Investigation of DMO algorithms during test-line processing: some recommendations

Abstract

The Dip Movement processor in seismic data processing, or partial migration before stack is an auxiliary data processing correction that attempts to improve the quality of a seismic stack in the presence of reflection point smearing and conflicting dips. Performed correctly, velocity analysis after DMO is supposedly independent of dips and thus would allow an easier decision in, and perhaps more correct velocity pick. DMO algorithms available sometime back in this region are essentially Fourier transform methods, usually with some logarithmic stretch formulation or Integral/Summation (Kirchhoff­ style) methods with provisions for spatial aliasing and dip constraints. Fourier transform methods are efficient and best applied to seismic data that are uniformly sampled in space. Kirchhoff-style methods are implemented instead with one input and one output trace at a time and are well suited for irregular survey geometries, missing shots, wide swaths, large variations in source-receiver distances and azimuths, large cable feathering angles, etc. We recommend during test-line evaluation to compare the velocity spectrum at a preselected CDP location without DMO from the same location with DMO. A velocity pick should be evident in the latter. We recommend next to subtract (a) the stacked section and (b) migrated section without DMO from the same with DMO. Assuming all non-DMO processing are identical, the difference sections should contain no horizontal reflections i.e. DMO should not in any way alter horizontal reflections. Diffraction hyperbolas will be better preserved with DMO in (a). Fault definitions are enhanced after migration with DMO in (b) because of this preservation. Lastly DMO should not be used solely for suppressing high velocity linear noise and lessening back scattered energy. Other filtering options are available.