| Abstrakt: |
Evacuation time estimate (ETE) is an essential part of the preparedness for tsunami disaster risk reduction. Existing tsunami ETEs have been implemented using GIS-based methods or agent-based simulations; however, their evacuation departure estimates have largely relied on hypothetical scenarios or stated-preference surveys because of the lack of sufficient empirical observations. In this study, we analysed a region-wide evacuation survey conducted for the 2011 Tohoku tsunami to fill this gap. The survey included over 20000 behavioural trips collected from 49 coastal cities, which had the greatest area coverage and number of samples in existing studies. The analyses revealed that there were large variations in tsunami evacuation departures in different areas; that is, faster evacuation tendencies were found in the northern ria coastal areas compared with the southern plain areas. In contrast to the evacuation behaviour, the departure timing for milling exhibited smaller variations, suggesting that these behaviours were driven primarily by strong ground shaking. Our analysis indicates that delayed evacuation in the plains is linked to longer milling trips, influenced by regional characteristics. For enriched scenarios in tsunami ETEs, this large amount of evacuation data was compiled into three tsunami evacuation departure curves with different evacuation departure tendencies. • Over 20000 behavioural trips during the 2011 Tohoku tsunami were analysed. • Delayed evacuation in plain areas is linked to longer milling trips. • The large amount of data was compiled into three evacuation departure curves. [ABSTRACT FROM AUTHOR] |
Full Text from ScienceDirect