The result \ref{758196} showed the better representation of peak structure compared to the migration after just NMO correction. In both NMO, DMO migrated data, the structure below the peak was poor which could be the effect of high-velocity zones, such as basaltic structure. Peak structure has a steeper dip and overlapping problem relaxed. Also, overall boundary representation was more clear compared to NMO-migrated result. However, the structure near CDP 4200 was hampered. The dip angle of the structure was nearly perpendicular, and we think this could be fixed with higher dip filter value of Kirchhoff migration. Another problem discovered was the strong overmigrated result on the seafloor. We thought this may be caused by DMO but from the result of another team, there was the same problem discovered, just with NMO migration. Since this phenomenon was universal among the different tapes, we thought this could be the limitation of post-stack migration. And time migration method cannot work properly if the lateral variation is not mild where hyperbolic moveout is invalid. But Shatsky rise was constructed from igneous activity, and the paper notes the inhomogeneity of the structure, which can be challenging structure to be processed by time migration. So we would like to suggest pre-stack migration as a resolution which is essential because of the existence of conflicting dips. In fact, we tried to run it on the workstation but the memory did not allow it.
Suggested Pre-stack processing
According to \citet{Yilmaz2001} the rigorous solution for conflicting dips with different stacking velocities is pre-stack migration. Fig. \ref{926605} presents 2 suggested processing flows. The main workflow (right) include DMO correction (CMPDMO) in time domain and pre-stack Kirchhoff migration (MIGTX). In general DMO correction and pre-stack migration is performed in common offset gathers. However Paradigm accepts CMP gathers given that they have the same number of traces. the UNIFORM module does this data treatment. The red box on the left side of Fig. \ref{926605} provided additional suggested steps before stacking such as NMO removal to re-pick velocities followed by NMO correction with the re-picked velocities and a second pre-stack migration to further improve the image.