Vertical Stratification of Seawater in the Southern South China Sea in Winter Based on Cluster Analysis

Temperature, salinity, density, and sound velocity in seawater are the main factors for the formation of seawater thermocline. According to these 4 parameters, the optimal number of clusters is estimated by the K-means clustering algorithm and Gaussian Mixture Model (GMM). Then, the K-means clustering algorithm is used to cluster 4 parameters and 2 parameters of temperature and salinity, respectively. Finally, we compare the clustering results and then stratify the seawater vertically. The study shows that the seawater about 2000 meters deep in the southern part of the South China Sea in winter can be divided into 4 types: surface water, subsurface water, middle water, and deep water. The results of seawater stratification by these 4 parameters are similar to the corresponding temperature-salinity graph (T-S point aggregation graph), which further verifies that the stratification structure of the South China Sea is mainly affected by temperature and salinity.