Work

Spatiotemporal characteristics of vegetation responses to compound drought and heatwave events in Central Asia

Year: 2026

Type: article

Abstract: With climate warming intensifying, compound drought and heatwave (CDHW) events have become more frequent and intense in arid Central Asia, severely affecting vegetation growth. However, systematic quantitative assessments of vegetation responses and their drivers under such compound extremes remain limited. Using meteorological and remote sensing data, we identified growing-season CDHW events across Central Asia from 1982 to 2023. Integrating the Kernel normalized difference vegetation index (KNDVI) and the long-term continuous SIF-informed photosynthesis proxy (LCSPP), we established a vegetation response indicator system including total loss (LOSStot), maximum decline (DECmax), and time to decline (TIMEdec) to characterize vegetation damage. We further applied the XGBoost model and SHAP analysis to quantify the environmental driving effects and interactions. The results indicate that CDHW events increased significantly during 1982–2023, with heatwave frequency rising at 0.38 days yr−1 and an abrupt shift around 1996. High-frequency CDHW events were primarily concentrated in north-central Kazakhstan and northern Xinjiang, China, while their frequency and duration generally increased with latitude, indicating a pronounced northward enhancement pattern. CDHW caused larger vegetation losses than individual drought or heatwave events. Event duration was the most important environmental variable of LOSStot, with the highest mean SHAP value (0.140), while precipitation anomalies during events showed a strong association with rapid vegetation decline, with a direct path coefficient of 0.333 for DECmax. This study clarifies vegetation loss characteristics and key contributing factors under compound extremes in Central Asia, providing a scientific basis for ecological risk assessment and climate adaptation in arid regions.

Source: GIScience & Remote Sensing

DOI: 10.1080/15481603.2026.2741443