Evaluation of Full Moment Tensor Inversion based upon Fiber-Optic Dataset
Abstract
Distributed Acoustic Sensing (DAS) systems offer a new and rich dataset recording seismic waves along fiber-optic cables. Recent studies have shown that DAS is sensitive to both induced microseismic events, which can occur in enhanced geothermal system and carbon sequestration settings, but also to ground motion response to remote earthquakes.
Here, we propose to evaluate the potential of DAS for full moment tensor inversion using a dataset collected in the Brady Hot Springs geothermal field, Nevada, in March 2016 under the PoroTomo project (Feigl and PoroTomo Team, 2017). To that end, we use a version of the Bayesloc method (Myers et al., 2007, 2009), a Bayesian hierarchical seismic event locator that incorporates differential-times measurements into the multiple event location algorithm, and adapt it to use strain waveform dataset to extract probability of event locations. Then, using these event locations, we show that we can recover the moment tensor coefficients through a linear least-squares inversion, where the objective is to minimize a residual between observed and predicted strain measurements. Finally, we use finite-frequency sensitivity kernels, calculated based on an adjoint method (Kim et al., 2011) adapted to DAS strain dataset, for a non-linear iterative inversion to refine both event location and moment tensor solution. Predicted strain measurements and sensitivity kernels are obtained using a spectral-element code. This work was performed under the auspices of the U.S. Department of Energy by Lawrence Livermore National Laboratory under Contract DE-AC52-07NA27344.- Publication:
-
AGU Fall Meeting Abstracts
- Pub Date:
- December 2019
- Bibcode:
- 2019AGUFM.S23B..07M
- Keywords:
-
- 7212 Earthquake ground motions and engineering seismology;
- SEISMOLOGY;
- 7260 Theory;
- SEISMOLOGY;
- 7280 Volcano seismology;
- SEISMOLOGY;
- 7294 Seismic instruments and networks;
- SEISMOLOGY