The U.S. Geological Survey, in cooperation with the San Antonio Water System, conducted a 4-year study during 2001– 04 to identify major ground-water flow paths in the Edwards aquifer in northern Medina and northeastern Uvalde Counties, Texas. The study involved use of geologic structure, surfacewater and ground-water data, and geochemistry to identify ground-water flow paths. Relay ramps and associated faulting in northern Medina County appear to channel ground-water flow along four distinct flow paths that move water toward the southwest.

The northwestern Medina flow path is bounded on the north by the Woodard Cave fault and on the south by the Parkers Creek fault. Water moves downdip toward the southwest until the flow encounters a cross fault along Seco Creek. This barrier to flow might force part or most of the flow to the south. Departure hydrographs for two wells and discharge departure for a streamflow-gaging station provide evidence for flow in the northwestern Medina flow path. The north-central Medina flow path (northern part) is bounded by the Parkers Creek fault on the north and the Medina Lake fault on the south.

The adjacent north-central Medina flow path (southern part) is bounded on the north by the Medina Lake fault and on the south by the Diversion Lake fault. The north-central Medina flow path is separated into a northern and southern part because of water-level differences. Ground water in both parts of the northcentral Medina flow path moves downgradient (and down relay ramp) from eastern Medina County toward the southwest. The north-central Medina flow path is hypothesized to turn south in the vicinity of Seco Creek as it begins to be influenced by structural features. Departure hydrographs for four wells and Medina Lake and discharge departure for a streamflow-gaging station provide evidence for flow in the north-central Medina flow path.

The south-central Medina flow path is bounded on the north by the Seco Creek and Diversion Lake faults and on the south by the Haby Crossing fault. Because of bounding faults oriented northeast-southwest and adjacent flow paths directed south by other geologic structures, the south-central Medina flow path follows the configuration of the adjacent flow paths—oriented initially southwest and then south. Immediately after turning south, the south-central Medina flow path turns sharply east. Departure hydrographs for four wells and discharge departure for a streamflow-gaging station provide evidence for flow in the south-central Medina flow path. Statistical correlations between water-level departures for 11 continuously monitored wells provide additional evidence for the hypothesized flow paths.

Of the 55 combinations of departure dataset pairs, the stronger correlations (those greater than .6) are all among wells in the same flow path, with one exception. Simulations of compositional differences in water chemistry along a hypothesized flow path in the Edwards aquifer and between ground-water and surface-water systems near Medina Lake were developed using the geochemical model PHREEQC. Ground-water chemistry for samples from five wells in the Edwards aquifer in the northwestern Medina flow path were used to evaluate the evolution of ground-water chemistry in the northwestern Medina flow path. Seven simulations were done for samples from pairs of these wells collected during 2001–03; three of the seven yielded plausible models.

Ground-water samples from 13 wells were used to evaluate the evolution of ground-water chemistry in the north-central Medina flow path (northern and southern parts). Five of the wells in the most upgradient part of the flow path were completed in the Trinity aquifer; the remaining eight were completed in the Edwards aquifer. Nineteen simulations were done for samples from well pairs collected during 1995–2003; eight of the 19 yielded plausible models. Ground-water samples from seven wells were used to evaluate the evolution of ground-water chemistry in the south-central Medina flow path. One well was the Trinity aquifer end-member well upgradient from all flow paths, and another was a Trinity aquifer well in the most upgradient part of the flow path; all other wells were completed in the Edwards aquifer. Nine simulations were done for samples from well pairs collected during 1996–2003; seven of the nine yielded plausible models. The plausible models demonstrate that the four hypothesized flow paths can be partially supported geochemically.