Repeated earthquakes: scientists recognize their limits

Is Mexico City doomed to suffer an earthquake every September 19? In the streets of the Mexican capital, some are already thinking about it as this date remains associated with natural disasters. Three times in the past (in 1985, 2017 and 2022) this fateful day was the scene of strong earthquakes, destroying homes and claiming lives. The probability of a third earthquake on September 19 was 0.000751%, according to scientist José Luis Mateos quoted by Mexican newspapers. So why not a fourth episode? This Thursday, September 22, a new earthquake of magnitude 6.9 hit the country, killing two people.

In their research laboratories, geophysicists prefer to rely on mathematical models rather than a calendar to predict the next disasters. However, everyone recognizes it: guessing precisely when an earthquake will occur is like an insoluble puzzle.

“We would like to have earthquake weather forecasts. Unfortunately, we are far from it,” says Romain Jolivet, university professor in the Geosciences department of the École Normale Supérieure in Paris. Of course, our knowledge is progressing. After the earthquake that caused the Fukushima nuclear accident in 2011, the discipline was called into question. Our understanding of what is physically happening in a fault, as well as our methods of establishing risk have improved.

Limited offer. 2 months for 1€ without commitment

Despite everything, scientists come up against an insurmountable barrier so far: the opacity of the Earth. “Unlike the atmosphere, whose composition and equations we know perfectly, the Earth’s mantle cannot be observed directly. As a result, we do not know it well”, explains Romain Jolivet. In recent years, research on slow earthquakes still gives hope. These events have the particularity of taking place in slow motion: instead of causing a slide of one meter per second, they move the land by one centimeter per year, sometimes with accelerations over a few days. Too slow to generate seismic waves, they are tracked using GPS stations.

The enigma of tremors and repetitive earthquakes

However, these observations have shown that slow earthquakes sometimes precede certain earthquakes: for example, that of Tohoku in Japan in 2011, or that of Iquique in Chile in 2014. The scientists also noticed that they s accompanied other phenomena such as “tremors” – tremors with a particular profile whose origin remains disputed – or repetitive earthquakes whose signatures, perfectly identical, are repeated at intervals of several years.

All this is not yet enough to be able to predict the arrival of a disaster to the day. “Even if we are sometimes able to say a posteriori that certain events were connected, what we observe before an earthquake also occurs at other times. We cannot therefore draw a law from it”, regrets Romain Jolivet. To try to see more clearly, the researcher accumulates data on slow earthquakes in order to see which ones are followed – or not – by a normal earthquake. Other New Zealand and Japanese scientists are working on the subject. Pending the outcome of their work, traditional risk assessment models remain our most valuable source of information.

They say there is a 70% chance of a magnitude 7.3 earthquake hitting greater Tokyo before 2050. Further south in New Zealand, research suggests an event of magnitude 8 or greater would have 26% chance of happening. The situation also seems tense on the side of the San Andreas fault in California. This zone accumulates the equivalent of 3 cm of landslide per year. Since the last earthquake in 1857, this represents 5 or 6 meters or roughly the equivalent of the last large landslide observed in the region.

L’application L’Express

To track analysis and decryption wherever you are

Download the app

Although very useful, these indications still lack precision. The monitoring of slow earthquakes over time or even the use of new, very sensitive measuring instruments such as quantum sensors, will no doubt make it possible to refine the diagnoses. But this will take time. “In the future, we will probably not be able to predict to the day the arrival of an earthquake of magnitude 8 in such and such a region. But it may be possible to identify periods of time more risky than d ‘others’, hopes Romain Jolivet. A necessity. About one hundred thousand earthquakes are recorded each year on the planet. And the ten largest of them have already killed two and a half million people.