Predicting earthquakes…a French faith turned science

After the Turkish earthquake last February, the scientific community specializing in earthquakes responded to the man’s connection with the appearance of Dutch earthquake forecaster Frank Hoogrebets, who claimed to have predicted the quake days before it happened. The movement of celestial bodies and earthquakes, he said, “unscientific talk, close to fraud,” and asked, “Have any scientific studies been published by this man?” They refused and asked.

This time the situation seems to be different, as the prediction claims come from a scientific study by French researchers of Geology (the science that studies topics related to Earth’s size, shape, dimensions, interior, magnetic field and interior temperature) fame. The study, published last July in the journal “Science,” was well-known and highly influential, drawing a lot of attention immediately after it was published. It quickly disappeared after being criticized by senior researchers who questioned it. Its results emphasize that publishing it in a popular journal does not protect it from criticism.

A half-century-old thirst

Before the scientific community began to investigate this further, it seemed like “hope” to quench the thirst for more than half a century, because for that long time scientists were looking for the early signals that earthquakes occur in minutes. , the French study said in its introduction that the signal can be tracked through movements recorded by GPS data hours or days before the Earth begins to vibrate, up to two hours before a major earthquake strikes. Nothing else.

What is scientifically proven is that it is impossible to predict an earthquake because there is no early sign of an earthquake until the ground actually starts shaking, but the authors of the study; In their study, researchers Quentin Beltieri and Jean-Mathieu Noquet from the University of Côte d’Azur in Guizus, France, analyzed 90 previous earthquakes, indicating that large earthquakes begin with a preliminary phase of slip. The earthquake itself occurs, and this is enough time for the authorities to issue a warning in the affected areas.

How did they discover “sliding”?

During the study, the researchers analyzed a history of highly accurate GPS data that showed movement along the fault line 48 hours before 90 major earthquakes, and found a persistent signal that appeared long before an earthquake struck, giving authorities time to issue a warning. .

On average, the GPS stations’ horizontal motions accelerated significantly in the last two hours before the quake, in a direction consistent with slow fault slip near the final nucleation point, two French scientists say, suggesting signs of a major earthquake. 110 minutes after the earthquake.

To make sure their finding wasn’t a random event, they analyzed 100,000 48 hours of GPS data that didn’t end in major earthquakes.

Noisy data

A few hours after the paper was published, researchers rushed to look at it, and its citation rates reached 38 thousand within 30 days, a very large rate, but there was no confidence in its results. It is in the interest of the research community to look at it and evaluate it.

Since Judith Hubbard, a researcher specializing in geosciences at Harvard University, and her collaborator Bradley published an article on the “substock” website, most of the assessments were in a negative direction. The researchers, and in their paper, said they used the same data provided by the French researchers and reanalyzed it to confirm their findings. Hubbard and Bradley’s efforts showed no initial drift within two hours and no clear signal to indicate it. An earthquake was imminent, and instead, they found that the hypothesized signal was just “noise data.” Received from the Global Positioning System, this is not a sign of an imminent earthquake. Noise data is “nonsensical extra information”.

Hubbard reaffirmed in exclusive reports via email to Al-Jazeera Net that the signal attributed to the French study is nothing more than this kind of data, and that “prediction two hours or 5 minutes before an earthquake is very high. Useful, but doesn’t seem like it’s possible.”

Japanese denial

Aiko Tohada, a professor of geotechnical engineering at Japan’s Tokyo University, criticized the study from another point of view, and said in an exclusive emailed statement to Al-Jazeera that it “finds a signal to predict an earthquake.” occurs, and requires — if we consider the accuracy of the data reached by the researchers — monitoring.” Faults are precise candidates for earthquakes, and this is difficult to achieve on the ground.

Explaining what the Japanese professor went through, Sherif El-Hadi, head of the earthquake department at the National Institute for Astronomy and Geophysical Research in Egypt, said in telephone reports to Al-Jazeera Net, “to obtain the data requested by the French researchers.” It would require installing a vast network of GPS sensors along major faults that form the source of earthquakes beyond the economic capacity of most countries.

The “noisy data” Hubbard refers to is caused by the fact that current global positioning systems are not sensitive enough to detect warning signs when they occur, in addition to earthquakes that may come from unusual locations, he explained. In the 1992 earthquake in Egypt, it did not come from one of the famous faults, but from an unusual area called “Tahshur” in the Egyptian Giza Governorate.

French Defense

For his part, the French researcher, Quentin Peltieri, denied that his study and that of his colleague, Jean-Matthew Noquet, claimed to predict earthquakes, as its content implied.

“We don’t claim in the study that our approach can be used to predict earthquakes. Our study only aims to highlight that faults begin to slide before large earthquakes occur, and we had to collect recorded data to monitor the signal. Before large earthquakes, the direction of expected displacement is projected, which we know in advance. Can not.”

He explained that the American researcher had gone as far as to assume that removing the common part of the global positioning system’s time series (what we call the common mode) would lead to signal disappearance, and they explained that common situation as such. Noisy data, which is an oversimplification.

So, despite geologists searching for a reliable method for decades, we cannot currently predict large earthquakes.