Biological and chemical analysis of water samples, collected from surface and subsurface streams was preformed to determine the quantity and biologic sources of enteric contamination in the Double Springs Groundwater Basin, an urban karstified carbonate aquifer under Bowling Green, Kentucky. Major flow paths, drainage basin boundaries and geologic/hydrologic relationships were identified for the basin. Analysis of biologic contaminates using fecal coliform /fecal streptococcus counts and ratios were conducted for baseflow and storm events and related to the Double Springs hydrographs. Analysis was also conducted to determine the source of sulfides responsible for the growth of sulfur fixing bacteria in the Overflow Spring resurgence. Research findings revealed that enteric contamination concentrations were below maximum acceptable levels during baseflow condition throughout the basin. During high discharge events, bacteria levels increased and exceeded acceptable levels by 12 to 30 times. Waste was determined to be animal or human in origin depending on storm event, intensity of rain, time of year and time between events. Bacteria counts from rural sources were insignificant and were exceeded by urban bacteria counts by 1000 times. It was determined that an urban environment can have a large impact on biologic water quality in karstified aquifers. The source of sulfides in the basin was determined to be caused by migration of high sulfate water from a lower confined aquifer.

Due to recent land use changes, runoff has greatly increased In Bowling Green, necessitating the drilling of nearly 600 drainage wells to direct storm water into the Lost River Karst Aquifer. Research was conducted to (1) differentiate areas of the city as optimal for removal of storm water runoff, (2) correlate lithologic features between wells, and (3) determine optimal depths to which new drainage wells should be drilled. This was accomplished using maximum capacity tests, correlation of lithologic features between wells and comparing void density to depth below the land surface. It was only possible to develop general trends in the areal distribution of voids throughout the city, hence: specific optimal areas for removal of storm water runoff could not be differentiated. The correlation of voids between wells was possible alone both the regional dip and strike of the beds. It was also determined that the greatest potential for intersecting voids was in the first 6 meters of borehole, but no clear cut depth criteria were developed, although recommendations were made pertaining to drilling new wells to an optimal, functional depth.

Southwest of Bowling Green, Kentucky, is the Western Pennyroyal, is a karstified area which has been neglected in the study of its ground water. About 100 square miles near Adairville were chosen for analysis to help rectify this lack of knowledge. The area is dominated by Sinking Creek, a surface-subsurface drainage system. The main purpose of the study was to map the underground flow in the Sinking Creek area. The drainage system seemed to be typical for the Western Pennyroyal and exhibited many similarities to flow paths known in the Central Kentucky Karst. The investigation proved to be of basic value to future studies dealing with water well location, pollution of subsurface streams and karst-related flooding problems. The geographic scope of the study area extends from the rise of Sinking Creek to the origin of its headwaters in the residual outlying knobs of the Dripping Springs Escarpment to the north. Most of the area is a sinkhole plain developed mainly on the Ste. Genevieve and St. Louis limestones of Mississippian age. Map and field reconnaissance of the study area revealed the presence of 12 sinking streams, 6 resurgences and 5 caves, all occurring at or below 600 feet in elevation. On the basis of the physical features mentioned, two hypotheses were devised to explain their development and their relationship to the hydrology of the area. The first hypothesis was that the sinking stream and resurgences in the study area are connected in a single drainage system. This hypothesis was testable by physically determining subsurface connections in the field. The second hypothesis was that the sinking streams are controlled by the stratigraphy of the lithologic units. Diversion of surface streams occurs at or near the 600 foot elevation level, upon flowing from the Ste. Genevieve to the St. Louis limestone. This hypothesis was field tested by standard geologic methods. The determination of surface-subsurface stream connections was carried out by standard water tracing techniques using Rhodamine W.T. dye (20 percent solution) and fluorescein dye. The dye was injected into the streams, and samples were collected and then analyzed in the laboratory with a Turner fluorometer. Positive dye connections were obtained in all 5 traces. In addition to dye tracing, a large amount of field reconnaissance and subsurface mapping was necessary to determine the nature of the geologic controls on the surface-subsurface drainage system. Numerous rock outcrops and 5 caves were explored; one cave was mapped for over 4,000 feet. The lithologic studies proved to be inconclusive for the entire study area, but enough evidence was collected to support lithologic stream control in one cave stream segment. The culmination of the research design was to present the results of dye tracing and field work with a written description, graphs, and especially a map showing the connections established by the research.

A field study of the lower reaches of the Long Creek drainage area in southeastern Allen County, Kentucky, established the karst character of that area. The area heretofore had been described as a non-karst area. Four swallow holes, which individually or collectively totally pirate Long Creek, were identified. Two major resurgences of the pirated flow were located and described. A detailed study of a portion of the Long Creek drainage area revealed thirty-four springs, all of which were pirated at least once, and no flow from these springs reached Long Creek by surficial routes. The field survey also revealed dolines and a major cavern, Carpenter’s Cave. The geologic formation responsible for the karst features within the Long Creek drainage area is the Louisville limestone of Silurian age. These strata are characteristically karstic wherever exposed. The impermeability of the Chattanooga shale which overlies the Louisville limestone was established by the analysis of spring piracy and resurgence. The phreatic character of the Carpenter’s Cave and other karst features led to the conclusion that these karst features were developed prior to the deposition of the impermeable Chattanooga shale during Devonian time.

Chemical analysis of samples collected from subsurface streams and wells were analyzed to identify nonpoint pollutants in karst aquifers resulting from agricultural land use. Major flow paths and travel times taken by nonpoint pollutants were identified. A land use limitation map showing areas where agricultural activities may results in nonpoint pollution of the karst aquifer was prepared. A groundwater susceptibility map was prepared to show areas of the karst aquifer most susceptible to groundwater contamination by nonpoint pollution resulting from agricultural land use. Research findings revealed that pollutant concentrations may reach levels thousands of times greater during high discharge events than during baseflow conditions. Approximately thirty-five percent of the drainage basin was considered to have serve limitations for conventional agricultural land use. Fifty-eight percent had moderate limitations while only seven percent had slight limitations and were considered acceptable for conventional agricultural land use. Lineaments, fracture traces and sinkholes were plotted on a map to revel those areas to the karst aquifer that are susceptible to groundwater contamination as a result of agricultural activities. Recommendations for agricultural best management practices were made in terms of reducing the potential for contributing pollutants to the karst aquifer.

Bowling Green, Kentucky is located in a distinctive karst region, characterized by subsurface drainage. The Lost River is a large subsurface stream which flows beneath the city. It receives much of the stormwater runoff from Bowling Green, since most of the city's runoff is directed underground. Significant pollutants in Bowling Green's stormwater runoff were identified from water quality test results of storm event grab samples, and a composite sample, of runoff entering the urban By-Pass Cave. Water quality test results were also obtained from storm event grab samples, and a composite sample, of the Lost River at the Blue Hole before it reached Bowling Green and at the Resurgence after it had passed beneath Bowling Green. Significant pollutant test results from the Blue Hole and Resurgence were analyzed to determine the impact of urban stormwater runoff on the water quality of the Lost River. Suspended solids were the only urban runoff pollutant identified as entering the Lost River in significant quantities. Animal waste, iron, and oil and grease were identified as stormwater runoff pollutants of the Lost River, but whose origin, either urban or rural, could not be conclusively determined from the available data.

The laboratory-derived calcite dissolution rate law of Plummer et al. (1978) is the most widely used and mechanistically detailed expression currently available for predicting dissolution rates as a function of water chemistry. Such rate expressions are of great use in understanding timescales associated with limestone karst development. Little work has gone into the field testing of the rate law under natural conditions. This work compared measured dissolution rates measured by a crystal weight loss experiment in Buffalo Creek within Fort’s Funnel Cave, which lies within a pristine, forested catchment of Mammoth Cave National Park. Continuous water chemistry sampling over the same period allowed a time-integrated prediction of the dissolution based on the Plummer et al (1978) expression. Results indicate that the rate law overpredicted dissolution by a factor of about ten. This concurs with earlier laboratory work suggesting that the law tends to overpredict rates in solutions close to equilibrium with respect to calcite, as were the waters in this study. Estimating dissolution rates with the expression under varying hydrologic conditions also allowed a prediction of storm scales change in cave forming processes. Neglecting effects of sediment masking on the bed, it was found that 78% of the work done in the dissolution of the cave passage during the study period occurred at or around baseflow conditions, with only a small amount during the effective but infrequent high flow conditions.

Eighty-nine caves and sinkholes were investigated in the Edwards Aquifer recharge zone in Bexar County, Texas. The study examined their hydrogeologic and topographic origins and distribution, relationships to major fracture traces, quantity of recharge into the aquifer and degree of sensitivity towards degradation of the aquifer’s water quality. Groundwater traces were attempted to determine aquifer flow routes, time of groundwater travel, groundwater volume within conduits, and the aquifer’s capacity for dilution and dispersion of recharged contaminants. Trends in water quality were examined to quantify the volume and variety of contaminants recharged into the aquifer and to determine the effects of urbanization upon the Edwards Aquifer. The Edwards recharge zone was hydrogeologically assessed to rate the sensitivity of its areas. Socio-political impacts on recharge zone development were also examined. Based on the results of the above outlined research method, the conclusions of this investigation are that caves and sinkholes contribute substantial recharge into the Edwards Aquifer, rapidly transmit that recharge to the aquifer and are sensitive sites for potential contamination. The entire recharge zone was determined to be very sensitive to contamination. No significant differences were found between areas within the recharge zone to scale their degree of sensitivity. Major conduit flow networks were found to exist within the aquifer and their groundwater flow paths could be traced. Urban development of the Edwards recharge zone was shown to decrease the volume of recharge and degrade the aquifer’s water quality. No significant detrimental effects on the aquifer were observed. The volume of diminished recharge and the concentration of recharged contaminants that were necessary to produce significant adverse effects on the aquifer were not determined due to lack of precipitation during the study period and inconclusive groundwater tracings. It was recommended that further development of the recharge zone be suspended until the effects of urbanization are quantified.

Sinkhole flooding on the urban karst area of Bowling Green, Kentucky, was studied by investigating past flood events in the South Sunrise/Media Drive Sinkhole. The actual flood levels of four floods were compared to levels calculated by an empirical runoff determination method. Outflow capacity tests and dye tracing revealed the nature of drainage within the karst depression. The Soil Conservation Service runoff determination method and the three-hour maximum rainfall event were used to estimate the level of flooding. Three of the four flood events studied provided deviations between the actual and calculated flood levels of less than 0.26 feet (0.08 meters). The fourth event produced a deviation of 1.45 feet (0.44 meters) which was explained by the intense nature of the precipitation event. Analysis of the data resulted in the conclusion that surface storm water runoff exceeded the outflow capacity of the sinkhole drains during all four flood events and that the peak flood elevations were not related to the level of the water table in the area. Evidence also suggested that drainage from the base of the South Sunrise/Media Drive Sinkhole occurs through interconnecting voids associated with a cave roof breakdown pile, and not through an efficient system of solution conduits.

The stability of Mammoth Cave’s microclimate has never undergone extensive investigation. The Snowball Dining Room area was chosen to measure the microclimate of the cave and to determine if food preparation, human presence and surface temperature variations alter this microclimate. Three portable weather stations containing a temperature / humidity probe, datalogger and microbarograph were placed in various locations along three passageways leading away from the dining room. Readings were taken 24 hours per day for four months. Plotting temperature readings in graph form show a correlation between temperature of the passageway and distance from the dining room.