High-Complexity Deterministic Simulation of the 1994 Northridge Mw 6.7 Quake

With the recent addition of realistic fault topography in 3D simulations of earthquake source models, ground motion can be deterministically generated more accurately up to higher frequencies. The synthetic ground motions have been shown to match the characteristics of real data, having a flat power spectrum up to some cutoff frequency (Shi and Day, 2013). However, the earthquake source is not the only source of complexity in the high-frequency ground motion; there are also scattering effects caused by small-scale velocity and density heterogeneities in the medium that can affect the ground motion intensity. Here, we dynamically model the source of the 1994 Mw 6.7 Northridge earthquake using the Support Operator Rupture Dynamics (SORD) code up to 8 Hz. Our fault model was input into a layered velocity structure characteristic of the southern California region characterized by self-similar roughness from scales of 80 m up to the total length of the fault. We extend the ground motion to further distances by converting the output from SORD to a kinematic source for the finite difference anelastic wave propagation code AWP-ODC. This code incorporates frequency-dependent attenuation via a power law above a reference frequency in the form Q0fn.

We model the region surrounding the fault with and without small-scale medium complexity, with varying statistical parameters. Furthermore, we analyze the effect of varying both the power-law exponent of the attenuation relation (n) and the reference Q (Q0) (assumed to be proportional to the S-wave velocity), and compare our synthetic ground motions with several Ground Motion Prediction Equations (GMPEs) as well as observed accelerograms. We find that the spectral acceleration at various periods from our models are within 1 interevent standard deviation from the median GMPEs and compare well with that of recordings from strong ground motion stations at both short and long periods. At periods below 1 second, Q(f) is needed to match the decay of spectral acceleration seen in the GMPEs as a function of distance from the fault (n~0.6-0.8).We find when binning stations with a common distance metric (such as Joyner-Boore or rrup) that the effect of media heterogeneity is canceled out, however, the similarity between the intra-event variability of our simulations and observations increases when small-scale heterogeneity is included.