Walkaway-VSP velocity modeling based on elastic full waveform tomography

Abstract Walkaway Vertical Seismic Profiling (VSP) can receive three-component full elastic wavefield information from a well, including up-going and down-going P- and SV-waves. This is beneficial for constructing accurate subsurface velocity models and improving the processing and interpretation quality of seismic data. Existing VSP velocity modeling methods are primarily based on traveltime tomography, and amplitude information has not been fully utilized. This is because the pre-stack preprocessing of VSP data often distorts the polarization characteristics of P- and SV-waves, resulting in insufficient fidelity of the P- and SV-wave amplitudes used for inversion. This study proposed a time-domain elastic full-waveform tomography (EFWT) strategy for Walkaway-VSP. Numerical tests on the VSP data of the Marmousi model showed that the EFWT method can produce stable inversion results for the P- and SV-wave velocity models even when the initial models are inaccurate. For application to real Walkaway-VSP data, we established a pre-stack processing flow for the EFWT that can maintain the polarization characteristics of vertical (Z)- and radial (R)-component seismic waves. The calibration results of the corridor stack section show that the P- and SV-wave velocity models generated by the EFWT can align PP- and PSV-wave images accurately at the correct depth. This highlights the feasibility and potential of the proposed EFWT strategy for VSP applications.