Smooth crustal velocity models cause a depletion of high-frequency ground motions on soil in 2-D dynamic rupture simulations [post]

Yihe Huang
2021 unpublished
35 A depletion of high-frequency ground motions on soil sites has been observed in recent large 36 earthquakes and is often attributed to a nonlinear soil response. Here we show that the reduced 37 amplitudes of high-frequency horizontal-to-vertical spectral ratios on soil can also be caused by a 38 smooth crustal velocity model with low shear wave velocities underneath soil sites. We calculate 39 near-fault ground motions using both 2-D dynamic rupture simulations and point-source models 40
more » ... source models 40 for both rock and soil sites. The 1-D velocity models used in the simulations are derived from 41 empirical relationships between seismic wave velocities and depths in northern California. The 42 simulations for soil sites feature lower shear wave velocities and thus larger Poisson's ratios at 43 shallow depths than those for rock sites. The lower shear wave velocities cause slower shallow 44 rupture and smaller shallow slip, but both soil and rock simulations have similar rupture speeds 45 and slip for the rest of the fault. However, the simulated near-fault ground motions on soil and 46 rock sites have distinct features. Compared to ground motions on rock, horizontal ground 47 acceleration on soil is only amplified at low frequencies, whereas vertical ground acceleration is 48 de-amplified for the whole frequency range. Thus, the horizontal-to-vertical spectral ratios on 49 soil exhibit a depletion of high-frequency energy. The comparison between smooth and layered 50 velocity models demonstrates that the smoothness of the velocity model plays a critical role in 51 the contrasting behaviors of horizontal-to-vertical spectral ratios on soil and rock for different 52 rupture styles and velocity profiles. Our results reveal the significant role of shallow crustal 53 velocity structure in the generation of high-frequency ground motions on soil sites. 54 55 56 57 58 59 60 61 62
doi:10.31223/x5bk58 fatcat:6vi3hquhuzfajn4dxvghgqzpz4