Geology Reference
In-Depth Information
Seismic wavelets and resolution
Chapter
3
3.1 Introduction
A fundamental aspect of any seismic interpretation in
which amplitudes are used to map reservoirs is the
shape of the wavelet. This chapter presents introduc-
tory material relating to the nature of seismic wave-
lets; how they are defined, described and manipulated
to improve interpretability. Seismic resolution, in
terms of recognising the top and base of a rock layer,
is controlled by wavelet properties. However, owing
to the high spatial sampling of modern 3D seismic,
resolution in the broadest sense also includes the
detection of geological patterns and lineaments on
amplitude and other attribute maps.
In practice, the amplitude spectrum is calculated
from the seismic trace using a Fourier transform over
a given seismic window, usually several hundred
milliseconds long. It is assumed to first order that
the Earth reflectivity is random and that the wavelet
is invariant throughout the window. The amplitude
spectrum of the wavelet is then assumed to be a scaled
version of the amplitude spectrum of the seismic
trace. Amplitude spectra are also commonly esti-
mated over short seismic segments using a range of
different approaches (e.g. Chakraborty and Okaya,
spectral decomposition techniques for use in detailed
In addition to the amplitude spectrum, the other
piece of information that is needed to define the shape
of the wavelet uniquely is the relative shift of the sine
wave at each frequency (i.e. the phase). The wavelet in
Fig 3.1
shows frequency components that have a peak
aligned at time zero. Such a wavelet is termed zero phase.
As described in
Chapter 2
a zero phase wavelet is ideal
for the interpreter because it has a strong dominant
central trough or peak at zero time. If this is the wavelet
in a processed seismic dataset, then an isolated subsur-
face interface between layers of different impedance will
be marked by a correctly registered trough or peak
(depending on the sign of the impedance contrast and
the polarity convention used). This makes it fairly easy
to relate the seismic trace to the subsurface layering,
even in
3.2 Seismic data: bandwidth and phase
The seismic trace is composed of energy that has a
range of frequencies. Mathematical methods of Four-
decomposition of a signal into component sinusoidal
waves, which in general have amplitude and phase that
vary with the frequency of the component. An example
is the seismic wavelet of
Fig. 3.1
, which can be formed
by adding together an infinite set of sine waves with
the correct relative amplitude and phase, of which a
few representative examples are shown in the figure.
The amplitude spectrum shows how the amplitude of
the constituent sine waves varies with frequency. In
Fig. 3.1
there is a smooth amplitude variation with a
broad and fairly flat central peak. This is often the case
as the acquisition and processing have been designed
to achieve just such a spectrum. The bandwidth of the
wavelet is usually described as the range of frequencies
above a given amplitude threshold. With amplitudes
that have been normalised, such as those shown in
Fig.
3.1
, a common threshold for describing bandwidth is
half the maximum amplitude. In terms of the logarith-
mic decibel scale commonly used to present amplitude
data this equates to
situations with overlapping reflec-
tions. The interpretation is made more difficult if the
seismic wavelet is not symmetrical but, for example, has
several loops with roughly the same amplitude; then the
interference between adjacent closely spaced reflectors
will be difficult to understand intuitively. The import-
ance of zero phase wavelets to the interpreter becomes
clear when considering well to seismic ties (
Chapter 4
).
Zero phase is a condition that requires processing of the
seismic data (see
Section 3.6.2
).
'
thin-layered
'
23
6 dB (i.e. 20 log
10
0.5).
