Incorporating rapidly developing thunderstorm data into a deep convection scheme for improving short-term prediction of heavy rainfall over South Korea

In this study, we examined the potential of the Korea Rapid-Development Thunderstorm (K-RDT) product obtained from a geostationary meteorological satellite to improve the short-term prediction of heavy rainfall caused by a mesoscale convective system over South Korea. Specifically, we utilized a simple nudging technique to integrate K-RDT data into the Simplified Arakawa Schubert (SAS) deep convection scheme of the Global/Regional Integrated Model System (GRIMs) Regional Model program (RMP). Our analysis focuses on selected cases of heavy rainfall. The nudging experiments outperformed the control experiments in terms of precipitation forecasts. Notably, the experiment that used longer nudging times produced the best results. Our results also demonstrate that the K-RDT, with its resolution of 1 km, can detect small-scale convective cells that have clear impacts on large-scale atmospheric fields. This suggests that incorporating such small-scale information into numerical weather prediction (NWP) models can significantly improve forecasting skill, especially when the model cannot represent subgrid-scale convection.