Ecological Indicators (Dec 2024)
Ecosystem stability assessment under hydroclimatic anomalies in the arid region of Northwest China
Abstract
Fragile dryland ecosystems face increasing destabilization risks due to frequent hydrometeorological anomalies. This study assessed ecosystem stability in the arid region of Northwest China (ARNWC) under hydroclimatic disturbances using the Autoregression (ARx) model. The assessment was based on the Normalized Difference Vegetation Index (NDVI) from 2001 to 2022 and ecosystem responses to changes in precipitation, soil moisture, and potential evapotranspiration. Key findings include: 1) Precipitation anomaly is the most important controlling factor affecting the ecosystem stability of ARNWC. 2) Ecosystem resilience and resistance exhibited a trade-off relationship, with 26.2 % of the area showing significant clustering of “high resilience - low resistance” or “low resilience - high resistance”. 3) Steppe and desert vegetation demonstrated lower resilience but higher resistance, while shrubs and meadows showed the opposite pattern. 4) Resilience generally decreased with increasing aridity (R2 = 0.584, p < 0.001), while resistance increased (R2 = 0.656, p < 0.001). 5) Lower altitude regions were more resistant but less likely to recover from disturbances, while higher altitude regions were more sensitive to hydroclimatic anomalies but more resilient; In different mountains, a complex nonlinear relationship between stability and altitude was observed. In addition, using an “exposure-resistance-resilience” framework, ecosystems in Northern Xinjiang were found vulnerable to precipitation anomalies, while the Hexi Region faced greater pressure from soil moisture anomalies. This study highlights the crucial role of wet and dry conditions in dryland ecosystem stability, suggesting that long-term ecosystem stability equilibrium is achieved through species adaptation to environmental conditions. These insights provide valuable guidance for ecosystem management in ARNWC and similar drylands.