Spatiotemporal dynamic evolution and driving factors of land use and carbon storage under multi-scenarios in the Huangshui River Basin, China

The Huangshui River Basin (HRB), a critical upper Yellow River ecological barrier, has carbon storage (CS) significantly affected by land use/cover change (LUCC). This study applied the PLUS and InVEST models with localized carbon density data to assess CS from 1980 to 2020 and project it for 2035 under four scenarios: natural development (ND), construction and expansion (CEs), ecological protection (EPs) and cultivated land protection (CPs). The geodetector and XGBoost + SHAP models were used to identify driving mechanisms. Results indicate: 1980–2020 grassland and cropland reduction and construction land expansion lowered total CS. By 2035, only EPs achieved CS recovery; CEs incurred the largest CS loss (1.08 × 10⁶ t), CPs saw the smallest (0.78 × 10⁶ t). CS showed multi-scale spatiotemporal differentiation, with Slope/DEM/NDVI as dominant factors and non-linear interactions between Slope-DEM and D-G.