Testing a new hybrid approach to seismic hazard assessment: an application to the Calabrian Arc (Southern Italy)

Abstract

We tested a new hybrid method for the evaluation of seismic hazard. A recently proposed fault segmentation and earthquake recurrence model of peninsular Italy suggests that the interval for which the local historical catalogue is complete is shorter than the mean recurrence time of individual large faults (∼1000 years), or at the most comparable. These new findings violate the fundamental assumption of historical probabilistic seismic hazard methods that the historical record is representative of the activity of all the seismogenic sources. The hybrid method we propose uses time-dependent modelling of the major earthquakes and catalogue-based historical probabilistic estimates for all minor events. We assume that the largest earthquakes are characteristic for individual discrete fault segments, model their probability of occurrence by a renewal process and compute the shaking associated with each of them with a simplified procedure. Then we calculate the probability of exceeding a given threshold of peak ground acceleration for specific sites as the aggregate probability of occurrence of large characteristic earthquakes and minor shocks. We apply the method to the Calabrian Arc (Southern Italy) performing the calculations for five major towns. The exposure to seismic hazard of Reggio Calabria, Catanzaro and Vibo Valentia, which locate close to recently activated large faults, decreases with respect to traditional time-independent estimates. On the contrary, an increase of seismic hazard is obtained for Castrovillari, which locates in an area where large faults displaying Holocene activity have been recently recognized but no significant earthquake is reported in the historical catalogue. Cosenza has the highest probability to experience a significant peak ground acceleration with both the new hybrid and the traditional approaches. We wish to stress that the present results should be interpreted only in terms of the differences between the new hybrid and the traditional approaches, not for their absolute values, and that they are not intended to be used for updating or modifying the current national seismic zonation.