中国气温降水综合极端事件的观测分析与模拟评估

In the past, most of the studies on extreme weather events, i.e., extreme temperature and precipitation, mainly focus on a single variable。This study first analyzes the compounding extreme weather events using Coupled Model Intercomparison Project Phase 5 (CMIP5) multiple-model ensembles and observations based on commonly used method and Copula method, focusing on the characteristics (i.e., affecting areas, duration, intensity and frequency) of simultaneous occurrence of wet/cold and dry/warm events. By using the method of Copula, the compoundprobability distribution of extreme precipitation and temperature was fitted, and the compound probability, conditional probability and return period of the compounding extreme events was analyzed. In addition to the historical extreme compounding events, this study also evaluate the availability of global climate models in simulating composite extreme events using CMIP5 outputs. Hopefully, this project can enhance the understanding of the ability of the most commonly used CMIP5 multiple global climate models in simulating the compounding extreme events in China, and further discuss the characteristics of the changes in these events, which play crucial roles in improving the global climate models and public policies on climate adaptation and mitigation.

过去的极端气候研究，较多地针对单个气候变量分析。本研究首先利用常规方法和Copula方法对Coupled Model Intercomparison Project Phase 5 （CMIP5）多模式和观测数据进行综合极端事件分析，主要是湿和冷、干和热在过去几十年（1975-2004）同时发生的特征（包括影响范围、持续时间、强度大小、发生频率）。并通过应用Copula方法，拟合极端降水和气温的复合概率分布，分析综合极端事件发生的复合概率、条件概率以及复合重现期(return period)。除研究过去综合极端事件的发生规律外，本研究还将结合CMIP5结果评估全球气候模式模拟综合性极端事件的能力。期望本项目的顺利开展能够深入了解当前最常用的CMIP5全球气候模式组合模拟我国综合极端事件的能力，这对未来全球气候模式的提高和气候适应和缓解政策有重大意义。

伴随全球增暖，灾害性天气日益频繁，比如洪涝、干旱、热浪、寒潮等。这些多为单一气象要素极值造成的，而更糟糕的事实是，当多个要素极值同时或者连续发生时，造成的灾害更为严重，比如连续无雨和高温热浪并存导致的干旱甚至火灾，或者低温和强降水同时发生导致的冻雨。为此，本项目从单一要素角度进一步到气温、降水两个要素通过Copula方法分析了我国极端灾害事件的变化规律。结果发现，我国降水谱有向强降水发展的趋势。随着气温升高，我国南部地区有强降水越来越多，中小雨越来越少，痕量降水全面减少的趋势。气温均值的升高以及气温变率的减小同时促成了高温事件的增多。另外，气温降水联合概率分布显示，低温少雨和高温大雨事件的发生概率明显较高，二者共存。在华南地区遭遇综合气象灾害的风险更高。