← Back to Sessions

Seismic Elastic Waveform Inversion for Characterizing Vp and Vs Models of Critical Zone

Session: Critical Zone, Environmental and Cryospheric Seismology

Type: Oral

Date: 4/22/2021

Presentation Time: 05:30 PM Pacific

Description: 

Seismic refraction methods play an important role in providing seismic properties of CZO structure. The important data processing following the seismic data collection is to conduct refraction tomography using picked traveltimes to provide a well smooth background P-wave velocity model that helps constrain the fluid properties (e.g., porosity) and fractures. While the P-wave velocity model by traveltime tomography provides a low resolution image by only preserving low wavenumber information, our previous work on acoustic full waveform inversion of early arrivals showed promising to invert high resolution P-wave velocity model of CZO subsurface (Huang et al., 2019). However, the amplitude-versus-offset (AVO) effect and residual S-wave energy in early arrival waveform could not be reconciled by acoustic full waveform inversion. Here we design an elastic waveform inversion procedure to tackle this issue by forward modeling of elastic waves to capture the AVO effect and S-waves. The nonlinearity of the elastic waveform matching is primarily caused by unknown source energy, complex AVO properties, and some errors in the initial velocity models. Several strategies are used to alleviate this nonlinearity. First, elastic seismic wave modeling data will be able to take the AVO effect and S-wave into account. We then estimate a source wavelet from high quality early arrivals in the seismic data. The waveform matching will adopt an optimized cross-correlation based misfit and the total-generalized variation based regularization. We demonstrate our elastic waveform inversion method using synthetic experiments and field data from the Garner Run in the Susquehanna Shale Hills Critical Zone Observatory.

Presenting Author: Xuejian Liu

Student Presenter: No

Authors