Our offset-dependent residual moveout picking can accurately describe complex events to yield accurate high-resolution velocity updates.

When common image gathers (CIGs) display moveout which coincides with low order polynomial terms, polynomial-based residual moveout (RMO) picking may effectively describe the smooth curvature. In situations where the curvature is complex, the polynomial assumption may often be inaccurate.

Complicated areas, such as those with high anisotropy, or showing heavy faulting, will often yield events with multiple turning points, which may actually be made worse by polynomial-based flattening.

The offset-dependent picking method developed by TGS represents the RMO as a continuous displacement field rather than as a series of discrete events. A multiscale, constrained solver is used to estimate the displacement field for each CIG.

A gradient constraint prevents reliance on crossed events and areas where waveform distortion leads to excessive far-offset event deformation. The events with maximum coherency are computed from the derived displacement field. This approach helps ensure consistency by allowing neighboring events to help constrain RMO picks.

Red curves: single-parameter picks
Green curves: offset-dependent picks
 

Hilburn, Guy, Yang He, Zengjia Yan and Francis Sherrill, 2014, High-resolution tomographic inversion with image-guided preconditioning and offset-dependent picking: 84th Annual International Meeting, SEG, Expanded Abstracts, 4768-4772.