The research laboratory of Professor Shunsuke Managi, Principal Investigator at Kyushu University Graduate School, in collaboration with aiESG Inc., conducted an analysis of the Kumamoto earthquake and showed that the loss of air conditioning may increase the 30-day mortality risk for elderly people affected by the disaster, and quantitatively estimated the necessary electricity and air conditioning resources.
In recent years, the increasing frequency of heat waves due to climate change has exacerbated health risks for vulnerable people such as the elderly who have lost access to air conditioning due to earthquakes and other disasters. However, disaster response has focused primarily on post-disaster recovery, and there are insufficient mechanisms to encourage proactive investment in disaster prevention and mitigation.
In this study, the research group spatially integrated reports and data to estimate a series of models ranging from housing loss to elderly exposure, air conditioning demand, and increased mortality risk. Their goal was to build a framework that utilizes AI and spatial data to visualize areas that should be prioritized for support during disasters and the amount of infrastructure needed.
The study focused on the Kumamoto earthquake, integrating satellite observation data, census mesh population data, building and shelter information, municipal damage reports, and meteorological observation data. Based on building exposure levels and distance from the epicenter, data was allocated to 36,657 census confidential processing groups. The study then estimated the relevant population aged 65 and over, their 30-day cooling needs, electricity demand, and the difference in mortality risk between those with and without effective air conditioning under the same outdoor temperature.
In the central scenario, an estimated 303.5 homes would be functionally lost, affecting approximately 246 elderly people. Protecting these homes would require a peak cooling capacity of 34.6 kW, or 622.5 kWh of electricity per day. Without effective cooling, the 30-day mortality load in the model would increase by 5.34%.
The research group emphasizes the importance of not only providing rapid assistance during disasters, but also using AI and spatial data to identify vulnerable areas in advance and promote disaster prevention investment and infrastructure decentralization.
