FUKUOKA - In the ongoing series on disaster preparedness, the focus this time is on how to make evacuation decisions. During heavy rains that threaten communities, timely evacuation information issued by local governments becomes the key to saving lives. In Fukuoka Prefecture, a pilot test of a new AI-powered system is underway to assist in making these difficult decisions.
Takashi Shiromizu, a 64-year-old farmer in Ogori City, Fukuoka Prefecture, has seen his vegetable fields flooded repeatedly by heavy rains. At times, the damage has reached as much as 20 million yen. "Once it floods, there’s no way to recover," he said. "Rather than constantly investing in measures like surrounding the fields to keep water out, I have to accept flood damage as a fact of life and find ways to generate income elsewhere."
Ogori City has experienced repeated flood damage to farmland and urban areas. Although there was no damage from yesterday’s rainfall, the city recorded a total of 820 water-related incidents over the past five years. In response, the city has introduced an AI-based municipal disaster management integration system for this fiscal year as part of a pilot project.
Professor Mitani, who developed the system, explained its strengths: "We can predict, up to six hours in advance, whether flooding will expand or not. This allows local governments to take preemptive measures."
The system combines weather data from the Japan Meteorological Agency with local conditions such as the proportion of elderly residents and the distance to evacuation centers. AI processes this vast amount of information to quickly forecast where and how severely flooding might occur.
"For example, if this river overflows, people living in the red zones on the map may face difficulties evacuating," said Mitani. "Authorities can issue early evacuation orders for these areas, especially for elderly residents."
The system displays risk levels for each area on a color-coded map showing predicted danger several hours ahead. Based on this information, municipalities can encourage residents to evacuate earlier and more precisely.
Until now, Ogori City relied on small teams to collect and assess large amounts of data manually before issuing evacuation instructions according to the city's disaster response manual. "We had to go out and check the situation, gather information, analyze it, and then forecast how conditions might develop before taking action," said one official.
The city now hopes that the new system will enable faster and more informed decisions. "By using AI analysis to forecast flooding and other developments, we hope to speed up the process of issuing evacuation information," the official added.
While expectations are high for the system’s ability to support quicker decisions, Mitani emphasized that residents also need to understand the risks in their areas. "It's important for people to recognize whether they live in high-risk zones and how vulnerable they may be," he said. "By being aware of these factors, residents can better prepare themselves and plan their evacuations accordingly."
As municipalities face the critical task of protecting lives, there is growing hope that the adoption of advanced systems like this will help eliminate evacuation delays.
Source: FBS
