Broad-band Modeling of Earthquakes in the Rock Valley, Nevada: Implication of Wave Propagation Effects on the P/S Source Discriminant

Analysis of seismic records of local earthquakes and a series of underground chemical explosions conducted during the Source Physics Experiment (SPE) at the Nevada National Security Site (NNSS) have shown that at local distances (<200 km) the effectiveness of the single-station P/S ratio source discriminant is reduced, especially when seismic recordings from a sparse network of stations is used.

We used high performance computing to model high-frequency (0-10Hz) waveforms for 12 selected local earthquakes, with magnitudes ranging from 2.05 to 3.54, recorded by a network of seismic stations in the Rock Valley at NNSS. In addition, we performed a series of simulations of collocated isotropic and double-couple explosion sources in the Rock valley. The high-frequency wave propagation scattering was simulated by adding correlated small-scale stochastic perturbations to the Seismic Velocity Model of the Rock Valley (SVM). The recorded and synthetic waveforms were then analyzed to investigate the effects of source radiation and wave scattering effects on the simulated waveforms and P/S source discriminant.

The inclusion of correlated depth-dependent stochastic velocity perturbations in the GFM, improved the quality of simulated source radiation and local waveforms, which resulted in better reproduction of the observed spatial variations of the P/S discriminant. We found that the shallow wave scattering deforms the radiation pattern and amplitude of source generated P and S waves, thus reducing the efficiency of the P/S discriminant. Our simulations suggest that a good azimuthal stations coverage and the network averaging can improve the performance of the P/S discriminant at local distances.