HR: 11:45h
AN: S32A-06    [Abstracts]
TI: A Model-Based Signal Processing Approach to Seismic Monitoring
AU: * Rodgers, A
EM: rodgers7@llnl.gov
AF: Lawrence Livermore National Laboratory, L-205 7000 East Avenue, Livermore, CA 94551 United States
AU: Harris, D
EM: harris2@llnl.gov
AF: Lawrence Livermore National Laboratory, L-205 7000 East Avenue, Livermore, CA 94551 United States
AU: Ford, S
EM: ford17@llnl.gov
AF: Lawrence Livermore National Laboratory, L-205 7000 East Avenue, Livermore, CA 94551 United States
AU: Pasyanos, M
EM: pasyanos1@llnl.gov
AF: Lawrence Livermore National Laboratory, L-205 7000 East Avenue, Livermore, CA 94551 United States
AB: Recent applications of correlation methods to seismological problems illustrate the power of coherent signal processing applied to whole waveforms. Examples of these applications include detection of events buried in ambient noise and cross-correlation of sets of waveforms to form event clusters and accurately measure delay times for event relocation. These methods rely on redundancy and have been successfully applied to large sets of empirical observations. However, in cases with little or no empirical event data, such as aseismic regions or exotic event types, correlation methods will likely fail due to the lack of previous observed similar waveforms. This study seeks to use model-based signals computed for three-dimensional Earth models to form the basis for correlation detection. To demonstrate the method we are modeling broadband regional seismograms for a moderate (M~5) earthquake near the China-North Korea border. Synthetic seismograms are computed with the Spectral Element Method for a suite of long-wavelength (2 degree) seismic velocity models inferred with the Markov Chain Monte-Carlo (MCMC) method. MCMC uses stochastic sampling to fit multiple data sets but rather than estimate a single "optimal" model, MCMC results in a suite of models that sample the model space and incorporates uncertainty through variability of the models. The variability reflects our ignorance of Earth structure, due to limited resolution, data and model errors, and produces variability in the seismic waveform response. Model-based signals are combined using a sub-space method where the synthetic signals are decomposed into an orthogonal basis by singular-value decomposition (SVD). The observed waveforms are represented with a linear combination of eigenvectors (signals) associated with the most significant eigenvalues of the SVD. We demonstrate the ability of model-based signals to represent intermediate period (down to 5 s) seismograms. Further work will require higher frequency synthetic seismograms and the inclusion of shorter wavelength velocity structure, whether inferred from various seismic data sets or generated stochastically.
DE: 7205 Continental crust (1219)
DE: 7219 Seismic monitoring and test-ban treaty verification
DE: 7290 Computational seismology
SC: Seismology [S]
MN: Fall Meeting 2005