TN295 No. 9015 ■ \> * ^ o v a *o . » * A ^°^ ■ ; ^ V % %^K ; <,^^\ °-^W** ^ V ^ %^K*° J'*'\ °W^/ A ^\ '°M^S & « „ *^» ^'T^T'' ^o> J ^>v «* . s • • , *>\ «V . t • » *P e>°^ "WW* A>"^ : .*°° .\/ ? ^ r, \^' V^^'V" \/^^\# V 0^ c'^,"^ *«*-6* °o 0«; *o » * • ■ U # * K €^f- \ #* •:-■ V*' \ % - « • iSfe-Xv* 4 " • ^ %*%*\S m c ^ : ^-/ V^V A/M'/\Wjn M' /\ W s\ '-sKf' S\.'ym .1 ^ < * ^ v* .♦•iC'* <^. ^ yflK-. ^ ^ .: .^ ^ C>°J» »WW' A>^. ^0< *<2* * A ^ ^ '"o V^ r ^c •4 C » aV <^ i. *V «C :» .«. v 4>- Bureau of Mines Information Circular/1985 Mining Deep Ocean Manganese Nodules Description and Economic Analysis of a Potential Venture By C. Thomas Hillman and Burton B. Gosling UNITED STATES DEPARTMENT OF THE INTERIOR T5| A*/NES 75TH A**^ ■Jt*t£ ZtzJCti. ^CxrcatA €>£ Hv.'u^ Information Circular 9015 )l Mining Deep Ocean Manganese Nodules Description and Economic Analysis of a Potential Venture By C. Thomas Hillman and Burton B. Gosling UNITED STATES DEPARTMENT OF THE INTERIOR Donald Paul Hodel, Secretary BUREAU OF MINES Robert C. Horton, Director Library of Congress Cataloging in Publication Data: THaisT -TWfyfc Hillman, C 4 Thomas Mining deep ocean manganese nodules: description and economic analysis of a potential venture. (Bureau of Mines information circular ; 9015) Bibliography: p. 15-16. Supt. of Docs, no.: I 28.27:9015. 1. Manganese nodules. 2. Manganese mines and mining, Submarine. I. Gosling, Burton B. II. Title. III. Series: Information circular (United States. Bureau of Mines) ; 9015. [TN490.M3] 622s 338.2'74 84-600337 CONTENTS - Page Abstract 1 Introduction 2 Resources 3 Location and geography 3 Geologic setting 5 Deposit description 5 Tonnage and grade estimates 6 Integrated recovery system and costs 7 System description 7 System costs 8 Economic analysis 9 Discussion 9 Base case description 10 Methodology and results 11 Summary and conclusions 14 References 15 Appendix. — Capital and operating cost detail for study area CI 17 ILLUSTRATIONS 1. Location of the northeast Pacific high-grade zone and DOMES Sites A, B, and C 3 2. Location of study area CI and sediment distribution in the northeast Pacific high-grade zone 4 3. Base case project development schedule 10 TABLES 1. Mean metal content, study area CI 6 2. Resource estimates and supporting data, study area CI 7 3. Transportation data summary, study area CI 8 4. Capital cost summary 9 5. Operating cost summary 9 6. Commodity data summary 11 7. Metal recovery options and resultant rates of return 12 8. Effects of grade variations on rates of return 12 9. Price options and resultant rates of return 13 10. Effects of capital and operating cost variations on rates of return 13 11. Effects of shortened development schedule on rates of return 13 12. Rates of return effected by favorable options 14 A-l. Mine costs 17 A-2. Transportation costs 18 A-3. Process costs 19 ^ UNIT OF MEASURE ABBREVIATIONS USED IN THIS REPORT cm centimeter m meter cm/s centimeter per second m 2 square meter °C degree Celsius mm millimeter dwt deadweight long ton m/s meter per second d/yr day per year nmi nautical mile g/cm 3 gram per cubic centimeter pet percent ha hectare t metric ton h/d hour per day t/km 2 metric ton per square kilometer kg/m 2 kilogram per square meter t/yr metric ton per year km kilometer wt pet weight percent km 2 square kilometer yr year Lt long ton MINING DEEP OCEAN MANGANESE NODULES Description and Economic Analysis of a Potential Venture By C. Thomas Hillman and Burton B. Gosling ABSTRACT This Bureau of Mines report describes an investigation of factors in- fluencing economic viability of a proposed system to mine and process manganese nodules. The system consists of hydraulic dredging and ammo- nia leach processing designed to recover three metals: nickel, copper, and cobalt. A ferromanganese recovery plant is a considered option. Annual capacity is 3 million dry metric tons (t) nodules. Aftertax rates of return (ROR's) of only 7.38 and 6.63 pet are pre- dicted for three- and four-metal base scenarios, respectively. Sensi- tivity analyses indicate that variations in commodity prices, metal recoveries, and deposit grades produce similar incremental changes in ROR. Estimated cumulative effects of price variations are greatest. While three-metal operations generally yield higher predicted returns, potential increases in manganese price and process recovery could make four-metal operations most profitable. Maximum variations in both capital costs and three-metal operating costs result in ROR changes of about 25 pet. In contrast, four-metal ROR's exhibit variations to 70 pet. Even so only a best case analysis, utilizing favorable variations in all categories tested, generated ROR's approaching a level thought necessary to trigger company interest. Con- sequently, it is concluded that nodule mining will not take place in the foreseeable future without significant financial incentives. Supervisory physical scientist, Western Field Operations Center, Bureau of Mines, Spokane, WA. INTRODUCTION Deep ocean manganese nodules represent an exceptionally large and potentially important mineral resource. Deposits of the highest grade and abundance occur in the northeast equatorial Pacific Ocean in an area of about 10 million km 2 . This area is of primary commercial interest and is known as the high-grade zone. Manganese nodules are also called poly- metallic nodules because they consist primarily of manganese and iron oxides that contain elevated concentrations of nickel, copper, and cobalt. Signifi- cantly, the United States is heavily de- pendent on foreign sources for all nickel, cobalt, and manganese. There- fore, it is in the Nation's best interest that these ocean resources be evaluated and integrated systems to mine and pro- cess them be proposed. In response to this need, the Bureau is involved in a continuing Minerals Avail- ability Program (MAP) effort to collect and analyze information on mineral re- sources and mining and processing sys- tems. Data have been gathered during the past 7 yr through grants to Scripps In- stitution of Oceanography and Washington State University and by contacts with the U.S. Geological Survey, the National Oceanic and Atmospheric Administration (NOAA), and private consultants. Much raw data came from project DOMES (Deep Ocean Mining Environmental Study) spon- sored by NOAA. DOMES involved a detailed investigation of the marine environment at three potential minesites (designated Sites A, B, and C) in the high-grade zone and a determination of possible environ- mental effects of manganese nodule min- ing. Figure 1 shows the high-grade zone and the three DOMES sites. nodule resources in the vicinity of the three DOMES areas and describes a pro- posed integrated system to mine, trans- port, and process nodules from one poten- tial minesite in each of the three areas. Cost estimates were made and discounted- cash-flow analyses were performed to evaluate potential profitability. Re- sults indicated an ROR of 6 pet or less after taxes, a figure that is only a small fraction of the estimated 25 to 30 pet ROR needed to attract risk capital. Because of these low predicted returns, the present study attempts to identify those factors having the most significant effects on the economics of proposed re- covery systems discussed in reference 1. The goal is to provide information to both government and industry with inform- ation that will be used to more accurate- ly plan for the eventual mining and util- ization of manganese nodules. This report first describes resources of a potential minesite (Subarea CI in reference 1), one of several previously evaluated. The area encompasses DOMES Site C and for this report is designated study area CI. All costing and subse- quent analyses are based on resource es- timates for this area. Because of high grade and abundance, nodule resources there are believed representative of de- posits that will be mined and processed before any others; the term frequently used for such a deposit is "first- generation minesite." Following resource definition, an integrated recovery system is described, and costs for the system are estimated and summarized. A cost sensitivity analysis using multiple runs of the Bureau's MINISIM08 computer pro- gram (2) is subsequently described along with results and conclusions. Several Bureau reports have resulted from the MAP effort. The latest report ( 1) , 2 contains estimates of manganese ^Underlined numbers in parentheses re- fer to items in the list of references preceding the appendix. In context of publicly available infor- mation, it is believed that this analy- sis, based on resource estimates for ac- tual deposits, should improve upon other financial analyses based on hypothetical resources. It is also reasonable to ex- pect mining consortia to delineate and • ASIA \ SOUTH AMERICA FIGURE 1. - Location of the northeast Pacific high-grade zone and DOMES Sites A, B, and C. mine those areas within potential mine- thus potentially exeeding economic pre- sites of highest grade and abundance, dictions in this study. RESOURCES LOCATION AND GEOGRAPHY Study area CI is in the north-central portion of the Pacific high-grade zone, approximately 2,200 km southest of Los Angeles, CA (fig. 2). The area is ap- proximately 57,600 km 2 and lies between latitudes 14.0° to 16.2° N and longitudes 124.8° to 127.0° W, at an average water depth of approximately 4,600 m. Climate is moderate throughout the year; rainfall is slight and air temper- atures average about 25° C for the entire year. Tropical storms and typhoons occur primarily during the summer months and usually last a day or two. An average of 15 storms per year occurred during the 1966-75 period, and an average of 11 of those occurred in the through September. months of July Measurements made in 1975-76 by NOAA indicate surface currents are moderate, averaging about 17 cm/s O ) . These cur- rents flow in a west-southwest direction, and are controlled by moderate trade winds that blow steadily throughout the year. Surface currents gradually de-- crease to zero at a depth of 160 m; at that approximate depth currents reverse direction and flow eastward at low velo- city. Although bottom current data for study area CI are unavailable, theoret- ical studies and water mass characteris- tics indicate a net eastward flow at low velocities (3). Q> C o N LU cc ID o Based on known data, it is assumed that neither surface nor bottom currents would significantly hinder mining in study area CI. Tropical storms can be expected to preclude mining for 30 to 40 days annual- ly, mostly during the summer months. GEOLOGIC SETTING Essentially all of the high-grade zone, including study area CI, occurs within the Eastern Pacific Sedimentary Basin, which extends from about 5° to 20° N lat- itude and eastward from 170° W longitude and is bounded both east and west by sea- mount provinces. Sediment-covered hills dominate the ocean floor topography and are characteristically elongated and par- allel (4_) , they trend north-south or northeast-southwest, and have slopes typ- ically between 2° and 3° ( 5_) . Local re- lief is generally low, except in areas of current scour or fault scarps. Parallel scarps impart a stairstep effect on hill- sides, and where accompanied by sediment slumping, underlying basalt bedrock is exposed. Water depths decrease toward the East Pacific Rise (fig. 2) , where oceanic crust is being formed within a narrow zone called a rift or spreading center. Progressive outward movement of newly formed crust results in successively old- er rocks in a westerly direction ( 6_) . Bedrock age in the area is mostly Oligo- cene, while further west crust is as old as Late Cretaceous. Ocean floor in the area is covered by a reddish-brown to chocolate-colored pelag- ic clay unit. This unit contains as much as 65 pet illite, kaolinite, and smec- tite; the remainder consists mostly of siliceous organic material, primarily ra- diolaria, diatoms, silicof lagellate skel- etal remains, and sponge spicules. Lo- cally, calcareous material comprises as much as 5 pet of the sediment, but for the most part is dissolved before it reaches the seafloor. Elsewhere in the high-grade zone, pelagic clay grades into siliceous ooze and clay that contains variable yet high percentages of si- liceous organic material. Calcareous sediments occur in the southeast and southwest parts of the zone (fig. 2). In study area CI, average bulk density of sediments is estimated to be about 1.3 g/cm 3 (_3 ) • Estimated sediment thickness is 100 m, and present day sedimentation rates probably range from 1 to 3 mm/1, 000 yr ( 6_) . Rates may have been greater in the past to account for the 100-m accumu- lation during the past 30 million yr (middle Oligocene). DEPOSIT DESCRIPTION Manganese nodule deposits occur as ir- regular, single-layer fields at the sediment-water interface. Typically, few nodules occur below a sediment depth of 1 m, and those within a meter of the ocean floor comprise an amount equal to about 25 pet of those at the sediment surface ( 6_) . Populations, defined as percentage of seafloor covered by nodules, range from zero to nearly 100 pet. Individual nodules have a dull luster and are earthy brown to bluish black col- or. Shapes of smaller nodules are mostly spheroidal, with progressively larger ones becoming ellipsoidal and then dis- coidal. This phenomenon is attributed to unequal growth rates; bottom portions, nested in sediment, are thought to ac- crete more rapidly than exposed tops (7^). Irregular shapes are common and result from either natural agglomeration of small nodules or the tendency of nodules to reflect morphology of irregularly shaped nuclei. Surface textures range from smooth to very rough and may be attributed to dif- ferential growth patterns of constituent oxide phases. Porosity and internal sur- face area are both high, typically 50 pet and 200 to 300 m 2 , respectively ( 8) . Consequently, individual nodules contain about 30 wt pet moisture. Wet specific gravity is generally between 2.0 and 2.5. Internally, nodules are composed of one or more nuclei surrounded by discontin- uous, concentric layers of manganese and iron oxides. Nuclei may be shark teeth, whale ear bones, small pieces of other nodules, or small rock fragments. Clay layers, generally present at Irregular intervals between oxide layers, are thought to indicate long periods of non- growth. Concentric and radial fractures are nearly universal in larger nodules. Nickel and copper probably occur within the manganese oxide minerals todorokite and birnessite by means of adsorption, lattice substitution, or ion exhange (8^). The occurrence of cobalt is less well known, but relatively recent work (9) in- dicates that cobalt in low-grade nodules is contained principally in iron oxide phases. In high-grade nodules, manganese phases are preferentially enriched in cobalt. TONNAGE AND GRADE ESTIMATES Quantity and grade estimates are based on information in reference 1. Evalu- ation of chemical analyses of samples from study area CI (10) show that the arithmetic mean of metal concentrations can be predicted within ±10 pet at a con- fidence level of 90 pet. Work by other investigators (11) using many hundreds of samples likewise indicates that grades within large areas of the high-grade zone can be predicted with similar accuracy. Nodule abundances (weight per unit area of seafloor) and nodule tonnages calcula- ted from those abundances are more diffi- cult to determine. Even though deposits cover very large areas, local nodule pop- ulations are extremely variable. Over distances of a few meters, abundances typically range from near zero to 10 to 15 kg/m 2 . Ideally, sampling should be conducted on a grid with coverage by tel- evision and seafloor photography between points. Because the purpose of most in- vestigations to date has been research rather than resource evaluation, sampling locations are generally random and photo- graphic coverage is along linear tracks. However, tonnage estimates are believed to be conservative, because photographic correction factors (12) that generally raise abundance-tonnage estimates could not be applied. These factors, used by mining consortia, require not publicly available. detailed data Boundaries of the area are primarily based on locations of available sample data. A rectangular shape is used for convenience and does not necessarily de- limit or enclose any single deposit. In reality, study area CI may include all or parts of several deposits separated by barren seafloor. Table 1 contains mean metal concentra- tions (grade) calculated from assays of available samples taken at 64 sample sites in the study area. Combined con- centrations of nickel plus copper (2.37 slightly greater than the regarded by some (13) to be requirement for a viable wt pet) is 2.3 wt pet the mininum minesite. TABLE 1. - Mean metal content, study area CI wt pet Cobalt 0.26 Copper 1.04 Iron 6.90 Manganese 26.80 Molybdenum .07 Nickel 1.33 Average abundance was calculated from estimates for individual ship stations. Estimates determined by sampling are not differentiated from those determined from photographs, nor is any greater signifi- cance attached to them. Assuming 30 pet water content, the calculated average of 11.7 wet kg/m 2 converts to 8.2 dry kg/m 2 or 8,200 dry t/km 2 . Gross dry tonnage can be easily figured by multiplying area size in square kilom- eters by the average dry abundance; in the case of study area CI, 57,600 km 2 times 8,200 dry t/km 2 . However, the re- sult can be misleading, because several practical considerations significantly lessen the amount of resource that can actually be recovered. Localized topo- graphic features such as fault scarps and steep slopes reduce the minable area to an estimated 75 pet of the minesite (14). Possibly one-third of the remain- ing area contains deposits with combina- tions of grade and abundance insufficient to warrant mining. As a result, it is reasonable to assume about half of the original site can be mined. Addition- ally, retrieval efficiencies of presently envisioned mine systems are expected to be about 90 pet. Also, ship maneuvering limitations indicate that only about 70 pet of the minable area will be travers- ed, resulting in a net mining efficiency of slightly more than 60 pet. Therefore, recoverable resources from first- generation mining may be only about 30 pet of the in situ tonnage. Minable and recoverable resources are listed in table 2, as well as average nodule abundance, mine size, and other pertinent data. Proposed mining- transportation-benef iciation systems , costs, and economic analysis in succeed- ing sections are based on this informa- tion, and on the mean metal content in table 1. TABLE 2. - Resource estimates and supporting data, study area CI Total area km 2 .. 57,600 Minable area 1 km 2 .. 28,800 Average abundance dry t/km 2 . . 8,200 Minable resource 10 6 dry t.. 236.2 Minable nodules traversed. .. .pet. . 70 Pickup efficiency pet.. 90 Mining efficiency 2 pet.. 63 Recoverable resource 3 .. 10 6 dry t.. 148.8 1 50 pet of total area. 2 Percent of nodules traversed times pickup efficiency. 3 Minable resource times mining efficiency. INTEGRATED RECOVERY SYSTEM AND COSTS SYSTEM DESCRIPTION The system proposed to recover nodules from study area CI is hypothetical be- cause there is presently little commer- cial experience from which to draw. How- ever, the system and costs are based on information from many knowledgeable sources, published and unpublished. Hy- draulic mining and ammonia leach (Cu- prion) processing are proposed because hydraulic mining systems have been tested somewhat successfully by both Interna- tional Nickel Co. (INCO) and Deepsea Ven- tures, and Kennecott has apparently dem- onstrated in a pilot plant the feasibil- ity of the ammonia leach process. Slurry transfer is the most probable method of transportation of nodule ore because the material is amenable, and considerable slurry handling experience exists in the minerals industry. For detail beyond what is given in the following discus- sion, the reader is referred to reference 1. Prior to mining, the explored minesite would be characterized in detail. Large-scale maps would be drawn showing locations of all bottom obstructions and specific mining blocks with detailed grade and abundance information. Mining plans would be drawn up at least a year in advance and would consider not only economics, but also licensing require- ments and other regulatory and environ- mental factors. Mining is scheduled 20 h/d, 300 d/yr, with a projected annual production of 3.0 million dry t. Equipment modification and repair, drydocking, and other nonmin- ing activity would take place during Au- gust and parts of July and September, when most major storms occur. Two ships, each towing hydraulic collectors at a velocity of 1.0 m/s , would conduct opera- tions independently of one another. Nod- ules would be dislodged, screened, and channeled to a large-diameter pipe con- nected to the ship. Submerged hydraulic pumps would maintain upward flow of a slurry composed of water, nodules, and nodule fragments. Aboard the ship, nodules would be screened, conveyed to storage, and dewatered by decantation. To prevent formation of a surface plume, decanted sediment and biogenic debris would be discharged through a pipe extending to a depth of about 200 m (15). No attempt to upgrade the ore is presently envisioned because it is not amenable to conven- tional flotation or other cost-effective means; the metals of interest are inti- mately associated with the oxide matrix. Every few days nodules would be reslur- ried and pumped through a flexible pipe to 70,000-dwt-capacity nodule transports where they would be dewatered and trans- ported to a terminal on the west coast of the United States. At the terminal, portable units would reslurry and pump the ore to holding ponds on shore. From there the nodule slurry would be pumped an assumed distance of 40 km inland to the processing plant. Table 3 is a sum- mary of pertinent transportation data. Three transport vessels, each making 23 trips annually, would be required to sup- port a 3.0-million-t-capacity plant. TABLE 3. - Transportation data summary, study area CI Transport capacity, t: Wet ore 1 64,000 Dry ore 44 , 800 Distance to port nmi.. 1,840 Cycle time days.. 13 Annual trips per vessel 23 Number of vessels require d 3 "90 pet of the nominal capacity of 70,000 Lt. Operations at the Cuprion plant would be conducted 24 h/d, 330 d/yr, at full capacity. The wet ore would be reclaimed from storage, and ground in a mixture of seawater and recycled process liquor con- taining dissolved copper and ammonium carbonate. Cuprous ions, generated by introduction of carbon monoxide in a series of reaction vessels, catalyze the reduction of manganese from the tetrava- lent to the divalent state, thereby re- leasing metals bound in the oxide matrix. The metals are separated from the solid residue by countercurrent washing, and the residue is processed with steam to recover ammonia and carbon dioxide. Nickel and copper would be extracted from solution by liquid ion exchange followed by electrowinning on high-purity cath- odes. Cobalt would be chemically precip- itated, purified, and recovered as metal- lic or oxide powder. A facility to recover manganese in the form of ferromanganese is a considered option. Carbonate residue from the ammo- nia recovery section of the Cuprion plant would be washed and centrifuged (16). Prior to flotation with saponified fatty acids, thickened residue would be mixed with fresh water, soda ash, caustic soda, and sodium silicate. Manganese carbonate would be recovered in the froth as a con- centrate assaying up to 40 pet manganese. The concentrate would be thickened to 50 pet solids , and then dried and calcined in a rotary kiln to make synthetic manga- nese oxide. This material would be stored or conveyed directly to submerged resistance furnaces charged with lime- stone, silica flux, coke, and iron ore. Slag would be skimmed off for disposal, and molten ferromanganese (78 pet Mn) poured into molds for cooling and ship- ment. The ferromanganese plant would operate 330 d/yr, and process an esti- mated 3 million t of Cuprion residue annually. Tailings from both Cuprion and ferro- manganese plants would be combined with other plant wastes and pumped as far as 100 km to a disposal site. At the dis- posal site, the tailings would be pumped into conventional waste impoundments. Granulated slag hauled from the ferroman- ganese plant would be a secondary source of bank material, if needed. Successive ponds would be stabilized in various ways in accordance with appropriate environ- mental regulations. SYSTEM COSTS All cost estimates in this study are in January 1983 dollars. Most are updated from January 1981 costs presented in re- ference 1. The time and tonnage basis for working capital calculations is changed, resulting in slight increases in mine and transportation working capital and a small decrease in processing capi- tal. Also, plant capital has been reduc- ed to reflect recent estimates (17) . In- dexing data used are from a Bureau cost index computer program and price indices published by the Bureau of Labor Statistics. Tables 4 and 5 are summaries of capital and operating costs for the proposed in- tegrated recovery system. Tables A-l through A-3 of the appendix contain addi- tional cost detail and descriptions of the various cost factors. TABLE 4. - Capital cost summary, million January 1983 dollars Investment Mining Transportation and transfer. . Cuprion plant and facilities. Total (3-metal) Ferromanganese plant Total (4-metal) $590, .6 310. .6 726, .9 1,628.1 215.3 1,843.4 TABLE 5. - Operating cost summary, January 1983 dollars Annual 10 6 Per dry t ore $76.5 36.7 110.9 $25.50 12.23 Transporation and Cuprion plant and 36.97 tl) 1) Total (3-meta Ferromanganese pla 224.1 216.6 74.70 72.20 Total (4-meta 440.7 146.90 Mine capital includes money for 6 yr of exploration and detailed site character- ization; once mining begins this expense is treated as an operating cost. The re- search and development budget of nearly $158 million is divided about evenly be- tween mining and processing. Working capital is based on full production for 15, 12, and 6 months, respectively, for mining, transportation, and processing. Additional capital of $23.2 million is allowed for exceptional expenses associ- ated with at-sea testing and modifying of the collector. Mine ships, on-board equipment, pipelines, and collectors are assumed to be new and constructed in the United States. Transportation investments include pur- chase of three new European-built ships for slightly more than $200 million. Al- so included are costs for construction of an offloading facility (slurry terminal) on a 10-ha site, 40 km of slurry pipeline to the plant, and purchase of a high- speed supply boat. Processing capital consists of all land purchases, equipment, buildings, a 100-km slurry line to tailing disposal, an 8-km railroad spur line, an 8-km access road, and installation of turbines for gener- ation of power for Cuprion plant opera- tion. Power for a ferromanganese plant would be purchased. Not included is capital for infrastructure such as townsite. Operating costs include allowances for wages and benefits, material and sup- plies, maintenance and repair, fuels and utilities, and insurance. Subsistence for ship's crew is included in mining and transportation, while costs associ- ated with operation of a small supply boat, unloading facility, and pipeline (40 km) to plant are charged to transpor- tation. Extraordinary expenses associ- ated with processing include maintenance of the railroad spur and operation of the 100-km pipeline to waste disposal. Oper- ation of a ferromanganese plant at full capacity incurs the largest single opera- ting expense, nearly $120 million for purchased power. ECONOMIC ANALYSIS DISCUSSION Financial analyses carried out during the initial Bureau study of three poten- tial minesites (1) resulted in low predicted ROR's. Two analyses for each site were reported; one for the Cuprion process, which recovers nickel, copper, and cobalt (three-metal), and one for Cuprion plus ferromanganese (four-metal) 10 recovered from processing about half of the carbonate residue available from Cuprion processing. Three-metal ROR's ranged from 4.1 to 6.0 pet, while four- metal projections ranged from 3.5 to 5.2 pet. Considering political and technical risks, these predictions are far below the 25 to 30 pet ROR that might attract venture capital. This study attempts to identify through sensitivity analysis, those factors that most affect economic viability of the proposed nodule mining venture; factors that should be subjected to close scruti- ny in future planning and evaluation. The sensitivity analysis addresses mining and processing of nodules from study area CI, an area representative of first- generation minesites, those having depos- its of relatively high grade and abun- dance. The proposed integrated recovery system and associated costs serve as the basis for two scenarios or base cases: a three-metal and a four-metal operation. Several cost factors associated with these base cases are varied through a series of financial analyses using the Bureau of Mines MINSIM08 computer program (2). The factors are varied independent- ly to determine effects on predicted ROR. BASE CASE DESCRIPTION Both the three- and four-metal oper- ations utilize descriptions and costs presented in preceding sections. The three-metal case is essentially the same as described in reference 1; costs are updated. However, the four-metal case involving the add-on f erromanganese plant is different. Instead of recovering fer- romanganese from only half the Cuprion process residue, the entire amount is treated to produce substantially more metal. Subsequent financial analyses in- dicate that the larger production scheme would be slightly more profitable. Figure 3 depicts the relatively aggres- sive project schedule assigned to both base cases. Project timing, tax treat- ment, and other assumptions are identical to one another. Research and development would begin as soon as possible in the first year and would continue through the seventh year. Exploration would also be- gin in the first year and continue through year 6, then resume in year 10 and be carried on through year 30. The exploration ship would assist in mining tests during years 7 through 9. Plant construction would last 4 yr (years 6 through 9) and ship construction about 5 yr. Keel for the first ship would be laid in year 6, with completion of con- struction in year 8. At-sea tests would be conducted late in year 8 and early year 9. A second ship would be con- structed in years 8 through 10, and im- provements made during testing of the first ship would be incorporated. Pro- duction would begin in the 10th year; 1 1 1 1 Research and | i i i development i | i l 1 1 | 1 1 1 ! | I 1 1 i | i I I l | Plant and ship construction Startup Full production Exploration Detailed exploration i i i i 1 . , , , 1 , , i i 1 i i i i 1 i i i i 1 i i i i 1 10 20 15 YEARS FIGURE 3. - Base case project development schedule. 25 30 11 full capacity of 3 million dry t annually is scheduled to begin in year 12 and con- tinue for 20 yr. Assumptions that materially affect results of the analysis include the following: 1. A go-ahead decision is made early enough to allow proper planning for construction. 2. Costs and commodity prices escalate at the same rate. 3. Equity capital is used, thus no fi- nance charges are incurred. 4. A 9-pct State income tax and 4 pet property taxes are included as well as Federal income tax. 5. A payment of 0.75 pet excise tax on gross value is assumed, which is mandated by the Deep Seabed Hard Mineral Resource Act. 6. Depletion is used; 15 pet for cop- per and 22 pet for nickel, cobalt, and manganese. Metal contents (grade), estimated re- coveries, and commodity prices for the two base cases are listed in table 6. Commodity selling prices for nickel, cop- per, and f erromanganese represent a 10-yr mean (1973-82) calculated in constant 1983 dollars. The average cobalt price was adjusted downward, because mean year- ly values for the 1978-81 period are con- sidered artificially high. This was done by assigning the mean of the two enclos- ing years, that is 1977 and 1982, to the four high years in the computation. METHODOLOGY AND RESULTS The primary tool used in the sensitiv- ity analysis is the Bureau's mine sim- ulation computer program (MINSIM08) de- vised by personnel of the Minerals Avail- ability Field Office, Denver, CO (2). This program, among other things, com- putes the ROR for potential investments when required operational parameters are supplied. This is the profitability yardstick used for comparison of the var- ious factors tested. Necessary data re- quired by the program include capital and operating costs, investment scheduling, metal processing recoveries, ore grade, and various product characteristics. TABLE 6. - Commodity data summary, base case conditions Cobalt. ... Copper. . . . Manganese 1 Nickel. . . . Metal Re covery , Price content, pet per wt pet lb 0.26 65 $8.53 1.04 92 1.17 26.80 44 .25 1.33 92 3.62 Recovered as f erromanganese , contain- ing 78 pet manganese; listed price is per pound f erromanganese. The first step toward sensitivity anal- ysis requires identification of param- eters to be tested. Those chosen include capital costs, operating costs, ore grade, process recoveries, metal prices, and the length of the preproduction peri- od. Additionally, best case runs were made to determine results of a combina- tion of favorable factors on the two base cases. Individual parameter ranges were estab- lished according to the uncertainty of the original estimate. For example, cap- ital and operating costs were varied up to 25 pet, while ore grade variance was limited to ±10 pet. Once a range was as- signed, analyses were completed using the extreme values of the range to demon- strate maximum effects of the parameter on ROR. Additional, intermediate runs were made for capital and operating cost parameters, because of the relatively large degree of uncertainty attached to costs. Base case analyses were run initially, and results served as standards for all subsequent analyses. Resulting ROR's are, in percent, • Three-metal (Cuprion)... 7.38 • Four-metal 6.64 12 These values are slightly higher than those in reference 1 because of minor differences in methodology and the change to full production of ferromanganese. The four-metal base case is slightly less profitable despite significant additional revenues, because of a near doubling of operating costs. Table 7 contains estimated ROR's from a series of runs testing variations in met- al recoveries. Logically, the greatest potential effect on return is exhibited by manganese, because there is a much greater potential for variance. Cobalt has the second greatest potential af- fect, resulting from a combination of high product value and relatively high variances. Incrementally, increased nickel recovery raises projected return the most; about 1.3 pet for every percent increase. Manganese raises the ROR about 1.0 pet for each percent rise in re- covery. For cobalt and copper the incre- mental rise in ROR is 0.44 and 0.31 pet, respectively. TABLE 7. - Metal recovery options and resultant rates of return, percent Recovery ROR Low Base High Low Base High 60 65 70 7.13 7.38 7.63 90 92 94 7.33 7.38 7.43 Manganese 1 . 35 44 60 4.93 6.64 9.08 90 92 94 7.16 7.38 7.59 ROR Rate of return. 'Recovery from Cuprion manganese car- bonate tailing. Processing feed grade ranges were set at only ±10 pet, because considerable confidence can be placed in grade esti- mates. However, as the nickel, copper, and cobalt grades are assumed directly proportional to one another, it is also appropriate to demonstrate the effect of varying all three recovered metals simul- taneously. Results of grade variation runs are listed in table 8. As might be expected, the ROR variation is a direct reflection of the relative level of con- tribution each metal makes in proportion to total revenues for the operation. Therefore, nickel has the greatest effect followed by cobalt and then copper. Although a much larger effect occurs when all three metals are varied simultaneous- ly, the change in ROR is still not great enough to suggest that slight grade changes will significantly affect propos- ed operations. TABLE 8. - Effects of grade variations on rates of return, percent Cobalt , Copper , Nickel , 3-metal variation. Low, High, -10 Base +10 pet pet 7.00 7.38 7.75 7.14 7.38 7.61 6.38 7.38 8.31 5.67 7.38 8.87 Similar analyses were not conducted for the manganese producing options, because expected grade variations are insignifi- cant when compared to wide limits placed on manganese recovery. The third parameter investigated was commodity price. Range end members were taken as the extreme upper and lower values for 1973-82 annual prices. In order to make direct comparisons, these values were escalated to constant 1983 dollars using price indexes published by the Bureau of Labor Statistics. Cobalt is the one exception; range was determin- ed from the highest and lowest annual average excluding years 1978-81, because of the previously mentioned artifically high selling price during those years. Table 9 summarizes the various price op- tions and ROR's resulting from MINSIM08 analysis. From this analysis it is seen that nickel price variations have the largest incremental effect on ROR. For instance a 1-pct rise in nickel price translates to about a 1.25-pct change in ROR. Again, this fact is a direct reflection of the relative importance of nickel rev- enue. It is also estimated that manga- nese, cobalt, and copper would add 1.1, 0.41, and 0.33 pet, respectively, per 1 pet rise in commodity price. 13 TABLE 9. - Price options and resultant rates of return Price per pound: 1 Cobalt Copper Manganese 2 Nickel ROR, pet: Cobalt Copper Manganese Nickel Low $6.75 $0.77 $0,186 $3.18 6.69 6.50 4.12 6.11 Base $8.53 $1.17 $0,250 $3.62 7.38 7.38 6.64 7.38 High 13.00 $1.57 0.358 $4.13 8.98 8.20 9.80 8.71 ROR Rate of return. "■Escalated to 1983 dollars. 2 Price per pound f erromanganese. Capital and operating costs were analy- zed next. Because of a relatively high uncertainty attached to estimates, end points were established at ±25 pet. Table 10 contains results of test runs includ- ing intermediate runs using ±10 pet. Al- though incremental effects were only slightly higher than factors previously discussed, the wide test ranges resulted in the greatest effects on ROR's of any of the four parameters tested. Signifi- cantly, variations in operating costs ex- erted much more influence on ROR than all other factors including capital costs. This is especially true of four-metal op- erations which included the only test run to exceed 10 pet return. However, the resultant ROR of about 11 pet is still far below reasonable expectations. TABLE 10. - Effects of capital and oper- ating cost variations on rates of return, percent Variation, Capital: Cuprion. 4-metal , Operating: Cuprion, 4-metal, +25 + 10 1 Base -10 -25 pet pet pet pet 6.08 6.70 7.38 8.14 9.21 5.31 5.99 6.64 7.36 8.28 4.96 6.15 7.38 8.21 9.39 2.18 5.49 6.64 9.02 11.17 It was considered possible that lengthy lead time results in such significant future cash flow discounting that varia- tions of previous parameters achieves at best only modest increases in the ROR. To examine this possibility the prepro- duction period was shortened from 9 to 5 yr, maintaining investment category to- tals constant, but accelerating expendi- ture rate. In table 11, returns of the resultant compressed schedule are compar- ed with base returns. Considering the original premise and the degree that the preproduction period is shortened, it is surprising that so little effect is noted. TABLE 11. - Effects of shortened develop- ment schedule on rates of return, percent Cuprion. 4-metal. Base 7.38 6.64 Shortened 8.10 7.25 Two additional best case runs were made in an attempt to evaluate the cumulative effect of varying all test parameters simultaneously. Worst case scenarios were not run because they would produce extremely low or negative returns. In addition to using high values for re- coveries, prices, and grades, the best case included 25 pet decreases in capital and operating costs, and a shortened de- velopment schedule discussed in the pre- ceding paragraph. As shown in the fol- lowing tabulation, there is a marked im- provement in ROR's. • Three-metal (Cuprion) 19.3 pet ROR • Four-metal 23.7 pet ROR Particularly interesting is the potential of four-metal operations to be slightly more profitable than three-metal. How- ever, these predicted returns would still be of marginal interest to potential ocean miners. Table 12 summarizes data pertaining to increased ROR's resulting from the most favorable options. Actual predicted re- turns are on the summary side of the ta- ble while figures on the right side (change) are percent increase, calculated by dividing the actual amount of increase by the base case ROR (either 7.38 or 6.64) and multiplying by 100. Unfavor- able options are not summarized, because of extremely low returns which preclude 14 TABLE 12. - Rates of return effected by favorable options, percent Summary ' Change (increase) 2 Metal recovery Deposit grade Commodity price Metal recovery Deposit grade Commodity price Cobalt 7.63 7.43 7.59 9.08 7.75 7.61 8.31 NAp 8.98 8.20 8.71 9.80 3.4 0.6 2.8 36.7 5.0 3.1 12.6 NAp 21.7 11.1 Nickel 18.0 Ferromanganese. . . 47.6 Capital costs Operating costs Shortened development Capital costs Operating costs Shortened development 9.21 8.28 9.39 11.17 8.10 7.25 24.8 24.7 27.2 68.2 9.8 9.2 NAp Not applicable. 1 Compare with base case — 3-metal, 7.38 pet; 4-metal, 6.64 pet. 2 Relative change, calculated by dividing the amount of increase by the base rate of return (ROR) and multiplying by 100. any thought of mining. Differences in ROR's resulting from independent changes in metal recovery, grade, and price vari- ations are a function of parameter range and revenue generated by the affected commodity. As an example, nickel con- tributes much more revenue to proposed operations, yet greater variance both in recovery and commodity price of cobalt accounts for larger potential increases in ROR. On the other hand, estimated grade variability for the two commodities is the same (10 pet), and the predicted ROR increase for nickel is two and one- half times that of cobalt. Rates of return associated with rela- tively large variations of capital and especially operating costs are also most significant. Predicted ROR increases range from about 25 pet, for 25 pet less capital investment, to nearly 70 pet for a 25-pct reduction in four-metal operat- ing expense. Shortened development peri- od (9 to 5 yr) has little effect on ROR; but in the analysis costs were not reduc- ed, only the rate spending was increased. A larger and possibly significant in- crease might result if the shortened de- velopment period was accompanied by less capital spending. Potential impacts related to copper are comparatively low, because of relatively low commodity value and low degree of variance. Conversely, changes in ferro- manganese parameters substantially influ- ence ROR's, both by contributing a large percentage of revenue and by having fair- ly high degrees of uncertainty associated with metal recovery and selling price. The best case scenarios yielded very large percentage increases in ROR's; ap- proximately 160 and 260 pet for three- and four-metal operations, respectively. While predicted returns are respectable, they are dependent on a series of fortu- nate circumstances, not likely to occur simultaneously. SUMMARY AND CONCLUSIONS The following statements summarize re- sults of the sensitivity analyses de- scribed in preceding sections: 1. Incremental changes in metal recov- eries, deposit grade, and commodity price affects ROR's, similarly. Accordingly, a 1-pct variance in grade has nearly the same effect on return as would a 1-pct change in price or recovery 2. Among deposit grade, metal recov- ery, and commodity price, the potential for changing ROR in relation to the base case is greatest for commodity price. This is because price estimates are less 15 reliable and consequently, were greater. test ranges 3. Comparisons of three- and four- metal operations indicate that under cer- tain circumstances (i.e., high manganese recovery and commodity price) four-metal operations could be slightly more profit- able. This is borne out in the best case scenarios . 4. Based on proposed integrated opera- tions, changes in capital costs affect both three and four-metal operations about equally. For a 1-pct change in in- vestment, there is about a 1-pct change in ROR. 5. Variation of operating costs af- fects four-metal operations dramatically, resulting in an estimated 2.7-pct change in ROR for every percentage change in op- erating expenses. Changes in three-metal ROR are about 1.1 to 1. 6. The sensitivity analysis shows that reasonable variations from the base case still result in ROR's well below the 25- to 30-pct thought to be required. Only the best case predictions approach that level. Based on the foregoing analysis it is most likely that nodules will not be mined and processed in the foreseeable future without significant financial in- centives. The incentives could be in the form of price supports, tax breaks, or other programs such as financing research and development. REFERENCES 1. Hillman, C. T. Manganese Nodule Resources of Three Areas in the Northeast Pacific Ocean: With Proposed Mining- Benef iciation Systems and Costs. BuMines IC 8933, 1983, 60 pp. 2. Davidoff, R. L. Supply Analyses Model (SAM): A Minerals Availability System Methodology. BuMines IC 8820, 1980, p. 14. 3. Ozturgut, E., G. C. Anderson, R. E. Burns, J. W. Lavaelle, and S. A. Swift. Deep Ocean Mining of Manganese in the North Pacific, Premining Environmental Conditions and Anticipated Mining Ef- fects. Dep. Commerce-NOAA Tech. Memoran- dum ERL MESA-33, 1978, 133 pp.; available upon request from C. T. Hillman, BuMines, Spokane, WA. 4. Menard, H. W. Marine Geology of the Pacific. McGraw-Hill, 1964, 271 pp. 5. Piper, D. Z. (comp.). Deep Ocean Environmental Study: Geology and Geo- chemistry of DOMES Sites A, B, and C, Equatorial North Pacific. U.S. Geol. Surv. OFR 77-778, 1977, pp. 217-266; available for consulation at U.S. Geol. Surv. libraries in Menlo Park, CA, Golden Co, and Reston, VA. 6. Ryan, W. B. T., and B. C. Heezen. Smothering of Deep Sea Benthic Communi- ties From Natural Disasters. 1976, 132 pp.; NTIS PB 279527/AS. 7. Sorem, R. K. , and R. H. Fewkes. Manganese Nodule Research Data, and Meth- ods of Investigation. IFI/Plenum, 1979, 723 pp. 8. Monhemius , A. J. The Extractive Metallurgy of Deep Sea Manganese Nodules. Ch. in Topics in Nonferrous Extractive Metallurgy. Soc. Chem. Ind. , 1980, pp. 42-69. 9. Frazer, J. Z., and M. B. Fisk. Geological Factors Related to Character- istics of Seafloor Manganese Nodule De- posits (grant GO264024, Scripps Inst. Oceanography). BuMines OFR 142-80, 1980, 41 pp.; NTIS PB 81-145831. 16 10. Fewkes, R. H. , W. D. McFarland, W. R. Reinhart, and R. K. Sorem. Develop- ment of a Reliable Method for Evaluation of Deep Sea Manganese Nodule Deposits (grant GO 274013, WA State Univ.). Bu- Mines OFR 64-80, 1979, 94 pp.; NTIS PB 80-182116. 11. Frazer, J. Z., M. B. Fisk, J. Elliott, M. White, and L. Wilson. Avail- ability of Copper, Nickel, Cobalt, and Manganese From Ocean Ferromanganese Nod- ules (grant G0264024, Scripps Inst. Oceanography). BuMines OFR 121-79, 1978, 141 pp.; NTIS PB 300 356. 12. Felix, D. Some Problems in Making Nodule Abundance Estimates From Seafloor Photographs. Marine Min. , v. 2, 1980, pp. 293-302. 13. Frazer, J. Z. Manganese Nodule Reserves: An Updated Estimate. Marine Min., v. 1, 1977, pp. 103-123. 14. Moncrieff, A. G. , and K. B. Smale- Adams. The Economics of First Generation Manganese Nodule Operations. Min. Congr. J., v. 60, No. 12, 1974, pp. 46-50. 15. Flipse, J. E. Deep Ocean Mining Pollution Mitigation. Paper in Proceed- ings, 12th Annual Offshore Technology Conference (Houston, TX, May 5-8, 1980). Offshore Technol. Conf., Dallas, TX, pp. 353-356. 16. Gale, G. D. , (U.S. Bureau of Mines). Private communication, 1981; available upon request from C. T. Hill- man, BuMines, Spokane, WA. 17. Flipse, J. E. An Economic Analy- sis of a Pioneer Deep Ocean Mining Ven- ture (partially supported by Sea Grant NA81AA-D00092, Texas A & M Univ.). TAMU- SG-82-201; COE Rep. 262, 1982, 131 pp. 17 APPENDIX. —CAPITAL AND OPERATING COST DETAIL FOR STUDY AREA CI TABLE A-l. - Mine costs, 1 million January 1983 dollars Costs Description CAPITAL Fixed capital: Exploration Research and development.... 2 mine ships 2 nodule collectors 3 pipelines 2 pumping systems 2 sets of on-board equipment, Total Startup costs Working capital Total investment $21.0 75.3 199.5 7.5 51.7 29.7 87.1 471.8 23.2 95.6 590.6 Initial 6-yr program. 7-yr program. Capacity: 1.5 million t/yr each. Approximate width, 10 m. 40-cm ID, including 1 spare. Pumps, connectors, valves. Nodule handling, storage. Equipment testing, redesign. Basis: 1.25 yr, 3.75 million dry t. OPERATING' Direct operating: $21.3 2.0 27.5 6.0 9.2 0.1 6.4 4 crews, 2 per ship. Food and miscellaneous supplies. Ship and equipment. Ship, equipment, and personnel. Approximately 51,000 t diesel. Ship and equipment. Mine characteristics and equipment improvement. Fuel Exploration and continuing research and development. Total 72.5 4.0 5.5 pet of direct costs. 76.5 'Estimates for 2 complete mining sys 2 Total mining cost per dry metric to terns. n of ore: $25.50. 18 TABLE A-2. - Transportation costs, million January 1983 dollars Costs Description CAPITAL Fixed capital: 3 transports , Slurry terminal , Slurry pipeline , Supply boat , Total Working capital , Total investment, Capacity: 44,800 dry t ore each, Dock and all facilities, 10 ha. 40 km, includes land purchase. High-speed shuttle. Basis: 1.0 yr, 3.0 million dry t, OPERATING ' Direct operating: Transport: Wages and benefits , Subsistence and supplies, Maintenance and repair. . , Insurance , Fuel, lubrication, oil.., Port charges , Helicopter , Total , Supply boat , Total vessel costs..., Transport unload and store, Slurry pipeline to plant.., Total , General and administrative.., Annual operating cost, $4.7 .5 6.2 1.4 8.9 1.0 .7 23.4 1.8 25.2 3.1 6.5 34.8 1.9 36.7 U.S. crews, 3 transports. Food and miscellaneous supplies. Ships and equipment. Ships, equipment, and personnel. Approximately 49,400 t diesel. Docking, tieup, and miscellaneous, Crew, fuel, maintenance, repair. Do. Operation, maintenance, repair. Do. 5.5 pet of direct costs. 'Total cost per day metric ton of ore: $12.23. 19 TABLE A-3 - Process costs, million January 1983 dollars Costs Description cap: [TAL CUPRION PLANT Fixed capital: $82.5 2.6 370.3 159.1 30.3 20.9 5.7 7 yr, including pilot plant. About 90 ha, including ferromanganese site. Installed cost plus auxiliaries. Installed cost. Land, piping, ponds. 100 km, includes land purchase. 8 km each, includes land. Total 671.4 55.5 Basis: 0.5 yr, 1.5 million dry t. 726.9 FERROMANGANESE PLANT Fixed capital: 129.6 31.1 Installed cost plus auxiliaries. Installed cost. Total 160.7 54.6 Basis: 0.5 yr, 0.7 million dry t. 215.3 942.2 OPERATING CUPRION PLANT Direct operating: $18.6 4.0 40.5 31.8 7.8 2.7 0.2 Operators, maintenance, technical, professional, management. Operating chemicals and reagents. Coal, power, petroleum, water. Fixed expenses: maintenance, materials, insurance. Operating new and existing ponds. Operating, repair, and maintenance. Do. Pipeline to waste disposal... Railroad spur and access road Total 105.6 5.3 5.0 pet of direct operating. 110.9 FERROMANGANESE PLANT 2 Direct operating: 32.0 43.6 118.6 12.1 Operators, maintenance, technical, professional, management. Operating chemicals and reagents. Coal, power, petroleum, water. Fixed expenses: maintenance, materials, Total 206.3 10.3 insurance. 5.0 pet of direct operating. 216.6 327.5 'Total cost per dry metric ton of ore: $109.17 — $36.97 (Cuprion) and $72.20 (ferromanganese) . 2 Includes flotation, calcining, ferromanganese to east-coast marke *U.S. CPO. 198S-50S-019a0.024 smelting, and handling as well as shipment of the ts. INT.-BU.OF MINES, PGH..P A. 27940 K442 D DD n O > "1 3 O PO Q_ o D o 0) CD n Q_ 03 3 -*■ 3 O -*• c 3 -i- o Q (/> o c Q S in n w _, 3T 3 T3 =T in ■D Q Q -♦■ fD Q_ 3 o 3 (O 37 Q -* Q_ n o (t Q • to Q o -i -*■ n a • 33 ft D 3 o < -t- 3- (/) < o w ^3 m w z CD ■< -n 3q 33 > ■v i- BUREAU OF IV 4800 FORBES A URGH, PENNSY >i m 2 i- < = c m < m Z £ w > Z m - m u> zc w (A £™ D m > H m Z oo "n ty> HH _ m zw -H m O 33 m O c > r - O -u TJ o 3J H C Z H -< m O •< m 33 c "0 S > w s ° m m Z z > — 1 -H 2 h. o o o! ■n Ti -I m I m m w 32 m O 33 o 33 A %» <* » ^V ' » t,, a* <* •^••' aO* ^ "••»* A <> ♦'TV?* .0* ^ '».>' A /\ : Jlt : /% flB /\ : llt : /% flS ,X . " A V ^£. •'^ii^'o cS^n, oW^^W" A^"^ •*^B^P' e ^°?n oW^^tf* A^ -^Ei^' ^^r. *^* A*' • A V "^ <£* r. ° H <>» <$>. ^-•^V V*-.. ^ V / V^-/ %»/ V^-> v^v v^ -^v ^VV^ o*j| ** A^ ^"7 C, if V^wvg^ \*-.^\^ ^v^^\^ X* : — : V I >0 o V °^. v«i 'S J 'A ^ ^5°^ ^ c^ /. a ^ ^ •'TV.' ,0 V r oV •^•o* cC^ 4 ^ 4 O 1 ^ .P ■..:• y \#/ v^V* \'-W-> v^ *v ^0 A^ • ••. <* ,* y ... **\^:.-:.\_ y,.^-.°o /,-^:-.\ rfi*...- P-% ^oV* - — - v* A.* ' ' .vV>u jp-v V LIBRARY OF CONGRESS lIlllllllM 002 955 967 3 PI