ESTIMATION OF GROUNDWATER MEAN RESIDENCE TIME IN UNCONFINED KARST AQUIFERS USING RECESSION CURVES ALIREZA KAVOUSI AND EZZAT RAEISI Dept. of Earth Sciences, College of Sciences, Shiraz University, 71454 Shiraz, Iran, kavousi@shirazu.ac.ir, e_raeisi@yahoo.com Abstract: A new parsimonious method is proposed to estimate the mean residence time of groundwater emerging at any specific time during recession periods from karst springs. The method is applicable to unconfined karstic aquifers with no-flow boundaries. The only required data are numerous consecutive spring hydrographs involving a wide range of discharge from high to low flow and the relevant precipitation hyetographs. First, a master recession curve is constructed using the matching-strip method. Then, discharge components corresponding to the individual hydrographs at any desired time are estimated by extrapolation of recession curves based on the master curve. Residence times are also taken from the time elapsed since the events’ centroids. Finally, the mean residence time is calculated by a discharge-weighted average. The proposed method was evaluated for the Sheshpeer Spring in Iran. There are 259 sinkholes in the catchment area of the Sheshpeer unconfined aquifer, and all the boundaries are physically no-flow. The mean residence time calculated by the proposed method was about one year longer than that of uranine dye tracer. The tracer mean time is representative of flowing water between the injection and emergence points, but the mean time by the proposed method is representative of all active circulating water throughout the entire aquifer. The proposed method is more appropriate, in practice, especially in groundwater recharge and contaminant vulnerability assessments, than isotopic methods, in cases where a small portion of water with exceptionally high residence time increases the mean. Moreover, isotopic methods are more expensive and not capable of directly determining the residence-time distribution. INTRODUCTION About 20 percent of the Earth’s land surface is covered by carbonate-karst formations (White, 1988), and roughly 20 to 25 percent of the global population depends largely or entirely on karst groundwater (Ford and Williams, 2007). Carbonate karst comprises about 11 percent (,185,000 km 2 ) of Iran’s land surface (Raeisi and Kowsar, 1997) and is one of the major fresh-water resources of the country. Groundwater residence time is defined as the elapsed time for a water molecule to travel from the recharge area to the discharge zone of the aquifer (Kazemi et al., 2006; Bethke and Johnson, 2008). Water emerging from a karst spring is composed of billions of water molecules from different precipitation events. Karst aquifer recharge water flows rapidly through conduits and fractures and very slowly through the small pores and fissures where most of the recharged water may be stored for a long period, which is especially the case in Iran. Consequently, water with a wide range of residence times may be observed in the spring water. The residence-time distribution represents the distri- bution of different times, which is calculated by mathemat- ical age-transport modeling (Kazemi et al., 2006). According to the age-mass concept (Goode, 1996), groundwater mean residence time (MRT) is the mass-weighted average residence time of all water emerging at a specific time. Knowledge of groundwater residence time is an invaluable tool for contaminant vulnerability assessment and evalua- tion of aquifer recharge rate during karst-water exploitation studies. The MRTs of groundwater in karst aquifers have been estimated routinely by environmental and artificial tracers. Radioisotopes are the commonly used environmental tracer for this purpose. Considering the amount of a radioisotope in precipitation as input and in spring water as output, mathematical lumped-parameter (e.g. piston flow, dispersion, etc.) models have been used to estimate the MRT (e.g. Maloszewski, 1994). Fluorescent dyes are the most practical and widely used artificial tracers in karst aquifers (Benischke et al., 2007). The mean transit time of dye can be approximated as the time difference between the injection and the centroid of the tracer-breakthrough curve (Benischke et al., 2007). For a sink-to-spring tracer test, the mean transit time represents the MRT for the part of the spring water that flows from the injection point during the conditions of the test. Worthington (2007) reported that MRTs determined from environmental tracers are typically one hundred times longer than the mean transit times of artificial tracers in A. Kavousi and E. Raeisi – Estimation of groundwater mean residence time in unconfined karst aquifers using recession curves. Journal of Cave and Karst Studies, v. 77, no. 2, p. 108–119. DOI: 10.4311/2014ES0106 108 N Journal of Cave and Karst Studies, August 2015