An Application of Cluster Analysis in Investigating Characteristics of the South and Southeast Asian Monsoon Onset in ENSO Years

In this study, the characteristics of the South and Southeast Asian summer monsoon onset in ENSO years were re-investigated using clustering analysis. Our approach focuses on detecting synoptic patterns, including identifying the pattern that best represents the monsoon onset condition. By projecting the real atmospheric conditions and their temporal evolutions into a recognizable and interpretable "weather pattern space" by using Self-organizing Maps and k-means, we can track the changing patterns and identify the initial occurrence of deep convection driven by large-scale circulation, which is considered the monsoon onset. The ENSO signal was defined by anomalous sea surface temperature over the Nino 3.4 region. The results revealed that the mean onset timing is later in warm phases than in the cold phases of ENSO. The onset in the warm phases is characterized by greater persistence of the subtropical high over the Vietnam East Sea (VES) and drier conditions over the equatorial Indian Ocean. Additionally, monsoon westerlies and deep convection are found to be weaker during the onset of the warm phases. The proposed approach can be considered a new option, an alternative to enhance the existing monsoon onset detection method and explore its characteristics.