长江中下游致洪大暴雨事件的关键环流型聚类分析

Clustering Study on Key Circulation Pattern of Flood-Waterlogging Rainfall Events in Middle and Lower Reaches of Yangtze River

摘要:

大范围的持续性暴雨事件会影响水文条件，造成洪水影响。本文针对20世纪60年代以来长江中下游致洪大暴雨事件个例，基于随机分化模拟退火聚类方法和扰动集合相似方法，研究致洪大暴雨事件的关键环流型及其对大暴雨的定量贡献。结果表明，长江中下游致洪大暴雨事件多发生在梅雨集中降水期，多个事件平均的日峰值降水强度达暴雨量级。致洪暴雨的500 hPa环流形势可聚类为4类，分别为：东亚南高北低型（1型）、东亚三明治型（2型）、南支槽型（3型）和高纬双阻型（4型）。4个聚类下的西太平洋副高和南支槽对致洪大暴雨的贡献较为稳定，约占3成和1.5成；而中高纬系统影响较不稳定，东北亚环流异常对聚类1、3和4型的贡献平均近2成，贝加尔湖阻塞异常对聚类2型有微弱贡献；中纬度西风槽异常对聚类4型的贡献约为2成。4个聚类下的10～30 d和30～60 d低频环流显著异常区和原始观测的环流异常关键区基本一致。西太平洋副高区的低频环流对所有事件均有正贡献，占比约2～7成；其中，除30～60 d低频环流对所有事件有影响外，其10～30 d低频活动对聚类1～2型事件也有明显影响。南支槽区的30～60 d低频环流对聚类3和4型事件的贡献占比为27%和16%。大暴雨事件相联系的高纬低频环流关键区分别位于贝加尔湖和鄂霍茨克海（1型）、乌拉尔山和西风槽区（2型）、东北冷涡（3型），且贡献各不相同（12%～31%）。上述关键环流型及其定量贡献评估结果可为加深对致洪大暴雨事件的形成认识和预报预测提供依据。

Abstract:

Prolonged and widespread heavy rainfall events can significantly impact hydrological conditions, leading to devastating floods. This study focuses on individual cases of flood-waterlogging rainfall events in the middle and lower reaches of the Yangtze River since the 1960s. We employed a combination of SAN (simulated annealing and diversified randomization) clustering method and perturbed ensemble analog method to investigate the key circulation patterns associated with these flood-waterlogging rainfall events and quantify their contributions to heavy rainfall. Results indicate that these flood-waterlogging rainfall events in the middle and lower reaches of the Yangtze River typically occur during the Meiyu season. Averaged over all events, the daily peak precipitation intensity reaches the level of a rainstorm. The 500 hPa circulation patterns associated with extreme rainfall were categorized into four classes: East Asian dipole mode (Class 1), East Asian Sandwich mode (Class 2), South Branch Trough (Class 3), and High-Latitude Double Block (Class 4). The key circulation features of each class are distributed in the Western Pacific Subtropical High, South Branch Trough, mid-latitude westerly trough, and high-latitude blocking activity areas, contributing to 40% to 70% of flood-waterlogging rainfall events. The Western Pacific Subtropical High and South Branch Trough contribute relatively consistently in all four categories, accounting for approximately 30% and 15%, respectively. However, the influence of mid-high latitude systems is less stable, with the Northeast Asian circulation anomaly contributing to an average of nearly 20% to Classes 1, 3, and 4, and the Baikal Lake blocking anomaly making a weak contribution to Class 2. The mid-latitude westerly trough anomaly contributes to approximately 20% of Class 4 events. The significant anomaly regions of low-frequency circulation at 10-30 days and 30-60 days in the four classes are generally consistent with the observed circulation anomaly key regions. Low-frequency circulation in the Western Pacific Subtropical High region has a positive contribution to all events, ranging from about 20% to 70%. Among these, the 10-30 day low-frequency activity also has a notable impact on Class 1 and 2 events. The 30-60 day low-frequency circulation in the South Branch Trough region contributes 27% and 16% to Classes 3 and 4 events, respectively. The key high-latitude low-frequency circulation relating to the flood-waterlogging rainfall events located in Lake Baikal and the Okhotsk Sea (Class 1), the Ural Mountains and the westerly trough region (Class 2), the cold vortex region in Northeast China (Class 3), with their contribution varying among different classes (12%-31%). The findings of this study on key circulation patterns and their quantitative contributions provide valuable insights for a deeper understanding of the formation and prediction of flood-waterlogging rainfall events.