Global synthetic seismograms using a 2-D finite-difference method

Abstract

finite-difference (FD) synthetics, which fill the gap between fast 1-D analytic synthetics and time-consuming full 3-D synthetics in our ability to model seismograms, have been used in many studies. We address several issues involving 2-D FD methods in generating global synthetic seismograms. These include: (1) interfacing point source excitation for earthquakes with 2-D FD methods; (2) out-of-plane spreading corrections and (3) reducing the spherical Earth to the flattened models. The first issue is tackled using two methods, a 'transparent source box' approach and a moment tensor excitation approach, where each has its own advantages. Moreover, our 'source box' excitation does not have the late-time drift problem that occurred in previous studies. The out-of-plane geometric spreading correction is accounted for by estimating the ray parameter and applying a post-simulation filter to 2-D synthetics. Finally, parameters of the Earth-flattening transformation are discussed and validated. The effectiveness of this method is demonstrated by comparing our synthetics with frequency-wavenumber summation, normal-mode and 3-D spectral-element synthetics.

Earthquake ground motion amplification for surface waves

“…Even for a variety of tested basin depths, impedance contrasts and even for more realistic geologic profiles [e.g., Boore and Joyner , ], the differences in amplification remain (see supporting information Figure S1). Lastly, we note that these transfer functions are confirmed through 2‐D finite difference simulations [ Li et al ., ], with only minor differences caused by path effects such as scattering and a conversion of wave types at the edge of the sedimentary basin.…”

Source Parameters of the Shallow 2012 Brawley Earthquake, Imperial Valley

“…A GPU version of the 8th order staggeredgrid 2D finite difference code is used for the numerical computation (Li et al, 2012). 1D synthetic waveforms are obtained through the frequency-wavenumber integration (Zhu and Rivera, 2002).…”

Mantle Transition Zone Structure Beneath Northeast Asia From 2‐D Triplicated Waveform Modeling: Implication for a Segmented Stagnant Slab

“…In contrast to previous studies using triplicated waveform modeling based on 1‐D velocity structure, we simulated the 2‐D regional elastic wavefield using the Spectrum‐Element Method with the implementation of Komatitsch and Tromp (, ). We inserted the moment tensor source following the approach of Li et al (). To resolve the synthetics with high‐frequency content of up to 0.8 Hz, we set the Spectrum‐Element Method grid to 2 km throughout the entire domain, and the time step to 0.005 and 0.01 s for P and SH wave, respectively, which transfers to 1,750 × 1,500 grid points (3,500 km in the horizontal and 3,000 km in the vertical).…”