Widespread strengthening vegetation responses and lengthening drought-accumulation timescales across China over the past four decades

Abstract

Drought is one of the most critical climatic disasters affecting the structure and functioning of terrestrial ecosystems. However, how vegetation response intensity and characteristic response timescales have evolved under long-term environmental change remains poorly understood. Here, we integrated multi-source vegetation and climate datasets across China from 1983 to 2022 and used moving-window and multi-timescale correlation analyses to quantify changes in maximum vegetation response intensity (R_max) and optimal drought-accumulation timescale (T_opt). We further compared two vegetation indices and examined patterns across aridity gradients, vegetation structures, and contrasting water-availability conditions, together with uncertainty and robustness analyses. Increasing R_max trends were more widespread than decreasing trends, although the national mean trend varied among vegetation indices, window lengths, and detrending treatments. In contrast, Topt showed a more consistent lengthening tendency, indicating increasing associations between vegetation variability and moisture conditions accumulated over longer timescales. Water-deficit and non-woody regions generally exhibited higher R_max and shorter T_opt, whereas water-surplus and woody-dominated regions showed lower R_max and longer T_opt. Under the most conservative block-permutation and BH-FDR classification, T_opt increased significantly in both water-deficit and water-surplus regions identified using kNDVI, whereas the corresponding NIRv-based trends were positive but not statistically significant. The spatial relationship between R_max and T_opt was generally weak and heterogeneous, with no consistent temporal change in their coupling. XGBoost-SHAP analysis suggested nonlinear associations of R_max and T_opt trends with solar radiation, atmospheric CO₂ concentration, aridity, atmospheric dryness, and precipitation. Overall, these results reveal widespread shifts in vegetation drought-response characteristics across China and provide a scientific basis for assessing ecosystem vulnerability and improving drought-risk management under ongoing climate change.