Assessing the Demographical Dynamics of Evacuations During Flood Hazard Using Mobile Spatial Statistics

This study proposes a method to quantitatively assess evacuation demographics during regional floods using Mobile Spatial Statistics (MSSs). It focuses on Koriyama City, affected by Typhoon Hagibis in 2019, as well as Yamagata City, which experienced torrential rains in July 2020, when COVID-19 infection risks in evacuation shelters could have affected evacuation behavior. Both events led to flooding. MSS provided by NTT Docomo were used to explore the dynamics of the population. Evacuees’ demographics, according to the changes in river levels, were presented by gender and age group. Our results show differences in evacuation dynamics between the two regions: In Koriyama, younger people were more likely to evacuate faster; in Yamagata, older people moved faster than other age groups. At the evacuation peak, the average relative evacuation ratios were 2.2 times higher for women in their 20s in Koriyama, and 2.5 times higher for women in their 30s in Yamagata. Gender differences indicate that women were more likely to evacuate than men. The effect of the COVID-19 pandemic on evacuation dynamics remains unclear; however, concerns about infection risk in shelters may have influenced people’s willingness to evacuate. MSSs provide a robust representation of evacuation dynamics in the local context and can help municipal governments develop more targeted evacuation policies, such as tailored warnings for specific age groups and enhanced support for older adults.