The Mediterranean Sea faces a significant risk of large, earthquake-triggered tsunamis in the coming decades. While historical records document the geographical occurrence of these events, the specific faults responsible for the seafloor displacement generating the tsunamis often remain unknown. This research focuses on identifying the submarine rupture responsible for the significant tsunami generated by the Amorgos earthquake on July 9, 1956—the largest Mediterranean tsunami in the past two centuries. The study utilizes advanced submarine exploration techniques to map and analyze undersea fault scarps and identify the fault plane responsible for the event. This detailed investigation is crucial for improving our understanding of tsunami generation and for more accurate assessments of future seismic and tsunami hazards in the region. The ability to identify and characterize the source of historical tsunamis, even decades after the event, provides invaluable data for refining hazard models and informing effective mitigation strategies. This work directly addresses the gap in knowledge concerning the precise sources of past tsunamis, thereby contributing significantly to a more robust and informed understanding of future risks.

Previous studies on the 1956 Amorgos earthquake have yielded a range of magnitude estimates (7.2–7.8) and hypocentral depth variations (10–45 km), reflecting the challenges of analyzing historical seismic data. Several focal mechanisms have been proposed, pointing towards a NE–SW striking normal fault. The resulting tsunami caused significant damage and fatalities on surrounding islands, with run-ups reaching up to 20 meters—the highest recorded in the Mediterranean in the 20th and 21st centuries. To explain the variability in tsunami run-up heights, a secondary source involving submarine landsliding has been suggested, although this remains debated due to limited data on the earthquake's epicenter, tsunami arrival times, and coastal bathymetry. On-land investigations, including the analysis of historical aerial imagery and archeological remains, indicate coseismic uplift of the southern coast and long-term subsidence of the northern coast of Amorgos, further supporting the presence of a significant normal fault. However, these on-land observations alone are insufficient to pinpoint the precise source of the earthquake and tsunami.

This study employed a multi-faceted approach combining advanced marine geophysical data with detailed in-situ submarine observations. High-resolution bathymetric data, acquired during the AMORGOS-22 and AMORGOS-23 cruises using multibeam echosounders onboard the R/V Europe and the AUV Idef, provided detailed maps of the seafloor around the proposed epicenter of the 1956 earthquake. The 1m resolution bathymetry from the AUV and 10m resolution bathymetry from the ship surveys were processed to produce digital elevation models (DEMs) for detailed fault analysis. To directly image and characterize the fault surfaces, the hybrid remotely operated vehicle (HROV) Ariane was deployed during the AMORGOS-23 cruise. The HROV was equipped with a high-resolution 4K camera to acquire video imagery of the fault scarps. The HROV dives targeted areas exhibiting the steepest and most linear cumulative scarps, minimizing interference from mass-wasting deposits. The collected video frames were processed using structure-from-motion techniques (MATISSE 3D and 3DF ZEPHYR software) to create three-dimensional digital outcrop models (DOMs), providing high-resolution 3D reconstructions of the fault surfaces, thereby allowing for precise measurements of the fault offsets. These DOMs, with resolutions of approximately 1 cm, enabled detailed analysis of fault plane features, including the identification of striae and fault gouge, offering critical insights into the timing and extent of the rupture. The accuracy of offset measurements was verified using multiple processing pipelines and algorithms, minimizing any inherent uncertainties. Multiple processing methods were employed to build the 3D models from the HROV video data, ensuring robustness and accuracy of measurements. The approach integrated multiple data sources and analytical methods, providing a comprehensive and rigorous methodology for studying the submarine rupture.

The high-resolution bathymetric data revealed a >600 m high submarine scarp along the Amorgos fault. Detailed analysis of the HROV imagery identified a well-preserved fault mirror at the base of the scarp, exhibiting numerous striae and a gouge coating, indicative of recent slip. The fault mirror’s smoothness and presence of undisturbed layers of sediment immediately above indicated a recent exhumation event. Measurements of the offset along the fault mirror ranged from 9.8 to 16.8 meters (mean 12.7 m), consistent with a significant earthquake. This large vertical displacement, located only 1 km from the Amorgos coastline, suggests that the tsunami was likely triggered primarily by the fault rupture and sudden vertical displacement of the seafloor, reducing the weight of the mass-wasting contribution to the overall tsunami's generation. The orientation of the fault system is compatible with focal mechanisms calculated for the 1956 earthquake. Analysis of the fault mirror's characteristics, including the presence of tectoglyphs (striae that are not related to sedimentary processes) and differential weathering of the fault surfaces (color changes), supports its association with the 1956 event. Applying scaling laws relating earthquake magnitude to rupture parameters, the observed displacement is consistent with a moment magnitude (Mw) of 7.5, aligning with seismological data estimates. Considering the uncertainty of earthquake location and incomplete tsunami data, three main faults in the system were initially considered possible sources. However, only the Amorgos fault showed the freshly exhumed fault mirror, firmly establishing it as the likely source of the 1956 event. The observed vertical displacement correlates with horizontal extension and suggests a recurrence time of approximately 2250 years for such a large earthquake on the Amorgos fault, highlighting a potential underestimation of the seismic hazard in the region.

The identification of the Amorgos fault as the source of the 1956 earthquake and tsunami has significant implications for seismic hazard assessment in the region. The large seafloor offset observed directly supports the hypothesis of tsunami generation primarily from the fault rupture, rather than solely relying on submarine mass-wasting. The estimated recurrence time of approximately 2250 years suggests a potentially longer-than-anticipated interval between such large events, while also highlighting the need for further paleoseismological studies to constrain the recurrence interval and seismic hazard more accurately. This research demonstrates the importance of detailed submarine geological investigations for understanding tsunami sources. The findings highlight the critical need for comprehensive and detailed submarine geological observations to improve the accuracy of tsunami triggering mechanisms and to better predict future underwater earthquake and tsunami sources, ultimately aiding in community adaptation and the reduction of regional vulnerability, especially in tourism-dependent areas.

This study provides conclusive geological evidence that the 1956 Amorgos earthquake, which generated the largest Mediterranean tsunami in recent history, originated from a large rupture along the Amorgos fault. The discovery of a well-preserved seafloor rupture, showing a significant vertical offset, resolves a long-standing debate on the tsunami's source. This finding emphasizes the critical role of submarine geological investigations in understanding tsunami generation and refining seismic hazard assessments. Future research should focus on paleoseismological studies to better constrain the recurrence interval of large earthquakes on the Amorgos fault and further explore other historical tsunamigenic events using similar submarine exploration techniques.

While the study provides strong evidence linking the Amorgos fault to the 1956 event, the lack of direct dating methods for the fault scarp prevents definitive confirmation of the precise timing of the exhumation. The study focuses primarily on the Amorgos fault; detailed investigation of other faults in the area is needed to completely rule out other sources. The assessment of recurrence interval is based on scaling laws and limited paleoseismic data, thus requiring further confirmation through more detailed studies.