Localized subsurface water flux estimation by simplified hydraulic tomography: Synthetic test case and outlook for field application

Floodplain aquifers receive substantial water input from adjacent hillslopes, most likely concentrated in hillslope hollows. Other localized inputs are expected from tributary valleys and along surface-water bodies. However, quantifying and localizing subsurface water fluxes is inherently difficult. A key requirement for flux estimation is the knowledge of the hydraulic-conductivity distribution at the scale of interest. Conventional hydrogeological investigation techniques, such as pumping and slug tests, may fail in the presence of heterogeneity and complex structural boundaries. While advanced 2-D and 3-D hydraulic tomography may resolve small-scale heterogeneity, it is typically limited to small spatial scales and require complex and costly field installations. We choose a simplified tomographic approach using a limited number of pumping and observation wells targeting spatial ranges in the order of 100 m. To infer the spatially variable hydraulic-conductivity field with its uncertainty, we apply an iterative ensemble smoother and pilot-point parameterization for dimensionality reduction. Subsequently the resulting conditional realizations of the hydraulic-conductivity field are used to calculate water fluxes applying mean hydraulic gradients. We test the approach by a synthetic scenario mimicking the conditions in a hillslope hollow connected to a floodplain aquifer. We aim at applying the approach as field method in a local floodplain aquifer in Southwest Germany near the city of Tübingen to quantify the lateral inflow from an adjacent hillslope hollow.