Spatiotemporal evolution of water conservation and its driving factors in the Weihe River basin based on the InVEST model

Elucidating the spatiotemporal dynamics and driving mechanisms of water conservation (WC) is fundamental to basin-scale ecological restoration. This study quantified an InVEST-based WC index across the Weihe River Basin from 2000 to 2025. Annual water yield was first estimated using the InVEST Annual Water Yield module and was subsequently adjusted using topographic, soil-hydraulic, and flow-velocity factors to characterize the relative water conservation potential. Geodetector was then used to examine spatial associations between the modeled WC index and eight selected factors. Eight physical and ecological factors, namely elevation (DEM), mean annual temperature (TEMP), slope, nighttime light (NTL), population density (POP), gross domestic product (GDP), net primary productivity (NPP), and the normalized difference vegetation index (NDVI), were evaluated at the pixel scale. The mean annual InVEST-based WC index was 120.78 mm, showing a long-term upward trend despite interannual climate variability. Spatially, WC exhibited a persistent south-high and north-low pattern, with high-value areas progressively extending northward. This spatial evolution coincided with pronounced land-use restructuring, characterized by cropland contraction and forest and grassland expansion. The corresponding decline in the cropland share of basin-total WC and increases in the forest and grassland shares indicate a close spatiotemporal association between land-use restructuring and WC redistribution. DEM and TEMP generally exhibited high spatial explanatory power, reaching q-values of 0.477 and 0.456 in 2020, respectively, whereas NDVI ranked first in 2010. The largest joint q-value occurred for DEM ∩ NPP in 2020 (q = 0.657), highlighting the spatial correspondence between WC and combined terrain-vegetation stratification. These findings characterize regional patterns in the modeled WC index and their associations with the selected factors.