The potential for subsurface storage and recovery, otherwise called aquifer storage and recovery, of effluent in the uppermost producing zone of the Upper Floridan aquifer in St. Petersburg, Florida, was studied by the U.S. Geological Survey, in cooperation with the city of St. Petersburg and the Southwest Florida Water Management District. The success of subsurface storage and recovery depends on the recovery efficiency, or the quantity of water, relative to the quantity injected, that can be recovered before the water that is withdrawn fails to meet salinity limits. The viability of this practice will depend upon the ability of the injected zone to receive, store, and discharge the injected fluid. A three-dimensional numerical model of ground-water flow and solute transport, incorporating available data on aquifer properties and water quality, was developed to evaluate the effects of changing various operational factors on recovery efficiency. The reference case for testing was a base model considered representative of the aquifer system underlying the Southwest St. Petersburg Water Treatment Facility. The base simulation used as a standard for comparison consisted of a single cycle of 90 days of simultaneous injection of effluent in three wells at a rate of 4.0 million gallons per day and then equal rate withdrawal of 4.0 million gallons per day until the pumped water in each well reached a dissolvedsolids concentration of 1,500 milligrams per liter. A recovery efficiency of 14.8 percent was estimated for the base simulation. Ten successive injection and recovery cycles increased recovery efficiency to about 56 percent. Based on model simulations for hypothetical conditions, recovery efficiency (1) increased with successive injection and recovery cycles; (2) increased when the volume of injectant increased; (3) decreased when storage time increased; (4) did not change significantly when the injection rate or recovery rate increased, or when the ratio of recovery rate to injection rate increased, and (5) was not significantly affected by any particular geometric arrangement of wells or by the number of wells when the volume of water injected remained constant. Recovery efficiency from multiple wells was nearly the same as from a single well. Recovery efficiency ranged from about 7 to 56 percent, in several tests. Sensitivity of recovery efficiency to variations in selected parameters such as dissolved-solids concentration of the injection zone, permeability, vertical anisotropy, longitudinal and transverse dispersivities, and effective porosity was tested. Changes in the dissolved-solids concentration of the injection zone produced the greatest change in recovery efficiency. Uniform changes in dispersivity values produced the second greatest change in recovery efficiency. Generally, recovery efficiency increased when the above parameter values were decreased and recovery efficiency decreased when these parameter values were increased. Density difference between native and injected waters was the most important factor affecting recovery efficiency in this study. For the base simulation, sensitivity tests indicated that recovery efficiency increased from about 15 to 78 percent when the dissolved-solids concentration of the native water decreased from about 7,800 to 500 milligrams per liter. Dispersivity is another important factor affecting recovery efficiency. For the base simulation, sensitivity tests indicated that recovery efficiencies from about 9 to 24 percent can be obtained for different dispersivity values. A field determination of dispersivity was not made as part of this study, and values used may not be representative of the actual dispersive characteristics of the aquifer system at the study site. However, dispersivity values tested are within the range of values used in previous studies.