Page 541 - Acquisition and Processing of Marine Seismic Data

532 11. SEISMIC MIGRATION

indicated by a dashed rectangle. In these zones, • Velocity analysis is improved, and the
it is not possible to pick velocity values charac- velocity field used for stacking becomes
terized by both enclosures with the same zero- usable for migration.
offset time, and the processor picks a velocity • Abnormally steep coherent noise is
value that conforms to the common trend of eliminated and fault planes become much
the semblance contours. After DMO processing, clearer.
however, the conflicting dips issue is solved and • DMO has no effect on horizontal events.
a more accurate velocity picking is possible, as in
Fig. 11.42B, which provides that the amplitudes
of the conflicting events are better preserved in 11.11 WHICH MIGRATION TO USE?
the stack section.
Fig. 11.43 compares the results of poststack
Different migration implementations applied
time migration (Fig. 11.43A) and DMO proces-
to prestack or poststack data, utilizing different
sing (Fig. 11.43B) utilizing the flow shown in
theoretical basis and algorithms, exist on the
Fig. 11.41. The differences between the conven-
market today, and each migration type has its
tional poststack time migration and DMO
own advantages and solution capabilities as
outputs are distinctive, especially along the
well as shortcomings. Table 11.2 compares gen-
fault planes, which are much sharper and more
eral characteristics of these different methods.
distinct in the DMO output. Horizontal layers
Although cost and data characteristics (such as
are not affected by DMO processing.
the quality of the input data, lateral velocity var-
Fig. 11.43C shows the Kirchhoff prestack time
iations, structural complexity, maximum dip,
migration image to compare the output of post- etc.) are the most effective parameters on the
stack migration followed by DMO processing. selection of the suitable algorithm to apply pre-
One can conclude that the DMO with a post- stack or poststack migration, the selection of
stack migration output is almost identical with optimal migration type appropriate for the input
the prestack migration image shown in data commonly depends on the experience of
Fig. 11.43C. the processor.
Deregowski (1986) summarizes the effects of
DMO processing on input seismic data: Farmer et al. (1993) proposes the characteris-
tics of an ideal migration method; no algorithm
• DMO moves each trace to its correct zero could do all of them in practice. That is, an out-
offset location; hence each common offset standing migration must
section becomes identical to a zero-offset
• handle full range of dips.
section.
• compensate for large velocity changes such as
• Crossline ties between the seismic data in a
>40%.
particular survey area are improved in the
• relocate all of the reflection events to their
case of 2D acquisition.
true subsurface locations.
• CDP smearing is eliminated and all CDPs
• preserve amplitude and phase information of
have reflections from the same common
the input data.
reflection points.
• generate no artifacts, spurious events
• Poststack time migration becomes almost
or noise.
identical to prestack time migration.
• be computationally efficient and fast.
• Stacking velocities become dip independent,
and the events with conflicting dips are Fig. 11.44 schematically summarizes the
stacked correctly. proper application of different migration