Monitoring Changes in the Landscape Water Balance: A Comparative Analysis of Satellite-Based Evapotranspiration Data in the Northern German Lowlands

The resilience of regional hydrology in human-influenced landscapes is a key challenge in the context of climate change. Lusatia in East Germany is an example of a region facing complex challenges in water management due to massive open pit mining activities as well as being subject to increasing water climate-induced scarcity. This study presents a comprehensive validation and comparative analysis of multi-temporal satellite-based evapotranspiration evapotranspiration (ET) data at multiple spatial resolutions including the 2000m Central Europe Refined Analysis (CERv2) – a product derived from the Weather Research and Forecasting (WRF) model forced by ERA5 reanalysis – alongside the 500m Moderate Resolution Imaging Spectroradiometer (MODIS) global product and 30m Landsat based ET, using lysimeter and eddy covariance measurements as ground-based references. This approach aims to assess the accuracy and practical utility of these data products for informing regional water management strategies. For the first time, a long-term analysis of landscape water balance changes and resilience is conducted, focusing on evapotranspiration as a central parameter for assessing the spatial and temporal variability of water dynamics. To compare the time series data, metrics such as Mean Absolute Error (MAE) were used to evaluate the agreement between satellite-based datasets and reference measurements. Our results reveal differences in the absolute values of evapotranspiration across the datasets. MODIS data, for instance, tend to underestimate evapotranspiration in water-saturated areas, while Landsat data appear to overestimate evapotranspiration in forested areas. These findings suggest the presence of systematic deviations influenced by specific hydrological conditions and land use types. Despite these differences, the datasets exhibit strong consistency in terms of spatial patterns as well as of generic temporal dynamics, suggesting that the key processes driving evapotranspiration are reliably represented. Analysis of long-term ET trends highlights the sensitivity of different land use types to climatic changes. Notably, all datasets indicate an increasingly earlier seasonal decline in ET on agricultural land over the past 20 to 30 years, reflecting shifts in water availability patterns. These findings provide a foundation for advancing water management models and developing sustainable management concepts. The insights not only support local management strategies but can also offer transferable frameworks for addressing similar challenges in comparable landscapes in Central Europe.