Are the source models of the M7.1 1908 Messina Straits earthquake reliable? Insights from a novel inversion and a sensitivity analysis of levelling data

For decades, many authors have attempted to define the location, geometry and kinematics of the causative fault for the 1908 December 28, M 7.1 earthquake that struck the Messina Straits between Sicily and Calabria (southern Italy). The coseismic displacement caused a predominant downwarping of the Straits and small land uplift away from it, which were documented by levelling surveys performed 1 yr before and immediately after the earthquake. Most of the source models based on inversion of levelling data suggested that the earthquake was caused by a low angle, east-dipping blind normal fault, whose upper projection intersects the Earth surface on the Sicilian (west) side of the Messina Straits.An alternative interpretation holds that the causative fault is one of the high-angle, west-dipping faults located in southern Calabria, on the eastern side of the Straits, and may in large part coincide with the mapped Armo Fault. Here, we critically review the levelling data with the aim of defining both their usefulness and limits in modelling the seismogenic fault. We demonstrate that the levelling data alone are not capable of discriminating between the two oppositely dipping fault models, and thus their role as a keystone for modellers is untenable. However, new morphotectonic and geodetic data indicate that the Armo Fault has very recent activity and is accumulating strain. The surface observations, together with appraisal ofmacroseismic intensity distribution, available seismic tomography and marine geophysical evidence, lends credit to the hypothesis that the Armo and possibly the S. Eufemia faults are part of a major crustal structure that slipped during the 1908 earthquake.Published1025-1041JCR Journalrestricte