Heavy Rainfall Frequency Analysis for Urban Flooding Using Rain Gauge Observations and d4PDF: A Case Study of Busan, South Korea

Recently, urban flooding has become more frequent due to short-duration, high-intensity rainfall. However, current design rainfall standards are based on limited observational records, which leads to large uncertainty when estimating extreme rainfall with long return periods.This study evaluates how design rainfall estimates change under increasingly variable future climate conditions. It focuses on differences in probability rainfall estimates depending on data sources and analytical methods, and examines how these differences affect the interpretation of recent flood-producing rainfall events. Busan was selected as the study area, and long-term hourly rainfall data from the ASOS station were used together with the large-ensemble climate dataset d4PDF (Database for Policy Decision-making for Future Climate Change). For the observational data, a parametric frequency analysis was applied, while for the d4PDF dataset, both nonparametric and parametric methods were used to estimate rainfall values for different return periods.In addition, major rainfall events that caused urban flooding in Busan between 2020 and 2025 were analyzed. The return periods of these events were estimated using the probability rainfall curves derived in this study, and the results from different data sources were compared.The results of this study suggest that the limitations of observation-based design rainfall can be improved by using large-ensemble climate data. This approach can provide useful information for urban flood risk assessment and for improving design standards under changing climate conditions.

Acknowledgments: The research for this paper was carried out under the KICT Research Program (Project no. 20260161–001, Development of Digital Urban Flood Control Technology for the Realization of Flood Safety City) funded by the Ministry of Science and ICT.