Unraveling the hydrological dimension of ecosystem resilience: Drought-induced responses of water retention and nonlinear drivers

Terrestrial ecosystem resilience is a key indicator for assessing sustainability under climate change. However, current evaluations largely rely on vegetation indices and often overlook the critical hydrological functions that sustain ecosystems during droughts. To bridge this gap, this study proposes an assessment framework based on water retention (WR), i.e., the capacity of ecosystems to intercept and retain precipitation. Focusing on the transboundary Yuanjiang–Red River Basin (YRRB), key basin-scale hydrological variables from 1973 to 2015 were simulated using the Variable Infiltration Capacity (VIC) model. The identified drought years were then used as the basis for calculating stability, resistance, and recovery metrics, which were integrated to construct a comprehensive resilience index (ERI). The results demonstrate a distinct “upstream-low, downstream-high” spatial pattern in ecosystem resilience. Notably, the upstream headwaters, despite having high vegetation coverage, exhibited lower ERI, suggesting that high vegetation density does not necessarily translate into enhanced hydrological stability in sub-basins with relatively lower precipitation and higher evaporative demand. Spatial analysis using the Optimal Parameter-based Geographical Detector (OPGD) showed that precipitation and potential evapotranspiration had the highest explanatory power for the spatial pattern of ERI, whereas land-surface factors showed nonlinear associations with ERI. Specifically, an approximately inverted U-shaped association was identified between forestland coverage and ERI. In some upstream sub-basins, ERI tended to be lower in the forest-cover stratum above approximately 87%. These findings highlight the importance of considering spatial vegetation–water relationships under different hydroclimatic conditions and provide a reference for ecosystem management in transboundary basins such as the YRRB.