Passive Imaging and Characterization of the Subsurface with Distributed Acoustic Sensing
- Thesis pdf
Table of contents
- Chapter 1: Introduction and background
- Chapter 2: Theoretical response of DAS
- Chapter 3: Experimental data overview
- Chapter 4: Observations of earthquakes at Stanford
- Chapter 5: Anthropogenic effects on noise correlation functions
- Chapter 6: Fast, scalable dispersion image calculation
- Chapter 7: Time-lapse interferometry throughout DAS arrays
- Appendix A: Review of particle velocity and strain rate rotations
- Appendix B: Deriving responses to plane waves
- Appendix C: Deriving responses to point sources
- Appendix D: 3D extension of corner data analysis
Active seismic surveys for subsurface imaging are expensive and logistically difficult in populated areas where they have potential to impact day-to-day life, so continuous monitoring experiments are rarely done. I combine two methods to make continuous subsurface monitoring cheaper: estimating wave equation Green’s functions from passive vibration recordings, and measuring meter-scale strain rate profiles along fiber optic cables which may already exist in urban areas or can easily be installed along infrastructure.
These methods may make continuous high-resolution subsurface imaging possible where it was not previously, but there are challenges. The shift from particle velocity data of seismometers to axial strain rates recorded by fibers leads to different responses to the same source. Additionally, signals extracted from ambient seismic noise interferometry are masked by a fundamentally different receiver response. I use data from multiple fiber optic surface arrays: two with fiber directly coupled to the ground intended for permafrost thaw monitoring, and one with fiber sitting loosely in existing telecommunications conduits. Although these data look different, the arrival times of earthquakes at known times verify that the arrays record vibrations over a wide range of frequencies.
Care must be taken to understand and mitigate the effects of non-ideal anthropogenic noise when doing ambient seismic noise in urban areas and around infrastructure. I use interferometry to extract repeatable signals for near-surface geotechnical characterization near infrastructure, even between fiber channels that are not collinear. I investigate temporal stability and changes in signals extracted throughout large arrays in the presence of a changing subsurface and noise field. As ambient noise practitioners begin using denser arrays, the typical cross-correlation process can become expensive, so I propose a new algorithm for dispersion image calculation that is an order of magnitude faster and parallelizable.