Quantification of temporal distribution of recharge in karst systems from spring hydrographs

Abstract

The estimation of the temporal distribution of recharge in karst aquifers is a challenge due to large heterogeneities in geometric and hydraulic parameters of the vadose and phreatic zone. This article provides a time-continuous approach for the estimation of inflow into the conduit system of a karst aquifer which consists of the sum of direct recharge and flow from the fissured matrix blocks into the conduit system. The approach employs the first time derivative of the spring hydrograph and the recession coefficient of the conduit system for the determination of this function. The first time derivative of the hydrograph describes the rate of change in spring discharge. It reflects the ratio of inflow to and outflow from the conduit system. The recession coefficient depends on the hydraulic diffusivity, which controls the velocity of the signal transmission through the system. As shown in parameter studies with a simplified serial two-reservoir model, direct recharge into the conduit system clearly dominates the early hydrograph response during recharge events even if the fraction of direct recharge represents just a few percent of total recharge. This behaviour is caused by a large contrast in recession coefficients between conduit system and fissured matrix blocks. The direct recharge component can therefore be separated from the estimated total inflow into the conduit system. Estimation of inflow into the conduit system of the Gallusquelle spring (Swabian Alb, Germany) after a storm event yields similar results as those obtained from parameter studies. The separation of the direct recharge component is in agreement with information from an independent isotope study. The methodology has been further applied to a recharge event initiated by snowmelt. As a result of daily variations of the air temperature, a clearly diurnal cycle of inflow into the conduit system is estimated that is not easily recognizable from the spring hydrograph. The applicability of the proposed methodology requires knowledge about the magnitude of the recession coefficient of the conduit system. The characteristics of the conduit system become also apparent in the rapid transport of tracers. The recession coefficient of the conduit system can, for example, directly be estimated from the reciprocal of the mean tracer travel time. However, for this type of analysis, only tracer experiments, covering the extent of the catchment should be taken into account. For the Gallusquelle catchment the recession coefficient obtained from the tracer experiment corresponds to that obtained from hydrograph recession analysis. The estimated inflow into the conduit system and interpretation of temporal distribution of recharge is therefore based on a plausible and physically based parameter.