Groundwater-Surface water exchanges in an alluvial plain subjected to pumping: a coupled multitracer and modeling approach

Abstract. Alluvial aquifers represent a vital water resource for many regions. However, understanding and characterizing the interactions between rivers and these aquifers is a major challenge for researchers and water managers. This characterization, in terms of flow velocity or water supply, is important to identify the vulnerability of the aquifers. In this study, our goal is to improve the understanding of interactions between rivers and alluvial aquifers by combining a multi-tracer approach with numerical modeling. By integrating these two complementary methods, we aim to accurately quantify the exchanges between groundwater and surface water, and to identify the water sources contributing to aquifer recharge. This combined approach allows a better quantification of river-aquifer interactions at local scale, in the context of groundwater exploitation by pumping along the river. A large drinking water catchment field located on the banks of the Rhône River, in the southeast of France, was chosen as the study site. This site consists of several pumping wells and observation piezometers parallel to the Rhône. As often, with alluvial aquifer exploitation close to a river, the pumping leads to flow from surface water to groundwater. Groundwater temperature, piezometric levels and river surface water levels were continuously recorded for an 18-month period. During field campaigns, conductivity, stable isotopes of water and radon activity of groundwater and surface water were measured. Radon was applied in a new way to measure the water flow from the river to the aquifer, which reduces the natural radon signal in the aquifer by radon-poor waters from a river. On the site, the radon data clearly delineates groundwater recharge from the river within 50 meters from the banks. The methodology to interpret periodic groundwater temperature signals was extended to the isotopic signal, making it possible to identify the dispersivity in addition to Darcy's velocity. All the experimental data were accounted for in a synthetic MODFLOW model, taking into account the boundary effect of the Ouvèze and Rhone rivers. Model calibration was made using the piezometric records and the PEST package. Reactive transport of radon was implemented using MT3DMS to ascertain the overall water balance of the study site. In addition to the Rhône as a supply, we have shown that the other river (Ouvèze) also contributes of the site’s supply (around 55%). By improving our understanding of the interactions between rivers and alluvial aquifers, this study offers valuable insights for sustainable water resource management in regions similar to our study site, i.e., the case of a losing river recharging an exploited aquifer and demonstrates the value of natural tracers, such as radon or stable water isotopes, in situations where the application of artificial tracers is impractical. Our results can guide policy decisions regarding groundwater development and river ecosystem protection.