Quantifying the spatiotemporal evolution of NDVI and Its response to driving factors based on the geographical detector model in the Northwest China

Under the combined impacts of global climate change and human activities, understanding vegetation spatiotemporal dynamics is critical for ecosystem management. This study quantifies NDVI changes in Northwest China from 2001 to 2021 using multitemporal remote sensing data and geographical detectors. The results indicated that: (1) Regional mean NDVI was low (0.2842), with 70.6% of areas below 0.4; (2) NDVI increased significantly at 0.0022/a (p < 0.05), especially post-2008; (3) NDVI stability was high, with a coefficient of variation of 0.254; (4) Precipitation (54.24%), sunshine hours (46.22%), and elevation (44.06%) were dominant drivers, while human factors had weaker effects (e.g. night light: 8.61%). The geographic detector model revealed synergistic interactions between surface and climate factors (e.g. elevation ∩ precipitation, q > 80%), overcoming spatial non-stationarity limitations of traditional regression. These findings highlight the need for targeted restoration in arid zones (e.g. Xinjiang deserts) and inform adaptive ecological strategies under climate change.