Seismic diffraction separation in the near surface: Detection of high-contrast voids in unconsolidated soils

Abstract

Seismic diffractions carry the signature of near-surface high-contrast anomalies and need to be extracted from the data to complement the reflection processing and other geophysical techniques. Because diffractions are often masked by reflections, surface waves, and noise, careful diffraction separation is required as a first step for diffraction imaging. A multiparameter time-imaging method is used to separate near-surface diffractions. The implemented scheme makes use of the wavefront attributes that are reliable fully data-derived processing parameters. To mitigate the effect of strong noise and wavefield interference in near-surface data, our workflow incorporates two wavefront-based parameters, dip angle and coherence, as additional constraints. The output of the diffraction separation is a time trace-based stacked section that provides the basis for further analysis and applications such as time migration. To evaluate the performance of the proposed wavefront-based workflow, it is applied to two challenging field data sets that were collected over small culverts in very near-surface soft soil environments. The results of the proposed constrained workflow and the existing unconstrained approach are presented and compared. The proposed workflow demonstrates superiority over the existing method by attenuating more reflection and noise, leading to improved diffraction separation. The abundance of unmasked diffractions reveals that the very near surface is highly scattering. Time migration is carried out to enhance anomaly detection by focusing the isolated diffractions. Although strong diffractivity is observed at the approximate location of the targets, there are other diffracting zones observed in the final sections that might bring uncertainties for interpretation.