A method for avoiding artifacts in the migration of deep seismic reflection data

Conventional wave-equation-based migration of deep seismic reflection data can produce severe artifacts, which appear as broad circular arcs or “smiles”, due to the existence of apparent truncations of reflections on the stack section arising from poor signal penetration, changes in orientation of the acquisition profile, and the existence of strong overlying lateral velocity variations. These artifacts limit the interpretation of deep seismic profiles, because they obscure weak reflections and reflection truncations that may, e.g., indicate the presence of subsurface faults. Here I present a new migration algorithm, in which each sample of the stack is migrated to a short linear segment whose position and dip are determined by its original position on the stack, an estimate of the local apparent dip at that point, and a user-specified migration velocity. No subjective interpretation of reflections on the stack section is required, and the algorithm produces no arc-like migration artifacts. The degree of lateral smearing can be easily controlled, allowing reflection truncations to be revealed. In practice, the algorithm is most effectively applied to data that have been coherency-filtered to remove low amplitude noise, which would otherwise be preserved.