Satellite-derived evapotranspiration for calibrating a spatially distributed hydrological model in a highly regulated Iranian river basin

Accurate calibration of hydrological models in heavily managed and data-scarce irrigation basins remains a major challenge due to unreliable or unavailable streamflow observations. Addressing this limitation is critical for improving water resource management and understanding hydrological processes in intensively managed agricultural regions. This study adopts a spatially distributed evapotranspiration (ET)-based calibration approach using the SWAT model and remotely sensed SEBAL ET data. The model is calibrated across 465 hydrological response units (HRUs) using the Kling-Gupta Efficiency (KGE) as the objective function and the DDS algorithm for optimization. Initial calibration with the original SEBAL dataset yields limited performance (mean KGE = 0.29), mainly due to insufficient representation of land-use-specific ET dynamics. To address this, a modified SEBAL dataset incorporating crop-specific variability is introduced, resulting in improved calibration performance (mean KGE = 0.42) and enhanced spatial consistency. Statistical tests confirm that these improvements are significant (p < 0.001). The results show that ET-based calibration effectively constrains model performance and captures irrigation-driven ET dynamics in highly regulated catchments. However, limitations related to spatial scale mismatch, irrigation uncertainty, and equifinality remain. Overall, the study highlights the potential of integrating remote sensing and hydrological modeling for improved water resource assessment in data-scarce regions.