Summary

The Marmara Fault, located immediately south of the megacity of Istanbul, poses the highest earthquake risk in the entire pan-European region. Until now, seismic hazard assessments have been based mainly on the concept of a single, continuous main fault running beneath the Sea of Marmara. Recent findings, however, point to a more complex network of interacting primary and secondary faults, each with distinct characteristics. According to Prof. Marco Bohnhoff of the GFZ Helmholtz Centre for Geosciences in Potsdam, Germany—who has been studying the region for decades—this necessitates a reassessment of the seismic hazard for the Istanbul area, likely resulting in a higher projected maximum earthquake magnitude than previously estimated. In a “Viewpoint” piece recently published in the Geophysical Journal International, Bohnhoff explains that the sequential activation of various segments within these fault zones could, in an extreme scenario, trigger a major earthquake with a magnitude of 7.6. Such an event would release up to twice as much energy as the extreme scenarios previously calculated, which estimated magnitudes in the 7.2–7.4 range. Consequently, Bohnhoff calls for improved monitoring directly along and within the fault zone beneath the Sea of Marmara.

Complexity of the Marmara Fault near Istanbul

Along the tectonic plate boundary immediately south of Istanbul—widely referred to as the “Marmara Fault”—the Eurasian and Anatolian tectonic plates slide past one another. Friction and the interlocking of these plates periodically trigger earthquakes, some of which can be extremely powerful. Contrary to earlier descriptions, the Marmara Fault should not be viewed as a single, continuous fault line. Instead, it consists of a complex network of faults with numerous branches, each possessing distinct characteristics. Recent geological, geophysical, and geodetic findings point to this reality.

The fact that this fault system lies beneath the seabed—and has therefore remained largely inaccessible to direct seismic monitoring—makes it exceptionally difficult to assess its structure and the resulting hazard potential for the region.

However, the mosaic of faults beneath the Sea of Marmara bears similarities to the San Andreas Fault in Southern California and the San Francisco Bay Area. These structures can serve as an analogue; because they are located on land—unlike the submarine Marmara Fault—they can be directly observed and measured. This provides a basis for drawing inferences about the Istanbul-Marmara region. Consequently, such comparisons have a direct impact on the modelling of potential earthquake scenarios in the Marmara region, thereby influencing the assessment of local seismic hazards and the overall seismic risk for Istanbul. 

This is the conclusion reached by Prof. Dr Marco Bohnhoff, who has been conducting research in the region for decades. At the GFZ Helmholtz Centre for Geosciences, Bohnhoff heads the “Geomechanics and Scientific Drilling” section. He is also the initiator and project leader of GONAF (Geophysical Borehole Observatory at the North Anatolian Fault)—a borehole-based earthquake observatory established for the Istanbul region in 2015, which the GFZ operates in collaboration with the Turkish disaster management agency, AFAD. GONAF now comprises ten sites located near the coast around the Sea of Marmara.

Re-evaluation of the fault system off Istanbul

GFZ researcher Marco Bohnhoff states: “For Istanbul, we must also consider earthquake scenarios that are more complex than those we have examined thus far. To best assess the resulting hazard, we need seismic monitoring equipment installed directly on and within the fault zone beneath the Sea of Marmara.”

Consequences of the re-evaluation

This re-evaluation of the fault system near Istanbul—now understood as a complex network of faults—has immediate implications for how earthquake risk is assessed. It necessitates direct monitoring of the fault zone on the seabed and beneath it—positioned right at and within the tectonic plate boundary.

According to Bohnhoff, the energy accumulated along the main fault remains a concern, with a 50% probability of an earthquake with a magnitude of 7.2 to 7.4 occurring in the coming decades. However, because the system includes complex, tectonically charged secondary branches—both near and further away from the main fault—there is a possibility that these branches could also be activated during a major earthquake. In an extreme scenario, this could push the estimated maximum magnitude as high as 7.6. This scenario also encompasses the section of the plate boundary west of the Sea of Marmara. This area was last activated in a 7.4-magnitude earthquake in 1912; it is now completely locked and has once again accumulated a significant amount of tectonic strain. For comparison, an earthquake with a magnitude of 7.6 releases approximately twice as much energy as one with a magnitude of 7.4. 

"The important point is that we cannot narrow down a time frame for when such an earthquake might occur—and also no specific point in time," says Bohnhoff. 

Historical records for the Marmara region over the past 2,300 years show no earthquakes with a magnitude greater than 7.4. However, as part of so-called “super-cycles,” much stronger earthquakes can occur at long intervals. Recent comparable scenarios include the 2011 Tohoku-Oki earthquake in Japan, which reached a magnitude of 9—a level previously thought impossible. This triggered a tsunami with unexpectedly high waves, leading to the well-known, dramatic consequences for the Fukushima nuclear reactor. Similarly, a highly complex earthquake struck near Kaikoura, New Zealand, in 2016. In that instance, a network of several secondary branches of a plate-boundary fault was activated, resulting in an unexpectedly high magnitude of 7.8—fortunately in a sparsely populated region.

System complexity complicates probabilistic forecasting

GFZ researcher Bohnhoff recently presented his reassessment of earthquake hazard in the greater Istanbul area at international conferences and has now published it as a scientific commentary in the journal “Geophysical Journal International”.

According to his findings, the complexity of the system makes probabilistic forecasting even more difficult than it already is. The reason: “Sooner or later, stress at some point within the network of geological faults will lead to rock failure. This can trigger isolated, local earthquakes of moderate magnitude. However, the energy could also propagate—potentially activating the entire system and releasing a significantly greater amount of energy,” explains Bohnhoff. This scenario could result in the afore mentioned earthquake, with a magnitude exceeding 7.4.

The need for measurements taken directly within the fault zone

An earthquake with a magnitude between 7 and 8 would, in all likelihood, have catastrophic consequences for the metropolis of Istanbul, which lies only a few tens of kilometres from the main Marmara fault. For this reason, Marco Bohnhoff advocates for more targeted seismic monitoring of the fault—specifically at the locations where the tectonic plates have become locked. “We need to deploy our measuring instruments and drill holes beneath the sea, directly at—or even into—the fault zone itself,” Bohnhoff says.

Background: The Marmara fault

The Marmara Fault poses the greatest seismic hazard and risk in Europe and its surrounding regions. It is a segment of the North Anatolian Fault Zone, where the Eurasian and Anatolian tectonic plates meet and slide past one another. Friction and locking occur along this boundary, periodically resulting in earthquakes. The Marmara Fault is the only section of this entire plate boundary that has not experienced a major earthquake (magnitude M > 7) since 1766.

The average recurrence interval for such large earthquakes in this region is approximately 250 years—a figure derived from historical records of tremors spanning the last 2,300 years, which allow for the estimation of earthquake magnitudes. Consequently, the main branch of the Marmara Fault is currently in the late stage of its seismic cycle, carrying the potential for a major earthquake with a magnitude of up to approximately 7.4.

In other words, while the tectonic plates have shifted by about 2.5 meters in other parts of the Marmara Fault system, no such movement has occurred off the coast of Istanbul; as a result, energy has accumulated there. This process will inevitably lead to rock failure, ultimately triggering an earthquake.

Previous studies led by researchers at the GFZ Helmholtz Centre for Geosciences have shown that the Marmara Fault is divided into several distinct segments. In the area west of Istanbul—covering the central and eastern parts of the Sea of Marmara—the degree of locking is low, and the tectonic plates partially "creep" past each other. In this zone, only a portion of the tectonic energy—roughly half—is stored. This was deduced from the detection and analysis of specific, small-scale earthquakes known as "repeaters." Beneath the eastern Sea of Marmara, south of Istanbul, the creep component is zero; the tectonic plates there are completely locked, leading to a maximum accumulation of tectonic energy.

Consequently, stress builds up to varying degrees along the individual segments of the Marmara Fault as a result of this tectonic loading process. Furthermore, relatively larger earthquakes trigger co-seismic and post-seismic stress redistributions—referred to as Coulomb stress—which can either increase or decrease the state of stress, depending on the type of rupture and its location.

Original publication:

Marco Bohnhoff, Rethinking the Main Marmara Fault—seismotectonic complexity and seismic hazard in the Istanbul–Marmara region, Geophysical Journal International, Volume 246, Issue 3, September 2026, ggag253, https://doi.org/10.1093/gji/ggag253

Previous press releases about GFZ studies on the Marmara region:

https://www.gfz.de/en/press/news/details/erdbeben-und-seismische-energieausbreitung-entlang-der-marmara-stoerungszone-wandern-ostwaerts-richtung-istanbul

https://www.gfz.de/en/sektion/erdbebengefaehrdung-und-dynamische-risiken/ueberblick/detailseite-sektionsnews/bruchmuster-von-erdbeben-verbessern-die-bewertung-seismischer-gefaehrdung

https://www.gfz.de/en/press/news/details/kriechen-und-verhaken-vor-istanbul-systematisches-bild-der-plattengrenze-im-marmara-meer