An ecological assessment model of water retention based on regional differences in characteristic using remote sensing data

Accurate assessment of Water Retention (WR), a critical ecosystem service, remains challenging due to regional heterogeneity. This study proposes the RD-WR model, a spatially explicit framework integrating regional differences (RD) to improve WR estimation. Leveraging remote sensing data, we develop two novel indicators (RC-Dif and EF-Dif) to quantify terrain and vegetation variability, alongside a boundary-adjustment metric (CRel) to account for cross-ecosystem dependencies. The model enhances InVEST Water Yield outputs by incorporating multi-source hydrologic data (precipitation, evapotranspiration) and DEM-derived landscape features. Validation in the Danjiangkou Ecological Reserve (2015 vs. 2020, contrasting humidity years) demonstrates that the RD-WR model outperforms conventional approaches (G-WR) in resolving fine-scale WR patterns, aligning closely with observed hydrologic gradients. Our findings highlight the importance of regionalized parameters in ecosystem service modeling, offering a potential framework for sustainable water resource management.