Abstract Frequent droughts pose serious threats to terrestrial ecosystems under global climate change. However, the spatial patterns of drought risk in China over the past few decades and the mechanisms of vegetation response to drought remain unclear, especially on large watershed scales. Here, we applied a cluster analysis method that captures the multidimensional characteristics of drought indicated by the self‐calibrating Palmer Drought Severity Index (scPDSI) to quantitatively assess drought risk in China from 1965 to 2018. Furthermore, we assessed the direct and legacy impacts of drought on vegetation dynamics across major river basins in China and identified the major drivers using random forest method. Results showed that high‐risk drought areas in China, accounting for ∼17.1%, were mainly located at the junctions of the Inland River and Yellow River basins, and the Yangtze River and Pearl River basins. Vegetation dynamics in northern China were more sensitive to drought, particularly in the Yellow River and the Hai River basins, with growing‐season detrended NDVI and scPDSI correlation coefficients of 0.68 and 0.73, respectively, during 1982–2015. The contribution of precipitation to vegetation growth was higher during droughts, particularly in high‐risk drought areas where the coupling between vegetation growth and precipitation increased with intensified drought. Limited by soil moisture availability, pre‐growing season drought in northern China exhibited legacy effects on spring phenology, particularly evident in the Hai River basin, where it resulted in an average decrease of 0.32 in spring NDVI anomalies. Overall, this study highlights the vulnerability of vegetation activities to drought stress in arid and semi‐arid regions.
Yin et al. (Wed,) studied this question.