东亚区域性强降水事件复杂网络的构建与分析

Strong precipitation is not only a low probability extreme value deviating from the mean state of climate, but also a regional strong precipitation event (RSPE) with spatial and temporal nonlinear characteristic. The prediction of RSPE has always been an important and difficult problem. However, at present simulation effect of climate models is not very ideal, which mainly because the interaction mechanism of strong precipitation has not been studied clearly. After building RSPE database using objective identification technology, we will analyze spatial and temporal continuity of RSPE from time maintain and space synchronization, then to calculate nonlinear correlation between different grid points by event synchronization method, and to construct strong rainfall network over eastern Asia. Besides, we will also get dynamics mechanism of strong rainfall by several complex network measures, such as degree, clustering coefficient, average path length, and so on. Based on the spatial distribution of these network measures, regions with similar rainfall dynamical behavior will be selected. Using temporal information about the occurrence of strong rainfall events between grid points in bilateral network, we will analyze the spatial distribution of out degree, in degree, and the difference of them, and then reveal the law of strong rainfall interaction between grid points. Those works will provide theoretical basis and technical support to the prediction of RSPE over eastern Asia.

强降水事件不仅是特定地区和时间发生的偏离气候平均态的小概率极值，更表现为具有时空非线性特征的区域性强降水事件（RSPE），而对RSPE的预测历来是气候学研究的热点和难点。当前动力模式对RSPE预报能力受限的重要原因在于格点间强降水相互作用机制还不够清晰，还需要从非线性相互作用的角度开展新的相关研究。本项目基于RSPE客观识别技术建立东亚地区夏季RSPE事件库，从时间维持和空间同步的角度分析RSPE的时空持续性特征；采用事件同步法建立格点之间的强降水非线性相关，进而构建能够描述强降水过程和邻近区域间相互作用的复杂网络模型；结合网络特征量的空间分布揭示RSPE的动力学结构新特征；利用强降水双向复杂网络中含有的格点间强降水事件发生的时序性信息，分析网络特征量出度、入度及两者差值，进而揭示格点间强降水相互传递作用规律，从另类途径为东亚夏季RSPE的模拟和预测提供理论依据和技术支持。

强降水的预测是气候学研究的热点和难点，当前动力模式对强降水预报能力受限的重要原因在于格点间强降水相互作用机制还不够清晰，还需要从非线性相互作用的角度开展新的相关研究。本项目采用理论分析与数值计算相结合的研究方法，对东亚地区强降水的区域性特征及可预测性进行了研究。建立了东亚区域性强降水事件库；获得了东亚夏季强降水的时空持续性规律；利用事件同步方法建立了强降水复杂网络，构建了能够刻画强降水复杂过程和格点间相互作用的简化数理模型；深入分析了复杂网络结构特征量的空间分布，获得了东亚地区强降水的动力学结构特征，从强降水动力学行为相似的角度将东亚地区进行了区域划分；结合强降水双向复杂网络中含有的格点间强降水事件发生的时序性信息，分析网络特征量出度、入度及两者差值，初步得到了格点间强降水相互传递作用规律；利用强降水复杂网络中格点的关联强度和关联方向信息，从强降水时空记忆性的角度建立动力预测模型，利用预测模型对典型强降水进行预测试验，为东亚夏季强降水的预测提供新的理论依据和技术支持。