Hydroclimatic associations with vegetation resilience in China's Loess Plateau

Understanding how recovery-related vegetation resilience varies along hydroclimatic gradients is important for evaluating vegetation sustainability in water-limited regions. Here, using satellite-derived vegetation indices and resilience metrics derived from critical slowing down theory, we assessed the spatial patterns and temporal dynamics of recovery-related vegetation resilience across China's Loess Plateau (LP) from 2000 to 2023 and examined how these patterns and dynamics relate to hydroclimatic conditions. Regional mean resilience shows vegetation-type-specific non-monotonic dynamics: forest resilience fluctuates after an initial increase and declines after 2014-2015, whereas grassland resilience peaks around 2010 before declining. Despite these regional-mean phase shifts, pixel-level resilience remains generally stable. Spatially, areas of relatively high resilience within each vegetation type occur mainly in forest interiors and in grasslands of the east-central LP and western highlands. Although resilience and vegetation greenness are positively associated in space, their temporal changes are only weakly synchronized. Soil moisture is the highest-ranked individual predictor of resilience for both forests and grasslands, although its importance may reflect integrated hydroclimatic conditions. Higher resilience generally occurs in regions characterized by greater water availability, lower temperatures, or weaker intra-annual temperature variability. These findings support a context-dependent interpretation of large-scale resilience metrics and provide a basis for adaptive management aimed at sustaining stable forest and grassland ecosystems in climate transition zones and water-limited regions.