Earthquake forecast: Predicting earthquakes: signals from the subsurface

This year, severe earthquakes struck Morocco, Turkey, the South Sea island of Vanuatu, Japan, Afghanistan and many other places around the world. Most recently, at the beginning of November, two severe earthquakes within three days brought death and destruction in the Jajarkot district in western Nepal.

Quakes occur wherever earth plates move, scrape past each other, get caught up and release the pent-up energy with enormous force at intervals of decades, centuries or even millennia. Germany is not immune to this either, as the, albeit mild, earthquakes between Aachen and Cologne or in the Upper Rhine Graben repeatedly show.

Seismologist Roger Bilham warned the “Nepali Times” that the severe earthquakes that shook Kathmandu in 2015, like the quakes in Jajarkot, could have been precursors to an overdue major earthquake. According to Bilham, enormous tectonic energy has accumulated in the seismic fault zone in western Nepal that has not been released for 500 years.

Nobody knows whether the big earthquake will hit Nepal tomorrow, next year or many years from now. Seismologists have still not found any means of predicting earthquakes, even though people around the world are keenly observing, measuring, experimenting and researching them.

Experts from the Georesearch Center (GFZ) in Potsdam have now presented data on seismic signals in a new study that could improve the forecast for some large earthquakes. To this end, study leader Patricia Martínez-Garzón and her colleagues from the USA and Turkey carried out extensive measurements in the Eastern Anatolian Fault Zone, where a double earthquake with magnitudes of 7.8 and 7.6 on the Richter scale in February 2023 claimed more than 60,000 lives has. 300,000 buildings were destroyed or damaged, with financial damage amounting to around $120 billion.

Premonitions of the earthquake in Turkey

According to the scientists found in the journal at the end of November »Nature Communications« published study found that since about eight months before the earthquake in Turkey, there was “an acceleration in seismic event rates and greater energy release in clusters within 65 kilometers of the epicenter.” Such a spatiotemporal focusing of seismicity was already known from controlled rock deformation experiments in the laboratory and was observed before the rock samples fractured. The same phenomenon has been observed in several, but by far not all, large earthquakes along continental fault zones in recent decades.

Measuring and identifying seismic signals and analyzing them using modern data processing techniques and machine learning is technically possible. The big question for both scientists and people in the earthquake regions concerns the specific predictive power of such data. »This is currently one of the biggest challenges: determining the potential of such signals to detect the potential of impending earthquakes. But we’re still not there yet,” Martínez-Garzón freely admits to “nd”.

The extent of destruction and the number of deaths depend on both the severity and type of earthquake, as shown, for example, by the triple disaster on the Indonesian island of Sulawesi in 2018: a magnitude 7.5 earthquake triggered a tsunami and in the Result in soil liquefaction. The ground became soft and wobbly like pudding, devouring entire villages like a greedy monster. The balance of the natural dramas: There were 4,300 deaths and 170,000 people lost their homes, fishing boats, businesses and thus their livelihoods.

Avoid deaths through safe construction methods

Another factor affecting the extent of earthquake damage is the construction of buildings. Old brick houses as well as many of the historic Hindu temples in the Nepalese capital Kathmandu collapsed like houses of cards in 2015. Modern buildings made of concrete, which were built according to earthquake safety rules, had cracks in the masonry and certainly suffered structural damage here and there. But they had not collapsed and had not buried people under the rubble.

“The Big One” is not only overdue in Nepal, but also in the Istanbul region with its around 20 million inhabitants. Classic seismic forecasts estimate the probability of such an earthquake of magnitude 7 or more by 2034 at 53 percent, with an error probability of between plus and minus 18 percent, explains Martínez-Garzón.

Short-term predictions as a goal

In the region, the GFZ operates the borehole-based geophysical observatory on the Anatolian Fault (GONAF). This has the task of “further reducing the observation gap between controllable laboratory experiments and uncontrollable natural earthquakes that are very dangerous for humanity and of better understanding the processes that – on a time scale of months – lead to large earthquakes,” says Marco Bohnhoff, Head of the GFZ section “Geomechanics and Scientific Drilling”. However, Bohnhoff warns against excessive expectations: “The short-term prediction of such events is currently still not possible and remains a long-term goal of our seismological research.”