The Unexpected Physics Behind Turkey’s Devastating 2023 Earthquakes

A new study from researchers at UC San Diego’s Scripps Institution of Oceanography uses a multidisciplinary approach to unravel the complexities of the two deadly earthquakes of nearly equal strength that struck Turkey and Syria on Feb. 6, 2023.

The research, published today in the journal Science, finds that each of the two quakes, measured at magnitude 7.8 and 7.7, respectively, had unexpected elements that added up to make the shaking even more destructive.

“The earthquakes occurred on known faults, and in this sense were expected,” said Scripps geophysicist Yuri Fialko, the paper’s co-author. “What was unexpected was their size – they were much bigger than any known past earthquakes on the same faults. This happened because these earthquakes did lots of unexpected things they weren’t supposed to do.”

The first of the two earthquakes increased in strength because of an improbable “cascade” of ruptures that broke through various fault bends and junctions that are normally expected to act as barriers to the propagating rupture. The second shake also packed an extra punch because of an unexpected phenomenon called a supershear rupture, in which the fault ruptures faster than the seismic shear waves can travel through Earth’s crust, creating a sonic boom-like effect that amplifies the earthquake’s destructive power.

Fialko also said that by enumerating the strange and uncommon aspects of these earthquakes the findings emphasize the need to include the possibility of similarly rare or extreme events in earthquake preparedness plans around the world. In particular, he said incorporating rare but possible scenarios would be a valuable exercise in areas near the San Andreas Fault in California, which he said is a structurally similar fault system to the East Anatolian Fault system that produced these two earthquakes.

Like California, Turkey is earthquake country. It sits in the middle of a complex series of strike-slip faults where the Eurasian, Arabian, and Anatolian tectonic plates meet. The Arabian plate is moving north towards the Eurasian Plate, with the whole of Turkey and its Anatolian plate being squeezed between the two.

“The simplest analogy is to watermelon seeds,” said Fialko. “They’re slippery and if you squeeze one between your fingers it wants to shoot out in one direction. There is compression on both sides of the Anatolian plate and it’s trying to shoot out in a western direction like a watermelon seed between two fingers.”

Along these tectonic boundaries, stress can accumulate in the rocky upper layers of Earth’s crust along the fault if the two sides get stuck due to friction and clamping pressure as the underlying plates continue to slide some 20 km (12 miles) deep underground. Earthquakes happen when the accumulated stress finally overcomes the friction and clamping pressures acting on the upper crust and the two sides of the fault suddenly and violently slip to catch up with the tectonic movement occurring at depth.

This all means that the region surrounding Turkey has a long history of big, deadly earthquakes, from the one that destroyed the Syrian city of Aleppo in 1138 to the 1999 quake that struck the Turkish city of İzmit.

Researchers at Scripps began studying the 2023 earthquakes almost immediately after they occurred in hopes of better understanding the processes that gave rise to these devastating seismic events that claimed the lives of more than 50,000 people across Turkey and Syria. The National Science Foundation, NASA, the U.S. Geological Survey, the Cecil and Ida Green Foundation and the European Union’s Horizon 2020 Research and Innovation Programme all contributed to the funding of the study.

The most obviously uncommon feature of the quakes was their virtually identical magnitude, which did not require much analysis to recognize. The second shake, which occurred some nine hours after the first, was not technically an aftershock but rather the second act of what is known to seismologists as a doublet – two earthquakes of similar magnitude that occur within a relatively short period of time.

To learn more, the team combined three disciplines: geodetic observations collected via satellites, on-the-ground seismic recordings, and two types of computer modeled simulations to recreate the earthquakes.

The geodesy component of the study used satellites to measure changes in Earth’s surface that could quantify slippage, allowing researchers to make inferences about the geologic changes that occurred beneath the surface during the earthquake. The seismic measurements came from hundreds of seismic instruments positioned in the immediate vicinity of the earthquakes and across the globe that recorded the seismic waves rippling through the Earth. The seismic aspect of the study showed not just the strength of the earthquakes but also how they progressed through time.

Finally, the team combined these observations with two types of computer modeling to attempt to tease apart how the two earthquakes happened and what caused them. The first type of model was what’s called a kinematic model. Essentially, the researchers added their observational data to the kinematic model to map how the ruptures moved along the faults with the goal of creating as detailed a picture of what happened underneath Earth’s surface as possible.

From this kinematic model of the fault’s rupture, the team was able to reverse engineer the configuration of initial stresses that caused the earthquakes. With these initial stresses in hand, the researchers then turned to what’s known as a 3D dynamic or physics-based model.

This type of computer model is more like running a 3D simulation. The team inputs the initial stresses and then, with the help of supercomputers running for several days in some cases, the model plays out the scenario based on complex equations that represent our knowledge of the governing physics for earthquakes.

“In this iterative way we can generate rupture scenarios that are consistent with our observations as well as mechanically consistent with what we know about friction and the dynamic stresses,” said Fialko. “This can help us understand the mechanisms behind the earthquake and recreate what happened.”

These multiple lines of analysis revealed that the initial magnitude 7.8 earthquake started as a magnitude 6.8 in a subsidiary of the East Anatolian Fault known as the Nurdağı-Pazarcık Fault and then spread to the main East Anatolian Fault. In addition to breaking through this junction with the East Anatolian Fault, the rupture broke through three other kinks and branching points in the fault, barriers that are typically assumed to put a stop to propagating ruptures. Breaking through these four fault barriers was part of what allowed the earthquake to grow in magnitude from 6.8 to 7.8.

This first quake unzipped the East Anatolian Fault’s clamped rocks in both directions for a total rupture length of roughly 300 kilometers (186 miles) – the longest known rupture along the fault. The fact that the rupture propagated in both directions when it reached the East Anatolian Fault was also unexpected, as only one of the directions was favored mechanically, said Fialko.

According to Zhe Jia, a postdoctoral scholar at Scripps and the study’s lead author, the magnitude 7.8 event caused static and dynamic stress perturbations in the fault system, including slippage of up to 8 meters (26 feet) in some places, that may have caused the second earthquake nine hours later.

Jia noted that the faults in a complex system like this one are always in conversation with one another.

“The release of stress in one section of the fault system impacts nearby sections, potentially adding stress or reducing clamping pressures in ways that make a rupture more likely,” said Jia.

The current study, however, can’t yet fully explain why the second magnitude 7.7 quake took nine hours to begin.

The team found that the second earthquake occurred on the Savrun-Çardak Fault, with a total rupture length of only around 150 kilometers (93 miles) going east to west. Though the rupture length was only about half that of the first earthquake, the second earthquake almost managed to equal the magnitude of the first because it featured more slippage (up to 10 meters or 32 feet) and because it featured a supershear rupture.

In supershear earthquakes, rupture progresses faster than the seismic shear waves can travel through Earth’s crust.

“The concept is similar to a sonic boom, in which an object travels faster than the speed of sound,” said Jia. “When supershear rupture speeds occur, the amplitude of seismic waves is higher and they keep their energy over larger distances, which can cause larger ground deformations and make ground shaking more destructive.”

Supershear ruptures are statistically uncommon and are even more uncommon on faults like the Savrun-Çardak, which is jagged and considered geologically immature because it has not yet accumulated much slippage from prior seismic activity. Generally, immature faults rupture more slowly than mature faults, making supershear events less likely.

The results highlight the advantages of combining multiple lines of observation with data-driven and physics-based computer models when it comes to understanding the complex dynamics that go into major earthquakes, said Fialko.

He said the study also highlights the necessity of planning for rare scenarios despite their more remote possibility.

“We can’t rely exclusively on what we see in the paleoseismic record when we assess seismic hazards,” said Fialko. “Past is prologue, but only to an extent. There is always a chance that the next earthquake is going to be bigger than what came before.”

This lesson has relevance to California, which Fialko said might do well to consider the possibility of the San Andreas Fault rupturing in statistically unlikely ways, especially given its structural similarities to the East Anatolian Fault system.

Scripps graduate students Zeyu Jin, Xiaoyu Zou, and John Rekoske, and faculty members Alice-Agnes Gabriel, Wenyuan Fan, and Peter Shearer contributed to the study. Mathilde Marchandon and Thomas Ulrich of Ludwig-Maximilians University, as well as Fatih Bulut and Asli Garagon of Bogazici University were also co-authors.