Zoning of the Water Area by Ocean Surface Temperature Using Cluster Analysis

In the work, the ocean temperature field is zoned for the selected research area. Zoning is performed by the cluster analysis method, called the “without a teacher” classification, the Ward method, the Euclidean metric, and the k-means method. The conclusion is made about the best partitioning variant, and a map with the boundaries of the obtained classes is constructed using the Surfer geoinformation system. Based on the results of the work, the optimal division of the temperature field into three classes is obtained, and the physical conclusions of the clustering result are explained. The map shows the boundaries of the areas of the water area that differ in hydrological characteristics. The cluster analysis results are analyzed from a physical point of view, and the correspondence of the division into classes of the influence of the system of currents, ocean fronts, and water circulation is revealed. The work aims to apply modern data mining technologies for hydrometeorological research in practice, which, in the future, will allow the correct study of a group of hydrometeorological parameters similar in the physical mechanism of formation or influencing factors. The application of the data mining algorithm is shown on the example of zoning the ocean surface temperature in the eastern Pacific Ocean for zoning coastal waters by the type of factors affecting them.