Simulation of Ground-Water Flow in the Mojave River Basin, California

By CHRISTINA L. STAMOS, PETER MARTIN, TRACY NISHIKAWA, and BRETT F. COX

U.S. GEOLOGICAL SURVEY

Water-Resources Investigations Report 01-4002 Version 2

Prepared in cooperation with the

MOJAVE WATER AGENCY 7208-20

Sacramento, California 2001

U.S. DEPARTMENT OF THE INTERIOR GALE A. NORTON, Secretary

U.S. GEOLOGICAL SURVEY Charles G. Groat, Director

The use of firm, trade, and brand names in this report is for identification purposes only and does not constitute endorsement by the United States Government.

For additional information write to: Copies of this report can be purchased from:

District Chief U.S. Geological Survey U.S. Geological Survey Information Services Placer Hall, Suite 2012 Box 25286 6000 J Street Federal Center Sacramento, CA 95819-6129 Denver, CO 80225

CONTENTS Abstract...... 1 Introduction ...... 2 Purpose and Scope ...... 3 Description of Study Area ...... 3 Acknowledgments ...... 5 Surface-Water Hydrology ...... 6 The Mojave River ...... 6 Ungaged Tributary Streams ...... 15 Ground-Water Hydrology ...... 16 Geologic Setting ...... 17 Stratigraphic Units...... 19 Definition of Aquifers ...... 24 Effects of Faulting on Ground-Water Flow ...... 24 Ground-Water Recharge and Discharge ...... 28 Recharge ...... 28 The Mojave River ...... 28 Mountain-Front Recharge ...... 30 Artificial Recharge ...... 31 Irrigation-Return Flow ...... 31 Fish Hatchery Discharge ...... 32 Treated Sewage Effluent ...... 32 Septic Systems ...... 36 Imported Water ...... 38 Discharge ...... 38 Pumpage ...... 38 Evapotranspiration ...... 42 Transpiration by Phreatophytes and Hydrophytes ...... 42 Bare-Soil Evaporation from Dry Lakes ...... 44 Free-Surface Evaporation ...... 44 Underflow at Afton Canyon ...... 44 Ground-Water Flow Model ...... 44 Model Grid ...... 45 Model Boundary Conditions ...... 48 Aquifer Properties ...... 49 Transmissivity ...... 49 Storage Coefficient ...... 49 Vertical Leakance ...... 53 Stream-Aquifer Interactions ...... 53 Simulation of Recharge ...... 58 Mountain-Front Recharge ...... 58 Artificial Recharge ...... 59 Irrigation-Return Flow ...... 59 Fish Hatchery Discharge and Imported Water ...... 59 Treated Sewage Effluent ...... 59 Septic Systems ...... 60 Simulation of Discharge ...... 60 Pumpage ...... 60 Recreational Lakes in the Baja Model Subarea ...... 61

Contents III

Transpiration by Phreatophytes and Bare-Soil Evaporation ...... 62 Dry Lakes ...... 63 Model Calibration ...... 63 Simulation of Steady-State Conditions ...... 64 Simulation of Transient-State Conditions ...... 66 Simulation Results ...... 75 Steady-State Ground-Water Flow Directions and Travel Times ...... 86 Evaluation of Effects of Regional-Scale Pumping ...... 88 Upper Region Pumping Only...... 91 Lower Region Pumping Only ...... 92 Summary of Effects of Regional-Scale Pumping ...... 92 Model Validation ...... 93 Simulated Changes in Hydraulic Head, 1931–99 ...... 97 Model Limitations ...... 100 Evaluation of Selected Water-Management Alternatives ...... 101 Management Alternative 1: Zero Percent of Artificial Recharge Allocation ...... 103 Management Alternative 2: 50 Percent of Artificial Recharge Allocation ...... 107 Management Alternative 3: 100 Percent of Artificial Recharge Allocation ...... 107 Discussion of Management Alternatives 2 and 3 ...... 107 Summary ...... 111 Selected References ...... 113 Appendix 1. Measured and model-simulated hydraulic heads at selected wells in the Mojave River ground-water basin, southern California, 1931–99 ...... 117 Appendix 2. Measured and model-simulated hydraulic heads at multiple-well completion sites, Mojave River ground-water basin, southern California, 1992–99 ...... 125

FIGURES 1. Map showing location of study area and subareas of the Mojave River ground-water basin, southern California...... 4 2–5. Graphs showing 2. Total annual discharge from the headwaters of the Mojave River in the Mojave River ground-water basin, southern California, 1931–94 ...... 9 3. Flow-duration curves for daily mean discharge in the West Fork Mojave River (gaging stations 10260950 and 10261000), Deep Creek (gaging station 10260500), and the Mojave River at the Lower Narrows near Victorville (gaging station 10261500) in the Mojave River ground-water basin, southern California ...... 10 4. Daily mean discharge in the Mojave River drainage basin, southern California...... 11 5. Total annual base flow for the Lower Narrows on the Mojave River (data from Lines, 1996) and discharge for selected gages in the Mojave River ground-water basin, southern California ...... 12 6. Map showing location of channel-geometry and artificial-recharge sites in the Mojave River ground-water basin, southern California...... 14 7. Map showing generalized geology of the Mojave River ground-water basin, southern California ...... 18 8. Cross section showing conceptualization of the ground-water flow system and model layers near Victorville, California ...... 20 9. Cross section showing conceptualization of the ground-water flow system and model layers at various locations along the Mojave River in southern California...... 22 10. Graph showing altitude of measured water levels at three multiple-well monitoring sites in the Mojave River ground-water basin, southern California ...... 25 11. Map showing altitude of water levels and generalized direction of ground-water flow in the Mojave River ground-water basin, southern California, November 1992...... 26 12. Graph showing estimated annual recharge to the Mojave River floodplain aquifer within the Alto subarea for 1931–94, within the Transition zone and Centro subarea combined for 1931–94, and within the Baja subarea for 1931–32 and 1953–94, Mojave River ground-water basin, southern California ...... 31

IV Contents

13. Map showing distribution of septic and sewer systems in the Alto subarea, Mojave River ground-water basin, southern California, for selected years between 1930 and 1990...... 37 14. Graph showing total pumpage and sources of pumpage data for the Mojave River ground-water basin, southern California, 1931–99 ...... 39 15. Map showing distribution of annual total pumpage in the Mojave River ground-water basin, southern California, 1931, 1951, 1971, and 1994 ...... 40 16. Graphs showing components of total pumpage by subarea for the Mojave River ground-water basin, southern California, 1931–99 ...... 42 17. Graph showing altitude of measured water levels at selected wells in the Mojave River ground-water basin, southern California ...... 43 18–22. Maps showing 18. Location of model grid and model boundaries and location of horizontal-flow barrier, mountain-front recharge, drain, evapotranspiration, stream, and artificial recharge cells of the ground-water flow model of the Mojave River ground-water basin, southern California...... 46 19. Areal distribution of transmissivity in the ground-water flow model of the Mojave River ground-water basin, southern California ...... 50 20. Areal distribution of specific yield for model layer 1 of the ground-water flow model of the Mojave River ground-water basin, southern California ...... 52 21. Areal distribution of anisotropy for model layer 1 of the ground-water flow model of the Mojave River ground-water basin, southern California...... 54 22. Schematic of simulated streamflow-routing network for the Mojave River ground-water basin, southern California...... 56 23. Graph showing total pumpage by model subarea for the Mojave River ground-water basin, southern California, 1931–99 ...... 60 24. Map showing measured water levels and simulated hydraulic head for 1930 for model layer 1 of the Mojave River ground-water basin, southern California ...... 65 25. Graph showing measured water levels and simulated hydraulic head and the root mean square error (RMSE) for each model subarea of the Mojave River ground-water basin, southern California, for 1930 steady-state conditions...... 66 26. Graphs showing mean daily discharge and average discharge for 1954 in the ground-water flow model of the Mojave River ground-water basin, southern California ...... 72 27. Map showing measured water levels, autumn 1992, and simulated hydraulic-head contours, 1992, for model layer 1 of the ground-water flow model of the Mojave River ground-water basin, southern California ... 73 28. Graph showing measured water levels and simulated hydraulic head and the root mean square error (RMSE) for each model subarea of the Mojave River ground-water basin, southern California, for 1992 transient-state conditions...... 74 29. Map showing measured ground-water levels and simulated hydraulic head for selected wells in the Mojave River ground-water basin, southern California ...... 76 30–33. Graphs showing 30. Measured and simulated discharge in the Mojave River ground-water basin, southern California, from the Lower Narrows near Victorville (gaging station 10261500), 1931–99; the Mojave River at Barstow (gaging station 10262500), 1931–99; and the Mojave River at Afton Canyon (gaging station 10263000), 1931–99 (measured data are for years 1931, 1953–78, and 1981–99) ...... 78 31. Volumetric difference between measured and simulated discharge in the Mojave River ground-water basin, southern California, for the Lower Narrows near Victorville (gaging station 10261500), 1931–99; the Mojave River at Barstow (gaging station 10262500), 1931–99; and the Mojave River at Afton Canyon (gaging station 10263000), 1931–99 (measured data are for years 1931, 1953–78, 1981–99)...... 79 32. Ground-water recharge to, and discharge from, the model subareas of the Mojave River ground-water basin, southern California, 1930 and 1994 ...... 84 33. Average annual ground-water recharge to, and discharge from, the model subareas of the Mojave River ground-water basin, southern California, 1931–90...... 85 34. Map showing simulated underflow between model subareas of the Mojave River ground-water basin, southern California, for 1930 and 1994, and the average underflow for 1931–90...... 87 35. Map showing simulated flow paths of selected particles for steady-state (1930) conditions initially placed at mountain-front recharge cells, and location of streamflow recharge cells of the ground-water flow model of the Mojave River ground-water basin, southern California ...... 89

Contents V

36. Graph showing average streamflow recharge from the Mojave River, ground-water discharge to the Mojave River, evapotranspiration, and storage for the analysis of regional-scale pumping effects on the Mojave River ground-water basin, southern California, 1931–90 ...... 91 37. Map showing measured water levels, spring 1998, and simulated hydraulic-head contours, 1998, for model layer 1 of the ground-water flow model of the Mojave River ground-water basin, southern California...... 94 38. Graph showing measured water levels, simulated hydraulic head, and the root mean square error (RMSE) for each model subarea of the Mojave River ground-water basin, southern California, for 1998 transient-state conditions...... 95 39. Maps showing simulated changes in hydraulic head for model layer 1 of the ground-water flow model of the Mojave River ground-water basin, southern California ...... 98 40. Graph showing cumulative departure from mean streamflow measured at the headwaters of the Mojave River, southern California, 1931–99...... 100 41. Map showing location of Mojave Water Agency artificial-recharge sites in the Mojave River ground-water basin, southern California...... 102 42. Map showing change in simulated hydraulic head in layer 1 of the ground-water flow model of the Mojave River ground-water basin, southern California, for three management alternatives, 1999–2019...... 104

TABLES 1. Annual total of the mean daily discharge at active gaging stations on, or on tributaries to, the Mojave River, southern California, 1931–94...... 7 2. Estimated annual inflow from selected ephemeral tributary streams to the Mojave River, southern California, 1931–99 ...... 16 3. Estimates of annual recharge to, and discharge from, the Mojave River ground-water basin, southern California, for selected periods...... 29 4. Sources and quantity of artificial recharge along the Mojave River, southern California, 1938–99...... 33 5. Population and estimated recharge from septic systems in the Alto subarea of the Mojave River ground- water basin, southern California, 1930–90 ...... 36 6. Hydraulic characteristics of horizontal-flow barriers used in the model of the Mojave River ground-water basin, southern California...... 48 7. Streambed-conductance values and associated flow conditions for stress periods used in the streamflow- routing package in the model of the Mojave River ground-water basin, southern California...... 58 8. Annual water consumption of recreational lakes in the Baja subarea of the Mojave River ground-water basin, southern California...... 62 9. Stress period lengths and specified inflows from The Forks (Deep Creek and West Fork) to the Mojave River, southern California, 1931–99...... 67 10. Simulated hydrologic budgets for model subareas of the Mojave River ground-water basin, southern California, 1930 (steady state), 1994, and 1931–90 average (adjudication period) ...... 81 11. Simulated hydrologic budget for model subareas of the Mojave River ground-water basin, southern California, 1995–99 average values...... 96 12. Simulated hydrologic budgets for model subareas of the Mojave River ground-water basin, southern California, for management alternatives 1, 2, and 3, 1999–2019 average values...... 108

VI Contents

CONVERSION FACTORS, VERTICAL DATUM, ABBREVIATIONS, AND WELL-NUMBERING SYSTEM

Multiply By To obtain acre 0.4047 hectare acre-foot (acre-ft) 0.001233 cubic hectometer acre-foot per year (acre-ft/yr) 0.001233 cubic hectometer per year foot (ft) 0.3048 meter foot per year (ft/yr) 0.3048 meter per year square foot per day (ft2/d) 0.09290 square meter per day square foot per second (ft2/s) 0.09290 square meter per second cubic foot per second (ft3/s) 0.02832 cubic meter per second gallon per day (gal/d) 0.003785 cubic meter per day gallon per day per foot (gal/d/ft) 0.01491 cubic meter per day per meter gallon per minute (gal/min) 0.003785 cubic meter per minute inch (in.) 2.54 centimeter inch per year (in./yr) 25.4 centimeter per year mile (mi) 1.609 kilometer square mile (mi2) 2.590 square kilometer

Temperature in degrees Fahrenheit (°F) may be converted to degrees Celsius (°C) as follows:

°C = (°F - 32)/1.8.

Vertical Datum

Sea level: In this report “sea level” refers to the National Geodetic Vertical Datum of 1929—a geodetic datum derived from a gen- eral adjustment of the first-order level nets of the United States and Canada, formerly called Sea Level Datum of 1929.

Abbreviations

ft-1 per foot

CSTR Streambed conductance GHB General-head boundary GIS Geographic information system HFB Horizontal-flow barrier ME Mean error MODFLOW Three-dimensional, finite-difference ground-water flow model MWA Mojave Water Agency RASA Regional Aquifer-System Analysis (southern California) RMSE Root mean square error STR1 Streamflow routing package SWP California State Water Project USGS U.S. Geological Survey USMC U.S. Marine Corps VVWRA Victor Valley Wastewater Reclamation Authority

Contents VII

Well-Numbering System

Wells are identified and numbered according to their location in the rectangular system for the subdivision of public lands. Identification consists of the township number, north or south; the range number, east or west; and the section number. Each section is divided into sixteen 40-acre tracts lettered consecutively (except I and O), beginning with “A” in the northeast corner of the section and progressing in a sinusoidal manner to “R” in the southeast corner. Within the 40-acre tract, wells are sequentially numbered in the order they are inventoried. The final letter refers to the base line and meridian. In California, there are three base lines and meridians; Humboldt (H), Mount Diablo (M), and San Bernardino (S). All wells in the study area are referenced to the San Bernardino base line and meridian (S). Well numbers consist of 15 characters and follow the format 004N002W10G001S. In this report, well numbers are abbreviated and written 4N/2W-10G1. Wells in the same township and range are referred to only by their section designation,10G1.The following diagram shows how the number for well 4N/2W-10G1 is derived.

RANGE R4W R3W R2W R1W R1E SECTION 10 T7N R2W D CBA 6 543 2 1 T6N E FGH 7 891011 12 T5N M L K J

TOWNSHIP 18 17 16 15 14 13 T4N T4N 19 20 21 22 23 24 NPQ R T3N San Bernardino Meridian 30 29 28 27 26 25 4N/2W-10G1

31 32 33 34 35 36

Well-numbering diagram (Note: maps in this report use abbreviated well numbers such as "10G1")

VIII Contents

Simulation of Ground-Water Flow in the Mojave River Basin, California

By Christina L. Stamos, Peter Martin, Tracy Nishikawa, and Brett F. Cox

ABSTRACT simulated streamflow hydrographs matched wet stress period average flow rates and times of no flow at the The proximity of the Mojave River ground-water Barstow and Afton Canyon gages. basin to the highly urbanized Los Angeles region has Steady-state particle-tracking was used to esti- led to rapid growth in population and, consequently, to mate travel times for mountain-front and streamflow an increase in the demand for water. The Mojave River, recharge. The simulated travel times for mountain- the primary source of surface water for the region, nor- mally is dry—except for a small stretch of perennial front recharge to reach the area west of Victorville flow and periods of flow after intense storms. Thus, the were about 5,000 to 6,000 years; this result is in reason- region relies almost entirely on ground water to meet able agreement with published results. Steady-state its agricultural and municipal needs. Ground-water particle-tracking results for streamflow recharge indi- withdrawal since the late 1800’s has resulted in dis- cate that in most subareas along the river, the particles charge, primarily from pumping wells, that exceeds quickly leave and reenter the river. natural recharge. To better understand the relation The complaint that resulted in the adjudication of between the regional and the floodplain aquifer sys- the Mojave River ground-water basin alleged that the tems and to develop a management tool that could be cumulative water production upstream of the city of used to estimate the effects that future stresses may Barstow had overdrafted the ground-water basin. In have on the ground-water system, a numerical ground- order to ascertain the effect of pumping on ground- water flow model of the Mojave River ground-water water and surface-water relations along the Mojave basin was developed, in part, on the basis of a previ- River, two pumping simulations were compared with ously developed analog model. the 1931–90 transient-state simulation (base case). The The ground-water flow model has two horizontal first simulation assumed 1931–90 pumping in the layers; the top layer (layer 1) corresponds to the flood- upper region (Este, Oeste, Alto, and Transition zone plain aquifer and the bottom layer (layer 2) corre- model subareas) but with no pumping in the remainder sponds to the regional aquifer. There are 161 rows and of the basin, and the second assumed 1931–90 pump- 200 columns with a horizontal grid spacing of 2,000 by ing in the lower region (Centro, Harper Lake, Baja, 2,000 feet. Two stress periods (wet and dry) per year Coyote Lake, and Afton Canyon model subareas) but are used where the duration of each stress period is a with no pumping in remainder of the basin. function of the occurrence, quantity of discharge, and length of stormflow from the headwaters each year. A In the upper region, assuming pumping only in steady-state model provided initial conditions for the the upper region, there was no change in storage, transient-state simulation. The model was calibrated to recharge from the Mojave River, ground-water dis- transient-state conditions (1931–94) using a trial-and- charge to the Mojave River, or evapotranspiration when error approach. compared with the base case. In the lower region, The transient-state simulation results are in good assuming pumping only in the upper region, there was agreement with measured data. Under transient-state storage accretion, decreased recharge from the Mojave conditions, the simulated floodplain aquifer and River, increased ground-water discharge to the Mojave regional aquifer hydrographs matched the general River, and increased evapotranspiration when trends observed for the measured water levels. The compared with the base case.

Abstract 1

In the upper region, assuming pumping only in hydraulic heads were lower than those for 1931 contin- the lower region, there was storage accretion, ued to increase until the end of the simulation (1999). decreased recharge from the Mojave River, increased Along the Mojave River, hydraulic heads fluctuated in ground-water discharge to the Mojave River, and the floodplain aquifer in response to recharge during increased evapotranspiration when compared with the years with large inflows with little apparent effect on base case. In the lower region, assuming pumping only the simulated hydraulic heads in the regional aquifer. in the lower region, there was less storage depletion, Three water-management alternatives were eval- increased recharge from the Mojave River, increased uated to determine their effect on ground-water ground-water discharge to the Mojave River, and resources using the calibrated ground-water flow increased evapotranspiration when compared with the model. The water-management alternatives consider base case. Overall, pumping in the lower region does the artificial recharge of imported water allocated to the not negatively affect the upper region; however, pump- Mojave Water Agency (MWA): the first assumes that ing in the upper region negatively affects the lower zero percent of the MWA allocation is available (alter- region by decreasing recharge from the Mojave River. native 1), the second assumes that 50 percent of the Streamflow, pumpage, and water-level data from MWA allocation is available (alternative 2), and the calendar years 1995–99 were used to validate the cali- third assumes that 100 percent of the MWA allocation brated ground-water flow model, that is, to test that the is available (alternative 3). Each of the three water- ground-water flow model will duplicate measured data management alternatives were evaluated for a 20-year for a noncalibration period without modification of the drought. Streamflow conditions were simulated using model parameters. In general, the simulated results are the 20-year drought of 1945–64 with associated in good agreement with the measured data, and the calibrated stream parameters. simulated hydrographs for wells in the floodplain and Management alternative 1 results in a reduction regional aquifers follow the measured water-level in ground-water recharge from the Mojave River com- trends. Simulated streamflow data for the 1995–99 wet pared with average recharge for 1995–99; this reduc- and dry stress periods at the Lower Narrows, Barstow, tion is reflected in simulated hydraulic-head declines and Afton Canyon were compared with the measured between 1999 and 2019 of as much as 45 feet. Manage- data for average streamflow for the same periods; in ment alternatives 2 and 3 result in no change in general, the model reflects 1995–99 streamflow condi- recharge from the Mojave River for management alter- tions. The simulation results also indicate that the native 2 and a small increase for management alterna- streambed conductance values calibrated to the 1931– tive 3 when compared with recharge for management 94 conditions reasonably simulate the 1995–99 condi- alternative 1. The artificial recharge of imported water tions and therefore can be used for predictive purposes. causes increases in simulated hydraulic head for both To visualize the magnitude, spatial distribution, management alternatives at each of the artificial- and timing of water-level changes in the basin through recharge sites. Some of the increases are related to time, simulated hydraulic heads for 1932–99 were water that recharges into areas of low transmissivity compared with simulated hydraulic heads for 1931. which implies that the recharge operations may benefit Greater than average annual inflows to the Mojave from being distributed over a larger area. River from the headwaters during the late 1930’s and throughout much of the 1940’s resulted in simulated hydraulic heads that were higher than the 1931 hydrau- INTRODUCTION lic heads along the Mojave River in most model subar- eas. Parts of the Baja and Harper Lake model subareas The proximity of the Mojave River ground-water had declines in the simulated hydraulic head because of basin to the highly urbanized Los Angeles region has the increase in agricultural pumpage. By 1960, the sim- led to rapid growth in population and, consequently, an ulated hydraulic heads were lower than the simulated increase in the demand for water. The Mojave River, hydraulic heads for 1931 in all model subareas of the the primary source of surface water for the region, nor- floodplain and the regional aquifers because of mally is dry—except for a small stretch that has peren- pumpage. After 1960, the size and the magnitude of the nial flow and periods of flow after intense storms. As a areas of the regional aquifer for which simulated result, the region relies almost entirely on ground water

2 Simulation of Ground-Water Flow in the Mojave River Basin, California

to meet its agricultural and municipal needs. Ground- data; (4) the distribution of aquifer properties, (5) water water withdrawal since the late 1800’s has resulted in levels using available and new monitoring wells, discharge, primarily from pumping wells, that exceeds (6) direction of ground-water flow; and (7) the loca- natural recharge. To plan for anticipated water demands tions of geologic barriers that may influence ground- and for the effects of imported water on the basin, water supply. This report summarizes the geohydro- methods are needed to evaluate and project ground- logic conditions of the Mojave River ground-water water conditions that result from present and planned basin. It presents the geology and hydrology of the changes in the Mojave River ground-water basin. This basin, which was used as the basis of the ground-water study is part of a series of studies started in 1992 by the flow model, presents the development of the regional U.S. Geological Survey (USGS) as part of southern ground-water flow model, and summarizes the calibra- California Regional Aquifer-System Analysis (RASA) tion, results, validation, limitations and needed future program, in cooperation with the Mojave Water refinements of the model. The simulated effects of pro- Agency (MWA). posed management alternatives during a 20-year drought with regard to artificial recharge of imported surface water are also presented. Purpose and Scope

The purpose of this report is to document the Description of Study Area numerical ground-water flow model of the Mojave River ground-water basin. The model was developed to The study area is the Mojave River ground-water update the analog model developed by Hardt (1971), to basin which, for the most part, is within the Mojave gain a better understanding of the relations between the River surface-water drainage basin as defined by the regional and the floodplain aquifer systems with regard Mojave Water Agency (1996). The surface-water drain- to the movement of ground water between manage- age basin encompasses about 3,800 mi2. The ground- ment subareas, and to develop a management tool that water basin covers about 1,400 mi2, is about 80 mi could be used to estimate the effects that future stresses northeast of Los Angeles, California, and is part of the may have on the ground-water system, specifically, Mojave Desert region. The Mojave River ground-water artificial recharge of imported water. Measured stream- basin is bounded by the San Bernardino and San flow, pumpage, and water level data for 1931–94 were Gabriel Mountains to the south, extending to Afton used to calibrate the model. Measured data for Canyon to the northeast, and is bounded by the Lucerne 1995–99 were then used to validate the calibrated flow Valley to the east, and the Antelope Valley to the west model. This study updates a previous analysis of the (fig. 1). The Mojave River ground-water basin basin completed by the USGS in the late 1960’s for boundary was defined initially by the California which an analog model was used to simulate ground- Department of Water Resources (1967) and later water flow (Hardt, 1971). All data and results from the modified by Hardt (1971) and Stamos and Predmore current study are presented in calendar year to coincide (1995). Generally, the boundary coincides with the with the previously published work by Hardt (1971). contact between the nonwater-bearing consolidated The analog model developed by Hardt (1971) rocks and the unconsolidated deposits. did not quantify the Mojave River’s effects on the In 1990, the city of Barstow and the Southern ground-water system nor did it sufficiently define the California Water Company filed a complaint that sources of recharge and discharge to the basin. Addi- alleged that the cumulative ground-water production tional geohydrologic information collected in the basin upstream of the city of Barstow had overdrafted the during this and concurrent USGS studies (Stamos and Mojave River ground-water basin (Mojave Basin Area Predmore, 1995; Izbicki and others, 1995; Lines, 1996; Watermaster, 1996a). In 1993, more than 200 parties Lines and Bilhorn, 1996; Densmore and others, 1997), stipulated to a “Physical Solution”; the stated purposes have helped to determine (1) the relations between the of the solution are (1) to ensure that downstream pro- ground-water system and the Mojave River; (2) the ducers are not adversely affected by upstream use, component of recharge from ungaged runoff; (3) the (2) to raise money to purchase supplemental water for age and rates of ground-water flow using geochemical the area, and (3) to encourage local water conservation.

Introduction 3

117 30' 117 00' 116 30'

Co SEE INSET

Kern Co Kern MAP BELOW

San Bernardino

West Cronese Lake

395 Coyote Lake Manix f Wash 35 Harper U.S. Marine Corps Afton Lake Canyon 00' 58 Nebo Yermo Centro subarea Annex Annex 15 Calico BajaMts subarea Barstow Camp e Cady Hinkley Valley Mojave Valley Daggett Newberry Troy Iron Springs Mtn 40 Lake Newberry Mts

Helendale Mojave River Channel Shadow Mts Alto Mojave Kane Transition Wash zone 247

El Mirage Desert Lake Lower Narrows Oeste Adelanto d Upper Narrows Este subarea sub- Southern Apple California Victorville Lucerne area Valley 34 Logistics Lake Airport 00' Sheep Cr Alto subarea California 18 Lucerne Hesperia Rabbit Valley 247 Lake Aqueduct Sheep Cr

Antelope Valley c Deep Creek b a 18 San Gabriel Mts The Forks 15 West Fork Silverwood Mojave River Lake 0 20 MILES

Co 0 20 KILOMETERS San Bernardino Mts

Los Angeles Co

San Bernardino

116 30' 116 05' EXPLANATION 35 0 10 MILES Silver 25' c Lake Mojave River Basin – Gaging station and 0 10 KILOMETERS 20' Mojave Water Ground-water basin identifier – Baker Agency (Study area) a 10260500 management b 10260950 and 15' San area Surface-water 10261000 East 15 Francisco drainage basin West Cronese c 10261100 Cronese Lake Soda CALIFORNIA d 10261500 Lake Lake 10' Mojave Water Agency e 10262500 boundaries – Mojave f 10263000 05' Desert Management area Los Angeles Dry lake (playa) f Subarea Rabbit Baja Subarea 35 San Diego Lake Area northeast of Manix Wash 00'

Figure 1. Location of study area and subareas of the Mojave River ground-water basin, southern California.

4 Simulation of Ground-Water Flow in the Mojave River Basin, California

The trial court ordered all parties either to stipulate to the mean annual precipitation in the nearby San the Physical Solution, file an answer to the cross- Bernardino Mountains to the south, a major source of complaint, or suffer default. From 1993 to 1998, the streamflow in the Mojave River, was more than 40 in. maximum annual ground-water production (the rate of (National Oceanic and Atmospheric Administration, production free of any fees for the stipulating parties) 1994). decreased from 100 percent to 80 percent; any pump- Land use in the study area is primarily agricul- age in excess of the maximum annual production was tural and residential; most residential development in assessed a fee. the Alto subarea occurred in the 1980’s (Umari and oth- For management purposes, MWA subdivided the ers, 1995, p. 4). Since the 1980’s, the population in this Mojave River surface-water drainage basin unit into subarea has increased from 44,230 in 1980 to 145,700 several subareas —Oeste, Alto, Transition zone of the in 1990 as growth in the Los Angeles area spread into Alto (hereinafter referred to as the Transition zone), the high desert (California Department of Finance, Este, Centro, and Baja (fig. 1). The Oeste subarea accessed November 28, 1998). Agriculture is concen- includes the Sheep Creek watershed because it is trated primarily along the Mojave River, near Harper within the MWA’s management area. The study area Lake (dry), and in the Mojave Valley (fig. 1). encompasses most of the subareas, except for the Este Ground water from wells is the sole source of subarea, of which only the southwestern part is water for public supply in the basin. In the upper part included. The eastern part of the Baja subarea is not part of the MWA’s management area but is included in of the basin (Alto subarea, fig. 1), water is pumped pri- this study because it is within the ground-water basin. marily for municipal, industrial, and agricultural uses. The California Department of Water Resources (1967) Ground-water withdrawal from private domestic wells referred to the Este, Alto (including the Transition constitutes only a small percentage of the total amount zone), and Oeste subareas as the upper Mojave Basin; of water withdrawn in the area. In the middle and lower the Centro subarea as the middle Mojave Basin; and the parts of the basin (Centro and Baja subareas, respec- Baja subarea as the lower Mojave Basin. tively), ground water is pumped primarily for The Mojave River is the principal source of agricultural irrigation. recharge to the basin; recharge occurs during sporadic stormflows. Generally, the river is dry, except for a small stretch of naturally occurring perennial flow Acknowledgments upstream of the Upper Narrows to the Lower Narrows Personnel from the Mojave Water Agency, the (fig. 1). (It should be noted that this reach ceased to flow for 3 days in September 1995, [Rockwell and oth- U.S. Marine Corps, the Victor Valley Water ers, 1999, p. 94]). The river is formed by two tributaries Reclamation Authority, the city of Barstow, Jess at the northern base of the San Bernardino Mountains Ranch, the Mojave River Fish Hatchery, James C. at an elevation of about 3,000 ft above sea level. The Hanson Engineering, and local municipalities and river bisects the study area and, when surface water is water companies provided hydrologic data and assis- present, flows northward through Victorville then east- tance in completing this study and are gratefully ward through Barstow. Any surface flow that does not acknowledged. The authors also thank the many private seep into the ground-water basin exits from Afton and public well owners that provided information and Canyon, which is at an elevation of about 1,400 ft allowed access to their wells for data collection. Appre- above sea level and about 100 mi downstream from the ciation is expressed to Steven K. Predmore, U.S. Geo- headwaters of the river. The study area contains five dry logical Survey, San Diego, California, whose lakes or playas—Rabbit, El Mirage, Harper, Coyote, innovative skills and expertise in Geographic Informa- and Troy Lakes. tion Systems made creation and manipulation of the The climate of the basin is typical of arid regions pumpage data base and visualization of the numerical of southern California; it is characterized by low pre- model input data possible. This report was completed cipitation, low humidity, and high summer tempera- with the help and suggestions of scientific illustrators tures. Between 1960 and 1991, the mean annual Anna C. Borlin, Larry G. Schneider, and Rudolph R. precipitation in most of the basin was less than 6 in. (Phil) Contreras, and editor Myrna L. DeBortoli with (James, 1992). Between water years 1931 and 1994, the U.S. Geological Survey, San Diego, California.

Introduction 5

SURFACE-WATER HYDROLOGY contribution of flow from these tributary streams has never been gaged or measured directly. The Mojave River ground-water basin is an allu- The Mojave River is formed by the confluence of vial plain sloping gently northward and eastward. The two smaller streams, West Fork Mojave River and Deep plain consists of valleys and closed basins separated by Creek, at a location known as The Forks (fig. 1). These hills and low mountains. The Mojave River is the prin- streams originate in the San Bernardino Mountains, cipal stream traversing the basin and is the main source where peaks reach elevations of 8,535 ft above sea of recharge to the underlying aquifers. Alluvial mate- level, and they join at The Forks, which is at an altitude rial beneath the floodplain of the Mojave River consti- of about 3,000 ft above sea level. Generally, the pres- tutes the most productive aquifer in the basin and yields ence of streamflow in the river results from storm run- most of the ground water pumped from the basin. off in the nearby mountains. From The Forks, the About 80 percent of the total ground-water recharge is Mojave River flows northward through Victorville, believed to be from leakage of floodflows in the Mojave then generally north and northeastward through River along its 100-mi reach between the San Bernar- Barstow, and finally eastward through Afton Canyon, dino Mountains and Afton Canyon. Some recharge is which is at an elevation of about 1,400 ft above sea contributed by small tributary streams along the San level. The river leaves the Mojave River ground-water Bernardino Mountain front. The presence of a large, basin through Afton Canyon, about 100 mi downstream ephemeral river makes the hydrology of this basin from The Forks. After it emerges from Afton Canyon, unique from other major ground-water basins in the the Mojave River splits into separate channels that ter- southern California desert. minate at Soda and East Cronese Lakes (fig. 1), which are dry lakes, except after major storms. Presently, streamflow along the West Fork The Mojave River Mojave River, Deep Creek, and the Mojave River is monitored by the USGS at six gaging stations (fig. 1, The major source of surface water in the basin is table 1). The streamflow records from these gages were the Mojave River, but it is unpredictable and unreliable used to estimate inflow, outflow, recharge, and base for direct water supply for most agricultural and munic- flow along the river. The progressive loss of water ipal uses because most of the river’s 100 mi of downstream is the result of recharge to the ground- streambed generally are dry. Historically, many reaches water system, phreatophyte use, and surface evapora- of the river had perennial flow; these reaches and their tion. Other gaging stations were operated along the locations are discussed in detail by Lines (1996, p. 31). Mojave River in the past; Lines (1996, p. 4) gives com- plete descriptions and histories of these gaging However, as pumping increased for agricultural pur- stations. poses, reaches that previously had perennial flow Inflow to the Mojave River at its headwaters can ceased to flow most of the year, and then flowed only in be estimated by combining the streamflow records response to storm runoff. By the mid 1900’s, only three from the gages on West Fork Mojave River (gaging reaches still had naturally occurring perennial flow. stations 10260950 and 10261000) and Deep Creek near Perennial flow now occurs only in the Mojave River Hesperia (10260500). Streamflow in West Fork Mojave upstream of the Upper Narrows to a short distance River (hereinafter referred to as West Fork) has been downstream from the Lower Narrows (fig. 1). recorded at two gaging stations about 0.6 mi apart. Natural, continuous surface-water flow along Gaging station 10261000 (West Fork Mojave River most of the river primarily occurs only when winter near Hesperia), about 0.5 mi above the confluence with storms produce runoff from the mountains, as shown Deep Creek, was operated from 1931 to 1971; gaging by isotopic data from Izbicki and others (1995). Flow station 10260950 (West Fork Mojave River above occurs along the entire reach of the river only during Mojave River Forks Reservoir, near Hesperia) has been episodes of floodflow. Runoff that enters the river in operation since 1975. The drainage areas of these through ephemeral tributary streams contributes to the two gage sites differ by about 5 mi2, but streamflow is surface water in the river during flooding. The considered equivalent for these gages (Lines, 1996,

6 Simulation of Ground-Water Flow in the Mojave River Basin, California

Table 1. Annual total of the mean daily discharge at active gaging stations on, or on tributaries to, the Mojave River, southern California, 1931–94 [Letters in parentheses correspond to location in figure 1. Discharge in acre-feet]

West Fork Combined discharge Mojave River at Lower Deep Creek Mojave River Mojave River Mojave River to Mojave River at The Narrows Calendar near Hesperia near Hesperia1 at Barstow at Afton Forks (Deep Creek and near Victorville year (10260500) (10261000 and (10262500) (10263000) West Fork) (10261500) (a) 10260950) (e) (f) (c) (d) (b) 1931 14,630 5,080 19,710 22,410 0 1,270 1932 64,390 32,560 96,950 84,340 37,480 (2) 1933 15,810 8,280 24,090 23,810 0 (2) 1934 14,730 4,970 19,700 23,590 0 (2) 1935 35,220 16,760 51,980 33,370 1,180 (2) 1936 21,020 7,790 28,810 21,280 0 (2) 1937 109,900 55,150 165,050 150,200 103,900 (2) 1938 144,900 79,240 224,140 189,300 138,100 (2) 1939 27,740 7,840 35,580 29,920 550 (2) 1940 30,630 8,460 39,090 28,010 0 (2) 1941 98,370 59,010 157,380 143,000 96,000 (2) 1942 15,310 5,620 20,930 24,600 100 (2) 1943 95,980 59,030 155,010 128,700 90,970 (2) 1944 50,390 40,990 91,380 76,770 36,250 (2) 1945 51,800 23,010 74,810 56,820 22,270 (2) 1946 44,010 27,890 71,900 51,570 14,570 (2) 1947 11,700 7,140 18,840 26,870 702 (2) 1948 10,210 3,120 13,330 25,250 0 (2) 1949 16,540 8,520 25,060 22,290 0 (2) 1950 7,580 2,640 10,220 21,130 0 (2) 1951 7,410 1,180 8,590 21,220 0 (2) 1952 55,010 42,970 97,980 66,780 12,550 (2) 1953 5,550 1,800 7,350 21,880 0 990 1954 38,660 17,080 55,740 31,800 0 930 1955 11,820 4,780 16,600 21,790 0 900 1956 14,010 2,110 16,120 21,440 0 900 1957 27,640 4,790 32,430 20,660 0 730 1958 94,390 44,400 138,790 97,640 20,070 2,770 1959 14,040 4,700 18,740 21,020 0 600 1960 9,270 230 9,500 18,730 0 720 1961 7,510 580 8,090 20,000 0 610 1962 46,770 15,810 62,580 24,340 730 660 1963 6,290 90 6,380 18,340 0 770 1964 9,800 730 10,530 15,560 0 500 See foonotes at end of table.

Surface-Water Hydrology 7

Table 1. Annual total of the mean daily discharge at active gaging stations on, or on tributaries to, the Mojave River, southern California, 1931–94—Continued

West Fork Combined discharge Mojave River at Lower Deep Creek Mojave River Mojave River Mojave River to Mojave River at The Narrows Calendar near Hesperia near Hesperia1 at Barstow at Afton Forks (Deep Creek and near Victorville year (10260500) (10261000 and (10262500) (10263000) West Fork) (10261500) (a) 10260950) (e) (f) (c) (d) (b) 1965 75,090 30,450 105,540 49,130 6,360 4,460 1966 55,850 18,860 74,710 40,240 7,160 1,700 1967 51,440 40,610 92,050 54,650 530 700 1968 13,520 4,790 18,310 17,520 0 210 1969 219,300 123,800 343,100 294,300 146,600 72,870 1970 15,800 5,630 21,430 21,250 0 480 1971 26,780 3,390 30,170 27,240 40 380 1972 7,050 (3) 11,670 15,530 0 590 1973 40,220 (3) 60,620 34,630 150 280 1974 18,480 (3) 27,210 17,020 0 390 1975 11,420 4,600 16,020 15,810 0 130 1976 18,070 6,200 24,270 25,850 0 380 1977 14,350 3,530 17,880 24,830 0 830 1978 231,400 133,200 364,600 209,600 50,460 (2) 1979 77,370 27,880 105,250 73,190 5,560 (2) 1980 194,100 113,600 307,700 229,300 137,700 (2) 1981 10,220 4,130 14,350 21,390 0 1,330 1982 51,550 20,160 71,710 37,360 0 970 1983 150,700 117,100 267,800 190,800 92,990 13,300 1984 11,470 3,860 15,330 24,300 40 1,810 1985 15,750 6,170 21,920 19,760 0 640 1986 30,590 12,590 43,180 15,750 0 550 1987 11,350 1,300 12,650 15,540 0 600 1988 10,950 4,350 15,300 15,070 0 830 1989 7,040 3,270 10,310 10,340 0 510 1990 6,230 1,370 7,600 8,420 0 440 1991 31,880 6,700 38,580 10,860 0 720 1992 51,510 34,550 86,060 25,760 30 850 1993 295,000 133,700 428,700 285,500 122,800 66,490 1994 20,470 6,020 26,490 9,390 0 490

1 Gaging station 10261000 was operated from October 1929 to September 1971. Gaging station 10260950 has been operated since October 1975. 2 Gaging station 10263000 was not operational between 1932–52 and 1978–80. 3 Inflow for 1972–74 was based on inflow at gaging station 102621100, Mojave River below Mojave River Forks Reservoir.

8 Simulation of Ground-Water Flow in the Mojave River Basin, California

p. 6). Between 1971 and 1974, there was no gage on to storm runoff and releases from the dam at Silver- West Fork Mojave River; estimated inflow to the wood Lake. This lake, formed by the construction of Mojave River was based on gage 10261100 (Mojave Cedar Springs Dam in 1971, is several miles upstream River below Mojave River Forks Reservoir), about 0.8 on West Fork and is used primarily for storage of mi downstream from The Forks. The total annual dis- imported water from the California Aqueduct as part of charge at the headwaters of the Mojave River is sum- the California State Water Project (SWP). Total annual marized in table 1 and shown in figure 2. inflows for West Fork (gaging station 10260950) Annual inflow from the headwaters averaged (fig. 2) include all releases from Cedar Springs Dam. about 70,000 acre-ft for 1931–94; however, because of The construction of this dam has not decreased the climatic conditions and river-channel characteristics, duration of flow in West Fork (Lines, 1996, p. 9). streamflow available for recharge can vary widely. Flow-duration curves are useful for predicting Extremes for 1931–94 for the combined inflows of the distribution of future flows for water supply and West Fork Mojave River and Deep Creek to the Mojave River ranged from about 6,380 acre-ft in 1963 to about hydrologic analysis and for demonstrating the hydro- 428,700 acre-ft in 1993, the wettest year of this period logic characteristics of the drainage area. A flow- (table 1). Inflow from the headwaters occurs primarily duration curve, or cumulative frequency curve, indi- during December through March. Most inflow to the cates the percentage of time that specified discharges river is from Deep Creek (fig. 2). The remainder of are equaled or exceeded in a given period (Searcy, inflow is from West Fork, which flows only in response 1959, p. 1). Flow-duration curves for flow, or

450,000

400,000 West Fork 350,000 Deep Creek

300,000

250,000

200,000

150,000 DISCHARGE, IN ACRE-FEET Average discharge 100,000

50,000

0 1931 1934 1937 1940 1943 1946 1949 1952 1955 1958 1961 1964 1967 1970 1973 1976 1979 1982 1985 1988 1991 1994 YEAR

Figure 2. Total annual discharge from the headwaters of the Mojave River in the Mojave River ground-water basin, southern California, 1931–94. Values are based on the combined annual flow at Deep Creek (gaging station 10260500) and West Fork Mojave River near Hesperia (gaging station 10261000) for 1931–71, on flow at Mojave River below Mojave River Forks Reservoir (gaging station 10261100) for 1972–74, and on combined flow at Deep Creek (gaging station 10260500) and at West Fork Mojave River above Mojave River Forks, near Hesperia (gaging station 10260950) for 1975–94.

Surface-Water Hydrology 9

discharge, measured at the West Fork and Deep Creek no contribution from ground water. Flow in West Fork gaging stations are shown in figure 3. The curve for also results from releases from Silverwood Lake. West Fork represents flow at gaging stations 10260950 Figure 3 indicates that the discharge at the gage in West and 10261000. The steep slope of the curve for West Fork equals or exceeds 0.1 ft3/s only about 40 percent Fork is typical of highly variable, ephemeral streams of the time. In comparison, the slope of the curve for with flows that are mainly from storm runoff Deep Creek, which is a perennial stream, is much (Searcy, 1959, p. 22) and with flows that have little or flatter, and discharge equals or exceeds 0.1 ft3/s more

100,000

10,000

1,000

Deep Creek 100

Lower Narrows

10 DISCHARGE, IN CUBIC FEET PER SECOND

1

West Fork 0.1

0.01 0.010.1 1 2 5 10 20 30 40 50 60 70 80 90 95 98 99 99.9 99.99 PERCENTAGE OF TIME DISCHARGE IS EQUALED OR EXCEEDED

Figure 3. Flow-duration curves for daily mean discharge in the West Fork Mojave River (gaging stations 10260950 and 10261000), Deep Creek (gaging station 10260500), and the Mojave River at the Lower Narrows near Victorville (gaging station 10261500) in the Mojave River ground-water basin, southern California.

10 Simulation of Ground-Water Flow in the Mojave River Basin, California

than 99.99 percent of the time. Over the entire period reach of naturally occurring perennial flow. In of record (1931–94), Deep Creek has ceased to flow September 1995, however, the river ceased to flow for only 2 days (July 17 and 18, 1961) (Rockwell and 3 days (Rockwell and others, 1999, p. 94). Chow others, 1999, p. 85). (1964) defines base flow as sustained or fair-weather Any water present in the streambed at The Forks runoff composed of ground-water discharge and travels only a few miles downstream before infiltrating delayed subsurface runoff. The perennial nature of the into the sandy streambed and recharging the ground- Mojave River near Victorville is apparent on the water system. Discharge from two fish hatcheries about hydrograph for the gaging station (fig. 4B). The aver- 9 mi downstream from The Forks also contributes flow age annual flow at the Lower Narrows near Victorville to this reach of the river for a short distance before it (gaging station 10261500) for 1931–94 is about rapidly percolates and disappears into the streambed. 54,000 acre-ft. Flow, or discharge, in this part of the At Victorville, shallow bedrock between the river is a combination of storm runoff and base flow Upper and Lower Narrows causes ground water to dis- (fig. 5A, B). Lines (1996) estimated the base flow for charge as base flow to the river channel, creating a the period 1931–94 in the Mojave River near

100,000 AC 10,000

1,000

100

10

1

0.1 1931 35 40 45 50 55 60 65 70 75 80 85 90 95 1931 35 40 45 50 55 60 65 70 75 80 85 90 1995 YEAR YEAR 100,000 BDNO 10,000 RECORD

DISCHARGE, IN CUBIC FEET PER SECOND 1,000

100 NO RECORD

10

1

0.1 1931 35 40 45 50 55 60 65 70 75 80 85 90 95 1931 35 40 45 50 55 60 65 70 75 80 85 90 1995 YEAR YEAR Figure 4. Daily mean discharge in the Mojave River drainage basin, southern California. A, Mojave River at The Forks (gaging stations 10260500, 10260950, and 10261000), 1931–94. B, Lower Narrows near Victorville (gaging station 10261500), 1931–94. C, Mojave River at Barstow (gaging station 10262500), 1931–94. D, Mojave River at Afton Canyon (gaging station 10263000), 1931–32, 1953–78, and 1981–94.

Surface-Water Hydrology 11

300,000 A Lower Narrows 250,000

200,000

150,000

100,000 Average discharge

DISCHARGE, IN ACRE-FEET 50,000

0 1931 34 37 40 43 46 49 52 55 58 61 64 67 70 73 76 79 82 85 88 91 1994 30,000 B Lower Narrows 25,000

20,000

15,000

10,000

5,000

0 1931 34 37 40 43 46 49 52 55 58 61 64 67 70 73 76 79 82 85 88 91 1994

150,000 C Barstow 100,000

50,000

0 1931 34 37 40 43 46 49 52 55 58 61 64 67 70 73 76 79 82 85 88 91 1994

100,000 Afton Canyon DISCHARGE, IN ACRE-FEETNO BASE FLOW, IN ACRE-FEET RECORD 50,000 NO D RECORD 0 1931 34 37 40 43 46 49 52 55 58 61 64 67 70 73 76 79 82 85 88 91 1994 YEAR Figure 5. Total annual base flow for the Lower Narrows on the Mojave River (data from Lines, 1996) and discharge for selected gages in the Mojave River ground-water basin, southern California. A, Discharge from the Mojave River at the Lower Narrows near Victorville (gaging station 10261500), 1931–94. B, Base flow in the Mojave River at the Lower Narrows, 1931–94. C, Discharge in the Mojave River at Barstow (gaging station 10262500), 1931–94. D, Discharge in the Mojave River at Afton Canyon (gaging station 10263000), 1931–32, 1953–78, and 1981–94.

12 Simulation of Ground-Water Flow in the Mojave River Basin, California

Victorville at the Lower Narrows. During this period, Narrows exceeded discharge at The Forks for most base flow averaged about 37 percent of the total flow days during each year and that the mean daily dis- (Lines, 1996, p. 29), but varied depending on the charge at the Lower Narrows exceeded mean daily dis- amount of storm runoff in the river. During years of low charge at The Forks for 23 of the 64 years between storm runoff, such as 1990, flow is predominantly base 1931 and 1994 (table 1). Daily discharge at The Forks flow. Comparison of the base-flow estimates with those exceeds the discharge at the Lower Narrows only dur- from previous studies would be misleading because the ing periods when storm runoff is concentrated in short previous studies included other surface-water sources pulses of floodflow (fig. 4A,B). and delayed subsurface runoff in their calculation of Streamflow leaving the Lower Narrows enters base flow (Lines, 1996). the Transition zone (fig. 1) and quickly infiltrates the Lines (1996, p. 29) reported that during water sreambed within a few miles of the Lower Narrows years 1900–01 and 1904–05, years when there was a gage owing to the absence of shallow bedrock in the gaging station at the Upper Narrows, annual base flow area. The streambed usually is dry for 4 mi; at this averaged about 30,000 acre-ft. Although average point, wastewater is discharged by the Victor Valley annual base flows are highly variable, they have Wastewater Reclamation Authority (VVWRA), which steadily declined since the 1950’s and early 1960’s. causes the river to flow again. This flow completely This decline was temporarily reversed in the late infiltrates the streambed about 4 mi downstream from 1960’s and the late 1970’s as a result of large inflows the discharge point (Lines, 1996, p.14). The remainder from the headwaters during water years 1969, 1978, of the streambed in the Transition zone and the Centro 1980, and 1983 (Lines, 1996, p. 29). The estimated subarea is normally dry. base flow reached an all-time low of about 4,000 acre- The gage on the Mojave River at Barstow (gag- ft in water year 1992, but increased to 11,000 acre-ft in ing station 10262500) flows in response to storm run- water years 1993 and 1994 following large inflows off. The ephemeral nature of the river at Barstow is from the headwaters (Lines, 1996, p. 29). apparent on the hydrograph for this gage (fig. 4C). The Note that Lines (1996) included other surface- average annual flow past the gage for 1931–94 was water sources, such as return flows from the fish hatch- about 18,330 acre-ft; however, average flow was about ery operated by California Department of Fish and 41,890 acre-ft for the 28 years when discharge reached Game (Mojave River Fish Hatchery) and the private the gage at Barstow (fig. 5C). Jess Ranch Fish Hatchery, in his calculations of base The river normally is dry in the Baja subarea but flow; therefore, his calculations may be overestimated. flows again through Afton Canyon where it exits the The estimated total return flow from the fish hatcheries Mojave River ground-water basin. Flow through Afton varied from about 300 acre-ft during water year 1949 to Canyon, when present, is a combination of base flow, about 18,000 acre-ft during water year 1991 (Lines, occasional storm runoff from the San Bernardino 1996, p. 21). Return flows from the fish hatcheries are Mountains, and local summer storm runoff. Ground- discharged to the Mojave River about 5 mi upstream of water discharge to the Mojave River begins about 1 mi the Upper Narrows (fig. 6), in a reach of the river where upstream of gaging station 10263000 in Afton Canyon. a shallow (about 40 ft beneath the channel bottom) clay Although the quantity of discharge is substantially less layer underlies the channel bottom. This clay layer at Afton Canyon (figs. 4D and 5D), geologic structures retards the deep infiltration of the return flow to the impede ground-water flow forcing ground water underlying aquifer; therefore, little deep infiltration of toward the surface, similar to what occurs at the Upper the fish hatchery return flow occurs prior to reaching and Lower Narrows. Ground-water discharge upstream the gage at the Lower Narrows. For example, in water of the gage in Afton Canyon has accounted for only year 1990, the quantity of estimated base flow (8,000 about 7 percent of the total river flow since water year acre-ft) was equal to the quantity of return flow from 1930 (Lines, 1996, p. 30). Annual base flow estimates the fish hatcheries (8,000 acre-ft) (Lines, 1996, figs. 20 range from as low as 130 acre-ft in 1976 to a high for and 26). the period of record (1930–32, 1953–78, and 1981–94) A comparison of mean daily discharge data of about 1,200 acre-ft in 1981 (Lines, 1996, p. 31). between The Forks and the Lower Narrows gages near Stormflow makes up the remaining volume of water Victorville indicates that base flow at the Lower that passes through Afton Canyon. During the years

Surface-Water Hydrology 13

Afton

10263000 Canyon (See 22 5 E R 6 E

40 21 Sewage ponds

Nebo Annex – Golf course Yermo Annex – Sewage ponds VVWRA – 30' 5 6 7 ° 116 15 20 R 4 ER 4 R

Troy Wash

Lake

Manix Cady EXPLANATION Camp 3 E

Upper sewage ponds Lower sewage ponds Effluent-irrigated field Sewage ponds City of Barstow – Nebo Annex – Springs (VVWRA)

Site of artificial recharge adjacent to the streambed Newberry Jess Ranch Fish Hatchery Mojave River Fish Hatchery State Water Project – Mojave Agency Morongo pipeline (imported water) Victor Valley Wastewater Reclamation Authority 20 MILES table 4 for amount of recharge) Dry lake (playa) Mojave River ground-water basin boundary Gaging station and number Channel-geometry site and number Site of artificial recharge to the streambed – Lake 1 2 3 4 Coyote 45'

1 2 W 1

19 FH FH

R 2 ER 2 R SWP 10260500 6 18 4 20 KILOMETERS R 1 E 18 5

10 Meridian San Bernardino San ermo 7 Annex Y 1 Mile Valley Kilometer

17 Lucerne 2 Enlargement of Nebo Annex area 0 01 10 ° R 1 W R 1 E R 1 W Nebo Annex 3 117 Lucerne Lake U.S. Marine Corps Nebo See enlargement Annex 1 USMC 10262500 Lake Rabbit 0 0 Barstow San Bernardino Mts R 2 W R 16 k e e r C p e

e

D Apple Valley 15 R 3 W 2 W R 3 W

1 6 2 1 18 2 SWP FH 10260500 FH 15' 14 4 8 5 Iron Mtn 3 7

10

r Valley

Lake

9

e Harper

v Summit Helendale

R 4 W i R 4 W

Upper Lower R 10261100 10260950 Narrows Narrows Hesperia

7

11

e

v a M j 13 o

12 W 10261500 River Victorville

15

Mojave

Wash West Fork

R 5 W R 5W Fremont Fremont Silverwood Lake R 42 E R43 E R44 E R 45 E R 46 E R 47 E Airport Southern Logistics Adelanto California 30' ° 117 R 6 W

R 6 W Sheep Creek Sheep Shadow Mts R 41 E R 7 W Lake R 7 W El Mirage San Gabriel Mts Location of channel-geometry and artificial-recharge sites in the Mojave River ground-water basin, southern California. R40E ° 45' 30' T3N 35 T4N ° T5N 32 S ° 12 N 10 N T8N T7N T9N T6N T T T 11 N T 11 T 34 34 Figure 6.

14 Simulation of Ground-Water Flow in the Mojave River Basin, California

when the gage was operational (1931, 1953–78, and transformation of annual mean discharge at 29 gaging 1981–94), the average annual flow in Afton Canyon stations in the Mojave Desert region to develop a was about 4,630 acre-ft (fig. 5D, table 1). During years relation between channel geometry and annual mean of low flow, summer evapotranspiration rates can flow. The discharge data were collected for various exceed discharge, resulting in no flow at the gage. This time periods for each of the 29 gaging stations for the has happened several times, most recently during the period 1900–93. Regression analysis showed that summer of 2000 (Jeffrey Agajanian, U.S. Geological channel depth was not a statistically significant vari- Survey, written commun., 2000). able in this area. For washes in the Mojave Desert The Mojave River emerges from Afton Canyon region, the relation between channel width and dis- and exits the ground-water basin, where it splits into charge is described by the following equation with a separate channels that lead to East Cronese Lake and coefficient of determination of 0.73 (Lines, 1996, Soda Lake (fig. 1), which are dry lakes except after p. 18): major stormflows. Not all winter storm runoff from the mountains reaches Afton Canyon. Stormflows are more Q = 28 × W 2.345 likely to reach Afton Canyon and beyond to the dry lakes during years when the streambed in the upper where reaches has been wetted several times before peak Q is the mean discharge, in acre-feet; and stormflows (Lines, 1996, p. 15). W is the channel width, in feet. On the basis of this relation, Lines (1996, p. 20) Ungaged Tributary Streams estimated that the ungaged tributary inflow to the Mojave River averaged about 8,700 acre-ft/yr for the Several major and minor ephemeral streams and period 1931–94. Inflow averaged about 2,400 acre-ft/yr washes lie within in the Mojave River ground-water to the Alto subarea (between the headwaters and Lower basin. For brief episodes after intense storms, Narrows gage), about 2,400 acre-ft/yr to both the precipitation and runoff from the mountains result in Transition zone and to the Centro subarea, and about ephemeral streamflow. Some of the storm runoff infil- 3,900 acre-ft/yr in the Baja subarea (Lines, 1996, trates into the upper reaches of the washes originating p. 20). However, because of the ephemeral nature of the in the southern part of the Alto subarea and into the ungaged tributary streams, runoff may not occur every channels of small streams that contribute flow directly year, or even following every storm. Since it is not pos- to the river. These washes and small streams are not sible to determine when the runoff from ungaged tribu- gaged, thus the amount of recharge they contribute to taries occurred in the past, Lines (1996) assumed that the basin has not been measured directly. During wetter the runoff in the ephemeral tributary streams occurred years, a significant quantity of this runoff is carried to at the same relative magnitude and during the same the Mojave River. years that ephemeral runoff in the Mojave River Lines (1996, p. 19) used channel geometry tech- occurred at the Barstow gaging station (10262500). niques to estimate the amount of water entering the The river flows as far as the Barstow gage only during Mojave River from 22 ephemeral streams (shown as large stormflows and, therefore, periods of flow at this “Channel-geometry site and number” in figure 6). gage were used as an indicator of periods of probable Channel-geometry techniques have been used by runoff from the tributary washes. Using this assump- several researchers to estimate various streamflow tion, it is possible to estimate the amount of inflow characteristics in the western United States (Lines, from tributary streams over the entire basin for years of 1996, p. 16). The width and depth of a channel or wash high discharge. For example, during 1969, about is strongly correlated with the annual mean discharge 340,000 acre-ft of water entered the basin at The Forks, in a region. Multiple linear regression can be used to which was about 820 percent of the average annual estimate the relation between annual mean discharge discharge; therefore, the estimated annual inflow from with channel width and depth. A complete discussion tributary streams was about 820 percent of average, or of the concept and methods of channel geometry is in about 70,000 acre-ft (Lines, 1996, p. 20). Table 2 shows Hedman (1970). Lines (1996) used the logarithmic the estimated annual inflow from tributary streams to

Surface-Water Hydrology 15

the Mojave River; the values were based on the ratio of basin and crop out in the surrounding mountains and average annual discharge of the Mojave River at the hills (fig. 7). In some places, the confining rocks at the Barstow gaging station between 1931 and 1994. sides of the basin are buried by unsaturated alluvial deposits. The unconsolidated deposits consist of gravel, sand, silt, and clay deposited by the recent GROUND-WATER HYDROLOGY Mojave River and the Pliocene-Pleistocene ancestral Mojave River, by tributary alluvial fans, and by older The aquifer system within the 1,400 mi2 Mojave River ground-water basin consists of unconsolidated streams and alluvial fans that predate the origin of the alluvial deposits. The basin boundary initially was Mojave River surface-water drainage basin. Also defined by the California Department of Water present are local deposits of silt and clay that accumu- Resources (1967) and later was modified by Hardt lated in lakes and playas along the margins of the basin. (1971) and Stamos and Predmore (1995). Generally, The consolidated deposits consist of pre-Tertiary the boundary is formed by nonwater-bearing consoli- igneous and metamorphic rocks and Tertiary volcanic dated rocks that underlie the alluvial deposits of the and sedimentary rocks.

Table 2. Estimated annual inflow from selected ephemeral tributary streams to the Mojave River, southern California, 1931–99 [Runoff data were compiled for 1931–99; missing years had no runoff. See figure 6 for location of channel-geometry sites. Site numbers are the same as those reported in Lines (1996, table 2); for the purposes of this current report, however, sites 6 and 7 are reversed. Number in parentheses is percent of con- tribution of tributary to total inflow into river reach. Number in bold italics corresponds to tributary segment number in streamflow-routing package of the model. Total inflow in acre-feet]

Runoff in the upper Mojave River reach (The Forks to Lower Narrows gaging station, 10261500)

Year Site 1 Site 2 Site 3 Site 4 Site 5 Site 6 Site 7 Site 8 Site 9 Site 10 (0.02) (0.01) (0.03) (0.24) (0.04) (<0.01) (0.04) (0.12) (0.02) (0.48) 4 6 8 10 13 14 16 22 24 26 1932 100 50 150 1,200 200 0 200 600 100 2,400 19353254060618 370 1937 280 140 410 3,310 550 0 550 1,660 280 6,620 1938 370 190 560 4,440 740 0 740 2,220 370 8,800 1939112183036 140 1941 260 130 390 3,100 520 0 520 1,550 260 6,200 1943 240 120 360 2,900 480 0 480 1,450 240 5,810 1944 100 50 150 1,200 200 0 200 600 100 2,400 1945 60 30 90 740 120 0 120 370 60 1,490 1946 30 17 50 410 70 0 70 200 30 820 19478461001601650 8 190 1952 30 17 50 410 70 0 70 200 30 820 1958 50 30 80 650 110 0 110 320 50 1,300 19622132040412 250 1966 18 9 30 220 40 0 40 110 9 430 1967 20 10 30 240 40 0 40 120 20 480 1969 390 200 590 4,700 780 0 780 2,350 200 9,410 1978 140 70 210 1,660 280 0 280 830 140 3,310 1979 14 7 20 170 30 0 30 80 14 340 1980 370 190 560 4,440 740 0 740 2,220 370 8,880 1983 250 120 370 2,980 500 0 500 1,490 250 5,950 1993 330 170 500 3,980 660 0 660 1,990 330 7,970 1995 30 15 44 355 59 0 59 178 30 710 1998 28 14 42 336 56 0 56 168 28 672

16 Simulation of Ground-Water Flow in the Mojave River Basin, California Geologic Setting Mojave River that originate in the western part of the San Bernardino Mountains. Drainage basins in the The ground-water basin is bordered to the south eastern part of the neighboring San Gabriel Mountains by the San Gabriel and San Bernardino Mountains— contribute significantly less water to the basin because segments of the central Transverse Ranges that were much of the runoff from these mountains is diverted uplifted along the San Andreas Fault during the past into other basins south of the study area by deeply several million years (Meisling and Weldon, 1989; incised streams along the San Andreas Fault Zone and Matti and Morton, 1993). These large ranges of granitic other northwest-trending faults (fig. 7). and metamorphic rocks of pre-Tertiary age contain the The ground-water basin arcs northward and east- main catchment areas of the ground-water basin. The ward amid low mountains of the southern and central basin is recharged primarily by tributaries of the Mojave Desert. These small ranges are composed of a

Table 2. Estimated annual inflow from selected ephemeral tributary streams to the Mojave River, southern California, 1931–99—Continued

Runoff in the middle Mojave River reach (Lower Narrows Runoff in the lower Mojave River reach (Barstow gaging sta- gaging station, 10261500, to Barstow tion, 10262500, to Afton Canyon gaging station (10263000) gaging station, 10262500) Total Year Site 11 Site 12 Site 13 Site 14 Site 15 Site 16 Site 17 Site 18 Site 19 Site 20 Site 21 Site 22 inflow (0.06) (0.01) (0.17) (0.17) (0.12) (0.47) (0.05) (0.25) (0.25) (0.31) (0.12) (0.20) 28 32 34 36 38 40 42 44 46 48 50 52 1932 310 50 880 880 620 2,440 400 2,000 2,000 2,480 960 160 18,180 1935 12 2 30 30 19 80 20 100 100 130 50 10 736 1937 600 100 1,700 1,700 1,200 4,700 1,140 5,670 5,670 7,040 2,720 450 46,490 1938 1,200 200 3,400 3,400 2,400 9,400 1,510 7,550 7,550 9,360 3,620 600 68,620 1939 6 1 14 14 10 40 14 70 70 90 30 6 440 1941 780 130 2,210 2,210 1,560 6,110 1,050 5,250 5,250 6,510 2,520 420 46,930 1943 780 130 2,210 2,210 1,560 6,110 1,000 5,000 5,000 6,200 2,400 400 45,080 1944 300 50 850 850 600 2,350 400 2,000 2,000 2,480 960 160 18,000 1945 180 30 510 510 360 1,410 250 1,250 1,250 1,550 600 100 11,080 1946 120 20 340 340 240 940 140 720 720 900 350 60 6,587 1947 20 4 70 70 50 190 40 200 200 240 90 16 1,588 1952 110 18 310 310 220 850 140 720 720 900 350 60 6,405 1958 170 30 480 480 340 1,320 200 1,000 1,000 1,240 480 80 9,520 1962 6 1 17 17 12 50 8 40 40 50 20 3 362 1966 50 9 150 150 90 420 60 320 320 400 160 30 3,065 1967 60 10 170 170 120 470 80 400 400 500 190 30 3,600 1969 1,200 200 3,400 3,400 2,400 9,400 150 7,500 7,500 9,300 360 600 64,810 1978 420 70 1,190 1,190 840 3,290 500 2,500 2,500 3,100 1,200 200 23,920 1979 50 8 140 140 100 380 50 250 250 310 120 20 2,523 1980 1,140 190 3,230 3,230 2,280 8,930 1,500 7,500 7,500 9,300 3,600 600 67,510 1983 600 100 1,700 1,700 1,200 4,700 1,000 5,000 5,000 6,200 2,400 400 42,410 1993 1,020 170 2,890 2,890 2,040 7,990 1,300 6,500 6,500 8,060 3,600 520 60,070 1995 89 15 252 252 178 696 120 602 602 746 289 48 5,367 1998 84 14 238 238 168 658 114 569 569 706 273 46 5,077 Total...... 558,370

Ground-Water Hydrology 17 T 10 N T8N T 12 N Ts Tv pTb Qp Tv R 5 E

ertiary Age Holocene Holocene to Pleistocene Holocene to Pliocene Pleistocene to Pliocene Holocene to Pliocene T Pre-Tertiary shown in figure 8, all

Tv Systems Research Institute, 1992 Ts A-A' 116 30' H H' Geology is modified from Environmental Qp Ts QTu QTu Section EXPLANATION Baja fault QTu 22R6,7 27J3-5

G' G pTb R 3E Fault Qya Ts

(delineated for this study) alley

V Newberry

Line of geologic section – other sections shown in figure 9

Fault fault Boundary of Mojave River ground-water basin Approximate outline of floodplain aquifer Well and identifier Mojave olcanic rocks ounger Mojave River alluvium Sedimentary rocks V Igneous and basement complex Playa deposits Recent Mojave River alluvium Y Undifferentiated alluvium Older alluvium of ancestral Mojave River Older lake deposits Unconsolidated deposits A' 45' Calico- Tv Qra Qp QTu QTu Ts Tv QpQp Qra pTb Qya QTu QTol QToa pTb Unconsolidated deposits - Floodplain aquifer Unconsolidated deposits - Regional aquifer Unconsolidated deposits - Outside of the Mojave River ground-water basin Consolidated deposits 1C3-5

Waterman E Fault A 10Q3,4 Ts Ts T2N T5N T4N T 6 N F F' F' Tv Valley HarperHarperaterman Lake Lake Fault Fault Fault) Zone Zone pTb Lucerne (Waterman(W Fault) Qp R 47 E R 1 E R 1 W R 1 E Ts

Camp Rock - pTb Barstow QTu QTu Ts 117 00' fault pTb QTu QTu

A' pTb

E'

FaultFault Mt General Qya R 45 E E B' R 3 W Tv QTol QTol 1R6,7

Fault

Mtn QTu Iron 15' pTb

Helendale Victorville Apple Valley fault

Narrows Fault fault Narrows Narrows

Ts

Qp

fault fault B

QToa QToa

D' 1C3-5

Iron Mtn Iron L oc k h ar t LockhartLockhart Mtn Iron Hesperia

C-C' San Bernardino Mts R 43 E D QTu QTu R 5 W pTb

20 MILES River

QTu QTu Mojave fault Adelanto 1C3-5 QToa QToa QTu QTu 23R2,3 pTb

Shadow Mts fault Tv

117 30' A Airport Southern Logistics California 20 KILOMETERS Ts

10 Shadow Mts R 41 E QTu QTu pTb pTb R 7 W Qp 10 T9N T 32 S T 12 N T 11 N T 10 N pTb San Andreas Fault Zone pTb Generalized geology of the Mojave River ground-water basin, southern California. 0 0 San Gabriel Mts 45' 35 00' 34 30' Figure 7.

18 Simulation of Ground-Water Flow in the Mojave River Basin, California diverse assortment of granitic and metamorphic rocks The pre-Tertiary basement complex (pTb) con- of pre-Tertiary age and volcanic and sedimentary rocks sists mainly of Mesozoic granitic rocks (Cretaceous of Tertiary age. East of Victorville, the mountains are and Jurassic age), accompanied by lesser amounts of clustered in an east-west-trending belt that is flanked to Proterozoic granitic and gneissic rocks (Precambrian the north and south by alluviated lowlands of the age); Mesozoic metavolcanic rocks (Jurassic age); and Mojave Valley and Lucerne Valley (fig. 7). This moun- Mesozoic, Paleozoic, and late-Proterozoic metasedi- tain belt represents a broad basement anticline that was mentary rocks. The rocks of the basement complex uplifted along with the nearby San Bernardino form the large mountains of the central Transverse Mountains (Howard and Miller, 1992; Cox and Hill- Ranges that border the south side of the Mojave River house, 2000). The arcuate path of the Mojave River and ground-water basin and also many of the smaller its ground-water basin across the Mojave Desert devel- mountains and hills that are distributed around, and oped as the ancestral Mojave River forged a route locally within the basin. Results of geophysical surveys across and around the margins of the anticline (Cox and (Subsurface Surveys, Inc., 1990; Zohdy and Bisdorf, Hillhouse, 2000). 1994) and cuttings from exploratory boreholes (Dibblee, 1967, table 4) indicate that the rocks of the The southern and central Mojave Desert is cross- basement complex beneath the Mojave River ground- cut by a series of northwest-trending faults, including water basin range from 1,000 to 4,000 ft below land the Helendale, Camp Rock-Harper Lake, and Calico- surface and consist of Tertiary volcanic and sedimen- Newberry Faults (fig. 7) (Dibblee, 1961; Dokka and tary rocks and overlying Quaternary (Pleistocene) and Travis, 1990). Geologic features and roads and fences Tertiary (Pliocene) alluvial deposits. However, at the that were offset following historical earthquakes show Upper Narrows near Victorville, granitic rocks lie a that these faults characteristically generate right-lateral mere 50 ft beneath the active channel of the Mojave strike-slip displacements consistent with those of the River (figs. 7 and 9, section C–C′). The pre-Tertiary nearby, more active San Andreas Fault. Some of the basement complex typically has low porosity and per- faults also show evidence of vertical displacement. meability, yielding only small quantities of water to wells; however, where the basement complex is intensely fractured, as along major faults, the bedrock Stratigraphic Units is more permeable. Unmetamorphosed sedimentary and volcanic A generalized surficial geology of the Mojave rocks of Tertiary age (Ts and Tv) crop out together in River ground-water basin is shown in figure 7. Alluvial several mountain ranges north and east of Barstow; deposits of the recent and ancestral Mojave River (Qra, sedimentary rocks also occur separately at the west end Qya, and QToa, respectively) are adapted from of the San Bernardino Mountains (fig. 7). Near California Department of Water Resources Bulletin 84 Barstow, the basement complex consists of two super- (1967, pl. 2) and Cox and Hillhouse (2000, fig. 2). The posed sequences with an aggregate thickness of about 10 units shown in figure 7 are (1) pTb—igneous and 7,000 ft (Woodburne and others, 1990; Fillmore and metamorphic rocks which compose the basement com- Walker, 1996). The lower sequence is early Miocene in plex (pre-Tertiary); (2) Tv, volcanic rocks (Tertiary); age and consists of volcanic intrusions, flows, and (3) Ts, sedimentary rocks (Tertiary); (4) QTol, older pyroclastic rocks interlayered with avalanche breccia, lake and playa deposits (Pleistocene to Pliocene); (5) sandstone and conglomerate, and limestone (Tv and QToa, older alluvium of the ancestral Mojave River Ts). This lower sequence is unconformably overlain by (Pleistocene to Pliocene); (6) QTu, undifferentiated an upper sequence of middle Miocene sandstone, con- alluvial deposits (Holocene to Pliocene); (7) Qya, glomerate, shale, limestone, and volcanic ash (Ts). younger alluvium of the Mojave River (Holocene to Both sequences underlie alluvial deposits of the Pleistocene); (8) Qra, recent alluvium of the Mojave Mojave River ground-water basin throughout the area River (Holocene); (9) Qp, playa deposits (Holocene); east of Barstow (fig. 7) (Densmore and others, 1997, and (10) undifferentiated unconsolidated deposits figs. 4 and 5, Tvs). The sedimentary rocks at the west (Holocene to Pliocene). Structural and stratigraphic end of the San Bernardino Mountains consists of a relationships within the Mojave River ground-water sequence of middle Miocene-age sandstone, siltstone, basin are presented in figures 8 and 9. and conglomerate that is as much as 3,200 ft thick in

Ground-Water Hydrology 19 places (Meisling and Weldon, 1989); below the subsur- deposits (QTol) near the east end of the ground-water face, the sedimentary deposits may extend northward basin (figs. 7 and 9, section H-H′) consists of clay and toward Hesperia and Victorville. The Tertiary volcanic silt, interspersed with lesser amounts of sand and rocks generally are nonwater-bearing. The Tertiary gravel. This stratigraphic section, exposed in the walls sedimentary rocks contain water-bearing strata, but of the incised Mojave River, is about 400 ft thick and such deposits typically yield only small quantities of consists of two subunits (Jefferson, 1985; Nagy and poor-quality water to wells. Murray, 1996). The lowermost 300 ft consists of Older lake deposits of Pleistocene to late gypsum-bearing clay, silt, and sand deposited in a playa Pliocene age (QTol) are exposed on the northeast edges basin between about 2.5 to 1.0 million years ago. The of Mojave Valley and in the bluffs of the Mojave River uppermost 100 ft, deposited between about 500,000 to near Victorville (fig. 7). The deposits near Victorville 15,000 years ago, consists of clay, silt, sand, and gravel consist of interbedded clay and freshwater limestone that accumulated in a freshwater lake at the terminus of that crop out for several miles, extending from the the ancestral Mojave River. Upper Narrows upstream to the boundary between Undifferentiated alluvial deposits of Holocene to townships 5 and 6 north (fig. 7). The unit of older lake late Pliocene age (QTu) form the bulk of the regional

Regional aquifer FEET A Section B-B' (see figure 9) A' 3,600 West East Land surface Floodplain aquifer 3,200 QTu Mojave River QTu 1992 water level QTu QToa Qra QToa 2,800 Layer 1 Layer 2 Qya QTu 2,400

2,000 QTu

1,600

pTb Apple Valley fault Valley Apple 1,200 Datum is sea level Vertical scale greatly exaggerated 0 5 MILES

0 5 KILOMETERS

EXPLANATION

Qra Recent Mojave River alluvium QTu Undifferentiated alluvium Holocene Holocene to Pliocene Floodplain Regional Qya Younger Mojave River alluvium aquifer QToa Older alluvium of ancestral Mojave River aquifer Holocene to Pleistocene Pleistocene to Pliocene

pTb Igneous and metamorphic basement complex Pre-Tertiary

Figure 8. Conceptualization of the ground-water flow system and model layers near Victorville, California. Location of section shown in figure 7.

20 Simulation of Ground-Water Flow in the Mojave River Basin, California aquifer, which unconformably underlies and surrounds middle Pleistocene in age and were deposited by the (figs. 8 and 9) the floodplain aquifer throughout most of ancestral Mojave River. The permeability of this older the Mojave River ground-water basin. These deposits alluvium unit generally is between that of the undiffer- consist of sand, gravel, and silt that accumulated in entiated alluvial deposits (QTu) and the younger and alluvial-fan, braided-stream, and playa or lacustrine recent alluvium of the Mojave River (Qya and Qra). environments. Most of the deposits formed in the Thick deposits of the older alluvium extend well below Pleistocene and late Pliocene, during and before the the water table. Deposits of the older alluvial unit development of the Mojave River surface-water drain- age basin. The unit is conspicuously faulted, tilted, and between the Southern California Logistics Airport and folded, and it typically is deeply eroded. Deposits Harper Lake lie mainly above the water table (fig. 9, exposed on hills and in ravines are as much as 350 ft section D-D′). thick, and subsurface data suggest that the unit may be The recent (Qra) and younger (Qya) Mojave as much as 1,000 to 2,000 ft thick in several deep struc- River alluvium units consist of granitic sand, silt, and tural depressions near Barstow, Harper Lake, and gravel deposited by the modern Mojave River during Victorville (Dibblee, 1967, table 4, locations 41, 49, the Holocene and late Pleistocene. The deposition of 52, 55; Densmore and others, 1997, fig. 4, QTof). The the younger alluvium unit followed a major episode of unit also includes surficial deposits of sand and gravel downcutting that excavated the Mojave River canyon that accumulated on alluvial fans and within incised during the late Pleistocene, about 60,000 to 70,000 drainages during the Holocene and late Pleistocene. Clay, silt, and fine sand deposited in modern playa years ago (Cox and Hillhouse, 2000). Qya typically is basins are mapped separately as Qp. The permeability about 200 ft thick indicating that nearly complete back- of the alluvial deposits (QTu) is lower than that of the filling of the Mojave River canyon occurred near fluvial sediments of the Mojave River (Qya and Qra) Hinkley Valley and Yermo Annex, where the canyon partly because of poor sorting on alluvial fans but also was about 200 ft deep, and partial backfilling occurred because of the widespread accumulation of secondary in areas upstream from Hinkley Valley, where the can- (pedogenic and diagenetic) clay and calcium-carbonate yon was about 350 to 400 ft deep. Radiocarbon ages cement. determined for several samples of detrital charcoal The ancestral Mojave River deposited alluvium recovered from sediments near the top of the younger consisting of granitic sand, silt, and gravel of alluvium unit indicate that the backfilling episode Pleistocene to Pliocene age (QToa) as it forged a route ended about 6,000 to 7,000 years ago (Rector and oth- northward and eastward across the Mojave Desert (Cox ers, 1983; Reynolds and Reynolds, 1985, 1991; and Hillhouse, 2000). The thickness and basal age of this older alluvial unit decrease from south to north and Densmore and others, 1997, fig. 7; Rector, 1999). then change abruptly near the Southern California Downcutting resumed about 6,000 years ago, as Logistics Airport (fig. 1). Deposits south of this loca- recorded by stream terraces perched about 25 ft above tion are about 400 to 500 ft thick and are of Pliocene the active river channel at several sites north of age (2 million years old or more) at their base (fig. 9, Victorville. section D-D′), whereas deposits north of the airport are The recent alluvium (Qra) fills a smaller mostly 25 to 80 ft thick and apparently are of middle channel-shaped incision that generally is inset into the Pleistocene age (about 0.5 million years old) near their younger alluvium unit; however, at the Upper and base (fig. 9, sections D-D′ and E-E′). Results of previ- Lower Narrows it is inset into granitic bedrock (fig. 9, ous studies indicate that the ancestral Mojave River section C-C′). In the Transition zone (figs. 7 and 9, reached Pleistocene Lake Manix in the eastern Mojave section D-D ), the recent alluvium is separated from the Valley about 0.5 million years ago, (Jefferson, 1985; ′ Nagy and Murray, 1996). Based on these results, it underlying younger alluvium by a unit of clay or clayey seems likely that poorly dated deposits of fluvial sand sand. The recent alluvium ranges from about 50 to and gravel buried about 200 to 400 ft beneath the land 70 ft in thickness, recording one or more second-order surface in western Mojave Valley (QToa) (fig. 9, sec- cycles of stream incision and backfilling that occurred tion F-F′) (Densmore and others, 1997, fig. 5, Qoa) are during the past 6,000 years.

Ground-Water Hydrology 21 Alto subarea Alto subarea FEET B West East B' FEET C C' 3,000 Land surface 3,100 Southwest Northeast QTu 4N/4W-1C3-5 Land surface 2,900 QTu QToa 3,000 QToa QTu Mojave River Mojave Water table 2,800 2,900 Qra

38/80 fault Apple Valley Layer 1 2,700 2,800 pTb Layer 2 Qya pTb

Mojave River Mojave Qra 43/190 Layer 1 Layer 1 2,600 QToa 2,700 Layer QTu (Inactive) 1 (Inactive) Layer 2 (Inactive) 2,500 44/330 2,600 QTu pTb

2,400 2,500 0 2 MILES 0 0.1 MILES

0 2 KILOMETERS 0 0.1 KILOMETERS

Alto Transition zone FEET D D' 2,800 West East

7N/5W-23R2,3 Land surface 2,700 QToa 7N/5W-24R6-8 Centro subarea FEET E E' 2,600 2,300 QTu Northwest Southeast 9N/3W-1R6,7 Mojave River Mojave Qra Land surface 2,500 2,200 River Mojave 9/50 Clay Qra 239/ Qya 305 Qya Layer 1 2,400 2,100 Layer 1 Layer 2 70/118 29/150 Layer 2

QTu 2,300 2,000 QToa 75/210

288/ 71/285 QTu 510 2,200 1,900 0 2 MILES 0 2 MILES

0 2 KILOMETERS 0 2 KILOMETERS Datum is sea level Vertical scale greatly exaggerated

Figure 9. Conceptualization of the ground-water flow system and model layers at various locations along the Mojave River in southern California. Location of sections and definitions of geologic terms are shown in figure 7.

22 Simulation of Ground-Water Flow in the Mojave River Basin, California Baja subarea Afton area FEET FF'FEET H H' 2,100 North North South 2,000 South

9N/1E-10Q3,4 Land surface 2,000 1,900 Land surface Mojave River Mojave

Qra 1,900 1,800 Qya Qya QTu boundary

Water table Ground-water basin 1,800 1,700 QTu 143/200 Layer 1 Qra Layer 2 QTol Mojave River Mojave 1,700 1,600 Tv QToa QTu

142/344 Layer 1 Layer 1 1,600 1,500 Layer 2 (Inactive) (Inactive) Tv 1,500 1,400

FEET G Baja subarea G' 1,900 North South 10N/3E-27J3-5 9N/3E-22R6,7 Land surface 1,800

Mojave River Mojave QTu Qra Qya 1,700 30/45 104/129 38/90 Layer 1 Layer 2 1,600

100/290 QToa 1,500 28/255 Tv QTu

1,400 0 2 MILES

Datum is sea level 0 2 KILOMETERS Vertical scale greatly exaggerated

EXPLANATION

Qra Recent Mojave River alluvium Unconsolidated deposits outside the Mojave River ground-water basin Holocene to Pliocene Holocene Floodplain Qya Younger Mojave River alluvium aquifer Tv Volcanic rock Tertiary Holocene to Pleistocene pTb Igneous and metamorphic basement complex QTu Undifferentiated alluvium Holocene to Pliocene Pre-Tertiary Older alluvium of ancestral Mojave River QToa Regional Multiple-well monitoring site – Sequence Pleistocene to Pliocene aquifer 9N/3E-22R6,7 numbers are from deepest to shallowest. Numbers indicate depth to water/depth QTol Older lake deposits of well, in feet. White box represents Pleistocene to Pliocene 142/344 screened interval of 20 feet for each well

Figure 9.—Continued.

Ground-Water Hydrology 23 Definition of Aquifers of Holocene to Pliocene age (QTu). These deposits have a combined thickness of more than 2,000 ft in The water-bearing deposits form two some places (fig. 8). Permeability generally decreases aquifers—a floodplain aquifer and a regional aquifer with depth and cementation occurs in some areas underlying and surrounding the floodplain aquifer. The (Hardt, 1971, p.12). On the basis of field observations, consolidated-rock and basement-complex units gener- the QToa deposits and the upper, more permeable ally are considered to be impermeable, forming the 300 to 800 ft of the QTu deposits constitute most of the base of the ground-water basin. regional aquifer. Data from multiple-well monitoring Perched water-table conditions exist east of the sites indicate large differences in water levels between city of Adelanto (Montgomery Watson, Consultants, the wells perforated in the lower QTu deposits and 1995), near El Mirage Lake (dry) in the Oeste subarea those perforated in the overlying deposits (fig. 10). The (Smith and Pimentel, 2000), and near Harper Lake differences in hydraulic head illustrate the poor (dry) (shown in fig. 11 in “Effects of Faulting on hydraulic connection between the lower QTu deposits Ground-Water Flow” section). During this study, we and the overlying deposits; as a result, the lower QTu focused on unconfined and confined ground-water con- deposits transmit very little, if any, water to the overly- ditions only and did not address areas of perched water. ing deposits. Although the lower QTu deposits contain The permeable recent river deposits of Holocene a substantial amount of ground water in storage, the age (Qra) and the younger river deposits of Holocene to low-permeability and fine-grained nature of the sedi- Pleistocene age (Qya) constitute the floodplain aquifer ments result in low well yields, generally poor-quality (figs. 8 and 9). In some areas, the floodplain aquifer water (high dissolved-solids concentrations), and large extends beyond the recent floodplain to include the drawdowns in wells. Estimated transmissivity values deposits of the ancestral Mojave River. Described in for the regional aquifer range from 1,000 to previous reports as the “shallow alluvial aquifer” and 13,000 ft2/d (Hardt, 1971). In the Alto subarea, trans- “Mojave River aquifer,” the floodplain aquifer is more missivity values are 20,000 ft2/d or greater as much as productive than the regional aquifer, yielding most of 5 mi away from the river and are related to the older the ground water pumped from the basin. These allu- alluvium of the ancestral Mojave River (QToa) (fig. 8). vial deposits are 100 to 200 ft thick and are within The older lake deposits (QTol) yield little water about 1 mi of the Mojave River (figs. 7–9). However, to wells and may act as confining layers between the the aquifer is much thinner in the area between the aquifer systems (California Department of Water Upper and Lower Narrows near Victorville because Resources, 1967, p. 23). Electric and geologic well consolidated rock formations are present at depths as logs indicate that these clay deposits underlie the active shallow as 50 ft below the streambed (fig. 9, section channel of the Mojave River at depths of 40 to 50 ft sev- C-C′) (Slichter, 1905, p. 55). Wells drilled in the river eral miles upstream from the Upper Narrows and also deposits typically yield between 100 and underlie the river throughout most of the Transition 2,000 gal/min, with reported rates as high as zone at depths of 50 to 80 ft. 4,000 gal/min (Hardt, 1971, p.11). These deposits accept most, if not all, of the recharge from the river. Hardt (1971) estimated transmissivity values from spe- Effects of Faulting on Ground-Water Flow cific capacity data at individual wells and reported val- ues for the floodplain aquifer between 13,000 to Faults and other geologic structures partially 27,000 ft2/d. For this report, Hardt’s (1971) transmis- control ground-water flow in the Mojave River ground- sivity estimates were supplemented using recent spe- water basin. The basin is dominated by extensive cific capacity data. Following the example by Driscoll right-lateral strike-slip faults that trend predominantly (1986, p. 1021) for unconfined aquifers, transmissivity northwest to southeast. The faults are barriers or partial was estimated by multiplying specific capacity data (in barriers to ground-water flow in the regional aquifer gal/d/ft) by 200 to obtain transmissivity in ft2/d. and, in many places, the floodplain aquifer, resulting in The regional aquifer extends throughout most of stairstep-like drops in the water table across the fault the study area and it consists of unconsolidated older zones (Stamos and Predmore, 1995). Between the fault alluvium of the ancestral Mojave River of Pleistocene zones, the water levels are relatively flat (fig. 11). to Pliocene age (QToa) and undifferentiated alluvium Historically, there were many perennial reaches along

24 Simulation of Ground-Water Flow in the Mojave River Basin, California the river where ground water was forced to the surface owing to the pumping of ground water in the Mojave upgradient of faults. Consolidated rocks at shallow River Basin. depths also obstruct ground-water movement and force Documented barriers to ground-water flow water to the surface, such as at the Upper Narrows, the include the Helendale Fault, the Lockhart Fault, the Lower Narrows, and Afton Canyon. Perennial reaches Calico-Newberry Fault, and the Camp Rock-Harper Lake Fault zone, also known as the Waterman Fault caused by faults and shallow bedrock were vital desert (Hardt, 1971). All documented barriers shown in watering places used by many Native Americans and figures 7 and 11 are denoted by an uppercase “F” in the early explorers who traveled through the region in the word “Fault.” Several known, but previously unnamed, late 1700’s (Thompson, 1929; Lines, 1996, p. 1). Most faults affect ground-water flow throughout the basin. of these historically perennial reaches are now dry Faults previously unnamed are referred to in this report

2,825 4N/4W-1C2-5 4N/4W-1C2-5 Alto subarea Well depth, Geologic in feet unit 2,800 C2 620 QTu

C3 330 QTu 2,775 C4 190 Qya

C5 80 Qra 2,750 2,825 4N/4W-3A2-5 4N/4W-3A2-5 Alto subarea Well depth, Geologic 2,800 in feet unit A2 790 QTu

A3 510 QTu 2,775 A4 360 QToa

A5 235 QToa 2,750 2,510 7N/5W-24R5-8 Well depth, Geologic in feet unit WATER-LEVEL ALTITUDE, IN FEET ABOVE SEA LEVEL 2,485 R5 550 QTu

R6 285 QTu 2,460 7N/5W-24R5-8 R7 150 Qya Transition zone R8 50 Qra 2,435

2,410 1992 1994 1996 1998 2000 YEAR

Figure 10. Altitude of measured water levels at three multiple-well monitoring sites in the Mojave River ground-water basin, southern California. Geologic units shown in figure 9; location of wells shown in figure 11.

Ground-Water Hydrology 25 Afton

Canyon Shows 20 MILES 5 E R 6 E

40 (water levels 116 30'

15 1,750 20 KILOMETERS

1,6001,650 ornia, November 1992 (Stamos and Predmore,

Troy Wash

Lake

Manix ? 10

Baja fault ? Newberry Fault EXPLANATION Cady Camp Delineated for this study 3 E R E 4 R

? Calico-? 10 – 1,700 1,800 Areas of perched water table Water-level contour (measured 1992) – altitude of water level. Contour interval 50 feet. Dashed where approximate, queried uncertain. Datum is sea level Generalized direction of ground-water flow Multiple-well monitoring site shown in figure 10) Well with water-level data Mojave River ground-water basin boundary Fault fault Subarea Dry lake (playa) Southern California Logistics Airport

1,750 Springs A Lake 45'

Coyote Newberry

1C2-5

0 0

? ? 3,000 Baja subarea R 2 ER 2 R

Camp Rock-Harper

Lake Fault Zone 18

? R 1 E 1,900 (Waterman Fault)

?

1,850 3,100

Meridian San Bernardino San 3,000

1,950 2,000 2,950 47 E

R Valley ? Este subarea Lucerne

? aterman E Fault ?

W R 1 W R 1 E

R 1 W 2,050 Lucerne Lake ?

117 00'

Lake

Barstow 2,900 Rabbit

2,850 ? ?

San Bernardino Mts R 2 W R ? Lenwood fault 2,800 k alley Fault e e V r 2,150

Mt. General 18

2,200 C 2,050 2,750 p 2,250 e

2,100 e

Hinkley D R 3 W 2 W R 3 W ? 2,300

Apple Valley

2,000

Apple Valley fault 1C2-5

15' ? fault ? Mtn Iron

Helendale Narrows fault

2,700 Fault ?

Adelanto fault 1,950 R 4 W

R 4 W

2,500 Lake 2,650 3A2-5

2,850

1,900 ?

? ? 2,450 Helendale

? ? Hesperia Lockhart 2,350 24R5-8 Harper

2,100 ? 2,250 River

Victorville Mojave 2,200 West Fork A Centro 2,400 2,750 subarea R 5 W R 5W Alto subarea

River ? ? Mojave ? 15 ? ? 2,800 Silverwood Lake

R 42 E R43 E R44 E R 45 E R 46 E 2,550 2,600 Alto ?

fault 117 30' Transition zone R 6 W

Shadow Mts R 6 W Adelanto

Shadow Mts 2,900 R 41 E R 7 W sub- area

? Oeste R 7 W ? Altitude of water levels and generalized direction ground-water flow in the Mojave River basin, southern Calif Lake San AndreasFault Zone ? San Gabriel Mts El Mirage R40E 45' 32 S T9N 12 N 10 N T3N T4N

T5N T7N T8N T6N T T T 11 N T 11 T 35 00' 34 30' Figure 11. 1995). (See figure 1 for location of subareas).

26 Simulation of Ground-Water Flow in the Mojave River Basin, California as the Apple Valley, Narrows, Shadow Mountains, southeast from the Narrows fault is another suspected Adelanto, Iron Mountain, Mt. General, and Baja faults barrier to flow on the east side of the Mojave River, and are denoted by a lower case “f” in the word “fault” which we refer to as the Apple Valley fault (fig. 11). in figures 7 and 11. The faults in the study area are dis- Ground water moves from the Transition zone to cussed in the following paragraphs roughly in the order the Centro subarea across the northern extension of the they are encountered in the ground-water basin, from Helendale Fault (fig. 11). Water-level data collected the upper subareas to the lower subareas. from USGS multiple-well monitoring sites and com- The southern extension of the Helendale Fault piled from historical sources indicate that this fault near the town of Lucerne Valley is an effective barrier restricts subsurface flow in the regional aquifer but not to subsurface flow and forms the southeastern bound- in the overlying floodplain aquifer (Hardt, 1971, p. 21). ary of the Mojave River ground-water basin. Water lev- To provide site-specific information for that area near els east of the fault and outside the study area are the Helendale Fault, monitoring wells were installed; between 60 to 100 ft lower than water levels west of the seismic refraction, water-level, and water-quality data fault (Schaefer, 1979, fig. 3; Stamos and Predmore, were collected; and hydraulic properties of the flood- 1995). West of the Helendale Fault, ground water flows plain and the regional aquifers were analyzed (Gregory westward from the southwestern part of the Este sub- O. Mendez, U.S. Geological Survey, written commun., area toward the Alto subarea (fig. 11); the water-table 1998). On the basis of these data, flow through the gradient in this area is relatively flat. The amount of floodplain aquifer near the Helendale Fault is estimated subsurface flow across the Alto/Este subareas boundary to be between 5,000 to 6,000 acre-ft/yr. Ground-water is estimated to be 300 to 600 acre-ft/yr (Stamos and flow through the surrounding and underlying regional Predmore, 1995). aquifer does not exceed 1,200 acre-ft/yr but probably is Hydrologic data indicate that faulting, possibly much less because the Helendale Fault is believed to be connected to the geologic formation of the Upper a barrier to flow in the regional aquifer (Gregory O. Narrows or subsurface structures associated with Mendez, U.S. Geological Survey, written commun., Shadow Mountains, affects ground-water flow in the 1998). Alto subarea. Faulting in this area is indicated by steep Ground water passing into the Centro subarea water-level gradients northwest of Victorville while a from the Transition zone flows around Iron Mountain relatively flat water-level gradient is maintained toward Harper Lake through Hinkley Valley on the east between the city of Adelanto and the northern edge of side and through a narrow gap between the Helendale the Southern California Logistics Airport (fig. 11). Fault and Iron Mountain on the southwest side. How- Water-level data collected near the city of Adelanto ever, steep water-level gradients between the Helendale (Stamos and Predmore, 1995; Mendez and Chris- Fault and Iron Mountain on the southwest side tensen, 1997) also indicate the probable presence of a (fig. 11)—shown by water-level declines of more than geologic structure controlling ground-water flow. 150 ft within a distance of only about 2 mi—indicate Although there is no surface expression to confirm the that subsurface faults or shallow geologic features presence of faults in this area, preliminary geologic probably impede subsurface flow to Harper Lake. We mapping by the USGS in the vicinity of the Lower refer to the barrier, or fault, affecting ground-water Narrows shows evidence of several north trending movement in this area as the Iron Mountain fault. faults that are exposed within the river terraces along The Lockhart Fault cuts through the northern the eastern boundary of the Southern California part of Iron Mountain and extends south of Harper Logistics Airport and an eastwest-trending fault that is Lake through Hinkley Valley and into the unconsoli- exposed within the terraces west of the Lower Narrows dated rocks south of the Mojave River in the Centro (Brett F. Cox, U.S. Geological Survey, written com- subarea (figs. 7 and 11). This fault appears to impede mun., 1997). To explain the anomalies in the water- the movement of ground water in the regional and the level data and possibly to explain the large differences floodplain aquifers although there is no evidence of this in water levels between adjacent wells as mentioned effect in the floodplain aquifer along the river (Gregory earlier in this report, we propose that there are three C. Lines, U.S Geological Survey, oral commun., 1996). separate faults in this area; for the purpose of this Data collected from USGS monitoring wells report, we refer to these three faults as the Narrows, installed in the Lenwood area during this study reveal a Shadow Mountains, and Adelanto faults. Extending previously unknown barrier between the Lockhart

Ground-Water Hydrology 27 Fault and the city of Barstow, which is referred to in recharge. Recharge to the aquifer system from direct this report as the Mt. General fault. This fault is an precipitation is considered minimal because precipita- effective barrier to ground-water flow in both the tion or runoff do not adequately meet evapotranspira- regional and the floodplain aquifers. tion and soil-moisture requirements. Mean annual The Camp Rock-Harper Lake Fault zone, also precipitation for 1960–91 was about 6 in. at Victorville known as the Waterman Fault, consists of five relatively and about 4 in. at Barstow and Afton Canyon (James, young strike-slip faults (Cox and Wilshire, 1993). 1992). The principal sources of ground-water dis- Water-level data collected from wells in this area indi- charge from the basin are pumpage, evapotranspira- cate that two of the five faults, Fault C and E (referred tion, and base flow at Afton Canyon. Previous to as the Waterman and the Waterman E Faults in this investigators have estimated selected sources of report) affect subsurface flow and cause abrupt, ground-water recharge to, and discharge from, the stairstep-like changes in the water table as ground Mojave River ground-water basin for various periods water flows eastward from the Centro to the Baja sub- (table 3) (California Department of Water Resources, area (fig. 11). Water-level data collected from multiple- 1967; Hardt, 1971; Mojave Basin Area Watermaster, well monitoring sites indicate that the two faults 1996b, Table C-1). These estimates for the upper, impede ground-water movement in the floodplain aqui- middle, and lower Mojave basins and the estimates of fer and the underlying regional aquifer. flow between them are presented in table 3. In the Baja subarea, ground-water flow is impeded by faulting and shallow, low-permeability Recharge deposits. Historical and recent water-level data show that the Calico-Newberry Fault has had a significant The Mojave River effect on the water table in the Baja subarea, causing a sharply lowered water table east of the fault. In 1992, The principal source of recharge to the basin is water levels were about 50 ft lower on the east, or derived from runoff in the San Bernardino and San downgradient side, than on the west, or upgradient Gabriel Mountains. Hardt (1971, p. 12) estimated that side, of this fault (Stamos and Predmore, 1995) 92 percent of total basin recharge originates in the San (fig. 11). Subsurface flow through the Baja subarea also Bernardino Mountains. The Mojave River is the natural is affected by low-permeability deposits at shallow conduit for most of the stormwater and snowmelt run- depths between Camp Cady and Afton Canyon and off from the mountains to the basin. Surface water infil- possibly by a previously unnamed fault, referred to as trates the permeable deposits of the river to recharge the Baja fault in this report. Near Camp Cady, fine- the floodplain aquifer. Recharge from the river is also grained unconsolidated deposits near land surface, termed stream leakage in this report. Although floods which are associated with ancient Manix Lake recharge the floodplain aquifer along the entire length (California Department of Water Resources, 1967, p. of the river, most of the water infiltrates the upper 23), cause an abrupt change in the water-table gradient. reaches of the river where flows occur more frequently Data from geologic well logs indicate that these and with larger magnitudes. During years of peak dis- deposits extend to Manix Wash and toward Afton charge, or floods (for example, 1969, 1983 and 1993), Canyon. At Afton Canyon, low-permeability deposits flow in the Mojave River can last several months result- at shallow depths below the Mojave River restrict sub- ing in significant ground-water recharge. Lines (1996) surface flow forcing ground water to the surface, gives a detailed description of a water-balance method resulting in base flow to the Mojave River before it used to estimate recharge from the river to the flood- exits the ground-water basin. plain aquifer. Estimates of recharge from the Mojave River range from 31,400 to 56,800 acre-ft/yr (table 3). The ever-changing physical character of the Ground-Water Recharge and Discharge Mojave River and the dynamic relationship between the river and the underlying aquifers greatly influence The principal sources of recharge are stream the amount of water exchanged between the two sys- leakage from the Mojave River, infiltration of storm tems. Many factors directly control the quantity, tim- runoff in ephemeral stream channels (termed ing, and distribution of ground-water recharge. These mountain-front recharge in this report), and artificial factors include (1) antecedent soil-moisture conditions;

28 Simulation of Ground-Water Flow in the Mojave River Basin, California Table 3. Estimates of annual recharge to, and discharge from, the Mojave River ground-water basin, southern California, for selected periods [Values in acre-feet per year. na, not applicable] 1930 [from Hardt (1971)] 1963 [from Hardt (1971)] Upper Middle Lower Upper Middle Lower Total Total Mojave1 Mojave2 Mojave3 Mojave1 Mojave2 Mojave3 Recharge Net stream leakage 11,550 9,800 10,050 31,400 20,600 13,490 11,650 45,740 Mountain front 9,550 1,100 350 11,000 9,550 1,100 350 11,000 Artificial Septic and sewage effluent 000000 00 Imported water 000000 00 Flow between subareas4 0 4,500 1,400 na 0 4,500 1,400 na Total...... 21,100 15,400 11,800 42,400 30,150 19,090 13,400 56,740 Storage 000010,700 27,060 16,900 54,660 Discharge Net pumpage 000025,400 36,300 20,600 82,300 Evapotranspiration Transpiration 16,600 11,500 8,500 36,600 10,950 6,950 6,400 24,300 Dry lakes 0 2,500 1,200 3,700 0 1,500 1,200 2,700 Underflow at Afton Canyon 0 0 2,100 2,100 0 0 2,100 2,100 Flow between subareas4 4,500 1,400 0 na 4,500 1,400 0 na Total...... 21,100 15,400 11,800 42,400 40,850 46,150 30,300 111,400

1937–61 Average [from California Department of Water 1931–90 Average [from Mojave Basin Area Watermaster Resources (1967)] (1996), table C-1] Upper Middle Lower Upper Middle Lower Total Total Mojave1 Mojave2 Mojave3 Mojave1 Mojave2 Mojave3 Recharge Net stream leakage 28,380 13,808 12,642 54,830 27,700 23,300 5,800 56,800 Mountain front5 9,846 1,896 1,272 13,014 9,700 0 400 10,100 Artificial Septic and sewage effluent 0 0000 0 0 0 Imported water 250 0 0 250 1,500 0 0 1,500 Flow between subareas 0 2,000 2,000 na 0 2,000 1,200 na Total...... 38,476 17,704 15,914 68,094 38,900 25,300 7,400 68,400 Storage 2,352 4,108 3,180 9,640 33,600 6,600 34,000 74,200 Discharge Net pumpage 16,728 12,494 7,308 36,530 65,400 29,800 39,900 135,100 Evapotranspiration Transpiration 22,100 7,318 11,786 41,204 5,100 900 1,500 7,500 Dry lakes 0 0000 0 0 0 Underflow at Afton Canyon 0 0000 0 0 0 Flow between subareas 2,000 2,000 0 na 2,000 1,200 0 na Total ...... 40,828 21,812 19,094 77,734 72,500 31,900 41,400 142,600

1 Upper Mojave includes the Este, Oeste, and Alto (including Transition zone) subareas. 2 Middle Mojave includes the Centro (including the Harper Lake area) subarea. 3 Lower Mojave includes the Baja (including the Coyote Lake and Afton Canyon areas) subarea. 4 1930 flow values are estimated such that the hydrologic budget for each subarea balanced. Hardt (1971) did not report interzonal flow values for 1963; they are assumed to be unchanged from 1930 values. 5 Mountain-front recharge includes ungaged surface water and deep percolation of precipitation.

Ground-Water Hydrology 29 (2) the width and permeability of the streambed; (3) the Annual recharge for 1931–94 averaged about magnitude, frequency, and duration of runoff; and 46,000 acre-ft in the Alto subarea and about (4) the volume of the unsaturated zone in the 39,000 acre-ft for the combined area of the Transition underlying floodplain aquifer. Infiltration of water zone and the Centro subarea (Lines, 1996, table 3). For through the streambed, although related to the physical the 44 years that the annual recharge could be esti- attributes of the streambed materials (porosity and ver- mated for the Baja subarea (1931–32 and 1953–94), tical hydraulic conductivity), is primarily a function of average annual recharge is about 11,000 acre-ft (1) the length of time that the channel contains water, (Lines, 1996, p. 33). (2) the total area of the channel that is wetted, and (3) whether the streambed has been prewetted by anteced- Mountain-Front Recharge ent flows (Durbin and Hardt, 1974, p. 14). With the exception of flows of very large magnitude, the dis- Recharge resulting from the infiltration of storm tance that surface water may flow is dependent on pre- runoff in ephemeral stream channels from the sur- ceding storms and the moisture content of the rounding mountains and highlands is termed mountain- unsaturated zone below the river. Reaches of the river front recharge for this report. Most mountain-front that are underlain by a thick unsaturated zone are capa- recharge occurs during wet years as storm runoff infil- ble of receiving more water. Areas along the river that trates the alluvial fan deposits of the regional aquifer receive the largest quantities of recharge are ephemeral located in the upper reaches of ephemeral streams and reaches within the Alto and Centro subareas. In some washes that lie between the headwaters of the Mojave areas, the water table is relatively deep, such as in Hin- River and Sheep Creek (fig. 1) (Izbicki and others, kley Valley in the Centro subarea. Following the record 1995). Recharge also may occur at the southern edge of high discharge in the Mojave River in the winter of the Este subarea, particularly in the area just west of the 1993, one well in Hinkley Valley had a water-level rise Helendale Fault. Near the mountain front, water infil- of almost 80 ft. Even though much of the aquifer in the trates the unsaturated zone, which is more than 1,000 ft Baja subarea between Daggett and Camp Cady is thick in places and consists of alternating layers of unsaturated, recharge is relatively small mainly owing gravel, sand, silt, and clay (Izbicki and others, 1995). to the presence of fine-grained, low vertical permeabil- The low, unsaturated hydraulic conductivities of the ity materials in the streambed and subsurface (Lines, fine-grained deposits, which range from about 1 to 1996, p. 40). 3 ft/yr, result in lateral spreading of the recharge and Recharge to the floodplain aquifer from infiltra- slow downward infiltration velocities (Michel, 1996). tion of Mojave River water was computed from mea- Caliche (calcrete) deposits that are near land surface in sured streamflow losses between gaging stations and much of the Alto subarea prevent the percolation of estimates of tributary inflow, base flow, anthropogenic rainfall and runoff from washes. discharges, and evaporation of river water between Tritium and chloride data collected from sites in gages (Lines, 1996, p. 31). Figure 12 shows the annual a wash about 9 mi west of the Mojave River indicate recharge estimated by Lines (1996, p. 32) to the flood- that most water entering the wash infiltrates the upgra- plain aquifer in the Alto subarea, the combined Transi- dient sites closer to the runoff source and almost no tion zone and Centro subarea, and the Baja subarea. It water infiltrates the downgradient sites (Michel, 1996). was not possible to distinguish separate recharge esti- Data from sites at the lower elevations of the wash indi- mates for the Transition zone and the Centro subarea cate that during periods of flow almost no water infil- because there is no gaging station at their boundary. trates and recharges the regional aquifer but is carried Although the Alto, Transition zone, and Centro subar- to the Mojave River as tributary inflow (tributary eas receive yearly recharge, recharge in the Baja sub- recharge is discussed in section titled “Ungaged area occurs only during years when flows are very large Tributary Streams”). Carbon-14 data collected from in magnitude. Recharge in the Alto, Transition zone, wells perforated in the lower parts of the regional aqui- and Centro subareas was comparable until the early fer in the Alto subarea suggest that some of the ground 1950’s. Since then, several thousand acre-feet of water water was recharged more than 20,000 years before from the Mojave River fish hatchery, operated by the present and that only minimal recharge occurs in the California Department of Fish and Game, has been lower reaches under present climatic conditions discharged annually to the river in the Alto subarea. (Izbicki and others, 1995).

30 Simulation of Ground-Water Flow in the Mojave River Basin, California The amount of discharge from these ephemeral connected to a municipal sewer system. Note that streams and washes has never been measured directly; previous researchers addressed only imported water as therefore, it is uncertain how much water infiltrates an artificial recharge because net values of pumpage their upper reaches to recharge the regional aquifer. were estimated. Estimates of total mountain-front recharge range from about 10,100 to 13,000 acre-ft/yr with most of the Irrigation-Return Flow recharge occurring in the Upper Mojave Basin (Oeste, Alto, and Este subareas) (table 3). Historically, the most significant component of ground-water discharge has been pumpage for agricul- ture (Hardt, 1971, p. 45). Depending on irrigation prac- Artificial Recharge tices and soil type, some of the water that is pumped Several sources provide artificial recharge to the and applied to crops returns to the ground-water sys- basin, including irrigation-return flow, fish hatchery tem; this is termed irrigation-return flow. Water that return flow, treated sewage and septic effluent, and does not return to the water table and is lost through imported water. With the exception of septic-tank dis- plant use and evaporation is considered net pumpage. charge, these sources discharge directly into, or adja- Net pumpage is a function of the consumptive use cent to, the river. The disposal of septic wastewater has applied to the total agricultural pumpage. Consumptive become a significant source of recharge to the aquifer use for agriculture is defined as the unit amount of in the Alto subarea where many residences are not water used on a given area in transpiration, building of

500,000

450,000

Baja subarea 400,000 Transition zone and Centro subarea

350,000 Alto subarea

300,000

250,000 No record 1933-52 200,000 (Baja subarea) RECHARGE, IN ACRE-FEET 150,000

100,000

50,000

0 1931 34 37 40 43 46 49 52 55 58 61 64 67 70 73 76 79 82 85 88 91 1994 YEAR

Figure 12. Estimated annual recharge to the Mojave River floodplain aquifer within the Alto subarea for 1931–94, within the Transition zone and Centro subarea combined for 1931–94, and within the Baja subarea for 1931–32 and 1953–94, Mojave River ground-water basin, southern California (Modified from Lines, 1996).

Ground-Water Hydrology 31 plant tissue, and evaporation from adjacent soil (Erie Valley (fig. 6) by the Crestline Sanitation District, and others, 1965, p. 5). Hardt (1971, p. 48) estimated which is located upgradient of the gage on West Fork; that the average consumptive use of total pumpage is this discharge is included in the streamflow measure- 40 to 45 percent and that irrigation-return flow is 55 to ments at the gage. Effluent from the Lake Arrowhead 60 percent. Modern farming methods have increased Community Services District near Hesperia is used for the efficiency of irrigation resulting in decreased irrigation near the Mojave River. In 1996, about irrigation-return flow rates. Estimates of irrigation- 1,450 acre-ft was applied to alfalfa (Ken Nelson, Lake return flow range from 46 percent in the Alto subarea to Arrowhead Community Services District, oral com- 29 percent in the Baja subarea (Robert Wagner, James mun., 2000), most of which was consumptively used C. Hanson Engineering, written commun., 1996). through transpiration; therefore, this source is not con- sidered a significant source of recharge in the Alto Fish Hatchery Discharge subarea. Two fish hatcheries, the Jess Ranch and the Treated wastewater from VVWRA, which has Mojave River Fish Hatchery, are adjacent to the been in operation since December 1981, is discharged Mojave River in the Alto subarea (fig. 6). These hatch- directly to the river through a pipeline about 3 mi eries pump ground water for circulation through fish- downstream from the Lower Narrows in the Transition rearing ponds; the effluent is used for irrigation on the zone (fig. 6). Metered discharge through the pipeline floodplain. Any excess effluent is discharged directly increased from about 2,680 acre-ft in 1982 (Neal B. into the river. The privately owned Jess Ranch Fish Allen, Victor Valley Wastewater Reclamation Hatchery, about 9 mi downstream from The Forks, has Authority, written commun., 1994) to about 7,450 been in operation since 1951. Fish-hatchery effluent acre-ft in 1999 (table 4) (Christine Nalian, Victor was reported to be used to irrigate nearby alfalfa fields Valley Wastewater Reclamation Authority, written between 1951 and 1990 (Gary Ledford, Jess Ranch commun., 2000). Effluent also has been discharged to Fish Hatchery, oral commun., 1996). Fish-hatchery six treatment ponds with flows ranging from about 420 effluent has been discharged directly into the river since acre-ft in 1982 to a high of about 1,000 acre-ft in 1997, 1990. Based on periodic discharge measurements and averaging about 530 acre-ft/yr for 1981–99. Discharge reported operations, estimated intermittent discharge values for VVWRA’s ponds (table 4) are based on total ranged from about 2,000 to 7,000 acre-ft/yr during reported flows minus free surface evaporation. The 1990–94 (table 4) (Lines, 1996, p. 20). California Department of Water Resources reports a The California Department of Fish and Game pan evaporation rate of 65.8 in./yr for 1995 in the has operated the Mojave River Fish Hatchery about Victorville area (David Inouye, California Department 10 mi downstream from The Forks since 1949. Ground of Water Resources, written commun., 1996). On the water pumped on site is circulated through the hatchery basis of about 2.3 acres of surface area of VVWRA’s and is discharged directly to the river. Until about 1994, ponds, evaporation is about 22 acre-ft/yr. Aside from 3,000 acre-ft/yr was diverted and used for irrigation the little discharge that is lost through evaporation, all (Lines, 1996, p. 20). Estimated fish-hatchery discharge wastewater effluent discharged to the ponds was ranged from about 300 acre-ft/yr in 1949 to about assumed to percolate to the ground-water system. 15,000 acre-ft/yr in the late 1970’s and mid 1980’s In the lower part of the Centro and the upper part (table 4). of the Baja subareas, effluent that percolates to the ground-water system is a significant source of Treated Sewage Effluent recharge. Historically, there have been six sources of Sewage effluent from several sources within the effluent recharge in the Barstow area: (1) the city of study area contribute recharge to the ground-water sys- Barstow upper sewage ponds, (2) the city of Barstow tem. Treated sewage from the Alto subarea at Victor lower sewage ponds, (3) the city of Barstow effluent- Valley Wastewater Reclamation Authority (VVWRA) irrigated field, (4) the USMC sewage ponds at Nebo is discharged in the Transition zone (fig. 6). In the Baja Annex, (5) the irrigation of treated wastewater from the subarea, however, treated sewage is discharged to treat- Nebo Annex ponds on the base’s golf course, and ment ponds or used for irrigation by the city of Barstow (6) the USMC sewage ponds at Yermo Annex (fig. 6). and the U.S. Marine Corps (USMC). Effluent is Sources 1 through 5 are located in the Centro subarea discharged to the Rancho Los Flores Ranch in Summit and source 6 is located in the Baja subarea. Estimated

32 Simulation of Ground-Water Flow in the Mojave River Basin, California site Hodge (56, 75) (Dec. 1999) site (49, 81) Lenwood (Feb. 1999) ater 11

W Mojave Agency pipeline (138, 65) Morongo 5 18 Fish Jess Ranch (127, 62) Hatchery 5 20 Fish River Mojave (125, 62) Hatchery San Area, (Silver Lakes) (71, 52) County County, Service Zone 70 Bernardino 4 30 sewage astewater (100, 48) pipeline Authority ictor Valley ictor Valley W V Reclamation 4 (99, 47) ponds sewage astewater Authority ictor Valley ictor Valley W V Reclamation 1,3 U.S. ermo Y Corps, Annex Marine sewage (51, 121) ponds 1,3 U.S. Corps, of total Marine (52, 108) applied) course Nebo golf (50 percent 1,3 410 410 410 410 U.S. Nebo 6 6 6 6 Corps, Annex Marine sewage (52, 109) ponds 2 field City of cial recharge. Numbers in parentheses are model cell numbers. Number in bold italics corresponds to tributary segment number in segment Numbers in parentheses are model cell numbers. Number bold italics corresponds to tributary cial recharge. (49, 103) Barstow irrigated effluent- fi g. 22). Values in acre-feet. —, no data] Values g. 22). fi 1,2 lower City of Railway sewage (52, 103) (52, 104) (52, 105) ponds Barstow Santa Fe opeka, and T and Atchison, 1 550 550 550 550 550 550 550 550 upper City of waste- ponds 6 6 6 6 6 6 6 6 effluent Railway (48, 100) Barstow Santa Fe Atchison, opeka, and T sewage and w-routing package of model ( Sources and quantity of artificial recharge along the Mojave River, southern California, 1938–99 o See footnotes at end of table. fl gure 6 for location of sources arti fi ear Y 1939 1940 1941 1942 1943 1944 1945 19461947 5501948 5501949 5501950 5501951 5501952 5501953 6401954 7501955 7501956 750 410 1957 750 410 1958 750 410 1959 860 410 1,200 410 410 410 463 463 463 463 463 610 350 150 300 900 700 6,000 9,000 8,000 7,000 8,000 9,000 9,000 8,000 1938 Table 4. [See stream

Ground-Water Hydrology 33 site Hodge (56, 75) (Dec. 1999) site (49, 81) Lenwood (Feb. 1999) ater 11 W Mojave Agency pipeline (138, 65) Morongo 5 18 Fish Jess Ranch (127, 62) Hatchery 5 20 Fish River Mojave (125, 62) Hatchery San Area, (Silver Lakes) (71, 52) County County, Service Zone 70 Bernardino 4 30 sewage astewater (100, 48) pipeline Authority ictor Valley ictor Valley W V Reclamation 4 (99, 47) ponds sewage astewater Authority ictor Valley ictor Valley W V Reclamation 1,3 707070 9,000 9,000 68 10,000 6868686868 2,000 4,000 15,000 15,000 13,000 15,000 U.S. ermo 6 6 6 6 6 6 6 6 6 110110110110110110110110110110110 10,000 10,000 10,000 9,000 9,000 9,000 8,000 2,000 3,000 9,000 1,000 Y Corps, Annex Marine 6 6 6 6 6 6 6 6 6 6 6 sewage (51, 121) ponds 1,3 U.S. Corps, of total Marine (52, 108) applied) course Nebo golf (50 percent 1,3 480480480 — 403 — 403 — 403 — 403 — 403 — 403 — — — U.S. Nebo 6 6 6 6 6 6 6 6 6 Corps, Annex Marine sewage (52, 109) ponds 2 field City of (49, 103) Barstow irrigated effluent- . 1,2 lower City of 1,860 1,860 1,734 1,750 1,795 1,795 1,687 2,126 2,312 Railway sewage (52, 103) (52, 104) (52, 105) ponds Barstow Santa Fe 6 6 6 6 6 6 6 6 6 opeka, and T and Atchison, 1 upper City of waste- ponds effluent Railway (48, 100) Barstow Santa Fe Atchison, opeka, and T sewage and Sources and quantity of artificial recharge along the Mojave River, southern California, 1938–99—Continued See footnotes at end of table ear Y 19601961 1,200 1,200 350 350 150 150 9,000 19621963 1,2001964 1,2001965 1,2001966 1,2001967 1,2001968 1,2001969 1,200197019711972 3501973 350 3,0001974 150 350 1,8001975 150 350 1,8001976 150 3501977 150 3501978 150 3501979 150 3801980 150 4801981 150 4801982 150 1983 150 1984 2,223 2,239 1,788 1,478 285 421 265 364 318 374 7 26 35 29 120 110 87 59 3 422 906 967 2,683 2,550 262 3,032 12,000 13,000 11,000 15,000 Table 4.

34 Simulation of Ground-Water Flow in the Mojave River Basin, California site 1,000 Hodge (56, 75) (Dec. 1999) 12 site 2,700 (49, 81) Lenwood (Feb. 1999) 10 300 5,000 4,500 2,100 7,100 2,200 ater 11 W Mojave Agency pipeline (138, 65) Morongo 10, 11 10, 11 10, 11 10, 11 10, 11 10, 11 5 18 Fish Jess Ranch (127, 62) Hatchery 5 20 Fish River 8,0007,0005,000 0 6,000 0 6,000 0 0 0 Mojave (125, 62) eline; and the Lenwood and Hodge sites are from the Mojave and Hodge sites are from the Mojave eline; and the Lenwood Hatchery 10 10 10 10 10 r only. uent volumes from preceding months. uent volumes fl ata were available). 450 350 400 480 520 San Area, (Silver Lakes) (71, 52) County County, Service Zone 70 10 10 10 10 10 Bernardino 4 30 sewage astewater (100, 48) pipeline Authority 7,300 7,970 7,843 8,080 7,450 ictor Valley ictor Valley W V Reclamation 9 9 9 9 9 4 530 120 130 800 (99, 47) ponds 9 9 sewage astewater 9 9 Authority 1,000 ictor Valley ictor Valley 9 W V Reclamation 1,3 96 96 96 96 96 U.S. ermo 8 8 8 8 8 Y Corps, Annex Marine sewage (51, 121) ponds 1,3 U.S. 23 140 470 5,259 11,000 72 72 72 72 72 7 Corps, of total Marine 8 8 8 8 8 (52, 108) applied) course Nebo golf (50 percent 1,3 U.S. 490 490 490 490 490 Nebo Corps, Annex Marine 8 8 8 8 8 sewage (52, 109) ponds 2 field 371 371 371 371 371 City of (49, 103) 8 8 8 8 8 Barstow irrigated effluent- 1,2 lower City of Railway sewage (52, 103) (52, 104) (52, 105) ponds Barstow 1,823 1,823 1,823 1,823 1,823 Santa Fe opeka, and 8 8 8 8 8 T and Atchison, 1 upper City of 550 550 550 550 550 waste- ponds effluent Railway (48, 100) Barstow 8 8 8 8 8 Santa Fe Atchison, opeka, and T sewage and Sources and quantity of artificial recharge along the Mojave River, southern California, 1938–99—Continued Data for 1938–71 from Robson (1974, p. 15) and Hughes (1975, 9–15). written commun., 1995). Data after 1973 from John Brand (City of Barstow, Cox (U.S. Marine Corps, Nebo, written commun., 1995). Data after 1980 from Peter Barella and Mike written commun., 1995). Authority, Reclamation Wastewater Valley Allen (Victory Data from Neal B. Data from Lines (1996, p. 20). Estimated. ef estimated using sewage was The quantity of fresh water applied during some months from 1988 through 1991. was Fresh water assumed to be the same as 1994 (the last year d and for the U.S. Marine Corps 1995–99 was for the city of Barstow Discharge Authority (Christine Nalian, written commun., 2000). Reclamation Wastewater Valley Victor for 1995–99 is from the Discharge Morongo pip Agency Water Fish Hatchery; the Mojave River Area, Zone 70; the Mojave for 1995–99 the County Service Discharge Morongo pipeline is during summer only. Agency Water for 1994 the Mojave Discharge written commun., 2000) during summe Wiegenstein, (Valerie Agency Water for 1999 the Hodge site is from Mojave Discharge 1 2 3 4 5 6 7 8 9 10 11 12 ear Y ater Agency (Valerie Wiegenstein, written commun., 2000). Wiegenstein, (Valerie Agency ater 1995 1996 1997 1998 1999 19891990199119921993 1,4381994 1,502 1,521 429 1,751 437 2,045 388 357 1,823 378 365 311 294 23 371 491 48 435 74 104 490 55 92 73 111 828 72 46 75 72 76 5,707 96 711 7,067 563 7,177 686 6,703 6,800 7,130 7,000 6,000 11,000 2,000 10,000 7,000 10,000 7,000 2,000 7,000 0 1985198619871988 1,647 1,270 1,630 383 1,548 524 428 482 421 296 388 26 369 4 8 80 53 122 757 619 338 3,399 3,683 4,395 15,000 6,000 7,000 Table 4. W

Ground-Water Hydrology 35 effluent recharge values in table 4 represent total flow water in the Mojave River does not reach to the ponds minus any surface-water evaporation. In Barstow—sewage effluent, though small in quantity, is nearby Newberry Springs, the California Department the only source of recharge to the Baja subarea other of Water Resources has reported pan evaporation rates than irrigation-return flow. of 79.18 in./yr (David Inouye, California Department of Water Resources, oral commun., 1996). The effluent Septic Systems from the Atchison, Topeka, and Santa Fe Railway at Although the VVWRA’s sewage treatment plant Barstow, now owned by the Burlington-Northern Rail- has been operational since December 1981, domestic way, is included in the data for the upper and lower Bar- wastewater in the Alto subarea is disposed of predomi- stow sewage ponds (table 4); data for 1938–71 are from nantly by septic systems, which have become a signif- Robson (1974). This area has been studied extensively icant source of recharge to the Alto subarea since the because the ground water has been contaminated by region was first settled. Septic recharge has been insig- industrial wastes and municipal sewage that has perco- nificant in other areas of the basin because housing lated into the floodplain aquifer (Robson, 1974; density has been low or because sewage-treatment Hughes, 1975). Effluent discharge at the Barstow plants have been operational (see “Treated Sewage ponds ranged from about 550 acre-ft/yr for 1938–51 to Effluent” section). In 1990, there were about 46,000 about 3,000 acre-ft/yr in 1969. Effluent discharge for residences in the Alto subarea near Victorville that dis- 1938–45, 1972–80, and 1995–99 were estimated for posed of domestic wastewater by discharging septic- Barstow’s upper and lower sewage ponds because no tank effluent to seepage pits (dry wells) (Umari and data were available. Effluent discharges also were esti- others, 1995, p. 2). Wastewater from the seepage pits mated for the Nebo Annex sewage ponds for 1942–45, percolates into the unsaturated zone. Rates for the ver- 1972–80, and 1995–99 and for the Yermo Annex sew- tical movement of water in the unsaturated zone from age ponds for 1961–80 because records for these years moisture profiles compiled by Umari and others (1995) were unavailable (Mike Cox, U.S. Marine Corps, oral indicate that wastewater from many of the disposal sys- commun., 1995). The total amount of estimated tems in the Alto subarea has reached the water table. recharge from sources of sewage effluent in 1999 was On the basis of population figures, the estimated about 2,760 acre-ft/yr in the Centro subarea and about recharge from wastewater in 1930 was about 100 acre-ft/yr in the Baja subarea. The estimated 200 acre-ft (table 5). The quantity of wastewater reach- amount of recharge from sewage effluent in the Centro ing the water table in 1990 was estimated to have subarea is about 15 percent of the total recharge from increased to about 9,980 acre-ft, about 18 percent of the the Mojave River estimated by Lines (1996, p. 32) for total recharge to the aquifer for that year (Umari and 1931–94. For years of low flow—years when surface others, 1995, p. 8).

Table 5. Population and estimated recharge from septic systems in the Alto subarea of the Mojave River ground-water basin, southern California, 1930–90 [See figure 13 for land use and distribution of septic tanks. Population estimates for 1970, 1980, and 1990 from the California Department of Finance, accessed November 28, 1998. —, no data]

Population City 1930 1940 1950 1960 1970 1980 1990 Adelanto ————2,115 2,164 8,517 Apple Valley ————6,702 14,305 46,079 Hesperia ————4,592 13,540 50,418 Victorville ————10,845 14,220 40,674 Total population ...... 12,650 13,250 18,400 125,000 24,254 44,229 145,688 Recharge (acre-feet) ...... 210 250 660 1,940 1,870 3,500 9,980

1Population estimate from California Department of Water Resources (1967, p. 65).

36 Simulation of Ground-Water Flow in the Mojave River Basin, California Estimates of recharge from septic systems for data were not available for years prior to 1983, we 1930–90 in table 5 were based on an average septic- based the historical distribution of septic recharge on tank discharge of 70 gal/d per person (Umari and oth- areas of residential land use and corresponding histori- ers, 1995, p. 8) and an assumed population density of cal population estimates. The increase in population four people per acre. Residential land-use data from since 1930 has lead to an expansion of residential land Southern California Edison (1983) were used to deter- use in the Alto subarea and, subsequently, an increase mine the areas with septic systems. Because land-use in recharge from septic systems (fig. 13). The historical

117 30' 117 116 30' EXPLANATION 1930 Victorville 35 15 Mojave River ground- 00' water basin 18 Apple Barstow Valley Newberry Helendale Springs Wastewater dispersement method – Area of map Mojave Adelanto Apple Septic tank 34 Victorville Valley 15

30' Hesperia Sewer Hesperia River

Mojave River ground-water basin 0510 MILES

0510 KILOMETERS

1940 1950 1960 Victorville Victorville Victorville Apple Apple Apple Valley Valley Valley

Mojave Mojave Mojave

Hesperia Hesperia Hesperia

River River River

1970 1980 1990 Victorville Victorville Victorville Apple Apple Apple Valley Valley Valley

Mojave Mojave Mojave

Hesperia Hesperia Hesperia

River River River

Figure 13. Distribution of septic and sewer systems in the Alto subarea, Mojave River ground-water basin, southern California, for selected years between 1930 and 1990. (See table 5 for population estimates.)

Ground-Water Hydrology 37 distribution of septic systems for the subarea was esti- Discharge mated by compiling the historical population estimate and by assuming a distribution for that population den- Pumpage sity starting in the older parts of towns and communi- Ground-water development in the study area ties. As the population increased, areas with septic started before the late 1880’s; Native Americans, pio- systems expanded (fig. 13). For this study, population neers, and early explorers dug shallow wells along the and the distribution of septic systems were assumed to Mojave River for their water needs. Ground-water remain constant for 10-year increments until new cen- pumpage in the region has increased with the popula- sus data updated the previous population estimates and, tion and has significantly affected the ground-water consequently, the extent of the septic recharge in the system since the early 1900’s. Pumpage data were not subarea increased. The population estimates applied to recorded before 1931, but about 30 wells reportedly were constructed along the Mojave River in the Alto the 5 years prior to and the 5 years after a reported year subarea by 1917 (Thompson, 1929). The wells were because the estimates were available only for the end of used to irrigate about 5,500 acres of mostly alfalfa. each decade. For example, population estimates for Thompson (1929) suggested that ground-water pump- 1950 were used for the period 1945–55. ing could have resulted in temporary declines in the The VVWRA sewage treatment plant was con- water table along the river in the Mojave Valley during structed in response to the failure of older septic sys- the early 1920’s, a period following several years of tems; it became operational in December 1981. After drought and the absence of recharge from the river. that time, areas on sewer systems (fig. 13) began Pumpage data were compiled for 1931–99 for sending wastewater to VVWRA, located in the Transi- this study (fig. 14). Except for municipal, military, and tion zone. Once an area was connected to a sewage sys- industrial wells, most wells in the Mojave River tem, the area was excluded from calculations for septic ground-water basin have never been metered; there- fore, pumpage estimates are based on data collected recharge (fig. 13). The amount of estimated septic from many sources, including previous studies (Dibble, recharge for 1990, therefore, is disproportionately 1967; Hardt, 1971), reported data (Mike Cox, U.S. lower than expected for the reported population with Marine Corps, written commun., 1994), field surveys, respect to previous decades (table 5). and indirect methods such as electric power consump- tion and water requirements of irrigated crops. An Imported Water assumed water-use rate of 7.0 ft was used to calculate total pumpage because alfalfa is the most extensive Imported water has been released periodically crop in the study area (Robert Wagner, James C. from Silverwood Lake to the West Fork Mojave River Hanson Engineers, written commun., 1996) (fig. 14). since February 1972. Through 1994, these releases For the years of the study period with missing or have totaled about 70,000 acre-ft (Lines, 1996, p. 21) incomplete data, pumpage values were extrapolated and are included in the flows measured at the West Fork using a Geographic Information System (GIS) to inter- gaging station (10260950). Except for a short period in polate pumpage between known values. Note that March 1983 when water flowed past Afton Canyon and Hardt (1971) reported net pumpage data; therefore, we out of the basin, all this water percolated into the applied an assumed consumptive use of 40 percent to Mojave River streambed primarily in the Alto subarea. the net pumpage values in the Alto subarea and 50 per- cent in all other subareas to estimate total pumpage Beginning in 1994, water also has been released from values for 1931–50. the California State Water Project (SWP) at the Mojave Initially, wells were constructed near the Mojave Water Agency’s Morongo Basin pipeline turnout in the River, but over time, the distribution of ground-water Alto subarea, which is about 4 mi downstream from pumpage spread to areas away from the river (fig. 15). The Forks (fig. 6). A total of about 21,200 acre-ft of Ground water was used primarily by agricultural and water was released from the turnout from August 1994 municipal and industrial users (fig. 16). In 1931, esti- to 2000 (table 4) (Norman Caouette, Mojave Water mated ground-water pumpage was about 40,000 acre-ft Agency, written commun., 2000). for the Mojave River ground-water basin (fig. 14), most

38 Simulation of Ground-Water Flow in the Mojave River Basin, California of which was used primarily for agriculture (fig. 16). in significant declines in the water table. Water levels in By the mid-1950’s, ground-water pumpage was about the Alto subarea have declined between 50 and 75 ft 190,000 acre-ft (fig. 14). This large increase in pump- since the mid-1940’s, about 100 ft in the Harper Lake age coincided with the widespread use of high- region in the Centro subarea since the early 1960’s, and capacity, deep-well turbine pumps for agriculture. almost 100 ft in the Mojave Valley in the Baja subarea Pumpage increased again in the 1970’s and through the since the early 1930’s (fig. 17). A possible consequence mid-1980’s, peaking at about 240,000 acre-ft. In the of ground-water pumping and, therefore, of water-level mid-1990’s, there was a substantial decrease in pump- decline is land subsidence. In alluvial aquifer systems, age to a low of about 150,000 acre-ft in 1998 (fig. 14). especially those that include relatively thick semi con- This reduction in pumpage coincided with the Physical solidated silt and clay layers, long-term ground-water- Solution of 1993 (Mojave Basin Area Watermaster, level declines can result in a one-time release of water 1996a). from compacting silt and clay layers, which results in By 1994, about half of the pumpage came from land subsidence (Galloway and others, 1999). wells located away from the river (fig. 15); therefore, a Ground-water pumping of the floodplain aquifer large quantity of ground water was withdrawn from the induces increased recharge to the ground-water system regional aquifer. Most of the water was pumped by from the Mojave River where streamflow is available. municipal suppliers. Wells perforated in the regional However, the increased amount of ground-water aquifer generally are drilled deeper and recover more recharge from the river in upstream reaches causes a slowly than wells in the floodplain aquifer, and they depletion in streamflow, thereby reducing the amount receive little, if any, local recharge which has resulted of streamflow available for ground-water recharge to

260,000

240,000

220,000

200,000

180,000 Source of pumpage data: 160,000 James C. Hardt (1971) Hanson 140,000 Engineers (including 120,000 Harper Lake area; Dibble (1967) not including incomplete 100,000 for some TOTAL PUMPAGE, IN ACRE-FEET Harper Lake area Mojave Estimated (as part years) Basin area 80,000 of this study) R. Wagner, written watermaster, commun., Mojave Water 60,000 1996 Agency (written commun., 40,000 2000)

20,000

0 1931 1934 1938 1942 1946 1950 1954 1958 1962 1966 1970 19741978 1982 1986 1990 1994 98 1999 YEAR

Figure 14. Total pumpage and sources of pumpage data for the Mojave River ground-water basin, southern California, 1931–99.

Ground-Water Hydrology 39 1931

Barstow

River

Mojave River ground-water basin Mojave

Southern Adelanto California Apple Logistics Victorville Valley Airport

Hesperia

015 01520 MILES 0 5101520 KILOMETERS 1951

Barstow River

Annual total pumpage – In acre-feet per year (model cell–see figure 18 for Mojave location) > 2,500 Adelanto > 2,000 and ≤ 2,500 Apple Victorville Valley > 1,500 and ≤ 2,000 > 1,000 and ≤ 1,500 > 500 and ≤ 1,000 > and Hesperia 0 ≤ 500 >, greater than ≤, less than or equal to

Figure 15. Distribution of annual total pumpage in the Mojave River ground-water basin, southern California, 1931, 1951, 1971, and 1994.

40 Simulation of Ground-Water Flow in the Mojave River Basin, California 1971

Barstow River

Mojave

Adelanto Apple Victorville Valley

Hesperia

015 01520 MILES 0 5101520 KILOMETERS 1994

Barstow

River

Mojave

Adelanto Apple Victorville Valley

Hesperia

Figure 15.—Continued.

Ground-Water Hydrology 41 downstream reaches. Withdrawals from the ground- Transpiration by Phreatophytes and Hydrophytes water system by both ground-water pumping and The phreatophytes and hydrophytes in the study transpiration by phreatophytes cause depletions in area are limited primarily to the floodplain and adjacent streamflow. The withdrawals from the aquifer may slopes and terraces along the Mojave River channel. cause river water to enter the floodplain aquifer, or they Distinctive associations or communities of native ripar- may “capture” ground water that normally would have ian plants grow in specific hydrologic environments or been discharged to the river. In either case, the net niches in the riparian zone depending on the availabil- effect is the same—a depletion in streamflow (Lines, ity of water and other environmental stresses (Lines 1996, p. 35). and Bilhorn, 1996, p. 4). Predominant plant communi- ties in the riparian zone include phreatophytes such as Evapotranspiration cottonwoods, willows, velvet ash, white alder, baccha- For the purpose of this study, evapotranspiration ris, mesquite, and saltcedar (Lines and Bilhorn, 1996). is the consumptive use of water by riparian plants (tran- Phreatophytes obtain their water supply from the satu- spiration), bare-soil evaporation, and free-surface evap- rated zone (and from shallow ground water) directly or oration. The riparian plants in the study area are by capillary action. Phreatophytes are capable of primarily phreatophytes and hydrophytes. Bare-soil extending their roots to the shallow water table and evaporation occurs primarily at the five dry lakes in the withdrawing water. Hydrophytes are dependent on sur- study area and free-surface evaporation occurs prima- face water for their survival and are limited to the rily in the reach between the Upper and Lower Narrows shoals and banks of the river in reaches where flow is of the Mojave River. perennial.

120,000 Alto subarea Centro subarea 100,000 Agriculture Municipal and industrial 80,000 Other 60,000 Total

40,000

20,000

0 120,000 Transition zone Baja subarea 100,000 PUMPAGE, IN ACRE-FEET 80,000

60,000

40,000

20,000

0 1931 1940 1950 1960 1970 1980 1990 2000 1931 1940 1950 1960 1970 1980 1990 2000 YEAR YEAR

Figure 16. Components of total pumpage by subarea for the Mojave River ground-water basin, southern California, 1931–99.

42 Simulation of Ground-Water Flow in the Mojave River Basin, California Most of the Mojave River floodplain is barren of Camp Cady, now barely support even the heartiest vegetation either because of periodical flooding or desert plants (Lines and Bilhorn, 1996, map). urbanization or because the depth to water is too deep Estimates of evapotranspiration can vary by at to support phreatophytes. In 1995, there were about least threefold depending on the prevailing hydrologic 13,000 acres of barren land in the riparian zone along conditions of the river and changes in riparian habitat the Mojave River; about 12,000 acres of the riparian (Lines, 1996, p. 40). Estimates also may vary because zone had been disturbed and was being used for agri- of the techniques used to determine evapotranspiration. In 1929, before significant ground-water development cultural, residential, and other uses (Lines and Bilhorn, in the area, the California Department of Public Works 1996, p. 6). Urbanization also affects the distribution of (1934) estimated that about 7,800 acres of phreato- phreatophytes because it affects the amount of water phytes consumed 40,000 acre-ft of water. The U.S. that phreatophytes use. In parts of the basin, ground- Bureau of Reclamation (1952) estimated that annual water pumping has lowered the water table below the evapotranspiration from about 11,000 acres of phreato- depth that the roots of most plants can reach. Many phytes, open water, and wetted stream channels con- areas that were once lush with vegetation, such as sumed about 35,000 acre-ft of water. In 1995, Lines upgradient of the Calico-Newberry Fault and near and Bilhorn (1996, p. 8) estimated that 10,000 acres of

2,875 2,000 4N/3W-1M1 11N/4W-29R1 2,850 Alto subarea 1,975 Centro subarea

2,825 1,950

2,800 1,925

2,775 1,900 2,850 1,875 5N/5W-22E1 2,825 5N/5W-22E2 1,850

2,800 1,875 TER-LEVEL ALTITUDE, IN FEET ABOVE SEA LEVEL TER-LEVEL ALTITUDE, WA 2,775 1,850 9N/2E-14N2 Alto subarea 9N/2E-14N1

2,750 1,825 1930 1940 1950 1960 1970 1980 1990 2000 YEAR 1,800

1,775 Baja subarea

1,750 1930 1940 1950 1960 1970 1980 1990 2000 YEAR

Figure 17. Altitude of measured water levels at selected wells in the Mojave River ground-water basin, southern California. (Location of wells shown in figure 29.)

Ground-Water Hydrology 43 riparian vegetation consumed about 17,000 acre-ft of GROUND-WATER FLOW MODEL water. A numerical ground-water flow model of the

Bare-Soil Evaporation from Dry Lakes Mojave River ground-water basin was developed to update the analog model developed by Hardt (1971), to There are five dry lakes in the study area— gain a better understanding of the relations between the Rabbit Lake in the Este subarea, El Mirage Lake in the regional and the floodplain aquifer systems, and to Oeste subarea, Harper Lake in the northern part of the develop a management tool that could be used to esti- Centro subarea, and Coyote and Troy Lakes in the Baja mate the effects that future hydrologic stresses may subarea (fig. 1). The dry lakes act as natural sinks to the have on the ground-water system. As a management local basins. Surface-water ponds after local flooding, tool, this model could be used to simulate ground-water and ground-water discharges to the lake surfaces, evap- conditions based on projected pumpage estimates and orates, and is lost from the ground-water system. also to simulate the effects of variations in natural and Ground-water development in the basin has resulted in artificial recharge in the basin. A numerical model is a change in the ground-water gradients and in the direc- based on assumptions and approximations that tion of ground-water flow toward pumping wells and simplify the actual system and cannot simulate exactly away from the dry lakes. Declining water levels the inherent complexity of the geohydrologic frame- probably have caused a decrease in ground-water work. The results of the model simulation are only an discharge to the dry lakes. approximation or an expectation of actual conditions and are only as accurate or realistic as the assumptions and data used in its development. The limitations of the Free-Surface Evaporation model are discussed later in this report. Lines and Bilhorn (1996, p. 5) estimated that in Hardt (1971) developed a two-dimensional, 1995, the total area of free-surface water and hydro- horizontal, electric-analog ground-water flow model of phytes was about 410 acres of which about 90 percent the Mojave River ground-water basin. The model of the area was free-surface water. The estimated total domain used by Hardt (1971) was the basis for this free-surface evaporation was about 2,200 acre-ft/yr for study. The analog model addressed the regional aquifer 1995 (Lines and Bilhorn, 1996). only and did not address the effects of variable stream- flow on the ground-water system. Hardt (1971, p. 2) Underflow at Afton Canyon concluded that because long-term pumping exceeded natural recharge, the water table was declining, and that Ground-water flows out of the study area only ground-water mining was depleting the aquifer storage. through Afton Canyon. During some years, the shallow The numerical ground-water flow model used bedrock forces ground water to the surface and sustains for this report is the three-dimensional, finite- flow through the narrow canyon. The thin veneer of difference ground-water flow model known as sediments below the streambed may allow some water MODFLOW (McDonald and Harbaugh, 1988). An to pass through as underflow. Because only a few wells explanation of the theoretical development of have been completed in the area and most of those MODFLOW, as well as the solution method and the wells do not have geologic records or construction mathematical basis of the model, is presented in information, it is difficult to estimate the thickness of McDonald and Harbaugh (1988). Additional model the alluvium and any component of underflow. The capabilities were incorporated into MODFLOW to California Department of Water Resources (1967, p. 53 simulate the routing of streamflow (Prudic, 1989) and and 59) reported that no subsurface outflow exits the to simulate faults as horizontal barriers to the flow of study area. Hardt (1971, p. 20) estimated that the recent ground water (Hsieh and Freckleton, 1993). Mojave River alluvium is 50 ft thick and annual under- The modeling process for the current study flow in the alluvium was less than a few hundred involved defining the model grid, model boundaries, acre-feet. Note that the analog model by Hardt (1971, aquifer properties, stream-aquifer interaction, and table 4) indicates 2,100 acre-ft/yr was discharged at recharge and discharge. The model was calibrated Afton Canyon (lower Mojave Basin) (table 3). using a trial-and-error approach. The period

44 Simulation of Ground-Water Flow in the Mojave River Basin, California 1931–94 was used to calibrate the transient-state from northward to eastward. Because of the character- model. Steady-state conditions for 1930 were used to istics of the finite-difference grid, cells representing provide initial conditions for the transient-state simula- smaller areas along the river would have greatly tion. The period 1995–99 was used to validate the increased the total number of cells required for the model. The calibrated model was used to simulate the model. Increasing the number of cells would have effects of proposed management alternatives on the required a substantial increase in computational time Mojave River ground-water basin during a 20-year and computer storage that would have made model drought (1999–2019). calibration unnecessarily cumbersome. The results of the model simulation provide To evaluate the simulated hydrologic budgets, information on probable hydrologic conditions prior to the model grid was divided into nine model subareas in the development of the basin, and aquifer-system layer 1 and eight model subareas in layer 2 (layer 2 was responses to changes in pumpage and recharge that not active in Afton Canyon area of the Baja subarea). have occurred since development began. Results of the The model subareas are subsets of the MWA-defined model calibration, sensitivity analysis, and selected subareas (Oeste, Alto, Este, Centro, and Baja) (fig. 18). simulations provide insight into the conceptualization Layer 1 represents the coarse materials of the of the regional ground-water flow system, as well as the floodplain aquifer, which include the recent Mojave limitations of this current model and potential future River alluvium (Qra) for the Alto, Centro, and Afton refinements. Canyon model subareas and most of the younger Mojave River alluvium (Qya) for the Baja model sub- Model Grid area (figs. 8 and 9). For the area outside the floodplain aquifer, layer 1 is assigned properties of the upper part The finite-difference model is represented by a of the regional aquifer system, which includes the rectangular grid discretized into rows and columns that undifferentiated alluvium (QTu) and the older alluvium form cells. When overlain onto a map of the study area, of the ancestral Mojave River (QToa) (figs. 7–9). The each cell of the model grid represents a small part of the more permeable deposits are grouped into layer 1 to region. Cells that coincide with areas of the aquifer sys- better simulate areas where the Qra and underlying tem are the “active” cells of the model grid. The values deposits are hydraulically separated by low- for the model input parameters assigned to each active permeability deposits (fig. 9, section D-D’). These low- cell represent the average value for each parameter for permeability deposits are also present just upgradient the ground-water system represented by that model of the Upper Narrows in the Alto model subarea and cell. As the cell size increases, the parameter values throughout most of the Transition zone model subarea. describing the actual aquifer properties, which vary The presence of these deposits causes differences in over the cell area, become more generalized. Every water levels in excess of 20 ft between the aquifer sys- active cell in the model area is assigned a value for all tems. These differences were observed in the multiple- necessary model input parameters thereby describing well monitoring sites in the Transition zone model sub- the areal distribution of the aquifer properties. area (wells 7N/5W-24R7, 8) (fig. 9, section D-D′; The finite-difference grid designed for this Appendix 2). model consists of 32,200 cells (161 rows and 200 col- Layer 2 is assigned properties of the younger umns oriented in an east-west and north-south direc- alluvium of the floodplain aquifer (Qya) for all the tion, respectively) for each of the two model layers model subareas, except Baja, and of the older alluvium (fig. 18). The area of active cells differ in each layer. of the ancestral Mojave River (QToa) and the undiffer- There are 9,898 active cells in the upper layer (layer 1) entiated alluvium (QTu) (figs. 8 and 9). In some areas, and 9,315 active cells in the lower layer (layer 2). The the QTu and QToa deposits are not present; for those area represented by each cell is 2,000 by 2,000 ft. Cells areas, layer 2 is assigned properties of the Tertiary vol- representing smaller areas would have allowed for a canic rocks (Tv). Cells in layer 2 are inactive between more detailed approximation of the flow system for the Upper and Lower Narrows, near Helendale south greater areal resolution of stresses, such as those along and northeast of Iron Mountain, west and east of the river; however, this was not possible because the Barstow along the river, and east of Camp Cady river changes course in the middle of the study area, (fig. 18).

Ground-Water Flow Model 45 117 30' 117 15' 117 00' 10 20 30 4050 60 70 80 90 100 110 120 MODEL COLUMN

10 Harper Lake Harper Coyote Lake 20 Lake 15'

35 30 00'

40

Centro Baja 50 IRON MTN Barstow Centro 60

70 River

34 MODEL ROW 45' Helendale 80 Transition zone 90 El Mirage Lake Mojave VVWRA pond 100 VVWRA El Mirage pipeline Lower Narrows Upper Narrows 110 Adelanto Oeste Victorville Southern Apple California Valley Lucerne 34 120 Logistics Lake 30' Airport

130 Fish hatchery Lucerne Alto Rabbit Valley Lake Este Hesperia 140 MWA Pipeline

Dee 150 est Fork p Cr W eek Mojave River

160

Figure 18. Location of model grid and model boundaries and location of horizontal-flow barrier, mountain-front recharge, drain, California. (VVWRA, Victor Valley Wastewater Reclamation Authority; MWA, Mojave Water Agency)

46 Simulation of Ground-Water Flow in the Mojave River Basin, California 116 45' 116 30' 130 140 150 160170 180 190 200 MODEL COLUMN EXPLANATION

Coyote Lake 10 Coyote Model Grid – Lake Afton Canyon Active areas 20 Inactive areas

Model cell description – Baja 30 35 00' Mountain-front recharge Drain

40 Evapotranspiration Camp Stream Cady Sewage (For identification of Baja 50 sites see table 4) Fish hatcheries and imported Troy water Lake 60 Newberry Septic tank recharge Springs 70 Dry lake (playa) 34 Rabbit ash Lake W 45' Kane 80 Boundaries – Model layer 1 Model layer 2

MODEL ROW 90 Model subarea Baja Name of model subarea 100 Horizontal flow barrier (Model fault) 110

120 34 30'

130

0 10 20 MILES

0 10 20 KILOMETERS 140

150

160 evapotranspiration, stream, and artificial recharge cells of the ground-water flow model of the Mojave River ground-water basin, southern

Ground-Water Flow Model 47 Model Boundary Conditions report). The Horizontal-Flow Barrier (HFB) package (Hsieh and Freckleton, 1993) was used to simulate The areal extent of the model coincides roughly these faults as horizontal-flow barriers. The HFB pack- with the ground-water basin boundary (fig. 18). The age allows for the simulation of thin, vertical, low- lateral boundary of layer 1 corresponds roughly to the permeability geologic features that impede horizontal contact between unconsolidated deposits and less per- ground-water flow in either one or both layers. The meable consolidated rocks and consolidated rocks that need to reduce the grid spacing in the region of the are not exposed in some areas but that are very near faults or to use variable grid spacing, which would ground surface (fig. 7). For some areas, such as north increase model size and associated computational of Apple Valley and west of Helendale, the boundary times, was avoided using this package. The faults are was determined from ground-water data (Stamos and approximated as sets of horizontal-flow barriers located Predmore, 1995). The southeastern boundary of the model coincides with, and is defined by, the Helendale on the boundary between pairs of adjacent cells in the Fault, which separates the Mojave River ground-water model grid. The width of the barriers in the model is basin from the Lucerne Basin to the east. Layer 2 of the assumed to be negligible compared with the horizontal model has the same lateral boundaries as layer 1, dimensions of the model cells. The function of each except as noted in the previous section (fig. 18). barrier is to lower the horizontal conductances between General-head boundaries are used to simulate the two adjacent cells. The barriers are defined by a underflow from the Mojave River at Afton Canyon hydraulic characteristic, which is the hydraulic conduc- using the General-Head Boundary (GHB) package tivity of the fault divided by the width of the fault. Each (McDonald and Harbaugh, 1988). The GHB package fault is represented as a horizontal-flow barrier and is used to simulate a source of water external to the Table 6. Hydraulic characteristics of horizontal-flow barriers used model area that either supplies water to, or receives in the model of the Mojave River ground-water basin, southern water from, the model at a rate proportional to the California hydraulic-head differences between the source and the [See figure 7 for location of barriers] model. The constant of proportionality is termed the Hydraulic characteristic, conductance. The general-head boundary controls the Horizontal flow barrier in feet squared per day (fault name) rate at which water is exchanged between the model Layer 1 Layer 2 cell and the external source. Because there often is flow Calico 2.0 × 10–9 2.0 × 10–9 at Afton Canyon, the altitude of the external source was in the floodplain 2.0 × 10–5 2.0 × 10–9 set equal to 1,320 ft which is the altitude of the near Newberry Spring 2.0 2.0 × 10–9 streambed at Afton Canyon. The estimated value of the Waterman 5.0 × 10–3 5.0 ×10–3 general-head boundary conductance (layer 1, row 26, –7 –7 column 197) was 2.0 ft2/s; this value was set such that Waterman E 5.0 × 10 5.0 × 10 –10 –8 the head differences closely matched the streambed Helendale 2.0 × 10 2.0 × 10 1 30 –8 gradient. in the floodplain 1.0 × 10 2.0 × 10 –8 –8 No-flow boundaries are used around and below Mt. General 1.0 × 10 1.0 × 10 –14 –14 the model area to represent the contact with consoli- Iron Mountain 1.0 × 10 1.0 × 10 –7 –7 dated deposits. Although the consolidated deposits are Apple Valley 5.0 × 10 5.0 × 10 not impermeable, the quantity of water contributed by Lockhart, upper 11.0 × 1030 1.0 × 10–6 them is probably negligible. A no-flow boundary was Lockhart, lower –4 –8 also used to simulate the ground-water divide between north of the river 1.0 × 10 1.0 × 10 the Oeste model subarea and Antelope Valley as indi- south of the river 1.0 × 10–8 1.0 × 10–8 cated by the perpendicular water-level contours near in the floodplain 11.0 × 1030 1.0 × 10–8 the boundary of the ground-water basin shown on Shadow Mountains 1.0 × 10–6 1.0 × 10–6 figure 11. Adelanto 1.0 × 10–6 1.0 × 10–6 Of the many faults transecting the basin, 12 were Narrows 1.0 × 10–6 1.0 × 10–6 considered to have a significant effect on the ground- Baja 1.0 × 10–8 (2) water system (see the discussion “Effects of Faulting 1Large value used to ensure no barrier to ground-water flow. on Ground-Water Flow” presented earlier in this 2Layer 2 not present.

48 Simulation of Ground-Water Flow in the Mojave River Basin, California assigned a hydraulic characteristic value (table 6); therefore, it is reasonable to simulate the system using these values were determined by model calibration. constant transmissivity values. The initial distribution of transmissivity used in Aquifer Properties this model was modified from Hardt (1971) and was augmented by transmissivity values estimated from The basic parameters that define the additional single-well aquifer tests and specific- geohydrologic properties of the aquifer are transmis- capacity data collected for this report. The initial esti- sivity, storage coefficient, and leakage between layers. mates were modified during the steady-state and the The values of transmissivity and storage coefficient transient-state simulations of the model until the final estimated by Hardt (1971) for the two-dimensional, distribution of transmissivity for both layers was horizontal analog model were used as initial values in derived (fig. 19). The estimated layer 1 transmissivity this current model; the values were modified using values for the Qra deposits in the floodplain aquifer available field data and model calibration. ranged from 1,000 to 60,000 ft2/d and from 50 to 2,500 ft2/d for the regional aquifer (fig. 19). In general, Transmissivity the estimated transmissivity values for the Qra deposits near the Mojave River (layer 1) were greater than the Transmissivity is the product of hydraulic values estimated by Hardt (1971); however, Hardt’s conductivity and the thickness of the aquifer material (1971) model did not explicitly consider these deposits. through which ground-water flows and, as such, trans- The estimated model transmissivity values for the missivity varies with saturated thickness. Transmissiv- regional aquifer (layer 2) ranged from 300 to ity values were held constant for both layers of this 17,000 ft2/d (fig. 19). The estimated transmissivity model during the entire simulation. When using a con- values for the regional aquifer are in good agreement stant transmissivity, errors are introduced where water- with those estimated by Hardt (1971). level changes are a significant percentage of the total saturated thickness of an unconfined aquifer. Storage Coefficient Water levels are relatively constant along the The storage coefficient values used for layer 1 of Mojave River throughout much of the Alto and Transi- the model initially were assumed to equal the specific- tion zone model subareas, and any water-level changes yield values estimated by Hardt (1971, fig. 8), varying are only a small percentage of the total saturated thick- from 25 percent in the floodplain aquifer in the Alto ness. However, significant water-level declines have model subarea to 12 percent in all areas of the regional occurred along the river in the Centro and Baja model aquifer system. Lines (1996), as part of a study of the subareas which may affect the values of transmissivity. ground-water and surface-water relations along the The version of the streamflow-routing package Mojave River, measured water-level and gravity (Prudic, 1989) used to simulate the Mojave River does changes at selected wells in the floodplain aquifer sys- not simulate the leakage of streamflow into or out of tem. From those measurements, he estimated specific the aquifer system once a model cell underlying the yield; estimates varied from 14 to 39 percent within the stream has gone dry. When model cells underlying the floodplain aquifer. Specific-yield estimates of the stream become dry, they are bypassed when stream- floodplain aquifer are largest within the Alto model flow is reintroduced, and any water in the stream is subarea and generally decreased in a downstream direc- routed to the next active downstream model cell. The tion. (Lines, 1996, p. 23). These values were modified streamflow-routing package allows only upward leak- for the current study during the transient-state calibra- age from the aquifer to the stream. These problems tion of the model; the final distribution is shown in fig- caused the model to become unstable and unable to ure 20. Calibrated specific-yield estimates were slightly converge to a solution. To overcome these problems, higher in the Baja model subarea than the estimates layer 1 is assigned a constant thickness and is not per- reported by Lines (1996), but they were similar to those mitted to go dry. In areas where the regional aquifer reported by Hardt (1971). (represented by both layers 1 and 2 in areas away from The calibrated values for layer 1 of the regional the river) is unconfined, measured water-level changes aquifer were 12 percent, except in the Oeste, western are less than 10 percent of the total saturated thickness; Alto, and Afton Canyon model subareas (fig. 20). The

Ground-Water Flow Model 49 , Model layer 2. B 180 20 MILES 116 30' , Model layer 1. 170 190 200 A a. 160 20 KILOMETERS In feet squared per day 10 150 EXPLANATION 10 140 Newberry Springs 45'

Greater than 50,000 10,000 - 20,000 1,000 - 2,500 250 - 1,000 Less than 250 Model grid Boundary of model layer 1 Boundary of model subarea Transmissivity – 0 0 Valley Lucerne 117 100 110 120 130 Barstow 90 MODEL COLUMN 80 Apple Valley 15' 60 Hesperia 50 70 Victorville Helendale Mojave River Airport Southern Logistics Adelanto California

117 30' 10 20 30 40 ROW MODEL A Areal distribution of transmissivity in the ground-water flow model Mojave River basin, southern Californi 120 110 100 10 20 30 40 50 60 70 80 90 140 130 150 160 35 34 45' 34 30' Figure 19.

50 Simulation of Ground-Water Flow in the Mojave River Basin, California 180 20 MILES 116 30' 170 190 200 160 20 KILOMETERS In feet squared per day 10 150 EXPLANATION 10 140 45' Newberry Springs

17,000 11,900 7,000 1,750-3,500 300-700 Model grid Boundary of model layer 2 Boundary of model subarea Transmissivity – 0 0 Valley Lucerne 117 100 110 120 130 Barstow 90 MODEL COLUMN 80 Apple Valley 15' 60 Hesperia 50 70 Victorville Helendale Mojave River Adelanto

117 30' 10 20 30 40 ROW MODEL B .—Continued. 120 110 100 10 20 30 40 50 60 70 80 90 140 130 150 160 35 34 45' 34 30' Figure 19

Ground-Water Flow Model 51 180 20 MILES 116 30' 170 190 200 southern California. See 160 20 KILOMETERS (Dimensionless) 10 150 EXPLANATION 10 140 Newberry Springs 45'

Model grid Boundary of model layer 1 Boundary of model subarea Specific yield – 0.38 - 0.39 0.26 0.20 - 0.23 0.12 0.05 0 0 Valley Lucerne 117 100 110 120 130 Barstow 90 MODEL COLUMN 80 Apple Valley 15' 60 Hesperia 50 70 Victorville Helendale Mojave River Airport Adelanto Southern Logistics California

117 30' 10 20 30 40 ROW MODEL Areal distribution of specific yield for model layer 1 the ground-water flow Mojave River basin, 120 110 100 10 20 30 40 50 60 70 80 90 140 130 150 160 35 34 45' 34 30' Figure 20. figure 18 for location of model subareas.

52 Simulation of Ground-Water Flow in the Mojave River Basin, California calibrated values for these three model subareas were where -1 significantly lower. This is possibly explained by the Vcont is the leakance between model layers [t ], high percentage of silt and clay in these areas, which B1 is the thickness of model layer 1 (assumed was determined from the geologic samples from equal to 100 ft), multiple-well monitoring sites installed during this B2 is the thickness of model layer 2 (assumed study. equal to 700 ft), Because the storage coefficient for layer 2 had T1 is the transmissivity of layer 1 [L/t], not been estimated during any previous study, it was T2 is the transmissivity of layer 2 [L/t], estimated for this study by multiplying the layer thick- a1 is the vertical-to-horizontal anisotropy of ness by a specific storage value of 1 × 10-6 ft-1. This layer 1, and value is representative of specific storage in most con- a2 is the vertical-to-horizontal anisotropy of fined aquifers (Lohman, 1972, p. 8) and was not varied layer 2. during model calibration. Although the total thickness The distribution of vertical-to-horizontal of the regional aquifer is more than 2,000 ft in some anisotropy for model layer 1 is presented in figure 21. places, an assumed average thickness of 700 ft was The vertical-to-horizontal anisotropy for model layer 2 used to estimate the storage coefficient. was assumed to equal 1:10 for all cells. Adjustments to Vcont were limited primarily to calibration for the Vertical Leakance transient-state conditions and involved adjusting esti- mates of vertical-to-horizontal anisotropy. The initial Vertical leakage of water between layers 1 and 2 estimate of the vertical-to-horizontal anisotropy was occurs whenever there is a vertical hydraulic-head dif- based on the presence and thickness of the silt and clay ference. The rate at which this leakage occurs is layers which was determined from geologic and geo- described by the equation physical logs of wells collected primarily during con- current USGS studies. Calibration was done by comparing simulated hydraulic-head differences QK= v ⋅⋅AH()()2 – H1 ⁄ B , between model layers with measured water-level dif- ferences between aquifers for selected multiple-well where 3 monitoring sites (fig. 10). The calibrated vertical-to- Q is the vertical leakage [L /t], horizontal anisotropy values for layer 1 ranged from KV is the effective value of vertical hydraulic con- 1:10,000 in the Transition zone model subarea, where ductivity between layers [L/t], the presence of clays causes large differences in A is the area of the cell [L2], hydraulic heads, to 1:10 in the regional-aquifer system. H2 is the hydraulic head in layer 2 [L], H is the hydraulic head in layer 1 [L], and 1 Stream-Aquifer Interactions B is the length of the vertical flow path [L]. Streamflow, tributary flow, and artificial recharge The quantity KV/B is referred to as the vertical along the Mojave River was simulated using the leakance term; in this report, it is designated as the Streamflow-Routing package (STR1) developed by leakage between model layers (Vcont). The ground- Prudic (1989). Though not a true surface-water flow water flow model requires that the user specifies the model, the Streamflow-Routing package simulates the term Vcont as input data. Vcont is calculated using the interaction between the river and the ground-water sys- following equation (modified from McDonald and tem, tracks the amount of flow in the river, and permits Harbaugh, 1988, p. 5–13): the river to go dry during certain stress periods in the model. This was helpful in simulating those reaches of the Mojave River that remain dry for long periods 2 V = ------, because surface-water flows are only sporadic. The cont 2 2 B B 1 2 Streamflow-Routing package simulates streamflow ------+ ------T ⋅ a T ⋅ a losses to the ground-water system, as well as stream- 1 1 2 2 flow gains from the ground-water system. The

Ground-Water Flow Model 53 180 20 MILES 116 30' 170 190 200 hern California. 160 20 KILOMETERS 10 150 EXPLANATION 10 140 Newberry Springs 45'

0.1 0.001 0.0001 Anisotropy Model grid Boundary of model layer 1 Boundary of model subarea 0 0 Valley Lucerne 117 100 110 120 130 Barstow 90 MODEL COLUMN 80 Apple Valley 15' 60 Hesperia 50 70 Victorville Helendale Mojave River Adelanto

117 30' 10 20 30 40 ROW MODEL Areal distribution of anisotropy for model layer 1 the ground-water flow Mojave River basin, sout 120 110 100 10 20 30 40 50 60 70 80 90 140 130 150 160 35 34 45' 34 30' Figure 21.

54 Simulation of Ground-Water Flow in the Mojave River Basin, California Streamflow-Routing package also simulates river HS is the sum of the elevation of the bottom of the fluctuations (wet and dry) when hydrologic conditions streambed and the stage in the river. The value assigned dictate; this could not be simulated using the analog to the stage for each stress period depended on the model developed by Hardt (1971). mean daily discharge measured at the Forks for each The Streamflow-Routing package assumes that year. (see “Simulation of Transient-State Conditions” water is available instantly to downstream reaches dur- section for further explanation of how the stress periods ing each stress period and that leakage between streams were defined). and the aquifer is instantaneous. These assumptions CSTR, also referred to as the streambed conduc- may not be reasonable for some stress periods, espe- tance, is equal to the product of the vertical hydraulic cially for areas of the model where the stream and conductivity of the streambed and the streambed area, underlying aquifer are separated by a thick unsaturated divided by the vertical thickness of the streambed. In an zone. The hydraulic and physical parameters assigned ideal system, where the river fully penetrates the aqui- to the cells that represent the river were the average fer and is not separated from the river by confining stream-reach properties of the actual system. material, the streambed conductance values are The Mojave River is represented by 330 assumed equal to the transmissivity of the aquifer of sequentially connected cells in downstream order from the model cell directly underlying the stream reach which there were no diversions, but many tributaries. divided by the thickness of layer 1, divided by the ratio The river is divided into 53 stream segments, whose of vertical-to-horizontal anisotropy for that cell. How- lengths and locations were defined by the tributaries. ever, the ephemeral nature of the Mojave River and the Each segment consists of a group of reaches connected large unsaturated zone beneath it in most areas greatly in a downstream order and each reach corresponds to affects the volume of water that infiltrates the individual cells in the model grid. Tributaries to the streambed. Infiltration of water through the streambed river were used to simulate ungaged runoff from local is not only related to the physical attributes of the washes to the river identified by Lines (1996, p. 19), streambed materials (porosity and vertical hydraulic discharge from fish hatcheries, discharge from the conductivity) it is primarily a function of the length of California State Water Project at the Mojave Water time that the channel contains water, the total area of Agency’s Morongo basin pipeline turnout, and dis- the channel that is wetted, and the soil-moisture content charge from VVWRA’s sewage pipeline. Figure 22 at the time the stream channel is wetted (Durbin and shows a schematic diagram of the Streamflow-Routing Hardt, 1974, p. 14). The amount of water passing package design and how it was used to incorporate the through the streambed materials also increases as the natural and artificial tributaries along the river. Leakage time interval between floodflows increases (Durbin and between the stream and aquifer is calculated for each Hardt, 1974, p. 14). Therefore, the values of CSTR reach based on the following equation when the assigned to the stream nodes in the model were based hydraulic head in the aquifer is greater than or equal to on the geologic materials of the streambed, the amount the elevation of the bottom of the streambed: of inflow (from the headwaters and ungaged QL = CSTR (HS - HA), tributaries), and the number of days of inflow. where The Mojave River was divided into 27 separate QL = leakage to or from the aquifer through the 3 sections, which were numbered sequentially in a down- streambed [L /t]; stream order, on the basis of similar geologic properties HS = hydraulic head in the stream [L]; of the streambed (fig. 22). By dividing the river into HA = hydraulic head in the aquifer side of the sections, it was possible to adjust streambed conduc- streambed [L]; and tance values along the river, basing those values on 2 CSTR = conductance of the streambed [L /t]. flow conditions. Table 7 shows the range of values of When the hydraulic head in the aquifer is less than the streambed conductance for each section of the river, elevation of the bottom of the streambed, the leakage is except the tributaries which had values of zero. QL = CSTR (HS - SBOT), Streambed conductance values for some sections were where constant during the entire model simulation (sections SBOT = the elevation of the bottom of the streambed 1,2 and 6–12); values for other sections changed [L]. depending on (1) the mean daily inflow and number of

Ground-Water Flow Model 55 117 30' 117 15' 117 00' 10 20 30 4050 60 70 80 90 100 110 120 MODEL COLUMN

10

Harper 20 Lake

35 30 00'

40 18 Barstow

50 42(1) IRON MTN 5.2 45(16) 41(33) 20 39(8) 40(5) 17 19 43(8) 16 44(2) 21 60 38(1) 37(11) 6.4 15 70 River 36(2) 35(14) 14 34(5)

MODEL ROW 11.74 34 32(7) 33(2) 13 45' Helendale 80 12 11 31(25) 90 Mojave

10

100 El Mirage 30(1)

Lake Lower Narrows 6 29(10) Adelanto 26(2) 110 28(2) 25(6) 8 Upper 27(3) Narrows Victorville 24(2) Southern 22(2) Apple California 7 23(2) Valley Lucerne 34 120 Logistics 6 Lake 30' Airport 21(10) 20(1) 5 19(3) 18(1) 17(3) 130 16(2) 15(3) Lucerne 13(2) 14(3) Rabbit Valley Hesperia Lake 4 12(8) 10(2) 11(1) 140 9(5)9(5) 8(2) 6(2) 7(3) 3 5(6) 3(2) 4(2) Dee 150 West Fork p Cr Mojave River eek 1(12) 2(3) 2 1 20.0 160

Figure 22. Schematic of simulated streamflow-routing network for the Mojave River ground-water basin, southern California.

56 Simulation of Ground-Water Flow in the Mojave River Basin, California 116 45' 116 30' 130 140 150 160170 180 190 200 MODEL COLUMN EXPLANATION

10 Coyote Model Grid Lake Active area 20 Inactive area 51(19) 52(1) 53(6) 6 50(1) 30 Model river section number 26 27 35 00' Model river cells 48(2) 5 49(10) 25 40 Specified-inflow tributary 47(43) 24 Artificial (see table 4 for site 10.0 23 10.0 names) 22 50 46(2) Natural (see tables 1 and 2 for site names) Troy Lake 60 20.0 Steady-state streamflow rate input In cubic feet per second Newberry locations – Springs 10.0 Steady-state injection well rate 70 input location – In cubic feet per second 34 45' 45(16) Streamflow segment number – 80 Number in parentheses is number of reaches Boundaries –

MODEL ROW 90 Model layer 1 Model Subarea 100 Gaging station

Dry lake (playa) 110 Rabbit Lake

120 34 30'

130

0 10 20 MILES

0 10 20 KILOMETERS 140

150

160

Figure 22.—Continued

Ground-Water Flow Model 57 days that the mean daily inflow from The Forks Simulation of Recharge exceeded 200 ft3/s during the year (sections 3–5), (2) the number of days that inflow from The Forks Recharge to the ground-water system includes exceeded 200 ft3/s during the year and whether there seepage loss from the Mojave River (discussed in the was inflow from ungaged tributaries (sections 13–18), preceding section), mountain-front recharge (infiltra- and (3) whether there was inflow from ungaged tribu- tion of runoff from selected washes and mountains taries (sections 19–27). (See “Simulation of Transient- along the southern boundaries) and artificial recharge State Conditions” for further discussion of the stress (irrigation-return flow, fish hatchery return flow, imported water, treated sewage, and septic effluent). periods). Years with similar flow regimes were grouped together in an effort to determine a relation between inflow and stream conductance. In doing so, the results Mountain-Front Recharge of the model simulations may not duplicate exactly the Most mountain-front recharge occurs during wet actual system for every year; therefore, these years as storm runoff infiltrates the alluvial fan deposits streambed conductance values should be considered of the regional aquifer. Recharge occurs mostly in the approximations to be improved upon by future studies. upper reaches of ephemeral streams and washes that lie

Table 7. Streambed-conductance values and associated flow conditions for stress periods used in the streamflow-routing package in the model of the Mojave River ground-water basin, southern California [See figure 22 for location of river sections. na, not applicable because flow conditions affecting streambed-conductance values during wet stress periods per- tain only to river sections 3–5 and 13–18; ft2/s, square foot per second; ft3/s, cubic foot per second; acre-ft, acre-foot; ≥, greater than or equal to]

Streambed Flow conditions affecting streambed- conductance conductance values during wet stress periods (ft2/s) River Number Comments Average daily inflow section of days of inflow Total inflow Wet stress Dry stress from from The Forks from The Forks period period The Forks (mean daily dis- (acre-ft) (acre-ft) charge ≥200 ft3/s) 1,2 0.2 0.2 na na na 3–5 .1 .8 0 1,800–3,600 10–20 .7 .8 1–3 400–1,400 400–700 .6 .8 1–10 1,400–10,000 450–2,650 3.0 .8 3–8 11,000–19,000 1,500–3,800 1.5 .8 15–20 10,000–23,000 600–1,100 1.8 .8 24–103 25,000–204,000 780–2,500 2.5 .8 108–138 245,000–400,000 1,980–3,700 6,7 1.0 .1 na na na 8,9 .1 .1 na na na 10 3.1 3.0 na na na 11,12 1.1 .1 na na na 13–18 3.5 2.5 0–6 na na 2.5 2.5 na na na Years with ungaged tributary flow to the river 2.0 2.5 na na na All other years 19–24 3.0 .1 na na na Years with ungaged tributary flow to the river 2.0 .1 na na na All other years 25–27 2.0 .1 na na na Years with ungaged tributary flow to the river .2 .1 na na na All other years

58 Simulation of Ground-Water Flow in the Mojave River Basin, California between the headwaters of the Mojave River and Sheep The estimated return flows were 46 percent for the Creek. In the Baja subarea, some recharge occurs near Alto, Transition zone, and Este model subareas; Coyote Lake and from Kane Wash (near Troy Lake) 35 percent for the Centro and Harper Lake model sub- (fig. 1). Mountain-front recharge was simulated as areas; and 29 percent for the Baja and Coyote Lake areal recharge to layer 1; the locations of the recharge model subareas. For 1951–73, the estimated return cells are shown in figure 18. According to concurrent flow for the area along the Mojave River between the studies by the USGS (Izbicki and others, 1995; John A. Jess Ranch and Mojave River Fish Hatcheries in the Izbicki, U.S. Geological Survey, oral commun., 1996; Alto model subarea was 70 percent. These higher esti- Michel, 1996; Gregory C. Lines, U.S. Geological mates were based on comparisons of land-use data Survey, oral commun., 1996), mountain-front recharge from historical areal photographs, consumptive-use occurs primarily in the upper reaches of the ephemeral rates of alfalfa (7.0 ft/yr), reported pumpage, and streams and washes and, therefore, recharge was model calibration. simulated in parts of the southern boundaries of the Recharge from irrigation-return flows to the Este, Alto, and Oeste model subareas (fig. 18). regional-aquifer system was not estimated for the Recharge also was applied to the Coyote Lake area and Oeste model subarea because of perched water-table at a few cells near the mouth of Kane Wash. Areal conditions (fig. 11). Smith and Pimentel (2000) recharge was applied at a constant rate and was deter- reported the mounding of ground water in a perched mined by model calibration. The model-calibrated aquifer system which probably is the result of areal recharge values, in acre-ft/yr, for the following irrigation-return flow. Although this water eventually are Oeste, 1,940; Alto, 7,760; Este, 1,030; Coyote may reach the regional aquifer system, model calibra- Lake, 260; and Kane Wash, 650. tion results indicate that the perched water is not a significant source of recharge to the regional system. Artificial Recharge Fish Hatchery Discharge and Imported Water The main sources of artificial recharge to the basin have been irrigation-return flow, fish hatchery Discharge from the Mojave River and Jess return flow, imported SWP water at the MWA Morongo Ranch Fish Hatcheries, and imported water from the basin pipeline turnout, treated sewage effluent, and MWA pipeline is released directly to the river, there- seepage from septic systems. fore, these sources were simulated in the model using the Streamflow-Routing package and treated as artifi- Irrigation-Return Flow cial tributaries (figs. 18 and 22). The annual release rates for the fish hatchery return flows and the imported Recharge from irrigation-return flows was water are presented in table 4. simulated in layer 1 using injection wells in the same areal location that the pumping occurred. For example, Treated Sewage Effluent when pumping for irrigation occurred in layer 2, row 125, column 60, the return-flow recharge was simu- Treated sewage effluent from VVWRA that is lated in layer 1, row 125, column 60. No return-flow discharged directly to the Mojave River in the Transi- recharge was applied to areas of perched water tion zone model subarea was simulated using the (fig. 11). Streamflow-Routing package and treated as an artificial As discussed earlier, Hardt (1971) reported only tributary (figs. 18 and 22). Sewage effluent that is net pumpage for 1931–50 and, therefore, 1931–50 routed to the VVWRA seepage ponds and thus not dis- irrigation-return flows were assumed to be 40 percent charged directly to the river was simulated as injection of the total agricultural pumpage in the Alto subarea wells at the corresponding model cells in layer 1. and 50 percent in all other subareas. For 1951–94, the Injection wells also were used to simulate the return-flow percentages were based on the method used sewage discharged to seepage ponds from the city of to calculate total agricultural pumpage for 1986–94 Barstow in the Centro model subarea, and sewage (Robert Wagner, James C. Hanson Engineering, writ- effluent in the USMC Nebo and Yermo Annexes in the ten commun., 1995) and consumptive-use rates in each Baja model subarea. The annual discharge rates for model subarea (U.S. Department of Agriculture, 1967). sewage effluent are shown in table 4.

Ground-Water Flow Model 59 Septic Systems pumpage is divided into five main categories of usage: Effluent from the septic systems in the Alto (1) agricultural, all water pumped for irrigation in the subarea was simulated as areal recharge to layer 1. basin; (2) municipal and industrial, water pumped by Areal recharge was applied to the number of acres nec- the various cities, individual water districts, and the essary to accommodate the population estimated for a military; (3) fish hatcheries, water pumped for circula- 10-year period (fig. 18 and table 5). tion in fish-rearing ponds; (4) lakes, recreational lakes in the Baja subarea; and (5) domestic. Generally, domestic pumpage is not a significant component of Simulation of Discharge the total annual ground-water production and thus is considered negligible for modeling purposes. All simu- The principal components of ground-water lated pumpage was extracted from layer 2 in the model. discharge from the aquifer system are pumpage, evapo- In areas where layer 2 did not exist, pumpage was transpiration, seepage to the Mojave River, and under- extracted from layer 1. Along the river, both layers have flow through Afton Canyon out of the basin. Seepage of similar hydrologic properties and most wells are perfo- ground water to the Mojave River is discussed in the rated in the younger alluvium (Qya) which extends to “Stream-Aquifer Interactions” section of this report, layer 2 (fig. 9). and underflow at Afton Canyon is discussed in the The estimated total annual pumpage from wells “Model Boundary Conditions” section. in each of the model subareas in the Mojave River ground-water basin for 1931–99 is shown in figure 23. Pumpage Annual pumpage in the Mojave River ground-water Ground-water pumpage is the principal source of basin was estimated during several previous studies; discharge from the aquifer system. For this report, however, the reports of these studies do not cover all

110

100 Subarea Oeste 90 Alto Este 80 Transition zone Centro Harper Lake 70 Baja Coyote Lake 60

50

40

30 PUMPAGE, IN THOUSANDS OF ACRE-FEET

20

10

0 1931 1935 1940 1945 1950 1955 1960 1965 1970 1975 1980 1985 1990 1995 2000 YEAR Figure 23. Total pumpage by model subarea for the Mojave River ground-water basin, southern California, 1931–99. (See figure 18 for location of model subareas.)

60 Simulation of Ground-Water Flow in the Mojave River Basin, California years of the period of this current study (1931–94), nor the net pumpage values reported for 1963 by Hardt were they complete (fig. 14). Production reported by (1971) were constant. Dibble (1967), the USMC (Mike Cox, written com- Municipal and military pumpage values were mun., 1994), the California Department of Fish and used in the model without modification for consump- Game (Richard Uplinger, written commun., 1994), and tive use because pumpage was distributed for public James C. Hanson Engineers, (Robert Wagner, written supply and, therefore, the water did not return to the commun., 1994) was total pumpage. It should be noted aquifer system at the point of discharge. Any reintro- that there are discrepancies between the pumpage val- duction of this water into the ground-water system, ues presented on page 36 and on table 2 of the Dibble such as through wastewater or irrigation, was (1967) report. The text of that report states that produc- accounted for at the point where the water was applied. tion was verified for 1,195 wells between 1951 and Municipal wastewater that is discharged directly to the 1965; however, the table in the accompanying data sec- river was included in the Streamflow-Routing package, tion of the report lists about 1,530 wells that were ver- and wastewater discharged to ponds at treatment plants ified, 1,522 of which lie within the model area. Because was simulated as injection wells (see “Treated Sewage the verified production data in the table were available Effluent” section). for more wells than presented in the text (Dibble, 1967, table 2), it is assumed that pumpage from some areas Recreational Lakes in the Baja Model Subarea was not included in the text for unknown reasons. All In the Baja model subarea, man-made the verified pumpage data were used in the model for recreational and private lakes were first constructed in this study. the late 1950’s and early 1960’s (table 8). Although it Hardt (1971) reported net pumpage for 1930–63. was assumed that most of the water pumped into the Those net-pumpage values were used in the model for lakes is recirculated, it was necessary to account for the the period 1931–50; the sources of the total pumpage volume of annual water consumption from evaporation values were not available. The amount of consumptive and the total volume of water needed to fill the lakes use applied to the total pumpage during this period was initially. Wells that operate to maintain the volume of not reported in detail, therefore, it was not possible to these lakes were assigned annual consumptive-use val- determine the corresponding total-pumpage values. For ues based on the lake surface area and an assumed the period 1951–63, when both net-pumpage data evaporation rate of 7.0 ft/yr (Robert Wagner, James C. (Hardt, 1971) and total pumpage (Dibble, 1967) were Hanson Engineering, oral commun., 1998). The available (fig. 14), total-pumpage values were used in amount of water necessary to fill the lakes, or fill vol- ume, was estimated by adding the volume of water con- the model. tained by the lake to the volume of water necessary to For years when no pumpage data were available, fill the underlying unsaturated zone. Once a lake was pumpage was estimated by linear interpolation using a filled, its volume was maintained by wells at the rate GIS. A linear estimate was made using available data determined by the annual water consumption for the year before and the year after the missing data. (evaporation) (table 8). To estimate the fill volume, the If the use of the water pumped at the well differed following assumptions were made: between years with known pumpage values, it was • average depth of lakes = 5 ft, assumed that a change in the use occurred at the mid- • average porosity of unsaturated zone point of the estimated period. In most cases, estimates sediments = 14 percent, and were made on a well-by-well basis; but when pumpage • depth to water table = 100 ft. at individual wells was not available, linear estimates Aerial photos were used to determine the year were made in the model on a cell-by-cell basis. Note that the lakes were constructed. However, aerial photos that data reported by Hardt (1971) were net-pumpage of the Baja model subarea were not available prior to values and, therefore, any pumpage that was estimated 1969; therefore, it was assumed that lakes present in the using data from 1931 or 1963 as endpoints was 1969 aerial photos were gradually filled over a 10-year estimated as net pumpage. period, from 1959 to 1969. When the length of time Estimates of net pumpage for 1964–85 for the that a lake existed was known, it was assumed that it Harper Lake model subarea were made by assuming was filled during the first year of record. On the basis

Ground-Water Flow Model 61 Table 8. Annual water consumption of recreational lakes in the of these assumptions, the total volume of the lakes is Baja subarea of the Mojave River ground-water basin, southern about 1,920 acre-ft, and the volume of water needed to California fill the lakes and the underlying unsaturated zone is [State well No.: See well-numbering system in text. acre-ft, acre-foot] about 7,300 acre-ft, spread over 40 years (table 8). Annual Period water Estimated Fill Area, Transpiration by Phreatophytes and Bare-Soil Evaporation State well No. of consump- volume, volume, acres record tion, acre-ft acre-ft acre-ft Transpiration by phreatophytes along the 8N/3E-3G 7 1981–99 49 35 133 Mojave River and evaporation from bare-soil areas in 8N/3E-3M 6 11959–69 42 30 114 the river channel are simulated in the model using the 8N/4E-6J 7 1987–99 49 35 133 Evapotranspiration (EVT) package developed by 9N/2E-5B,F 24 1985–99 168 120 456 McDonald and Harbaugh (1988). Consumptive use of 9N/2E-7P 3 1972–99 21 15 57 water by riparian vegetation for 1995 was computed 9N/2E-10N 3 1973–99 21 15 57 using water-use estimates for various plant species and 9N/2E-13R 7 11959–71 49 35 133 areal densities by Lines and Bilhorn (1996). Evapo- 9N/2E-24N 4 11959–77 28 20 76 transpiration was assumed to be at a maximum rate 9N/2E-24R 4 11959–89 28 20 76 when the water table was at land surface and to 9N/3E-2L 6 1987–99 42 30 114 decrease linearly to zero when the water table was 9N/3E-3G 3 11959–99 21 15 57 25 ft below land surface. The extinction depth of 25 ft 9N/3E-3H 6 1985–99 42 30 114 represents an average depth for deep-rooted (saltcedar, 9N/3E-4G,H 30 11959–99 210 150 570 desert willow, and mesquite) and shallow-rooted (cot- 9N/3E-8K,Q 24 1985–99 168 120 456 tonwoods, baccharis, and willows) riparian vegetation 9N/3E-8K,Q 15 1989–99 105 75 285 along the Mojave River channel. 9N/3E-10C 6 1972–99 42 30 114 The maximum water-use rate (evapotranspira- 9N/3E-13C 5 11959–77 35 25 95 tion rate) in the Alto and Transition zone model subar- 9N/3E-19A,H 17 1985–99 119 85 323 eas was 5.6 ft/yr, and the maximum rate for the Centro, 9N/3E-19N 3 11959–99 21 15 57 Baja, and Afton Canyon model subareas was 6.7 ft/yr 9N/3E-20B,G 14 1977–99 98 70 266 (Lines, 1996). These rates were applied to the model 9N/3E-22D 6 1985–99 42 30 114 for 1931–49, the period prior to the significant lower- 9N/3E-25K,Q 17 11959–99 119 85 323 ing of the water table by pumpage when the water table 9N/3E-26C 4 11959–91 28 20 76 was most likely at, or very near, land surface. Water- 9N/3E-28M 4 11959–73 28 20 76 use rates for 1950–94 were estimated for each model 9N/3E-36G 4 11959–85 28 20 76 subarea on the basis of areal densities of various plant 9N/4E-7Q 15 11959–73 105 75 285 species as reported by Lines and Bilhorn (1996, 9N/4E-8E 5 11959–99 35 25 95 table 6). The amount of evapotranspiration estimated 9N/4E-18A 11 11959–99 77 55 209 by Lines and Bilhorn (1996, table 7) for each subarea 9N/4E-18D 15 1971–99 105 75 285 in 1995 was used as a guide when selecting water-use 9N/4E-21J 4 11959–85 28 20 76 rates for each model subarea. The water-use rates used 9N/4E-32D 3 1985–99 21 15 57 in the model were 5.6 ft/yr for the Alto and Transition 9N/4E-20D 10 11959–87 70 50 190 zone model subareas; 1.5 ft/yr for the Centro model 10N/3E-14L,P 5 1987–99 35 25 95 subarea (based on the predominance of saltcedar); 10N/3E-15M 3 11959–83 21 15 57 1.3 ft/yr for the Baja model subarea west of Camp 10N/3E-20G 8 11959–99 56 40 152 Cady (based on the predominance of mesquite); and 10N/3E-30L,P 52 1981–99 364 260 988 1.7 ft/yr for the remainder of the river, downstream 10N/3E-36N 5 11959–77 35 25 95 from Camp Cady through the Afton Canyon model 10N/4E-20B 7 11959–99 49 35 133 subarea (based on various types of vegetation). It was 10N/4E-31C,D 12 11959–85 84 60 228 evident during model calibration that to achieve the Total ...... 384 1,920 7,296 evapotranspiration rates estimated by Lines and Bilhorn (1996) for the Alto model subarea, it was 1Present on 1969 aerial photographs, assumed to have been constructed in 1959. necessary to use the maximum water-use rate of

62 Simulation of Ground-Water Flow in the Mojave River Basin, California 5.6 ft/yr. The need to use a high water-use rate in this above sea level: Rabbit Lake, 2,936; El Mirage Lake, area possibly is due to an overestimation of pumpage in 2,834; Harper Lake, 2,020; Coyote Lake, 1,706; and the area upstream from the Upper Narrows. A lower Troy Lake, 1,773. Flow out of the drain is controlled by water-table altitude would act to limit the amount of the conductance between the aquifer and the drain and water that could be removed from the model by the by the effects of the hydraulic head at each cell. The simulation of evapotranspiration. estimated drain conductances, in foot squared per day, The acreages and areal densities to which the for Rabbit, El Mirage, Harper, Coyote, and Troy Lakes water-use rates were applied were estimated by Lines were 2.0 × 10-3, 2.0 × 10-3, 1.0, 1.0 × 10-3, and and Bilhorn (1996, tables 1–6). The total evapotranspi- 2.0 × 10-2, respectively. The conductance was ration rates used in the model, therefore, are a product determined by model calibration because ground-water of the water-use rates assigned to each model subarea discharge to the dry lakes is not measured directly. and the number of acres of riparian vegetation and open water represented by each model cell. The total number of acres in the model is slightly lower than those Model Calibration reported by Lines and Bilhorn (1996) because some of their estimates of acreage were for areas of vegetation Ground-water conditions during the period outside the active area of the model. The area, in acres, 1931–94 were used to calibrate the transient-state of vegetation and open water in the following model model of the Mojave River ground-water basin. A subareas was about 1,320 for Alto; 2,580 for the Tran- steady-state simulation was made to provide initial sition zone; 2,740 for Centro, except for 1931–49; conditions for the transient-state simulation. The model 2,760 for Baja; and 350 for Afton Canyon. During was iteratively calibrated using a trial-and-error pro- model calibration, it became evident that historical cess during which initial estimates of the aquifer prop- water levels in the Centro model subarea were higher than those indicated by the model results. This proba- erties were adjusted to improve the match between bly is due to a change in the amount of acreages of simulated and measured ground-water levels, and some riparian vegetation that once existed but has since been water-budget items were reviewed. The iterative cali- reduced by development and now is being used for bration process involved three steps: (1) calibrating the agricultural, residential, and other uses. Indeed, Lines steady-state (initial-condition) model, (2) using the and Bilhorn (1996, plate) show that the greatest num- parameter estimates from the steady-state model in the ber of acres of disturbed land—about 6,300 acres—is transient-state model, and (3) calibrating the parame- in the Centro model subarea. Therefore, to increase the ters specific to the transient-state model. If a satisfac- amount of evapotranspiration and thus lower hydraulic tory match between measured and simulated results heads in the model for 1931–49 in the Centro model was not obtained, the process was restarted at step 1. subarea, the area of riparian vegetation was increased The initial estimates were adjusted within limits of the to include the entire area of each cell used in the stream geologic and hydrologic properties of the aquifer package regardless of whether or not vegetation was system. The closeness of the final match is controlled present in 1995. by the complexity of the real system not addressed by the model, the quality and availability of data to char- Dry Lakes acterize the system, and the time constraints on the The five dry lakes in the study area were study. Data for calendar years 1995–99 were used to simulated as drains using the Drain (DRN) package validate the calibrated ground-water flow model, that developed by McDonald and Harbaugh (1988). This is, to test that the flow model will duplicate measured package allowed ground water to discharge at the dry data for a non-calibration period without modification lakes only when the hydraulic head in the aquifer was of the model parameters (see discussion in “Model Val- greater than the altitude of the drain. When the hydrau- idation” section). Many of the figures presented in the lic head in the model cell was less than the altitude of “Simulation of Transient-State Conditions” and the the drain, there was no flow into the drain. The altitudes “Model Validation” sections show results of both the of the drain cells in the model are set equal to the transient-state period (1931–94) and the model average altitude of the surface of the dry lakes, in feet validation period (1995–99).

Ground-Water Flow Model 63 Simulation of Steady-State Conditions for the Alto, Centro, and Baja model subareas, respectively (fig. 22). These rates were applied to each A steady-state simulation of 1930 conditions was made to provide initial conditions for the transient- model subarea using the Streamflow-Routing package state simulation. The year 1930 was chosen because (Prudic, 1989). To ensure that flow occurred in the river pumpage prior to this year probably did not signifi- in the lower Baja model subarea and to better match the cantly affect the aquifer system and because it was the measured data, it was necessary to input an additional first year that a comprehensive water-level data base 7,200 acre-ft/yr using an injection well (fig. 22). had been compiled for the study area (California Streambed conductance affects the leakage of stream- Department of Public Works, 1934, pl. 5). Although flow into the ground-water system from the headwaters ground water was pumped in the area prior to 1931, we to Afton Canyon. In order to simulate the measured assumed that because the pumping rates and water lev- water levels, the streambed conductance values were els were relatively constant between 1931 and 1945 calibrated; however, these estimates had no transfer (fig. 14) the ground-water-system was at or near a value to the transient-state model because of the steady-state condition. A steady-state condition occurs manner in which the streamflow was distributed. when the inflow and outflow of an aquifer system are equal and thus the volume of water stored in the system Calibration of the steady-state conditions was remains constant over the long term. The steady-state done and goodness-of-fit was determined by compar- simulation consisted of modifying initial estimates of ing simulated hydraulic heads and measured water lev- transmissivity, vertical conductance between layers, els. Simulated steady-state hydraulic heads for layer 1 hydraulic characteristics of the faults, drain conduc- and measured water levels for 1930 are shown in tances, stream conductances, general-head boundary figure 24. In general, the simulated hydraulic-head con- conductances, streamflow, mountain-front recharge, tours are similar to the measured 1930 water-levels. and evapotranspiration. Storage coefficients are not The largest differences were in the Transition zone used in steady-state simulations. Ground-water level model subarea west of the Mojave River and in the measurements made prior to 1931, most of which were Centro model subarea west of Iron Mountain (fig. 24). made in 1917, were used to determine if the steady- Although Hardt (1971) showed contoured water levels, state simulation provided reasonable initial conditions there were no supporting data for these contours; there- for the transient-state simulation. fore, the differences between the contours from Hardt Simulating streamflow in the Mojave River for (1971) and the modeled data from this report can not be steady-state conditions is not straightforward. The steady-state water levels in the Mojave River ground- interpreted in detail in these areas. The large differ- water basin are the result of a series of wet and dry peri- ences between the simulated hydraulic head and ods. During some wet periods, flow is present in the measured water levels for 1930 in the area near the Mojave River from the headwaters to Afton Canyon for Southern California Logistics Airport (fig. 24) are due short periods of time; however, the Mojave River is to the perched water table which is not representative of usually dry over most of its reach. Recharge from the the regional aquifer and, therefore, was not simulated Mojave River to the aquifer system occurs primarily in the model. during these infrequent wet periods. Streamflow vari- The simulated hydraulic heads and measured ability, however, could not be simulated because a water levels for 1930 are plotted together for compari- steady-state simulation is independent of time and spo- son purposes in figure 25. The overall root mean square radic episodes cannot be incorporated into a constant error (RMSE) equaled 16.7 ft and the measured minus estimate of recharge. Therefore, in order to simulate the measured steady-state ground-water levels, it was nec- simulated mean error (ME) equaled 7.4 ft. The largest essary to distribute Mojave River streamflow in the RMSE value was for the Centro model subarea, which Alto, Centro, and Baja model subareas such that the had a value of 28.6 ft (fig. 25). The correlation coeffi- simulated steady-state water levels in each model sub- cient between the simulated steady-state hydraulic area matched the measured values. The streamflow head and the measured water levels for 1930 conditions rates were about 14,500, 16,900, and 7,200 acre-ft/yr equaled 0.999.

64 Simulation of Ground-Water Flow in the Mojave River Basin, California 1400

1,500 20 MILES 1,500 30' ° 116

1,600 California.

1,700 7E1

1,779 20 KILOMETERS

12G1 1,774 1,600 Shows altitude of water level. 10 EXPLANATION 3D1 1,700 1,781 34D1 1,774 With abbreviated State well number.

2111 1,734

1,800 Springs Newberry 10 ater-level contour (measured 1930) from ater-level

ell – Model faults W Hardt, (1971) – Contour interval is 100 feet. Datum sea level Hydraulic-head contour (simulated 1930) – Shows altitude of simulated hydraulic head. Contour interval is 100 feet. Datum sea level W Lower number is measured altitude of water level. Datum is sea level 1,800 Boundary of model layer 1

26D1 1,869 45'

14N1 1,862 3K1 1,720 1,852 7E1

27D1 1,881 1,779 2,800 2,800 0 0

22D1 1,894 1,882 20M1

13E2 1,900 1,887 1,900 R 1 E R 2 E R 3 E R 4 E R 5 E R 6 E

35P4 1,885 3,200

21C1 1,916 3,100

Valley

2,000 Lucerne 2,000 3,000 ° R 1 W 117 6B1 32K1 2,069 2,088

2,100

Barstow 36P1 2,106 2,900

10G1 2,874 2,100 2,900

34F1 2,057 Apple Valley

2,200 1M1 2,849 12H1 2,840

30N3 2,180

18A1 2,180 2,200 2,100

27N2 2,047

10R1 2,874 28H1 2,041 9J1 28A1 2,840 2,828 34N4 2,264 5A1

2,836 30G1 2,881

R 3 W 2 W Airport 2,100 Southern Logistics 7F5 California 18R1 2,804 2,300 2,940 2,300 7A1 2,834

Victorville

15' 20N1 2,029 2,300 10E1 2,710

Iron Mtn Helendale

22A1 2,353 30C1 2,507 6E9 2,545

2823 1C1 36N1 2,858

32N3 2,710

2,000 2,710 14M1 7Q3 2,466 R 4 W

2,100 12R1 2,795 Hesperia

2,800 12P1 2,862

2,200 2,400 2,500 2,700 2,800

2,100 7B1 2,900 2,446 5 W R 4 W R 3 W R 2 W R 1 W R 1 E 2H1 2,781 5R1

2,400 2,880

2E1 2,900 2,100 2,576

2,600 22E1 2,824

R 5W 2,500

16R1 2,736

12F2 2,854 Mojave River 2,600 30' °

2,,700 20R1 2,768 117

22J1 2,924 R 6 W 14P1 2,790 3,000 2Q1 R 41 E 2,923

12N1 2,839

2,865 24M1

3,100 W 7 R R 6 W R R 7 W

18Q1 2,864 Measured water levels and simulated hydraulic head for 1930 model layer 1 of the Mojave River ground-water basin, southern 5P1 2,923 R40E R 42 E R43 E R44 E R 45 E R 46 E R 47 E ° 45' 30' 35 ° T8N ° T9N 32 S T4N 12 N T3N 10 N T5N T7N T6N T T T 11 N T 11 34 T 34 Figure 24.

Ground-Water Flow Model 65 3,000 Oeste (11.5) Subarea – (Number is RMSE value) Este (1 value)

2,800

Transition zone (11.7) Alto (17.3)

2,600

2,400

1:1 Correlation line 2,200 Harper Lake (8.9) Centro (28.6) SIMULATED WATER LEVEL, IN FEET WATER SIMULATED 2,000

1,800

Baja (12.9)

1,600 1,600 1,800 2,000 2,200 2,400 2,600 2,800 3,000 MEASURED WATER LEVEL, IN FEET Figure 25. Measured water levels and simulated hydraulic head and the root mean square error (RMSE) for each model subarea of the Mojave River ground-water basin, southern California, for 1930 steady-state conditions. (See figure 18 for location of model subareas.)

Simulation of Transient-State Conditions Each calendar year of the transient-state simulation was represented by two stress periods, a wet Ground-water conditions for the 64-year period period and a dry period, for a total of 128 stress periods of 1931–94 were used to calibrate the transient-state (table 9). The duration of each stress period was a func- model. Transient-state conditions occur in an aquifer tion of the occurrence, quantity, and length of storm- system when inflows do not equal outflows. The flow from the headwaters of the Mojave River each transient-state calibration consisted of modifying esti- year. The actual number of days in each stress period mates of transmissivity, vertical conductance between during 1931–94 was based on combined streamflow layers, hydraulic characteristics of the faults, drain con- discharge records from the headwaters (West Fork and ductances, stream conductances, general-head bound- Deep Creek tributaries). Inflows to the Mojave River ary conductances, streamflow, mountain-front from the headwaters are highly seasonal and vary in volume from year to year. Peak discharges, or floods, recharge, evapotranspiration, and storage coefficients. generally occur during the winter and early spring, These parameters were modified using a trial-and-error although some isolated flood events do occur in the approach until simulated hydraulic heads and fluxes summer. Each stress period was simulated in the model reasonably matched measured values. The steady-state with four equal-length time steps. A maximum of simulated hydraulic heads were used as initial 12 equal-length time steps were tested, but the conditions. additional time steps did not improve the results.

66 Simulation of Ground-Water Flow in the Mojave River Basin, California Table 9. Stress period lengths and specified inflows from The Forks (Deep Creek and West Fork) to the Mojave River, southern California, 1931–99 [Number of days for the wet stress period represents the number of days that average combined inflow at The Forks is greater than or equal to 200 cubic feet per second; exceptions noted in footnote. ft3/s, cubic foot per second; acre-ft, acre-foot; w, wet; d, dry]

Inflow from Deep Creek Inflow from West Fork Combined mean Stress daily Stress Number Year period Start date discharge period No. of days 3 3 (wet/dry) (ft /s) (acre-ft) (ft /s) (acre-ft) at The Forks (acre-ft) 1 1931 w 1/01/31 10 344.40 6,831 149.80 2,971 9,802 2 1931 d 1/11/31 355 11.07 7,794 3.00 2,110 9,905 3 1932 w 1/01/32 71 352.87 49,694 205.42 28,929 78,623 4 1932 d 3/12/32 295 25.12 14,698 6.21 3,635 18,333 5 1933 w 1/01/33 1 603.00 1,196 166.00 329 1,525 6 1933 d 1/02/33 364 20.24 14,609 11.01 7,952 22,561 7 1934 w 1/01/34 6 642.83 7,650 284.50 3,386 11,036 8 1934 d 1/07/34 359 9.95 7,082 2.22 1,580 8,662 9 1935 w 1/01/35 28 303.14 16,836 149.46 8,301 25,137 10 1935 d 1/29/35 337 27.51 18,387 12.66 8,459 26,846 11 1936 w 1/01/36 15 241.53 7,186 155.27 4,620 11,806 12 1936 d 1/16/36 351 19.87 13,837 4.55 3,169 17,006 13 1937 w 1/01/37 103 492.05 100,524 247.56 50,577 151,101 14 1937 d 4/14/37 262 17.99 9,350 8.80 4,573 13,923 15 1938 w 1/01/38 88 741.25 129,382 423.27 73,880 203,262 16 1938 d 3/30/38 277 28.32 15,558 9.76 5,362 20,920 17 1939 w 1/01/39 18 255.56 9,124 52.06 1,859 10,982 18 1939 d 1/19/39 347 27.05 18,617 8.69 5,979 24,596 19 1940 w 1/01/40 18 408.22 14,575 116.00 4,141 18,716 20 1940 d 1/19/40 348 23.26 16,057 6.26 4,322 20,378 21 1941 w 1/01/41 101 423.07 84,754 274.46 54,982 139,736 22 1941 d 4/12/41 264 26.00 13,614 7.69 4,028 17,643 23 1942 w 1/01/42 3 185.67 1,105 39.33 234 1,339 24 1942 d 1/04/42 362 19.78 14,201 7.50 5,388 19,589 25 1943 w 1/01/43 79 512.87 80,364 349.24 54,724 135,088 26 1943 d 3/21/43 286 27.53 15,618 7.59 4,308 19,926 27 1944 w 1/01/44 74 233.11 34,215 246.26 36,145 70,360 28 1944 d 3/15/44 292 27.93 16,176 8.37 4,847 21,023 29 1945 w 1/01/45 54 343.02 36,740 167.87 17,980 54,720 30 1945 d 2/24/45 311 24.42 15,061 8.15 5,026 20,087 31 1946 w 1/01/46 37 395.57 29,030 309.49 22,713 51,743 32 1946 d 2/07/46 328 23.03 14,982 7.96 5,182 20,164 33 1947 w 1/01/47 2 143.00 567 127.50 506 1,073 34 1947 d 1/03/47 363 15.46 11,133 9.22 6,638 17,772 35 1948 w 1/01/48 4 230.75 1,831 131.75 1,045 2,876 36 1948 d 1/05/48 362 11.67 8,377 2.89 2,074 10,450

Ground-Water Flow Model 67 Table 9. Stress period lengths and specified inflows from The Forks (Deep Creek and West Fork) to the Mojave River, southern California, 1931–99—Continued Inflow from Deep Creek Inflow from West Fork Combined mean Stress daily Stress Number Year period Start date discharge period No. of days 3 3 (wet/dry) (ft /s) (acre-ft) (ft /s) (acre-ft) at The Forks (acre-ft) 37 1949 w 1/01/49 7 152.71 2,120 119.29 1,656 3,777 38 1949 d 1/08/49 358 20.31 14,422 9.67 6,863 21,285 39 1950 w 1/01/50 2 383.50 1,521 121.00 480 2,001 40 1950 d 1/03/50 363 8.41 6,056 3.00 2,156 8,212 41 1951 w 1/01/51 2 1,030.00 4,086 293.00 1,162 5,248 42 1951 d 1/03/51 363 4.62 3,328 0.02 14 3,342 43 1952 w 1/01/52 74 280.53 41,175 252.22 37,020 78,194 44 1952 d 3/15/52 292 23.89 13,836 10.27 5,949 19,785 45 1953 w 1/01/53 1182 3.84 1,385 1.24 448 1,833 46 1953 d 7/02/53 183 3.84 1,392 1.24 450 1,843 47 1954 w 1/01/54 38 364.37 27,463 183.24 13,811 41,274 48 1954 d 2/08/54 327 17.27 11,201 5.04 3,270 14,471 49 1955 w 1/01/55 4 150.75 1,196 84.00 666 1,862 50 1955 d 1/05/55 361 14.84 10,628 5.75 4,118 14,746 51 1956 w 1/01/56 3 1,652.33 9,832 238.33 1,418 11,250 52 1956 d 1/04/56 363 5.80 4,175 0.97 696 4,872 53 1957 w 1/01/57 8 1,044.50 16,574 127.84 2,028 18,602 54 1957 d 1/09/57 357 15.62 11,063 3.90 2,762 13,826 55 1958 w 1/01/58 71 587.23 82,697 285.99 40,274 122,971 56 1958 d 3/13/58 294 20.06 11,696 7.07 4,123 15,819 57 1959 w 1/01/59 6 475.83 5,663 262.33 3,122 8,785 58 1959 d 1/07/59 359 11.77 8,378 2.21 1,577 9,955 59 1960 w 1/01/60 1183 6.38 2,316 0.16 56 2,373 60 1960 d 7/02/60 183 6.38 2,316 0.16 56 2,373 61 1961 w 1/01/61 2 946.00 3,753 122.00 484 4,237 62 1961 d 1/03/61 363 5.21 3,753 0.14 103 3,855 63 1962 w 1/01/62 30 546.03 32,491 192.67 11,464 43,956 64 1962 d 1/31/62 335 21.49 14,276 6.54 4,347 18,623 65 1963 w 1/01/63 1 214.00 424 7.00 14 438 66 1963 d 1/02/63 364 8.12 5,862 0.10 71 5,933 67 1964 w 1/01/64 2 261.50 1,037 102.50 407 1,444 68 1964 d 1/03/64 364 12.14 8,767 0.45 325 9,092 69 1965 w 1/01/65 36 918.36 65,576 370.03 26,422 91,997 70 1965 d 2/06/65 329 14.58 9,512 6.17 4,028 13,540 71 1966 w 1/01/66 24 802.67 38,210 255.08 12,143 50,352 72 1966 d 1/25/66 341 26.08 17,637 9.94 6,720 24,357 73 1967 w 1/01/67 72 242.19 34,588 218.92 31,263 65,851

See footnote at end of table.

68 Simulation of Ground-Water Flow in the Mojave River Basin, California Table 9. Stress period lengths and specified inflows from The Forks (Deep Creek and West Fork) to the Mojave River, southern California, 1931–99—Continued Inflow from Deep Creek Inflow from West Fork Combined mean Stress daily Stress Number Year period Start date discharge period No. of days 3 3 (wet/dry) (ft /s) (acre-ft) (ft /s) (acre-ft) at The Forks (acre-ft) 74 1967 d 3/14/67 293 28.99 16,850 16.08 9,345 26,195 75 1968 w 1/01/68 2 143.00 567 148.00 587 1,154 76 1968 d 1/03/68 364 17.93 12,948 5.83 4,208 17,156 77 1969 w 1/01/69 120 868.92 206,817 511.22 121,678 328,495 78 1969 d 5/01/69 245 25.61 12,444 4.30 2,089 14,533 79 1970 w 1/01/70 6 272.17 3,239 247.33 2,943 6,182 80 1970 d 1/07/70 359 17.64 12,562 3.78 2,689 15,251 81 1971 w 1/01/71 6 1,235.00 14,698 561.87 6,687 21,384 82 1971 d 1/07/71 359 16.96 12,078 4.62 3,288 15,365 83 1972 w 1/01/72 2 480.00 1,904 240.00 952 2,856 84 1972 d 1/03/72 364 8.14 5,877 4.07 2,938 8,815 85 1973 w 1/01/73 49 262.80 25,542 131.04 12,736 38,277 86 1973 d 2/19/73 316 23.84 14,942 11.92 7,471 22,414 87 1974 w 1/01/74 7 240.67 3,342 120.33 1,671 5,012 88 1974 d 1/08/74 358 20.84 14,798 10.42 7,399 22,197 89 1975 w 1/01/75 2 185.00 734 151.50 601 1,335 90 1975 d 1/03/75 363 14.84 10,685 5.55 3,998 14,683 91 1976 w 1/01/76 7 562.00 7,803 235.29 3,267 11,070 92 1976 d 1/08/76 359 14.42 10,269 4.13 2,937 13,207 93 1977 w 1/01/77 5 556.20 5,516 198.00 1,964 7,480 94 1977 d 1/06/77 360 12.38 8,838 2.19 1,563 10,401 95 1978 w 1/01/78 138 806.20 220,671 473.43 129,586 350,257 96 1978 d 5/19/78 227 23.75 10,695 8.07 3,633 14,327 97 1979 w 1/01/79 86 348.28 59,409 142.40 24,290 83,699 98 1979 d 3/28/79 279 32.46 17,963 6.48 3,586 21,549 99 1980 w 1/01/80 110 818.68 178,621 506.16 110,436 289,057 100 1980 d 4/20/80 256 30.56 15,517 6.19 3,145 18,661 101 1981 w 1/01/81 1182 7.05 2,547 2.85 1,028 3,575 102 1981 d 7/02/81 183 7.05 2,561 2.85 1,034 3,595 103 1982 w 1/01/82 50 355.48 35,254 175.48 17,403 52,657 104 1982 d 2/20/82 315 26.07 16,290 4.41 2,756 19,046 105 1983 w 1/01/83 124 537.92 132,301 460.00 113,137 245,439 106 1983 d 5/05/83 241 38.40 18,356 8.32 3,978 22,333 107 1984 w 1/01/84 4 280.50 2,225 226.00 1,793 4,019 108 1984 d 1/05/84 362 12.87 9,241 2.87 2,063 11,305 109 1985 w 1/01/85 2 347.50 1,379 10.25 41 1,419 110 1985 d 1/03/85 363 19.96 14,370 8.52 6,134 20,504

See footnote at end of table.

Ground-Water Flow Model 69 Table 9. Stress period lengths and specified inflows from The Forks (Deep Creek and West Fork) to the Mojave River, southern California, 1931–99—Continued Inflow from Deep Creek Inflow from West Fork Combined mean Stress daily Stress Number Year period Start date discharge period No. of days 3 3 (wet/dry) (ft /s) (acre-ft) (ft /s) (acre-ft) at The Forks (acre-ft) 111 1986 w 1/01/86 24 399.38 19,012 159.25 7,581 26,592 112 1986 d 1/25/86 341 17.13 11,584 7.40 5,006 16,590 113 1987 w 1/01/87 2 524.00 2,079 31.00 123 2,202 114 1987 d 1/03/87 363 12.88 9,272 1.64 1,178 10,450 115 1988 w 1/01/88 3 151.67 902 98.67 587 1,490 116 1988 d 1/04/88 363 13.96 10,049 5.22 3,761 13,810 117 1989 w 1/01/89 3 146.00 869 169.33 1,008 1,876 118 1989 d 1/04/89 362 8.60 6,175 3.15 2,265 8,440 119 1990 w 1/01/90 1182 4.30 1,553 0.95 341 1,894 120 1990 d 7/02/90 183 4.30 1,561 0.95 343 1,904 121 1991 w 1/01/91 20 489.35 19,412 66.45 2,636 22,048 122 1991 d 1/21/91 345 18.21 12,464 5.94 4,062 16,526 123 1992 w 1/01/92 33 525.30 34,383 365.46 23,921 58,304 124 1992 d 2/03/92 333 25.92 17,122 16.09 10,630 27,752 125 1993 w 1/01/93 108 1,305.26 279,606 553.95 118,665 398,271 126 1993 d 4/19/93 257 30.15 15,369 29.43 15,003 30,372 127 1994 w 1/01/94 8 385.25 6,113 189.25 3,003 9,116 128 1994 d 1/09/94 357 20.27 14,353 4.26 3,017 17,370 129 1995 w 1/01/95 92 680.18 124,120 294.74 53,784 177,903 130 1995 d 4/03/95 273 28.42 15,390 10.75 5,820 21,211 131 1996 w 1/01/96 11 804.73 17,558 240.16 5,240 22,798 132 1996 d 1/12/96 355 17.54 12,347 5.00 3,520 15,867 133 1997 w 1/01/97 7 575.14 7,985 272.71 3,786 11,772 134 1997 d 1/08/97 358 15.15 10,758 8.60 6,107 16,865 135 1998 w 1/01/98 104 533.36 110,021 194.45 40,111 150,132 136 1998 d 4/15/98 261 28.09 14,542 10.00 5,178 19,720 137 1999 w 1/01/99 1182 4.64 1,673 0.89 322 1,995 138 1999 d 7/02/99 183 4.64 1,682 0.89 323 2,006

1 Years when combined mean daily discharge did not exceed 200 cubic foot per second were divided into two stress periods of equal length.

70 Simulation of Ground-Water Flow in the Mojave River Basin, California Discharge records for the Mojave River gages less total discharge, the simulated recharge was less (fig. 4) indicate that during periods when mean daily than measured discharge in the Centro and Baja model discharge is greater than or equal to 200 ft3/s at the subareas. Although total discharge is higher for the dry headwaters, streamflow commonly extends significant periods, usually the average discharge rate is not high distances downstream into the Centro and Baja model enough for streamflow to extend past the Alto model subareas. During periods when mean daily discharge subarea resulting in simulated recharge to the Alto was less than 200 ft3/s, streamflow normally did not model subarea being higher than measured recharge. extend past the Alto model subarea. The wet period for Pumping in the Mojave River ground-water each calendar year was defined for this model as the basin varies on a seasonal basis; streamflow-based number of days during which the combined mean daily stress periods do not match the seasonal pumping discharge at The Forks was greater than or equal to cycles. The use of stress periods of shorter durations 200 ft3/s. The remaining number of days in the year (daily or weekly) would be required to simulate the defined the dry period. For years when the mean daily variability of streamflow and pumping more accurately. discharge did not exceed 200 ft3/s, the year was divided This would require large amounts of computer- into two equal stress periods. The average discharge for processing time and computer storage which are each wet or dry period was computed by dividing the beyond the scope of this project. Modeling streamflow total discharge for each period by the number of days variability was deemed of greater importance than in the period. For example, figure 26 shows how stress modeling seasonal pumping cycles; therefore, for the periods 47 and 48 were defined on the basis of dis- purposes of this study, pumping was assumed to be charge at the headwaters for calendar year 1954. The constant on an annual basis. values assigned for the stage were 5.0 ft for the wet Calibration of the transient-state model was done stress period, and 0.25 ft for the dry stress period, and goodness-of-fit determined by comparing simu- except for those years when the mean daily discharge lated hydraulic-head contours for 1992 with measured did not exceed 200 ft3/s. For those years (1953, 1960, water-level contours, long-term (1931–94) and short- 1981, and 1990) (table 9), the stage value for the wet term (1992–94) simulated hydraulic heads in relation stress period was 1.0 ft. to measured water levels, and streamflow hydrographs. The mean daily discharge rate used to define the In addition, simulated water-budget components of wet and dry periods (200 ft3/s) was determined during recharge and discharge were compared with published model calibration and is referred to in this report as the values (table 3). “wet-period cutoff.” If a lower value is used for the wet- The simulated hydraulic-head contours for layer period cutoff, the number of days in the wet period 1 for the end of 1992 were compared with the measured increases resulting in lower average discharge rates, water-level contours for autumn 1992 (fig. 27). In gen- greater total discharge for the wet periods, and a lower eral, the simulated results are in good agreement with total discharge for the dry periods. When this lower the measured data except for the Oeste model subarea average computed discharge rate was input into the for which simulated hydraulic heads show a pumping model for the wet-period cutoff, the simulated stream- depression near El Mirage dry lake that is not shown by flow for the wet period did not extend as far down- the measured data. Water-level measurements made in stream as actual streamflow in the Mojave River, as 1998, however, do indicate the existence of such a indicated by the gage data. As a result, the model sim- depression (Smith and Pimentel, 2000). In the Transi- ulated too much recharge in the Alto model subarea and tion zone model subarea, measured water-level data for too little recharge in the Centro and Baja model subar- 1992 indicate a depression near the Southern California eas compared with measured values for wet periods. If Logistics Airport that is not well simulated. Pumpage a higher value is used for the wet-period cutoff, the data may be underestimated for this area. number of days in the wet period decreases, resulting in Measured water levels and simulated hydraulic higher average discharge rates and lower total dis- heads for 1992 are shown in figure 28. The overall charge for the wet period, and a higher total discharge RMSE equaled 23.6 ft and the measured minus simu- for the dry periods. The higher average discharge rates lated ME equaled −3.3 ft. The largest RMSE value, for the wet periods resulted in the simulated streamflow 34.3 ft, was in the Oeste model subarea (fig. 28). The extending further downstream; however, because the correlation coefficient between the measured water lev- wet period defined by the higher wet-period cutoff has els and simulated hydraulic head for 1992 was 0.999.

Ground-Water Flow Model 71 10,000 A (257) Days of flow 15 Average discharge – In cubic feet per second WET PERIOD Summation of: 1,000 = (38) – Days Q Average discharge – (in this 9 3 wet Inflow = 1.8 x 10 ft example for wet period) s Seconds

200 ft3/s 100 DRY PERIOD

Days = (327) 8 3 Inflow = 6.4 x 10 ft 10

(2) 356 (3) (4) (29) MEAN DAILY DISCHARGE, IN CUBIC FEET PER SECOND MEAN DAILY 1 1,676 548 444

(18) (3) (17) (32) (257) 7 63 64 67 15 0.1 10,000 B

9 3 – Inflow 1.8 x 10 ft 1 day 3 Qwet = = = 548 ft 1,000 Days 38 days 86,400 s

38 100 days – 327 days Stress period 47 8 3 – Inflow 6.4 x 10 ft 1 day 3 10 Qdry = = = 22.7 ft Feb. 7 Days 327 days 86,400 s VERAGE DISCHARGE (Q), IN CUBIC FEET PER SECOND A Stress period 48

1 JAN FEB MAR APR MAY JUNE JULY AUG SEPT OCT NOV DEC 1954 Figure 26. Mean daily discharge and average discharge for 1954 in the ground-water flow model of the Mojave River ground- water basin, southern California.

72 Simulation of Ground-Water Flow in the Mojave River Basin, California 1,400

1,500 20 MILES

116 30' 1,700 del of the Mojave River

1,600 1,700

? 20 KILOMETERS 10

1,600 EXPLANATION

? 1,800 Measured for water level in autumn 10 ater-level contour (measured 1992) – ater-level ell – W Shows altitude of water level. Contour interval is 100 feet. Dashed where approximate. Queried where uncertain. Datum is sea level Hydraulic-head contour (simulated 1992) – Shows altitude of simulated hydraulic head. Contour interval is 100 feet. Datum sea level W 1992 Boundary of model layer 1 Generalized direction of ground-water flow Inactive areas Model faults

? 45' Springs Newberry

1,800 2,800 0 0 2,800

1,900 R 1 E R 2 E R 3 E R 4 E R 5 E R 6 E

3,200 3,100

1,900 ? Valley

Lucerne 3,000

2,000 2,000 2,950

R 1 W 2,100 3,100

117

?

2,500 3,000 ? ?

Barstow 2,900

2,900 ? 2,850

Apple Valley

2,200 2,200 2,800 2,000

2,700 R 3 W 2 W 2,300 2,300

2,800

? 2,100 2,000 Victorville

15' 3,000 2,200 Mojave River

?

1,900

1,900 ? 2,600

2,100

R 4 W 2,500 ?

? 2,000 2,100 Hesperia

2,400 2,700

2,600 5 W R 4 W R 3 W R 2 W R 1 W R 1 E ?

? ?

R 5W 2,800 Helendale

2,500

2,400 2,700 117 30' R 6 W 2,700 2,800

2,9002,900 3,100

R 41 E 3,000 2,700

2,900 W 7 R R 6 W R

R 7 W ? Measured water levels, autumn 1992, and simulated hydraulic-head contours, for model layer 1 of the ground-water flow mo

? R40E R 42 E R43 E R44 E R 45 E R 46 E R 47 E 35 T8N T9N 32 S T4N 12 N T3N 10 N T5N T7N T6N T T T 11 N T 11 T 34 45' 34 30' Figure 27. ground-water basin, southern California.

Ground-Water Flow Model 73 Long-term (1931–94) simulated hydraulic heads after about 1950 (fig. 29). Discrepancies between the were compared with measured water levels using simulated hydraulic heads and the measured water hydrographs from 42 wells (14 in the Alto model sub- levels in the regional aquifer may be due, in part, to the area, 3 in the Transition zone model subarea, 6 in the assumption that pumping is constant for each calendar Centro model subarea, and 19 in the Baja model sub- year. For some areas, the match could be improved with area) and are presented in Appendix 1. Eleven of the 42 better estimates of the quantity and distribution of long-term hydrographs are shown in figure 29. The pumping. hydrographs in figure 29 are, for the most part, grouped Simulated hydraulic heads for a short-term by wells in the floodplain aquifer and wells in the (1992–94) period were compared with measured water regional aquifer. In general, the simulated hydraulic levels using hydrographs from 26 multiple-well moni- head for wells in the floodplain aquifer matched the toring sites (9 in the Alto model subarea, 3 in the Tran- measured water-level decline, which began in the mid- sition zone model subarea, 7 in the Centro model 1940’s, and the measured water-level rises, which subarea, and 7 in the Baja model subarea) installed by resulted from floodflow recharge. Simulated hydraulic the USGS. The short-term hydrographs are presented heads for wells in the regional aquifer generally follow in Appendix 2; these data were used to calibrate the the measured water-level trends and start to decline vertical conductance between the floodplain and the

3,000 Este (1 value) Subarea – (Number is RMSE value) Oeste (34.3)

Transition zone (23.9) Alto (18.6) 2,500

Centro (24.0)

Harper Lake (25.4)

2,000

Coyote Lake (1 value)

Baja (24.8)

SIMULATED WATER LEVEL, IN FEET WATER SIMULATED Afton Canyon (31.0) 1,500

1:1 Correlation line 1,000 1,000 1,500 2,000 2,500 3,000

MEASURED WATER LEVEL, IN FEET

Figure 28. Measured water levels and simulated hydraulic head and the root mean square error (RMSE) for each model subarea of the Mojave River ground-water basin, southern California, for 1992 transient-state conditions. (See figure 18 for location of model subareas.)

74 Simulation of Ground-Water Flow in the Mojave River Basin, California

regional aquifers by adjusting estimates of vertical-to- Simulation Results horizontal anisotropy. Lower values of anisotropy result in greater hydraulic-head differences between For this study, the simulated hydrologic budgets the model layers, and higher values result in smaller for 1930 (steady state), 1994, and 1931–90 average head differences. The greatest measured vertical-head were used to describe the flow characteristics of the differences (as much as 25 ft) were in the Transition model subareas; the hydrologic budgets are presented zone model subarea; the anisotropy of model layer 1 in table 10, and figures 32 and 33. The 1930 hydrologic was calibrated to a value of 0.0001 (fig. 21) to match budget represents the state of the ground-water system these measured differences (Appendix 2). prior to significant ground-water development. The 1994 hydrologic budget represents the state of the Simulated streamflow data for the wet and dry ground-water system after 64 years of water-resources stress periods at the Lower Narrows, Barstow, and development in the basin. The average 1931–90 hydro- Afton Canyon were compared with average measured logic budget represents the 60-year adjudication streamflow discharge data for the period 1931–94 period. The 1931–90 period was chosen to determine (fig. 30). In general, the model simulations reflect the the average annual obligation of runoff and underflow streamflow conditions at the Barstow and Afton from one model subarea to another in accordance with Canyon gages matching the peak wet-stress periods, the Stipulated Judgement in The City of Barstow et al. flow rate, and times of no flow (fig. 30B,C). At the vs. The City of Adelanto et al. (Mojave Basin Area Lower Narrows, the simulated and measured stream- Watermaster, 1996a). flow discharges for the dry stress periods were about Results of the model simulations show that total 30 ft3/s between 1931 and about 1950 (fig. 30A). Since inflow, or recharge, for 1930 steady-state conditions 1950, the simulated and measured streamflow dis- was about 74,320 acre-ft (table 10). About 62,680 charge for dry stress periods has decreased with time; acre-ft, or 84 percent of the total recharge, was from the lowest amount of streamflow discharge for a dry stream leakage from the Mojave River and about stress period occurred in 1990 and was less than 11,640 acre-ft, or about 16 percent of the total recharge, 10 ft3/s. Any differences between the simulated and was from mountain-front recharge. Total outflow, or measured results may have been caused by the manner discharge, was about 74,000 acre-ft, most of which was in which the stress periods were defined; in order to discharge owing to evapotranspiration. Evapotranspi- better match times of flow in the upstream model sub- ration was about 52,880 acre-ft, or 71 percent of the areas, daily stress periods may be required. total discharge; stream leakage from the Mojave River The volumetric differences (gage data minus was about 16,820 acre-ft, or 23 percent of the total dis- model simulated values) for the transient-state simula- charge and drains about 4,280 acre-ft, or 6 percent of the total discharge (table 10). Base flow is included in tion between the simulated and measured streamflow, the total amount of stream leakage from the river. The or discharge at the Lower Narrows, Barstow, and Afton distribution of recharge and discharge by model sub- Canyon gaging stations are presented in figure 31. The area is presented in figure 32. The difference between model underestimates measured streamflow at the recharge and discharge is about 330 acre-ft (0.4 percent gages for the entire transient-state simulation by of total discharge). Theoretically, under steady-state 1,400 acre-ft, or 0.04 percent, at the Lower Narrows; conditions, recharge should equal discharge; any dif- 49,400 acre-ft, or 4 percent, at Barstow; and ferences may be due to the accumulation of small 70,800 acre-ft, or 38 percent, at Afton Canyon. The numerical errors in the model and to the rounding of underestimates of streamflow at the Lower Narrows large numbers. during the early 1990’s (fig. 31A) may be related to the Total simulated recharge to the basin for 1994 2 use of a constant stream conductance value (0.8 ft /s) was about 117,520 acre-ft (table 10). Most of the for the dry stress periods. Most of the underestimation recharge was from stream leakage (about 61,600 for Afton Canyon was for 1969 (fig. 31C), a large- acre-ft, or 52 percent of the total recharge), irrigation- stormflow year. In addition, inaccuracies in measured return flow (about 30,780 acre-ft, or 26 percent), and gaged streamflow and in estimated ungaged runoff mountain-front recharge (about 11,640 acre-ft, or (total estimated ungaged runoff is about 10 percent). Total discharge from the basin was about 558,370 acre-ft, table 2) probably contributed to the 216,900 acre-ft (table 10). Most of the discharge was underestimation of streamflow. attributable to pumpage (about 194,400 acre-ft, or

Ground-Water Flow Model 75 R 40 E R 41 E 117°30' R 42 E R 43 E R 44 E117°15' R 45 E R 46 E 117°00' R 47 E R 1 E 10 20 30 4050 60 70 80 90 100 110 120

MODEL COLUMN 2,075 T 32 S 11N/3W-28R1,2 10 2,050 2,025 R7 W R 6 W R 5W R 4 W R 3 W R 2 W 28R1R 1 W 2,000 Regional 28R2 20 1,975 T 11 N 2,175 10N/3W-35N1 Harper 28R1,2 2,150 29R1 Lake 35° 30 30A1,2 2,125 00' 2,100 Regional 2,075 T 10 N 193040 50 60 70 80 90 2000 40

2,475 10262500 8N/4W-31R1 20M1-2 4K1-3 2,450 3H1,5 35N1 3E1-3 4R 2-4

2,425 IRON MTN 1R5-7 11R2 T9N Barstow 12N 4-7 10L1 2,400 9D5-8 Floodplain 19B1 4M 4-7 15H1 2,375 13R1 22D1 193040 50 60 70 80 90 2000 11M1,11 13E2 16F1-4

21H1 T8N River 12Q1 70 20N1 2,080 21M1-4 9N/1W-11M1,11 2,055 11M1

MODEL ROW 20Q7-12 34° 11M11 45' Helendale 2,030 80 31R1 2,005 Floodplain 1,980

Mojave 1,900 T7N 30C1 90 23R1-3 1,875 13E2 El Mirage 24R5-8 15H1 Lake 1,850 Floodplain 1,825 9N/1E-13E2, 15H1 100 1,800 T6N El Mirage 2,850 5N/5W-22E1,2 Adelanto 2,825 10261500 110 2,800 Apple Regional 22E1 2,775 22E2 11P1,3 Valley 13D1 2,750 T5N Southern 22E1-3 California 14D1-4 14G1 193040 50 60 70 80 90 2000 22J1 YEAR 34° 120 Logistics Victorville 30A1-3 22E1,2 30' Airport 36N1 24N1 Lucerne 1D2 27E3-6 23R1 Lake 1C2-5 6A1 3A2-5 1M1 6D1,2 12A1-3 130 8N2 Lucerne 8G1 18R1,R2,R3 T4N 18E1 10R1 Rabbit Valley Lake Hesperia 19G2-6 20P1

140 31L6-9 2,850 2,875 Floodplain 10260950 4N/3W-1M1 2,825 2,850 10261100 2,800 2,825Deep 150 West Fork C T3N 5N/4W-36N1 reek 2,775 Mojave River 2,800 4N/4W-1D2 Regional 2,750 10260500 2,875 193040 50 60 70 80 90 2000 193040 50 60 70 80 90 2000 160 R 7 W R 6 W R 5 W R 4 W R 3 W R 2 W R 1 W R 1 E Figure 29. Measured ground-water levels and simulated hydraulic head for selected wells in the Mojave River ground-water basin, southern California.

76 Simulation of Ground-Water Flow in the Mojave River Basin, California R 2 E116°45' R 3 E R 4 E116°30' R 5 E R 6 E 130 140 150 160170 180 190 200 EXPLANATION MODEL COLUMN T 32 S 10 200 Coyote Model Grid – Lake Active areas 20 10263000 Inactive areas T 11 N

30 Dry lake (playa) 35° Rabbit 00' Lake 15Q1 21A1 T 10 N Fault 1,760 40 10N/3E-21A1, -15Q1 1,735 Model fault 3K 5-9 27J 1-5 3G 6-9 1,710 21A1 3A2 15Q1 Boundaries – 11H1,3 3D1,2 3K1 1,685 50 18F1 12N1 Regional T9N Model layer 1 19E1 1,660 14N1,2 22R 4-7 193040 50 60 70 80 90 2000 YEAR 20Q1 20K1 34D1 Model layer 2 Newberry 60 34N1 31K1 Model subarea Springs 7E1

Troy T8N Well and abbreviated State Lake 70 ash well number – W 34 ° 20K1 45' Used to construct hydrographs Kane 80 shown in Appendix 1 1,900 1,800 12A1-3 9N/2E-20Q1, -20K1 8N/4E-7E1 Multiple-well monitoring site 1,775 1,850 used to construct hydrographs 1,750 T7N shown in Appendix 2 20Q1 MODEL ROW 90 1,800 1,725 Regional 20K1 Regional 10260500 Gaging station and number 1,750 1,700 193040 50 60 70 80 90 2000 193040 50 60 70 80 90 2000 YEAR YEAR 100 Explanation of hydrograph

2,080 9N/1W-11M1,11 110 2,055 11M1 11M11 2,030 2,005 Floodplain 120 34° 1,980 HYDRAULIC HEAD, IN 30' FEET ABOVE SEA LEVEL 193040 50 60 70 80 90 2000 YEAR

Floodplain 130 Aquifer designation Model layer 1 0 10 20 MILES Model layer 2

0 10 20 KILOMETERS 140 11M1 11M11 Data point for well indicated

150

160

Figure 29.—Continued.

Ground-Water Flow Model 77 1,000 A

100

10

Simulated Measured

1 1,000 B

100

10 DISCHARGE, IN CUBIC FEET PER SECOND

1 1,000 C No record of No record measured of measured streamflow streamflow

100

10

1 1931 1934 1937 1940 1943 1946 1949 1952 1955 1958 1961 1964 1967 1970 1973 1976 1979 1982 1985 1988 1991 1994 1997 2000 YEAR

Figure 30. Measured and simulated discharge in the Mojave River ground-water basin, southern California, from (A) the Lower Narrows near Victorville (gaging station 10261500), 1931–99; (B) the Mojave River at Barstow (gaging station 10262500), 1931–99; and (C) the Mojave River at Afton Canyon (gaging station 10263000), 1931–99 (measured data are for years 1931, 1953–78, and 1981–99).

78 Simulation of Ground-Water Flow in the Mojave River Basin, California 30,000 A

20,000

10,000

0

-10,000

-20,000 Wet stress period DISCHARGE, IN ACRE-FEET Dry stress period

-30,000

-40,000

-50,000

80,000 B 60,000

40,000

20,000

0

-20,000 DISCHARGE, IN ACRE-FEET

-40,000 Wet stress period

Dry stress period -60,000

-80,000

-100,000 1931 1936 1941 1946 1951 1956 1961 1966 1971 1976 1981 1986 1991 1996 2001 YEAR

Figure 31. Volumetric difference between measured and simulated discharge in the Mojave River ground-water basin, southern California, for (A) the Lower Narrows near Victorville (gaging station 10261500), 1931–99; (B) the Mojave River at Barstow (gaging station 10262500), 1931–99; and (C) the Mojave River at Afton Canyon (gaging station 10263000), 1931–99 (measured data are for years 1931, 1953–78, 1981–99).

Ground-Water Flow Model 79 90 percent), evapotranspiration (about 15,760 acre-ft, pumpage has resulted in declines in ground-water lev- or 7 percent), and stream leakage to the Mojave River els which, in turn, have resulted in the depletion of (about 5,980 acre-ft, or 3 percent) accounted for most ground-water storage (treated as recharge on of the remainder. Recharge and discharge are presented figure 32A) in all the model subareas. in figure 32 by model subarea. The difference between Simulated total average recharge to the ground- recharge and discharge, which is the contribution from water system for the adjudication period of 1931–90 ground-water storage, was about 99,370 acre-ft was about 150,310 acre-ft/yr (a total volume of (46 percent of total discharge). The declining water lev- 9.0 million acre-ft) (table 10), slightly more than twice els in the basin (fig. 29) are indicative of this change in the steady-state recharge computed for 1930. Most of storage. the recharge was from stream leakage Results of the simulations for 1930 and 1994 are (87,410 acre-ft/yr, or 58 percent of the total recharge), presented on figure 32. It was assumed that 1930 repre- irrigation-return flow (47,220 acre-ft/yr, or 31 percent), sented pre-development conditions (steady state) and and mountain-front recharge (11,650 acre-ft/yr, or therefore pumpage was not simulated for that year. As 8 percent). For this same period, the average total dis- shown by the simulation results for 1994 (fig. 32), charge from the ground-water system was about pumping occurred in most of the model subareas 189,720 acre-ft/yr (a total volume of 11.4 million resulting in irrigation-return flows for most of the acre-ft) (table 10). Most of the discharge was attribut- model subareas. Recharge from the Mojave River able to pumpage (151,740 acre-ft/yr, or 80 percent of (stream leakage) in the Alto, Centro, and Baja model the total discharge), evapotranspiration (24,670 subareas decreased between 1930 and 1994 (fig. 32A). acre-ft/yr, or 13 percent), and discharge to the Mojave The total simulated evapotranspiration decreased from River (stream leakage) (10,550 acre-ft/yr, or 6 percent). about 52,880 acre-ft/yr in 1930 to about The distribution of simulated annual recharge and 15,760 acre-ft/yr in 1994 (table 10). The increase in discharge by model subarea is presented in figure 33.

70,000 C Wet stress period 60,000 Dry stress period

50,000

40,000 No record of No record of measured measured streamflow streamflow 30,000

20,000 DISCHARGE, IN ACRE-FEET

10,000

0

-10,000 1931 1936 1941 1946 1951 1956 1961 1966 1971 1976 1981 1986 1991 1996 2001 YEAR Figure 31.—Continued.

80 Simulation of Ground-Water Flow in the Mojave River Basin, California otal T Afton Canyon Lake Coyote e), 1994, and 1931-90 average (adjudication Baja Lake Harper 1930 (steady-state) Este Oeste Alto Zone Transition Centro . Simulated hydrologic budgets for model subareas of the Mojave River ground-water basin, southern California, 1930 (steady stat otal...... 1,019 2,858 27,563 12,931 17,468 2,934 24,872 980 384 73,995 otal...... 1,035 2,893 27,858 12,943 17,469 2,934 24,873 980 384 74,321 T T See footnotes at end of table DrainsEvapotranspirationHead-dependent boundaryStream leakage between subareasFlow 0 0 921 98 0 0 0 2,342 516 13,597 0 4,321 0 10,424 0 2,507 9,645 0 12,069 5,399 0 0 0 0 0 0 16,453 0 0 1,269 2,934 0 0 255 0 6 332 7,144 0 0 52,875 725 0 na 26 0 26 4,279 26 16,815 Mountain front Stream leakage between subareasFlow 1,035 0 0 1,941 952 7,763 0 3,008 17,087 0 3,721 9,222 2,762 0 14,707 2,934 0 0 2,721 21,505 647 721 259 0 229 155 0 na 62,676 11,645 Discharge Recharge period) Table 10. na, not applicable] are in acre-feet per year. Values values. for 1931–90 are average [Values

Ground-Water Flow Model 81 otal T Afton Canyon tate), 1994, and 1931–90 average (adjudication Baja Coyote Lake Lake Harper 1994 24 0 7,110 3,922 6,662 1,052 11,988 16 0 30,774 513 6,348 76,745 18,902637 33,172640 4,328 10,258 4,343 36,035 48,452 35,782 -14 56 -14 18,016 17,524 0 4,916 4,911 194,446 34,340 34,261 924 927 192 192 99,374 98,566 Este Oeste Alto Zone Transition Centro 1 3, 4 3 2 Simulated hydrologic budgets for model subareas of the Mojave River ground-water basin, southern California, 1930 (steady s otal ...... 1,061 3,534 48,499 30,580 26,293 5,342 18,342 423 312 117,524 T otal...... 1,698 7,862 84,534 30,566 44,309 10,258 52,682 1,347 504 216,898 See footnotes at end of table. T and discharge Sewage pondsSewage Mountain front Septic tank Stream leakage between subareasFlow 1,035 0 1,940 0 2 0 0 1,594 7,758 0 0 2,668 0 0 21,152 9,811 3,257 686 22,547 0 1,627 168 2,264 15,740 4,290 0 0 0 0 3,117 647 2,004 0 586 149 259 160 0 0 0 0 152 na 0 0 11,638 61,595 3,536 0 9,981 Irrigation return Irrigation Pumpage DrainsEvapotranspirationHead-dependent boundaryStream leakage between subareasFlow 0 0 1,116 69 0 0 1,514 0 0 4,308 3,481 0 0 1,301 8,977 0 0 0 6,859 1,839 0 1,386 0 0 0 2,439 0 1,068 1,158 0 0 696 0 0 0 2,004 0 305 0 0 0 15,760 595 na 47 152 0 5,981 47 664 Difference between recharge between recharge Difference Storage depletion Discharge Recharge period)—Continued Table 10.

82 Simulation of Ground-Water Flow in the Mojave River Basin, California otal T Afton Canyon tate), 1994, and 1931–90 average (adjudication and 50 percent return elsewhere. Baja Coyote Lake Lake Harper 1931–90 average Centro Zone ransition T 12 0 20,900 7,270 9,671 330 9,029 12 0 47,224 129 2,196 55,835 19,637172 32,654172 1,598 8,990 1,604 6,307 32,253 6,212 1,843 43 1,813 7,952 7,811 6,474 0 6,482 14,463 151,737 14,490 762 765 -167 -170 39,404 39,179 Este Oeste Alto 1 3, 4 3

2 Simulated hydrologic budgets for model subareas of the Mojave River ground-water basin, southern California, 1930 (steady s alues of storage differ as a result of accumulation of small, consistent errors in the model and rounding of large numbers. as a result of accumulation small, consistent errors in the model and rounding large alues of storage differ otal ...... 1,047 2,990 66,509 28,967 36,928 3,666 24,987 775 1,332 150,311 otal...... 1,219 4,588 72,816 30,810 44,880 10,140 39,450 1,537 1,165 189,715 Irrigation return for 1931–50 calculated from Hardt’s (1971) adjusted net pumpage: 60 percent return in the Alto model subarea (1971) adjusted net pumpage: 60 percent return in the return for 1931–50 calculated from Hardt’s Irrigation Alto model subarea, 50 percent elsewhere. 1931–50 net pumpage adjusted by 40 percent in the indicates storage accretion. storage value indicates storage depletion; negative storage value Positive V T T 1 2 3 4 and discharge Irrigation return Irrigation Sewage pondsSewage Mountain front Septic tank Stream leakage between subareasFlow 1,035 0 1,941 0 0 0 0 7,763 1,049 0 0 2,886 0 0 32,593 2,367 3,745 90 17,845 0 2,279 17 1,179 23,799 3,336 0 0 0 0 2,921 647 12,015 0 504 375 259 0 170 0 0 0 1,162 na 0 0 11,645 87,414 1,644 0 2,384 Pumpage DrainsEvapotranspirationHead-dependent boundaryStream leakage between subareasFlow 0 0 995 95 0 0 0 2,088 304 5,187 0 4,446 0 0 8,785 1,717 7,348 0 6,508 5,511 0 671 0 0 0 207 0 0 3,653 1,340 1,150 0 793 0 0 2,204 1 539 0 0 701 0 24,672 504 122 na 0 10,552 504 2,251 Storage depletion Difference between recharge between recharge Difference Discharge Recharge Table 10. period)—Continued

Ground-Water Flow Model 83 MODEL SUBAREA

ESTEOESTE ALTO TRANSITION CENTRO HARPER BAJA COYOTE AFTON ZONE LAKE LAKE CANYON 90,000 A

80,000 Recharge Storage depletion Mountain-front 70,000 Interzonal Sewage ponds Stream leakage Irrigation return Septic tank 60,000

50,000

40,000

30,000

20,000

10,000

0 90,000 B Discharge 80,000 Storage accretion Evapotranspiration Interzonal Drains 70,000 Stream leakage Pumpage Underflow at 60,000 Afton Canyon

50,000

40,000

DISCHARGE, IN ACRE-FEET PER YEAR30,000 RECHARGE, IN ACRE-FEET PER YEAR

20,000

10,000

0 1930 1994 1930 19941930 1994 1930 1994 1930 1994 1930 1994 1930 1994 1930 1994 1930 1994 YEAR

Figure 32. Ground-water recharge to, and discharge from, the model subareas of the Mojave River ground-water basin, southern California, 1930 and 1994. A, Recharge. B, Discharge.

84 Simulation of Ground-Water Flow in the Mojave River Basin, California MODEL SUBAREA

TRANSITION HARPER COYOTE AFTON ESTEOESTE ALTO CENTRO BAJA ZONE LAKE LAKE CANYON 80,000 A Recharge 70,000 Storage depletion Mountain-front Interzonal Sewage ponds Stream leakage 60,000 Irrigation return Septic tank

50,000

40,000

30,000

RECHARGE, IN ACRE-FEET PER YEAR 20,000

10,000

0 80,000 B Discharge Storage accretion Evapotranspiration 70,000 Interzonal Drains Stream leakage 60,000 Pumpage Underflow at Afton Canyon 50,000

40,000

30,000

20,000 DISCHARGE, IN ACRE-FEET PER YEAR

10,000

0

Figure 33. Average annual ground-water recharge to, and discharge from, the model subareas of the Mojave River ground- water basin, southern California, 1931–90. A, Recharge. B, Discharge.

Ground-Water Flow Model 85 The difference between recharge and discharge, which Oeste model subarea to the Alto model subarea at a is the contribution from ground-water storage, aver- lower flow rate (1,162 acre-ft/yr in 1930 and aged about 39,400 acre-ft/yr (a total of 2.4 million acre- 315 acre-ft/yr in 1994). In 1994, there was a reversal of ft) (table 10). This value is slightly different from the flow from the Transition zone model subarea into the storage value for the simulated change in storage Oeste model subarea. Ground water continued to flow (39,180 acre-ft/yr) because of the accumulation of downstream from the Transition zone model subarea small numerical errors in the model and to the rounding into the Centro model subarea; however, the flow rates of large numbers. This change in storage is indicated by decreased (2,444 acre-ft/yr in 1930 and 720 acre-ft/yr declining ground-water levels in the basin (fig. 29). in 1994). The ground-water flow rate from the Centro The simulated total cumulative volume of water to the Harper Lake model subarea increased from 2,934 pumped from the ground-water basin for 1931–90 was to 4,290 acre-ft/yr; this increase in flow rate was caused about 9.1 million acre-ft, or an average of about by pumping in the Harper Lake area. Ground water 151,740 acre-ft/yr (table 10). Of this total pumpage, continued to flow downstream from the Centro model most was contributed by irrigation-return flow (about subarea into the Baja model subarea; however, the flow 47,220 acre-ft/yr, or 31 percent), ground water from rates decreased (2,146 acre-ft/yr in 1930 and 1,662 storage (about 39,180 acre-ft/yr, or 26 percent), a acre-ft/yr in 1994). Ground-water flow was from the decrease in evapotranspiration compared with steady Baja to the Coyote Lake model subarea in 1930; how- state (about 28,200 acre-ft/yr, or 18 percent), and an ever, there was a reversal of flow in 1994. Ground water increase in recharge from stream leakage compared exited the basin from the Afton Canyon model subarea with steady state, (about 24,740 acre-ft/yr, or 16 per- at a higher flow rate in 1994 than in 1930 (26 acre-ft/yr cent). The remainder was from sewage ponds, septic compared with 47 acre-ft/yr). tanks, and a decrease in discharge to the drains and to The simulated rates of underflow for 1931–90 the Mojave River. are the average rates for that period. The direction of ground-water flow between the model subareas for the The simulated net ground-water underflows 1931–90 period was the same as that simulated for between model subareas for 1930 and 1994 and the 1994, except between the Transition zone and the Oeste average underflow for 1931–90 (the adjudication model subareas where underflow again reversed direc- period) is shown in figure 34. The results of the 1930 tion, flowing from the Oeste model subarea to the Tran- simulation indicate that ground water flowed from the sition zone (fig. 34). A comparison between the Este and Oeste model subareas into the Alto model sub- simulated 1931–90 average and the steady-state rates area, from the Oeste model subarea into the Transition of ground-water underflow indicates that underflow zone model subarea, and downstream through the between the Centro and Harper Lake model subareas Centro and Baja model subareas; only a small amount was about 400 acre-ft/yr less for the steady state; of flow (about 26 acre-ft/yr) exited from the Afton underflow between the Transition zone and the Centro Canyon model subarea. Ground water moved from the model subareas was about 880 acre-ft/yr less for 1931 Centro to the Harper Lake model subarea and from the –90; and underflow between the Centro and Baja model Baja to the Coyote Lake model subarea. The ground subareas about 680 acre-ft/yr less for 1931–90; there water that flowed toward the dry lakes exited the basin was a reversal of flow between the Baja and Coyote as evapotranspiration (simulated as drain discharge). Lake model subareas (a net change of about Note that the underflow from the Centro to the Harper 760 acre-ft/yr). The average 1931–90 underflow exit- Lake model subarea actually flows through two bound- ing the flow system from the Afton Canyon model aries on either side of Iron Mountain (fig. 34) but that subarea was about 480 acre-ft/yr greater than the the separate flow rates were not distinguished for this steady-state value. study; therefore, the flow rates through the two bound- aries are shown as a single flow rate for illustrative purposes. Steady-State Ground-Water Flow Directions and In 1994, ground water continued to flow from the Travel Times Este model subarea as underflow into the Alto model subarea; however, there was a slight increase in flow The computer program MODPATH (Pollock, rate (920 acre-ft/yr in 1930 to 1,116 acre-ft/yr in 1994) 1994) was used in this study to simulate the direction (fig. 34). Ground water continued to flow from the of particles of ground-water flow and their travel times.

86 Simulation of Ground-Water Flow in the Mojave River Basin, California 26

47

504 Afton Canyon 180 20 MILES 116 30' the average underflow for 1931–90. 170 190 200

229 170

160 Baja 160 20 KILOMETERS 150 10 In acre-feet per year EXPLANATION Baja

289 547

466 140 Newberry Springs 10 45'

Average for 1931-90 Direction of underflow Lake Active area Inactive area 1930 1994 Coyote Model grid Boundary of model layer 1 Boundary of model subarea Model subarea name Underflow –

Este 3,434

3,071 3,501 0 0 Valley Lucerne Este 117

2,146

1,462 1,662 100 110 120 130

920 995

1,116 90 Barstow MODEL COLUMN 80

2,934

4,290 3,336 Apple Valley 15' 60 Helendale Victorville Harper Lake Hesperia 50 70 Centro

720

1,566 2,444 Alto

Zone 3,434 3,071

3,501 ransition Mojave River Adelanto T

117 30' MODEL ROW MODEL 93

224 395 10 20 30 40 110 Oeste

946

315 1,162 Simulated underflow between model subareas of the Mojave River ground-water basin, southern California, for 1930 and 1994, 120 110 100 140 10 20 30 40 50 60 70 80 90 150 130 160 35 34 45' 34 30' Figure 34.

Ground-Water Flow Model 87 MODPATH is a three-dimensional particle-tracking main stem of the Mojave River (below The Forks) and post-processing program designed for use with output within the Alto model subarea, left the river, traveled from ground-water flow simulations obtained using north within the floodplain aquifer, and reentered the MODFLOW. The results from this program represent river at the Upper Narrows (fig. 35B); travel times for ground-water travel times and pathlines for advective particles in this model subarea were about 1,000 years. transport only. A complete description of MODPATH’s Particles for which tracking started in the river within theoretical development, solution techniques, and the Transition zone model subarea quickly left and limitations is presented by Pollock (1994). reentered the river or never left the river system at all. Two particle-tracking simulations were made for Particles for which tracking started in the river within the 1930 steady-state conditions; the first simulation the Centro model subarea either traveled to the Harper tracked mountain-front recharge and the second Lake model subarea to be discharged as evapotranspi- tracked stream leakage to the ground-water system ration, quickly left and reentered the river, or traveled (fig. 35). The mountain-front recharge particle-tracking downstream staying within the floodplain aquifer, results are presented in figure 35A. Particles were reentering the river near the Waterman Fault. Particles tracked from the mountain-front recharge-site cells for- for which tracking started in the river within the Baja ward along flowpaths in layer 1 of the model; one par- model subarea either traveled to the Coyote Lake ticle was located in the center of each cell. By using one model subarea to be discharged as evapotranspiration, particle per cell, the program allows one to infer flow quickly left and reentered the river, or left the river and directions and travel times, but no statistics can be gen- traveled southward through the Mojave Valley, reenter- erated from the results. In general, most of the particles ing at Camp Cady where ground water is forced to the traveled downstream and discharged to the river at the surface because of decreased transmissivity and Upper Narrows in the Alto and Transition zone model faulting (figs. 19 and 35). subareas upstream from the Helendale Fault. Izbicki and others (1995) analyzed the source, movement, and Evaluation of Effects of Regional-Scale Pumping age of ground water in the Alto subarea. Using carbon- 14 data from production and monitoring wells, Izbicki The complaint that resulted in the adjudication and others (1995) estimated that water in the regional of the Mojave River ground-water basin alleged that aquifer west of Victorville was recharged from 10,000 the cumulative water production upstream of the city of to 20,000 years before present. The simulated travel Barstow had overdrafted the Mojave River ground- times for mountain-front recharge to reach the area water basin (Mojave Basin Area Watermaster, 1996a). west of Victorville were about 5,000 to 6,000 years; A Physical Solution was developed to ensure that this result is in reasonable agreement with the results of downstream producers are not adversely affected by Izbicki and others (1995). The simulated travel times upstream use. The Physical Solution requires that each did not include the travel times through a thick (greater management subarea within the basin provides a spe- than 1,000 ft) unsaturated zone. cific quantity of water to the adjoining downstream For the particle-tracking simulation of stream subarea. The water supply for the city of Barstow is leakage, one particle was placed in the center of every ground water pumped from the Centro subarea. river cell of model layer 1 and tracked forward along The calibrated ground-water flow model was the flowpaths (fig. 35B). All particles for which track- used to determine the effects of pumping in the upper ing started in the West Fork of the Mojave River (fig. 1) region (Este, Oeste, Alto, and Transition zone model left the river, traveled north outside of the floodplain subareas) on the lower region (Centro, Harper Lake, aquifer, and reentered the river at the Upper Narrows Baja, Coyote Lake, and Afton Canyon model subareas) (fig. 35B). Using carbon-14 data from production and and to determine the effects of pumping in the lower monitoring wells, Izbicki and others (1995) estimated region on the upper region. The results of each simula- that water along this flow path was recharged less than tion were compared with the simulated results from the 2,400 years before present. The simulated travel times adjudication period (1931–90), or base-case period, for particles started in West Fork of the Mojave River when there was pumping in all model subareas. to reach the Upper Narrows were about 2,000 years; For the first simulation, 1931–90 pumping rates this result is in reasonable agreement with the results of were maintained in the upper region (total average Izbicki and others (1995). Particles tracked from the pumpage equaled 77,850 acre-ft/yr) with no pumping

88 Simulation of Ground-Water Flow in the Mojave River Basin, California 1,400

1,450 Afton Canyon 20 MILES

180 1,500 location of streamflow recharge cells

116 30' (B) 1,550 Shows altitude 170 190 200 Baja , and 1,700 20 KILOMETERS

160 1,600 Each interval between dots 10 Baja 150 EXPLANATION 10 Active area Inactive area Springs Newberry Model faults Boundary of model layer 1 Boundary of model subarea Simulated steady-state hydraulic head contour (model layer 1) – of simulated steady-state hydraulic head. Contour interval 50 feet. Datum is sea level Flow paths – represents 1,000 years of travel time Streamflow-recharge cell Drain cell Model grid

140 45' Coyote Lake

1,750 1,800 0

1,850 0 3,200

1,900 3,200

1,950 3,150

3,100 2,000 Valley Lucerne

2,050 3,000 2,100 3,050 117

100 110 120 130 Barstow 2,950 Este

90 MODEL COLUMN 2,900

Centro 2,200

80

2,850 2,100 Apple Valley 2,050 2,300

15' 2,950 2,300 60 2,500

2,550 Harper Lake Hesperia 50 70 Centro Victorville

3,000 Helendale Alto 2,100 zone ransition Mojave River

2,400 T 2,600 2,050 2,700 117 30' 2,450 2,800 3,100

Simulated flow paths of selected particles for steady-state (1930) conditions initially placed at mountain-front recharge cells MODEL ROW MODEL

10 20 30 40 2,850 A Oeste (A) 2,900 120

3,050 10 20 30 40 50 60 70 80 90 140

3,000 150 160 100 110 130 35 34 45' 34 30' of the ground-water flow model Mojave River basin, southern California. Figure 35.

Ground-Water Flow Model 89 1,400

1,450 Afton Canyon 20 MILES

180 1,500 116 30' 1,550 Shows altitude 170 190 200 Baja 1,700 20 KILOMETERS

160 1,600 Each interval between dots 10 150 EXPLANATION Baja 10 Active area Inactive area Springs Newberry Model faults of simulated steady-state hydraulic head. Contour interval 50 feet. Datum is sea level represents 1,000 years of travel time Streamflow-recharge cell Drain cell Boundary of model layer 1 Boundary of model subarea Simulated steady-state hydraulic head contour (model layer 1) – Flow paths – Model grid

140 45' Coyote Lake

1,750 1,800 0 3,200 0

1,900

3,200 3,150 1,950 3,100

Valley

Lucerne

3,000 2,000 3,050 117

100 110 120 130 Barstow 2,950 Este

90

MODEL COLUMN 2,100 2,900 2,850

Centro 2,200

80 2100 Apple Valley

15' 2,300 60 2,500 Victorville

2,550 Harper Lake Hesperia 50 70 2,100 2,800 Centro 2,050 2,850 3,000 2,400 2,900 Alto zone 2,950 Helendale ransition T

Mojave River 2,600 2,700 117 30'

2,450

3,100 10 20 30 40 ROW MODEL B Oeste .—Continued. 120 10 20 30 40 50 60 70 80 90 140 150 160 100 35 110 130 Figure 35 34 45' 34 30'

90 Simulation of Ground-Water Flow in the Mojave River Basin, California in the lower region. For the second simulation, the Upper Region Pumping Only 1931–90 pumping rates were maintained in the For the simulation with pumping only in the lower region (total average pumpage equaled upper region, simulated recharge to the ground-water 73,890 acre-ft/yr) with no pumping in the upper region. system from the Mojave River was equal to that for the The simulated recharge from the Mojave River, base case in the Alto and Transition zone model subar- ground-water discharge to the Mojave River, evapo- eas, less than that for the base case in the Centro model transpiration, and change in storage for the Alto, subarea (about 6,630 acre-ft/yr), and less than that for Transition zone, Centro, and Baja model subareas are the base case in the Baja model subarea (about presented in figure 36. 460 acre-ft/yr) (fig. 36). The simulated ground-water

MODEL SUBAREA TRANSITION ALTO CENTRO BAJA ZONE Upper Lower Upper Lower Upper Lower Upper Lower region region region region region region region region Base pumping pumping Base pumping pumping Base pumping pumping Base pumping pumping case only only case only only case only only case only only 40,000

30,000

20,000

10,000 RECHARGE, IN ACRE-FEET PER YEAR

0

10,000

20,000

30,000

DISCHARGE, IN ACRE-FEET PER YEAR 40,000 EXPLANATION

Storage depletion Evapotranspiration Recharge from Mojave River Storage accretion

Ground-water discharge to Mojave River

Figure 36. Average streamflow recharge from the Mojave River, ground-water discharge to the Mojave River, evapotranspiration, and storage for the analysis of regional-scale pumping effects on the Mojave River ground-water basin, southern California, 1931–90.

Ground-Water Flow Model 91 discharge to the Mojave River was about equal to depleted from storage for the base case, this resulted in discharge for the base case in the Alto and Transition a net increase in storage of about 3,770 acre-ft/yr in the zone model subareas and discharge to the Mojave River Transition zone model subarea when compared with was higher in the Centro model subarea (about the base case. In the Centro model subarea, about 4,630 acre-ft/yr) and in the Baja model subarea (about 1,230 acre-ft/yr was depleted from storage; since about 4,190 acre-ft/yr) than for the base case. Evapotranspi- 7,810 acre-ft/yr was depleted from storage for the base ration was about equal to that for the base case in the case, this resulted in about 6,580 acre-ft/yr less water Alto and Transition zone model subareas and higher being depleted from storage in the Centro subarea than that for the base case in the Centro when compared with the base case. In the Baja model (1,910 acre-ft/yr) and Baja (700 acre-ft/yr) model sub- subarea, about 11,260 acre-ft/yr was depleted from areas. In the Alto and Transition zone model subareas, storage; since about 14,490 acre-ft/yr was depleted the simulated storage depletion was approximately from storage for the base case, this resulted in about equal to the simulated storage depletion for the base 3,230 acre-ft/yr less water being depleted from storage case (fig. 36). In the Centro model subarea there was in the Baja model subarea when compared with the about 870 acre-ft/yr in storage accretion; since about base case. 7,810 acre-ft/yr was depleted from storage for the base case, this resulted in a net increase in storage of about Summary of Effects of Regional-Scale Pumping 8,680 acre-ft/yr into the Centro model subarea when compared to the base case. In the Baja model subarea In summary, the simulation with pumping only there was about 3,800 acre-ft/yr in storage accretion; in the upper region showed that there was no change in since about 14,490 acre-ft/yr was depleted from stor- storage, recharge from the Mojave River, discharge to age for the base case, this resulted in a net increase in the Mojave River, or evapotranspiration in the Alto and storage of about 18,290 acre-ft/yr into the Baja model Transition zone model subareas when compared with subarea when compared with the base case. the base case. In addition, the simulation with pumping only in the upper region showed storage accretion, a decrease in recharge from the Mojave River, an Lower Region Pumping Only increase in discharge to the Mojave River, and an For the simulation where there was pumping increase in evapotranspiration in the Centro and Baja only in the lower region, simulated recharge to the model subareas when compared with the base case. ground-water system from the Mojave River was lower These changes in the Centro and Baja model subareas in the Alto and the Transition zone model subareas are the result of no pumping in the lower region, caus- (about 6,890 and 3,200 acre-ft/yr, respectively) and ing the simulated hydraulic heads to rise throughout the higher in the Centro and the Baja model subareas lower region. (about 13,110 and 3,860 acre-ft/yr, respectively) The simulation with pumping only in the lower compared to the base case. The simulated ground-water region showed storage accretion, a decrease in recharge discharge to the Mojave River was higher in the Alto from the Mojave River, an increase in discharge to the (about 15,360 acre-ft/yr), Transition zone (about Mojave River, and an increase in evapotranspiration in 2,250 acre-ft/yr), Centro (about 3,430 acre-ft/yr), and the Alto and Transition zone model subareas when Baja (about 770 acre-ft/yr) model subareas than for the compared with the base case. In addition, the simula- base case. Evapotranspiration was higher in the Alto tion with pumping only in the lower region showed that (about 1,010 acre-ft/yr, Transition zone (about there was less storage depletion and that there were 3,320 acre-ft/yr), Centro (about 2,480 acre-ft/yr), and increases in recharge from the Mojave River, discharge Baja (about 260 acre-ft/yr) model subareas than for the to the Mojave River, and evapotranspiration when com- base case. In the Alto model subarea there was about pared with the base case in the Centro and Baja model 5,760 acre-ft/yr in storage accretion; but about subareas. The greatest changes occurred in the Centro 6,210 acre-ft/yr was depleted from storage for the base model subarea. The changes in the Centro and Baja case, which resulted in a net increase in storage of model subareas were the result of the simulated about 11,970 acre-ft/yr in the Alto model subarea when hydraulic head in the Alto and Transition zone model compared with the base case. In the Transition zone subareas being near the altitude of the streambed model subarea there was about 1,960 acre-ft/yr of throughout most of the upper region. This caused storage accretion; since about 1,810 acre-ft/yr was potential recharge from the Mojave River to be rejected

92 Simulation of Ground-Water Flow in the Mojave River Basin, California in the upper region thereby allowing more streamflow changed depending on the daily inflow and the number to reach and recharge the lower region. of days that inflow from The Forks exceeded 200 ft3/s Overall, pumping in the lower region does not during the year (sections 3–5), the number of days that negatively affect the upper region; however, pumping inflow from The Forks exceeded 200 ft3/s during the in the upper region does negatively affect the lower year and whether there was inflow from ungaged tribu- region by decreasing recharge from the Mojave River taries (sections 13–18), or whether there was inflow in the lower region. The decrease in Mojave River from ungaged tributaries, without regard to inflow from recharge results in increased storage depletion and The Forks (sections 19–27). As in the calibration decreases in discharge to the Mojave River and evapo- period, the CSTR values for river sections 1 and 2 and transpiration in the Centro and Baja model subareas, 6–12 were constant. The CSTR values for sections 3–5 which can be seen by comparing the results of the base were specified on the basis of the total inflow at The case and the lower region pumping only simulations in Forks. The CSTR values for sections 13–18 and 19–27 the Centro and Baja model subareas (fig. 36). were assigned the values for “all other years” (table 7) because inflow from The Forks exceeded 200 ft3/s more than six days during the year and because it was Model Validation assumed that there was no ungaged tributary flow. Contours of measured water levels for spring Streamflow, pumpage, and water-level data for 1998 and simulated hydraulic heads for 1998 are calendar years 1995–99 were used to validate the shown in figure 37. In general, the simulated hydraulic calibrated ground-water flow model, that is, to test that heads are in good agreement with the measured water the flow model will reasonably simulate hydrologic levels, except for the water levels for the Oeste subarea. observations for a non-calibration period without mod- In this subarea, the model simulates the general trend ifying the model parameters. A wide range of stream- of the water-level contours, but the model overesti- flow conditions were used for the 5-year validation mates the hydraulic head. Similarly, the water-level period—two relatively wet winter stress periods [1995 depression in the Transition zone near the Southern (about 178,000 acre-ft) and 1998 (about 150,000 acre- California Logistics Airport is not well simulated, pos- ft)] and one dry winter stress period [1999 (about sibly due to an underestimation of pumpage in the 2,000 acre-ft)]. The ground-water flow model was cal- model for this area. ibrated using measured and approximated data for 1931–94. Simulated hydraulic heads at the end of 1994 Measured water levels and simulated hydraulic were used as initial conditions for the validation. Mea- head for 1998 data are shown in figure 38. The root sured pumpage for 1995–99 (fig. 14) and inflows mean square error (RMSE) for all model subareas is (table 9) were used to validate the calibrated model. 29.1 ft and the measured minus simulated mean error Values for mountain-front, septic-tank, and sewage- (ME) is −5.9 ft. The Oeste model subarea had the larg- pond recharge values were assumed to equal the 1994 est RMSE of 55.1 ft (fig. 38). The correlation coeffi- values, and irrigation-return flow was based on cient between the measured water levels and the measured agricultural pumpage (fig. 16). simulated hydraulic head was 0.998. Table 7 was used to specify the CSTR values for Simulated hydraulic heads and measured water the 1995–99 streamflow data based on the criteria used levels also were compared for 1995–99 using the same to assign values for the calibration period, 1931–94. As 42 hydrographs used for the transient-state model cali- discussed in the “Stream-Aquifer Interactions” section bration. The 42 hydrographs are presented in Appendix of this report, the Mojave River was divided into 27 1, 11 of which are also shown in figure 29. The 11 separate sections (fig. 22), which were numbered hydrographs are grouped by floodplain aquifer and sequentially in a downstream order, based on similar regional aquifer wells. In general, the hydrographs geologic properties of the streambed. Table 7 shows the show that the simulated hydraulic heads for wells in the range of values of streambed conductance (CSTR) for floodplain and the regional aquifers follow the the wet and dry periods for each section of the river, measured water-level trends (fig. 29). excluding the tributaries which had a value of zero. Simulated 1995–99 hydraulic heads were com- While streambed conductance values for some sections pared with short-term (1992–99) water levels measured were constant for all wet and dry stress periods (sec- at USGS-installed multiple-well monitoring sites along tions 1 and 2 and 6–12), the values for other sections the Mojave River (fig. 29, Appendix 2). In general, the

Ground-Water Flow Model 93 1,400

1,500 20 MILES 1,500

1,600 116 30' 1,750

1,600 1,700 del of the Mojave River ground-water basin,

? ? 1,700

? 20 KILOMETERS

10 1,725 1,700 EXPLANATION 1,725 Measured for hydraulic head in Springs 1,700 10 Newberry ater-level contour (measured 1998) – ater-level ell – 1,750 W Shows altitude of water level. Contour interval 100 feet. Selected intermediate contours 25 and 50 feet. Dashed where approximate, queried where uncertain. Datum is sea level Hydraulic-head contour (simulated 1998) – Shows altitude of simulated hydraulic head. Contour interval 100 feet. Datum is sea level W spring 1998 Active area of model Inactive area of model Model faults Boundary of model layer 1 Generalized direction of ground-water flow 45'

1,800 2,800 2,800 1,800 0 0

1,900

R 1 E R 2 E R 3 E R 4 E R 5 E R 6 E 1,900 2,000 3,000

Valley

Lucerne

3,200 3,000 3,100

? R 1 W

3,100

117

2,900

2,000 2,000

Barstow 2,100 2,900

2,100

2,000

2,100 Apple Valley 2,800 2,200

2,200 2,000

? 2,750

R 3 W 2 W 2,300 2,300

2,800 2,300 Victorville 15'

2,150 2,150 2,250 2,250

2,200 2,100 Helendale 2,350 2,500 ? 2,000 2,000

2,000

1,950 2,450 3,000 2,600 2,700 2,750

R 4 W 2,700 Hesperia

?

1,950 2,100 2,000

?

1,900

2,800 5 W R 4 W R 3 W R 2 W R 1 W R 1 E 2,400 2,900 2,850 R 5W Mojave River

2,400

2,600

2,600 2,550 2,500 2,925 117 30' R 6 W

2,700 3,000

2,800 ?

R 41 E 3,100 2,900 2,950 R 7 W 7 R R 6 W R R 7 W Measured water levels, spring 1998, and simulated hydraulic-head contours, for model layer 1 of the ground-water flow mo

R40E R 42 E R43 E R44 E R 45 E R 46 E R 47 E ? 35 T8N T9N 32 S

2,700 T4N 12 N T3N 10 N

2,600 T5N T7N T6N T T T 11 N T 11 T 34 45' 34 30' southern California. Figure 37.

94 Simulation of Ground-Water Flow in the Mojave River Basin, California simulated hydraulic heads follow the measured for the Lower Narrows, Barstow, and Afton Canyon water-level trends of the shallowest wells at the gaging stations for the 1995–99 period averaged multiple-well monitoring sites; however, the pumpage- −8,100; −2,960; and 600 acre-ft/yr, respectively. The induced seasonal water-level fluctuations were not sim- largest difference for the Lower Narrows gaging station ulated because only constant pumpage data were used was for the 1996 winter (wet) stress period for which in the model. The deeper wells at the multiple-well the model overpredicted streamflow by about monitoring sites along the Mojave River are not well 12,000 acre-ft (fig. 31). The largest difference for the simulated because layer 1 of the model simulates the Barstow gaging station was for the 1995 winter (wet) floodplain aquifer and, in most areas, layer 2 simulates stress period for which the model overpredicted the younger alluvium of the floodplain aquifer and, streamflow by about 22,000 acre-ft (fig. 31). These therefore, the underlying older units are not simulated. results indicate that the streambed conductance values Simulated streamflow data for 1995–99 wet and calibrated to the 1931–94 conditions (table 7) reason- dry stress periods at the Lower Narrows, Barstow, and ably simulate the 1995–99 conditions and therefore can Afton Canyon gaging stations were compared with be used for predictive purposes. average measured streamflow for the same periods During the period of 1995–99, the total average (fig. 30). In general, the model reflects measured inflow, or recharge, to the ground-water system was 1995–99 streamflow conditions (fig. 31). The differ- about 164,500 acre-ft/yr (a total volume of ence between measured and simulated streamflow rates 822,600 acre-ft) (table 11). Most of the recharge was

3,000 Alto (21.8) Oeste (55.1) Subarea – (Number is RMSE value) Este (20.5)

2,500

Transition zone (35.8) Harper Lake (30.53)

2,000 Centro (30.5)

Coyote Lake (9.4)

Baja (21.5)

SIMULATED WATER LEVEL, IN FEET WATER SIMULATED Afton Canyon (1 value) 1,500

1:1 Correlation line 1,000 1,000 1,500 2,000 2,500 3,000

MEASURED WATER LEVEL, IN FEET

Figure 38. Measured water levels, simulated hydraulic head, and the root mean square error (RMSE) for each model subarea of the Mojave River ground-water basin, southern California, for 1998 transient-state conditions. (See figure 18 for location of model subareas.)

Ground-Water Flow Model 95 otal T Afton Canyon Lake Coyote lues Baja Lake Harper Centro 13,43813,643 2,113 2,113 27,289 27,302 959 961 -299 -300 32,582 32,309 − − Zone ransition T 594594 3,439 3,443 11,726 11,664 199 176 Este Oeste Alto 1 1, 2 Simulated hydrologic budget for model subareas of the Mojave River ground-water basin, southern California, 1995–99 average va alues of storage differ as a result of accumulation of small, consistent errors in the model and rounding of large numbers. as a result of accumulation small, consistent errors in the model and rounding large alues of storage differ Positive storage value indicates storage depletion; negative storage value indicates storage accretion. storage value indicates storage depletion; negative storage value Positive otal...... 1,654 6,944 82,402 28,940 39,526 7,381 45,728 1,460 558 197,096 otal...... 1,060 3,505 70,676 28,741 52,964 5,268 18,439 501 857 164,515 V 1 2 T T recharge and discharge recharge Difference between Difference PumpageDrainsEvapotranspirationHead-dependent boundaryStream leakage between subareasFlow 0 442 0 1,149 63 5,430 0 0 1,514 0 73,983 0 4,185 4,234 15,732 0 0 1,930 9,873 28,232 0 0 0 7,147 2,013 7,381 0 1,405 41,141 0 0 0 2,134 262 0 1,131 0 956 0 0 0 616 0 0 2,500 172,603 0 355 0 0 0 17,606 582 45 na 158 0 6,197 45 645 Storage depletion Irrigation returnIrrigation PondsSewage Mountain front Septic tank Stream leakage 23 between subareasFlow 1,035 0 0 0 1,941 2 0 0 1,564 4,454 7,762 0 0 3,278 2,670 0 0 45,974 9,816 5,648 3,101 956 21,238 1,052 0 168 2,157 3,005 42,154 10,091 4,216 0 0 0 0 3,463 76 647 0 3,652 586 166 0 259 0 159 0 0 24,622 0 698 0 0 na 11,644 113,716 4,547 0 9,986 Discharge Recharge Table 11. na, not applicable] in acre-feet per year. [Values

96 Simulation of Ground-Water Flow in the Mojave River Basin, California from the Mojave River (about 113,700 acre-ft/yr, or attached CD-ROM (in pocket) on a computer capable 69 percent), irrigation-return flow (about of reading CD-ROM’s. 24,600 acre-ft/yr, or 15 percent), mountain-front The 1940 simulated hydraulic heads were 20 to recharge (about 11,600 acre-ft/yr, or 7 percent), and 30 ft higher than the 1931 simulated hydraulic heads septic-tank recharge (about 10,000 acre-ft/yr, or 6 per- along most of the Mojave River (fig. 39A) because cent). During this same period, the total average out- there were large inflows to the Mojave River from the flow, or discharge, from the ground-water system was headwaters (Deep Creek and West Fork) in 1937 and about 197,000 acre-ft/yr (a total volume of 985,500 1938 (about 165,000 and 224,100 acre-ft, respectively) acre-ft) (table 11). Most of the discharge was attribut- (fig. 40, table 1). These large inflows recharged the able to pumpage (about 172,600 acre-ft/yr, or 86 per- underlying floodplain aquifer and increased the cent), evapotranspiration (about 17,600 acre-ft/yr, or hydraulic head along the Mojave River in 1937 and 9 percent), and discharge to the Mojave River (about 1938 (see CD-ROM). There were little or no changes in 6,200 acre-ft/yr, or 3 percent). The difference between simulated hydraulic head along the Mojave River in the recharge and discharge, which is the contribution from eastern part of the Baja model subarea and throughout ground-water storage, averaged about 32,600 acre-ft/yr the regional aquifer (fig. 39A and CD-ROM). (a total volume of 162,900 acre-ft). The 1950 simulated hydraulic heads were about The average pumping rate simulated for 20 ft higher than the 1931 simulated hydraulic heads 1995–99 was about 172,600 acre-ft/yr (table 11) com- along the Mojave River from the headwaters to about pared with the average of 151,700 acre-ft/yr simulated 10 miles east of Barstow (fig. 39B). The simulated for 1931–90 (table 10). Although the average pumping hydraulic heads were higher in 1950 because total rate for 1995–99 was slightly greater than the average annual inflow to the Mojave River from the headwaters for 1931–90, the average storage depletion for throughout much of the 1940’s was greater than the 1995–99 was less than the average storage depletion average annual inflow shown in the cumulative depar- for 1931–90 (about 32,600 acre-ft/yr compared with ture from mean streamflow for Deep Creek (fig. 40). 39,400 acre-ft/yr). This difference was made up by Large inflows from the headwaters in 1941 and 1943 increased stream leakage in 1995–99 (113,700 (about 157,400 and 155,000 acre-ft, respectively) acre-ft/yr for 1995–99 compared with 87,400 acre-ft/yr (table 1) recharged the floodplain aquifer and resulted for 1931–90). in increased simulated hydraulic head throughout In general, the match between the simulated much of the floodplain aquifer (see CD-ROM). Simu- hydraulic heads for 1995–99 and measured water levels lated hydraulic heads for 1950 were about 15 ft lower was good. The ability to reasonably match data from than the simulated hydraulic heads for 1931 in the another time period implies that the ground-water flow southeastern part of the Baja model subarea and in the model may be used to predict the response of the aqui- western part of the Harper Lake model subarea fer system to stresses that are similar in type and mag- (fig. 39B) as a result of increased agricultural pumpage nitude to those used during the calibration process. after 1945 (figs. 16 and 23). The 1960 simulated hydraulic heads were more Simulated Changes in Hydraulic Head, 1931–99 than 30 ft lower than the 1931 simulated hydraulic heads in the Alto (near Victorville), Transition zone The spatial and temporal distribution of recharge (near Helendale), Centro and Harper Lake model sub- and pumpage results in water-level changes in the areas (about 30, 45, 55, and 80 ft, respectively) Mojave River ground-water basin. To help visualize the (fig. 39C). The 1960 simulated hydraulic heads were 10 magnitude, spatial distribution, and timing of these to 25 ft lower than the 1931 simulated hydraulic heads water-level changes, simulated hydraulic heads from in the Oeste (near El Mirage dry lake), Alto (near Apple 1932–99 were compared with simulated hydraulic Valley), and Baja model subareas (about 15, 15, 25 ft, heads for 1931 on an annual basis. The simulated respectively). The lower simulated hydraulic heads changes in hydraulic head for model layer 1 are pre- (drawdown) correspond with increased agricultural sented in figure 39 for 10-year increments; the annual and municipal pumpage (figs. 16 and 23). A large changes in simulated hydraulic head for model layer 1 inflow from the headwaters in 1958 (about 138,800 can be viewed on a personal computer by playing the acre-ft; table 1) recharged the floodplain aquifer and

Ground-Water Flow Model 97 5 5 -5 -5 90 50 30 90 50 30 -30 -50 -90 -30 -50 -90 200 130 200 130 -130 -200 -130 -200 Afton Afton , 1940. Canyon Canyon A Troy Lake Troy Lake In feet In feet 20 MILES 20 MILES (1960) – (1970) – in hydraulic head in hydraulic head Simulated change Simulated change outhern California. 20 KILOMETERS 20 KILOMETERS Springs Springs Newberry Newberry

10 10 10 10 Boundary of Boundary of model subarea model subarea Valley Valley Lucerne Lucerne Model fault Model fault 0 0 0 0 Barstow Barstow Apple Valley Apple Valley Boundary of Boundary of model layer 1 model layer 1 Helendale Helendale Victorville Victorville Hesperia Hesperia River River Mojave Mojave Lake Lake Harper Harper Airport Southern Logistics California Airport Adelanto Adelanto Southern Logistics California Lake Lake El El El Mirage

C El Mirage Mirage Mirage 5 5 -5 -5 90 50 30 90 50 30 -30 -50 -30 -50 -90 -90 130 130 200 200 -130 -200 -130 -200 Afton Afton Canyon Canyon Troy Troy Lake Lake In feet In feet 20 MILES 20 MILES (1940) – (1950) – in hydraulic head in hydraulic head Simulated change Simulated change 20 KILOMETERS 20 KILOMETERS Springs Springs Newberry Newberry

10 10 10 10 Boundary of Boundary of model subarea model subarea Valley Valley Lucerne Lucerne , 1999. (See Compact Disk for video version showing simulated change in hydraulic head years 1931–99.) Model fault Model fault G 0 0 0 0 Barstow Barstow Apple Valley Apple Valley Boundary of Boundary of model layer 1 model layer 1 , 1990. F Helendale Helendale , 1980. Victorville Victorville E Hesperia Hesperia , 1970. D River River Mojave Mojave Lake Lake Harper Harper Airport Airport Adelanto Adelanto Southern Southern Logistics Logistics California California Simulated changes in hydraulic head for model layer 1 of the ground-water flow Mojave River basin, s ,1960. C Lake Lake El El

El Mirage A El Mirage BD Mirage Mirage , 1950. B Figure 39.

98 Simulation of Ground-Water Flow in the Mojave River Basin, California 5 -5 90 50 30 -30 -50 -90 200 130 -130 -200 Afton Canyon Troy Lake In feet 20 MILES (1999) – in hydraulic head Simulated change 20 KILOMETERS Springs Newberry

10 10 Boundary of model subarea Valley Lucerne Model fault 0 0 Barstow Apple Valley Boundary of model layer 1 Helendale Victorville Hesperia River Mojave Lake Harper Airport Adelanto Southern Logistics California Lake El El Mirage Mirage 5 5 -5 -5 90 50 30 90 50 30 -30 -50 -30 -50 -90 -90 130 130 200 200 -130 -200 -130 -200 Afton Afton Canyon Canyon Troy Troy Lake Lake In feet In feet 20 MILES 20 MILES (1980) – (1990) – in hydraulic head in hydraulic head Simulated change Simulated change 20 KILOMETERS 20 KILOMETERS Springs Springs Newberry Newberry

10 10 10 10 Boundary of Boundary of model subarea model subarea Valley Valley Lucerne Lucerne Model fault Model fault 0 0 0 0 Barstow Barstow Apple Valley Apple Valley Boundary of Boundary of model layer 1 model layer 1 Helendale Helendale Victorville Victorville Hesperia Hesperia River River Mojave Mojave Lake Lake Harper Harper Airport Airport Adelanto Adelanto Southern Southern Logistics Logistics .—Continued. California California Lake Lake El El El Mirage El Mirage EG F Mirage Mirage Figure 39

Ground-Water Flow Model 99 resulted in a temporary increase in simulated hydraulic drawdowns, beneath the Mojave River (see CD-ROM). heads directly beneath the Mojave River (CD-ROM). However, these large inflows to the Mojave River had After 1960, the areas in the regional aquifer with little apparent effect on the simulated drawdowns in the lower simulated hydraulic heads than the simulated regional aquifer (fig. 39D–G, CD-ROM). hydraulic heads (drawdowns) for 1931 were essentially unchanged from the 1960 simulation (fig. 39C); how- Model Limitations ever, the areas of drawdown increased in size and mag- nitude (fig. 39D–G). By 1999, simulated drawdowns Although a ground-water flow model can be a exceeded 50 ft in the Alto (near Victorville), Oeste useful tool for investigating aquifer response, it is a (south of El Mirage dry lake), Harper Lake, Centro, and simplified approximation of the actual system and is Baja model subareas (80, 175, 120, 55, and 90 ft, based on average or estimated conditions; the accuracy respectively) (fig. 39G). The simulated hydraulic head of its predictions are dependent on the availability and along the floodplain aquifer fluctuated in response to accuracy of the input data used to calibrate the model. large inflows (in excess of 160,000 acre-ft) to the For the study area of this report, the model is able to Mojave River at the headwaters in 1969, 1978, 1980, duplicate hydraulic heads fairly accurately for the 1983, and 1993 (table 1) and 1995 and 1998 (table 9). floodplain aquifer because long-term measured water- These large inflows recharged the floodplain aquifer levels are available. However, in areas where there are and resulted in increased simulated heads, or lessened sparse or no data, such as is the case for most areas of

100,000

Mean 0

-100,000

-200,000

-300,000 CUMULATIVE DEPARTURE, IN ACRE-FEET DEPARTURE, CUMULATIVE

-400,000

-500,000 1930 1940 1950 1960 1970 1980 1990 2000 YEAR Figure 40. Cumulative departure from mean streamflow measured at the headwaters of the Mojave River, southern California, 1931–99. Values are based on the annual flow from Deep Creek (gaging station 10260500).

100 Simulation of Ground-Water Flow in the Mojave River Basin, California the regional aquifer, the accuracy of the model is model simulations which could lead to errors when reduced. Another model limitation is that model cali- water-level declines are large compared to the saturated bration, or the “inverse problem,” yields non-unique thickness of the aquifer. However, when this assump- sets of parameter estimates because different combina- tion was not made and the floodplain aquifer (model tions of hydrogeologic conditions may lead to similar layer 1) was allowed to have a variable saturated thick- observations of water level (Sun, 1994). ness, the simulated hydraulic heads declined below the Possible sources of inaccuracies related to input bottom altitude of this aquifer during some dry periods. data include the estimates of pumpage. Most of the As discussed in the “Transmissivity” section of this wells in the Mojave River ground-water basin have report, the version of the streamflow-routing package never been metered and, therefore, the assumed water- used to simulate the river does not allow the leakage of use rate of 7.0 ft applied to all agricultural land use may streamflow into or out of the aquifer system once a be an overestimation or an underestimation of pump- stream cell has gone dry. Stream cells that have become age for some areas, depending on crop type and irriga- dry are bypassed when streamflow is reintroduced, and tion practices. In addition, constant pumping was any water in the stream is routed to the next active assumed for the entire calendar year, which does not downstream reach; only upward leakage from the aqui- reflect seasonal pumping practices; therefore, the fer to the stream is allowed. Because of this, it was nec- model will not simulate the maximum and minimum essary to hold the transmissivity values constant over drawdowns in the basin. Estimation of the distribution time. and quantity of ungaged tributary flow is another During the course of a year, evapotranspiration source of model inaccuracy. can vary by as much as 50 percent depending on the The most significant limitations of this model availability of surface water and the altitude of the were its sensitivity to streambed conductance and the water table (Lines and Bilhorn, 1996, p. 1). This avail- assumptions made in the streamflow-routing package. ability of water fluctuates with streamflow and the time of year. Most evapotranspiration from the water table The sensitivity of the model to streambed conductance occurs when surface water is not readily available, such was such that any change in other parameters (trans- as in the hotter summer months, and tapers off in winter missivity, fault hydraulic characteristic, or evapotrans- when water is available from the river and air tempera- piration rate, for example) required the recalibration of tures are cooler. The evapotranspiration package the streambed conductance values. In order to keep assumes a constant rate and does not allow for an streambed conductance values constant throughout a increase in water-use rates during the summer and a stress period, constant width and stage values were decrease in rates during the winter. In fact, it acts in the assigned to the model even though they vary in the opposite manner and withdraws more water from the actual system depending on volume of flow and they ground-water system when the water table is highest can change markedly during a single flood event. (winter, or wet, stress periods) and less water when it is In order to address the ephemeral nature of long lowest (summer, or dry, stress periods). This does not reaches in the river during floodflows and to match the allow the model to accurately simulate the timing of measured streamflow at the gages, two stress periods evapotranspiration or the amount withdrawn from the per year (winter and summer) were used. The winter ground-water system. stress period was defined by any discharge in excess of 200 ft3/s as measured at the headwaters, referred to as the “wet-period cutoff” and the summer stress period EVALUATION OF SELECTED WATER- was the remainder of the year. In general, this resulted MANAGEMENT ALTERNATIVES in an overestimation of streamflow for the winter stress period and an underestimation of streamflow for the The MWA has the authority to artificially summer stress period. To more accurately model the recharge the Mojave River ground-water basin with Mojave River streamflow, weekly, or perhaps daily, imported water from the State Water Project (SWP). stress periods are required. The MWA has constructed, or has proposed to con- The transmissivity values of the aquifers used in struct, eight artificial recharge sites within the Mojave the model were assumed to be constant over time. This River ground-water basin (fig. 41). Artificial recharge assumption implies that the saturated thickness of the to the ground-water basin initially occurred through model layer does not change significantly during releases from Silverwood Lake into the Mojave River.

Evaluation of Selected Water-Management Alternatives 101 Afton

10263000 Canyon 5 E R 6 E

40 number (98, 48) (56, 75) (49, 81) (138, 65) (121, 12) (54, 118) (50, 141) (66, 154) Model cell 30' ° 116 15 3,427 5,483 5,483 4,798 3,770 6,512 13,709

Troy Wash

Lake

Manix

Newberry Entitlement

Cady EXPLANATION Camp Fault

Baja fault Wash 1,714 2,742 2,742 2,399 6,854 1,885 3,256 50 % 100 % 10,967 21,934

3 E R E 4 R Calico- Kane (values in acre-feet/year) Fault Dry lake (playa) Mojave River ground-water basin boundary Gaging station and number Site of recharge and name Minneola Springs

Newberry Kane Wash Lake Coyote 45' ransition zone 10260500 Recharge site name Rock Springs Road outlet El Mirage T Hodge Lenwood Daggett Minneola Kane Wash R 2 ER 2 R

Camp Rock-Harper 18 Lake Fault Zone pipeline

(proposed)

Mojave River R 1 E

(Waterman Fault) 20 MILES

Meridian San Bernardino San Lucerne Lake Valley

Lucerne Daggett

Waterman E Fault ° R 1 W R 1 E R 1 W 117 20 KILOMETERS 10262500 Lake Rabbit 10 Fault Barstow San Bernardino Mts R 2 W pipeline R

Morongo Basin k e 10 e r Mt. Generalfault C Hodge p e

e

D Apple Valley R 3 W 2 W Lenwood R 3 W 18 10260500 0 0 Apple Valley fault 15' Iron Mtn 10261100

Helendale Narrows fault

Fault r

e

v

Lake i

Harper 10260950 R 4 W R Lower R 4 W Upper Narrows Narrows

fault

Iron Mtn

Lockhart Hesperia

Road outlet

M Rock Springs

e v a o j River

10261500 Aqueduct Victorville Mojave

West Fork

15 Wash Helendale R 5 W R 5W

zone pipeline Fremont

Adelanto fault Transition

Mojave River Silverwood Lake R 42 E R43 E R44 E R 45 E R 46 E R 47 E Adelanto 30' °

117 California R 6 W

fault R 6 W

Shadow Mts Sheep Creek Sheep Shadow Mts R 41 E Logistics Airport Southern California R 7 W El Mirage Lake Location of Mojave Water Agency artificial-recharge sites in the River ground-water basin, southern California. R 7 W El Mirage San GabrielSan Andreas MtsFault Zone R40E ° 45' 30' T3N 35 T4N ° T5N 32 S ° 12 N 10 N T8N T7N T9N T6N T T T 11 N T 11 T 34 Figure 41. 34

102 Simulation of Ground-Water Flow in the Mojave River Basin, California From 1978 through 1994, these releases totaled about headwaters of the Mojave River [Deep Creek 70,000 acre-ft (Lines, 1996, p. 21). Beginning in 1994, (10260500) gaging station] for 1931–99 followed a SWP water has been released to the Mojave River from decreasing trend for which only short periods (1-year a turnout in the Morongo Basin pipeline at the Rock or less) of this decreasing trend were reversed. The Springs Road outlet (fig. 41). A total of 21,200 acre-ft cumulative departure from mean streamflow shown in of water was released at the Rock Springs Road outlet figure 40 indicates that a 20-year drought started in between 1994 and 2000. In 1995, construction began about 1945 and ended in about 1964. on the Mojave River pipeline which was designed to In order to evaluate the three water-management provide delivery capabilities of SWP water along the alternatives, the 1999 rates and distributions for Mojave River past Barstow (fig. 41). In addition, the mountain-front recharge, sewage-pond recharge, MWA has the authority to take SWP water from the septic-tank recharge, irrigation-return flow, and pump- California Aqueduct near El Mirage (fig. 41). age were used. Streamflow conditions were simulated Three water-management alternatives were used using the specified inflows from The Forks (Deep to evaluate the effect of artificial recharge on the Creek and West Fork) to the Mojave River for 1945–64 ground-water resources of the Mojave River ground- (table 9) with associated calibrated stream parameters water basin using the calibrated ground-water flow (table 7). model developed for this study. The three water- management alternatives considered the artificial recharge of SWP water allocated to the MWA at the Management Alternative 1: Zero Percent of eight existing or proposed recharge sites. In 2000, the Artificial Recharge Allocation total MWA allocation was 75,800 acre-ft, but about 65,000 acre-ft of the allocation was available for use Management alternative 1 evaluated the within the model area owing to water delivery obliga- response of the ground-water system assuming current tions in other parts of the MWA management area. The (1999) rates of pumpage (about 165,900 acre-ft/yr) first water-management alternative evaluated in the (fig. 14) during a 20-year drought with no MWA allo- model assumed that zero percent of the MWA alloca- cation of water available for artificial recharge to miti- tion was available (alternative 1), the second assumed gate the effects of the drought. Simulated recharge that 50 percent of the MWA allocation, about from the Mojave River (stream leakage) was about 30,000 acre-ft/yr, was available (alternative 2), and the 52,300 acre-ft/yr, which is a 61,400 acre-ft/yr reduction third assumed that 100 percent of the MWA allocation, in recharge compared with the average recharge for about 60,000 acre-ft/yr, was available (alternative 3). 1995–99 (tables 11 and 12). Most of this reduction in The artificial recharge site locations and the amount of recharge occurred in the Alto (19,400 acre-ft/yr), Tran- entitlement for each location are shown in figure 41. sition zone (3,600 acre-ft/yr), and Centro (35,600 The simulated hydraulic-head changes for the three acre-ft/yr) model subareas. This reduction in recharge water-management alternatives are shown on figure 42, was reflected in simulated hydraulic-head declines and the resulting hydrologic budgets are presented in between 1999 and 2019 of as much as 50 ft (fig. 42). table 12. Simulated evapotranspiration decreased about 5,600 Each of the three water-management alternatives acre-ft/yr and ground-water discharge (stream leakage) were evaluated using streamflow conditions that to the Mojave River decreased about 3,900 acre-ft/yr existed during an extended 20-year drought that compared with the average discharge for 1995–99 occurred in the Mojave River ground-water basin from (tables 11 and 12); these reductions were related to the 1945 to 1964 along with the 1999 artificial recharge declines in simulated hydraulic heads. values from the Mojave River Fish Hatchery and The hydraulic-head decline simulated at the VVWRA (table 4). A drought condition was chosen boundary between the Oeste model subarea and because it represents a time of minimal natural stream- Antelope Valley may be overestimated because a no- flow recharge to the ground-water system and, there- flow boundary was being used to simulate the ground- fore, would represent a worst-case scenario based on water divide between the Oeste model subarea and the available data. For the purposes of this study, a Antelope Valley. In reality, water from the Antelope drought is defined as a period where the cumulative Valley may have been a source of water to the Oeste departure from mean streamflow measured at the model subarea under simulated pumping conditions

Evaluation of Selected Water-Management Alternatives 103

200

0 0 20 MILES Interval 180 0 116 30'

0

170 190

10 10 California, for three management alternatives,

0 0

20 20 KILOMETERS 160 Kane Wash 10 150 EXPLANATION

Minneola 30 20 30 10

10 10 Active area Inactive area Springs Newberry Model grid – Model fault Boundary of model layer 1 Boundary of model subarea Line of equal simulated hydraulic head change between 1999 and 2019 for management alternative 1 – 10 feet Artificial-recharge site

140

45'

40 40

20 Hodge 0 30 0 20

50 50 Daggett

30 30

40

40 10

20 20 60

60 0 20 20

Valley

Lucerne

30 30

40 40 117

100 110 120 130

Barstow

10 10 90 MODEL COLUMN Lenwood , Management alternative 3.

C

20 20 20 20

10 80 10

40 40

30 30 Apple Valley

10 20

0 0

10 10

10 10 15'

40

20 Hodge 20 60 0 zone

20 Transition 0 10 0 20 30

30 Road outlet Hesperia Rock Springs

50 70 0

, Management alternative 2. 0 0 20 10 10 B Victorville

Helendale

0 0 Mojave River 117 30' 0 -10 -20

10 10 MODEL ROW MODEL , Management alternative 1. 10 20 30 40 A A El Mirage Change in simulated hydraulic head layer 1 of the ground-water flow model Mojave River basin, southern

20 20

30 30 10 20 30 40 50 60 70 80 90 140 150 160 100 110 130 35 34 45' 34 30' 1999–2019. Figure 42.

104 Simulation of Ground-Water Flow in the Mojave River Basin, California

0

0 0 20 MILES 180 0 116 30' 170 190 200 Interval 10 feet

395

1010 20 KILOMETERS

160 2020

3030 3030

5050 5050

10

8080 8080 Springs

Newberry

4040 4040

150 EXPLANATION

Minneola

100100 100100

6060 6060 10

20 Active area Inactive area Model grid – Line of equal simulated hydraulic head change for 2019 between management alternatives 1 and 2 – Artificial-recharge site 10 Model fault Boundary of model layer 1 Boundary of model subarea 140 45' 0 Kane Wash 20

Hodge

0 0

0 0

0 0

Daggett

10 10

93

40 40

30 30

20 20 0

60 80 80 Valley Lucerne

0 117 100 110 120 130 Barstow

90 10

MODEL COLUMN

10 80 10

10 10

20 20 Apple Valley Lenwood 30

15' Hodge

10 10

60 20

0 0 10 Road outlet Rock Springs 50 70 Hesperia Victorville Helendale zone Transition Mojave River 117 30'

El Mirage

20 20

10 20 30 40 ROW MODEL

60

10 B 10 —Continued.

124 40 40

30 30 10 20 30 40 50 60 70 80 90 140 150 160 100 110 130 35 34 45' 34 30' Figure 42.

Evaluation of Selected Water-Management Alternatives 105 0 20 MILES 180 116 30' Kane Wash

170 190 200

Interval 10 feet

10 655 10

0 20 20 KILOMETERS 160

Springs 40

Newberry

60

10 80

150 EXPLANATION 100

Minneola 200 10 Active area Inactive area

30 120

Model grid – Model fault Boundary of model layer 1 Boundary of model subarea Line of equal simulated hydraulic head change for 2019 between management alternatives 1 and 3 – 20 Artificial-recharge site

140 140

45'

20 20

10 10

20 Hodge 0 40 0

60

20 Daggett 186

10

180 180

80 80

160 160 Valley

Lucerne

140 140

120 120

100 100

10 Barstow 10 117 100 110 120 130

20

30 90 MODEL COLUMN

Lenwood

10 10

30

80 20

20

10 50

75 40 40 Apple 0 Valley 40 0 50 40 40

60 60

30

0

15' 0 0

10

60 10 Hodge 20 30 20

Road outlet 10 Rock Springs 50 70 Hesperia

0 Victorville Helendale zone Transition Mojave River 100 100 80 80 60 60 120 140 117 30'

El Mirage

10 10

20 20 10 20 30 40 ROW MODEL 30 C —Continued.

10 20 30 40 50 60 70 80 90 40 40 140 150 248 160 100 110 130 35 34 45' 34 30' Figure 42.

106 Simulation of Ground-Water Flow in the Mojave River Basin, California and thus may have reduced the actual water-level was available for artificial recharge to mitigate the declines. effects of the drought. The model simulated about a 5,600 acre-ft/yr increase in recharge from the Mojave River (stream leakage) (about 58,000 acre-ft/yr for Management Alternative 2: 50 Percent of management alternative 3) compared with recharge for Artificial Recharge Allocation management alternative 1 (about 52,300 acre-ft/yr) (table 12). The effects of the artificial recharge were Management alternative 2 evaluated the indicated by the increases in simulated hydraulic heads response of the ground-water system assuming current at each of the artificial-recharge sites at the end of the (1999) rates of pumpage (about 165,900 acre-ft/yr) simulation period (2019) when compared with the sim- continued during a 20-year drought and 50 percent of ulated hydraulic heads for management alternative 1 the MWA allocation of water (about 32,500 acre-ft/yr) (fig. 42C). The model results indicated that evapotrans- was available for artificial recharge to mitigate the piration increased about 2,900 acre-ft/yr compared effects of the drought. The model simulated very little with the evapotranspiration for management alternative change in recharge from the Mojave River (stream 1 (12,000 acre-ft/yr). Ground-water discharge to the leakage) (about 54,400 acre-ft/yr) compared with Mojave River (stream leakage) increased about recharge for management alternative 1 (about 5,600 acre-ft/yr compared with the discharge for man- 52,300 acre-ft/yr) (table 12). The effects of the artificial agement alternative 1 (2,300 acre-ft/yr) (table 12). recharge were indicated by the increases in simulated These increases were related to the increases in hydraulic heads at each of the artificial-recharge sites at simulated hydraulic heads. the end of the simulation period (2019) when compared As described in the other management alterna- with the simulated hydraulic heads for management tives, the hydraulic-head increase simulated at the alternative 1 (fig. 42B). The model results indicated boundary between the Oeste model subarea and that evapotranspiration increased about 1,400 acre-ft/yr Antelope Valley may have been overestimated because compared with the evapotranspiration for management a no-flow boundary was used to simulate the ground- alternative 1 (12,000 acre-ft/yr). Ground-water dis- water divide between the Oeste model subarea and charge to the Mojave River increased about Antelope Valley. In reality, water from the Oeste model 2,000 acre-ft/yr compared with the discharge for man- subarea may have flowed into the Antelope Valley agement alternative 1 (2,300 acre-ft/yr) (table 12). under the simulated recharge conditions, which may These increases were related to the increases in have reduced the actual water-level increase. simulated hydraulic heads. As with management alternative 1, the hydraulic-head increase simulated at the boundary Discussion of Management Alternatives 2 and 3 between the Oeste model subarea and Antelope Valley may have been overestimated because a no-flow The largest increases in simulated hydraulic boundary was used to simulate the ground-water divide heads for management alternatives 2 and 3 are at the El between the Oeste model subarea and Antelope Valley. Mirage, Kane Wash, and Daggett artificial recharge In reality, water from the Oeste model subarea may sites (fig. 42B, C). Simulated hydraulic heads increased have flowed into the Antelope Valley under the simu- more than 390 ft under management alternative 2 and lated recharge conditions which may have reduced the more than 650 ft under management alternative 3. actual water-level increase. These high increases at El Mirage and Kane Wash were the result of water recharging into areas of low trans- missivity for model layers 1 and 2 (transmissivities less Management Alternative 3: 100 Percent of than 2,000 ft2/d) (fig. 19). The recharged water did not Artificial Recharge Allocation tend to spread in areas of low transmissivity compared with recharge water in areas of high transmissivity, Management alternative 3 evaluated the such as the sites along the Mojave River (figs. 19 and response of the ground-water system assuming current 42). The Daggett site is located on the Mojave River in (1999) rates of pumpage (about 165,900 acre-ft/yr) a relatively narrow area of high transmissivity (about continued during a 20-year drought and 100 percent of 37,500 ft2/d); the recharge rate at this site was relatively the MWA allocation of water (about 65,000 acre-ft/yr) high (about 13,700 acre-ft/yr for management

Evaluation of Selected Water-Management Alternatives 107 otal T Afton Canyon cial recharge allocation. Values in Values allocation. cial recharge fi Lake Coyote ternatives 1, 2, and 3, 1999–2019 average Baja Lake Harper Management alternative 1 cial recharge allocation; and 3, 100 percent of Mojave Water Agency arti Agency Water allocation; and 3, 100 percent of Mojave cial recharge fi 692697 2,802 2,817 31,283 31,241 2,037 1,895 15,368 15,251 -425 -428 27,569 27,659 1,154 1,163 -32 -34 80,448 80,260 Este Oeste Alto Zone Transition Centro cial recharge; 2, 50 percent of Mojave Water Agency arti Agency Water 2, 50 percent of Mojave cial recharge; fi 1 1, 2 Simulated hydrologic budgets for model subareas of the Mojave River ground-water basin, southern California, management al cial 0 0 0 0 0 0 0 0 0 0 fi otal...... 1,058 3,519 50,669 24,156otal...... 1,750 16,652 6,321 4,641 81,952 14,141 26,193 353 32,020 580 4,216 100,403 41,710 1,507 548 180,851 See footnotes at end of table. T T recharge and discharge recharge Irrigation returnIrrigation pondsSewage Mountain front Septic tank Stream leakageArti 21 1,035 0 1,941 0 2 0Pumpage 0DrainsEvapotranspiration 7,763 3,845Head-dependent boundary 0 0Stream leakage 2,851 between subareasFlow 0 0 0 26,596 9,817 433 0 between Difference 1,275 5,344 650 17,651 42 4,949 0 0 168 0 1,372 1,052 0 2,265 6,507 75,262 0 9,745 4,085 0 2,605 0 0 0 14,834 0 0 2,504 72 7,691 26,369 0 647 0 0 0 1,159 586 4,865 4,216 596 259 0 1,164 0 0 0 0 0 39,610 0 22,930 0 0 190 247 427 0 0 0 0 529 11,645 823 52,340 0 736 0 0 3,501 0 0 0 748 165,920 9,987 0 0 0 261 0 524 137 11,976 na 150 0 137 2,252 566 Flow between subareasFlow 0 1,578 2,648 2,836 2,536 3,589 2,004 22 153 na Storage depletion values Table 12. 1, No arti [Management alternatives: acre-feet per year. na, not applicable] acre-feet per year. Recharge Discharge 108 Simulation of Ground-Water Flow in the Mojave River Basin, California otal T Afton Canyon Lake Coyote ternatives 1, 2, and 3, 1999–2019 average Baja Lake Harper Centro Management alternative 2 Zone ransition T 0000005240566 0000000137137 Este Oeste Alto 688692 1,603 1,612 20,678 20,637 1,188 1,094 8,928 8,754 -560 -563 15,672 15,711 1,088 1,096 -31 -34 49,254 48,998 1 1, 2 Simulated hydrologic budgets for model subareas of the Mojave River ground-water basin, southern California, management al cial recharge 0 1,715 10,975 2,744 5,145 0 12,004 0 0 32,583 fi otal ...... 1,750 6,321 82,511 28,463 32,449 4,216 41,851 1,507 549 184,251 otal ...... 1,058 5,234 61,362 26,844 24,023 4,641 26,268 353 580 134,997 See footnotes at end of table. T T recharge and discharge recharge EvapotranspirationHead-dependent boundaryStream leakage between subareasFlow 0 0 1,275 between Difference 0 1,372 0 4,085 3,164 0 2,504 8,188 0 4,865 911 2,937 0 0 304 529 842 0 736 0 870 0 262 0 13,367 na 150 4,261 Irrigation returnIrrigation pondsSewage Mountain front Septic tank Stream leakage 21Arti 1,035 0 1,941 0 2 0Pumpage 0Drains 7,763 3,845 0 0 2,851 0 0 26,314 9,817 433 5,344 650 17,595 42 4,949 168 0 1,052 2,265 8,733 75,262 9,745 0 0 14,834 0 0 72 26,369 647 0 1,282 586 4,216 0 259 0 39,610 0 0 22,930 0 427 247 0 0 11,645 54,351 0 0 3,501 165,920 9,987 Flow between subareasFlow 0 1,578 2,648 2,836 2,536 3,589 2,004 22 153 na Storage depletion values—Continued Table 12. Recharge Discharge

Evaluation of Selected Water-Management Alternatives 109 otal T Afton Canyon Lake Coyote ternatives 1, 2, and 3, 1999–2019 average Baja Lake Harper Centro Management alternative 3 Zone ransition T 683685 403 407 10,353 10,320 1,026 912 1,666 1,385 -699 -700 3,735 3,729 1,022 1,025 -31 -34 18,159 17,729 Este Oeste Alto 1 1, 2 Simulated hydrologic budgets for model subareas of the Mojave River ground-water basin, southern California, management al cial recharge 0 3,429 21,949 5,487 10,288 0 24,008 0 0 65,161 alues of storage differ as a result of accumulation of small, consistent errors in the model and rounding of large numbers. as a result of accumulation small, consistent errors in the model and rounding large alues of storage differ fi Positive storage value indicates storage depletion; negative storage value indicates storage accretion. storage value indicates storage depletion; negative storage value Positive V otal...... 1,058 6,948 71,801 29,529otal...... 1,750 33,224 6,321 4,641 83,097 38,438 31,406 353 33,841 581 4,216 171,207 42,044 1,507 550 189,366 1 2 T T recharge and discharge recharge Storage depletion Irrigation returnIrrigation pondsSewage Mountain front Septic tank Stream leakageArti 21 1,035 0 1,941 0 2 0Pumpage 0DrainsEvapotranspiration 7,763 3,845Head-dependent boundary 0 0Stream leakage 2,851 between subareasFlow 0 0 0 25,779 9,817 433 0 between Difference 1,275 5,344 650 17,537 42 4,949 0 0 168 0 1,052 1,372 0 12,791 2,265 75,262 9,745 0 4,085 0 3,750 0 0 0 14,834 0 0 2,504 72 8,500 647 26,369 0 1,448 0 0 586 0 4,865 1,468 4,216 0 259 5,568 0 0 0 39,610 0 0 0 0 22,930 1,139 0 428 247 0 0 0 529 0 869 11,645 0 57,983 0 0 736 3,501 0 0 1,036 0 165,920 9,987 0 0 262 0 0 524 14,849 137 na 151 0 7,894 137 566 Flow between subareasFlow 0 1,578 2,648 2,836 2,536 3,589 2,004 22 153 na Table 12. values—Continued Recharge Discharge

110 Simulation of Ground-Water Flow in the Mojave River Basin, California alternative 2), which resulted in the increases in simu- calibrated to steady-state and transient-state conditions lated hydraulic head. These results imply that the using a trial-and-error approach. recharge operations at El Mirage, Kane Wash, and Dag- The simulated steady-state hydraulic heads were gett may have benefitted from being distributed over a in good agreement with the measured 1930 water lev- larger area. els. The root mean square error (RMSE) was about 17 ft and the mean error (ME) was about 7 ft. The simulated transient-state hydraulic heads are SUMMARY in good agreement with measured 1931–94 water lev- els. The RMSE using 1992 measured water levels is The proximity of the Mojave River ground-water about 24 ft and the ME is about −3 ft. The hydrographs basin to the highly urbanized Los Angeles region has of the simulated floodplain and regional aquifer match led to rapid growth in population and, consequently, an the general trends of the measured water levels. The increase in the demand for water. The Mojave River, hydrographs of the simulated streamflows match mea- the primary source of surface water for the region, nor- sured peak flow rates and periods of no flow at the mally is dry—except for a small stretch with perennial Barstow and the Afton Canyon gages. The model flow and periods of flow after intense storms. Thus, the under- estimated streamflow over the entire simulation: region relies almost entirely on ground water to meet streamflow was underestimated by only 1,400 acre-ft, its agricultural and municipal needs. Ground-water or 0.04 percent, of measured streamflow for the Lower withdrawal since the late 1800’s has resulted in dis- Narrows; 49,400 acre-ft, or 4 percent, for Barstow; and charge, primarily from pumped wells, that exceeds nat- ural recharge. To plan for anticipated water demands 70,800 acre-ft, or 38 percent, for Afton Canyon. Most and for the effects of imported water on the basin, a of the underestimation at Afton Canyon was for 1969, ground-water flow model (MODFLOW-based) was a large stormflow year. Inaccuracies in measured developed to evaluate the geohydrologic conditions in streamflow and in estimated ungaged runoff also prob- the Mojave River ground-water basin and to project ably contribute to the underestimation of streamflow. ground-water conditions that will result from present A particle-tracking model was used to estimate and planned changes in the basin. steady-state ground-water flow directions and travel This study updates a previous analysis of the times in the Mojave River ground-water basin. Two basin completed by the U.S. Geological Survey in particle-tracking simulations were made; the first 1971. The effects of intermittent flows in the Mojave tracked mountain-front recharge in the ground-water River were incorporated into this study to help better system and the second tracked streamflow recharge. understand the relations between the regional and the The results of mountain-front recharge particle floodplain aquifer systems and to develop a tool for tracking were in good agreement with other published anticipating the effects of future stresses on the ground- results in terms of travel times from the recharge sites water system. to the area west of Victorville (5,000 to 6,000 years). The ground-water flow model has two horizontal The results of the particle tracking of streamflow layers, the top layer (layer 1) corresponds to the flood- recharge indicate that most particles quickly leave and plain aquifer and the bottom layer (layer 2) corre- reenter the river, except the particles starting in the sponds to the regional aquifer. The area represented by West Fork of the Mojave River. each cell in the model is 2,000 by 2,000 ft. Each calen- The complaint that resulted in the adjudication of dar year of the transient-state simulation was repre- the Mojave River ground-water basin alleged that the sented by two stress periods (wet and dry). The cumulative water production upstream of the city of duration of each stress period was a function of the Barstow had overdrafted the Mojave River ground- occurrence, quantity of discharge, and length of storm- water basin. In order to ascertain the effect of pumping flow from the headwaters of the Mojave River each on the ground-water and surface-water relations along year. The model boundary types were no flow and gen- the Mojave River, two pumping simulations were com- eral head. The model incorporated the following pared with the 1931–90 transient-state simulation (base optional MODFLOW packages: horizontal flow bar- case). For the first simulation, 1931–90 pumping rates rier, evapotranspiration, stream, drain, recharge, and were maintained in the upper region (Este, Oeste, Alto, well. The recharge component was subdivided into and Transition zone model subareas) with no pumping mountain-front and artificial recharge. The model was in lower region, and for the second simulation,

Summary 111 1931–90 pumping rates were maintained in the lower To visualize the magnitude, spatial distribution, region (Centro, Harper Lake, Baja, Coyote Lake, and and timing of water-level changes in the basin through Afton Canyon model subareas) with no pumping in the time, simulated hydraulic heads for 1932–99 were upper region. compared with simulated hydraulic heads for 1931. In the upper region, assuming pumping only in Greater than average annual inflows to the Mojave the upper region, there was no change in storage; in River from the headwaters during the late 1930’s and recharge from, and discharge to, the Mojave River; and throughout much of the 1940’s resulted in simulated in evapotranspiration compared with the base case. In hydraulic heads that were higher than the 1931 hydrau- the lower region, assuming pumping only in the upper lic heads along the Mojave River in most model subar- region, storage increased, recharge from the Mojave eas. Parts of the Baja and Harper Lake model subareas River decreased, and discharge to the Mojave River and had declines in simulated hydraulic head because of the evapotranspiration increased compared with the base increase in agricultural pumpage. By 1960, the simu- case. lated hydraulic heads were lower than the simulated In the upper region, assuming pumping only in hydraulic heads for 1931 in all model subareas of the the lower region, storage increased, recharge from the floodplain and the regional aquifers because of pump- Mojave River decreased and discharge to the Mojave age. After 1960, the size and the magnitude of the areas River and evapotranspiration increased compared with of the regional aquifer that had simulated hydraulic the base case. In the lower region, assuming pumping heads lower than those for 1931 continued to increase only in the lower region, there was less storage deple- until the end of the simulation (1999). Along the tion and recharge from, and discharge to, the Mojave Mojave River, hydraulic heads fluctuated in the flood- River and evapotranspiration increased compared with plain aquifer in response to recharge during years with the base case, with the greatest change occurring in the large inflows with little apparent effect on the simulated Centro model subarea. Overall, pumping in the lower hydraulic heads in the regional aquifer. region does not negatively affect the upper region; Three water-management alternatives were however, pumping in the upper region negatively evaluated to determine their effect on ground-water affects the lower region by decreasing recharge from resources using the calibrated ground-water flow the Mojave River. model. The water-management alternatives were simu- Streamflow, pumpage, and water-level data for lated assuming artificial recharge of imported calendar years 1995–99 were used to validate the cali- California State Water Project water allocated to the brated ground-water flow model, that is, to test that the Mojave Water Agency (MWA); the first simulation flow model will duplicate measured data for a non- assumes that zero percent of the MWA allocation is calibration period without modification of the model available for recharge (alternative 1); the second parameters. In general, the simulated results are in assumes that 50 percent of the MWA allocation is avail- good agreement with the measured data except for the able (alternative 2); and the third assumes that 100 per- Oeste model subarea where simulated hydraulic heads cent of the MWA allocation is available (alternative 3). show a pumping depression near El Mirage dry lake Each of the three water-management alternatives were that was overestimated by model at the pumping center. simulated for a 20-year drought. Streamflow conditions The RMSE, using 1998 measured water levels, is about were simulated using the 20-year drought of 1945–64 29 ft and the ME is about −6 ft. In general, the simu- with associated calibrated stream parameters. lated hydrographs for wells in the floodplain and the For management alternative 1, the response of regional aquifers follow the measured water-level the ground-water system was simulated assuming cur- trends. Simulated streamflow data for the 1995–99 wet rent (1999) rates of pumpage continue during a 20-year and dry stress periods at the Lower Narrows, Barstow, drought and no MWA allocation of water available for and Afton Canyon were compared with average mea- artificial recharge to mitigate the effects of the drought. sured streamflow data for the same periods and gener- The model simulated recharge from the Mojave River ally reflect 1995–99 streamflow conditions. The results of about 52,300 acre-ft/yr; this is a 61,400 acre-ft/yr indicate that the streambed conductance values cali- reduction in recharge compared with the 1995–99 aver- brated to the 1931–94 conditions reasonably simulate age. This reduction in recharge is reflected in simulated the 1995–99 conditions and therefore can be used for hydraulic-head declines between 1999 and 2019 of as predictive purposes. much as 45 ft. The model simulated evapotranspiration

112 Simulation of Ground-Water Flow in the Mojave River Basin, California decreases of about 5,600 acre-ft/yr and ground-water Chow, V.T., 1964, Handbook of applied hydrology: discharge to the Mojave River decreases of about McGraw-Hill, New York, 1418 p. 3,900 acre-ft/yr compared with the 1995–99 averages; Cox, B.F., and Hillhouse, J.W., 2000, Pliocene and Pleis- these reductions are related to the declines in simulated tocene evolution of the Mojave River, and associated hydraulic heads. tectonic development of the Transverse Ranges and Mojave Desert, based on borehole stratigraphy studies For management alternatives 2 and 3, the near Victorville, California: U.S. Geological Survey response of the ground-water system was simulated Open-File Report 00-147, 66 p. assuming current (1999) rates of pumpage continue Cox, B.F., and Wilshire, H.G., 1993, Geologic map of the during a 20-year drought and 50 or 100 percent (man- area around the Nebo Annex, Marine Corps Logistics agement alternatives 2 and 3, respectively) of the MWA Base, Barstow, California: U.S. Geological Survey allocation of water is available for artificial recharge to Open-File Report 93-568, 36 p., 1 map sheet, scale mitigate the effects of the drought. The model simu- 1:12,000. lated very little change in recharge from the Mojave Densmore, J.N., Cox, B.F., and Crawford, S.M., 1997, Geo- River for management alternative 2 and about a hydrology and water quality of Marine Corps Logistics 5,600 acre-ft/yr increase for management alternative 3 Base, Nebo and Yermo Annexes, near Barstow, when compared with management alternative 1. The California: U.S. Geological Survey Water-Resources Investigations Report 96-4301, 116 p. artificial recharge results in increases in simulated Dibble, E.F., 1967, Mojave Water Agency, water production hydraulic head for management alternatives 2 and 3 at verification program: Consultant’s report prepared for each of the artificial-recharge sites. The simulated and in the files of Mojave Water Agency, 39 p. increases in hydraulic head result in increased evapo- Dibblee, T.W., Jr., 1961, Evidence of strike-slip movement transpiration and ground-water discharge to the on northwest-trending faults in the Mojave Desert: U.S. Mojave River when compared with management alter- Geological Survey Professional Paper 424-B, native 1. p. B197–B199. The largest increases in simulated hydraulic ———1967 Areal geology of the western Mojave Desert, heads for management alternatives 2 and 3 were for the California: U.S. Geological Survey Professional Paper El Mirage, Kane Wash, and Daggett artificial recharge 522, 153 p. Dokka, R.K., and Travis, C.J., 1990, Late Cenozoic strike- sites. The increases at El Mirage and Kane Wash are the slip faulting in the Mojave Desert, California: Tecton- result of recharging water into areas of low transmissiv- ics, v. 9, no. 2, p. 311–340. ity for model layers 1 and 2. Although the Daggett site Driscoll, F.G., 1986, Groundwater and wells: St. Paul, is located on the Mojave River in an area of high trans- Minnesota, Johnson Division, 2nd ed., 1089 p. missivity, the area of high transmissivity is relatively Durbin, T.J., and Hardt, W.F., 1974, Hydrologic analysis of narrow and the recharge flow rate is relatively high the Mojave River, California, using a mathematical resulting in the increases in simulated hydraulic head. model: U.S. Geological Survey Water-Resources Inves- These results imply that the recharge operations at El tigations Report 17-74, 50 p. Mirage, Kane Wash, and Daggett may benefit from Environmental Systems Research Institute, 1992, ARC/ being distributed over a larger area. INFO layers CD-ROM, in Mojave Desert Ecosystem Program: [Redlands, Calif.] [disc 6], ARC/INFO data format. Erie, L.J., French, O.F., and Harris, Karl, 1965, Consumptive SELECTED REFERENCES use of water by crops in Arizona: Arizona University Technical Bulletin 169, 44 p. California Department of Finance, Historical census popula- Fillmore, R.P., and Walker, J.D., 1996, Evolution of a supra- tions of places, towns, and cities in California, 1850– detachment extensional basin: The lower Miocene 1990: accessed November 28, 1998, at URL Pickhandle basin, central Mojave Desert, California, in http://www.dof.ca.gov/html/Demograp/histtext.htm Beratan, K.K., ed., Reconstructing the History of Basin California Department of Public Works, Division of Water and Range Extension Using Sedimentology and Resources, 1934, Mojave River investigation: Califor- Stratigraphy: Boulder, Colorado, Geological Society of nia Department of Public Works Bulletin 47, 249 p. America Special Paper 303, p. 107–126. California Department of Water Resources, 1967, Mojave Galloway, D.L., Jones, D.R., Ingebritsen, S.E., eds., 1999, River ground-water basins investigation: California Land subsidence in the United States: U.S. Geological Department of Water Resources Bulletin 84, 151 p. Survey Circular 1182, p. 177.

Selected References 113 Hardt, W.F., 1971, Hydrologic analysis of Mojave River Palinspastic Reconstruction, and Geologic Evolution: basin, California, using electric analog model: U.S. Boulder, Colorado, Geological Society of America Geological Survey Open-File Report, 84 p. Memoir 178, p. 107–159. Hedman, E.R., 1970, Mean annual runoff as related to chan- McDonald, M.G., and Harbaugh, A.W., 1988, A modular nel geometry of selected streams in California: U.S. three-dimensional finite-difference ground-water flow Geological Survey Water-Supply Paper 1999-E, 17 p. model: U.S. Geological Survey Techniques of Water- Howard, K.A., and Miller, D.M., 1992, Late Cenozoic fault- Resources Investigations, book 6, Chap. A1, 576 p. ing at the boundary between the Mojave and Sonoran Meisling, K.E., and Weldon, R.J., 1989, Late Cenozoic tec- blocks: Bristol Lake area, California, in Richard, S.M., tonics of the northwestern San Bernardino Mountains, ed., Deformation associated with the Neogene Eastern southern California: Geological Society of America California Shear Zone, southeastern California and Bulletin, v. 101, p. 106–128. southwestern Arizona: Redlands, California, San Mendez, G.O., and Christensen, A.H., 1997, Regional water Bernardino County Museum Special Publication, 92-1, table (1996) and water-level changes in the Mojave p. 37–47. River, the Morongo, and the Fort Irwin ground-water Hsieh, P.A., and Freckleton, J.R., 1993, Documentation of a basins, San Bernardino County, California: U.S. computer program to simulate horizontal-flow barriers Geological Survey Water-Resources Investigations using the U.S. Geological Survey’s modular three- Report 97-4160, 34 p., 1 pl. dimensional finite-difference ground-water flow model: Michel, R.L., 1996, Use of tritium to estimate recharge under U.S. Geological Survey Open-File Report 92-477, 32 p. an intermittent stream wash in the Mojave Desert, Hughes, J.L., 1975, Evaluation of ground-water degradation Southern California: Eos, Transactions of the American resulting from waste disposal to alluvium near Barstow, Geophysical Union, Abstracts, v. 7, Abstract H32G-12, California: U.S. Geological Survey Professional Paper p. F263. 878, 33 p. Mojave Basin Area Watermaster, 1996a, Second annual Izbicki, J.A., Martin, Peter, and Michel, R.L., 1995, Source, report of the Mojave Basin area watermaster: Apple movement and age of groundwater in the upper part of Valley, California, 133 p. the Mojave River basin, California, U.S.A., in Adar, ———, 1996b, Table C-1 Sample calculation, Mojave Basin E.M., and Leibundgut, Christian, eds., Application of area adjudication subarea hydrological inventory based tracers in arid zone hydrology: International on long-term average natural water supply and outflow Association of Hydrological Sciences, no. 232, and current year imports and consumptive use in Supe- p. 43–56. rior Court of California, Riverside County, case no. James, John, 1992, Precipitation/evaporation climatology of 208565, exhibit C-1 to City of Barstow et al. v. City of the Mojave Water Agency: Apple Valley, California: Adelanto et al.: 7 p. Mojave Water Agency, 21 p., 1 pl. Mojave Water Agency, 1996, First annual engineer’s report Jefferson, G.T., 1985, Stratigraphy and geologic history of on water supply for water year 1994–95: Victorville, the Pleistocene Manix Formation, central Mojave California, 55 p. Desert, California, in Reynolds, R.E., ed., Geologic investigations along Interstate 15, Cajon Pass to Manix ———1999, Fourth annual engineer’s report on water Lake, California: Redlands, California, San Bernardino supply for water year 1997–1998: Apple Valley, County Museum, p. 157–169. California, 77 p. Lines, G.C., 1996, Ground-water and surface-water relations Montgomery Watson, Consultants, 1995, Installation along the Mojave River, Southern California: U.S. Restoration Program, Draft Operable Unit 1 Pre-Design Geological Survey Water-Resources Investigations Study, George Air Force Base, California, vol. 1. Report 95-4189, 43 p. Nagy, E.A., and Murray, Bruce, 1996, Plio-Pleistocene Lines, G.C., and Bilhorn, T.W., 1996, Riparian vegetation deposits adjacent to the Manix fault: implications for and its water use during 1995 along the Mojave River, the history of the Mojave River and Transverse Ranges southern California: U.S. Geological Survey Water- uplift: Sedimentary Geology, v. 103, nos. 1–2, p. 9–21. Resources Investigations Report 96-4241, 10 p., 1 pl. National Oceanic and Atmospheric Administration, 1994, Lohman, S.W., 1972, Ground-water hydraulics: U.S. Climatological data—California: v. 98, nos. 1–12. Geological Survey Professional Paper 708, 70 p. Pollock, D.W., 1994, User’s guide for MODPATH/ Matti, J.C., and Morton, D.M., 1993, Paleogeographic evo- MODPATH-PLOT, version 3: a particle tracking post- lution of the San Andreas Fault in southern California: processing package for MODFLOW, the U.S. A reconstruction based on a new cross-fault correlation, Geological Survey finite-difference ground-water flow in Powell, R.E., Weldon, R.J., II, and Matti, J.C., eds., model: U.S. Geological Survey Open-File Report The San Andreas Fault System: Displacement, 94-464.

114 Simulation of Ground-Water Flow in the Mojave River Basin, California Prudic, D.E., 1989, Documentation of a computer program Smith, G.A, and Pimentel, M.I., 2000, Regional water table to simulate stream-aquifer relations using a modular, (1998) and ground-water-level changes in the Mojave finite-difference, ground-water flow model: U.S. River and the Morongo ground-water basins, San Geological Survey Open-File Report 88-729, 113 p. Bernardino County, California: U.S. Geological Survey Rector, Carol, 1999, Human and animal trackway at Oro Water-Resources Investigations Report 00-4090, 107 p. Grande, in Reynolds, R.E., ed., Fossil footprints: Southern California Edison, 1983, Eastern Mojave Resource Redlands, California, San Bernardino County Museum Inventory database. Association, Abstracts from proceedings, the 1999 Stamos, C.L., and Predmore, S.K., 1995, Data and water- Desert Research Symposium, p. 53–55. table map of the Mojave River ground-water basin, San Rector, C.H., Swenson, J.D., and Wilke, P.J., 1983, Archae- Bernardino County, California, November 1992: U.S. ological studies at Oro Grande, Mojave Desert, Geological Survey Water Resources Investigations California: Redlands, California, San Bernardino Report 95-4148, 1 pl. County Museum Association, 181 p. Subsurface Surveys, Inc., 1990, Inventory of groundwater Reynolds, R.E., and Reynolds, R.L., 1985, Late Pleistocene stored in the Mojave River Basins: Consultant’s report faunas from Daggett and Yermo, San Bernardino prepared for and in the files of Mojave Water Agency, County, California, in Reynolds, R.E., compiler, geo- 47 p. logic investigations along Interstate 15, Cajon Pass to Sun, N-Z., 1994, Inverse problems in groundwater model- Manix Lake, California: San Bernardino County ing: Kluwer, Norwell, MA, 337 p. Museum, p. 175–191. Thompson, D.G., 1929, The Mohave Desert region, ———1991, Structural implications of late Pleistocene fau- California—A geographic, geologic, and hydrologic nas from the Mojave River Valley, California, in reconnaissance: U.S. Geological Survey Water-Supply Woodburne, M.O., Reynolds, R.E., and Whistler, D.P., Paper 578, 759 p. eds., Inland southern California: The last 70 million years: San Bernardino County Museum Association Umari, A.M.J., Martin, Peter, Schroeder, R.A., Duell, L.F.W. Quarterly, v. 38, nos. 3 and 4, p. 100–105. Jr., and Fay, R.G., 1995, Potential for ground-water Robson, S.G., 1974, Feasibility of digital water-quality mod- contamination from movement of wastewater through eling illustrated by application at Barstow, California: the unsaturated zone, upper Mojave River Basin, U.S. Geological Survey Water-Resources Investiga- California: U.S. Geological Survey Water-Resources tions Report 46-73, 66 p. Investigations Report 93-4137, 83 p. Rockwell, G.L., Anderson, S.W., and Agajanian, J., 1999, U.S. Bureau of Reclamation, 1952, Report on Victor Project, Water resources data, California, water year 1998. California: Regional Director’s Report, Boulder City, Volume1. Southern Great Basin from Mexican Border Nevada, 42 p. to Mono Lake Basin, and Pacific Slope Basins from U.S. Department of Agriculture, Soil Conservation Service, Tijuana River to Santa Maria River: U.S. Geological 1967, Irrigation Water Requirements, Technical Survey Water-Data Report 98-1, 427 p. Release No. 21, Revised September 1970. Schaefer, D.H., 1979, Ground-water conditions and potential Woodburne, M.O., Tedford, R.H., and Swisher, C.C., III, for artificial recharge in Lucerne Valley, San Bernardino 1990, Lithostratigraphy, biostratigraphy, and geochro- County, California: U.S. Geological Survey Water- nology of the Barstow Formation, Mojave Desert, Resources Investigations Report 78-118, 37 p. southern California: Geological Society of America Searcy, J.K., 1959, Flow-duration curves: U.S. Geological Bulletin, v. 102, p. 459–477. Survey Water-Supply Paper 1542-A, 33 p. Zohdy, A.A.R., and Bisdorf, R.J., 1994, A direct-current Slichter, C.S., 1905, Field measurements of the rate of move- resistivity survey near the Marine Corps Logistics ment of underground waters: U.S. Geological Survey Bases at Nebo and Yermo, Barstow, California: U.S. Water-Supply and Irrigation Paper 140, 122 p. Geological Survey Open-File Report 94-202, 155 p.

Selected References 115 Page 116 LEFT BLANK INTENTIONALLY APPENDIX 1

Appendix 1 117 Alto Subarea (Regional system) 2,875 2,875

4N/3W-1M1 5N/3W-13D1, 14G1 2,850 2,850

2,825 2,825

13D1 2,800 2,800 14G1

2,775 2,775 2,875 2,900

4N/3W-10R1 5N/3W-22J1 2,850 2,875

2,825 2,850

2,800 2,825

2,775 2,800 2,900 2,875 4N/3W-20P1 5N/3W-24N1, -23R1 2,875 2,850

2,850 2,825

24N1 2,825 2,800 23R1

2,800 2,775 2,875 2,800 4N/4W-8G1, 8N2 5N/4W-11P1, 11P3 2,850 2,775 11P1

HYDRAULIC HEAD, IN FEET ABOVE SEA LEVEL 11P3 2,825 2,750

8G1 2,800 2,725 8N2

2,775 2,700 2,850 1930 1940 1950 1960 1970 1980 1990 2000 5N/5W-22E1, 22E2 YEAR 2,825 EXPLANATION

2,800 Simulated – Model layer 1 22E1 Model layer 2 2,775 22E2 Measured – 11P1 2,750 Data point of well indicated 1930 1940 1950 1960 1970 1980 1990 2000 11P3 YEAR

Appendix 1. Measured and model-simulated hydraulic heads at selected wells in the Mojave River ground-water basin, southern California, 1931–99. (See figure 29 for location of wells).

118 Simulation of Ground-Water Flow in the Mojave River Basin, California Alto Subarea (Floodplain aquifer system) 2,875 2,850 4N/3W-18R1, R2, R3 4N/3W-6A1 2,850 2,825

2,825 18R2 2,800 18R2 18R3 2,800 2,775

2,775 2,750 2,875 2,850 4N/4W-18E1 5N/4W-36N1, 4N/4W-1D2 2,850 2,825

2,825 2,800 36N1 2,800 2,775 1D2

2,775 2,750 1930 1940 1950 1960 1970 1980 1990 2000 2,850 YEAR 4N/3W-6D1, 6D2 HYDRAULIC HEAD, IN FEET ABOVE SEA LEVEL 2,825 EXPLANATION 2,800 Simulated – Model layer 1 6D1 2,775 Model layer 2 6D2 Measured – 6D1 2,750 Data point of well indicated 1930 1940 1950 1960 1970 1980 1990 2000 6D2 YEAR

Appendix 1.—Continued.

Appendix 1 119 Alto Subarea (Transition Zone) 2,550 2,450 7N/4W-30C1 8N/4W-20N1 2,525 2,425

2,500 2,400

2,475 2,375

2,450 2,350 1930 1940 1950 1960 1970 1980 1990 2000 2,475 YEAR 8N/4W-31R1 2,450 EXPLANATION Simulated – 2,425 Model layer 1 HYDRAULIC HEAD, IN FEET ABOVE SEA LEVEL Model layer 2 2,400 Measured –

2,375 Data point of well indicated 1930 1940 1950 1960 1970 1980 1990 2000 YEAR

Appendix 1.—Continued.

120 Simulation of Ground-Water Flow in the Mojave River Basin, California Centro Subarea

2,350 2,075 8N/4W-12Q1 11N/3W-28R1,28R2 2,325 2,050

2,300 2,025 28R1 28R2 2,275 2,000

2,250 1,975 2,200 2,050 9N/2W-19B1, 13R1 11N/3W-30A1,30A2 2,175 2,025

2,150 2,000 19B1 30A1 2,125 13R1 1,975 30A2

2,100 1,950 2,175 2,050 10N/3W-35N1 11N/4W-29R1

HYDRAULIC HEAD, IN FEET ABOVE SEA LEVEL 2,000 2,150 1,950 2,125 1,900

2,100 1,850 Note: scale change 2,075 1,800 1930 1940 1950 1960 1970 1980 1990 2000 1930 1940 1950 1960 1970 1980 1990 2000 YEAR YEAR

EXPLANATION Simulated – Measured – 30A1 Model layer 1 Data point of well indicated Model layer 2 30A2

Appendix 1.—Continued.

Appendix 1 121 Baja Subarea (Regional system) 1,900 1,875 9N/1E-3H1, 3H5 9N/2E-11H1, 11H3 1,875 1,850

1,850 1,825

3H1 11H1 1,825 1,800 3H5 11H3

1,800 1,775 1,900 1,875 9N/2E-20Q1, 20K1 9N/2E-12N1 1,875 1,850 1,850

1,825 1,825

1,800 20Q1 1,800 1,775 20K1 Note: scale change 1,750 1,775 1,875 1,875 9N/2E-14N1, 14N2 9N/3E-19E1 1,850 1,850

1,825 1,825

1,800 1,800 14N1 1,775 14N2 1,775 Note: scale change Note: scale change

1,750 1,750 1,760 1,800 10N/3E-21A1, 15Q1 9N/4E-31K1 1,735 1,775 HYDRAULIC HEAD, IN FEET ABOVE SEA LEVEL

1,710 1,750

21A1 1,685 1,725 15Q1

1,660 1,700 1,825 1,800 9N/3E-34D1, 34N1 8N/4E-7E1 1,800 1,775 1,775

1,750 1,750

1,725 34D1 1,725 34N1 1,700 Note: scale change 1,675 1,700 1930 1940 1950 1960 1970 1980 1990 2000 1930 1940 1950 1960 1970 1980 1990 2000 YEAR YEAR EXPLANATION Simulated – Measured – 34D1 Model layer 1 Data point of well indicated Model layer 2 34N1

Appendix 1.—Continued.

122 Simulation of Ground-Water Flow in the Mojave River Basin, California Baja Subarea (Floodplain aquifer) 2,080 1,900 9N/1W-11M1, 11M11 9N/2E-18F1

2,055 11M1 1,850 11M11 2,030

1,800 2,005 Note: scale change 1,980 1,750 2,060 1,875

9N/1W-11R2 9N/2E-3K1 2,035 1,850

2,010 1,825

1,985 1,800

1,960 1,775 1,925 1,850 9N/1E-21H1, 22D1 9N/2E-3A2 1,900 1,825

1,875 1,800

1,850 21H1 1,775 22D1 1,825 1,750 Note: scale change Note: scale change 1,800 1,725 1,920 1,800 9N/1E-10L1 9N/3E-3D1, 3D2 1,895

HYDRAULIC HEAD, IN FEET ABOVE SEA LEVEL 1,750

1,870 3D1 1,700 3D2 1,845 Note: scale change 1,820 1,650 1,900 1930 1940 1950 1960 1970 1980 1990 2000 9N/1E-13E2, 15H1 YEAR 1,875 EXPLANATION Simulated – 1,850 Model layer 1 13E2 Model layer 2 1,825 15H1 3D1 Measured – 3D2 Data point of well indicated 1,800 1930 1940 1950 1960 1970 1980 1990 2000 YEAR

Appendix 1.—Continued.

Appendix 1 123 Page 124 LEFT BLANK INTENTIONALLY APPENDIX 2

Appendix 2 125 Alto Subarea 2,925 2,825 4N/3W-31L7-9 4N/4W-3A3-5 2,900 2,800

2,875 2,775

2,850 L7 380 ft 2,750 A3 510 ft L8 260 ft A4 360 ft L9 140 ft A5 235 ft 2,825 2,725 2,900 2,875 G2 600 ft 4N/3W-19G2-6 4N/3W-12A1,2 G3 375 ft 2,875 G4 195 ft 2,850 G5 95 ft G6 55 ft 2,850 2,825

2,825 2,800 A1 600 ft A2 345 ft

2,800 2,775 2,825 2,850

4N/4W-1C2-5 E3 236 ft 5N/3W-27E3-6 E4 194 ft 2,800 2,825 E5 168 ft E6 148 ft

2,775 2,800

C2 620 ft 2,750 C3 330 ft 2,775 C4 190 ft C5 80 ft 2,725 2,750 2,775 2,825

HYDRAULIC HEAD, IN FEET ABOVE SEA LEVEL 5N/4W-14D1-4 5N/3W-30A1-3 2,750 2,800

2,725 2,775

D1 340 ft A1 280 ft 2,700 D2 200 ft 2,750 A2 228 ft D3 100 ft A3 195 ft D4 50 ft 2,675 2,725 1992 1994 1996 1998 2000 2,950 YEAR 5N/6W-22E1-3 2,925 EXPLANATION Simulated – 2,900 Model layer 1 Model layer 2 E1 750 ft 2,875 E2 565 ft Measured – E3 400 ft A1 280 ft Multiple-well completion site A2 228 ft and depth – In feet (ft). Datum is 2,850 A3 195 ft sea level 1992 1994 1996 1998 2000 YEAR

Appendix 2. Measured and model-simulated hydraulic heads at multiple-well completion sites, Mojave River ground-water basin, southern California, 1992–99. (See figure 29 for location of wells.)

126 Simulation of Ground-Water Flow in the Mojave River Basin, California Alto Subarea (Transition zone) 2,510 2,450 7N/5W-23R3 8N/4W-20Q7-12 Q7 460 ft Q8 350 ft 2,485 2,425 Q9 270 ft Q10 160 ft Q11 50 ft 2,460 2,400 Q12 139 ft R3 315 ft

2,435 2,375

2,410 2,350 1992 1994 1996 1998 2000 2,510 YEAR

2,485 7N/5W-24R7,8 EXPLANATION Simulated – 2,460 Model layer 1

HYDRAULIC HEAD, IN FEET ABOVE SEA LEVEL Model layer 2 2,435 R7 150 ft R8 50 ft Measured – Multiple-well completion site R7 150 ft 2,410 R8 50 ft and depth – In feet (ft). Datum is 1992 1994 1996 1998 2000 sea level YEAR

Appendix 2.—Continued.

Appendix 2 127 Centro Subarea 2,400 2,100 8N/4W-21M1-4 9N/1W-9D5-8 2,375 2,075

2,350 2,050 M1 370 ft M2 230 ft 2,325 M3 140 ft 2,025 M4 40 ft D5 500 ft D7 190 ft D6 300 ft D8 80 ft 2,300 2,000 2,175 2,100

R5 330 ft 9N/3W-1R5-7 M4 440 ft 9N/1W-4M4-7 R6 210 ft M5 250 ft 2,150 R7 120 ft 2,075 M6 160 ft M7 80 ft

2,125 2,050

2,100 2,025

2,075 2,000 2,150 2,075 9N/2W-3E1-3 9N/2W-3E1-3 9N/1W-4R2-4 2,125 2,050

2,100 2,025 HYDRAULIC HEAD, IN FEET ABOVE SEA LEVEL 280 ft 230 ft R2 E1 R3 140 ft 2,075 E2 185 ft 2,000 R4 40 ft E3 120 ft

2,050 1,975 1992 1994 1996 1998 2000 2,025 YEAR YEAR 9N/1W-12N4-7 2,000 EXPLANATION Simulated – 1,975 Model layer 1 Model layer 2 N4 640 ft 1,950 N5 310 ft Measured – N6 170 ft N7 80 ft R2 280 ft Multiple-well completion site R3 140 ft and depth – In feet (ft). Datum is 1,925 1992 1994 1996 1998 2000 R4 40 ft sea level YEAR

Appendix 2.—Continued.

128 Simulation of Ground-Water Flow in the Mojave River Basin, California Baja Subarea 1,900 1,900 F1 410 ft F3 250 ft 10N/1E-20M1,2 F2 340 ft F4 150 ft 1,875 1,875 9N/1E-16F1-4

1,850 1,850

1,825 1,825 M1 350 ft M2 285 ft

1,800 1,800 1,875 1,775

9N/1E-4K1-3 R4 610 ft 9N/3E-22R4-7 1,850 1,750 R5 510 ft R6 290 ft R7 129 ft

1,825 1,725

1,800 K1 470 ft 1,700 K2 340 ft K3 195 ft 1,775 1,675 1,875 1,800 K5 650 ft 9N/2E-3K5-9 9N/2E-3G6-9 K6 510 ft 1,850 K7 340 ft 1,775 K8 210 ft K9 65 ft 1,825 1,750 HYDRAULIC HEAD, IN FEET ABOVE SEA LEVEL G6 600 ft 1,800 1,725 G7 490 ft G8 300 ft G9 140 ft 1,775 1,700 1992 1994 1996 1998 2000 1,775 YEAR J1 570 ft J4 90 ft J2 370 ft J5 45 ft 1,750 J3 255 ft EXPLANATION Simulated – 1,725 Model layer 1 Model layer 2 Measured – 1,700 10N/3E-27J1-5 K1 470 ft Multiple-well completion site K2 340 ft and depth – In feet (ft). Datum is K3 195 ft 1,675 sea level 1992 1994 1996 1998 2000 YEAR

Appendix 2.—Continued.

Appendix 2 129