Sensitivity of Ground-Motion Simulations to Earthquake Source Parameters: A Case Study for Istanbul, Turkey

Abstract

Following the disastrous earthquakes in Izmit and Duzce along the North Anatolian fault in 1999, the earthquake hazard in the Istanbul area became a great concern. In this study we simulate strong ground motions caused by a scenario earthquake (M 7.5) in the Marmara Sea, and investigate the effect of varying the input parameters on the broadband frequency ground motion. Simulations are based on a multiasperity source model that involves the combined rupture of the North Anatolian fault segments beneath the Marmara Sea. We use a hybrid model combin- ing a deterministic simulation of the low frequencies (0.1-1.0 Hz) with a semisto- chastic simulation of the high frequencies (1.0-10.0 Hz). Computation at each fre- quency range is performed separately and the total ground motion is combined in the time domain. Computations are linear and are performed at bedrock level, thereby not taking any effect of local geological conditions into account. We calculate a total of 17 earthquake scenarios corresponding to different source and attenuation param- eters to study their effect on the ground motion. The most significant parameters in terms of ground-shaking level are the rise time, rupture velocity, rupture initiation point, and stress drop. The largest variability of strong ground motions is observed in regions adjacent to asperities and is associated with frequencies higher than 5 Hz. For lower frequencies our simulated velocity spectra within the Istanbul area are fairly stable among scenarios. The average standard deviations of all ground-motion measures are less than 35% of the mean. Online material: Figures of peak ground acceleration and peak ground velocity and their differences to the reference scenario values.