GNGTS 2013 - Atti del 32° Convegno Nazionale

Earthquake Forecasting Models and their Impact on the Ground Motion Hazard in the Marmara Region, Turkey A. Akinci 1 , M. Murru 1 , R. Console 2 , G. Falcone 1 , S. Pucci 1 1 Istituto Nazionale di Geofisica e Vulcanologia, Rome, Italy 2 Center of Integrated Geomorphology for the Mediterranean Area, Potenza, Italy Introduction. We calculate the probability of occurrence of earthquakes Mw > 6.5 for individual fault sources in the Marmara region for 5, 10, 30 and 50-year periods (starting from January 1 th , 2013) using time-independent and time-dependent earthquake forecasting models. The use of geologic data incorporates the long-term recurrence behavior of active faults and the earthquake recurrence rates of individual fault sources derived from slip rate and magnitude using the characteristic earthquake model (Wesnousky, 1994). In order to express the time dependence of the seismic processes to predict the future ground motions that will occur across the region, we used a Brownian Passage Time (BPT) model (Matthews et al. , 2002) that is characterized by a mean recurrence time, uncertainty in the recurrence distribution, and elapsed time since the last earthquake. We also used a methodology based on the fusion of the BPT renewal model with a physical model that considers the earthquake probability perturbation for interacting faults by static Coulomb stress changes (King et al. , 1994). We treat the uncertainties in the slip rate, characteristic magnitude and aperiodicity of the statistical distribution associated to each examined fault source, by a Monte Carlo technique. The Monte Carlo samples for all these parameters are drawn from a uniform distribution within their uncertainty limits. In order to evaluate the sensitivities of the earthquake probability models to ground motion hazard we attempt to calculate the probabilistic seismic hazard in the study region. The probabilistic seismic hazard maps of Marmara region are generated by the same procedure used for seismic hazard assessment in United States (Petersen et al. , 2008). We adopted empirical Ground Motion Prediction Equations, GMPEs, for assessing the ground shaking hazard as defined by Akkar and Bommer (2010). We observed that the impact of the different occurrence models on the seismic hazard estimate is quite high; the hazard may further increase up to 50% or more or may decrease by as much as 50% depending on the selected occurrence model in the selected sites. This difference mostly depends on the time elapsed after the latest major earthquake on a specific fault. In this study, we give the probabilities of occurrence for the next characteristic earthquake, considering the 10 th , 50 th and 90 th percentiles of the Monte Carlo distribution, over the future 50 years, starting on 1 January 2013, considering the information on used parameter uncertainties in the Marmara region. We then attempt to calculate the fault-based probabilistic seismic hazard maps (PSHA) of mean Peak Ground Acceleration (PGA) having 10% probability of exceedence in 50 years on rock site condition from those three forecasting models. Finally, we demonstrate the earthquake occurrence model uncertainty, and the sensitivity of the ground motion hazard for different earthquake recurrence models reporting those of the percentage ratio between the Poisson, BPT and BPT+∆CFF models. Tectonic setting. The Marmara region is located at the western end of the North Anatolian Fault Zone (NAFZ). NAFZ splays into three strands in the Marmara region (Barka 1991). The lack of macro-seismicity is located between the 28°-29° longitudes which correspond to the northern border of the Marmara Sea (Ambraseys, 1970; Stein et al. , 1997). This gap has a length of 150 km and is therefore capable of generating an earthquake with magnitude similar to that of the Izmit earthquake (Hubert-Ferrari et al. , 2000; Parsons et al. , 2000). The return period for large earthquakes south of Istanbul varies between 100-1000 years depending on the magnitude of earthquakes and the considered slip rates (GPS=15 mm/y; geological=2-4 mm/y). The latest major large earthquake on this segment was probably in 1509 AD, (intensity, IX). 5 GNGTS 2013 S essione 2.1