The demand for climate and meteorological information prediction is increasing due to climate change. However, there are many difficulties in predicting extreme rainfall, which is concentrated in the summer or in the rainy season. Extreme rainfall characterized by a large amount of rainfall in a short period of time is a major cause of various water disasters, such as embankment failures, overflows, landslides, inundation of agricultural areas, and urban inundation due to sewer system overflow.

To date, disaster mitigation and management have focused mainly on building dikes or designing sewers to minimize maximum flood damage by probability rainfall through frequency analysis. Recently, however, the frequency of rainfall has been increasing, and the development of rainfall forecasting and precautions technology that predicts rainfall is urgently needed.

This study developed a method for a precipitation retrieval algorithm consisting of combined COMS rainfall and radar rainfall (Rain-1), COMS brightness temperature (TBB), and radar reflectivity (dBz) with a database of co-located Global Precipitation Mission (GPM) Dual Precipitation Radar (DPR) Level 2 precipitation, rain flag, precipitation type flag, land/ocean flag, convective/stratiform flag, and cloud top height flag. The Rain-1 and Rain-2 algorithms were compared with the COMS rainfall algorithm and radar algorithm for the case of the Busan city flood of August 25, 2014. The results showed that the rainfall of the Rain-2 algorithm based on blended TBB and radar reflectivity was more accurate than the rainfall of the Rain-1, COMS, and radar rainfall retrieval algorithms.

Rainfall data obtained from satellite data were used as input data for the 1D–2D Coupled Urban iNundation Analysis model (CUNA) to evaluate the applicability for urban flood analysis. For a test case, the rainfall incident that caused inundation damage in the Oncheon River basin of Busan on August 25, 2014, was selected. As a result, it was determined that the rainfall data using the satellite data showed sufficient accuracy to analyze the cause of inundation through urban flood analysis. It also showed potential for use in inundation vulnerability analysis as well as in flood analysis. The results of this study can contribute to the development of a real-time urban inundation forecasting system through the combination of predicted rainfall data using satellite rainfall and an urban inundation model.