Weather Clustering of Waves and Abrupt Change Analysis of Typical Extreme Waves in the Shelf Seas of China

Abstract Wave characteristics can vary significantly under different weather systems, necessitating weather clustering of waves to understand ocean processes. In this study, a hybrid clustering method is used to identify the atmospheric circulation patterns driving ocean waves over the shelf seas of China. Instead of the traditional random initialization method, the self‐organizing map method is used to systematically determine the reasonable initial centers for K ‐means method, avoiding local optima and achieving globally optimal clustering. By applying this hybrid clustering method, the regional clustering results are derived, such as the number pattern of clusters and the specific weather types (i.e., cold surge, summer monsoon, extratropical cyclone and Borneo vortex). The directional convergence characteristic, the temporal convergence characteristic and the typical atmospheric circulation pattern indicate the necessity and rationality of the wave clustering. By analyzing the intensity, frequency and spatial extent of these types of waves, it is discovered that the cold surge is a dominant weather system responsible for generating extreme waves throughout the entire study region. The annual frequency of extreme cold surge waves from 1996 to 2018 is generally higher than that from 1979 to 1995, and a significant abrupt change in the mean frequency of such waves is identified in the northern South China Sea, with their mean frequency even doubling. This increase may be attributed to the weak polar vortex and the significant “trough‐ridge‐trough” pattern near the Ural region, providing favorable cold sources and dynamic conditions for cold surges to drive extreme waves in the study region.