Using common-reflection-surface stack for enhanced near-surface seismic reflection imaging: Examples from consolidated and unconsolidated environments

Abstract

The small geophone spacing and spread lengths commonly used to investigate ultrashallow layers in conjunction with the limited bandwidth of surface impact sources do not allow for clear separation of different seismic arrivals. The interference of events and the high noise levels due to near-source offsets make shallow seismic data processing a challenging task. The wavefront attributes of seismic waves commonly used in conventional seismic processing are suggested to improve near-surface seismic reflection imaging. These wavefront attributes are often used as the stacking parameters in multidimensional time imaging methods such as common-reflection surface (CRS). It is shown in this paper that the CRS method can improve near-surface seismic data processing by enhancing: (1) unstacked gathers via a CRS-based local stacking scheme, (2) semblance picking for velocity model building, and (3) zero-offset stacked data via the CRS global stacking. The CRS-based local stacking can mitigate the loss of data associated with optimum-windowing-based muting of coherent noise. The CRS local stacking infills the muted zones via its robust data interpolation and regularization features and enhances the image quality. Furthermore, the CRS-based enhanced data allow for improved near-surface velocity model building by producing higher coherence and more focused semblance peaks. The CRS global stacking is shown to further smooth and provide more continuity of events in the zero-offset data. Applications to a synthetic and two field data sets collected from high- and low-velocity environments indicate the efficiency and feasibility of the proposed approach.