Machine Learning-Driven Dynamic Classification of Sri Lanka into Meteorologically Homogeneous Regions Based on Current Climate Conditions
Conference
Iresh, ADS, Athapattu, BCL, Fernando, WCDK et al. (2025). Machine Learning-Driven Dynamic Classification of Sri Lanka into Meteorologically Homogeneous Regions Based on Current Climate Conditions
. 10.1109/ICATC68823.2025.11407538
Iresh, ADS, Athapattu, BCL, Fernando, WCDK et al. (2025). Machine Learning-Driven Dynamic Classification of Sri Lanka into Meteorologically Homogeneous Regions Based on Current Climate Conditions
. 10.1109/ICATC68823.2025.11407538
This study illustrates the regionalization of Sri Lanka into meteorologically homogeneous areas by employing Ward's hierarchical cluster analysis, seasonality index, rainfall indices, as well as factors such as altitude, latitude, and longitude. It also includes evaluations of discordancy and heterogeneity. The analysis utilized total rainfall data collected from 352 stations across the country. The discordancy among the clustered regions was assessed using a discordancy measure, followed by an examination of the heterogeneity in the non-discordant regions through a heterogeneity measure. Additionally, the initial regions were further subdivided to create meteorologically homogeneous groups that met the criteria for both non-discordant and heterogeneity. Ultimately, Sri Lanka has been divided into 11 meteorologically homogeneous regions. These regions are crucial for regional studies, especially in countries like Sri Lanka, where rainfall data may be scarce.
