Y. C. Kim, J. R. Krebs
Integral based common-offset time migration is a robust and cost-effective tool for migration velocity analysis. In practice this method is not applied iteratively, and travel times are often computed without properly accounting for ray bending in a layered earth. For dipping reflectors these practices can lead to erroneous velocity estimates.
The velocity of an event measured by common-offset time migration velocity analysis is fairly insensitive to the initial migration velocity even for dipping reflectors. However, the time at which that event occurs is very sensitive to the initial migration velocity. This is because errors in the initial velocity result in under or over migration of the event. For example, if the initial velocity is too slow, the data will be under migrated, and the velocity event from a dipping reflector would appear at approximately correct velocity but at a time that is too early. When velocities increase with depth, this error would cause us to pick a velocity function that is faster than the correct function. However, this new velocity function will generally be closer to the correct velocity than the initial migration velocity. Convergence to the correct velocity usually occurs after a few iterations.
It is important to account for ray bending when computing travel times, even for time migration. Coincident reflections having different dips focuses at different velocities if ray bending is ignored. Thus, we must choose between the velocity that optimizes one dip or the other but not both. By properly accounting for ray bending, all dips are focused at the same velocity, producing significant improvements in the final image.
AAPG Search and Discovery Article #90981©1994 AAPG Pacific Section Meeting, Ventura, California, April 27-29, 1994