Spatiotemporal vegetation dynamics and climate drivers in arid regions: a case study in Gansu

As Western China's crucial ecological security barrier, the ecosystems of Gansu province remain vulnerable for heatwaves and prolonged droughts in recent years. This study investigates vegetation dynamics and their drivers in Gansu Province using MODIS NDVI, land cover, and climate datasets from 2019 to 2023. Theil-Sen trend analysis and Hurst exponent were employed to assess spatiotemporal patterns and persistence, while partial correlation analysis quantified climatic controls. Results reveal dominant vegetation degradation (59.98% of areas), driven by a net expansion of non-vegetated lands (+4273 km²), despite localized forest recovery. Precipitation emerged as the primary driver (partial correlation: 64.22%), with spatial heterogeneity in climate responses: warming enhanced NDVI in cool-humid zones but suppressed growth in arid regions. Hurst-based predictions indicate persistent polarization, with 47.16% of areas facing continued degradation versus 26.56% sustaining improvement. These findings underscore the critical role of land cover-climate interactions in shaping vegetation trajectories, necessitating spatially adaptive strategies to address water constraints and ecosystem fragility in transitional landscapes.