MODELING RESIDENCE TIME DISTRIBUTION OF A KARST AQUIFER USING A PHYSICALLY BASED APPROACH

Karst systems can be conceptualized by five compartments: (i) a network of highly conductive flow conduits, (ii) a saturated porous fissured system that is usually overlain by (iii) a thick unsaturated zone, (iv) an epikarst layer that redistributes recharge and stores water near the surface, and (v) a thin top soil layer. Sampling water of karst springs allows isotopic analyses in an integrative way because karst springs are usually connected to conduits that drain the entire aquifer. Water flow through conduits and exchange between conduits and the porous fissured system lead to mixing of water, making determination of residence times a tedious task because residence times are different in each compartment.

The objective of this paper is to determine residence times and their distribution in a karst aquifer using observed tritium concentrations at the karst spring “Gallusquelle”, located on the Swabian Albs, South Germany. The spring comprises a catchment area of 45 km2 and the thicknesses of the saturated (compartment ii) and unsaturated (iii and iv) zones are about 30 m and 100 m, respectively. The control volume finite element FRAC3DVS/HydroGeoSphere model (Therrien and Sudicky, 1996, JCH) is used to simulate recharge, saturated/unsaturated water flow, and discharge at the Gallusquelle for a period of 50 years. Available monthly precipitation data are transformed using a water balance approach that accounts for evaporation, interception and snow storage (Sauter, 1992, TGA). We also simulate transport of tritium found in young rain water and evaluate isotope signals at the spring. This paper shows how to use a physical-based approach to interpret residence times in karst systems based on isotopic data.