GNGTS 2013 - Atti del 32° Convegno Nazionale

Bibliografia Allan N., Cantle N., GodFrey P., Yin Y.; 2011: A review of the risk of complex systems applied to risk appetite and emerging risks in ERM practice. (London) Institute and Faculty of Actuaries, Ania; 2011: Danni da eventi sismici e alluvionali al patrimonio abitativo italiano: studio quantitativo e possibili schemi assicurativi AON; 2012: Annual Global Climate and Catastrophe Report – Impact forecasting 2012 EIOPA; 2013: DOC-13/061 28 - LTGA technical specification (part I) Luraschi P., Corrigan J.; 2013: Operational risk modelling framework, Capitolo 6 Swiss Re; 2013: Sigma No 2/2013 Taleb N. N; 2009: Il cigno nero. Come l’improbabile governa la nostra vita An alternative approach for the estimate of seismic hazard: applications to Italian Regions and comparison with the present hazard maps E. Mantovani 1 , D. Babbucci 1 , M. Baglione 2 , N. Cenni 3 , V. D’Intinosante 2 , L. Martelli 4 , C. Tamburelli 1 , M. Viti 1 1 Dipartimento di Scienze Fisiche, della Terra e dell’Ambiente, Università di Siena, Italy 2 Coordinamento Regionale Prevenzione Sismica, Regione Toscana, Italy 3 Dipartimento di Scienze Biologiche, Geologiche e Ambientali, Università di Bologna, Italy 4 Servizio Geologico, Sismico e dei Suoli, Regione Emilia-Romagna, Italy Introduction. The estimate of seismic hazard in a zone is a scientific objective which has very important social implications. So, any effort must be made to identify the most reliable evaluation, through a critical analysis of the results obtained by different approaches. The hazard maps presently adopted for the mitigation of the seismic risk in Italy are based on the probabilistic approach (PSHA, e.g., Stucchi et al. , 2011), whose limitations have been pointed out by several researchers (Albarello and D’Amico, 2008; Wyss, 2012; Stein et al. , 2012; Mantovani et al. , 2012, 2013; Viti et al. , this volume). A significant example of the major problems cited above can be taken from the present hazard maps concerning the Toscana and Emilia-Romagna Regions (Fig. 1). This result provides that in such Regions the maximum expected intensity (with a exceedence probability of 10% in 50 years) is VIII °MCS (Mercalli- Cancani-Sieberg scale). This implies that a reasonable defence from earthquakes in those Regions, also considering social and economic criteria, can be obtained by making building able to resist the shaking caused by such intensity. Since the above two Regions have been hit by several shocks with greater intensity (IX and X MCS), the people living in those zones could be worried about the possibility that the PSHA prediction underestimates the real risk. In our opinion, the above fear can hardly be removed since the reliability of the procedure used to achieve the present hazard maps cannot be evaluated, being based on a number of arbitrary assumptions about the behaviour of seismicity. For instance, the hypothesis that earthquakes are casual and independent events is not compatible with the nature of such phenomenon (e.g., Castanos and Lomnitz, 2000; Mantovani et al. , 2012, 2013; Stein et al. , 2012). Furthermore, the assumption that the available data set (some centuries long) is representative of the future behaviour of seismicity can hardly be verified, considering that the deformation and the consequent fracturing of rocks has developed over geological times. Thus, the probability that the space-time distribution of earthquakes during the next 50 years will just resembles the one that developed in the last 300-400 years is extremely low (e.g., Swafford and Stein, 2007). On the contrary, it is reasonable to think that in the next 50 years the faults activated in recent times have a low probability of being reactivated. The above remarks imply that any statistical estimate of seismicity parameters in a given zone (such as return periods, starting time of the seismic cicle, etc.) may be affected by unknown uncertainty, which may reflect in a biased estimate of earthquake probability. 422 GNGTS 2013 S essione 2.3