Article on Instantaneous tracking of earthquake growth with elastogravity signals
The research article on "Instantaneous tracking of earthquake growth with elastrogravity signals" by Prof. Bertrand Rouet-Leduc, Disaster Prevention Research Institute (DPRI), Kyoto Univers et.al. was published at Nature in May 2022.
Everyone wants to know the occurrence of a huge earthquake as soon as possible. Yet it is difficult to estimate the magnitude of large earthquakes. The recent breakthrough of "speed-of-light prompt elastogravity signals (PEGS)" raises hopes of an answer to this situation while remaining to be tested on early warning systems.
The team from Kyoto University and G¨¦oazur has developed a new method based on deep learning AI for detecting "prompt elasto-gravity signals" (PEGS). PEGS is a gravitational change generated by the massive motion of mega earthquakes and can be recorded by a seismograph. PEGS also conveys information about ongoing earthquakes at the speed of light, arriving much faster than the fastest seismic waves.
"These signals allow us to track the extent of an earthquake in real time as soon as it reaches magnitude 8 and above," says author Bertrand Rouet-Leduc.
In the past, however, the small amplitude of PEGS has hindered its application to earthquake and tsunami warning systems. Standard early warning systems based on seismic waves estimate magnitude directly from shaking but fail to detect rapidly the size of an earthquake. Magnitude 9 earthquakes have 30 times more energy than magnitude 8 earthquakes, which saturates the early warning system and tends to make it impossible to estimate the magnitude for larger earthquakes.
"When our research team developed a deep learning AI model using the information carried by PEGS, we were pleasantly surprised by the success of the first prototype to classify earthquakes according to magnitude," said Rouet-Leduc.
The research team was then able to demonstrate the power of their deep learning model to instantly track earthquakes in real time after they reach a certain size. Although the algorithm needs to be further tested on raw data, scientists believe the results could improve seismic and tsunami warning systems.
"Our new model may help people to know quickly about the tsunami that will occur after a major earthquake," the author concludes.
This information appeared at the DPRI, Kyoto University; and Kyoto University websites.
