Cluster analysis of tropical cyclone path in Southern Indian Ocean Based on finite mixture model

Abstract Based on the JTWC 1983-2024 tropical cyclone best-path dataset and NCEP reanalysis data, this paper used the finite mixture model to classify 691 tropical cyclones in the southern Indian Ocean into five types according to their path characteristics. The mean regression path of each tropical cyclone type was calculated by cluster analysis. The interannual variability and intermonthly variability of each type were statistically analyzed. Ten typical cases of each type were selected from each type, and the typical guiding flow field characteristics of each tropical cyclone type in the early, middle and late stages were obtained by synthetically analyzing the upper-air situation field and the tropospheric main body (850-300 hPa) guiding airflow.