Richard F Ott , Karl William Wegmann, Sean F Gallen, Frank J Pazzaglia, Mark T Brandon, Kosuke Ueda, Charalampos Fassoulas

The hallmark of great earthquakes (Mw ≈ 8.3-8.5) in the Mediterranean is the 21 July AD 365 earthquake and tsunami that destroyed cities and killed thousands of people throughout the Eastern Mediterranean. This event is intriguing because most Mediterranean subduction forearcs exhibit pervasive crustal extension and minimal definitive evidence exists for great subduction megathrust earthquakes, consistent with weak seismic coupling. This conundrum has led many to favor rupture of a previously unrecognized upper plate splay fault south of Crete in an Mw 8.3-8.5 earthquake, uplifting a Holocene paleoshoreline on Crete by up to 9 m. Similar source mechanisms have been adapted and commonly used for seismic and tsunami hazard estimation in the region. We present an alternative model for the uplift of the Cretan paleoshoreline and the AD 365 tsunami that centers on known active normal fault systems offshore of western and southwestern Crete. We use new and published radiocarbon dates, together with historical records, to show that uplift of the Cretan paleoshoreline likely occurred during two or more earthquakes within 2 to 3 centuries. Visco-elastic dislocation modeling demonstrates that the rupture of these normal faults fits observed data equally as well as reverse fault models but requires less slip. Tsunami modeling shows that normal-fault ruptures produce strong tsunamis that better match historical reports than a hypothetical reverse fault. Our findings collectively favor the interpretation that damaging earthquakes and tsunamis in the Eastern Mediterranean originate on normal faults and highlight the potential hazard from tsunamigenic upper plate normal fault earthquakes.