An Investigation Into the Long-Chain Propagation Patterns and Driving Factors of Multiple Types of Droughts in Xinjiang

The process of drought propagation is extremely complex and involves all aspects of the water cycle. Clarifying the mechanism of drought propagation is of great significance for understanding the evolutionary pattern of drought and constructing a drought early warning system. However, due to the complexity of droughts and the numerous factors influencing them, there are certain limitations and shortcomings in studying the intrinsic links, propagation characteristics and triggering mechanisms of multiple drought types. To this end, this paper takes the Xinjiang region as the research object, identifies the drought events based on the run theory and determines the drought propagation time by matching the drought events; risk probabilities, trigger thresholds and drivers between different drought types were explored using copula functions, Bayesian theory and random forest models. The results show that (1) the average drought propagation time from meteorology to agriculture and from meteorology to hydrology in Xinjiang is 2.1 and 2.8 months, respectively. (2) The probability of the occurrence of agricultural (hydrological) drought increased with the severity of meteorological (agricultural) droughts. Specifically, the average probability of extreme meteorological (agricultural) droughts triggering mild agricultural (hydrological) droughts reached 89% (41%), with an average trigger threshold of −1 (−2.2). (3) Mild meteorological drought on the northern slopes of the Tianshan mountains was the primary trigger for mild agricultural drought, with an average trigger probability and threshold of 45% and −0.89, respectively. In contrast, moderate meteorological drought on the southern slopes of the Tianshan mountains was the main trigger for mild agricultural drought, with an average trigger probability and threshold of 56% and −1.1, respectively. (4) Meteorological drought triggers a significant reduction in the threshold for agricultural drought, mainly influenced by precipitation and potential evapotranspiration. Agricultural drought triggers a significant reduction in the threshold for hydrological drought, with temperature and potential evapotranspiration being the main drivers of the threshold change.