STUDY ON INFLUENCE OF PRIOR RECOGNITION OF FLOODING STATE ON EVACUATION BEHAVIOR

Abstract In recent years, disasters have caused extensive water damage in various parts of Japan. Flooded evacuation routes and ineffective hazard maps commonly constrain evacuees’ behavior, which results in casualties. In this study, storm surge flooding analysis was carried out using OpenFOAM, which is a computational fluid dynamics software that monitors flooding state as a function of time, and evacuation simulations using a multi-agent system that considers evacuation behavior knowing the flooding state established by the flooding analysis. In some of the scenarios, many residents were caught in the flooding and could not be evacuated. Additional evacuation simulations revealed that evacuation behavior is more effective if residents have advance knowledge of the area expected to flood. Identifying these flood hazard areas and recognizing these areas in advance as impassable enables evacuees to avoid flood hazard areas; consequently, there is a decrease in the number of evacuees who get caught in the flooding. Anticipating the area expected to be flooded suggests the best evacuation locations and routes. Using these simulation results with useful hazard maps and disaster education, residents will correctly recognize the risk of a storm surge disaster and prepare for it, which will assist recovery. Finally, our analytical and simulation approach to disaster management is amenable to long-term planning for disaster-resistant cities.