Monitoring vegetation sensitivity to drought events in China.

The frequency and severity of drought events have increased over the decades under the influence of global warming. Continued drought increases the risk of vegetation degradation. Many studies have investigated the responses of vegetation to drought but rarely from the perspective of drought events. Moreover, the spatial distributions of vegetation sensitivity to drought events are not well understood in China. Thus, the spatiotemporal patterns of drought events were quantified based on the run theory at different time-scales in this study. The relative importance of drought characteristics for vegetation anomalies during drought events were calculated by using the BRT model. Then, the sensitivity of vegetation anomalies and vegetation phenology was quantified by dividing standardized anomalies of vegetation parameters (NDVI and phenological metrics) and SPEI during drought events for different regions in China. The results show that Southern Xinjiang and Southeast China experienced relatively higher values of drought severity, especially at the 3-month and 6-month scales. Most arid areas experienced more drought events but of low severity, while some humid zones underwent few drought events but of high severity. Notable negative NDVI anomalies appeared in the Northeast China and Southwest China, while positive NDVI anomalies were observed in Southeast China and Northern central region. Drought interval, intensity and severity contributed approximately 80 % of the model's explained vegetation variance in most regions. The sensitivity of vegetation anomalies to drought events (VASD) varied regionally in China. The Qinghai-Tibet Plateau and Northeast China tended to exhibit higher sensitivity to drought events. Vegetation in these regions with high sensitivity faced a high risk of degradation and could function as warning signals of vegetation degradation. Drought events at high timescales had a greater impact on vegetation sensitivity in dry zones, while they had a smaller impact on humid areas. With the increase in drought degree of climate zones and the decrease in vegetation coverage, VASD showed a gradual increase. Furthermore, a strong negative correlation between VASD and the aridity index (AI) was observed in all vegetation types. The change in VASD for sparse vegetation was the largest with the change in AI. For vegetation phenology, drought events in most regions delayed the end of the growing season and extended the length of growing season, especially for sparse vegetation. The start of the growing season was advanced in most humid areas, while being delayed in most dry areas during drought events. Knowledge of vegetation sensitivity to drought events will be beneficial to provide decision-making references for the prevention and control of vegetation degradation, especially in the ecological fragile regions.