The proximity of the Mojave River ground-water basin to the highly urbanized Los Angeles region has led to rapid growth in population and, consequently, to an increase in the demand for water. The Mojave River, the primary source of surface water for the region, normally is dry-except for a small stretch of perennial flow and periods of flow after intense storms. Thus, the region relies almost entirely on ground water to meet its agricultural and municipal needs. Ground-water withdrawal since the late 1800's has resulted in discharge, primarily from pumping wells, that exceeds natural recharge. To better understand the relation between the regional and the floodplain aquifer systems and to develop a management tool that could be used to estimate the effects that future stresses may have on the ground-water system, a numerical ground-water flow model of the Mojave River ground-water basin was developed, in part, on the basis of a previously developed analog model. The ground-water flow model has two horizontal layers; the top layer (layer 1) corresponds to the floodplain aquifer and the bottom layer (layer 2) corresponds to the regional aquifer. There are 161 rows and 200 columns with a horizontal grid spacing of 2,000 by 2,000 feet. Two stress periods (wet and dry) per year are used where the duration of each stress period is a function of the occurrence, quantity of discharge, and length of stormflow from the headwaters each year. A steady-state model provided initial conditions for the transient-state simulation. The model was calibrated to transient-state conditions (1931-94) using a trial-and-error approach. The transient-state simulation results are in good agreement with measured data. Under transient-state conditions, the simulated floodplain aquifer and regional aquifer hydrographs matched the general trends observed for the measured water levels. The simulated streamflow hydrographs matched wet stress period average flow rates and times of no flow at the Barstow and Afton Canyon gages. Steady-state particle-tracking was used to estimate travel times for mountain-front and streamflow recharge. The simulated travel times for mountain-front recharge to reach the area west of Victorville were about 5,000 to 6,000 years; this result is in reasonable agreement with published results. Steady-state particle-tracking results for streamflow recharge indicate that in most subareas along the river, the particles quickly leave and reenter the river. The complaint that resulted in the adjudication of the Mojave River ground-water basin alleged that the cumulative water production upstream of the city of Barstow had overdrafted the ground-water basin. In order to ascertain the effect of pumping on ground-water and surface-water relations along the Mojave River, two pumping simulations were compared with the 1931-90 transient-state simulation (base case). The first simulation assumed 1931-90 pumping in the upper region (Este, Oeste, Alto, and Transition zone model subareas) but with no pumping in the remainder of the basin, and the second assumed 1931-90 pumping in the lower region (Centro, Harper Lake, Baja, Coyote Lake, and Afton Canyon model subareas) but with no pumping in remainder of the basin. In the upper region, assuming pumping only in the upper region, there was no change in storage, recharge from the Mojave River, ground-water discharge to the Mojave River, or evapotranspiration when compared with the base case. In the lower region, assuming pumping only in the upper region, there was storage accretion, decreased recharge from the Mojave River, increased ground-water discharge to the Mojave River, and increased evapotranspiration when compared with the base case. In the upper region, assuming pumping only in the lower region, there was storage accretion, decreased recharge from the Mojave River, increased ground-water discharge to the Mojave River, and increased evapotranspiration when compared with the base case. In the