Ground-water levels in Lancaster, California, declined more than 200 feet during the 20th century, resulting in reduced ground-water supplies and more than 6 feet of land subsidence. Facing continuing population growth, water managers are seeking solutions to these problems. Injection of imported, treated fresh water into the aquifer system when it is most available and least expensive, for later use during high-demand periods, is being evaluated as part of a management solution. The U.S. Geological Survey, in cooperation with the Los Angeles County Department of Public Works and the Antelope Valley-East Kern Water Agency, monitored a pilot injection program, analyzed the hydraulic and subsidence-related effects of injection, and developed a simulation/optimization model to help evaluate the effectiveness of using existing and proposed wells in an injection program for halting the decline of ground-water levels and avoiding future land subsidence while meeting increasing ground-water demand. A variety of methods were used to measure aquifer-system response to injection. Water levels were measured continuously in nested (multi-depth) piezometers and monitoring wells and periodically in other wells that were within several miles of the injection site. Microgravity surveys were done to estimate changes in the elevation of the water table in the absence of wells and to estimate specific yield. Aquifer-system deformation was measured directly and continuously using a dual borehole extensometer and indirectly using continuous Global Positioning System (GPS), first-order spirit leveling, and an array of tiltmeters. The injected water and extracted water were sampled periodically and analyzed for constituents, including chloride and trihalomethanes. Measured injection rates of about 750 gallons per minute (gal/min) per well at the injection site during a 5-month period showed that injection at or above the average extraction rates at that site (about 800 gal/min) was hydraulically feasible. Analyses of these data took many forms. Coupled measurements of gravity and water-level change were used to estimate the specific yield near the injection wells, which, in turn, was used to estimate areal water-table changes from distributed measurements of gravity change. Values of the skeletal components of aquifer-system storage, which are key subsidence-related characteristics of the system, were derived from continuous measurements of water levels and aquifer-system deformation. A numerical model of ground-water flow was developed for the area surrounding Lancaster and used to estimate horizontal and vertical hydraulic conductivities. A chemical mass balance was done to estimate the recovery of injected water. The ground-water-flow model was used to project changes in ground-water levels for 10 years into the future, assuming no injection, no change in pumping distribution, and forecasted increases in ground-water demand. Simulated ground-water levels decreased throughout the Lancaster area, suggesting that land subsidence would continue as would the depletion of ground-water supplies and an associated loss of well production capacity. A simulation/optimization model was developed to help identify optimal injection and extraction rates for 16 existing and 13 proposed wells to avoid future land subsidence and to minimize loss of well production capacity while meeting increasing ground-water demands. Results of model simulations suggest that these objectives can be met with phased installation of the proposed wells during the 10-year period. Water quality was not considered in the optimization, but chemical-mass-balance results indicate that a sustained injection program likely would have residual effects on the chemistry of ground water.