ABSTRACT Fractional vegetation coverage (FVC) is an important indicator for measuring ecosystem function and environmental quality. This study, using MODIS data from 2001 to 2022, investigates the spatiotemporal dynamics, driving factors, and predictive performance of FVC in Anhui Province of China. The results show that over the past 20 years, FVC in Anhui has significantly increased, with more than 80% of the region showing an upward trend, and 47.72% of the area experiencing highly significant increases, while less than 7% showed a decline. The annual mean FVC increased from 0.64 to 0.74, exhibiting a clear spatial gradient of “high in the south, low in the north.” Feature correlation analysis revealed that temperature and evapotranspiration are the main environmental drivers of FVC. Temperature can promote the growth of vegetation and increase its coverage, while evapotranspiration is closely related to regional water stress. In the established predictive modeling, the XGBoost algorithm outperformed random forest and ridge regression, demonstrating the best predictive performance on the validation set ( R 2 = 0.83). While the individual analytical methods employed are well‐established, the primary innovation of this study lies in constructing an integrated framework that bridges traditional long‐term spatiotemporal statistics with advanced ensemble machine learning. This methodological integration successfully transitions regional vegetation analysis from retrospective observation to high‐precision dynamic forecasting. Ultimately, by accurately forecasting vegetation trends under varying thermal and moisture conditions, this research fills a critical gap in Anhui Province, providing robust, data‐driven support for climate change adaptation strategies, targeted afforestation planning, and sustainable land‐use policymaking.
Zhang et al. (Mon,) studied this question.