SEPARATION OF CONDUCTIVITY FREQUENCY DISTRIBUTIONS (CFD): A METHOD FOR CHARACTERIZING THE STRUCTURE AND HYDROLOGIC BEHAVIOR OF KARST SYSTEMS

Variations in the electrical conductivity of karst spring waters can give insight into karst flow systems. Use of the coefficient of variation of measured conductivities has been a simple and popular approach to inferring degree of karstification; however, problems with this approach arise because conductivity frequency distributions (CFDs) are usually multi-modal, and the method sometimes erroneously classifies karstic aquifers as non-karstic. Here we demonstrate that a more rigorous analysis of the CFD can give insight into different water types contributing to spring flow. The CFD for a given water year can be separated into an additive series of normal distributions, each related to a water type representing a more or less homogeneous hydrogeochemical population. For each water type, a coefficient of variation, an area coefficient (a measure of the contribution of each curve to the overall CFD area), and an amplitude can be determined. These parameters can be compared between water types for a given year, and can be also used to evaluate the evolution of a water type from one year to the next. We applied this method to four years of conductivity data collected at Barton Springs, Austin, Texas. Although the overall shape of the CFD changed from year to year, it could consistently be separated into five normally distributed populations. We suggest that each population represents a water type resulting from a particular mode of aquifer functioning (e.g., storm-derived waters, equilibrium flow, intrusion of saline waters brought on by low flow). Changes in the parameters describing the five curves reflect aquifer response to climatic variations. Similarly, when the CFD for a karst chalk aquifer in Normandy, France, was analyzed for two water years, it could be separated into six curves, each one related, as for Barton Springs, to a specific hydrogeochemical process. The consistent number of curves and the evolution of the parameters describing them suggest that this approach can be very useful in gaining insight into the functioning of karst aquifers.