Nonlinear vegetation dynamics driven by socio-ecological coupling in the alpine canyon region of northwestern Yunnan, China

Northwestern Yunnan, located in the core of the Hengduan Mountains, is both part of a global biodiversity hotspot and a vulnerable socio-ecological system. In this region, complex topography and human activities jointly shape vegetation change, which exhibits pronounced nonlinear dynamics. Using long-term remote-sensing data from 2000 to 2023, this study integrates the BFAST01 model with the optimal parameters-based geographical detector (OPGD) to characterise the coupled gradual-abrupt dynamics of fractional vegetation cover (FVC) and identify their associated driving mechanisms. The results show that: (1) regional vegetation conditions have generally improved but remain strongly constrained by vertical zonation; the 1500–3500 m belt functions as the core ecologically active zone, whereas areas above 4500 m exhibit an abrupt decline in FVC under extreme environmental stress. (2) Nonlinear analysis reveals an apparent vegetation recovery accompanied by substantial instability: ~80% of areas identified as significantly increasing display highly unstable growth trajectories, corresponding to a potential degradation-risk estimate of 73.3%. (3) Temperature and net primary productivity (NPP) form the fundamental natural baseline regulating vegetation dynamics, while the explanatory power of human activities—proxied by night-time light intensity—increased by 87.9% over the past two decades. Through nonlinear interactions with this baseline, human activities appear to exert a disproportionately strong restructuring influence on local ecological patterns. These findings reveal a mismatch between vegetation evolution and environmental pressures in alpine valley systems and provide cross-disciplinary evidence supporting differentiated governance strategies that reconcile energy development with long-term ecosystem resilience.