Protecting as many lives as possible from disasters: Analyzing evacuation behavior and solving social issues with Fujitsu technology

Hello! This is the Fujitsu PR note editorial department.

Today marks 13 years since the Great East Japan Earthquake occurred on March 11, 2011. Various studies have been conducted on the behavior of people who evacuated from the tsunami at that time, and it has become clear that complex factors, including where people were when the earthquake struck, what they were doing such as working, and the need to confirm the safety of family members, as well as various circumstances and psychological factors, influenced their actions.

At Fujitsu, there is someone who has been researching the behavior of people who evacuated from the tsunami of the Great East Japan Earthquake for many years, starting from when they were a graduate student. In this interview, I would like to share what has been learned through this research and the possibilities that can be expanded by combining those results with Fujitsu's technology.

Interviewee
Fumiyasu Makinoshima
Profile: Completed the doctoral program at the Graduate School of Engineering, Tohoku University in 2019 with research on tsunami evacuation. Ph.D. (Engineering). Currently a researcher at the Social Digital Twin Core Project, Converging Technologies Laboratory, Fujitsu Limited, and a specially appointed associate professor (visiting) at the International Research Institute of Disaster Science, Tohoku University.

When did you start your research, and what kind of research are you conducting?

Makinoshima: I have been researching the behavior of people when a tsunami strikes since I entered graduate school at Tohoku University in 2014. My grandparents were affected by the 2004 Niigata Chuetsu Earthquake, which sparked my interest in the field of disaster prevention. Later, in 2011, I participated in the Joint Survey Team for the Great East Japan Earthquake Tsunami Evacuation led by Professor Fumihiko Imamura of Tohoku University, which led me to enter graduate school at Tohoku University and begin my research on tsunami evacuation. I continue my research with the desire to help save as many lives as possible when a disaster occurs in the future.

I will introduce an example of research on evacuation behavior during the Great East Japan Earthquake. People take various actions between the time an earthquake or tsunami occurs and the time the tsunami reaches the coast. Recent surveys and analyses have revealed that before people take evacuation actions, they engage in very complex actions that are different from evacuation, which delays the start of their evacuation. However, the background of these complex actions, such as "why people chose those actions" and "what they were thinking at the time," had not been sufficiently clarified. To clarify the detailed actual behavior at the time, including the psychological state of people when complex actions were taken, I visited Ishinomaki City, Miyagi Prefecture, which suffered enormous damage in the Great East Japan Earthquake, many times. I collaborated with NPOs and residents working on disaster legacy preservation in the area and conducted detailed interviews with about 100 people who experienced evacuation from the tsunami. I analyzed the background of the complex actions where psychological and social factors are intertwined, and have published the results in papers and other formats.

What did you learn from this behavioral analysis?

Makinoshima: First, through interviews with people who evacuated from the tsunami, I clarified where people were moving at that time. Figure 1 shows the results. It shows the flow of people at that time, along with a breakdown of where they were immediately after the earthquake occurred and where they were heading as their destination. If many people evacuated to high ground immediately after the earthquake, the number of people heading to Hiyoriyama (high ground) as their first destination would be high in this figure. However, out of the 101 people surveyed, only 35 people moved to Hiyoriyama in their first movement. It was found that others moved to their homes or elementary schools, or returned to their workplaces if they were out for work. Even after the first movement, some people continued to move to other locations multiple times, such as visiting acquaintances' homes or returning home, revealing a very complex evacuation process where about half of the people had two or more destinations.

What is the background to this?

Makinoshima: The analysis results showed that actions between the occurrence of the earthquake and the arrival of the tsunami were complicated by various purposes such as confirming the safety of family members, checking for damage to homes and workplaces, and work. If people were out when the earthquake occurred, there were many actions of heading home to confirm safety or damage. Also, if they were out for work, there were many actions of returning to their workplace or home. It is thought that such actions were likely to occur because telephones and other means of communication were not working at the time, making it difficult to get in touch. It can be seen that in addition to where they were when the earthquake occurred, their situation and social factors, such as whether they had family, where their family was, and whether they were at work, had a major impact on their evacuation behavior.

Also, the analysis results revealed that evacuation behavior can influence the movements of others and potentially trigger a chain reaction. In the case of Ishinomaki City at the time of the disaster surveyed this time, it is known that the initiative evacuation behavior of elementary school teachers who immediately evacuated children to high ground promoted the evacuation behavior of their families and people around them. People evacuate while having various social connections. Therefore, it is possible that one person's decision-making can influence the actions of others and even affect the evacuation trends of the region. This means that the decision of oneself or an organization to flee or not to flee has a greater significance than just determining the success or failure of one's own evacuation. In the event of a tsunami disaster, I believe that if we understand the tendency of people to take certain actions during a disaster, based on the principle of immediate evacuation to high places such as high ground or the upper floors of robust buildings, it will lead to the creation of social mechanisms and technological development that promote evacuation behavior. I plan to continue my research to contribute to reducing damage during disasters as much as possible.

What kind of initiatives are you currently working on following these research results?

Makinoshima: Currently, I am working on research and development related to social digital twins at Fujitsu's Converging Technologies Laboratory, and I am also considering applications for disaster prevention based on the knowledge gained from my previous research. Recently, based on the actual evacuation situation during the Great East Japan Earthquake, I have devised an evacuation simulation method that considers the influence of communication between evacuees and psychological factors, and have published it as a paper (Figure 2). When I evaluated the effect of initiative evacuation using this simulation, I found that the trend of evacuation behavior could change significantly due to a small number of initiative evacuees. I believe this is knowledge that will contribute to the consideration of new evacuation promotion methods and the formulation of effective evacuation plans.

In recent years, with the advancement of human behavior observation technology using AI, various data related to human behavior, such as human flow observation information, have become available. By analyzing such data, understanding human behavior, and modeling it to combine with simulations, it has become possible to reproduce and predict human behavior with higher accuracy. If we can understand and predict human behavior, it may become possible to proactively consider effective measures in digital space and actually work toward solving problems before they manifest in real society. At Fujitsu, research and development of social digital twins are being advanced as a technology to support policy formulation for solving diverse and complex social issues. I believe that such technology will lead to the solution of a wide range of social issues, not just disaster prevention. I would like to continue to advance technological development and contribute to the solution of various social issues, including disaster prevention.