Structural uncertainty and surface water–groundwater interactions in the probabilistic delineation of well capture zones

In the central Venetian plain (Northeastern Italy), drinking water demand is mainly met by groundwater abstraction from the complex underlying aquifer system. From the foothills of the Prealps towards the coast, this system consists of a thick unconfined aquifer made up of coarse sediments – mainly of fluvial or fluvioglacial origin – transitioning into a multi-layered aquifer system with progressively thicker clay layers. The deep confined aquifers in the latter region are heavily exploited, as they represent a valuable source of high-quality drinking water. In this context, proper delineation of well capture zones and wellhead protection areas (WHPAs) is critical to ensure drinking water quality. However, for wells exploiting the deep confined aquifers, especially in such a complex geological context, relying on the simple geometric or analytical criteria seems inadequate. Moreover, due to the high level of uncertainty involved, a deterministic definition of the spatial continuity, extent, and connectivity of structures with different permeabilities could lead to misleading results. In this work, we adopted a stochastic approach that allows for geological realism and for uncertainty quantification. Using an extensive dataset of borehole stratigraphic information, we set up a geostatistical model for the hierarchical simulation of lithofacies and validated it by means of K-fold cross-validation. Through subsequent groundwater flow modeling of multiple equiprobable realizations, we assessed the impact of structural uncertainty on the groundwater dynamics. Then, through the application of backward particle tracking techniques we analyzed the uncertainty in the preliminary delineation of WHPAs for deep wells. Furthermore, this study presents one of the first real-world applications of particle tracking that integrates the displacement of particles along surface watercourses. This latter method allows us to account for the high dependence of the groundwater system under investigation on the dynamics of the Piave River, and sheds light on the relevance of surface water–groundwater interactions in the problem of capture zones identification.