GNGTS 2013 - Atti del 32° Convegno Nazionale

Probability of occurrence for the fault segmentation in the Marmara region. We calculated the probability of occurrence for the fault segmentation in the Marmara region using the Poisson model for time intervals of 5-10-30 and 50 years starting from January 1, 2013. These probabilities are shown in Tab. 1. We note that the maximum values of the Poisson probability are on those faults that have a high long-term slip rate value. This value has an influence on the mean recurrence time and consequently high values of slip rate reduce the inter-event time. In fact if we examine the latest two events occurred in the Marmara region on the Izmit (August 17, 1999, M w 7.4) and Duzce (November 12, 1999, M w 7.1) faults, we will note on these fault sources two of the highest Poisson values. For both faults the value of long-term slip rate is equal to 15±3 mm per year. The BTP probability is larger than the Poisson probability in the case when the characteristic event has occurred long time ago and the elapsed time is closer to the inter event time. If we consider the 50 th percentile for the next 50 years, the largest values will be for West Marmara and Cinarcik faults, in the northern part of Marmara Sea, near the city of Istanbul. For Iznik fault, whose latest event is reported in 121 AD, if we consider the 50 th percentile we will obtain for the next 50 years, a BPT probability equal to 8.01% with respect to a 3.78% Poisson probability. The larger value for the BPT time-dependent model is explained by the very long time elapsed after the latest characteristic event on this fault. The BPT probabilities for the Izmit and Duzce faults are lower than the Poisson values due to the short elapsed time (14 years) after the occurrence of their last characteristic events. The time–dependent hazard rate obtained by the BPT distribution on each fault is successively modified by the inclusion of a permanent physical effect due to the Coulomb static stress change caused by failure of neighboring faults since the latest characteristic earthquake on the fault of interest. We treat again in this step the uncertainties in the recurrence time, co-seismic slip, and magnitude, by a Monte Carlo technique, related to each fault. The Monte Carlo samples for all these parameters are drawn from a uniform distribution within their uncertainty limits. The probability values for Cinarcik, and South Cinarcik show a small difference (±1%) between the two models. This variation is due to a positive (Cinarcik) or negative (South Cinarcik) fault interaction. The largest difference between the two models is visible on the North Saroz fault. In this last case, the BPT probability is larger (22.6%) than the BPT probability with interaction (16.5%). Implications on the probabilistic seismic hazard maps. We also demonstrate the earthquake occurrence model uncertainty, and the sensitivity of the ground motion hazard for different earthquake recurrence models. Fig. 1a, b and c reports the percentage ratio between the seismic hazard computed with the three earthquake occurrence models (Poisson, BPT and BPT+∆CFF models, respectively). It is presented as a relative difference in percentage, computed as: [PSHA model1 – PSHA model2 ] / PSHA model2 * 100. In generally, the PSHA results based on the time-independent and time-dependent earthquake occurrence models display the effect of the fault recurrence rate and the regency of fault rupture by drastically reducing hazard levels along the eastern part of the North Anatolian Fault Zone by up to 50% nearby Izmit (1), Duzce (11), Gonen (22), Biga (23) and Pazarkoy (24) areas. The strongest effect may be caused by the recurrence rates and the lapse time ratio of the faults and rarely by the maximum magnitude since the models associate same magnitudes to each fault. The areas affected by the recurrence rate change get larger and/or bigger in the time-dependent earthquake occurrence model than the time-independent one. The faults where the BPT probabilities are smaller than the Poisson ones can easily be seen in Fig. 1a, b and c; areas with blue color indicate the decrease in the seismic hazard in terms of ground accelerations. The seismic hazard around faults number 1, 10 and 11 decreases strongly from 0.55 to 0.25 g (up to 50% of change) when the earthquake occurrence model changes from the 7 GNGTS 2013 S essione 2.1