Seismic Waveform Tomography Using Multi-component Data at a Shallow Groundwater Contamination Site

Seismology is becoming a more commonly used technique for characterizing contaminated groundwater aquifers and other near-surface problems. Currently, active-source seismic studies typically focus on the P- wave field, however SH-wave data may offer benefits such as having higher intrinsic resolution at a given frequency and being relatively independent of pore fluid, potentially allowing for a less ambiguous determination of lithology. We have acquired a 2D, nine-component seismic dataset over a shallow (~10 m depth) contaminated aquifer at Hill Air Force Base, Utah. To date, work has focused on two components: a vertical source and receiver (P-wave dataset), and a horizontal transverse source and receiver (SH-wave dataset). Previously, the P- and SH-wave datasets have been processed using post-stack depth migration and traveltime inversion to produce a structural image and velocity model. These previous results indicate that the quality of the SH-wave and P-wave datasets are comparable. Although noisy, both datasets are sufficient to produce a useful image of the subsurface. In the current study we use full waveform tomography to produce high-resolution velocity models of the shallow subsurface for interpretation. Full waveform tomography (FWT) has so far been developed primarily under the acoustic approximation, and several previous studies have used acoustic FWT to image single-component surface seismic data. To our knowledge there have been no previous attempts to use FWT on an SH-wave surface seismic dataset. We have developed an alternate version of the acoustic FWT method based on the SH-wave equation. Acoustic and SH-wave FWT are applied to the P- and SH-wave datasets respectively. The quality of the resulting images will be compared to determine what benefits are obtained using an SH dataset for FWT. In particular, we expect SH-wave FWT to provide better resolution immediately below the water table where the S-wavelength is much shorter than the P-wavelength. Furthermore, the P- and S-wave velocity models may be jointly interpreted to further constrain elastic properties in the subsurface beyond the velocity information provided by acoustic tomography alone. This should enable us to enhance our understanding of near- surface structure and lithology at the contamination site and provide valuable counsel for efficient targeting of remediation efforts.