과거 초단기 강수 예보 데이터를 활용한 도시침수 모의 및 선행 예측 시간 분석

Abstract

Due to sudden localized heavy rainfall under recent climate conditions, the risk of urban flooding has been increasing. Urban flooding occurs rapidly within a few hours, making timely response challenging. To enable proactive measures, securing advance lead time through urban flood prediction is critical. The study here utilized archived very short-range precipitation forecast data from Korea Meteorological Administration as inputs for an urban flood simulation model. The model simulated flooding scenarios and analyzed the time required to reach hazardous inundation depths of 20 cm, 30 cm, and 50 cm, defined in the study here, thereby estimating lead time. Prediction accuracy was validated by comparing the precipitation forecast data with observed hourly precipitation measurements from the Automated Synoptic Observing System. Simulations employed a 1D–2D coupled hydrodynamic model that utilizes the TUFLOW module in XP-SWMM developed in prior studies by others to represent both inland and riverine inundation dynamics. The study area was the lower basin of the Naengcheon Stream in Pohang, Gyeongsangbuk-do, which experienced flood damage due to Typhoon Hinnamnor on September 6, 2022. Results indicated that certain areas achieved over one hour of lead time prior to reaching the 50 cm hazard threshold. These findings provide foundational insights for leveraging very short-range forecast precipitation data in urban flood forecasting systems.