A local ground-water flow model is used to simulate four recharge scenarios along Rillito Creek in northern Tucson to evaluate mitigating effects on ground-water deficits and water-level declines in Tucson's Central Well Field. The local model, which derives boundary conditions from a basin-scale model, spans the 12-mile reach of Rillito Creek and extends 9 miles south into the Central Well Field. Recharge scenarios along Rillito Creek range from 5,000 to 60,000 acre-feet per year and are simulated to begin in 2005 and extend through 2225 to estimate long-term changes in ground-water level, ground-water storage, ground-water flux, and evapotranspiration. The base case for comparison of simulated water levels and flows, referred to as scenario A, uses a long-term recharge rate of 5,000 acre-feet per year to 2225. Scenario B, which increases the recharge along Rillito Creek by 9,500 acre-feet per year, has simulated water-level rises beneath Rillito Creek that range from about 53 feet to 86 feet. Water-level rises within the Central Well Field range from about 60 feet to 80 feet. More than half of these rises occur by 2050, and more than 95 percent occur by 2188. Scenario C, which increases the recharge along Rillito Creek by 16,700 acre-feet per year relative to scenario A, has simulated water-level rises beneath Rillito Creek that range from about 71 feet to 102 feet. Water-level rises within the Central Well Field range from about 80 feet to 95 feet. More than half of the rises occur by 2036, and more than 95 percent occur by 2100. Scenario D, which initially increases the recharge rate by about 55,000 acre-feet per year relative to scenario A, resulted in simulated water levels that rise to land surface along Rillito Creek. This rise in water level resulted in rejected recharge. As the water table continued to rise, the area of stream-channel surface intersected by the water table increased causing continual decline in the recharge rate until a long-term recharge rate of about 34,000 acre-feet per year was sustained. The long-term recharge rate for scenario D is about 29,000 acre-feet per year greater than the long-term recharge rate for scenario A. Simulated long-term water-level rises beneath Rillito Creek range from about 97 feet to 131 feet, resulting in water levels near or at the land surface. Shallow depths to water associated with this scenario have implications for contamination owing to the presence of landfills within or adjacent to Rillito Creek. Water-level rises for cells within the Central Well Field range from about 96 feet to 109 feet. More than half of the water-level rises occur by 2018 and more than 95 percent occur by 2041. Almost all the increased water added to the ground-water system in the recharge scenarios can be accounted for by a combination of increased storage near Rillito Creek, ground-water flux to the south, ground-water flux to the northwest, and increased discharge as evapotranspiration along Rillito Creek. The percentage of newly added water accounted for by storage changes is large relative to the percentage accounted for by changes in flux and evapotranspiration at the onset of each scenario; however, the changes in storage become smaller throughout the simulation, and the long-term component accounted for by storage is minimal. Long-term ground-water fluxes to the south increase by about 3,300, 4,840, and 7,500 acre-feet per year for scenarios B, C, and D, respectively. The percentage of increased recharge that flows south toward the Central Well Field, therefore, is 35, 29, and 26 percent for scenarios B, C, and D, respectively. Long-term ground-water fluxes to the northwest increase by about 3,100, 3,900, and 6,980 acre-feet per year for scenarios B, C, and D, respectively. The long-term percentage of increased recharge flowing northwestward is about 31, 25, and 21 percent for scenarios B, C, and D, respectively. Shallow ground-water evapotranspiration along Rillito Creek incr