Nonlinear responses and differentiated pathways of driving mechanisms underlying the spatiotemporal evolution of water retention in the Qinba Mountains

Study Region The Qinba Mountains represent a critical water retention area characterized by complex terrain and hydrological heterogeneity. Study Focus This study distinguished high, moderate, and low water retention states and integrated the InVEST model, random forest regression, and piecewise structural equation modeling to systematically analyze the spatiotemporal evolution, key drivers, and causal pathways of water retention. To address the limited monthly representation of baseflow in the InVEST-SWY module, an observation-constrained seasonal allocation approach was further used to disaggregate simulated annual baseflow into monthly values, improving the representation of intra-annual water retention dynamics. New Hydrological Insights for the Region Water retention showed a south-high and north-low spatial pattern, with moderate-to-high water retention levels covering 59.01% of the region. From 2000–2023, water retention remained generally stable, showing a nonsignificant increase of 0.58 mm yr −1 . Precipitation was the strongest positive driver, whereas potential evapotranspiration imposed the strongest negative constraint. The dominant regulatory pathways exhibited clear state dependence, shifting from precipitation dominance with vegetation-wind speed synergy in high water retention states to precipitation dominance with strong evapotranspiration limitation in low states, with moderate states representing a transitional regime in which precipitation control coexisted with coordinated topography-vegetation regulation. These findings clarify the dynamic regulation of water retention in the Qinba Mountains and provide scientific support for regional water resource management.