Tatsuya SUZUKI, Masanobu KII, Mitsuhiro MIYAMOTO
Program in Architecture, Civil and Environmental Engineering, Dept. of Eng. and Design, Kagawa Univ.

The outbreak of a new coronavirus infection (Covid-19) has necessitated infection control measures in various locations. In the midst of this situation, the torrential rains that hit Kumamoto Prefecture in July 2020 forced shelters to implement infection control measures. According to the Japan Medical Association's manual, it is recommended that the distance between family members should be at least 1 meter and the distance between beds should be at least 2 meters. This is expected to drastically reduce the original capacity of shelters. In order to cope with unexpected disasters, which have been increasing in recent years, it is necessary to properly operate shelters during emergencies. The purpose of this study is to evaluate the scale and location of shelters during emergencies by measuring the access to available shelters when the capacity of the original shelters is reduced.
　In this study, we set up two types of evacuation scenarios. One scenario assumes that evacuees do not know the congestion status of the shelter until they arrive there. In this scenario, the evacuees go to the nearest shelter from their residence and check the availability of the shelter. If it is available, they use this shelter. If there is no vacancy, the evacuees move from the shelter to the nearest shelter and repeat the same process to check the availability of the shelter (hereinafter referred to as the search scenario). The other method is to assign evacuees and shelters in advance (hereinafter referred to as the allocation scenario). Using the above two scenarios, we measure the number of people who have difficulty evacuating and the distance they travel when the capacity of the shelter is set to 1/2 or 1/3.
　The results of the analysis showed that in both scenarios, the average travel distance increases as the number of people who can use shelters decreases, and the number of people who have difficulty evacuating increases. The average travel distance in the search scenario is about 10 to 15 km, but the maximum travel distance is more than 1000 km, resulting in many cases of actual evacuation difficulties. On the other hand, in the allocation scenario, the average travel distance was about 8 km to 13 km, and the maximum travel distance was about 96 km. It is clear that the allocation scenario can reduce the travel distance more than the search scenario.
　We found that a reduction in the capacity of shelters affects the number of people who have difficulty evacuating and the distance they travel to shelters. The reduction in the capacity of shelters due to Covid-19 is likely to have an impact on evacuees who have to stay for a long time. It is difficult to maintain a large shelter for such emergencies, so it is suggested that the use of private facilities such as hotels is necessary. Furthermore, the establishment of shelters dispersed by region will help to reduce the number of evacuees who have difficulty evacuating or who are forced to travel long distances.