Stochastic Strong Ground Motion Simulation of the Southern Aegean Sea Benioff Zone Intermediate‐Depth Earthquakes

Abstract We employ the stochastic finite‐fault modeling approach of Motazedian and Atkinson (2005), as adapted by Boore (2009), for the simulation of Fourier amplitude spectra (FAS) of intermediate‐depth earthquakes in the southern Aegean Sea subduction (southern Greece). To calibrate the necessary model parameters of the stochastic finite‐fault method, we used waveform data from both acceleration and broadband‐velocity sensor instruments for intermediate‐depth earthquakes (depths ∼45–140 km) with M 4.5–6.7 that occurred along the southern Aegean Sea Wadati–Benioff zone. The anelastic attenuation parameters employed for the simulations were adapted from recent studies, suggesting large back‐arc to fore‐arc attenuation differences. High‐frequency spectral slopes (kappa values) were constrained from the analysis of a large number of earthquakes from the high‐density EGELADOS (Exploring the Geodynamics of Subducted Lithosphere Using an Amphibian Deployment of Seismographs) temporary network. Because of the lack of site‐specific information, generic site amplification functions available for the Aegean Sea region were adopted. Using the previous source, path, and site‐effect constraints, we solved for the stress‐parameter values by a trial‐and‐error approach, in an attempt to fit the FAS of the available intermediate‐depth earthquake waveforms. Despite the fact that most source, path, and site model parameters are based on independent studies and a single source parameter (stress parameter) is optimized, an excellent comparison between observations and simulations is found for both peak ground acceleration (PGA) and peak ground velocity (PGV), as well as for FAS values. The final stress‐parameter values increase with moment magnitude, reaching large values (>300 bars) for events M≥6.0. Blind tests for an event not used for the model calibration verify the good agreement of the simulated and observed ground motions for both back‐arc and along‐arc stations. The results suggest that the employed approach can be efficiently used for the modeling of large historical intermediate‐depth earthquakes, as well as for seismic hazard assessment for similar intermediate‐depth events in the southern Aegean Sea area.