The characterization of intensity decay with distance from the source is of paramount importance for seismic hazard assessment and to earthquake damage scenarios necessary for the management of prevention activities. In the recent literature, a renewal of interest about this problem is occurring in particular as concerns the Italian region where different parameterizations have been proposed. The significant scatter between these relationships suggested a re-analysis of the problem in order to overcome most significant drawbacks that characterised previous approaches, as discussed in a companion presentation. As a first step the analysis aimed at the definition of an isotropic attenuation relationship valid for the whole Italian area. Since this attenuation relationship has to be primarily used in probabilistic seismic hazard assessment, major attention has been devoted to evaluate attenuation relationship in its complete probabilistic form. An important aspect has been the preliminary evaluation of the “intrinsic” scattering of data (i.e. the one independent of specific attenuation relationship to be used), which represents the lowest threshold for the variance to be explained with the attenuation relationship. Furthermore, to reduce possible biases, completeness of the available database has been checked and a suitable data selection has been performed. Since epicentral intensity cannot be defined unambiguously from the experimental point of view, the attenuation relationship has been scaled with a new variable more representative of the earthquake dimension. Several criteria have been considered to evaluate competing attenuation formulas (ANOVA, BIC, AIC, etc.) while statistical uncertainty about empirical parameters has been evaluated by using standard approaches and bootstrap simulations. At last, the performance of the selected relationship has been analysed by using a distribution-free approach.
The attenuation of seismic intensity in Italy: modeling and validation / Pasolini C.; Albarello D.; Gasperini P.; D'Amico V.; Lolli B.. - STAMPA. - (2006), pp. 124-124. (Intervento presentato al convegno First European Conference on Earthquake Engineering and Seismology tenutosi a Geneva, Switzerland nel 3-8 September 2006).
The attenuation of seismic intensity in Italy: modeling and validation
PASOLINI, CHIARA;GASPERINI, PAOLO;LOLLI, BARBARA
2006
Abstract
The characterization of intensity decay with distance from the source is of paramount importance for seismic hazard assessment and to earthquake damage scenarios necessary for the management of prevention activities. In the recent literature, a renewal of interest about this problem is occurring in particular as concerns the Italian region where different parameterizations have been proposed. The significant scatter between these relationships suggested a re-analysis of the problem in order to overcome most significant drawbacks that characterised previous approaches, as discussed in a companion presentation. As a first step the analysis aimed at the definition of an isotropic attenuation relationship valid for the whole Italian area. Since this attenuation relationship has to be primarily used in probabilistic seismic hazard assessment, major attention has been devoted to evaluate attenuation relationship in its complete probabilistic form. An important aspect has been the preliminary evaluation of the “intrinsic” scattering of data (i.e. the one independent of specific attenuation relationship to be used), which represents the lowest threshold for the variance to be explained with the attenuation relationship. Furthermore, to reduce possible biases, completeness of the available database has been checked and a suitable data selection has been performed. Since epicentral intensity cannot be defined unambiguously from the experimental point of view, the attenuation relationship has been scaled with a new variable more representative of the earthquake dimension. Several criteria have been considered to evaluate competing attenuation formulas (ANOVA, BIC, AIC, etc.) while statistical uncertainty about empirical parameters has been evaluated by using standard approaches and bootstrap simulations. At last, the performance of the selected relationship has been analysed by using a distribution-free approach.I documenti in IRIS sono protetti da copyright e tutti i diritti sono riservati, salvo diversa indicazione.
