Conclusions

In this short note, I have presented the extension of the velocity independent Stolt residual migration to common-azimuth data. Equations (5) can be used to obtain images corresponding to new velocities defined by a given ratio to the original velocity.

Such a residual migration method inherits the advantages and disadvantages of Stolt residual migration. Some of the most notable characteristics are:

• Speed: Because CA Stolt residual migration is essentially a Stolt migration, its cost is comparable to that of any Stolt migration.
• Constant velocity: Stolt migration assumes constant velocity. In theory, CA residual Stolt migration assumes constant velocity, too. In practice, however, if we residually migrate an image migrated with a variable velocity, we will obtain another image that has a corresponding variable velocity. In this case, however, the actual relationship between the velocities before and after residual migration is still not clarified and remains a subject for further research.
• Artifacts: Since Stolt CA residual migration is a frequency-domain method, it is susceptible to the same interpolation and wrap-around artifacts as the usual Stolt migration. Nevertheless, any technique used to reduce the Stolt artifacts Fomel (1998); Levin (1994) can also be used for CA residual migration with the same efficiency.