Impacts of mining on regional groundwater systems are commonly assessed using numerical groundwater flow models which require an understanding of the strata matrix hydraulic conductivities. A rapid, reasonably accurate and relatively low cost option for assessing the conductivity is laboratory testing of rock core obtained from exploration boreholes. Core measured hydraulic conductivities can also be correlated to wireline logs that are routinely run in exploration boreholes. Simple bi-variate correlation can provide a means of assessing the uniformity or otherwise, of the hydraulic conductivity distribution but multi-variate correlations using genetic algorithms can offer an alternative and possibly superior methodology for predicting hydraulic conductivities. Core tests undertaken at Ulan Coal Mine as part of a groundwater management program have been correlated to natural gamma, density and neutron logs using a genetic algorithm employing a backward propagation neural network. The procedure has been used to generate continuous vertical profiles of hydraulic conductivities in numerous exploration holes where coring was not undertaken.