Elastic full waveform imaging of Critical Zone

Xuejian Liu, Tieyuan Zhu

Research output: Contribution to journalConference articlepeer-review


Seismic refraction tomography has been the primary tool to build the P-wave velocity (Vp) model of Critical Zone. Conventionally, refraction tomography searches for an optimal Vp model to fit picked first-arrival traveltimes. Full waveform inversion (FWI) has then been utilized to improve the Vp model with better resolution. However, acoustic FWI methods for processing near-surface seismic datasets do not consider shear waves that exist in the data, and even worse mistakenly handle them as leaked P-waves, generating artifacts in the Vp model. In this abstract, we present a practical elastic full waveform inversion (EFWI) procedure to characterize not only a Vp model but also a S-wave velocity (Vs) model. Our practical EFWI workflow is briefly described as follows: First, we carefully window seismic data to preserve elastic waves mainly including P-wave, P-to-S (PS) converted wave, and S-wave. Second, we utilize a correlative misfit instead of the classic L2 misfit in the framework of EFWI to alleviate the interference of possible unreliable amplitudes in observed real data. Third, we perform sequential inversion of several low-pass frequency groups to avoid cycle skipping. Coherent experiments with synthetic data and the field data acquired from the Garner Run in the Susquehanna Shale Hills Critical Zone Observatory show that our EFWI approach can provide Vp & Vs models with high-resolution details.

Original languageEnglish (US)
Pages (from-to)617-621
Number of pages5
JournalSEG Technical Program Expanded Abstracts
StatePublished - 2021
Event1st International Meeting for Applied Geoscience and Energy - Denver, United States
Duration: Sep 26 2021Oct 1 2021

All Science Journal Classification (ASJC) codes

  • Geotechnical Engineering and Engineering Geology
  • Geophysics


Dive into the research topics of 'Elastic full waveform imaging of Critical Zone'. Together they form a unique fingerprint.

Cite this