551.49 Un3426W 89-4174 PREDEVELOPMENT HYDROLOGY OF THE GILA RIVER INDIAN RESERVATION, SOUTH-CENTRAL ARIZONA f ^ ^ U.S. GEOLOGICAL SURVEY fc Water-Resources Investigations Report 89—4174 ■ ljIL ’-.s- ’ . BEWMITORY 7 1991 OF IILINOIS - at MP^ Prepared in cooperation with the U.S. BUREAU OF INDIAN AFFAIRS UNIVERSITY 0P ILLINOIS LISBABV AT URBANA-CHAM^A^ GEOLOSY PREDEVELOPMENT HYDROLOGY OF THE GILA RIVER INDIAN RESERVATION, SOUTH-CENTRAL ARIZONA By B.W. THOMSEN and J. H. EYCHANER U.S. GEOLOGICAL SURVEY Water-Resources Investigations Report 89—4174 Prepared in cooperation with the U.S. BUREAU OF INDIAN AFFAIRS Tucson, Arizona June 1991 U.S. DEPARTMENT OF THE INTERIOR MANUEL LUJAN, Jr., Secretary U.S. GEOLOGICAL SURVEY Dallas L. Peck, Director For additional information write to: Copies of this report can be purchased from: District Chief U.S. Geological Survey 375 South Euclid Avenue Tucson, Arizona 85719 U.S. Geological Survey Books and Open-File Reports Section Federal Center, Box 25425 Denver, Colorado 80225 CONTENTS Page Abstract . 1 Introduction . 1 Location, physiography, and climate . 2 Purpose and scope . 4 Previous investigations . 5 History of water development . 6 Geology . 8 Hydrology . 8 Precipitation . 9 Streamflow . 10 Ground-water flow . 17 Underflow . 19 Hydraulic characteristics of the ground-water reservoir .. 20 Evapotranspiration . 22 Water budget . 24 Simulation of ground-water flow . 26 Computer program . 27 Estimation of model parameters . 29 Calibration . 30 Simulation results . 31 Sensitivity analysis . 32 Summary . 38 Selected references . 38 ILLUSTRATIONS [Plates are in pocket] Plates 1-2. Maps showing: 1. Finite-difference grid and boundary conditions used in the mathematical model, Gila River Indian Reservation 2. Calibrated transmissivity, simulated water levels, and measured water levels, Gila River Indian Reseirvation III IV ILLUSTRATIONS Page Figure 1. Map showing location of study area. 3 2. Graph showing precipitation at Tucson, 1876-1984. 11 3. Map showing location of streamflow and precipitation stations . 13 4-6. Graphs showing: 4. Annual runoff of the Gila River at Kelvin. 15 5. Average base flow of the Gila River near Gila Crossing . 18 6. Relation between depth to ground water and annual water use by mesquite. 24 7. Schematic diagram of conceptual longitudinal section showing initial estimates of average annual values, in acre-feet, of water-budget components . 25 8-12. Graphs showing: 8. Functions for estimating evapotranspiration. 31 9. Sensitivity of water-budget components to variations in transmissivity. 35 10. Sensitivity of water-budget components to variations in evapotranspiration . 36 11. Sensitivity of water-budget components to variations in the leakance coefficient. 37 12. Sensitivity of water-budget components to variations in mountain-front recharge and underflow . 39 TABLES Page Table 1. Precipitation data . 10 2. Streamflow data . 14 3. Comparison of basin characteristics . 17 4. Simulated predevelopment ground-water budget . 33 V CONVERSION FACTORS For readers who prefer to use the International System (SI) units, the conversion factors for terms in this report are listed below: Multiply inch pound unit By To obtain metric unit inch (in.) 25 foot (ft) 0 mile (mi) 1 square mile (mi^) 2 acre 0 gallon per minute 0 (gal/min) acre-foot (acre-ft) 0 acre-foot per acre 304 (acre - ft/acre) acre-foot per square mile 0 (acre-ft/mi^) cubic foot per second 0 (ftVs) foot squared per day 0 (ftVd) foot per mile 0 (ft/mi) degree Farenheit (°F) °C = 5/ 4 millimeter (mm) 3048 meter (m) 609 kilometer (km) 590 square kilometer (km^) 4047 square hectometer (hm^) 06309 liter per second (L/s) 001233 cubic hectometer (hm®) 8 millimeter (mm) 47625 millimeter (mm) 02832 cubic meter per second (m^/s) 0929 meter squared per day (m2/d) 1894 meter per kilometer (m/km) X (°F-32) degree Celsius (°C) Sea level : In this report "sea level" refers to the National Geodetic Vertical Datum of 1929 (NGVD of 1929)—geodetic datum derived from a general adjustment of the first-order level nets of both the United States and Canada, formerly called "Sea Level Datum of 1929." Digitized by the Internet Archive in 2018 with funding from University of Illinois Urbana-Champaign Alternates https://archive.org/details/predevelopmenthy8941thom PREDEVELOPMENT HYDROLOGY OF THE GILA RIVER INDIAN RESERVATION, SOUTH-CENTRAL ARIZONA By B.W. Thomsen and J.H. Eychaner ABSTRACT Indians have occupied and irrigated the reservation lands along the Gila River for centuries. Knowledge of the hydrologic conditions that existed before development of the water resources by non-Indian settlers is needed to aid in evaluation of water-right claims. Water resources of the Gila River Indian Reservation for the period before development by non-Indian settlers were characterized by perennial flow in the Gila River and ground water near the land surface. Depths to water ranged from 10 to 70 feet and the direction of ground-water movement was from the northeast, east, and southeast toward the northwest. The ground-water reservoir was essentially in equilibrium and was sustained mainly by the infiltration of water from the Gila River. Mesquite thickets and groves of cottonwood trees covered large parts of the flood plain and other lowlands along the Gila River. Evapotranspiration averaged 200,000 acre-feet per year. On the basis of available streamflow data and long-term runoff estimates from tree-ring data, the mean annual flow of the Gila River upstream from the reservation was estimated to be 500,000 acre-feet and the median annual flow 380,000 acre-feet. Tree-ring data do not indicate a significant change in precipitation from 1602 to 1970. Ground water was discharged by evapo¬ transpiration by phreatophytes where water levels were shallow and by return to surface flow mainly in the western third of the reservation where bogs and sloughs were common. A numerical model was developed to simulate ground-water flow, s t re am - aqui f e r connection, and evapotranspiration for purposes of evaluating predevelopment hydrologic conditions. The model represents average conditions in the ground-water system before the system was affected significantly by diversions upstream from the reservation. Average values for components of ground-water flow determined from the model include recharge by infiltration from the Gila River, 94,000 acre-feet per year; evapotranspiration from ground water, 96,000 acre-feet per year; and discharge to surface flow in the western third of the reservation, 29,000 acre-feet per year. INTRODUCTION The Gila River Indian Reservation was established in 1859 in an area along the Gila River that had been occupied and irrigated by the 1 2 Indians for centuries (Bancroft, 1889; Haury, 1976, p. 357; Ezell, 1963). Non-Indian settlers arrived in Arizona in large numbers in the 1860's and 1870's and began diverting water from the Gila River and its tributaries upstream from the Gila River Indian Reservation. The development and activities that have occurred since that time have significantly changed the hydrology of the area. The flow of the Gila River and recharge to the ground-water system on the reservation have been greatly diminished as a result of upstream diversions and storage. Water levels in wells have declined and the direction of ground-water flow has changed as a result of pumping for irrigation in areas adjacent to the reservation. General adjudication to determine water rights of water users in the Gila River watershed is being conducted in the superior courts of Arizona under authority established by Arizona Revised Statutes Title 45, Chapter 1, Article 6. In order to develop data pertinent to the adjudication process, the U.S. Bureau of Indian Affairs entered into a cooperative agreement with the U.S. Geological Survey to evaluate the hydrologic conditions that existed prior to the arrival and development of the area by non-Indian settlers. Location, Physiography, and Climate The study area includes about 2,200 square miles in south-central Arizona, of which about 580 square miles is in the Gila River Indian Reservation (fig. 1). The reservation consists of a parcel of land about 50 miles long and 3 to 10 miles wide on both sides of the Gila River. The study area includes part of the Salt River Valley north of the reservation and the lower Santa Cruz basin south of the reservation. The Ak-Chin Indian Reservation, the north tip of the Papago Indian Reservation, and the south edge of the Salt River Indian Reservation are within the study area. The study area is characterized by broad desert plains dissected by arroyos, and valleys separated by rugged low-lying mountains. The altitude of the desert plains ranges from 1,600 feet above sea level east of the reservation to less than 1,000 feet at the northwest corner. To the east and northeast of the reservation, the terrain slopes irregularly upward to an altitude of more than 5,000 feet in the Superstition mountains. The dominant native vegetation types are mesquite and saltbush along the washes and palo verde and cacti on the hills. Creosotebush covers most of the desert floor except where it has been replaced by crops. Dense thickets of mesquite and groves of cottonwood and willow covered large areas along the rivers when the non-Indian settlers arrived ('-ee, 1905) but most have been removed. Surface drainage includes parts of three major rivers the Gi..a, Salt, and Santa Cruz (fig. 1). The Salt and Santa Cruz Rivers are tributaries to the Gila River and both join the Gila near the northwest corner of the Gila River Indian Reservation. The Gila River drains more than 18,000 square miles east and southeast of the reservation. The Salt River and its major tributary, the Verde River, drain more than 12,000 square miles north and northeast of the reservation. The Santa Cruz 3 I-1-*—I-r* 0 50 100 150 KILOMETERS Figure 1.--Location of study area (shaded) 4 River drains more than 8,000 square miles south and southeast of the reservation (fig. 1). The Gila and Salt Rivers contributed perennial flow to the study area prior to the arrival of non-Indian settlers but now are only intermittent streams. The Santa Cruz River flows mainly in response to intense rainfall. The climate is dry and incapable of supporting more than a minimum vegetation growth without irrigation. Summers are hot, and daily temperatures generally exceed 100 °F from mid-June through August. Mean daily temperatures range from about 64 "F to 105 °F. The relative humidity generally is low, ranging from about 20 to 50 percent (Sellers and Hill, 1974). Winters are mild, and average temperatures range from 60 ®F to 80 ®F in the afternoons and from 30 “F to 40 ®F in early mornings. Subfreezing temperatures occur on only a few days during an average year (Sellers and Hill, 1974). Mean daily temperatures range from about 33 ^F to 70 “F. Annual precipitation averages about 8 inches and results mainly from two types of storms. Summer thunderstorms, which develop as a result of the flow of moist tropical air from the Gulf of Mexico, make July and August the wettest months. Regional storms from the Pacific Ocean produce gentle widespread showers during the fall and winter months. April, May, and June are the driest months. Occasional tropical storms produce large amounts of rain in the fall. Wind movement in the area is relatively light. In 1895, the monthly average wind speed was about 5 miles per hour at Phoenix (Davis, 1897a, p. 31). U.S. Weather Bureau records for January 1948 through December 1955 at Phoenix show that average wind speeds did not exceed 8.3 miles per hour (Sellers and Hill, 1974, p. 30). Purpose and Scope The purpose of this report is to describe the hydrologic conditions that existed in the area of the Gila River Indian Reservation prior to development by non-Indian settlers. Non-Indian settlers were diverting significant quantities of water from the Gila River and its tributaries upstream from the reservation in the 1870's (Olberg, 1919). No pre- 1870 hydrologic data are available; therefore, data collected since 1870 were used to evaluate predevelopment conditions. The results of the evaluation represent average hydrologic conditions during the 100-year period prior to 1870. The 100-year average was used in order to dampen the effect of short-term variations in hydrologic conditions. The evaluation of hydrologic conditions prior to 1870 required estimating the flow of the Gila River upstream from the Gila River Indian Reservation and defining the ground-water system in and adjacent to the reservation. Estimates of average flow of the Gila River were made from recorded data with adjustments to represent predevelopment conditions. The adjustments were based on the effects of development on river flows and the mathematical evaluations of climatic trends. Studies of relations between streamflow and tree rings were used to help substantiate estimates of the predeve 1 opment flow of the Gila River. The ground-water system was 5 evaluated by the use of a numerical model. The model parameters were estimated initially from published values and recorded field data; each parameter was estimated independently. Evapotranspiration was calculated by using old maps and photographs to determine types and areas of vegetation and applying evapotranspiration rates determined in recent studies. The model covers an area slightly larger than the reservation (fig. 1) in order to encompass parts of the mountain ranges that form physical boundaries to much of the ground-water system. Previous Investigations A reconnaissance of the water supply for the Gila River Indian Reservation was made by the U.S. Geological Survey in 1886-87 (Davis, 1897b, p. 8). A streamflow-gaging station was established on the Gila River at the Buttes 12 miles upstream from Florence in 1889, and records were obtained for 1 year. The station was discontinued in 1890, reestablished in 1895, and operated through September 1899 (Lippincott, 1900). An investigation of the water supplies available for irrigation in the Salt and Gila Valleys near Phoenix, Arizona, was made in 1896 by Arthur P. Davis. This investigation dealt mainly with surface-water supplies and the results, as they pertained to irrigation waters for the Gila River Indian Reservation, were reported to the U.S. Senate (Davis, 1897b). In a more general report, Davis (1897a) described the topographic and climatic conditions, irrigation works in use and under construction, facts relating to water supply, underground waters, evaporation, silting of reservoirs, and legal problems related to storing and diverting water. Davis (1897a) recommended a more detailed investigation of potential reservoir sites on the Gila River, which was conducted by Lippincott (1900). W.T. Lee investigated the "underground waters" of the Gila Valley (1904) and the Salt River Valley (1905). The reports present tabulations of well records, water levels, and chemical quality of ground water. Also included are descriptions of the geology, physiography, and economics of pumping ground water. Ground waters of the Arizona territory were examined as to their suitability for sanitary, irrigation, and technical uses (Skinner, 1903). A study of the geology and water resources of the Gila River Indian Reservation was made by the U.S. Department of Agriculture (Rott, 1936). A quantitative study of the ground-water resources of the Eloy district in Pinal County, Arizona, was conducted by the Agricultural Experiment Station of the University of Arizona (Smith, 1940). The U.S. Geological Survey investigated the ground-water resources of the Queen Creek area (Babcock and Halpenny, 1942), the Santa Cruz basin (Turner and others, 1943), and western Pinal County (Hardt and others, 1964; Hardt and Cattany, 1965; Kister and Hardt, 1966). An electrical-analog model of the ground-water system in central Arizona was constructed to determine the probable future effects of continued ground-water withdrawal (Anderson, 1968). The model was 6 constructed by using the known hydrologic characteristics of the water¬ bearing rocks and the pumping history through 1964. Predictions of future water levels were based on an extension of the pumping rate and areal distribution that existed in 1964. Maps showing water-level changes for 1923-47 and 1923-64 were prepared and estimates of ground-water pumpage and ground water in storage were made for the Gila River Indian Reservation by Babcock (1970). Water- level altitudes were updated for the eastern part of the Salt River Valley (Laney and others, 1978) and for the lower Santa Cruz area (Konieczki and English, 1979). Descriptions of hydrologic conditions in alluvial basins (Freethey and Anderson, 1986) and distribution of aquifer materials in alluvial basins (Freethey and others, 1986) are pertinent to the study area. History of Water Development Prior to the arrival of non-Indian settlers, the Indians diverted water from the Gila River for the irrigation of cropland but the amount of land that was irrigated is uncertain. Olberg (1919) estimated that the largest area cultivated by the Indians at any one time (about 1885) was 15,800 acres and that an additional 11,315 acres had been previously irrigated. Land classification by the U.S. Reclamation Service shows that 44,900 acres of reservation land was cultivated or formerly cultivated (unpublished reports on proposed allotments for Gila River Indian Reservation, C.R. Olberg, Superintendent of Irrigation, September 1913). Other documents describe 25,000 to 28,000 acres of land under this classification (Paul Gregory, hydrologist, U.S. Bureau of Indian Affairs, written commun., 1987). The development of water resources by settlers began with the diversions of water from the Gila River and its tributaries for irrigation. Diversions started in the 1860's and increased rapidly in the 1870's and 1880's. Water in the Gila River was diverted for irrigation in the valleys of the upper Gila River above Coolidge Dam, along the San Pedro River, and in the Gila Valley near Florence just upstream from the Gila River Indian Reservation. Diversions from the Gila River near Florence after 1864 probably had the most direct effect on the quantity of streamflow available for use on the reservation. By 1885, twelve ditches—four on the north side of the Gila River and eight on the south side—were capable of irrigating 6,000 acres (Olberg, 1919). The major diversion was to the Florence Canal, which first carried water in 1887. The Florence Canal and Land Company filed water claims in 1885, started construction in 1886, and diverted water into the first 16 miles of the Florence Canal in 1887. By 1889, the length of the canal had been extended to 50 miles and Picacho Reservoir, a shallow storage reservoir on McClellan Wash south of Florence, had been constructed (Southworth, 1919). The capacity of the Florence Canal was 440 cubic feet per second. The capacity of Picacho Reservoir was 24,500 acre-feet (U.S. Senate, 1890, p. 455). When the canal was built, the plans specified the irrigation of 60,000 acres (U.S. Senate, 1890). The maximum area under 7 irrigation however was about 7,000 acres (Southworth, 1919). The total diversion into the Florence Canal during the irrigation season of 1896 was 64,444 acre-feet and the area irrigated was 6,472 acres (Davis, 1897a). The quantity of water used was great, about 10 acre-feet per acre, because of large seepage and evaporation losses. In contrast, the average water use on the Salt River for 60,000 acres irrigated in 1895 was 4.6 acre-feet per acre (Davis, 1897a). The number of acres irrigated from the Florence Canal was much smaller than planned probably because the quantity of water available was less than anticipated. Irrigation systems were being constructed simultaneously in the upstream valleys of the Gila River. By the end of the 19th century, at least 40 canals were being used to divert water from the Gila River and its tributaries for the irrigation of about 40,000 acres upstream from the Buttes (summation from Southworth, 1919). Diversions upstream from the Buttes probably were at least 60,000 acre-feet per year because a minimum of 14 feet of water was needed to produce a crop (Davis, 1897a). Water shortages that resulted from the diversion of water from the Gila River upstream from the reservation caused some Indians to move from the area near the Gila River to near the Salt River in 1872 (Olberg, 1919). The Indians abandoned old fields along the Gila and reclaimed lands at places favorable to the use of return flows of the river upstream from Sacaton. Irrigation of lands in the Gila Crossing area, not formerly used because of poor-quality water, began in the 1870's owing to the shortage of water on the Gila River farther upstream (Southworth, 1919). As diversions from the Gila River increased, the available supply of water in the river decreased. Decreases were most noticeable during seasonal low-flow periods in the areas downstream from Florence, and the need for water storage soon became evident (Davis, 1897a). Ashurst-Hayden Dam, a diversion structure on the Gila River, was completed in 1922 and Coolidge Dam, a storage reservoir, was completed in 1928. Because of the lack of streamflow for irrigation during parts of some years, wells were constructed for irrigation in the late 1890's (Davis, 1897a), but only small quantities of ground water were withdrawn until many years later. The demand for agricultural products during World War I caused an increase in ground-water withdrawals. The types of pumping plants available at that time had physical limitations and high operating costs; hence, quantities of water pumped remained relatively small, less than 100,000 acre-feet per year in the Florence-Casa Grande area prior to 1925. Electric power became available in the late 1920's and the quantity of water pumped increased gradually to about 350,000 acre-feet per year by 1940. These increases resulted from high cotton prices, improvement in the design and efficiency of pumping plants, and decreases in electric-power rates (Turner and others, 1943). During the 1940's, rates of ground-water withdrawal continued to increase and reached the present-day level of about 1 million acre-feet per year from the lower Santa Cruz basin in 1948 (U.S. Geological Survey, 1983). 8 GEOLOGY The study area lies within the Basin and Range physiographic province (Fenneman, 1931), which is characterized by broad alluvial valleys separated by rugged mountains. The mountains are composed mainly of granitic, volcanic, and metamorphic crystalline rocks that yield very little water. The valley floors are underlain by a wide variety of sedimentary deposits that constitute the main ground-water reservoirs. These deposits consist of unconsolidated to variably consolidated sediments that are several thousand feet thick in places. The sediments were deposited in river beds, flood plains, lakes, fans at the foot of mountain slopes, and estuaries. Sediments include unconsolidated clay, silt, sand and gravel, caliche, gypsum, mudstone, siltstone, sandstone, conglomerate, and anhydrite. The degree of sorting and cementation and the distribution of the different materials vary areally and with depth. Interbedding and lensing are common, and lateral discontinuities caused by high-angle faults may be present in some older units (Laney and Hahn, 1986). The sediments have been divided into four geohydro1ogic units—pre-Basin and Range deposits, lower basin-fill deposits, upper basin-fill deposits, and stream alluvium—on the basis of their geologic and hydrologic properties (Freethey and others, 1986). The pre-Basin and Range deposits range from silt, clay, and claystone to gravel and conglomerate. The lower basin-fill sediments were deposited in closed basins and consist of weakly to highly consolidated gravel, sand, silt, clay, and evaporites. The upper basin-fill sediments consist of unconsolidated to moderately consolidated gravel, sand, silt, and clay and represent deposition during and after the transition from a closed-basin to an integrated-drainage environment. The stream alluvium consists typically of well-sorted sandy gravel with some silt; the alluvium is areally extensive along the Gila and Salt Rivers and in some places exceeds 300 feet in thickness. HYDROLOGY The hydrologic cycle is a term used to denote the circulation of water from the ocean, through the atmosphere, to the land, and back to the ocean. The movement of water over and through the land is the main concern of this study. Water that moves over the land surface tends to collect and become streamflow. The quantity and duration of streamflow, in general, depend on the amount, intensity, and type of precipitation and on the nature of the material over which the water passes. As streamflow moves along natural channels, some water may evaporate and thus be lost from the local system, or part or all of it may percolate into receptive materials and become either soil moisture or ground water. Water that percolates into the earth from either precipitation or streamflow and reaches the water table, or the zone of saturation, is 9 called ground water. Water that is retained in the unsaturated zone above the water table is called soil moisture. Water in the subsurface may return to the land surface and become streamflow where the water table intersects the land surface. The water may move into the unsaturated zone to become soil moisture or it may be removed from the local system by evapotranspiration or by pumping. Precipitation Precipitation is the source of water, but not all the precipitation that reaches the land surface is available for man's use. Water that reaches the land surface as precipitation may proceed along any of three general paths. It may evaporate soon after contact with the land surface, move across the land as surface runoff, or penetrate the earth to become either soil moisture or ground water. Recorded precipitation data indicate that the quantity of precipitation may be extremely different from year to year, and studies of past climates show long-term changes in precipitation amounts (Sellers, 1965). Precipitation data were collected at U. S. Army Posts on the Gila River watershed as early as 1867 (Davis, 1897a). However, most long-term precipitation records in Arizona began between 1895 and 1915, at least 25 years after the period of interest for this study. Hence, an evaluation of the possibility of a climatic change was needed. Precipitation in the study area averages about 8 inches per year and occurs mainly as rain. Snow falls in the upper reaches of the rivers that flow into the study area. Total precipitation averages nearly 1 million acre-feet per year, of which about 250,000 acre-feet per year falls on the reservation. Most of the rainfall on the flatlands of the study area evaporates or is used by vegetation, and virtually none reaches the ground-water reservoir. Precipitation on the mountains tends to collect in channels as run off and may be sufficient in quantity at times to provide recharge to the ground-water system along the mountain fronts. Annual precipitation at Sacaton for 1931-72 ranged from 1.85 to 17.21 inches and averaged 8.37 inches (Sellers and Hill, 1974). Precipitation is less than potential evapotranspiration in all months , but particularly so in April, May, and June. Two sets of annual precipitation data in and near the study area were examined for any indications of long-term climate changes. One set contains data from nine precipitation stations operated for several years prior to 1900 and again during 1941-70 (table 1). The average annual precipitation prior to 1900 is greater than that for 1941-70 for five stations and less for four stations. The average for the nine stations is 9.79 inches for the period prior to 1900 and 9.93 inches for the later period. The other set of precipitation data (fig. 2) is for Tucson and is the longest continuous precipitation record in Arizona. During 109 years (1876-1984), annual precipitation at Tucson ranged from 5.07 to 24.17 inches and averaged 11.41 inches. The median or middle-rank value was 10.94 inches. Averages for 10-year periods were all within 25 percent 10 Table 1-Precipitation data [Prior to 1900 from Lee, 1904; after 1900 1974. See figure 3 for location from Sellers and Hill, of stations] Station Number of years of record Average annual precipitation, in inches 1873-98 1941-70 1873-98 1941-70 Casa Grande. 14 28 5.33 8.12 Florence. 11 30 9.74 9.50 Fort Grant. 24 23 15.34 12.52 Fort Thomas. 10 7 12.32 9.01 Phoenix. 16 17 7.35 7.59 San Carlos. 17 30 12.31 11.88 San Simon. 14 18 4.65 8.83 Tucson. 23 30 11.68 10.73 Willcox. 17 30 9.42 11.19 Average. 9.79 9.93 of the overall average. The data show no obvious long-term trend, although periods such as 1942-56 and 1964-72 were consistently above or below the average (fig. 2). Kendall's tau-b statistic (SAS Institute, 1982, p. 501-512) was used to test for trends in the data. The statistic measures the degree to which annual precipitation increases or decreases from the beginning to the end of the record. The computed tau-b was 0.0075, a value that could be expected in 90 percent of samples having no trend. Thus, the Tucson data indicate no trend in precipitation at a 90- percent confidence level. Fritts and others (1979) used tree-ring data to evaluate climatic variations over a longer time period (1602-1970) and showed that average winter precipitation during 50-year intervals can vary by 20 percent over much of the United States. Percentage of agreement, however, between reconstructed and observed precipitation was greatest in southwestern United States, including Arizona. Precipitation and tree-ring data indicate that the precipitation regime in the study area before 1870 probably was similar to the current regime. Therefore, precipitation estimates that are based on recorded data are considered to be representative of predevelopment time. Streamflow The Gila River was the main source of water for native inhabitants when the Gila River Indian Reservation was established. The river probably flowed perennially through the reservation before non-Indian settlers arrived (Brown and others, 1981). Streamflow records are 11 o cr» o> o 00 cr> o hx o cr> o CO CT» o CM CT» o o> o o CTt o cr> 00 o 00 00 o 00 CM CM S3H3NI NI ‘N0IiViIdID3yd Figure 2.--Precipitation at Tucson, 1876-1984. 12 available for several sites on the Gila River and major tributaries (fig. 3). Streamflow measurements were made as early as 1888, but most long-term records started in 1911 or later (table 2). However, diversions of water from the Gila River upstream from the reservation began long before the earliest streamflow measurements were made and had depleted the flow of the river by 1895 (Davis, 1897b). The channel was dry throughout most of the reach from east of Florence to near Pima Butte. An exception was near Coolidge where ground water reached the surface and formed a stream of several cubic feet per second (Lee, 1904). The nearest gaging-station sites on the Gila River upstream from the reservation are at the Buttes—12 miles east of Florence—and at Kelvin—28 miles east of Florence (fig. 3). The station at the Buttes was operated for a total of 5 years—1889-90 and 1895-99—and the station at Kelvin has been operated continuously since 1912. The records at the two sites are virtually equivalent because the drainage areas differ by less than 2 percent. Beginning in 1928, the flow of the Gila River at Kelvin was affected by water storage upstream in the San Carlos Reservoir. Annual runoff data used in this study were adjusted for storage. Annual runoff of the Gila River at Kelvin upstream from the Gila River Indian Reservation ranged from 16,700 acre-feet in 1974 to 2,246,000 acre-feet in 1915 (fig. 4). For the period of record 1912-84, the mean annual runoff was 359,000 acre-feet and the median annual runoff was 238,000 acre-feet. The mean is the arithmetic average of the annual values. The median is the middle-ranked value; that is, half the values are larger and half are smaller than the median. Streamflow data for the Gila River at the Buttes or Kelvin for three periods of record are summarized in table 2. For each period, the mean is larger than the median, which is the typical relation for runoff data. In addition, median runoff during later periods is smaller than earlier periods. The flow of the Gila River upstream from the Gila River Indian Reservation prior to 1870 was estimated from recorded streamflow with adjustments to account for changes since 1870 (fig. 4). A plot of the annual runoff data indicates a decreasing trend of runoff with time (fig. 4). A test of the data for 1912-84 using Kendall's tau-b statistic (SAS Institute, 1982, p. 501-512) showed a highly significant trend of decreasing runoff. The decrease in runoff probably was caused by increasing upstream diversions and not by decreasing precipitation. Median flows for 5-year periods were computed from the 1912-84 data, and the trend of the medians was estimated (fig. 4). The trend lines represent long-term median runoff; the 5-year medians fall both above and below the trend lines during some periods. Median runoff about 1870 was estimated from the trend at 360,000 ±80,000 acre-feet per year. Tree-ring data were also used to estimate runoff prior to 1870. Trees preserve a record of climatic variations in their annual growth rings. Long-term growth records and a shorter term streamflow record can be used to estimate streamflow for the longer period using statistical multiple regression (Fritts, 1976). Long-term runoff estimates using tree-ring data were made for three sites on tributaries of the Gila River (Stockton, 1971, 1975; Smith, 1981; Smith and Stockton, 1981), but no equivalent estimates are available for the Gila near the reservation. The three sites with long-term rn fO ro CsJ ro 14 Table 2. --SCreamflow data Annual runoff Station number Station name Water years Mean Median Acre-feet Inches' Acre-feet 09474500 Gila River: at the Buttes 1890, 1896-99 466,000 0.48 450,000 09474000 at Kelvin 1912-28 550,000 .57 405,000 at Kelvin^ 1912-84 359,000 .37 238,000 09479500 near Laveen 1941-46, 1949-84 22,700 .02 8,900 09500500 Salt River: at Roosevelt 1888-1907, 1910-13 756,000 2.44 491,000 09498500 near Roosevelt 1913-84 646,000 2.81 502,000 09502000 below Stewart Mountain Dam^ 1931-84 713,000 2.15 430,000 09508500 Verde River: below Tangle Creek 1945-84 409,000 1.39 292,000 09510000 below Bartlett Dam^ 1888-1984 498,000 1.51 396,000 09486500 Santa Cruz River: at Cortaro 1940-46, 1951-82 33,200 .18 23,200 09489000 near Laveen 1941-46, 1949-84 16,300 .04 8,000 09444000 San Francisco River: near Glenwood, NM 1927-84 57,400 .65 36,000 'One inch of runoff is the volume equivalent to a layer of water one inch deep over the entire watershed. ^Data adjusted for changes in storage in major upstream reservoirs. The complete 8-digit station number for each station, such as 09474500, includes the 2-digit part number "09" plus the 6-digit downstream order number "474500," of which only the first four digits are used on figure 3. (2,246) 15 o o ID o o o o o o LO o CT> o 00 cr> o cn o VO cn o ID cn o cn o CD cn o c\j cn o - 133d-3y3V JO SQNVSnOHl NI ‘JdONny Figure 4.--Annual runoff of the Gila River at Kelvin. 16 streamflow estimates are Salt River near Roosevelt, Verde River below Tangle Creek, and San Francisco River near Glenwood, New Mexico (fig. 3). The Verde River basin is most similar to the Gila River basin (table 3), but the long-term streamflow estimates are for winter only. Because summer streamflow records on the Verde River are affected by increasing irrigation diversions, a consistent base period is not available for calibrating the regression equations (Smith, 1981, p. 40-41). Estimates of annual runoff using tree-ring data are available for the Salt and San Francisco Rivers. The Salt and San Francisco River basins are smaller, wetter, at higher altitudes, and more heavily forested than the Gila River basin (table 3). The Salt River basin is more similar to the Gila River basin in size, elevation, and channel slope, however, and is least influenced by irrigation diversions. Thus, the tree-ring estimates of runoff for the Salt River were selected as a basis for estimating predevelopment flow of the Gila River at Kelvin. The 100-year period 1770-1869 was selected to represent predevelopment conditions. The ratio of median annual estimated runoff during 1770-1869 to median annual recorded runoff during two periods for the Salt River near Roosevelt was used to estimate predevelopment flow of the Gila River at Kelvin. The earliest recorded runoff was least affected by diversions and therefore provides the most reliable estimate of median runoff during 1770-1869. The resultant estimate of median annual runoff of the Gila River is 525 cubic feet per second or 380,000 ±50,000 acre-feet, which is within the range of 360,000 ±80,000 acre-feet estimated from the trend of median runoff (fig. 4). Mean runoff on the Gila River for the period of record generally is 20 to 40 percent higher than the median. Mean annual runoff for predevelopment conditions probably was also 20 to 40 percent higher and therefore about 500,000 ±75,000 acre-feet. The estimated mean is consistent with the estimate of 510,600 acre-feet per year as the average runoff of the Gila River at Kelvin during 1868-1941 (Gorps of Engineers, unpublished report, 1945). Ground water was discharging into the Gila River channel in the western third of the reservation in 1895 when the channel was dry in most other places on the reservation (Davis, 1897b). Discharge measurements made on June 1, 1903, of the flow in six diversion ditches and the river channel near Gila Grossing totaled about 51 cubic feet per second (Lee, 1904). Fifty measurements of flow of the Gila River at Gila Grossing, including flow in irrigation ditches, made during 1934-40 ranged from 6.1 to 48.4 cubic feet per second (G.J. Moody, U.S. Indian Irrigation Service, written commun., 1940). A streamflow-gaging station was established on the Gila River near Laveen (at Gila Grossing) in 1940. At that time, flow in the river was perennial and base flow was estimated from daily flow records for the gaging station and measurements of diversion upstream from the gage (U.S. Geological Survey, 1941). The average annual base flow at Gila Grossing decreased gradually from a high of 25 cubic feet per second in the 1941 water year to 0 in 1957 (fig. 5). The decrease in base flow (ground-water discharge) resulted from lowering of the ground-water level. During 1929-40, miscellaneous measurements, which were made monthly of the flow of the Gila River at the junction with the Salt River, indicate that the base flow was about twice the base flow at Gila Grossing and decreased with time. Average discharge was 57.2 cubic feet per second 17 Table 3.-- Comparison of basin characteristics stream Drainage area, in square mi les Annual precipi- tat ion, in inches Annual snowfal1, in inches Mean altitude, in feet Forest area, in percent Channel slope, in feet per mile Irrigated area, 1982, in acres per square mile Gila River at Kelvin. 18,011 15.7 23 5,150 13 18 4.6 Salt River near Roosevelt. 4,306 22.0 44 6,190 71 23 0.9 Verde River below Tangle Creek. 5,499 18.4 32 5,470 67 16 2.3 San Francisco River near Glenwood, New Mexico. 1,653 17.6 52 7,780 85 55 1.2 for 1929-34 and 45 cubic feet per second for 1934-40 (P.V. Hodges, U.S. Indian Irrigation Service, written commun., 1941). Other streams that affect the reservation's water resources are the Salt and Santa Cruz Rivers; Queen Creek; Santa Rosa, Greene, and Vekol Washes; and many small streams along mountain fronts. The Salt River was perennial when non-Indian settlers arrived. The Santa Cruz River was ephemeral except near the confluence with the Gila River (Brown and others, 1981). Queen Creek was perennial in the upper reaches but most of the flow infiltrated into the alluvium near the contact with the bedrock. Santa Rosa, Greene, and Vekol Washes and other ephemeral channels that drain the mountains in and around the study area carry water only in response to heavy rainfall. Flow in the ephemeral channels generally travels only a short distance onto the valley floor before all the water infiltrates into the underlying material. Ground-Water Flow Recharge to the ground-water system occurs mainly from infiltration of streamflow. Prior to major diversion of water for irrigation, the Gila River was the main source of recharge. Queen Creek; the Santa Cruz River; Santa Rosa, Greene, and Vekol Washes; and other ephemeral channels that drain the mountains in and around the study area probably contributed minor amounts of recharge periodically. The Salt River was a source of recharge from the head of the valley—near the site of Granite Reef Dam—to a point about 10 miles downstream (Lee, 1905). A map of predevelopment water levels (Thomsen and Baldys , 1985) indicates that recharge from the Salt River may have reached the Gila River in the western third of the reservation. Water is discharged from the ground-water system by surface flow, underflow, and evapotranspiration. Discharge of ground water into the Gila and Salt Rivers occurred regularly prior to development but seldom occurs now. In the western third of the reservation, water was discharged from 18 — T" , 4/> •4-» C 0) E a> — i. •o 3 9) (/) r« ea fO r— 9> E • E •1- to c ■4-> X o •A k. o •<- 9> o C 4- o 1 ■*-> O «/» N a> i/> r— -O •»- 4- L. c« CT* O «£ — .a c 0) u (/) 0) «' a» rj _ c o <« .O (U <7> X 0) X o f— o» 9) ^ (/) (« a> ^ u c o •>- Z bO = 0S6I ^C6I :;udujdjnse3ui auQ 0961 9961 = 9^61 - 0^61 - 9961 o LO o CO o CNJ 0961 QNOOBS y3d i33d 31900 NI ‘39aVH0Sia Figure 5.--Average annual base flow of the Gila River near Gila Crossing. 19 the aquifer to the Gila River channel and adjacent bogs and sloughs (Lee, 1904). As water in the aquifer moved northwestward between Sierra Estrella and South Mountain, the reduced cross-sectional area of the aquifer forced part of the water to the surface. The predevelopment ground-water discharge to the Gila and Santa Cruz Rivers in the western third of the reservation is estimated on the basis of streamflow measurements to have averaged 33,000 acre-feet per year. Ground-water discharge near Coolidge probably averaged about 3,000 acre-feet per year. Ground water occurs mainly under water-table or unconfined conditions in the sedimentary material that underlies much of the Gila River Indian Reservation and the surrounding area. The water table is that surface in an unconfined water body at which the pressure is atmospheric. The water table is defined by the levels at which water stands in wells that penetrate the water body. Prior to development of the area by non- Indian settlers, water levels were 10 to 70 feet below the land surface and the water table was a surface of low relief, sloping in general with the grade of the river (Lee, 1904). The water level at Sacaton was 15 feet below the land surface. The direction of ground-water movement in 1900 was from the northeast, east, and southeast toward the northwest, and ground-water discharge occurred through the gap between South Mountain and Sierra Estrella (Thomsen and Baldys, 1985). The ground-water reservoir was in equilibrium and was sustained mainly by the infiltration of water from the Gila River. Ground water was discharged to the Gila River in the western part of the reseirvation. The movement of water from the Salt River to the ground-water reservoir and back to the Salt River affected a small part of the study area but had little effect on the overall ground-water system. Mountain ranges that lie within the study area impede the movement of ground water. The rocks that form the mountains are generally not water bearing but may, where fractured, yield as much as a few tens of gallons per minute of water to wells. The stream alluvium and the upper basin fill yield as much as 4,000 gallons per minute of water to wells (Laney and Hahn, 1986). In places, the upper basin fill contains layers of fine-grained material that restrict the downward migration of water. The lower basin fill generally yields less than 50 gallons per minute of water to wells. The pre-Basin and Range unit is more permeable than the lower basin fill. Water in the pre-Basin and Range unit is under confined conditions in areas where the unit is overlain by the lower basin fill. Elsewhere, water in the upper basin fill and the pre-Basin and Range unit is in a common and generally unconfined water body. Underflow Underflow through permeable materials underlying the surface drainages recharges the ground-water system in the study area. Underflow is the subsurface movement of water from one basin to another. The Gila and Salt Rivers and Queen Creek enter from areas underlain by crystalline rocks; hence, the underflow from these drainages probably was negligible. The Santa Cruz River and Santa Rosa, Greene, and Vekol Wash drainages are underlain by thick alluvium, so the potential exists for a moderate amount of underflow through the alluvium. Underflow along the Santa Cruz, Santa 20 Rosa, and Vekol drainages is indicated by predevelopraent water levels (Thomsen and Baldys, 1985). Underflow from the Santa Cruz River drainage probably was about 25,000 acre-feet per year (Turner and others, 1943; Hardt and Cattany, 1965). Underflow from Santa Rosa and Vekol Washes was about 1,500 and 500 acre-feet per year, respectively (Turner and others, 1943). Underflow derived principally from runoff in the Picacho Mountains may range from 2,000 to 3,000 acre-feet per year. Total underflow into the area probably is about 30,000 acre-feet per year. Underflow from the area occurs mainly through the gap between South Mountain and Sierra Estrella. The quantity of underflow may have been as much as 10,000 acre-feet per year on the basis of transmissivity data and estimates of the cross-sectional area. The alluvial material is about 1,000 feet thick at the narrowest part of the gap (H.J. Thiele, consultant, written commun., 1958). The deepest wells in the area are about 700 feet deep. Hydraulic Characteristics of the Ground-Water Reservoir The hydraulic characteristics of the ground-water reservoir are the physical properties that control the ability of the material to store and transmit water. These properties depend mainly on the size of openings or interstices and their shape, arrangement, and interconnection. The hydraulic characteristics commonly used to describe ground-water reservoirs are storage coefficient and transmissivity, which provide a measure of the amount of water stored in the reservoir and the rate at which the reservoir will transmit water. The movement of ground water through a section of aquifer can be expressed by the equation: Q » TIW, (1) where Q - flow, in cubic feet per day; T - transmissivity, in feet squared per day; I = hydraulic gradient (dimensionless); and W - width of section, in feet. Transmissivity is a function of the hydraulic conductivity and saturated thickness of the reservoir and may be expressed by the equation: T - KM, (2) where K - hydraulic conductivity, in feet per day, and M = saturated thickness, in feet. 21 Hydraulic conductivity is the volume of water at the existing kinematic viscosity that will move in unit time under a unit of hydraulic gradient through a unit area measured at right angles to the direction of flow (Lohman and others, 1972). Hydraulic conductivity is expressed in units of length per unit time, such as feet per day. Transmissivity is the rate at which water at the existing kinematic viscosity is transmitted through a unit width of aquifer (saturated thickness of ground-water reservoir) under a unit hydraulic gradient. Transmissivity is expressed in consistent units of volume (L^) per unit time (T) per unit width (L), which reduces to L^T ^. In the English system, transmissivity is expressed in cubic feet per day per foot, which reduces to feet squared per day. The storage coefficient is the volume of water an aquifer releases from or takes into storage per unit surface area of the aquifer per unit change in head (Lohman and others, 1972). In an unconfined water body, it is virtually equal to the "specific yield," which is the volume of water that water-bearing material will yield by gravity drainage. The storage coefficient is expressed as a ratio of the volume of water removed to the volume of the material dewatered and is, therefore, dimensionless. Quantitative data on the hydraulic characteristics of ground- water reservoirs are obtained from field data on water levels, water-level fluctuations, and natural or artificial discharges (Ferris and others, 1962; Bentall, 1963). In general, transmissivity values in the study area range from about 1,500 to 30,000 feet squared per day (Anderson, 1968), but values of about 100,000 feet squared per day have been determined for some local areas. Anderson (1968) defined a regional pattern of average transmissivity values for 1923-64. Many of these values were obtained after the upper part of the aquifer, which in many areas is the most transmissive, had been at least partly dewatered; hence, the values are less than they would have been prior to dewatering. The greatest values of transmissivity are in the area from Mesa southward through Chandler westward toward Gila Crossing and in an area south of Maricopa (Anderson, 1968). Stratification in alluvial material causes transmissivity values to be much greater parallel to the bedding plane than perpendicular to the bedding plane. The average storage coefficients for central Arizona range from 0.15 to 0.20 (Anderson, 1968, p. 10). In computing the amount of recoverable ground water in storage beneath the Gila River Indian Reservation, Babcock (1970) used a storage coefficient of 0.15 for the upper 600 feet of sediments and 0.10 for the 600- to 1,000-foot interval. Anderson (1968) used a storage coefficient of 0.19 for most of the water¬ bearing zones in central Arizona but reduced the value to 0.15 in the area east of Mesa along the southwest slopes of the Superstition Mountains and to 0.10 in the area west of Casa Grande extending from the Sacaton Mountains on the north to the Table Top Mountains on the south. Storage coefficients, or more appropriately "specific yields," are relatively low in water-bearing zones that contain a large proportion of fine-grained material. The horizontal hydraulic gradient averaged about 0.001 and ranged from 0.0006 to 0.008 prior to ground-water withdrawals by the non-Indian 22 settlers. Horizontal hydraulic gradients of 0.02 are common in heavily pumped areas today. EvapoCransDiration Evapotranspiration is defined as "water withdrawn from a land area by evaporation from water surfaces and moist soil and plant transpiration” (Langbeln and Iseri, 1960). Evaporation is commonly measured by noting the change in water level in an open pan during a given time period. Such measurements, however, do not accurately reflect evaporation from natural water bodies because of differences in water temperature, vapor pressure, and water - surface roughness. The rate of evaporation from a small pan generally far exceeds that from a large reservoir or lake. The ratio of lake evaporation to pan evaporation is referred to as the pan coefficient. Annual evaporation from a U.S. Weather Bureau Class A pan at Mesa during 1963-73 averaged 106.31 Inches (Sellers and Hill, 1974). In the study area, the pan coefficient is about 0.67 and the average annual lake evaporation is about 70 to 75 inches (U.S. Department of Commerce, 1968). Annual lake evaporation, in feet, multiplied by an area of water surface, in acres, is equal to the volume of water evaporated, in acre-feet per year. Plants obtain water from precipitation and soil moisture, and deep-rooted plants called phreatophytes obtain much of their water from the capillary fringe and the saturated zone. The rate of transpiration by phreatophytes depends on the availability of water and on the species, the cover density and size, and the stage of maturity of the plants. The quantity of water withdrawn from the ground-water reservoir by phreatophytes depends on the depth to ground water. The use of water is greatest when ground water is shallow and decreases as depth to water increases (fig. 6). The relation between water use and depth to water is not well defined for all phreatophyte species but is fairly well defined for mesquite (Anderson, 1976). Little is known about the relation between ground-water use and the total water use by phreatophytes. The most common species of phreatophytes indigenous to southern Arizona are cottonwood, willow, baccharis (seepwillow), and mesquite (Gatewood and others, 1950). These species probably were the main woodland types of vegetation on the Gila River Indian Reservation prior to the arrival of the non-Indian settlers. In 1936, mesquite thickets covered about 22,500 acres, tamarisk 6,500 acres, cottonwood 5,000 acres, and scattered mesquite 20,000 acres for a total of about 54,000 acres of phreatophytes on the reservation (U.S. Soil Conservation Service, 1936). Rott (19 36) referred to the tamarisk as a recent invasion, having come in during the 6 to 10 years prior to 1936. Robinson (1965) stated that the date of introduction of tamarisk, better known as saltcedar, is not known, but it had spread enough to become noticeable in the 1920's. Turner (1974) noted that saltcedar did not occur in the upper Gila River Valley until after the flood of 1916 and that the increase in saltcedar has been almost wholly at the expense of baccharis. An investigation of the consumptive use of water by phreatophytes was made in 1963-71 to determine evapotranspiration before and after clearing phreatophytes on 15 miles of Gila River flood plain (Culler and 23 ANNUAL WATER USE BY MESQUITE, IN INCHES Figure 6.--Relation between depth to ground water and annual water use by mesquite (from Anderson, 1976, fig. 10, p. 46). others, 1982). Results of the study showed that the annual evapotranspira¬ tion averaged 3.7 feet and ranged from 4.7 feet for dense stands of phreatophytes to 2.1 feet on areas of no phreatophytes. Vegetation consisted mainly of saltcedar and mesquite with scattered cottonwood, seepwillow, seepweed, and arrowweed. Depth to ground water on the flood plain ranged from 5 feet near the river to 20 feet near the outer boundaries of the flood plain. Removal of phreatophytes resulted in a reduction in evapotranspiration that averaged 1.6 feet per year and ranged from 1.2 to 2.2 feet per year owing to the differences in the density of phreatophytes. Evapotranspiration after the removal of phreatophytes consisted of evaporation from bare ground and transpiration from annual vegetation. Because phreatophytes obtain their water supply primarily from ground water, the reduction in evapotranspiration that resulted from 24 removal of the phreatophytes is considered to represent a measure of ground-water withdrawal by phreatophytes. Precipitation and soil moisture provide a significant part of the water removed by evapotranspiration during the period of high potential evapotranspiration (Culler and others, 1982) . Evapotranspiration by phreatophytes was responsible for the withdrawal of large quantities of ground water along the Gila River when water levels were shallow. Evapotranspiration along the Gila River from near Coolidge to the Salt River probably ranged from 100,000 to 150,000 acre-feet per year (Turner and others, 1943). On the basis of the assumption that the number of acres of phreatophytes was about the same in 1870 as in 1936, the average evapotranspiration in 1870 was about 200,000 acre-feet per year. Of that amount, about 100,000 acre-feet per year was withdrawn from the ground- water reservoir by phreatophytes and 100,000 acre-feet per year was from precipitation and soil moisture. Total evapotranspiration was estimated by using the maximum evapotranspiration rate of 4.7 feet per year (Culler and others, 1982) for about 34,000 acres of cottonwood, tamarisk, and mesquite thickets (Rott, 1936) and the minimum evapotranspiration rate of 2.1 feet per year for 20,000 acres of scattered mesquite. The magnitude of the withdrawal from the ground-water reservoir by phreatophytes, 100,000 acre-feet, was estimated by using the the maximum reduction in evapotranspiration of 2.2 feet per year that resulted from the removal of phreatophytes on the upper Gila River (Culler and others, 1982) for the 34,000 acres of dense phreatophytes and the minimum reduction in evapotranspiration of 1.2 feet per year for 20,000 acres of scattered mesquite. The average annual precipitation of 8 inches for the 54,000 acres of phreatophytes accounts for 36,000 acre-feet per year of the total evapotranspiration. The remaining 64,000 acre-feet per year probably was from soil moisture that was replenished periodically by surface flow. WATER BUDGET Water budgets were prepared for the ground-water reservoir and for the study area as a whole. A water budget must account for all inflows, outflows, and changes in storage. Because aquifers were in equilibrium prior to development by non-Indian settlers, the average change in ground-water storage during the 100-year period prior to 1870 was considered zero. There was no surface storage. Hence, the sum of all inflows must have equaled the sum of all outflows. Most of the inflows and outflows for the study area as a whole, as well as for the ground-water reservoir, occurred along the Gila River. Thus, the components considered in the water budgets relate to the area along the Gila River and to the ground-water reservoir that underlies the study area (fig. 7). Precipita¬ tion on the area of evapotranspiration—the 54,000 acres used in computing evapotranspiration—was included in the water budget of the study area because precipitation provides a significant part of the evapotranspiration (Culler and others, 1982). Precipitation on the rest of the study area was not a significant factor in the water budget. One intermediate component of water movement appears in figure 7 but is not included in either water budget. This component, represents Gila River flow that entered the P = 36,000 ET - 200,000 25 356,000* Figure 7.--Conceptual longitudinal section showing initial estimates of average annual values, in acre-feet, of water-budget components. soil zone and subsequently was evaporated or used by plants. The value of Q was calculated as a residual of the evapotranspiration computation. s The average annual water budget for the ground-water reservoir is expressed by the equation G. + Q 1 G o + Q, + ET , ' I I o K> U.S. DEPARTMENT OF THE INTERIOR U.S. GEOLOGICAL SURVEY PREPARED IN COOPERATION WITH THE U.S. BUREAU OF INDIAN AFFAIRS WATER-RESOURCES INVESTIGATIONS REPORT 89-4174 PLATE 1 EXPLANATION BLOCKS—Each block includes 0.5 to 6 square miles of the aquifer and Is represented In the model by a node at the center of the block. The fwdel simulates the movement of water between blocks. Inflow and outflow processes simulated are: Constant flu*—Water entering the model area at a constant rate as underflow from adjacent basins or as recharge along mountain fronts Constant head—Water entering or leaving the model area at a sufficient rate to maintain a constant water level along the Salt River and at the north end of Sierra Estrella Evapotranspiratlon—Water leaving the water table by evaporation or by transpiration of plants • Streambed leakage—Hater entering or leaving the aquifer along the channels of the CHa and Santa Cruz Rivers BOUNDARY OF MATHEMATICAL MODEL—Simulated ground- water flow across boundary Is zero BOUNDARY Of CONSOLIDATED ROCK TOPOGRAPHIC CONTOUR INTERVAL 200 FEET national geodetic vertical datum of 1929 FINITE-DIFFERFNCE GRID AND BOUNDARY CONDITIONS USED IN THE MATHEMATICAL MODEL. GILA RIVER INDIAN RESERVATION rub! I’ L / h ' h ji j^Lj h ^ ijS ih /