THE FLUX OF WATER AND BOTH NON-POINT SOURCE AND POINT SOURCE CONTAMINANTS DURING STORM EVENTS IN FREE-FLOW KARST AQUIFERS

Storm hydrographs from observation wells in the Mammoth Cave-Bowling Green, Kentucky area which were drilled into trunk cave streams are compared with those of nearby farmers' water wells. Hydrographs for observation wells into large cave streams show that during major flood-producing rains the stage can rise over 30 m (100 ft) while water wells located 183 m (600 ft) distant indicate that the hydrograph crest lags several hours behind that of the main conduit and only rises approximately 2/3 as high in elevation. This evidence supports the hypothesis that storm water from large cave streams flows back into karst aquifers and is analogous to river bank storage in alluvial aquifers.

During heavy rains, surface runoff flowing rapidly into cave streams from sinking streams and open-throat sinkholes carries more suspended sediment and other non-point source contaminants than water percolating through the soil. The flow of storm water from cave streams back into the aquifer appears to be one reason why some farmers' water wells often become muddy during and after heavy rains. Also, an increase in non-point source contaminants occurs in water wells similar to that of karst springs during and after heavy rains.

Water samples collected from cave streams and springs during and after storm events also shows an increase in point source contaminants. Rather than a decrease in point source contaminants due to dilution, data show an actual increase in concentration of several different contaminants from a known contaminated site during and after storm events compared to their concentrations during base flow. This clearly indicates the need for storm event monitoring in free-flow karst aquifers. Finally, the paper discusses an investigation to determine the best time for sampling at a spring during the rise and fall of a storm hydrograph for point source contaminants. This study involved tagging groundwater flow from a contaminated site to a spring by injecting three different fluorescent dyes. One dye was injected into a sinking stream, a second dye into a sinkhole and a third dye into the soil at the beginning of several heavy rains. Dye breakthrough curves were then monitored at the contaminated spring. Recommendations concerning when and how often to collect samples at springs during the rise and fall of storm hydrographs are presented.