Geoscience Reference
In-Depth Information
the data to a zero-offset section, thus allowing conventional post-stack migration to be
applied. This is how the pre-stack migration process in step 17 is applied. In this step,
we are not attempting to get a final migration; instead we are trying to move events with
potential stacking conflicts apart sufficiently so that we can determine a unique stacking
velocity. This allows us to use a computationally very fast constant velocity form of
migration - in this example performed with a velocity of 1600 m/s. It is important that
this step does not use too high a velocity since then events can disappear completely
if they are steeply dipping. Generally one uses a velocity that is close to the slowest
seen in the section (1480 m/s for water velocity). Step 18 removes the approximate
correction for NMO applied during the DMO process (step 16) and this allows for the
detailed velocity analysis to be performed and applied in steps 19 and 20. The data
are usually saved to tape after DMO and the removal of the initial NMO correction
to allow any later re-processing to begin from this stage. It is common practice to
perform the final velocity analysis at a spatial interval of 250-500 m though this too is
reducing in the desire to retain the higher frequencies. The spacing needed depends on
the variability in the velocity field; large variations require more frequent analysis than
more gradual changes. For really detailed work, the advent of better automatic velocity
picking algorithms means that it is possible to pick velocities for every CMP location
within the target area.
After the final velocity analysis and moveout correction the data are stacked. Stacking
together traces that contain the same reflection information both improves the signal to
(random) noise content (by the square root of the number of traces stacked) and reduces
any residual coherent noise such as multiples which stack at velocities different from
the primary events. During stacking, mutes (zeroing the data within specified zones)
are applied to the data to ensure that NMO stretch is not a problem and that any residual
multiples left on the near-offset traces do not contaminate the stacked section
(fig. 2.35)
.
There may be some amplitude variation with offset (AVO) effects in the data, which
can be used as hydrocarbon indicators, so this information is retained by stacking the
data over selected offset ranges.
Chapter 5
contains more details on the physics and
use of AVO data. An offset to angle relationship is used that depends on the subsurface
velocities to define mute patterns to produce two or more angle stacks as well as the
stack over all usable offsets. All processing from this point onwards needs to be applied
to each of the individual stacked sections.
The result of the stacking after step 17 is a number of partially migrated sections
requiring the removal of the constant velocity migration. This is done by step 22, an
inverse migration or demigration using the same constant velocity as used for the initial
migration.
Step 23 adds a static correction based on the depth of the source and the receiver so
that the datum is now at sea-surface. Step 24 removes some of the effects of the pattern
left in the data from the acquisition variability. Since different CMPs will contain a
different combination of traces in a regular pattern this may show itself in the final
