Evaluation of SUEWS sensitivity to vegetation parameters in a residential area in Berlin, Germany

Cities are in the spotlight of environmental research in the face of climate change and local populations’ vulnerability. Land surface models are constantly being improved to better capture urban dynamics with better accuracy and scaling potential to address these challenges. In this study, the Surface Energy and Water Balance Scheme (SUEWS) is evaluated using one year (2023) of energy balance fluxes in Berlin while exploring vegetation parameters derived from near-surface remote sensing and satellite observations.

As the latent heat flux (QE) is strongly influenced by vegetation phenology—particularly leaf area index (LAI)—we evaluate the model’s ability to capture seasonal variations in QE using one year (2023) of 60-minute eddy covariance flux measurements in Berlin. Hourly comparisons show a mean absolute error (MAE) of 25 W m⁻² and a mean bias error (MBE) of approximately 0.5 W m⁻², underscoring the challenges in accurately representing Leaf Area Index (LAI) dynamics and vegetation state throughout the seasons. In contrast, net all-wave radiation exhibits a systematic yet overall low MAE and MBE (~10 W m⁻²), demonstrating the model’s robustness in these calculations.

We further explore the impact of varying data coverage, resolution, and accuracy on model vegetation-parameters sensitivity. Our findings underscore the trade-offs between local and global/regional data inputs and their implications for model accuracy, which is especially relevant for modelling applications in data-sparse regions. The study also emphasizes the potential of Earth observation products—such as ESA World Cover and Copernicus services—to enhance large-scale urban climate modeling. These insights are invaluable for future efforts aiming to improve the accuracy and scalability of predictions across diverse urban environments, especially when accounting for seasonal vegetation dynamics.