GNGTS 2017 - 36° Convegno Nazionale

GNGTS 2017 S essione 2.1 305 Model A2 . Source zones used by A2 are the same of the A4 model described later, but A2 has adopted an original approach to estimate seismicity rates. The most innovative part of the approach consists of the joint estimation of the time completeness and of the seismicity rate of the complete part of the historical catalogue. Specifically, assuming that the main shocks follow a homogeneous Poisson process, the problem consists of analysing the completeness of the earthquake occurrences and calculating the rates for the complete part of the CPTI15 catalogue. The issue has been tackled as a problem of identification, in a Bayesian framework, of the change-point between two Poisson processes. The rate of occurrence of part considered as complete of the data set of events has been then assumed as representative of the long-term seismic rate occurrence. Considering both the rates and the time when the process changes rate as random variables, the method computes, the entire probability distribution of these variables. Models A3a and A3b . Two models, A3a and A3b, are here described together as they use the same approach to evaluate seismicity rates, but with different definitions of area sources. Both models have been based on a first-order geodynamic subdivision that identifies areas characterized by a common tectonic setting under the current stress regime and plate-tectonic setting. Then, these zones have been split into smaller zones, which have been drawn differently. The annual rates in each zone have been estimated using the declustered CPTI15 catalogue, considering the two completeness studies (historical and statistical). For each zone, the b value and the corner magnitude of a Tapered GR (Pareto) distribution (Kagan 2002) have been estimated with the maximum-likelihood algorithm by Jackson and Kagan (1999). Successively, annual rates of each zone have been partitioned onto the smaller zones. ModelA4 . Source zones ofA4have designedusing themost recent input data, in a conventional PSHA scheme. The model represents an update of the previous model ZS9 (Meletti et al. , 2008); it includes all recent advances on the understanding of the active tectonics of the Italian territory and on the distribution of seismogenic sources, which are described by the Database DISS 3.2.1 and other active fault compilations at the national and regional scales; it accounts for information that are derived from the investigation of the most recent seismic sequences that have struck Italy after the compilation of ZS9. To estimate the expected seismicity rates, the zones have been grouped into macro areas, according to their tectonic features. Assuming that the distribution of the earthquake magnitudes follows a truncated GR model, the b and a values have been estimated applying the Weichert (1980) approach to each macro area, using CPTI15. Then, the seismicity rate for each zone has been estimated from the rates of the macro areas using different approaches. The same procedure has been applied directly also to the zones, but imposing upper and lower bounds for the b-value. Fault based models. Model F1 . it is a fault-based seismicity model that uses exclusively geological information taken from the DISS 3.2.1. This characteristic ensured the independence of the seismicity rates predictedby thismodel fromthe earthquake catalogue.The seismicmoment rate Ṁs of a seismogenic fault is derived from the geologic moment rate. Using the moment conservation principle, the seismic moment rate of any seismogenic fault has been converted in the annual earthquake rate of a Tapered GR (Pareto) magnitude-frequency distribution (MFD) using the formulations by Kagan (2002) and assuming a unique Beta equal to 0.667 for all the faults. To evaluate off-fault seismicity rates F1 follows an empirical approach aimed to capture the natural distribution of observed earthquakes around faults and the inherent location uncertainty of both earthquakes and faults. To this aim, an Empirical Proximity Function (EPF), which represents the frequency at which any observed earthquake has occurred at any distance from any mapped fault, has been firstly developed. Then, the earthquakes predicted by the MFD described above are symmetrically spread around the fault according to the EPF. Model F2 . the model combines the seismic rates obtained by three different databases: the individual sources of DISS 3.2.1, the CPTI15 catalogue and an instrumental catalogue. The seismic moment rate of each individual source has been computed using the seismic-moment