Synconvergent Extension and the AD 365 earthquake: Implications for seismic and tsunami hazards in the eastern Mediterranean

The largest historical earthquake to strike the Mediterranean occurred on July 21, 365 AD. This event, along with an associated tsunami, destroyed several cities and killed thousands of people in the coastal regions throughout the eastern Mediterranean. The source and magnitude of this event have been debated for several decades. A Holocene bio-erosional tidal notch preserved for >250 km along the western coastlines of Crete, Greece, locally reaching ~10 m above present-day sea level, provides evidence of coseismic uplift around the first centuries AD. Current models assume that this bio-erosional notch was uplifted in a single event in 365 AD, requiring an M_w 8.3 - 8.5 earthquake. Recognizing that a megathrust rupture of the Hellenic subduction zone cannot reproduce the spatial and vertical deformation of the Cretan notch, the currently accepted source of this event is a northward dipping reverse fault that daylights offshore of southwestern Crete. Here, we use new and published radiocarbon data together with historical records to show that this uplift likely occurred in two or more events, distributed within 2-3 centuries, on two known active normal faults offshore of southwestern Crete. We evaluate this hypothesis by inverting the Holocene uplift data for fault rupture length, depth, dip and slip using a visco-elastic dislocation model to simulate co-and post-seismic deformation. Results show that rupture of these faults fits the elevation of the deformed paleoshoreline as well as previous models; however, relative to the reverse fault hypothesis, the fault dimensions are more realistic, and their kinematics are consistent with long-term extensional deformation of the Hellenic forearc. Furthermore, tsunami propagation modeling shows that these faults are capable of producing strong tsunamis that better match historical reports than the hypothesized reverse fault event. Our results demonstrate that uplift during at least two historical earthquakes provides a more parsimonious explanation for the observed Holocene uplift on Crete. These findings suggests that these historical eastern Mediterranean earthquakes did not exceed M_w 7.8, implying a lower, but still substantial, regional earthquake and tsunami hazards relative to previous work.