Using self-organizing maps to identify the South Asian seasonal cycle

Abstract

A synoptic climatological approach using the self-organizing maps (SOMs) technique has been used to identify South Asian seasonality, by categorizing atmospheric patterns for each of five near-surface variables. This research mainly focused on showing the general climatology of the monsoonal seasonal cycle through four traditional seasons (winter, pre-monsoon, monsoon, and post-monsoon) and the transitions between each of these seasons, highlighting spatiotemporal variability. For each variable, a 9 × 9 SOM was created using surface-level reanalysis data. Each node in each SOM is assigned to a season or a seasonal transition, based on defined standard criteria. The SOM-based synoptic approach helps to increase our understanding and visualization of the different atmospheric conditions and their seasonal cycle of the South Asian climate, such as a complete reversal of mean sea-level pressure over the course of the year, with high pressure over the land in winter and low pressure in the summer, and a clear seasonal cycle in wind direction with cyclonic and anticyclonic circulations in different seasons creating a greater wind speed over the ocean than the land.