Abstract
Lumped parameter modeling approach has been widely applied in karst hydrology for, among other applications, groundwater availability assessment in a context of global change. Nonetheless, such approach generally does not account for land-cover land-use (LCLU) and its impacts on recharge processes. Then, considering a semi-distributed recharge constitutes a relevant approach to capture the impacts of LCLU on flow dynamics but also introduce more complexity in the modeling approach. The present study consists of a multiple hydrological model calibration to assess the implication of considering a semi-distributed recharge in a lumped parameter model and focuses on the uncertainty originating by the model structure. The modeling results are discussed to evaluate the trade-off between hydrological model complexity and hydrological model performance. The study focuses on forest dominated karst areas with three karst catchments: Kerschbaum (Lower Austria), Baget (French Pyrenees) and Oeillal (southern France). Considering a semi-distributed recharge gives better performance for both Kerschbaum and Oeillal catchments hydrological models, while no improvement is obtained for Baget catchment. Systematic consideration of LCLU is thus not necessarily worthwhile in karst environmen ts as it can bring contradictory results in terms of hydrological model performance.