Seismic method - Geophysics for the Mineral Exploration Geoscientist - page 377

from each gather and the arrows indicate the locations of
their common midpoints.
For 2D surveys, the result is a seismic pseudosection, the
term being used because the vertical scale is in TWT, not
depth. For a 3D survey, a pseudovolume is produced. Com-
pare the cross-section and the pseudosection in Fig. 6.20d .
In this simple situation the two closely resemble each other,
although the vertical scale is in different units. As is
described next, this is not the case in real-world situations.

their actual form. The distortions arise because the ray-
paths are non-vertical ( Fig. 6.22 ), whereas each trace in the
stacked section is plotted vertically. Consequently, the dis-
play wrongly implies that the re ector or diffractor respon-
sible for the arrival is located directly below the source

-
detector midpoint. The various components of the stacked
data need to be moved to their correct relative positions to
produce a more geologically realistic section. This is
achieved using a process known as migration.
Depending on the type of migration used, the vertical
display axis may remain as two-way time (time migration)
or it may be converted to depth (depth migration). If the
data remain in two-way time they may subsequently be
depth converted. In both cases, information about velocity
variations in the subsurface is required, and the variations
in velocity must be completely known for the conversion to
be totally accurate. Again we are faced with the geophysical
paradox (see Section 1.3 ) , i.e. that the information required
to fully understand the geophysical response would render
the geophysical survey unnecessary. In practice, velocities
are estimated from whatever sources of information are
available, usually stacking velocities (see Correcting for
normal moveout in Section 6.5.2.4 ) and sonic logs (see
Section 6.6.9 ) , and collated to create a velocity model.
Fortunately, some types of migration are not too sensitive
to the choice of velocities. In contrast, depth conversion is
highly sensitive to this choice. So rather than degrade the
data through an incorrect choice of velocities, seismic
reflection data are usually left as a TWT pseudosection.
An example of depth conversion is presented by Zhou and
Hatherly ( 2004 ), who worked with data from near the
Oaky Creek coal mine, located in Queensland, Australia.
Migration may be applied pre- or post-stack and applied
to both 2D and 3D data. Here we describe only post-stack
migration in any detail. Pre-stack migration is more com-
plex than post-stack migration, but is based on the same
concepts. Also, there are many ways to implement migra-
tion, varying from the conceptually straightforward to
those of considerable mathematical complexity. Again we
describe an implementation that is easily understood at a
conceptual level. For a comprehensive discussion of the
subject of migration the reader is referred to Yilmaz
( 2001 ). Lillie ( 1999 ) provides a useful,

6.5.2.5 Migration and depth conversion
As described above, the conversion from finite- to zero-
offset data transforms the seismic section into a form that
begins to resemble the subsurface geology; for example,
compare the cross-section and the stacked section in
Fig. 6.20d . In fact, normal-incidence data correctly
represent the geometry of the subsurface only when the
responses are re

ections from a horizontal interface and
the velocity is constant throughout the entire subsurface
( Fig. 6.22a ).
Figures 6.22b to g illustrate how the subsurface is mis-
represented in the stacked data when these limiting condi-
tions are not satis ed. For example:

￿ Lateral changes in seismic velocity create apparent verti-

cal displacements in re ectors. Higher velocity decreases
the re ection travel times producing an apparently shal-
lower re ector ( Fig. 6.22b ). The effect is known as
velocity pull-up. Note the phantom diffractor where the
velocity changes.

Diffractions appear as concave-downwards events, but
they originate from

￿

sources and from abrupt
terminations or changes in the dip of a reflecting horizon
( Figs. 6.22c and d ) , note the diffractors at the edges of
the steps.

' point '

Dipping re ectors are in the wrong place. Their dip is

reduced, they are lengthened and they are laterally dis-
placed down-dip ( Fig. 6.22e ). These effects are exacer-
bated with increasing dip.

￿

For the case of folded re ectors ( Fig. 6.22f ) , the diffrac-
tion and dipping re ector responses combine to create
the most obvious discrepancy between actual re ector
geometry and the seismic response. Anticlines appear
too open with geologically impossible cross-cutting
events on their flanks if folding is tight. Worse still are
synclines, where a

￿

less technical,

description.

' '

bow-tie pattern is formed ( Fig. 6.22g ).

Post-stack time migration
It was previously demonstrated (see Section 6.3.4.3 ) that a
reflecting interface is equivalent to a series of closely spaced

Clearly, realistically complex geological structures will
produce seismic responses that show little resemblance to