What is the contribution of localized recharge to spring flows in a binary karst aquifer? A response using a multi-scale physico-chemical approach

The objective of this work is to investigate infiltration processes, flows on both saturated and unsaturated zones, and quantify their contributions to spring discharge, in a binary karst aquifer (recharged by the karst massif, as well as sinking streams). The study site is the Verneau karst aquifer located in the Jura Mountains (France). The recharge area covers 15 km², with half consisting of marl outcrops, where water enters through five losses (localized infiltration), and the other half consisting of limestone massif covered by a soil layer (diffuse infiltration).First, to analyze the spatial variability of flows generated by diffuse and localized infiltration at various depth, we performed hydrochemical analyses in soil lysimeters, caves and at the spring, conducting to characterize the physico-chemical end-members of the various compartments: soil layer, unsaturated zone in the karst massif and in the conduit network, saturated zone. Second, we used continuous high-frequency (1hour) time series (3 years) of semi-conservative tracers (electrical conductivity and nitrate concentrations) to characterize discharge response throughout the seasonal cycle. A End-Member Mixing Analysis (EMMA) was conducted on 40 flood events to determine the contribution of the infiltrations types to spring discharge. Our results show that the localized infiltration shows a relatively homogeneous spatial signal, characterized by low values of electrical conductivity and nitrate content. Diffuse infiltration is spatially variable due to anthropogenic activities and contrasted residence time within the massif. Results of the EMMA method reveal that during flood events, approximately 1/3 of spring discharge comes from localized infiltration, while the majority comes from diffuse infiltration and pre-event water stored in the massif. A seasonal variability is evidenced in link with lower stream losses and storage in the unsaturated zone. A hydrogeological conceptual model is finally proposed, allowing us to discuss the origin of spring waters, given new elements on drivers controlling infiltration modalities, and the role of transfer and storage in the unsaturated zone.