Groundwater recharge estimates in agriculturally managed site in Northeast Germany: combining Cosmic ray neutron sensing and soil hydrological modelling

Brandenburg is among the driest regions in Germany, and heavily relies on groundwater resources for both agricultural and drinking water supply. Already suffering from declining groundwater tables, climate change is expected to exacerbate the situation. For a sustainable management of groundwater resources, the rate of groundwater recharge (GWR) is a key variable. Yet, its quantification remains a challenge, as it cannot be measured directly at the field scale One way to estimate GWR is using vadose zone models to simulate the local water balance and the vertical percolation of water towards the groundwater. Observations of soil moisture (SM) in the root zone can provide a means to calibrate such models so that they can adequately represent the local water balance. However, conventional point-scale SM observations notoriously suffer from a lack of horizontal and vertical representativeness, compromising the validity of the calibration. In this study, we explore the potential of cosmic-ray neutron sensors (CRNS) to address this issue. CRNS allow for non-invasive SM monitoring of the shallow root zone at the hectare-scale. We use daily CRNS-based soil moisture estimates to calibrate the vadose zone model HYDRUS-1D, and hence to derive daily estimates of the downward water fluxes below the root-zone, as an approximation of GWR. For this purpose, we explore a unique dataset that was obtained in a research site near Potsdam, Brandenburg, over a period of more than three years. The site features a diversity of agricultural plots, and sits on a gentle hillslope over a glacial till aquifer, with the groundwater table at depths between 1 to 10 m. In an area of around 10 ha, we operated eight CRNS sensors and 27 SM profile probes, complemented by measurements of soil texture and soil hydraulic properties, among others. In various simulation experiments, we evaluate the added value of using CRNS-based soil moisture estimates for model calibration, as a replacement or as a supplement of conventional profile probes. Based on a calibrated model, we also assess long-term (centennial) changes of GWR.