Stochastic ground motion simulation of the 26 September 2019 Mw 5.8 Silivri (Istanbul) earthquake

The aim of using simulation techniques to provide generated ground motion data is to extend our knowledge on the effect of earthquakes and understanding their physical properties. High-frequency accelerations have incoherent behavior because of unpredictable irregularities and heterogeneities associated with faulting and wave propagation. Simulations of ground motions frequencies beyond > 1 Hz can be represented with stochastic methods using simplified model representations of source, path and site effects. In this paper, stochastic simulations are performed for the recordings of the 26 September 2019 Silivri, Istanbul earthquake, using a finite fault simulation approach with a dynamic corner frequency. The main target is to create a valid synthetic model database with consistent source, path, and site parameters in the region that can be implemented in future simulation efforts. In calibration, we have used the recordings at 59 widely distributed stations in Istanbul located on different site conditions with epicentral distances ranging from 23 to 101 km. Four different frequency-dependent Q models were tested to obtain the best fit with the observations. By comparing generated ground motions to the observed ones, optimum source parameters and crustal characteristics were estimated. The calibrated model parameters have been obtained from the set of best-fit data with observed ground motion in frequency domain. Synthetic PGAs have been compared with the NGA-West2 Ground Motion Models (GMMs). Furthermore, spatial distributions of the ground motion intensity parameters were obtained and compared with available damage observations in Istanbul due to this earthquake. In conclusion, the results of the simulation were in good agreement with the recorded ones, both in time and frequency domains. The results indicate that the proposed stochastic model can be used to simulate ground motion distributions in Istanbul and beyond from past and future events in the region.