Modeling the usefulness of spatial correlation analysis on karst systems

Cross-correlation analyses on field data collected in karst aquifer systems can be used to develop a conceptual understanding of the aquifer. This includes the use of many data sets from the same aquifer to develop an understanding of how properties vary spatially. We focus on a method for characterizing the distribution of recharge, which is becoming increasingly important in regions where urban development encroaches on these important sources of water. Spatially varying precipitation data and cross-correlation analysis provide a means of spatially characterizing recharge locations on a karst aquifer. Our work expands on the numerical experiments conducted by Padilla and Pulido-Bosch (1995) using the numerical ground water model MODFLOW to introduce spatially varying parameters. The numerical experiments include conduit-controlled, matrix-controlled, and mixed karst systems with more than one precipitation time series input. The results show that spatially varying parameters can be inferred based on the cross-correlation of precipitation data and spring discharge. Simulations were completed using aquifer parameters derived from studies of the Barton Springs segment of the Edwards Aquifer. The simulations indicate that spatial variability within an aquifer can be inferred using cross-correlation analysis. A field study using these methods is summarized for Barton Springs near Austin, Texas.