复合热浪事件不仅会危害人类健康，而且不利于社会经济和生态系统的可持续发展。近年来川渝地区复合热浪事件显著增多，探究影响川渝地区夏季复合热浪的因素及模式模拟差异的原因对研究川渝地区夏季复合热浪具有重要意义。利用最新的观测和再分析资料，本文研究了川渝地区夏季复合热浪与青藏高原大气热源的关系，评估了28个CMIP6（Coupled Model Intercomparison Project Phase 6）模式对两者关系的模拟能力，进一步探讨了模拟差异产生的可能原因。结果表明，川渝地区东部的夏季复合热浪日数年际变化最为显著，其与同期青藏高原东南部的大气热源存在显著的负相关关系。当青藏高原东南部夏季大气热源偏弱时，川渝地区东部高空出现异常的反气旋。该异常环流伴随着强烈的下沉运动一方面通过绝热增温，另一方面通过减少云量、增加地面净短波辐射，进而有利于川渝地区东部复合热浪事件的发生。大部分CMIP6模式能够模拟出川渝地区夏季复合热浪日数和青藏高原夏季大气热源气候态的空间分布特征，但模式对青藏高原夏季大气热源年际变率的模拟普遍优于其对川渝地区夏季复合热浪日数的模拟，且模式对川渝地区夏季复合热浪与高原大气热源间的负相关关系的模拟存在较大的差异。进一步的研究表明，模式能否模拟出青藏高原东南部夏季大气热源异常时川渝地区东部的垂直运动和云量异常，是影响模式对川渝地区夏季复合热浪与青藏高原东南部大气热源关系模拟能力的主要原因。

Compound heatwave events are heatwaves that occur simultaneously during the day and at night， which not only threaten human health but also impede sustainable socio-economic and ecosystem development. In recent years，the frequency of compound heatwave events in the Sichuan-Chongqing（SC）region has in‐ creased significantly，and an advanced understanding of the plausible factors of compound heatwave variations over the SC region is of great importance for a better understanding of their impacts. Based on observational and reanalysis data，the relationship between the summer compound heat wave frequency（HWF）in SC and the anomalous atmospheric heat source（AHS）over the Qinghai-Xizang（Tibetan）Plateau（QXP）is statistically analyzed in the present study. Then，this study evaluates the performance of 28 climate models from the Coupled Model Intercomparison Project Phase 6（CMIP6）in simulating the relationship between the SC HWF and the QXP AHS of 1985-2014 and explores the key factors causing the differences in the simulation capability among the models. The results show that both the climate mean and the standard deviation of the HWF have relatively large values in the eastern SC，where the HWF variability shows a significant negative correlation with the summer AHS over the southeastern QXP. When the AHS over the southeastern QXP is weaker in summer，an anomalous anticyclone is present over the eastern SC，leading to an anomalous downward motion. This anomalous descent induces cloud cover and surface shortwave changes，leading to the formation of heatwave events over the eastern SC. Most CMIP6 models can essentially simulate the spatial distribution characteristics of the observed climate states of the summer HWF in the SC region and the summer AHS in the QXP. However，the models perform better in simulating the standard deviation of the AHS over the QXP than in simulating the HWF in the SC regions. There are significant differences in the simulated relationship between the SC HWF and the QXP AHS between different models. This is mainly due to the different abilities of the models to simulate the anomalous vertical motions over the SC regions corresponding to the anomalous summer AHS over the QXP. Further analysis shows that a good performance of the models usually depends on their ability to simulate the anomalous vertical motions and cloudiness in the SC regions，which can lead to air temperature variations. Therefore， the simulated vertical motion and cloudiness anomalies in the eastern SC in response to the QXP AHS anomalies are the key circulation factors affecting the model’s performance in simulating the HWF-AHS relationship.