Japanese lab creates technology to detect major tsunami impact in less than a second

The RIKEN Prediction Science Laboratory in Japan has developed technology that uses machines to detect tsunami impacts in less than a second.

According to a press release by RIKEN, this is an improvement of what was formerly at least a half-hour of warning. The new technology is a potential life-saving measure, which is especially notable after the massive tsunami that killed 18,500 people in Japan in March 2011.

“The main advantage of our method is the speed of predictions, which is crucial for early warning,” said Iyan Mulia of RIKEN. “Conventional tsunami modeling provides predictions after 30 minutes, which is too late. But our model can make predictions within seconds.”

The RIKEN team used more than 3,000 computer generated tsunami events to test this technology. It was then tested with 480 other tsunami scenarios, as well as three real tsunamis. The model had a 1% computational effort as far as accuracy. Mulia said this approach also can be used for other similar disaster scenarios as well.

"The sky's the limit. You can apply this method to any kind of disaster predictions where the time constraint is very limited," he said.

Mulia has been studying tsunamis since 2004, when the Indian Ocean tsunami ravaged his home country of Indonesia. 

He is currently working on a storm surge prediction through machine learning as well. He added that the method is only accurate for large tsunamis that are more than 1 1/2 meters high and that his team is looking to improve the model so they can detect smaller ones.