The study shows that GPS can predict earthquakes two hours in advance, but the accuracy of the sensor needs to be improved by 100 times.

CTOnews.com July 24 news, the British "New Scientist" Weekly recently reported that the latest research by Contan Breitley and Jean-Mathieu Nokai of the University of the Azure Coast in Nice, France, shows that theoretically GPS can predict earthquakes two hours in advance, but this requires the development of GPS sensors with 100 times the accuracy of the current level.

▲ map source Pixabay two scholars chose to use GPS data to identify the gradually accelerated relative slip between plates before the earthquake. These slips are too small to appear on the seismograph, but if they can be detected, they can predict when the earthquake will begin. Similar attempts have been made by scientists in the past, but previous studies have only analyzed a small number of earthquakes, and early warning signals have appeared in the absence of earthquakes, or early warning signals have been observed at an uncertain point in time before the earthquake.

The two scholars used GPS measurements collected over a five-minute period and made them accurate to less than 1cm, which were collected within 48 hours of 90 independent earthquakes. They built a joint data set with more than 3000 measurements and compared the recorded stratum movement with the expected direction of each point at the time of the earthquake.

The test results show that from about two hours before the earthquake, there has been a gradual, slow and otherwise imperceptible relative slip between geological plates-which could lead to a reliable seismic detector.

However, there is also a problem. Blattelli said that the current noise level of the GPS sensor means that this movement can only be detected based on the big data set, rather than relying on any point. Relying on a point detection requires GPS sensors to be able to detect movements as small as 0.1mm, he said.

CTOnews.com learned from the report that Breitri believes that current sensors cannot detect this movement in an independent earthquake and therefore cannot predict earthquakes. But this test shows that some activity does occur before an earthquake, and if people can make significant progress in measurement (improving sensor accuracy or deploying more sensors), it is possible to detect the activity and make predictions.