Comparison of ground motion hybrid simulations to NGA modified GMPE in the Marmara Sea region (Turkey) in a directivity rupture context.

Abstract

We have simulated strong ground motions for two Mw>7.0 rupture scenarios on the North Anatolian Fault, in the Marmara Sea within 10-20 km of Istanbul. This city is characterized by one of the highest levels of seismic risk in Europe and the Mediterranean region. The increased risk in Istanbul is due to the high probability of the occurrence of a large earthquake, which stands at about 65% during the coming years. To estimate the ground motion characteristics and its variability in the region we have adopted physics-based rupture scenarios, simulating hybrid broadband time histories. We have merged two simulation techniques: a full 3D wave propagation method to generate low-frequency seismograms (Aochi and Ulrich, 2015) and the stochastic finite-fault model approach based on a dynamic corner frequency (Motazedian and Atkinson, 2005) to simulate high-frequency seismograms. They are merged to compute realistic broad band hybrid time series. The comparison of intensity measures (PGA, PGV, SA) on our simulations with recently proposed Ground Motion Prediction Equations (GMPEs) in the region (Boore & Atkinson, 2008; Chiou & Young, 2008; Akkar & Bommer, 2010; Akkar & Cagnan, 2010) points out the rupture directivity and super-shear rupture effects associated to these cases. In order to improve the comparison, we use the GMPE proposed by Boore & Atkinson (2008) with the directivity correction proposed by Spudich & Chiu (2008). This study highlights the importance of the directivity of the rupture for the hazard estimation in the region of the Marmara Sea and especially for the city of Istanbul.