Contents lists available at ScienceDirect International Journal of Disaster Risk Reduction journal homepage: www.elsevier.com/locate/ijdrr Simulated tsunami evacuation behavior of local residents and visitors in Kamakura, Japan Tomoyuki Takabatake a, ⁎ , Tomoya Shibayama a , Miguel Esteban b , Hidenori Ishii a , Go Hamano a a Department of Civil and Environmental Engineering, Waseda University, 3-4-1 Okubo, Shinjuku-ku, Tokyo, Japan b Graduate Program in Sustainability Science - Global Leadership Initiative (GPSS-GLI), The University of Tokyo, Chiba, Japan ARTICLE INFO Keywords: Tsunami Evacuation Agent-based modeling 2011 Tohoku Earthquake tsunami Human behavior Visitor ABSTRACT Currently, it is believed that evacuation is the most effective method of protecting lives from tsunamis, in particular after the events of the 2011 Tohoku Earthquake Tsunami. In many coastal areas tsunamis pose a threat not only to local residents, who know the area well, but also to visiting tourists, and thus it is important to consider both these groups when preparing evacuation plans. However, while numerous studies and simulations on tsunami evacuation have been made, research focusing on the influence of visitors on evacuation processes is limited. To clarify this the authors developed an agent-based tsunami evacuation model which considers the different behavior of local residents and visitors, which can estimate the evacuation time, number of individuals reaching each evacuation area, the location of bottlenecks and the number of casualties. The model was applied to study the case of Yuigahama Beach, Kamakura, Japan, with results indicating that the behavior and number of visitors have a significant impact on evacuation processes, especially the location of bottlenecks and the number of casualties. Results also show that heavy congestion will occur during evacuation in places where there are many visitors and thus, in such a situation, the decrease of the moving speed owing to the congestion needs to be appropriately taken into account to simulate the evacuation process. It can be concluded that reducing congestion (i.e., widening roads, guiding visitors to less congested roads) is a crucial countermeasure for a sight- seeing location to reduce the casualties that can result from a tsunami. 1. Introduction The 2011 Tohoku Earthquake Tsunami devastated large parts of Japan's northeastern coastline and claimed over 15,800 lives [23]. During this event coastal infrastructure such as breakwaters, seawalls and coastal dykes sustained significant damage [20,21]. Although it was recognized that reinforced concrete buildings were often strong enough to withstand tsunami forces [7], many of these were severely damaged and some were washed away [21,24]. One important lesson that can be drawn from this devastating tsunami is that it is difficult to protect coastal areas and human life from extreme tsunamis using only hard measures (e.g., coastal dykes, seawalls, buildings). Following this event the Japanese Coastal Engineering Community started classifying tsunami events into two different levels based on their level of severity and intensity (as reported by Shibayama et al. [30]). Level 1 events have a return period of several decades to 100 years or more and generate relatively lower inundation heights (typically less than 7–10 m). Level 2 events are less frequent, with return periods ranging from a few hundred to a few thousand years. The tsunami inundation heights would be much higher (typically, over 10 m), possibly even reaching 20–30 m in height. The construction of hard measures to protect human life and property is impractical for the case of Level 2 events, given the expense involved [6]. Therefore, evacuation is the most effective way to protect against the loss of human life for Level 2 tsunamis [30]. This in turn has highlighted the need for further research into designing and establishing good evacuation plans. For some coastal areas not only local residents but also visitors are at risk of suffering the consequences of a tsunami. Generally speaking, visitors have less knowledge than local residents about the location of tsunami shelters and the safer routes to reach them. Their presence and behavior should therefore affect evacuation time, the number of evacuees reaching each shelter, the locations of bottlenecks and casualties. Thus, in order to establish effective evacuation plans it is important for disaster risk managers to consider both local residents and visitors. One effective way of investigating human behavior under extreme conditions is to conduct evacuation drills. Through such drills, not only is it possible for participants to learn evacuation routes and the location of safe areas [35], but disaster risk managers can also clarify http://dx.doi.org/10.1016/j.ijdrr.2017.04.003 Received 7 January 2017; Received in revised form 12 April 2017; Accepted 12 April 2017 ⁎ Corresponding author. E-mail address: tomoyuki.taka.8821@gmail.com (T. Takabatake). International Journal of Disaster Risk Reduction 23 (2017) 1–14 Available online 13 April 2017 2212-4209/ © 2017 Elsevier Ltd. All rights reserved. MARK