When trying to improve, refine, and seek more efficiency in disaster management measures, predicting the extent of disaster damage more accurately and communicating the assessments in a coherent manner become crucial elements for the task. This is why there are high hopes for the role of computational science, in particular simulations, and information technology in this endeavor. Professor OISHI Satoru, who specializes in social infrastructure and hydroengineering research at the Research Center for Urban Safety and Security, aims to help society change and enhance disaster resilience by building digital twins. We asked him what we can anticipate from the use of digital twin technology, and how it can contribute to disaster measures.
How is information technology being used to make cities safer?
Oishi:
The “Cross-ministerial strategic innovation promotion program (SIP),” which is led by the Cabinet Office’s Council for Science, Technology and Innovation, lays out key issues for solving societal problems and maintaining economic and industrial competitiveness in Japan, and one of them is to build a smart disaster prevention network. This issue has adopted five agendas, one of which is to enable organizations involved in disaster response to share and leverage information in a cross-organizational manner. Since the first phase of the SIP (2014-2018), this agenda has been about building a system so that the Cabinet Office, Self-Defense Forces, Fire and Disaster Management Agency, Ministry of Land, Infrastructure, Transport and Tourism, and others can share information that they gather across their organizations immediately after a disaster and allow swift search and rescue operations as well as early recovery efforts to happen. Parts of this system were recently put into practical use, so you could say that we have only just begun to use information technology in disaster management.
Rapid information-sharing is essential for overall optimization
The internet was still in its early days when the Great Hanshin-Awaji Earthquake struck in 1995.
Oishi:
At the time of the Great Hanshin-Awaji Earthquake, it took more than half a day for on-the-ground information to reach the Prime Minister's Office, and this subsequently delayed decisions that were needed for recovery efforts. I was at the Kyoto University Disaster Prevention Research Institute studying water-related disasters when it all happened, and watching the whole situation unfold made me realize that the inability to rapidly assess overall situations and systematically issue commands based on that information could result in the loss of lives that could be saved. This became the starting point for my research.
In the dam business, growing opposition in the mid-1990s was making it difficult to construct new dams, and so the research group I belonged to presented a way to control floods by using existing dams. The idea was to reduce reservoir water levels before it rained, based on weather forecast information, so that they could store the excess rainfall. At the time, restrictions based on the “River act” prevented dam operators from adjusting water levels according to the weather forecast, and the academic community was also very critical of our proposal. But later, the idea actually became the government’s policy for river basin management and flood control. This development convinced me that our approach of trying to gain an understanding of the overall situation in order to maximize efforts to address the problem would eventually be appreciated someday, and that is why I expanded my scope of research to include not only disasters involving dams and rivers, but also earthquakes, tsunamis, and storm surges since coming to Kobe University.
Getting results automatically and visualizing the data on maps
Tell us about the digital twin technology you are currently working on in disaster management.
Oishi:
The digital twin technology is a way for us to replicate real-world buildings and land features in a digital space so that we can perform various simulations and prepare for disasters. In particular, our goal is to enhance our ability to respond to larger, more complex, and diversified natural disasters.
To be specific, we look at public information for details regarding a building’s structure or earthquake resistance, as well as traffic volume by the time of day, and population by day and night; and all of this data is stored on the cloud. Then, depending on what you want to know, such as what kind of damage to expect in the event of an earthquake, tsunami, or storm surge, the necessary information is automatically selected from the enormous amount of cloud data, calculations are made, and results are produced to simulate the damage on the digital twin.
It used to be that, if you wanted a certain result, you needed to collect the necessary information yourself and build a program to produce the calculation results. But with digital twins, there is a built-in program that automatically picks out the information necessary for the calculation. Even with data entry, which used to be performed manually, technology has made it possible to automatically extract the necessary data from printed information, making the process much more efficient.
What kind of results can be derived from the calculations?
Oishi:
By inputting seismic acceleration data, it is possible to show how many buildings will collapse across a city on a map. We can also visualize the impact of a tsunami in terms of the areas and extent of flooding with three-dimensional information, or see how various precipitation scenarios could cause flooding in which areas and by how much in minute detail. It is also possible to learn how many people will be stranded in specific areas when a disaster happens.
Let’s say there is heavy rainfall. We would be able to see how some areas will be completely submerged, while others sustain less damage. We can also learn that water does not simply flow from upstream to downstream areas of the river basin because embankments, railroad tracks, and other structures can change the direction of the flow. A characteristic of digital twin technology is that we can produce objective results according to the laws of physics, without being distracted by other aspects.
Providing information that can help disaster management and recovery measures with digital twins
What is the ultimate goal of utilizing digital twin technology?
Oishi:
In fiscal year 2023, our research was selected for an agenda of the SIP initiative to build a smart disaster prevention network, and the ultimate goal is to help society change and enhance disaster resilience. In other words, we hope to provide material that can help people determine the measures they should take to minimize potential damage before a disaster occurs, and — assuming we can assess the full extent of a disaster six hours after it happens — help decide how to allocate people, money, or goods so that recovery efforts can proceed as quickly as possible.
To this end, we are working to provide three outputs. The first is the probabilistic hazard map. If we can calculate the likelihood of buildings collapsing in various types of earthquakes, or areas being submerged in different rainfall scenarios, we could consider more cost-effective measures for areas highly prone to disasters. This is now more or less available for earthquakes, tsunamis, and storm surges.
The second is the dynamic hazard map, which can simulate evacuation routes with avatars. Let’s say the city you live in experiences torrential rain. The dynamic hazard map can show the expected flooding for each road of that area, and for each amount of rainfall in advance. You can also use the avatar on the screen to really see which path you should choose when trying to evacuate. The avatar can be set as a child or an elderly person, so you can see how long each would take to evacuate.
The third is the multi-hazard map, which simulates multiple disasters. For example, earthquakes can damage sea walls. At the time of the Great Hanshin-Awaji Earthquake, they managed to patch up the damaged sea walls before the next storm surge. But it is estimated that a massive event on the scale of the Nankai Trough earthquake would cause more widespread damage, making temporary restoration difficult. This is why we are trying to make it possible to calculate which sea walls will be affected, the extent of damage, and the number of days required for repairs that would enable them to withstand storm surges again through the use of digital twins.
Also discussing how to address the public
What challenges does your research face in the future?
Oishi:
If we were to disclose the raw simulation data generated by the digital twins to the public, that could cause chaos as people would become aware of the realities surrounding their homes. Or, if municipalities started prioritizing certain areas to implement cost-effective measures, there is sure to be backlash from residents of areas that are left behind. So I believe that one of the important themes going forward is finding a way to convey information that would be acceptable to the public.
We want to make it possible for everyone to share the digital twin’s simulation results, so that when municipal leaders present disaster management plans, everyone can participate in an open discussion based on the same sets of information. We call this the democratization of disaster management measures.
Kobe University has a framework called the “Research alliance for multidisciplinary integration for resilience and innovation,” where five disciplines, including humanities and social sciences, natural and health sciences, and urban resilience studies, work together to create a grand design for safe and secure living spaces and environments. I would say that this initiative has been possible because Kobe University is a place where members of different disciplines can go back and forth and see each other. We hope to take advantage of this environment and engage in further discussions about how we should address the public and other important issues.
Resume
Graduated from the Kyoto University Faculty of Engineering in 1991 and the Kyoto University Graduate School of Engineering in 1993. Received his Ph.D. in engineering from Kyoto University in 1998 and is a certified weather forecaster. Became a research associate at the Kyoto University Disaster Prevention Research Institute and an associate professor at University of Yamanashi, before becoming a professor at Kobe University in 2009. He has also been a team leader at the RIKEN Center for Computational Science since 2017.
