Surface-wave tomography for near-surface characterization with continuous wavelet transform for two-station cross-correlation

The multichannel analysis of surface waves (MASW) is the most popular approach to estimate dispersion curves of surface waves for near-surface S-wave velocity characterization. The MASW is robust because fundamental mode of surface waves could be distinguished from noise such as body waves or higher modes in the frequency-wavenumber domain. However, sharp lateral variation is difficult to accurately retrieve by MASW-based approach. We proposed a workflow to effectively estimate lateral variation of shallow subsurface from surface-wave tomography using active-source seismic data. In our approach, we stacked cross-coherence between two stations from different shot data to improve the stability of phase velocity estimation. We then applied continuous wavelet transform (CWT) for cross-coherence data to reduce the influence of noise by applying a time-domain filter. Lateral variation of phase velocities was finally estimated from dispersion curves between two stations by tomography. Numerical experiments demonstrated that our CWT-based approach could be used to accurately estimate phase velocity between two stations reflecting lateral variation of the simulated model. The lateral boundary could be well retrieved in the velocity model estimated by tomography. Application of our approach to field data indicated the high stability of phase velocity estimation between two stations and tomography analysis, compared to conventional MASW-based approach.