82 Interpretation of Water Chemistry and Stable Isotope Data from a Karst Aquifer According to Flow Regimes Identified through Hydrograph Recession Analysis By D. H. Doctor 1 and E. C. Alexander, Jr. 2 1 U.S. Geological Survey, 345 Middlefield Rd., MS 434, Menlo Park, CA, 94025 2 Univ. of Minnesota, Dept. of Geology and Geophysics, 310 Pillsbury Dr. SE, Minneapolis, MN, 55414 ABSTRACT In this study the relation between flow regime and chemistry of a major karst groundwater resurgence zone in southwestern Slovenia was examined using spring hydrograph recession analysis. Long-term (>2 weeks) recession periods were isolated from 6 years of flow data. Breaks in slope on a plot of the natural log of the discharge versus time allowed for the identification of four separate flow regimes of the aquifer outflow. Major ion chemistry and stable isotopic composition (δ 18 O of water and δ 13 C of DIC) of samples collected twice monthly for two years were then grouped according to where they had been collected within each identified flow regime. Patterns in the chemical and isotopic data emerged which indicated shifting sources of water contributing to the outflow of the spring under different hydrologic conditions. This type of analysis may be a valuable water resource management tool in other karst regions. INTRODUCTION A primary challenge for the management of karst water resources is to characterize water quality changes with discharge variability. In order to accomplish this goal, managers must be able to effi- ciently assess two aspects of the karst aquifer system that interact and determine overall water quality: the hydrologic and the hydrochemical variability. Often, however, resources for characterizing water quality across the full range of hydrologic variability are limited, resulting in a frequency of water sampling that is far lower than the actual time scale of chemi- cal changes taking place at the point of measure- ment. Therefore, a need exists for a method through which relatively infrequent water quality data can be used to accurately understand and possibly predict major changes in water quality as the hydrologic conditions change. In this paper, we describe a technique in which long-term records of discharge and relatively infre- quent water quality sampling can be combined for the purpose of studying water quality changes with flow. The steps are not mathematically complex, allowing for straightforward and rapid culling of information from data which already exists for many springs. The analysis begins with examination of the recession limbs of a long-term (several years) record of discharge. First suggested by Maillet (1905), sev- eral authors have since proposed that the recession limb of a karst spring hydrograph can be approxi- mated by a function that is the sum of several expo- nential segments of the total recession (Forkasiewicz and Paloc, 1967; Hall, 1968, Mil- anovic ´ , 1981; Bonacci, 1993; Tallaksen, 1995). Thus, the entire discharge-time relationship of the recession is expressed as: Qt () q i 0 e α i ( ) – t i 1 = N ∑ = . (eq, 1) Where Q is the discharge at time t, N is the num- ber of exponential segments of the recession, q o i is the discharge at the beginning of each recession seg- ment, and α i is the recession coefficient for each seg- ment. In this model, each exponential segment is interpreted to represent the depletion of an aquifer reservoir, with the rate of depletion of that reservoir being represented by the recession coefficient (α i ). Accordingly, the segment with the greatest recession coefficient would represent the most rapid drainage of the karst network (presumably surface runoff or displacement of water into the largest conduits) and the recession segment with the smallest coefficient would represent the baseflow (i.e., the slow drainage of that portion of the aquifer with the lowest trans- missivity). The latter is often termed the diffuse flow