Autonomous evacuation and local community

Purpose: The Aso area of Kumamoto Prefecture and the western part of Oita Prefecture in Japan experienced heavy rainfall from midnight until morning on 12 July 2012. Flooding and landslides caused by this torrential rain killed 31 people and injured 11. This paper aims to analyze a time series of flood risk perception and evacuation behavior, and to identify factors that promote effective autonomous evacuation. Design/methodology/approach: A time series of flood risk perception and evacuation behavior following the 2012 flood was developed and systematically analyzed. Differences between the data sets (compiled from survey data) for the evacuated group and the non-evacuated group were statistically evaluated. Then, an evacuation behavior model was developed to simulate which households would be likely to evacuate in different scenarios. The relationship between disaster prevention and activities of a local community were also statistically assessed. Findings: This study concludes that an assessment of river conditions and evacuation advice from fellow local community members are the factors that most strongly influence and promote autonomous evacuation. This study also revealed that the everyday activities of a local community have the potential to foster effective disaster prevention and emergency responses if they promote the building of relationships between community members. Originality/value: The research focused on actual decision-making and autonomous evacuation behavior. Whereas previous studies were limited to the analysis of activities of disaster prevention on a normal day, this study proved that the usual activity level in local community activities and relationships significantly affected evacuation behavior.