Abstract

Edwards Air Force Base (EAFB) in southern California historically has relied on ground water for its water-supply needs. Pumping of ground water at the base has led to problems such as declining water levels and land subsidence. For this study, a MODFLOW-based ground-water flow model was developed for EAFB to estimate the effects of pumping and injection strategies on water levels and on land subsidence.

The ground-water flow model grid has 154 rows and 126 columns of 660- by 660-foot cells. and the boundary of the active model grid basically corresponds to bedrock outcrops. The model has seven layers of varying thickness; model layers 2, 3, and 4 correspond to a thick clay layer in the southern half of the model area and model layer 5 corresponds to the middle aquifer which is the primary source of water supply for EAFB. The model was calibrated using a trial-and-error approach. Because relatively little ground-water development took place prior to the establishment of EAFB in 1947, assumed (predevelopment) steady-state conditions for model calibration were represented by a compilation of data from 1913-46. The steady-state simulated hydraulic gradient was northward [or all model layers. The transient-state 1996 simulated hydraulic heads indicate that a groundwater divide is located near the middle of the basin and that south of the divide ground water flows southward and north of the divide ground water flows northward. There are two subsidence centers located in the vicinity of the two primary pumping centers on the base. The maximum simulated hydraulic head change was about 150 feet and the maximum simulated subsidence was about 5 feet. The simulated results indicate that the greatest amount of compaction occurs in the middle aquifer (model layer 5). The simulated water budget rates for 1947,1961, and 1996 indicate that about 97, 98, and 76 percent, respectively, of the net discharge was derived from storage.

Results of the sensitivity analysis indicate that the model was sensitive to changes in the hydraulic-characteristic values of Faults 1 and 4 and the hydraulic conductivity values and inelastic storage values in layer 5. Specifically, increasing or decreasing the hydraulic-characteristic value of Fault 1 by an order of magnitude affected simulated hydraulic heads by as much as 45.0 feet and simulated subsidence by as much as 0.5 foot. Decreasing the hydraulic-characteristic value of Fault 4 by an order of magnitude affected simulated hydraulic heads by as much as 10.0 feet and simulated subsidence by as much as 0.5 foot. Decreasing the hydraulic conductivity values for layer 5 by 50 percent affected simulated hydraulic heads by as much as 10.0 feet and simulated subsidence by as much as 0.8 foot, and decreasing the inelastic storage values for layer 5 by an order of magnitude affected simulated hydraulic heads by as much as 45.0 feet and simulated subsidence by as much as 10.0 feet.

The simulated hydraulic head and subsidence results were not sensitive to vertical conductance areas 2 and 4; however, the results indicate that the clay layers in area 4 hydraulically control water levels and subsidence in these areas. When the vertical conductance was varied in area 2, the resulting simulated hydraulic heads and subsidence showed little change in either area; however, when the vertical conductance was varied in area 4, simulated hydraulic heads and simulated subsidence were affected in both areas.

Three water-management scenarios were tested for Edwards Air Force Base: for scenario 1 (base case), 1997 pumping rates were maintained for 10 years (1997-2006); for scenario 2, water was injected steadily into the middle aquifer at well 8N/10W-1C2 in the South Tract well field between December and February, concurrent with base-case pumping; and for scenario 3, water was injected steadily into the middle aquifer at well 9N/1OW-24E3 in the South Base well field, concurrent with base-case pumping. For scenarios 2 and 3, two separate cases were simulated: in the first case, about 3 acre-feet per day of water was injected, and in the second case about 30 acre-feet per day of water was injected. Injecting 3 acre-feet per day had little effect on simulated hydraulic heads; however, injecting 30 acre-feet per day raised simulated hydraulic heads more than 100.0 feet for both scenarios. In general, the injection of 3 acre-feet per day of water had little or no effect on the total subsidence over the lO-year simulation. Subsidence still accumulated with time when 30 acre-feet per day was injected at either site, but at a much lower rate than when 3 acre-feet per day was injected. Simulated subsidence decreased 60 percent more at the S-5 production well when 30 acre-feet per day of water was injected at well 8N/lOW-1C2 than when 3 acre-feet per day was injected, and simulated subsidence decreased 90 percent more at the South Base well field when 30 acre-feet per day of water was injected at well 9N/1OW-24E3 than when 3 acre-feet per day was injected.