Su Zhen, Meyerhenke Henning, Kurths Jürgen
Potsdam Institute for Climate Impact Research, 14473 Potsdam, Germany.
Department of Computer Science, Humboldt-Universität zu Berlin, 12489 Berlin, Germany.
Chaos. 2022 Apr;32(4):043126. doi: 10.1063/5.0077106.
The identification of regions of similar climatological behavior can be utilized for the discovery of spatial relationships over long-range scales, including teleconnections. Additionally, it provides insights for the improvement of corresponding interaction processes in general circulation models. In this regard, the global picture of the interdependence patterns of extreme-rainfall events (EREs) still needs to be further explored. To this end, we propose a top-down complex-network-based clustering workflow, with the combination of consensus clustering and mutual correspondences. Consensus clustering provides a reliable community structure under each dataset, while mutual correspondences build a matching relationship between different community structures obtained from different datasets. This approach ensures the robustness of the identified structures when multiple datasets are available. By applying it simultaneously to two satellite-derived precipitation datasets, we identify consistent synchronized structures of EREs around the globe, during boreal summer. Two of them show independent spatiotemporal characteristics, uncovering the primary compositions of different monsoon systems. They explicitly manifest the primary intraseasonal variability in the context of the global monsoon, in particular, the "monsoon jump" over both East Asia and West Africa and the mid-summer drought over Central America and southern Mexico. Through a case study related to the Asian summer monsoon, we verify that the intraseasonal changes of upper-level atmospheric conditions are preserved by significant connections within the global synchronization structure. Our work advances network-based clustering methodology for (i) decoding the spatiotemporal configuration of interdependence patterns of natural variability and for (ii) the intercomparison of these patterns, especially regarding their spatial distributions over different datasets.
相似气候行为区域的识别可用于发现长距离尺度上的空间关系，包括遥相关。此外，它为改进大气环流模型中的相应相互作用过程提供了见解。在这方面，极端降雨事件（ERE）相互依存模式的全球图景仍需进一步探索。为此，我们提出了一种基于自上而下的复杂网络聚类工作流程，结合了一致性聚类和相互对应关系。一致性聚类在每个数据集下提供可靠的群落结构，而相互对应关系则在从不同数据集获得的不同群落结构之间建立匹配关系。当有多个数据集可用时，这种方法确保了所识别结构的稳健性。通过将其同时应用于两个卫星衍生的降水数据集，我们识别出了北半球夏季全球范围内ERE的一致同步结构。其中两个显示出独立的时空特征，揭示了不同季风系统的主要组成部分。它们在全球季风背景下明确体现了主要的季节内变化，特别是东亚和西非上空的“季风跳跃”以及中美洲和墨西哥南部的仲夏干旱。通过一个与亚洲夏季风相关的案例研究，我们验证了全球同步结构内的显著联系保留了高层大气条件的季节内变化。我们的工作推进了基于网络的聚类方法，用于（i）解码自然变率相互依存模式的时空配置，以及（ii）这些模式的相互比较，特别是关于它们在不同数据集上的空间分布。
