An algorithm to detect an earthquake as quickly as possible

The Earth is shaken every day by many earthquakes, more or less powerful. While most are not even felt by the population, some can be very destructive and deadly, especially when they generate tsunami waves.

However, predicting major earthquakes remains, for the time being, in the realm of fiction. Today, the fight against seismic risk is mainly based on the early detection of earthquakes, and the implementation of warning systems. And in this context, time matters a lot. For example, when of the Sumatra earthquake in 2004, the devastating wave arrived on the Indonesian coast only a few minutes after the shock. This tragic case, which caused the disappearance of more than 250,000 people, shows how crucial it is to develop methods to characterize as quickly as possible the magnitude of an earthquake and issue an alert if necessary.

minutes that matter

Currently, early warning systems based on the reception and analysis of seismic waves do not make it possible to estimate precisely and quickly, in a few seconds or minutes, the magnitude great earthquakes. However, this is essential data that allows us to estimate the risk of a tsunami. The problem is that seismic waves travel at a certain speed underground (6 km/s for the fastest) and that there is therefore a latency between the occurrence of an earthquake and its detection by seismological stations. It is then necessary to analyze the received signal to estimate fracture characteristics. Even if this process only takes a few minutes, it is a waste of time for the announcement of a potential alert.

Having access to the magnitude of an earthquake in real time, when it occurs, is therefore of paramount importance for the backup populations located in high-risk areas. Some studies have thus focused on detection methods other than those using seismic waves. Recently, scientists have thus discovered that large earthquakes can produce disturbances in the gravity earthly. Indeed, an earthquake is characterized by the sudden displacement of large masses rocky. This movement inevitably induces a variation in density within the subsoil, which, in turn, modifies the earth’s gravity field. This disturbance of gravity occurs simultaneously with the earthquake and its signal, which propagates to the speed of light can, in theory, be detected instantly.

An elastogravity signal that arrives before the seismic waves

In theory, because there is currently no equipment capable of directly detecting these tiny variations in gravity. While waiting for them to be developed, researchers from Géoazur, in Sophia Antipolis (in the Alpes-Maritimes), have developed a system ofartificial intelligence to process the elastogravity signals received by the seismometers already in place. These instruments, which normally record ground movements, are indeed also capable of measuring gravity disturbances combined with the elastic response of the ground. However, detecting these signals is very difficult. The amplitude of the gravity disturbance is in fact very low and the associated elastic response tends to erase the traces of this signal on the recording of the seismometers. However, this signal arrives well and truly before P waves, which are the fastest. A time saving that could go up to a few tens of seconds.

Obtain the magnitude of an earthquake in real time and immediately issue a tsunami warning

The second advantage of this method is that these signals cannot saturate the seismometers and therefore produce complete information, which makes it possible to estimate the magnitude of an earthquake precisely. This is not always the case for seismic P waves, which tend to saturate the recordings of seismometers located near theepicenter large earthquakes. The analysis of the elastogravity signal from a single source would therefore make it possible to immediately have a precise and reliable value of the magnitude and the mechanism of the fault the origin of the earthquake.

In an article published in the journal Nature, the researchers from Géoazur show how this signal, until now considered as noise on the seismograms, can make it possible to detect an earthquake in real time. Their method is based on a new algorithm that allows live analysis of this elastogravity signal to immediately obtain a magnitude value and issue a tsunami warning at the same time.