Massive earthquakes do not just move the earth – they make adjustments to the speed of light in the Earth’s gravitational field. Now researchers have trained computers to identify these small gravitational signalsdemonstrating how signals can be used to mark the location and magnitude of a strong earthquake almost instantly.

This is the first step towards creating a very early warning system for the most powerful earthquakes on the planet, scientists said on May 11 in nature.

Such a system could help solve a difficult problem in seismology: how to quickly determine the true magnitude of a massive earthquake as soon as it occurs, says Andrea Liquiardi, a geophysicist at the Université Côte d’Azur in Nice, France. Without this capability, it is much more difficult to issue life-saving warnings quickly and effectively.

As large earthquakes are destroyed, tremors and tremors send seismic waves through the earth that look like large shakes of seismometers. But current detection methods based on seismic waves unfortunately have difficulty distinguishing, say, between a 7.5 and 9 magnitude earthquake within seconds of such an event.

This is because the initial magnitude estimates are based on the height of the seismic waves, called P waves, which are the first to arrive at the monitoring stations. However, for the strongest earthquakes, these initial amplitudes of the P wave are maximized, which makes earthquakes of different magnitudes difficult to distinguish.

But seismic waves are not the earliest signs of an earthquake. This whole mass, which moves around in a big earthquake, also changes the density of the rocks in different places. These changes in density translate into small changes in the Earth’s gravitational field, producing “elastogravity” waves that travel through the earth at the speed of light – even faster than seismic waves.

Such signals were once thought to be too small to detect, said seismologist Martin Vale of the Institut de Physique du Globe de Paris, who was not involved in the new study. Then in 2017, Vale and his colleagues were the first to report seeing these elastogravity signals in the seismic station data. These findings have shown that you have a window between the beginning of the earthquake and the moment you receive it [seismic] waves, ”says Vale.

But researchers were still pondering how to turn these elastogravity signals into an effective early warning system. Because gravitational movements are small, they are difficult to distinguish from background noise in seismic data. When scientists looked back, they found that only six mega-earthquakes in the last 30 years had generated recognizable elastogravity signals, including a magnitude 9 Tohoku-Oki earthquake in 2011 that triggered a devastating tsunami that flooded two nuclear power plants. in Fukushima, Japan (SN: 16.03.11). (The initial estimate of the magnitude of this earthquake, based on AP waves, was 7.9.)

Computers can come here, Likiardi says. He and his colleagues created PEGSNet, a machine learning network designed to identify “fast ElastoGravity signals”. The researchers trained the machines on a combination of real seismic data collected in Japan and 500,000 simulated gravitational signals for earthquakes in the same region. Synthetic gravity data is essential for training, Likiardi says, because real data is so scarce and the machine learning model requires enough input to find models in the data.

After being trained, the computers received a test: Follow the origins and evolution of the 2011 Tohoku earthquake as if it were happening in real time. The result was promising, Likiardi said. The algorithm was able to identify both the magnitude and the location of the earthquake five to 10 seconds earlier than other methods.

This study is proof of concept and we hope it is the basis for a prototype early warning system, says Likiardi. “He is currently fit for work … in Japan. We want to build something that can work in other areas known for powerful earthquakes, including Chile and Alaska. Ultimately, the hope is to build a system that works globally.

The results show that PEGSNet has the potential to be a powerful tool for early earthquake warnings, especially when used in conjunction with other earthquake detection tools, Vallée said.

However, more work needs to be done. On the one hand, the algorithm is trained to look for a point of origin of the earthquake, which is a reasonable approximation if you are far away. But in the near future, the origin of the earthquake no longer seems to be a point, but is in fact a larger region that has been torn apart. If scientists want an accurate estimate of where a rupture has occurred in the future, the machines must look for regions, not points, Vale added.

More progress may come in the future, as researchers are developing much more sensitive instruments that can detect even minor earthquake disturbances in the Earth’s gravitational field while filtering out other sources of background noise that can hide signals. The earth, says Vale, is a very noisy environment, from the oceans to the atmosphere.

“It’s a bit like the challenge physicists face when trying to observe gravitational waves,” says Vale. These space-time waves caused by colossal cosmic collisions are a very different type of gravitational wave (SN: 2/11/16). But the signals of gravitational waves are also blurred by the Earth’s noise – in this case microtremors in the earth.

Machine learning and gravity signals could rapidly detect big earthquakes

Previous articleMusk, ever the erratic visionary, revels in contradictions
Next article2022 Display Week: Samsung, LG Introduce Future Foldable Screens in California