Article Highlight | 16-Nov-2026

The Marmara Seabed reveals 2,500 years of major earthquakes

Istanbul Technical University

Istanbul, Türkiye – For centuries, significant earthquakes beneath the Marmara Sea have been recorded not only in historical sources but also in sediment layers buried on the seafloor. By analyzing nine sediment cores from various parts of the Marmara Sea alongside historical and geological earthquake data, researchers have reconstructed the impact of major earthquakes on different segments of the Main Marmara Fault over roughly 2,500 years.

 

The overall picture indicates that large earthquakes often involved multiple fault segments simultaneously, rather than breaking only one section at a time. This reconstruction also clarifies which parts were active during specific historical earthquakes and, together with long-term recurrence data, has identified the Central High segment southwest of Istanbul as currently the most earthquake-prone part of the fault. However, this does not predict the timing or location of future earthquakes. The study, led by researchers at Istanbul Technical University (ITU), was published in Marine Geology.

 

Connecting earthquake records across an entire fault

The Main Marmara Fault is the most active northern segment of the North Anatolian Fault beneath the Marmara Sea. Various parts of it have caused destructive earthquakes throughout history. Previously, scientists used underwater sediment deposits to study past Marmara earthquakes, but many studies focused on specific basins or fault sections.

 

The recent research combines records from multiple areas to explore how earthquakes have propagated across the fault system over time. The team analyzed cores from the Çınarcık Basin, Central High, Central Basin, Western High, and Tekirdağ Basin, integrating these with earlier palaeoseismological data.

 

Prof. M. Namık Çağatay of the Applied Research Centre and the Geological Engineering Department at Istanbul Technical University explains, “One sediment core gives us one piece of the earthquake story. Connecting records from across the Marmara Sea reveals which fault sections ruptured together and how these rupture patterns evolved over centuries.”

 

How the seabed records an earthquake

Strong earthquake shaking can destabilize sediments on underwater slopes, causing sand, silt, and mud to flow downslope into deeper basins. These deposits can be buried by later sediment and preserved for thousands of years. The researchers identified these earthquake-related layers through X-ray imaging, physical property measurements, elemental analyses, and grain-size evaluations. Radiocarbon dating and short-lived radioactive isotopes helped determine the timing of individual deposits.

 

The Marmara Sea is ideal for this approach because its deep basins accumulate sediment quickly enough for individual events to be distinguishable. It also has a long documented earthquake history, allowing comparisons between dated sediment layers and approximately 2,500 years of recorded earthquakes. Instead of analyzing each deposit separately, the researchers examined where matching earthquake records appeared or were missing across different basins. This geographical pattern offers clues about the extent of individual earthquake ruptures.

 

Following ancient ruptures across the Marmara Sea

The reconstruction shows that several significant historical earthquakes left sedimentary evidence across multiple sections of the fault. For example, deposits linked to the major earthquakes of 1766 are found over a wide area of the Marmara Sea. Sedimentary records from the 1509 and 740 earthquakes stretch from the İzmit Gulf in the east to the Central Basin farther west.

 

Different earthquakes display varying geographical patterns. Overall, the distribution of these deposits indicates that large earthquakes often involved ruptures across multiple segments. This also helps researchers identify where ruptures may have halted. For example, the 1912 Mw 7.4 Şarköy–Mürefte earthquake is recorded in the sediments of the western Marmara region but not in the Central Basin, suggesting the rupture did not extend east beyond the western end of the Western High. In this way, the seafloor offers more than a record of past earthquakes; it enables researchers to reconstruct their spatial footprints along the fault.


“If the same earthquake appears in sediment records from several parts of the sea, we can begin to trace how widely that event affected the fault. The pattern of where an earthquake is recorded and where it is not helps us reconstruct the history of individual ruptures,” explained Prof. M. Namık Çağatay.

 

A fault that does not follow a clock

The extended sedimentary record reveals how often major earthquakes have struck different parts of the fault. Average recurrence intervals, estimated at about 220–300 years for various Main Marmara Fault segments, align broadly with geological slip rates and GPS data. However, these averages can be misleading if seen as exact schedules. In reality, the time between earthquake-related deposits varies significantly, ranging from a few decades to several centuries. This study underscores that an average recurrence interval does not predict the timing of the next earthquake. The Western Marmara segment exemplifies the complexity of long-term fault behavior. While it currently exhibits significant creeping, earlier sedimentary evidence suggests it experienced periods of more frequent large earthquakes, implying alternating episodes of locking and creeping over time.

 

Why one segment southwest of Istanbul stands out

The comprehensive reconstruction of the fault is especially vital because much of the submerged Main Marmara Fault runs between regions affected by two major earthquakes from the twentieth century: the 1912 Mw 7.4 Şarköy–Mürefte earthquake in the west and the 1999 Mw 7.4 İzmit earthquake in the east.

 

Sedimentary records show that recent large earthquakes (magnitude 7 or higher) in different sections occurred in 1999 in the İzmit Gulf, 1894 along the Prince Islands segment, 1766 in the Kumburgaz and Central basins, and both 1912 and 1766 along the Western Marmara segment.

 

By combining these observations with longer earthquake histories of these segments, researchers found that the Central High segment, southwest of Istanbul, is currently the most earthquake-prone part of the Main Marmara Fault.

 

The authors emphasize, however, that palaeoseismic records cannot predict the exact timing of future earthquakes. Instead, these records provide long-term evidence useful for probabilistic seismic hazard assessments.

 

Giving the fault a longer memory

Modern instruments have only monitored earthquakes for a small fraction of the active fault's lifetime. While historical records extend this view further into the past, they eventually become incomplete. The seabed offers another valuable archive. By linking earthquake layers found across different areas of the Marmara Sea, researchers have reconstructed not just when major earthquakes occurred, but also how various sections of the fault have repeatedly ruptured together. This distinction is important: sediment layers cannot predict when the next big earthquake will happen. However, they reveal which fault segments have repeatedly broken, which sections have ruptured together, and how irregularly major earthquakes have returned. For communities around the Marmara Sea, this extended geological memory can help scientists develop a more accurate understanding of the fault's behavior, and the seismic risk it continues to pose.


“The seabed gives the fault a memory much longer than modern instruments can provide. It cannot tell us when the next earthquake will happen, but it helps us understand how this fault has behaved, and what it has been capable of over centuries and millennia,” said Prof. M. Namık Çağatay.

 

Disclaimer: AAAS and EurekAlert! are not responsible for the accuracy of news releases posted to EurekAlert! by contributing institutions or for the use of any information through the EurekAlert system.