Implications of earthquake recurrence models to the seismic hazard estimates in the marmara region, Turkey

Abstract

In this study, we show the effect of time-independent and time-dependent occurrence models on the seismic hazard estimations. The time-dependency is introduced by 1) the Brownian Passage Time (BPT) probability model that is based on a simple physical model of the earthquake cycle 2) the fusion of the BPT renewal model (BPT+ΔCFF) with a physical model that considers the earthquake probability perturbation for interacting faults by static Coulomb stress changes. To do so, we calculate the probability of occurrence of earthquakes Mw > 6.5 for individual fault sources in the Marmara region for the 5-10-30 and 50-year periods (starting from January 1, 2013). We treat the uncertainties in the fault parameters (e.g. slip rate, characteristic magnitude and aperiodicity) of the statistical distribution associated to each examined fault source by a Monte Carlo technique. Then the probabilities of occurrence for the next characteristic earthquake are calculated from three different models (Poisson, BPT, and BPT+ΔCFF) considering the 10th, 50th and 90th percentiles of the Monte Carlo distribution. In order to evaluate the impact of the earthquake probability models to ground motion hazard we attempt to calculate the fault-based probabilistic seismic hazard maps (PSHA) of mean Peak Ground Acceleration (PGA) having 10% probability of exceedance in 50 years on rock site condition. We note that in the present study we did not take in to account the ground motion variability caused by the different GMPE choices. In fact only one GMPE model is chosen as defined by Akkar and Cagnan (2010) (hereafter, AC10) for the active shallow crustal region for assessing the ground shaking hazard in the Marmara region in order to avoid those variability’s effect the final seismic hazard estimations in the study region. We observed that the impact of the different occurrence models on the seismic hazard estimate of selected sites is quite high: the hazard may increase by more than 70% or decrease by as much as 70%, depending on the applied model. We demonstrated that the estimated average recurrence time and the associated magnitude together with the elapsed time are crucial parameters in the earthquake probability calculations.