Page 528 - Acquisition and Processing of Marine Seismic Data
P. 528

11.6 POSTSTACK MIGRATION                          519
























           FIG. 11.28  Ray path of a prism wave reflected from the flank of a salt dome.


           account for extreme lateral velocity variations  11.6 POSTSTACK MIGRATION
           and provide excellent amplitude-preserving
           results with a higher S/N compared to other     In poststack migration, the seismic data is
           migration algorithms. It is also suitable to use  migrated after it has been stacked. It can be
           with wide azimuth (WAZ) seismic data. The    implemented in the time or depth domains for
           most important disadvantage of RTM, however,  2D and 3D seismic data. The most important
           is its extremely high computational cost; it takes  advantage of the poststack migration originates
           several times longer to run the RTM algorithm  from stacking: removal of significant amounts of
           than a conventional finite-difference algorithm,  coherent and random noise embedded in the
           which has prevented RTM from being used      prestack data, suppression of multiple reflec-
           widely by the seismic industry until recent  tions, increased S/N ratio before migration,
           years. To overcome the high computational cost,  and inexpensive processing due to the reduced
           RTM can also be implemented in the time-     data volume after stacking.
           wavenumber (T-K) domain, known as reverse       During the migration process, what we have is
           time TK migration, which is relatively faster  only the seismic traces and RMS velocity field
           and more economical, but accounts for only ver-  associated with the input data. The correct reflec-
           tical velocity variations.                   tion point of a reflection from a dipping reflector
              Accuracy of the RTM output also depends on  in a 2D earth model can be found by using the
           the accuracy of the interval velocity field used  arrival times and propagation velocities of the
           for the migration. A slower velocity causes  reflected events, as schematically illustrated in
           frowns and the inclined events are not relocated  Fig. 11.31. Considering the reflection event with
           to their true lateral positions, while a faster  an arrival time of t(x) at trace number 7, for
           velocity results in overmigration, resulting in  instance, we can conclude that this event is orig-
           smiles. Fig. 11.29 shows the collapse of a theoret-  inated from a reflection point in the subsurface
           ical diffraction hyperbola in a 1500-m/s constant  along a semicircle of radius Vt/2 (red dashed
           velocity medium using RTM. Fig. 11.30 shows a  curve), centered at the zero-offset location of that
           poststack reverse time migration output of a  trace indicated by X, where V is the propagation
           marine stack section.                        velocity of the reflection event. This implies that
   523   524   525   526   527   528   529   530   531   532   533