Geology Reference
In-Depth Information
angles of incidence than others) and the imperfect
fidelity of the recording equipment. Marine seismic
has the advantage that sources and receivers are very
repeatable in their characteristics. This is not true for
land data, where the coupling of source and receivers
to the ground may be quite variable from one shot to
another, depending on surface conditions. However,
these effects can be estimated and allowed for by the
seismic processor.
Some amplitude effects are features of the subsur-
face that are of little direct interest and ideally would
be removed from the data during processing if pos-
sible; these include divergence effects, multiples, scat-
tering, reflection curvature and rugosity, and general
superimposed noise. Depending on the individual
data set, it may be quite difficult to remove these
without damaging the amplitude response of interest.
For example, processors often have difficulty in
attenuating multiple energy whilst preserving the
fidelity of the geological signal. Other effects on seis-
mic amplitude, for example related to absorption and
anisotropy, might be useful signal if their origin were
better understood. The processor clearly faces a tough
challenge to mitigate the effects of unwanted acquisi-
tion and transmission factors and enhance the geo-
logical content of the data.
Interpretation of seismic amplitudes requires a
model. A first order aspect of the seismic model is
that the seismic trace can be regarded as the convolu-
tion of a seismic pulse with a reflection coefficient
related to contrasting rock properties across rock
boundaries. This idea is an essential element in seismic
processing as well as seismic modelling. Given that the
seismic processor attempts to remove unwanted
acquisition and propagation effects and provide a
dataset in which the amplitudes have
physical terms. In the context of the small stresses and
strains related to the passage of seismic waves, rocks
can be considered perfectly elastic (i.e. they recover
their initial size and shape completely when external
forces are removed) and obey Hooke
s Law (i.e. the
strain or deformation is directly proportional to the
stress producing it). An additional assumption is that
rocks are to first order isotropic (i.e. rocks have the
same properties irrespective of the direction in which
the properties are measured). Experience has shown
that in areas with relatively simple layered geology
this isotropic/elastic model is very useful, being the
basis of well-to-seismic ties (
Chapter 4
) and seismic
inversion (
Chapter 9
).
There are, however, complexities that should not
be ignored. These complexities can broadly be char-
acterised as (a) signal-attenuating processes such as
absorption and scattering and (b) anisotropic effects,
related to horizontal sedimentary layering (vertical
polar isotropy) and vertical fracture effects (azimuthal
anisotropy) (see
Section 5.3.7
). One effect of absorp-
tion is to attenuate the seismic signal causing changes
in wavelet shape with increasing depth and this is
usually taken into account. Attenuation is difficult to
measure directly from seismic but at least theoretically
this information could have a role in identifying the
etical understanding about anisotropy (Thomsen,
ledge of how to exploit it for practical exploration
purposes. One problem is the availability of data with
which to parameterise anisotropic models. Practical
seismic analysis in which anisotropic phenomena
are exploited has so far been restricted to removing
horizontal layering effects on seismic velocities and
moveout in seismic processing (i.e. flattening gathers
particularly at far offsets) (
Chapter 6
) and defining
vertical fracture presence and orientation (
Chapter 7
).
'
'
'
correct
relative
scaling, the interpreter
s approach to modelling tends
(at least initially) to focus on primary geological signal
in a target zone of interest. Of course, one eye should
be on the look out for
'
remaining in the section
that has not been removed (such as multiple energy
and other forms of imaging effects). The presence
of such effects might dictate more complex (and
more time consuming) modelling solutions and can
often negate the usefulness of the seismic amplitude
information.
From a physical point of view the geological com-
ponent of seismic reflectivity can be regarded as
having various levels of complexity. For the most part,
the geological component can be described in simple
'
noise
'
2.3.1 The convolutional model, wavelets
and polarity
The cornerstone of seismic modelling is the convolu-
tional model, which is the idea that the seismic trace
can be modelled as the convolution of a seismic pulse
with a reflection coefficient series. In its simplest form
the reflection coefficient is related to change in acoustic
impedance, where the acoustic impedance (AI) is the
product of velocity (V) and bulk density (
7
ρ
)(
Fig. 2.6
):
