Page 437 - Acquisition and Processing of Marine Seismic Data

428 9. VELOCITY ANALYSIS

Thesuitable velocityfor a primaryreflectionisthe stack stripes are plotted to pick the velocities
one that perfectly flattens the hyperbola of that that produce the best stack sections. In Fig. 9.4,
specific reflection. For instance, NMO correction NMO correction is applied to 24 successive
is applied to the input CDP for different velocities CDPs for different velocities ranging from
ranging from 1480 to 2400 m/s. A reflection 1480 to 2500 m/s, and prepared stack sections
hyperbola indicated by A is flattened for for each velocity are plotted side by side. The
1480 m/s RMS velocity, while reflections B and most suitable velocities are those that produce
C are flattened for 1550 and 1650 m/s, respec- the stacks with highest amplitudes. For instance,
tively. Therefore, the estimated velocities for 1480 m/s velocity produces the highest ampli-
events A, B and C are 1480, 1550 and 1650 m/s, tudes around A in the first stack stripe, while
respectively. If the suitable velocities are picked the highest amplitudes in the second stripe are
for each distinctive reflection hyperbola on the obtained for 1550 m/s velocity around B. There-
CDP gather, then a velocity function can be fore, the suitable velocities for reflections at zero
obtained for that specific CDP, indicated by the offset times A and B are 1480 and 1550 m/s,
dashed blue line in Fig. 9.3. respectively.
Each reflection hyperbola is perfectly flat- If the suitable velocities are picked for each
tened if an NMO correction is applied with that distinctive high-amplitude zone on the stack
reflection’s dedicated seismic velocity, provid- stripes, then a velocity function can be obtained
ing that the reflections are hyperbolic on CDP as indicated by the dashed blue line in Fig. 9.4
gathers. The hyperbola is overcorrected if the and is assigned to the location of the central
velocity is low and is undercorrected if the CDP gather. This function can be derived for a
velocity is high after NMO correction. range of CDP gathers by testing several different
A velocity function for the whole recording time velocity values as shown by the solid blue line in
on the CDP can be derived for that specific Fig. 9.4. If this constant velocity stacking proce-
ensemble by testing several different velocity dure is repeated for several successive sets of
values, as shown by the solid blue line in CDPs along the line, a 2D velocity field can be
Fig. 9.3. If this velocity scanning procedure for obtained for the whole seismic line.
constant velocities is repeated for several CDPs Determination of velocity range and incre-
along the line, a 2D velocity field can be obtained ment is important in CVS calculations and must
for the whole seismic line. be selected carefully depending on the require-
ments of the input seismic dataset. Dipping
events and sideswipes require abnormally high
9.2.2 Constant Velocity Stack
stacking velocity values. If such high velocities
The theoretical basis of the velocity estima- are observed in relatively shallow parts of the
tion by the constant velocity stack (CVS) method data, side effects may be suspected. Practical
is that an optimal velocity function derived from applications indicate that the CVS method is par-
the velocity analysis is the one that yields the ticularly useful in geologically complex areas.
best stack section. Therefore, stack sections con-
sisting of 8–24 CDPs, stacked after NMO correc- 9.2.3 Velocity Spectra
tions for different constant velocities, can be
used to pick the optimal velocity values. Velocity spectra are the most common
Fig. 9.4 schematically shows the constant veloc- method in deriving 2D and 3D velocity func-
ity stack (CVS) procedure. A number of succes- tions from multichannel seismic data. This
sive CDPs are stacked after NMO correction method is based on the variations in the curva-
with a range of constant velocity values. These ture of the reflection hyperbolas due to the