Kushiro town center is expected to experience a major tsunami in the future. This study used a multi-agent simulation comprising walking residents to elucidate the changes in the spatial patterns of tsunami shelter buildings in this location and examined their effectiveness as well as difficulties. The conducted simulation modeled residents in a district with no high-rise buildings. The behavior of residents was assumed as a horizontal movement toward tsunami shelter buildings located in residential neighborhoods where they would access the top floor via stairs. Tsunami inundation projections for this region were updated after the Great East Japan Earthquake. As a part of the tsunami countermeasure, multiple high-capacity shelters were constructed. The results of this study confirmed that such efforts improved the evacuation environment. However, the simulation also revealed that residents of areas distanced from the tsunami shelters would slow down on climbing stairs, and walking would likely to be obstructed near the stairs. These people would find it difficult to reach the upper floors of the buildings by the time the tsunami arrives. Therefore, the study results indicate that the effectiveness of tsunami shelters should not be judged solely by their holding capacities; rather, their usefulness should be evaluated by the number of evacuees they can accommodate before the tsunami hits. This scope was estimated by considering the horizontal and vertical movements of the residents, and the simulation outcomes evinced that the study method was apt for the performed appraisal.

The communication of information on mountain climbing safety is one of the preventive measures against mountain accidents. This study focused on Asahidake in Daisetsuzan National Park, the highest mountain in Hokkaido, where trekking/hiking accidents occur frequently, and clarified the factors leading to mountain accidents. Based on interviews with 75 visitor groups in August 2016, and 300 comments on social networking sites, we identified the specific risks behind the accidents. These consisted of problems with signages on the trails, and risks caused by low body temperatures (hypothermia), sudden changes of weather, and trails conditions. Regarding the problems in information communication on safety, the responses to the interviews revealed the following problems: the web information did not convey specific dangers in the mountain environment to visitors; the difficulty level of trails (Daisetsuzan Grade) was not indicated on signages at the time of the survey; there was a lack of information in English for foreigners. We suggest the following improvements in the information communication: (1) providing explicit information on the web to help visitors better understand the dangers of the mountain environment, for example, by showing the temperature differences compared with other places, (2) color coding the trail signages for visitors to more easily recognize the level of trail difficulty (Daisetsuzan Grade), and (3) providing information about the day’s trail and weather conditions to foreign visitors just before the start of trekking or hiking.