Seismologists have created a new algorithm that recreates what the tsunami looked like when it was born, to determine what it will do in the future.
The development could lead to a new system that gives coastal cities early warning of incoming tsunamis and where the most dangerous places to be.
Current tsunami warning systems rely on scenarios based on previous patterns in that area. But that’s not always reliable, as the tsunami’s behaviour when and where it hits the land, varies according to where and how it formed.
Now, scientists at the Australian National University can use real-time data to look back in time to when the tsunami formed and use that data to predict how it behave when it breaks on shore – all as the wave speeds towards its destination.
They call the algorithm the “Time Reverse Imaging Method” and say it is fast enough to compete with existing algorithms but is much more accurate.
“It is not based on some guess, it’s based on real-time information,” said Jan Dettmer, a seismologist at the ANU.
He and his colleagues presented the algorithm at the 171st meeting of the Acoustical Society of America this week in Salt Lake City.
They created the algorithm using data from plate tectonics in the Japan Trench. Tsuanmis are formed when plates move and push against each other resulting in earthquakes underwater, displacing the water above. They kill an average of 8,000 people every year, according to the United Nations Office for Disaster Risk Reduction, making early warning vital.
“Once the earthquake happens, then we have minutes,” Dettmer says. But often the exact trajectory of the wave is unknown until it is too late. In order to predict its course, you need know the initial sea surface displacement, or, what the wave looked like when it first started.
While there are a lot of sensors in the Pacific Ocean, they do not cover the entire seafloor. So Dettmer took the information gathered from the 11 March , 2011, Tohoku-Oki earthquake and tsunami and used it to go backward in time mathematically, calculating what the tsunami looked like when it first started.
Once he had the information from the beginning of the tsunami, he added it to the sensor data and projected what the tsunami would look like once it hit land. By checking his results against what actually happened in 2011, Dettmer was could also perfect his algorithm.
The plan is to apply test his method on other recorded earthquakes and fine-tune the technology until it is ready for implementation, which he says could be in less than five years.
“This is a step forward,” Dettmer adds. “This research can be part of the next generation of tsunami warning systems that are based on real time information.”