Trade‐Offs in Constraining Evapotranspiration Estimates in Complex Landscapes Using a Tracer‐Aided Ecohydrological Model

ABSTRACT

Evapotranspiration (ET) links vegetation and soils to the atmosphere and can be the dominant water flux in many areas. Whilst the estimation of ET can be constrained by different observations (e.g., discharge, stable water isotopes, remote sensing products etc.), their information content differs and their trade‐offs in model calibration are usually unknown. This is particularly the case in ET‐dominated, intensively managed catchments where ET strongly affects water supply and ecosystem services. Here, we investigated the heavily‐managed Middle Spree catchment (MSC, ~2800 km ² ) in NE Germany, a sub‐basin of the River Spree that includes the large Spreewald wetland (~472 km ² ) and serves as a key water resource for Berlin. The catchment is strongly impacted by groundwater abstraction and drainage. We applied multiple calibration schemes that combined different observations in a tracer‐aided, spatially distributed catchment model (STARR) over a 20‐year period, to investigate linkages between ecohydrological functioning and ET processes. Importantly, trade‐offs among performance metrics were subsequently analysed in the solution space of multi‐variable calibrations using Shapley additive explanation (SHAP) analysis. We found that discharge and temporal ET patterns can constrain water balance volumes, though their influence was modulated by ET‐dominated wetlands and human management. In contrast, incorporating spatial ET patterns and isotopic observations led to more plausible spatial representations of hydrological processes (i.e., ET and subsurface storage). Isotopic data uniquely captured relative contributions of evaporation and transpiration to total ET (ET partitioning). However, the constraining power of these observations was limited by trade‐offs, which were reflected in contrasting hydrological processes. The wetlands and human management could also influence these constraints. Although multiple calibration schemes revealed competing hypotheses of local hydrological functioning, they collectively provided richer and more integrated insights into how human management influences ecohydrological functioning in complex ET‐dominated catchments.