近几十年来，中亚干旱区增温速率约为全球平均速率的两倍，降水模式变化明显，干旱高温复合事件对当地农业生产和生态环境造成了严重影响。明晰中亚干旱区干旱高温复合事件的时空特征及植被响应过程可为中亚干旱区减轻气候变化影响与防灾减灾提供有效的理论基础。目前，有研究已分别对中亚干旱区干旱、高温与植被间响应关系开展了探讨，但对干旱与高温事件的复合效应、干旱高温复合的多种类型、植被在干旱高温复合事件过程中的响应差异仍待补充。本文以中亚干旱区为研究对象，以揭示并发、续发两类干旱高温复合事件特征及植被对干旱高温复合事件的响应差异为研究目标。首先，本文基于多源数据识别了1950~2023年并发与续发两类干旱高温复合事件，分析了两类干旱高温复合事件在单一事件组成、时空特征与变化趋势上的差异；进一步，基于PKU GIMMIS NDVI数据，提取了1982~2022年植被对干旱高温复合事件的响应参数，探讨了植被响应与干旱高温复合事件、植被类型的关系；最后，基于可解释的机器学习方法，给出了中亚干旱区植被对干旱高温复合事件响应转变的持续时间与强度阈值。本文的主要结论如下：

（1）中亚干旱区极易发生干旱高温复合，约46.02%的干旱事件、54.87%的高温事件发生了复合，干旱高温复合事件的持续时间相对干旱事件延长1.61倍，相对高温事件延长6.14倍，显著放大了单一事件的影响。在事件组成上，续发干旱高温复合事件中高温事件更强，干旱事件较弱，其高温事件持续时间比并发干旱高温复合事件多2天，强度高0.25；干旱事件持续时间短8天，强度低7.54。在复合类型上，69.62%的并发干旱高温复合事件中高温发生于干旱事件内部，57.61%的续发干旱高温复合事件由高温引发干旱导致。

（2）1950~2023年中亚干旱区干旱高温复合事件的频次、持续时间与强度明显上升。在事件比例上，并发干旱高温复合事件是中亚干旱区干旱高温复合事件主要类型，约99.61%的区域并发干旱高温复合事件的比例高于续发干旱高温复合事件；在增长趋势上，续发干旱高温复合事件增速更为明显，约92.43~97.55%的区域续发干旱高温复合事件增长显著性更高，1950~2023年续发干旱高温复合事件平均增速是并发干旱高温复合事件的1.12~1.84倍，近40年趋势差异进一步扩大至1.28~2.33倍；在极端复合事件上，中亚干旱区极端干旱高温复合事件以强度增加为主，在塔里木河流域，哈萨克西部、帕米尔高原西部等地区极端干旱高温复合事件已经常态化。

（3）在剔除中亚干旱区植被不对称变化趋势后，1982~2022年中亚干旱区植被对干旱高温复合事件响应自后2005年从积极转变为消极。在植被类型上，干旱高温复合事件对林地具有积极作用，其他类型植被主要为消极作用，其中耕地与草地对干旱高温复合事件最为敏感。植被响应主要受干旱高温复合事件时间、强度及空间位置调控，当干旱高温复合事件持续时间>38天、SMI>0.4时，植被响应由积极转变为消极。

本文将并发干旱高温复合事件的强度描述方法扩展到续发干旱高温复合事件中，系统探讨了并发与续发两类干旱高温复合事件的特征，并在事件角度揭示了中亚干旱区植被对干旱高温复合事件的响应特征，给出了植被干旱高温复合事件响应的关键因子及转变阈值。这有助于全面认识中亚干旱区干旱高温复合事件特征及其影响，为中亚干旱区干旱高温复合灾害的预防与预警提供科学依据。

In recent decades, the rate of warming in arid Central Asia (ACA) is about twice the global average, with significant changes in precipitation patterns leading to more frequent compound droughts and hot events (CDHEs) causing serious impacts on agricultural production and the ecological environment. Revealing the spatio-temporal characteristics of CDHEs and the vegetation response to CDHEs in the ACA provides an effective theoretical basis for understanding climate change in the region, its impact, and drought prevention and mitigation. Currently, several studies investigate the relationship between drought, heat stress, and vegetation in the ACA. However, the different types of CDHEs, as well as the differences in vegetation response in the ACA, remain unclear. This paper focuses on variations in vegetation response to different types of CDHEs in the ACA. Firstly, based on the ERA5-land data, Compound CDHEs (Com_CDHEs) and cascading CDHEs (Cas_CDHEs) are identified for the period from 1950 to 2023. Then, the difference in the composition of the single events, the spatio-temporal characteristics of the two types of CDHEs are analyzed. The influence of different threshold values and different datasets on the extraction of CDHEs has also been examined. Subsequently, based on the PKU GIMMIS NDVI data, the characteristics of vegetation response during the CDHEs are identified from 1982 to 2022. The contributions of CDHEs and vegetation types to the vegetation response are quantified. Lastly, the thresholds for the duration and magnitude of the CDHEs that cause negative vegetation anomalies in the ACA are provided. The main conclusions of this paper are as follows:

(1) Droughts and heatwaves are highly susceptible to compounding in the ACA, with about 46.02% of drought events and 54.87% of heatwaves compounding. And the duration of CDHEs was 1.61 times longer than that of drought events and 6.14 times longer than that of heatwaves, which significantly amplifies the impacts of a single event. The duration of heatwaves in the Cas_CDHEs was 2 day longer than that of Com_CDHEs, with a higher intensity of 0.25. In terms of compound type, 69.62% of Cas_CDHEs were caused by heatwaves triggering droughts, and heatwaves occurred within droughts in 57.61% of Com_CDHEs.

(2)The frequency, duration, and magnitude of CDHEs in the ACA from 1950 to 2023 rise markedly. Now, Com_CDHEs are the main type of CDHEs, about 99.61% of regions have a higher proportion of Com_CDHEs than Cas_CDHEs. The increasing trend of the frequency, duration, and magnitude of the Cas_CDHEs is much higher than another, with 92.43-97.55% of regions have a higher increasing trend of Cas_CDHEs. The growth rate of Cas_CDHEs in the period of 1950-2023 is 1.12-1.84 times that of Com_CDHEs, and the difference between the trends of the two types of CDHEs further increases to 1.28-2.33 times in the last 40 years. The magnitude of extreme CDHEs increased most in the ACA, and in the Tarim River Basin, extreme CDHEs have been normalized in western Kazakhstan, western Pamir Plateau and other regions.

(3) From 1982 to 2022, a significant change in vegetation response to CDHEs is observed in 2005 in the ACA, under removing the trend of vegetation. The duration, magnitude, and location characteristics of the CDHEs explain 83% of the variation in vegetation response. Vegetation response shifts from positive to negative when the duration of CDHEs exceeds 38 days and the SMI value exceeds 0.4. All vegetation types show negative anomalies except for the positive effect of CDHEs on forest, with cropland and grassland being the most sensitive to CDHEs characteristics.

This paper presents a comprehensive analysis of the characteristics of Cas_CDHEs and Com_CDHEs. It extends the magnitude description method of Com_CDHEs to Cas_CDHEs and elucidates the response characteristics of vegetation to CDHEs in the ACA in the event-specific perspective. Additionally, it identifies the key factors that influence vegetation response and the transition threshold of vegetation response in CDHEs. This enables a comprehensive understanding of the characteristics of CDHEs and provides a scientific basis for the prevention and early warning of CDHEs in the ACA.