Real - Time Tsunami Prediction System Developed
Researchers at two Japanese universities and a private-sector lab have jointly developed technology that predicts tsunami flooding in near real time.
The system was developed by Tohoku University, the University of Tokyo and Fujitsu Laboratories.
Researchers carried out simulations on tsunamis and coastal flooding conditions using Japan's supercomputer Fugaku and then trained an artificial intelligence model to learn how the two data sets relate.
They say by inputting data on tsunamis that have occurred after earthquakes, the AI model can predict flooding conditions in coastal areas before landfall.
The technology was applied to predict tsunami flooding in Tokyo Bay triggered by a theoretical megaquake around the Nankai Trough off Japan's Pacific Coast.
Researchers say a highly accurate prediction was made in several seconds using an ordinary PC.
They say they aim to put the technology into practical use in two years' time.
Expectations are high that the system will result in safer evacuations of residents in coastal areas.
Fujitsu Laboratories' Oishi Yusuke expressed hopes to develop a system that will be applicable to various situations, saying AI technology has progressed over the past decade to allow for high-resolution predictions.
Tsunami flooding prediction technologies have also been developed by NEC and Mitsubishi Electric.