Analysis of summer heavy rainfall pattern types in East Asia of the Korean Integrated Model (KIM)

Climate change has intensified localized heavy rainfall events during summer, leading to frequent damage in urban areas, and the diversity of these occurrences poses significant forecasting challenges. This study employs the K-means clustering technique in conjunction with the Korean Integrated Model (KIM) ensemble prediction system to categorize recent summer precipitation cases on the Korean Peninsula, and to test its potential as an ensemble prediction diagnostic tool. Heavy rainfall during summers from 1991 to 2020 was analyzed using Global Precipitation Measurement (GPM) and ECMWF Reanalysis v5 (ERA5) data. By applying K-means clustering, we identified three representative synoptic patterns. With We specifically examined two recent cases: July 6th, 2021, and August 8th, 2022. Case 1 was classified as an upper- and low-level jet synoptic pattern, generating precipitation as a low-level jet coupled with an upper-level jet over the peninsula. Case 2 was categorized by an 850 hPa GPH meridional gradient, resulting in precipitation due to cold air from a blocking event over the Urals and Kamchatka and a warm moist low-level jet from China. The classification using K-means clustering enhances our understanding of heavy rainfall mechanisms and offers new diagnostic tools for ensemble forecasting. This approach not only diminishes forecast uncertainty but also contributes to reducing the impacts of extreme precipitation, underlining the potential of the KIM ensemble prediction system in probabilistic weather forecasting.