The Fall 2017 CIG-LLNL Computational Seismology Workshop was key in getting new users up and running quickly on seismic wave-propagation codes. The lead author of Dunn et al. (2018), a workshop participant, presented results from investigations of basin amplification effects due to local and regional earthquakes around the Reno-area basin at the annual SSA meeting the following Spring 2018. The study generated 3D SW4 synthetic seismograms for frequencies up to 3 Hz and 1 Hz from two events, the local 2008 Mogul sequence and the regional 2016 Nine Mile Ranch event. Several alternative basin models were investigated (Figure). Simulations showed higher than observed peak amplitudes for the Mogul sequence, which may be due to inaccuracies in the source or geological model and/or the lack of topography. Nine Mile Ranch simulations indicate that basin amplification is sensitive to the global minimum shear velocity. While computing higher frequencies becomes computationally expensive at lower velocities, basin effects are not modeled correctly when shallow low velocities are excluded. Lastly, the study observed no clear correlations between PGV, and Vs30 or basin thickness.
Mogul simulations were computed on Amazon Web Services on a 40 km x 40 km x 30 km grid with two mesh refinement layers and 25 meter grid spacing at the surface for a total of 310 million grid points. Minimum velocity was about 600 m/s.
Dunn, M., J. Louie, K.D. Smith, S. Dickenson (2018), Investigating Basin Amplification Factors for Shaking in the Reno, Nevada, Region for Local and Regional Events, poster presented at the 2018 Seismological Society of America Meeting, Miami, FL.
Figure. The maps compare shaking results from the different Reno-area basin models of Abbott and Louie, Saltus and Jachens, and Widmer. The top row shows the basin shear-velocity model at the surface, and the second and third rows the model PGV for horizontal and vertical motions respectively, all for the 2008 M5.0 Mogul earthquake in west Reno, Nevada.