COMPARISON OF CONTRIBUTING AREAS TO A PUBLIC SUPPLY WELL USING TWO STOCHASTIC METHODS THAT INCORPORATE UNCERTAINTY OF PARAMETERS AND KARST FEATURES

Contributing areas to a public-supply well in Tampa, Florida were determined using two stochastic methods. The first method used parameter estimation to estimate values and statistical distributions of 18 hydraulic and recharge parameters. The model calibration was based on hydraulic head and flow observations from over 400 monitoring wells and 10 river flow measurements from the year 2000. One-hundred realizations of hydraulic and recharge parameters were generated for simulation with MODFLOW and MODPATH to obtain particle starting locations. Particle starting locations were accumulated and mapped to determine the contributing area to the public-supply well. The second method employed a program that generates random fractures based on statistical information such as length, orientation, diameter, and density of the karst features. A local-scale model with small cell-size (25 x 25 m) was created around the public supply well. Hydraulic conductivity where the random fractures were located was increased by two orders of magnitude compared to the local hydraulic conductivity. MODFLOW and MODPATH were run 100 times, and starting particles were accumulated and mapped to identify probable contributing areas to the public-supply well. There are many sources of uncertainty in modeling, whether incorporating fractures, estimating parameters, setting hydraulic conductivity or recharge values, or selecting boundary conditions. These two methods address different aspects of uncertainty. The first method addresses the inherent uncertainty in hydraulic and recharge values; the second method focuses the uncertainty on physical features of the aquifer that are highly likely to affect the transport of contaminants to a public supply well. Both of these sources of uncertainty will affect the size and shape of the contributing area. However, the incorporation of karst features may improve estimates of source areas and therefore potential contaminants that could reach the public supply well.