Ground water from the Upper Floridan aquifer is the sole source of water supply for Tallahassee, Florida, and the surrounding area. The City of Tallahassee (the City) currently operates 28 water-supply wells; 26 wells are distributed throughout the City and 2 are located in Woodville, Florida. Most of these wells yield an ample supply of potable water; however, water from several wells has low levels of tetrachloroethylene (PCE). The City removes the PCE from the water by passing it through granular-activated carbon units before distribution. To ensure that water-supply wells presently free of contamination remain clean, it is necessary to understand the ground-water flow system in sufficient detail to protect the contributing areas.

Ground-water samples collected from four public-supply wells were analyzed for tritium (³H), chlorofluorocarbons (CFCs), and sulfur hexafluoride (SF₆). Using data for the CFC compounds, apparent ground-water ages ranged from 7 to 31 years. For SF₆, the apparent ages tended to be about 5 to 10 years younger than those from CFCs. Apparent ages based on the tritium/tritiogenic helium-3 (³H/³He_trit) method ranged from 26 to 33 years. The three dating methods indicate that the apparent age of ground water generally decreases from northern to southern Leon County. This southward trend of decreasing ages is consistent with increasing amounts of recharge that occur as ground water moves from north to south.

The ground-water age data derived by geochemical and tracer analyses were used in combination with the flow model and particle tracking to determine an effective porosity for the Hawthorn clays and Upper Floridan aquifer. The effective porosities for the Upper Floridan aquifer that resulted in best model matches were averaged to produce an effective porosity of 7 percent, and the effective porosities for the Hawthorn clays that resulted in a match were averaged to produce an effective porosity of 22 percent.

Probabilistic contributing areas were determined for 26 City wells using MODFLOW and MODPATH. For each probabilistic contributing area delineated, the model was run 100 times and the results were analyzed statistically. For each of the 100 runs, a different hydraulic conductivity for each of the zones was assigned to the Upper Floridan aquifer. The hydraulic conductivities were generated randomly assuming a lognormal probability distribution; the mean of the distribution was equal to the hydraulic conductivity from the calibrated model.

The 5-year time-dependent capture zones (TDCZs), assuming effective porosities of 0.1, 1, and 7 percent for four representative wells, were delineated. The higher probabilities of capture (greater than 40, 60, and 80 percent) were similar for all effective porosities, and the TDCZ delineated using a 7-percent porosity was slightly smaller; the lower probabilities of capture (greater than 10 and 20 percent) showed a large range of variability.