Identification of riverbank filtration sites at watershed scale: A geochemical and isotopic framework

Identifying groundwater wells performing riverbank filtration (RBF) is crucial to ensure safe drinking water through vulnerability assessment plans adapted to these hybrid water sources. Nonetheless, RBF is often unintentional or insufficiently documented and official inventories are scarce. We developed a user-friendly geochemical and isotopic framework for the in-situ identification of RBF facilities. It includes an interpretation abacus for non-specialists. While most studies using tracers are site-specific and/or based on discrete samples, we propose a novel multi-site characterization where time-series of EC, δ²H and δ¹⁸O are directly used as proxies of surface water infiltration at the watershed-scale. The basic statement is that time varying signal of raw water from a groundwater pumping facility reveals a significant induced infiltration of surface water. The framework was applied on nearly 2000 samples from 40 pumping wells and 4 neighboring rivers (<500 m), collected through collaborative sampling on a weekly to monthly basis for 18 months. Despite proximity to surface water, two-third of the complete dataset (19 facilities) were revealed not to benefit from significant contribution of surface water, demonstrating location criteria to be insufficient to identify RBF sites. Permanent RBF was evidenced at 5 facilities, where year-long seasonal variation of tracers in raw groundwater highlighted a continuous high proportion of infiltrated surface water. Unexpectedly, time-series also unveiled a third category: occasional RBF, where induced infiltration occurred only when specific hydrodynamic conditions were met (4 facilities). This study also provided concrete illustrations on how climate change may impact the efficiency of RBF to naturally attenuate microbiological contaminants and how geochemical and isotopic time-series considerably help at anticipating the evolution of contaminant attenuation capacity of RBF sites. Finally, by highlighting the existence of occasional RBF, this study tackles the common oversimplification that groundwater facilities can be binarily and classified either as RBF or groundwater.