LLNL-PRES-XXXXXX This work was performed under the auspices of the U.S. Department of Energy by Lawrence Livermore National Laboratory under contract DE-AC52-07NA27344. Lawrence Livermore National Security, LLC
Comparison of 3D and 1D Wave Propagation Effects in the San Francisco Bay Area
Arben Pitarka
SWUS GMC Workshop 2 October , 2013
Lawrence Livermore National Laboratory LLNL-PRES-xxxxxx 2
Motivation
Lawrence Livermore National Laboratory LLNL-PRES-xxxxxx 3
Objective: Compare the efficiency of 3D and 1D velocity models at capturing wave propagation effects in areas with complex underground structure and surface topography.
Approach:
Simulate ground motion from Loma Prieta earthquake in the Bay Area in the long period range 1-10 s
Use kinematic rupture scenarios generated on the SCEC BB Platform
Analyze 3D wave propagation effects due to :
- large-scale structural complexities (3D-USGS)
- small-scale velocity variations (random perturb.)
- surface topography
SW4-3DFDM (Sjogreen and Petesson, 2009):
40m grid spacing, 3000 cores, fmax=1.3hz
Lawrence Livermore National Laboratory LLNL-PRES-xxxxxx 4
Random Velocity Perturbations (Pitarka et. al., 2009)
Lh=2000m Lz=200m Perturb=30% Vsmin=400m/s
3D USGS Velocity Model Version 8.3.0
SW NE SW NE
Lawrence Livermore National Laboratory LLNL-PRES-xxxxxx 5
SCEC-BBP Kinematic Rupture Model Graves&Pitarka (2010) Method
Lawrence Livermore National Laboratory LLNL-PRES-xxxxxx 6
E-W N-S Ground Motion Velocity
Lawrence Livermore National Laboratory LLNL-PRES-xxxxxx 7
1D 3D 3D+Topo+Stoch
RotD50 Acceleration Response Spectra 5% Damping, 1-10 s
- Significant improvements at long periods (> 2s) with 3D modeling - No visible effects at long periods due to topography and small scale velocity perturbations
Lawrence Livermore National Laboratory LLNL-PRES-xxxxxx 8
1D 3D
3D + Stoch
3D + Topo
Both 3D and 1D simulated motions are systematically lower than the recorded motion at periods > 3s.
- slip depth ? - inadequate 1D and 3D velocity models ? - 3D attenuation?
Lawrence Livermore National Laboratory LLNL-PRES-xxxxxx 9
3D simulations reveal redistribution of seismic energy among different wave types due to very complex propagation scattering and conversions of P and Rg energy to S-wave energy Max Frequency : 10Hz
P pS Rg
P
pS
S
Rg 5 km
5 km
10 km
pP
div curl
Lawrence Livermore National Laboratory LLNL-PRES-xxxxxx 10
Conclusions: 1. 3D modeling performs better than 1D modeling at periods longer than 2s
in the San Francisco Bay area.
2. Not enough stations to draw conclusions about basin depth dependency of the 3D/1D ground motion differences.
1. No visible effects due to surface topography and small scale velocity
perturbations in the long period range (1 – 10 s). Future Work : 1. Increase the maximum frequency in the 3D simulations for the San
Francisco Bay area to 3Hz. Use a series of BB rupture scenarios. Model smaller recent events.
2. Investigate correlation of 3D /1D effects with larger slip and basin depth. (San Francisco Bay, LA Basin).
Lawrence Livermore National Laboratory LLNL-PRES-xxxxxx 11
Acknowledgments: We thank - Christine Goulet at PEER Center for helpful discussions and suggestions during this work, and for providing the recorded ground motion data for the Loma Prieta Earthquake. - Robert Graves and Norm Abrahamson for their suggestions. - Jeff Bayless and Fabio Silva for their help in the preparation of some of the GOF plots.
Thank You !
Lawrence Livermore National Laboratory LLNL-PRES-xxxxxx 12
References 1. Graves, R.W, and A. Pitarka (2010). Broad-band ground motion simulation using an hybrid approach, Bull. Seism. Soc. Am., 100, 2095-2123.
2. Pitarka, A. (2009). Simulating forward and backward scattering in viscoelastic 3D media with random velocity variations and basin structure, Tech. Rep., USGS, award number 06HQGR0042. 3. Sjogreen B. and N. A. Petersson (2010). A fourth order accurate finite difference scheme for the elastic wave equation in second order formulation. J. Sci. Comput., 52:17–48, 2012. DOI 10.1007/s10915-011-9531-1.