Valutazioni sperimentali di probabilità di occorrenza dei terremoti utilizzando metodologie non parametriche applicate a zonazioni diverse

Abstract

In 2003 a new building code for Italy has been released. Subsequently a new seismic hazard reference map of Italy (MPS04) have been compiled for definition of seismic zones. In order to define priority area for any short-term policy in loss reduction in Italy, different approach have been followed and compared. One of this approach is based on the clustered feature of the earthquake occurrence for events greater than M5.5, according to the Proportional Hazard Model. The analysis of the distribution of large events is composed by several ingredients. In fact, besides the statistical distribution of events, the catalogue and the zonation play an important role. In this work the same input data used for the MPS04 have been introduced in the model, in order to investigate its sensitivity and the stability of the results and to check the influence into the probability distribution of factors like the catalogues, the zonations and the magnitude-temporal completeness. Then, the predictive ability of this model has been tested and compared to the one of the Poisson distribution, which is used in standard hazard analysis. The results show the same pattern for earthquake occurrence in all the applications, indicating a cluster properties for earthquake occurrence. The cluster characteristic, in terms of time duration and intensity, may change adopting a different catalogue, but it is not significantly influenced by the three zonations adopted in the test.