Vegetation plays a crucial role in nature, with intricate interactions between it and the geographical environment. The Yangtze River Basin (YRB) refers to the third-largest river basin globally and an essential ecological security barrier in China. Monitoring vegetation dynamics in the basin is of profound significance for addressing climate change, soil erosion, and biodiversity loss in the basin’s ecosystems. Here, we investigate the spatiotemporal variations of vegetation at both the basin and land-cover scales in the YRB from 2000 to 2020. We elucidate the determinants driving the changes and explore future NDVI trends. The results indicate that NDVI in the YRB increased at a rate of 0.0032 yr⁻¹ (P < 0.01) over the past 21 years, and it is anticipated to maintain an upward trend in the future. Regions in the upper and middle reaches of the YRB demonstrated higher NDVI, whereas regions in the headwater area and the lower reaches showed lower NDVI. Significant vegetation improvement was primarily concentrated in the central part of the basin, while noticeable vegetation degradation was observed in the eastern region. Temperature and wind speed were identified as the primary controlling factors affecting vegetation greenness. Global-scale climate oscillations played a significant role in driving periodic variations in NDVI, with La Niña events tending to increase NDVI, while El Niño events hindered its rise. Land cover types were influenced by long-term interactions between natural factors and human activities, although short-term vegetation variations might be more affected by the latter. Our findings provide valuable insights into the mechanisms behind vegetation variability driven by multiple variables, and the strong vegetation carbon sink capacity advances the conservation and development of ecosystems.

MODIS NDVI time series data from February 2000 to December 2020 were selected, and the MODIS Reprojection Tool was used for image mosaic, format and projection transformation. For missing data in January and February 2000, mean values from the corresponding period over multiple years were utilized for interpolation. Referring to the quality control file, a maximum-value composites procedure was employed to generate monthly NDVI dataset to eliminate the influence of invalid values or outliers. Furthermore, the mean method was utilized to generate annual NDVI datasets. Using ArcGIS software, the NDVI time series dataset of the study area from 2000 to 2020 was clipped and true value conversion was performed.

Meteorological data was adopted from the annual climate dataset of 704 Chinese national meteorological stations in the YRB between 2000 and 2020. The dataset includes average  temperature, average relative humidity, average wind speed, annual total precipitation, and annual total sunshine duration. These data were sourced from the Hunan Meteorological Bureau and have been undergone quality control procedures. The meteorological stations in the YRB are uniformly distributed spatially, although the upper reaches (including the headwater region) has relatively sparse station coverage. Linear interpolation was conducted to fill in missing data for some years to ensure data integrity. IDW interpolation implemented in ArcGIS software was performed to obtain raster images consistent in projection and resolution with NDVI for the above-mentioned climate variables.