and rutile (Miser and Stevens, 1938; Fryklund and Holbrook, 1950). Signifi- cant concentrations of vanadium and columbium accom- pany the taeniolite. Hermanjackson (written commun., 1973) found that the quartz-brookite veins crystallized at a temperature under 440°C. Taeniolite generally forms less than 1 percent of the veins or of altered chert-novaculite; but locally there are concentrations as high as 20 percent (Fryklund and Holbrook, 1950). Erickson and Blade (1963) also reported 0.03 percent lithium in some molybdenite- and rutile-bearing veins at Magnet Cove, but the lithium mineral has not been identified. At the “V” intrusive, taeniolite occurs in a milky- quartz-fluorite vein, associated with opaline silica or chalcedony, pyrite, sulfur, gypsum, rutile, anatase, goethite,jarosite, sphalerite, dolomite, zircon, sanadine, and tourmaline; present in lesser amounts are barite, witherite, galena, brookite, and zeolite minerals; there are also small quantities of gold and silver. In 1955, taeniolite was synthesized; it was the first synthetic mica. Miller and Johnson (1962) reported a synthetic fluor—mica intermediate between taeniolite and hectorite (LingfixSigOgo (OH)4, AMPLITUDE—a — Good conductor —Poor conductor IO MILES l l IO KILOMETRES FIGURE 83.—Contoured airborne electromagnetic survey data for the Willcox Playa area, Arizona. Figure 84 also shows the location of a prominent total magnetic field anomaly defined by the airborne magne- tic survey referenced previously. This is a magnetic high of about 300 gammas whose contours define a ridge- shaped feature. The trend of the feature is oblique to the gravity and topographic trend of Sulphur Springs Valley. A general interpretation of the magnetic anomaly’is that it is due to a volcanic intrusive dikelike feature related to the volcanic activity at the south end of the valley. We note that the local gravity contours transect Willcox Playa and tend to parallel the trend of the magnetic anomaly. The magnetic feature then is of significant dimensions to alter the main trend of the local gravity low. The following statements summarize the important geological and geophysical features of the previous dis- cussion and the implications for lithium occurrences. 1. A thick section of conductive sediments on the order of a few hundred meters is located at the south end of Willcox Playa. Though the sediments occur within a gravity low they are not associated with closed gravity lows. This corresponds roughly to the relationship between the lithium occurrence at Clayton Valley and the associated gravity. The resistivities of the conductive rocks are the same generally as those associated with the lithium pro- ducing area at Clayton Valley. . A thick accumulation of sediments is indicated by the gravity low northwest of Willcox. This sedimentary sequence could be the site of evapo- rite deposits. . Volcanic activity seems to have been frequent in the southern portion of Sulphur Springs Valley. . The airborne magnetic survey indicates the pres- ence of a dikelike volcanic body beneath the Willcox Playa. This volcanic body could be a source of hydrothermal solutions. . Base metal mining has taken place most intensively i LITHIUM RESOURCES AND REQUIREMENTS BY THE YEAR 2000 45‘ 09° 30‘ l l EXPLANATION emu/m CONTOUHSerushea when appmimucw loomed. Shading mum-s areas 0! lower gramy. Contour mverval 2 milligals GRAwW STATION Magnetic anomaly axis Conductive ‘ sediments sz‘j 00“ i‘ i O 5 l0 MlLES ‘ o 5 FFGURE 84.—Generalized Bouguer gravity map for the Sulphur ‘ Springs Valley, Ariz. IO KILOMETRES l l within the exposed rocks at the south end of the valley. Base metal mining has also been extensive l in the exposed strata adjacent to Clayton Valley. l The overall picture which emerges from this analysis is that the south end of Sulphur Springs Valley seems to be the most likely area for the occurrence of lithium. he interpreted thick sequence of sediments north of illcox should be investigated for potential evaporite ‘eposits. An integrated electromagnetic survey and eismic survey would provide the best information about jhis sedimentary sequence. DISCUSSION The interpretation of geophysical surveys given in the revious discussion indicates areas for research in ithium exploration methods. As discussed in the intro- uction the most important general research objective is 0 define physical models for lithium deposits. We will oncentrate on the geophysical aspects of this general esearch objective. One aspect of a model for lithium exploration is the lnature and character of basin sediments. Clearly an im- lportant question is how typical or atypical is Clayton lValley, the only valley of its kind producing lithium ‘brine, in comparison to other sedimentary basins within ‘the Basin and Range Province. The characteristics of the sedimentary sequence is a consideration in seeking an ‘answer to this question. The previous examples of elec- 153 trical surveys have shown that they are valuable tools in defining the nature of a sedimentary sequence. How- ever, a very real limitation exists in their interpretation since only general conclusions can be made about the lithology itself. Fortunately the state-of—the art for elec- trical methods has made rapid advances in recent years. For example, using alternating electrical currents and measuring the resistivity as a function of the frequency has been advocated as a means to distinguish copper sulfides from noncopper sulfides (Zonge and Wynn, 1975). We are currently starting a research program which will employ such a method to try to deduce more about the nature of the basin sediments. An important part of this research is the measurement of resistivity as a function of frequency in the laboratory for small rock samples whose mineralogy and geochemistry are rela- tively well known. We are confident that this laboratory study, when used in conjunction with field measure- ments, will enable a geophysical estimate to be made of the nature of the sedimentary sequence. Another class of geophysical methods which can be applied in lithium exploration are the bore-hole logging techniques. These methods have an application in sup- port of surface electrical measurements and in supple- menting lithologic logging. For example, in situ meas— urements of the physical properties of a given sedimen- tary sequence greatly aids interpretation of geophysical surveys used to map the lithologic variations. Combined use of borehole logging methods and a surface geophys- ical survey has proven to reduce drilling costs signifi- cantly. Neutron activation methods are one recently de- veloped class of borehole logging methods which have a potentially interesting application to lithium explora- tion. The use of lithium in shielding for nuclear reactors (for example, activation potential) may also prove to lead to a method for in situ detection of lithium. Another broad class geophysical method which will prove useful in characterizing sedimentary deposits in the Basin and Range Province is the remote sensing method. One obvious application is thermal or infrared remote sensing. If indeed lithium has a hydrothermal origin or has a correlation with hydrothermal activity, then shallow lithium deposit target areas may be as— sociated with thermal anomalies. CONCLUSION Lithium deposits will not be defined by a characteristic geophysical anomaly at least with present geophysical methods and interpretation techniques. Consequently geophysical methods cannot be thought of as a black box which produces a geophysical anomaly directly indicat- ing the location or presence of a mineral deposit. Geophysics must be thought of as analogous to a tool box with each tool being a geophysical method appro- priate to the geological problem at hand. The most suc- 154 LITHIUM RESOURCES AND REQUIREMENTS BY THE YEAR 2000 cessful application of geophysical methods to any min- eral exploration problems will involve integration of several geophysical and geological techniques. REFERENCES Dempsey, W. ]., Fackler, W. D., and others, 1963, Aeromagnetic map of the Cochise quadrangle, Cochise County, Arizona: US. Geol. Survey Geophys. Inv. Map GP—4l3. Dobrin, M. B., 1960, Introduction to geophysical prospecting [2d ed.]: New York, McGraw-Hill, 446 p. Griffiths, D. H., and King, R. F., 1965, Applied geophysics for en- gineers and geologists: London, Pergamon Press, 223 p. Keller, G. V., and Frischknecht, F. C., 1966, Electrical methods in geophysical prospecting: Oxford, Pergamon Press, 519 p. Parasnis, D. S., 1962, Principles of applied geophysics: London, Met- huen, 176 p. Peterson, D., 1966, Principle facts for gravity stations in Sulphur Springs Valley, Arizona: US. Dept. Commerce Natl. Tech. Inf. Service, Springfield, Va., Rpt. PB212—99, 12 p. Wilson, C. W., 1975, Bouguer gravity map of Clayton Valley, Nevada: US. Geo]. Survey open-file rept. 75—333. Wynn, ]. C., 1974, Electromagnetic coupling in induced polarization: Univ. Arizona, Ph.D. thesis, 150 p. Zohdy, A. A. R., Eaton, G. P., and Mabey, D. R., 1974, Application of surface geophysics to ground-water investigations: U.S. Geol. Survey Water—Resources Inv. Techniques, Book 2, Chapter D1, 116 p. Zonge, K. L., and Wynn,]. C., 1975, Recent advances and applications in complex resistivity measurements: Geophysics, v. 40, no. 5, p. 851—864. \f‘ LITHIUM—DATA BASES AND RESOURCE ESTIMATES: PROBLEMS AND POTENTIAL By ALLEN L. CLARK, JAMES A. CALKINS, DONALD SINGER, and MARY A. URICK, U.S. GEOLOGICAL SURVEY, RESTON, VA ABSTRACT Data files contain raw or disaggregated information on mineral de- posits and commodities. Computerized data files offer the only means presently available by which all data items (fields) are individually ad- dressable. One operational data file is CRIB (Computerized Resources Information Bank), which is a general purpose inventory and refer- ence file on metallic and nonmetallic mineral deposits. Other files are being constructed for special purposes. A computer file on lithium resources should contain detailed infor- mation related to the following main categories: record identification, name and location, description of deposit, analytical data, and produc- tion and (or) reserves. The US. Geological Survey is developing improved methods for the handling of mineral resource data. These include computerized data files and predictive resource models. A resource model is a system of postulates, data, and inferences which serves to predict the resource-related outcome when key vari- ables affecting a mineral commodity are changed. The objective is to provide several different kinds of mineral availability estimates, in- cluding: geological availability (search and occurrence model), technological availability (exploration models), and economic availabil— ity (economics models). Information on mineral resources should be or- ganized more efficiently and made more easily accessi- ble so that this information can be used to greater ad- vantage by interested parties both in government and in the private sector. As stated by McKelvey (1972), “Better methods for estimating the magnitude of potential min- eral resources are needed to provide the knowledge that should guide the design of many key public policies.” The rapid changes in the lithium picture in recent years—new sources of supply from subsurface brines and the likelihood of greatly increased demand due to new uses—emphasize the general need for rapid access to current information relating to our position toward minerals and for long range mineral resource apprai- sals. The US. Geological Survey is engaged in collecting and organizaing computerized information on mineral resources and in developing predictive resource models which will provide improved capability for making pol- icy decisions on this subject. This paper describes the methods being developed and suggests that some of these methods may be applic- able tO lithium resource studies. Mineral resources data files, loosely called data bases, contain the basic raw or disaggregated information on mineral deposits and mineral commodities. This infor- mation is rigorously and logically organized down to the smallest unit of information for which computer access is desired; each individual datum is stored in a computer field, which is individually addressable. It is important to decide beforehand the smallest level of information that will be required because, while it is possible to aggregate small components of data into larger groups (grouped 1 l LITHIUM RESOURCES AND REQUIREMENTS BY THE YEAR 2000 da a), it is not possible to do the reverse. lthough organizing the information into a com- p terized file is tedious, so is the use of a manual file. In the case ofa subject as complex as mineral resources, the cotnputerized operation offers some unique advantages n t available elsewhere. For example, the computer file ca be addressed down to the level of the individual field, plus those two or three fields within the record, th t have been sorted in advance, whereas the manual filp can only be addressed down to the level of the re- cqrd (document). Therefore only the computer file can disassemble the record into its individual parts, and for mlost uses of data this disassembly process is an initial requirement. lOne of the operational data files being used for the stbrage of data on mineral resources is called CRIB (Computerized Resources Information Bank) (Calkins and others, 1973). CRIB is a general-purpose inventory and reference file on metallic and nonmetallic mineral deposits. As ofjanuary, 1976, this file contained about 49,000 records of varying quality and completeness; these records were supplied to the file voluntarily from various sources. Very few records specific to lithium had bleen contributed to the file. An example of a typical rbcord stored in the CRIB file is shown in table 64. This rbcord contains certain key information on the Spor Mountain beryllium district, which contains lithium as a potential resource. A second example, the printed out- ut resulting from a search for lithium deposits in the wnited States, is shown in table 65. In this type of out- put, selected parts of the retrieved records are rear- anged into fixedlength fields, sorted, and then passed tb a report processing program which provides for pecialized output arrangements. Several other data files are being constructed to serve pecial purposes, including (1) a file specific to the min- ral economics of metallic mineral deposits, (2) a special le on the 900 largest producing mines of the world, (3) file on coal resources, and (4) a file on geothermal esources. In order to use the computer to store and manipulate ithium data, it will be necessary to organize the data for computer operation. In this regard we suggest the ollowing as a preliminary list of important categories of nformation that should be organized and collected: l Record Identification l Record number Deposit number l Cross index number l Date ‘ Name and Location l Deposit name ‘ Mining district/area/subdistrict 155 Country State County Latitude Longitude Description of Deposit Commodities present Deposit type Size of deposit (large, medium or small) Ore minerals Status of exploration/development Property (active or inactive) Workings (surface, underground, or both) Analytical Data Production and (or) reserves Annual production Cumulative production Reserves Reserves and potential resources The compilation of resource data into data bases pro- vides information which can be used in resource models to create long-term-supply curves. This ultimate pur- pose is directly analogous to the ultimate purpose of geologic data collection, which is the historical interpre- tation of the earth and its phenomena. In both cases, it is imperative that the ultimate purpose be clearly estab- lished so that data collection and data banks do not be- come an end unto themselves; rather, they should be regarded as necessary building blocks of analysis, and used like any other tool of geologic studies. Numerous steps must be taken, however, before one can produce the end product—a long-term-supply curve—and the first and most critical step is a mineral resource inventory. The most important reasons for such inventories are: (1) the increasing demands on world’s resources by both the developed and developing nations; (2) the need for more accurate data for short- and long-range planning; (3) the need to assess national endowment in order to effectively allocate other scarce commodities such as manpower and capital; and (4) the need to ascertain the best trade-offs in terms of world food demands, environmental degradation, rational growth programs, and multiple land use within de- veloped and developing nations. Once an inventory of a mineral resource has been completed and a data base developed, a large number of studies can be undertaken utilizing this inventory. Within the US. Geological Survey and elsewhere in government and industry, much emphasis has recently been placed on resource models of various types. This emphasis has caused a great deal of confusion, con- tradiction, and controversy over the use of models. I56 LITHIUM RESOURCES AND REQUIREMENTS BY THE YEAR 2000 TABLE 64,—An example of a typical record stored in the U.S. Geological Survey’s CRIB file CRIB MINERAL RESOURCES FILE, REVISION 8 RECORD IDENTIFICATION RECORD NOOOOOOOOQOOOO “000026 RECORD TYpEogoouooooo A U COUNTRY/ORGANIZATION. USGS FILE LINK IDoooooooo. RASS DEPOSIT NOOOOOOOCOOOO 496 REPORTER NAME: WORLo RONALD 6. DATE! 71 04 NAME AND LOCATION STATUS OF EXPLOR. 0R DEV. 4 DISTRICT/AREAIO00000000.. SPOR MOUNTAIN COUNTRY CODE..o.oooooonoo US STATE CODEoooooooooooooao 49 STATE NAMES UTAH COUNTY-nun.oo-ooooooaoooo JUAB QUAD SCALE QUAD N0 0R NAME 12 62500 TOPAZ MOUNTAIN LATITUDE LONGITUDE ALTITUDE 39-43-00N 113-12-00W COMMODITY INFORMATION COMMODITIES PRESENT: F GE U LI SIGNIFICANCE: MAJORoooooo F MINOR-00000 U BE COPRODUCTuo BYPRODUCTo- POTENTIALoo LI ORE MATERIALS (MINERALS.ROCKS¢ETC.): FLUORITE DEPOSIT TYPES: PIPESy VEINS PRODUCTION YES However, the purpose of data, data bases, and re- trate on lithium, but the general philosophy we present source models is to provide an indication of the long can apply to almost any mineral or energy commodity. term supply of a commodity. In this paper, we concen- The least that is required of any model is that it serve l LITHIUM RESOURCES AND REQUIREMENTS BY THE YEAR 2000 157 i TABLE 64.—An example of a typical record stored in the US. Geological Survey’s CRIB file—Continued NNUAL PRODUCTION (ORE AND COMMODITIES) ‘ ITEM ACC AMOUNT THOUSoUNITS 1 ORE 3 BE 4 U YEAR GRADE OR USE lCUMULATIVE PRODUCTION (ORE AND COMMODITIES) ‘ ITEM ACC AMOUNT THOUS.UNITS 8 F ACC 00000144 TONS SOURCE OF INFORMATION.. DASCH! 1964 ‘PRODUCTION COMMENTS. 9 a o RESERVES AND POTENTIAL RESOURCES ITEM ACC AMOUNT 1 CAFZ EST 00000362 TONS 2 BE LGE l 3 LI \ 4 U LGE SOURCE OF INFORMATION.. DASCHo 1964 ‘ COMMENTSIOIOOOOOIOOOOOQ EEOLOGY AND MINERALOGY LARGE LOW GRADE CAFZ YEARS GRADE OR USE 1943-1962! 75% CAFZ U FROM YELLONCHIEFI BE BY BRUSH BERYLLIUM CD.’ SMALL FLUORSPAR OPERATIONS CONTINUING FROM 1962 TO PRESENT. THOUS.UNITS GRADE OR USE 40% CAFZ LOW GRADE LOH GRADE (5%) ASSOC. HITH BE DEPOSITS. HOST AND/OR COUNTRY ROCKS AND AGEoo LIMESTONE ‘ ASSOCIATED IGNEOUS ROCKS AND AGE... DIKES i AGE OF MINERALIZATIONoogcooqooo0.00 TERTIARY GEOLOGICAL DESCRIPTIVE NOTESoouono- LIMESTONE. FLUORITE (5%) ‘REFERENCES l) STAATZ. USGS PROF. PAPER 415 ‘to predict actual physical happenings in some way relev- ant to man, either by allowing him to anticipate future iuncontrollable events or by demonstrating the possible lconsequences of various decisions. To date, many min- ‘ era] resource models have been deficient in meeting this requirement. i Reasons for past deficiencies include misunderstand- l ings between geologists and economists, inadequate re- ‘ source data, and little research on certain fundamental relationships. Some of these deficiencies, however, may be resolved by working backwards from the policy ques- FLUORITE IN IRREG PIPES AND VEINS IN IN HATER-LAID TUFF WITH BE! LI! AND MN OXIDES. tions to determine the type of information which is needed to answer the questions—the approach attemp- ted here. This approach leads to an overview of the type of mineral resource data which should be gathered. The purpose of resource estimates is to provide un- biased mineral resource information to decision makers, enabling them to make rational decisions which would insure the orderly supply of mineral materials without neglecting the proper use of land and the health and welfare of the people. Thus the need of the resource estimator to anticipate all future policy questions is im- LITHIUM RESOURCES AND REQUIREMENTS BY THE YEAR 2000 158 Oz 02 oz uhmhdxcun wn>h .0011 h~mouuo oz mw> nw> .mwnnn wn>k .ooam humonwo ww> uhuh¢xowm mm) mhxhdxomm mm» whnhqxown oz 02 mu> oz mhnhdxama wn>h oooza humoawo mm> uhuh wh~hh .ooau humonwo mw> ukuhdxown oz whuh wh~hb ooozu huwomuo 02 oz 02 02 oz 02 02 oz ma>h .ooaa humonwc m4 2: 3 mm ma 2: H4 on 3 m» @< D< “4 mwnhnooxxou nJ mm ”A 3 um u mw~h~oo:zoo n4 ~J n4 H4 H4 n4 H4 mwuhaoo:xoo H4 k mm on: cam zmw > "4 mm mm Nbo QJm UH: OJu on: “4 muuhnooxzou ~4 :uw um mm "4 m on: two mm H; 3 Nae NAc IoOINHIMAH oz: 13» w03»~9204 Ith xomlmmlmho 3mmu¢h~4<0¢4 oszz<2 wsz «thHJ NGQMJ awmzou uqz wzny (azhud wGQwJ xzbzmaomn maaaTUz~ we“ mew )Mo Zho :mo zhe :nc zoo 3no 25o zco znu mcapah<4 $2: a3» webbmeOJ z resource availability at present . . _ Discovery-- 0 price and technology Exploration drilling" No of D=f(W N __) w=f(F:C,d,S,l,—-) ' Q'f(W'S,-—) ' 1 1 . NUMBER OF WELLS —————r———— ————-——-——-———————-—-——————————————4 Govern-I- mento policies EXPLANATION World Fi, Frontier region 1 SUPP'Y i, Number of regions Substi- price m, One through any number "11'0" t, Function . _ effects Reserve accu- Production model-—Aggregatlon . W, Wells Feedback "‘U'Ol'O" PM P Price gives predicted production at gggiruction C: Cost various assumed soclo-economic d, Depth conditions 8, Deposit size I, Investment 2’ No. of D, Number of discoveries o . ; Reserves and N, culmber :f deposns go resources 0, oume 0 resource é recovery a TIME .J FIGURE 85.—General model for estimating resource availability. tion of the data from the search-model portion of the model (fig. 85) gives the resource availability at present price and technology. ' Finally, it is necessary to introduce the econometric model in order to devise a time-dependent production model. This aspect of the model is normally outside the expertise of the US. Geological Survey but must be in- tegrated with the data from the search and occurrence submodels in order to develop a long-term supply model for use by decision makers. In summary, decisions on mineral resource policy should reflect the major factors which affect the availa- bility of resource commodities: (1) geological availabil— ity, (2) technological availability, (3) economic availabil- ity, and (4) alternate sources of supply. Reliable infor- mation on these factors should be developed through research on quantitative resource occurrence models, models of the exploration process, mineral economics models, and gathering of resource data at a disaggre— gated level. Systems design and software development in progress should allow users without computer ex- perience to use the system. Update procedures, along with retrieval and display capabilities should be pro- vided for in the software being developed. References Calkins, J. A., Kays, Olaf, and Keefer, E. K., 1973, CRIB—The min- eral resources data bank of the US. Geological Survey: US. Geol. Survey Circ. 681, 39 p. McKelvey, V. E., 1972, Mineral resource estimates and public policy: Am. Scientist, v. 60, no. 1, p. 32—40. Repr. in Brobst, D. A., and Pratt, W. P., eds, 1973, United States mineral resources: U.S. Geol. Survey Prof. Paper 820, p. 9—19. m l LITHIUM RESOURCES AND REQUIREMENTS BY THE YEAR 2000 161 ‘ ANALYTICAL METHODS AND PROBLEMS OF LITHIUM DETERMINATION IN ROCKS, SEDIMENTS, AND BRINES ‘ By ALLEN L. MEIER, U.S. GEOLOGICAL SURVEY, DENVER, CO ABSTRACT xploration for lithium resources involves reliance on geochemical me hods. In turn, the precision and accuracy of the analytical proce- du es determine the usefulness of the resultant data for exploration. T 0 published analytical methods and an experimental technique weie applied to a variety of rocks, sediments, and brines to determine th ir relative precision and accuracy. The method used for analysis of bri 6 showed good precision and accuracy. Improved precision and accliliracy were obtained on rocks and sediments by an experimental tec nique that uses boric acid when compared to a method that does no[“ use boric acid. INTRODUCTION Potential lithium deposits are not easily recognized by p+ysical evidence alone. Chemical analysis must be used to detect and determine the distribution of the element. The success of lithium exploration depends on the abil— ity of the analytical technique to provide the data neces- sqry to recognize and evaluate a potential resource. The most frequently asked question of an analyst is, “What is the precision and accuracy of the method used?” Precision and accuracy do not mean the same thing but are mistakenly used interchangeably. Precision is a measurement of reproducibility when replicate analysis or measurements are made. Precision c n be determined by repeating an analysis and examin- i‘g the variability. Variability is usually expressed as dilandard deviation, relative standard deviation, confi- dence intervals, or range. l Accuracy is the nearness of a value to the true value. Comparison of a value obtained to an accepted value for a sample determines accuracy. A percentage of the true vialue is an expression of accuracy. A true or accepted alue for the material under study is often unknown. Accuracy can be determined in this case, however, by t e addition of a known amount of the element to the reviously analyzed sample. The percent recovered in the re—analysis is an indication of the accuracy (Fletcher and Collins, 1974). ‘ Precision and accuracy are difficult to determine in analysis of geologic material for the following reasons: 1. Each sample is unique. l Variability within the sample is unknown. ‘ The matrix of each sample is different. ‘ Interaction of the elements present is often un- known. l 5. The element of interest normally occurs in trace l amounts. Wherefore, an analytical technique that works well on l l H‘E-filo one sample type may not work on another. In this work the precision and accuracy of two pub- lished analytical methods and an experimental technique were studied. Several representative samples of different types were analyzed for lithium. They in- cluded clays, tuffs, carbonates, water, and brine. PROCEDURES Atomic absorption was used for the determination of lithium in each of the methods. The water and brine samples were analyzed by a pro- cedure described by Brown, Skougstad, and Fishman (1970). Water samples were analyzed directly without dilution. Brine samples were diluted to concentrations of less than 15,000 mg/l total dissolved solids. (Results are given in table 66.) Rock and sediment samples were analyzed by a mod- ified method reported by Ward, Nakagawa, Harms, and VanSickle (1969). Decomposition was accomplished by a mixture of nitric and hydrofluoric acids. The use of perchloric acid was eliminated from the procedure for exploration and field use. Samples were taken to dry- ness on a hotplate, redissolved in 6-normal hydrochloric acid, and diluted. (Results are given in table 67.) Rock and sediment samples were also analyzed by an experimental technique using boric acid. The method is essentially the same as the one previously described, ex- cept that 5 ml of a saturated boric acid solution was added after decomposition. (Results are given in table 67; see columns that refer to boric acid method.) To study precision, each sample was analyzed five times by each method. The mean and confidence limits at the 95 percent level, and the percent relative standard deviation are reported (tables 66 and 67). Accuracy was determined by adding a known amount of lithium to each of the rock and sediment samples before the decomposition step. Percent recovery was calculated by using the mean of five replicate determina- tions of each sample with the addition and the mean of the sample as previously determined by each of the methods (Fletcher and Collins, 1974) (tables 66 and 67). RESULTS AND DISCUSSION Results obtained on the water and brine samples show both good precision and accuracy by the method studied (table 66). Sample number MAF 810 has poorer preci— 162 LITHIUM RESOURCES AND REQUIREMENTS BY THE YEAR 2000 TABLE 66.—Precision and accuracy of lithium determinations on brine and water samples Ran e Mean and (mg ) confidence Relative S cific limits at standard Accuracy Sample con uctance 95 rcent deviation (percent Sample N°~ type (micromhos) LOW High leve (mg/l) (percent) recovery) MAF 109 __________ Water ________ 6,800 0.99 1.03 1.01:0.02 1.5 97 115 __________ Brine ________ 58,300 6.8 7.2 7 : .2 2.2 99 810 __________ __ do __________ 180,000 10 12 11 :1 7.7 109 797 __________ __ do __________ 160,000 65 68 67 :1 1.7 98 798 __________ __ do __________ 160,000 68 69 68 :1 1.9 100 MAI 095 __________ __ do __________ 190,000 300 310 308 :6 1.5 98 TABLE 67.—Comparison of precision and accuracy determinations on rock and sediment samples using a method without boric acid and a method with boric acid Method without boric acid Method with boric acid Mean and Mean and confidence Relative confidence Relative limits at standard Accuracy limits at standard Accuracy 95 rcent deviation (percent 95 ercent deviation (percent Sample No. Sample type leve (ppm) (percent) recovery) leve (ppm) (percent) recovery) MAI 180 __________ Clay ______________ 3841- 56 11.7 66 815 t25 3.0 100 MAI 177 __________ __ do ____________ 73: 7 7.2 95 74: 6 6.3 106 MAI 445 __________ __ do ____________ 272 i- 47 13.9 80 420 :22 5.0 99 MAF 432 __________ Tuff ____________ 642:182 22.7 71 12:33:54 4.2 95 MAF 593 __________ __ do ____________ 1620: 56 2.8 95 1705:42 2.3 98 MAH 429 ________ __ do ____________ 338:127 30.1 48 700:18 2.4 99 MAC 754 __________ Carbonate ________ 252 t 27 8.6 62 588 :20 2.8 96 MAG 886 __________ __ do ____________ 448: 20 3.7 100 456:24 4.3 97 MAC 714 _________ '_ __ do ____________ 12101-494 21.7 108 1467:54 3.5 100 sion and accuracy than the other samples, because of the high ratio of salinity to lithium concentration in the sample. Dilution was necessary to reduce the dissolved- solids concentration for atomic absorption determina- tion. The lithium concentration was also reduced by this dilution to near the detection limit. The precision and accuracy normally decrease as the detection limit is ap- proached. A comparison of the results obtained by the two methods for rocks and sediments indicates that the method using boric acid has better precision and accu- racy than the method without boric acid (table 67). The mean values obtained by the boric acid technique are higher than the mean values obtained by the method without boric acid. Stable fluoride compounds can be formed when hydrofluoric acid is used for decomposi- tion (Dolezal and others, 1968). Boric acid can'be used to complex residual fluoride (Bernas, 1969). These fac- tors explain the lower recoveries using the technique without boric acid and the higher recoveries with the use of boric acid. CONCLUSIONS The method used in this study for water and brine samples shows adequate precision and accuracy to be useful for exploration for lithium. An improved technique for the determination of low concentrations of lithium in concentrated brines would be useful. The poor accuracy of the method without boric acid indicates low recovery of lithium for some samples. The greater recovery observed for these samples using the method with boric acid suggests that the formation of stable fluorides does occur when hydrofluoric acid is used for dissolution. Boric acid inhibits this formation and improves the recovery. The use of boric acid with hydrofluoric acid is under study‘at the present time. REFERENCES CITED Bernas, Bedrich, 1968, A new method for decomposition and com- prehensive analysis of silicates by atomic absorption spectrometry: Anal. Chemistry, v. 40, no. 11, p. 1682—1686. Brown, Eugene, Skougstad, M. W., and Fishman, M. J., 1970, Methods for collection and analysis of water samples for dissolved minerals and gases: U.S. Geol. Survey Water-Resources Inv. Techniques, Book 5, Chapter A1, 160 p. Dolezal, Jan, Provondra, P., and Sulcek, Z., 1968, Decomposition techniques in inorganic analysis: New York, Elsevier, 224 p. Fletcher, G. E., and Collins, A. G., 1974, Atomic absorption methods of analysis of oilfield brines—Barium, calcium, copper, iron, lead, lithium, magnesium, manganese, potassium, sodium, strontium, and zinc: U.S. Bur. Mines Rept. Inv. 7861, 14 p. Ward, F. N., Nakagawa, H. M., Harms, T. F., and VanSickle, G. H., 1969, Atomic-absorption methods of analysis useful in geochemi- cal exploration: U.S. Geol. Survey Bull. 1289, 45 p. N