Professional Documents
Culture Documents
Hamid N. Alsadi
Seismic
Hydrocarbon
Exploration
2D and 3D Techniques
Advances in Oil and Gas Exploration &
Production
Series editor
Rudy Swennen, Department of Earth and Environmental Sciences,
K.U. Leuven, Heverlee, Belgium
The book series Advances in Oil and Gas Exploration & Production pub-
lishes scientic monographs on a broad range of topics concerning geo-
physical and geological research on conventional and unconventional oil and
gas systems, and approaching those topics from both an exploration and a
production standpoint. The series is intended to form a diverse library of
reference works by describing the current state of research on selected
themes, such as certain techniques used in the petroleum geoscience business
or regional aspects. All books in the series are written and edited by leading
experts actively engaged in the respective eld.
The Advances in Oil and Gas Exploration & Production series includes
both single and multi-authored books, as well as edited volumes. The Series
Editor, Dr. Rudy Swennen (KU Leuven, Belgium), is currently accepting
proposals and a proposal form can be obtained from our representative at
Springer, Dr. Alexis Vizcaino (Alexis.Vizcaino@springer.com).
Seismic Hydrocarbon
Exploration
2D and 3D Techniques
123
Hamid N. Alsadi
Data Processing Section
Ministry of Oil
Baghdad
Iraq
The favorable reception of the rst edition of this book (seismic Exploration,
published by Birkhauser Verlag in 1980) stimulated my belief in the need of
an updated book that includes the advances in the techniques which have
taken place during the past three decades. In preparing the present updated
volume, I have taken into consideration the remarks and suggestions of the
users of the 1980 edition from both of the academic and industrial work
domains.
Since 1980, when the rst edition of this book was published, great
developments in the seismic exploration technology have taken place. These
developments have occurred in all of the three exploration phases: acquisi-
tion, processing, and interpretation techniques. The most prominent advances
which have taken place in these years are the widespread implementations
of the 3D surveying, pre-stack migration, and growing interpretation tech-
niques in both structural and stratigraphic exploration. As it is familiar with
the exploration geophysicists, this subject (seismic exploration) is fully dealt
with in many original and authentic internationally known text books. In this
publication, no new subjects added to those found in the other standard books
which are well known in the geophysical library. In fact, these and other
related scientic papers and research reports formed solid references for the
present book. There are, however, differences in the design and presentation
approach.
In its design, the book is intended to be a comprehensive treatise of the
seismic exploration tool, addressing audiences in both of the academic and
industrial establishments. It is made up of 12 chapters covering the basic
aspects of the seismic reflection exploration subject, starting with the basic
theory, followed by the applied data acquisition technology, and ending with
the processing and interpretation. In presenting the subject matter, emphasis
is made on the practical aspects of the subject, using clear and simplied
presentation, avoiding excessive descriptions and unnecessary lengthy
comments. Numerous illustration gures (>390 gures) have been used
throughout the book to aid in clarifying the concepts and procedures involved
in any standard seismic exploration survey. In this way, the book can be
considered as a very useful introductory teaching manual for university
students taking seismic reflection exploration as part of a postgraduate
course.
v
vi Preface
The chapters of the book are sequenced in the order of the activities
normally executed in a standard seismic exploration survey: eld acquisition,
processing, and interpretation. Chapter 1 is an introductory chapter in which
a brief historical note and short review of the geophysical exploration
methods are given. This is followed by four chapters covering the theoretical
aspect of the subject including basic principles and denitions of seismic
waves, with a special chapter assigned for the seismic wave propagation
velocity. The propagation phenomena, reflection, diffraction, transmission,
and refraction, are dealt with in Chaps. 4 and 5. The following two chapters
are devoted to the two main tools applied in seismic exploration, namely the
2D and 3D surveying techniques.
Due to its important role in understanding of the processing steps applied in
seismic data processing, a chapter (Chap. 8) is assigned solely for the seismic
signal. This chapter is structured on the theme of considering the seismic
reflection wavelet as a propagating signal in the same way as the electro-
magnetic signal is treated by the communication theory applied in electro-
magnetic wave propagation. Including a chapter on seismic signals, preceding
the processing chapters, is a feature by which this book has deviated from
other conventional publications. Data processing is presented under two
headings: processing tools (Chap. 9) and the normally applied processing
sequence (Chap. 10). Chapter 11 covers some specialized seismic exploration
tools sometimes used in support of the conventional seismic reflection and
refraction surveying. The book is concluded with Chap. 12 on interpretation.
I would like to express my gratitude to my wife Asira and my sons (Eng.
Majid, Ph.D., Eng. Muhannad, M.Sc., Thurayah, B.Sc., and Eng. Mahir,
B.Sc.), for their continuous support and help throughout the past three years.
My work in the writing of the book has incurred an additional burden to the
family especially during the abnormally difcult times, which our country
has experienced in the past twelve years.
1 Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1
1.1 Historical Review . . . . . . . . . . . . . . . . . . . . . . . . . . 1
1.1.1 Elasticity and Seismic Waves (Bth 1979) . . . 1
1.1.2 Earthquake Seismology (Richter 1958). . . . . . 2
1.1.3 Exploration Seismology
(Telford et al. 1990) . . . . . . . . . . . . . . . . . . 2
1.1.4 Summary of Exploration Seismology History
(Sheriff and Geldart 1995) . . . . . . . . . . . . . . 3
1.2 The Geophysical Exploration Project . . . . . . . . . . . . . 3
1.3 Geophysical Exploration Methods . . . . . . . . . . . . . . . 4
1.3.1 The Seismic Method . . . . . . . . . . . . . . . . . . 4
1.3.2 The Gravity Method . . . . . . . . . . . . . . . . . . 4
1.3.3 The Magnetic Method . . . . . . . . . . . . . . . . . 6
1.3.4 The Electical Method. . . . . . . . . . . . . . . . . . 7
1.3.5 The Radioactivity Method . . . . . . . . . . . . . . 9
1.4 Oil Well Drilling. . . . . . . . . . . . . . . . . . . . . . . . . . . 9
1.4.1 Drilling of the Exploration-Well . . . . . . . . . . 10
1.4.2 The Oil-Well Rotary Drilling . . . . . . . . . . . . 10
1.4.3 The Drill-Hole and Well Casing . . . . . . . . . . 11
1.5 Well Geophysical Logging . . . . . . . . . . . . . . . . . . . . 11
1.5.1 Electrical Logging . . . . . . . . . . . . . . . . . . . . 12
1.5.2 Radioactivity Logging . . . . . . . . . . . . . . . . . 14
1.5.3 Acoustic Logging . . . . . . . . . . . . . . . . . . . . 16
1.5.4 Log Interpretation . . . . . . . . . . . . . . . . . . . . 19
1.6 Latest Developments in Well Logging . . . . . . . . . . . . 19
1.6.1 Logging While Drilling Technique . . . . . . . . 20
1.6.2 Borehole Imaging Tools. . . . . . . . . . . . . . . . 20
1.7 Well Completion . . . . . . . . . . . . . . . . . . . . . . . . . . . 20
2 Seismic Waves . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 23
2.1 The Fundamental Conditions . . . . . . . . . . . . . . . . . . 23
2.2 Theory of Elasticity . . . . . . . . . . . . . . . . . . . . . . . . . 23
2.2.1 Stress . . . . . . . . . . . . . . . . . . . . . . . . . . . . 24
2.2.2 Strain . . . . . . . . . . . . . . . . . . . . . . . . . . . . 26
2.2.3 Common Types of Strain . . . . . . . . . . . . . . . 26
2.2.4 The Volume-Changing Strain . . . . . . . . . . . . 27
2.2.5 The Shape-Changing Strain . . . . . . . . . . . . . 28
2.2.6 The Cubical Dilatation . . . . . . . . . . . . . . . . . 28
vii
viii Contents
References. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 321
Index . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 325
Introduction
1
waves, named after its discoverer, the English- In recent times, we may add Agadir-Morocco
man A. E. H. Love. in 1960, Kobe-Japan in 1995, Fukushima-Japan
in 2011, and Katimando-Nipal in 2015.
seismic
reflection depth
raypath
reflector
seismic
refraction depth
raypath
reflector
Fig. 1.2 Essence of the two seismic techniques; reflection (a) and refraction (b)
change (called the gravity anomaly) is greater The measuring instruments, normally used in
over the anticline axis falling to lower values gravity surveying, are called the gravity meters
over the flanks. A structural feature showing (or gravimeters). These are designed to measure
relative density-deciency (such as a salt dome), gravity variations rather than absolute gravity
gives an inverted bell-shaped anomaly normally values. These instruments are capable to measure
referred to as (negative anomaly) which is of gravity changes to less than a tenth of a milligal.
opposite shape-changes to that exhibited by the The gravity unit (the gal) is dened to be equal to
relatively denser-rock geological anomaly (posi- 1 cm/s2. Thus the gravimeter is capable of mea-
tive anomaly). The principle is shown in Fig. 1.3. suring gravity variations to within about
0 0
Sea level
geological anomaly
1 2 1 2
Fig. 1.3 Dependence of the algebraic sign of surplus where body density (q2) is greater than that of the
gravity-anomaly on whether the contrast is surplus or host medium (density, q1), and vice versa, given (q2 > q1)
deffeciency. Positive sign is given to case of density
6 1 Introduction
L L L
Fig. 1.5 Principle of the Wenner conguration used in electrical resistivity surveying. I and V denote electric current
and voltage respectively. L is the distance between electrodes
accumulations.
Natural-Source Electrical Methods
time, t There is another less commonly applied electrical
t0
group of methods which are dependent on mea-
Fig. 1.6 The procedure used in the IP method. Voltage surements of the earth natural electric eld. Self
drop after the applied electric current is switched off at Potential (SP), Telluric Current and Magneto-
time (t0)
Telluric methods are typical examples of such
techniques. These methods are used to determine
back to normal state after the applied electric large-scale crustal structural variations as sedi-
current is suddenly stopped at a certain instant. The mentary basins and regional-scale geological
process is sketched in Fig. 1.6. variations.
The method is used mainly for purposes of (i) Self Potential (SP) Surveying
metallic minerals and ground water exploration.
Mineralized zones found in the upper part of the
For the eld surveying, the same electrode set up
Earth crust and some ore bodies found at shallow
for resistivity is also used. Thus using four depths, develop their own natural electric elds.
electrodes (called dipole-dipole array) is used in
In particular, a metallic sulde body which has
IP surveying. The polar dipole system is pre-
part of it above the water table and the rest of it, is
ferred where the voltage electrode spacing is below it. With this condition, such a phenomenon
decreased to minimize the effect of wire-currents.
is created. The body acts as a natural battery
(iii) Equipotential and Electromagnetic where ions, due to differential oxidation, move
(EM) Surveying from one part of the body to the other. The
Equipotential surveing uses the electric potential electrical currents are formed as a result of
eld generated by a xed electric-current electrode chemical reaction that takes place with the aid of
while the electric potential is mapped by a moving the electrolytes present in the host medium.
electrode. Instead of using direct current The potential difference measured on the sur-
(DC) electrodes, another method, the Electro- face will map the potential anomaly caused by the
magnetic (EM) method, uses an electric alternating body. The behavior of the electric potential due to
current (AC) of few hundred-to-few kilohertz, presence of a sulde body is shown in Fig. 1.7.
induced into the ground by a source-coil and
received by another coil, the receiver-coil. The electrical
transmitter coil induces an alternating (primary) potential
magnetic eld, while the receiver coil senses the +
ground-generated alternating (secondary) mag- 0
netic eld combined with the primary. The com-
bined magnetic elds induce AC in the secondary -
coil. This current is then measured and converted ground level
into the combined magnetic eld at the location of
oxidization
the receiver. From comparison of the combined region
magnetic eld with the primary magnetic eld + water table
S
(which is known), the contribution of the anoma- -
lous body is determined and then interpreted into
current
geological information. flow lines
In addition to ground surveying, there is the
airborne EM surveying where the transmitter coil Fig. 1.7 Sulde body (S) acting as a natural battery,
is xed to the aircraft and receiver is mounted in causing abnormality in the generated electrical potential
1.3 Geophysical Exploration Methods 9
The SP method is typically used to map the summarized in the following table (Maton et al.
electrochemical potential variations, generated 1995, p. 269).
by buried sulde bodies.
Radiation Radiation Energy
(ii) Telluric Currents Surveying type characteristics characteristics
The term (telluric currents) is used for the natural
Alpha Positively-charged Can be stopped
electric currents that flow horizontally in the particles protons by a sheet of
upper part of the Earth crust. Variation of the paper
current-density, over the earth surface is gov- Beta Negatively-charged Can pass
erned by the rock-resistivity changes. Thus, if a particles electrons through as
salt dome, for example, is found imbedded much as 3 mm
of Aluminum
within relatively high conductivity-formations,
the lines of currents flow will by-pass the salt Gamma High-frequency Can pass
rays electromagnetic through several
body causing distortions in the potential gradient waves centimeters of
in the overlying surface cover. From the detected Lead
anomalous potential gradient, the geological
changes causing the changes in the potential All of the three types of radiation have the
gradient, can be determined. ionizing ability when colliding with atoms. This
(iii) Magneto-Telluric Currents Surveying effect is the basis of radioactivity detection.
Another related method, the so-called A detection instrument (the Geiger counter)
(magneto-telluric method), involves simultane- was designed by Hans Geiger in 1928. It consists
ous measurements of both voltage created by the of a tube lled with an inert gas (helium or
telluric currents and the magnetic eld induced argon) at a reduced pressure. When radiation
by these currents. Plots of the alternating voltage enters the tube, it removes electrons from the
and that of the associated magnetic eld (as atoms of the gas, making them positively
function of frequency) can give information on charged ions. The electrons move to wire in the
the resistivity as function of depth. tube, setting up an electric current which is
amplied and fed into the recording device. The
(iv) Audio-Frequency Magnetic Surveying current produces clicking sound that pulsates at a
There is still another closely related method rate which is depending on the radiation strength.
which uses the audio-frequency variations in the A counting unit (counter), attached to the wire,
Earth electric eld, to study the earth resistivity measures the created current and gives a pro-
variations. This method is called Audio- portional radioactivity-reading.
Frequency Magnetic method, termed (AFMAG)
method.
Fig. 1.9 A simplied sketch showing the main parts of a 1.5 Well Geophysical Logging
rotary drilling rig
S
r1
r2
normal-sonde lateral-sonde
configuration configuratio
sonde
SP-Log
only in an uncased holes lled with conductive be drawn for sandstone SP-value. The inflection
mud. The measuring sonde is of simple congu- points in the SP-log indicate formation
ration. It consists of only two electrodes: one is boundaries.
lowered into the well by an insulated cable and
the other is xed at the ground surface
(Fig. 1.15). 1.5.2 Radioactivity Logging
This method, which needs no articial current,
can detect natural potential differences which Radioactivity is the phenomenon of emission of
develop at formation boundaries. Compared with particles and photons of electromagnetic energy
resistivity logs, the SP logs give more sharp from an atom. This radiation process occurs
changes and hence more accurate formation- either naturally from unstable nuclii or induced
boundaries. Like lateral logging mentioned by bombarding of stable nuclii by photons or
above, there is a variation made on induction atom particles. Examples of such radioactive
logging to give focused current radiation for elements are Uranium, Thorium Rubidium and
getting sharper boundary indication. Potassium 40 which is most commonly found in
Interpretation of the SP logs depends on the shale and clay and less in sandstone and
manner of the electrical potential variation. As a limestone.
general rule, changing of the log values towards There are three types of radiation: Alfa parti-
positive potentials is considered as an indicative cles formed of charged helium nuclii, Beta par-
of impermeable rocks as shale or tight limestone ticles formed of high speed electrons, and
or tight sandstones. When the change is towards Gamma rays of electromagnetic wave-energy.
negative potential, it is interpreted as being due Out of these, only gamma ray is used in well
to porous sandstones or porous limestone. In this radioactivity logging. The detection instrument is
way, the SP logs give useful indications on a Geiger or more usually is the scintillation
lithology, water salinity and help in determina- counter which consists of a special crystal (like
tion of formation boundaries. In particular, it sodium iodide), which emits flashes of light as
shows the boundaries of formations in sand-shale they absorb gamma-ray photons, hence the name
sequences. These are determined with the help of (scintillation counter). A photoelectric tube con-
drawing lines (sand and shale base-lines) in the verts these flashes into electric currents which are
produced SP log (shown in Fig. 1.15). It is displayed in the form of a continuous chart (the
commonly observed that shale beds give the radioactivity log).
same level of SP-readings allowing for drawing a Radioactivity logs provide important infor-
straight line indicating shale SP-value. This is mation on rock lithological types, especially on
called (the shale line). Similarly a (sand line) can those containing certain concentrations of
1.5 Well Geophysical Logging 15
shale
S S
limestone
R R
R R sandstone
S
shale
about 3 ft from the nearest receiver. To correct Sonic logs provide the interval transit time
for tilting and hole-irregularities effects, a dual which represent travel-time (usually in
source sonde is used, making what is called a microsecond units) of the P-wave in covering a
borehole-compensated sonde (Fig. 1.18). distance (usually 1 ft). This parameter (the
The electronic structure of the sonde is interval transit time) is useful in computing the
designed in such a way, that the output is made porosity () by using the following relationship:
to be the difference in the travel-times to the two
receivers. The time difference, measured in Dtl Dtm =Dtf Dtm
time-units per 1-ft, called (interval transit time),
is plotted (normally in micro seconds) against where, (Dtm), (Dtf), and (Dtl), represent transit
depth to give the continuous wiggly curve known times of matrix material, of pore fluid, and log
as the (sonic log). In the compensated sonic read-transit time respectively.
logging, seismic pulses are emitted alternately
(ii) Well Velocity Surveying
from the two sources and the transit times from
A hydrophone-type detector is lowered down the
the two oppositely traveling refracted P-waves
well which is lled with the drilling fluid. The
are averaged electronically. The output (transit
travel time of the seismic wave generated by a
time) is plotted against depth, giving the
surface-placed shot and received by a detector
borehole-compensated (BHC) sonic log
placed at a formation boundary is recorded. The
(Fig. 1.19).
full log is obtained by repeating the recording at
The borehole compensated sonde (BHC)
each boundary traversed by the well. From the
gives an average interval transit time which is
measured travel time of direct arrivals, the av-
plotted on a paper strip. The produced log in this
erage velocity and interval velocity are plotted as
case is normally referred to as BHC sonic log
function of well-depth. The velocity survey is
which is used to identify lithologies, determine
also called (check-shot survey). Principle of the
formation boundaries, and in computing syn-
velocity survey setup and velocity-depth plot is
thetic seismograms. The interval transit time can
shown in Fig. 1.20.
be integrated down the well to give the total
The obtained velocity information is used to
travel time. This type of logging can only be run
calibrate the sonic log and check the sonic-log
in an open (uncased) hole.
18 1 Introduction
seismic
detector
depth
integrated time (hence the name, check-shot normally applied to refer to the direct and
surveying). reflected waves respectively. The ray paths of the
VSP shooting and the corresponding travel-time
(iii) Vertical Seismic Proling (VSP) plot of the recorded waves are schematically
Basically the measurement set-up is the same as shown in Fig. 1.21.
the check-shot recording system. The difference A seismic pulse is generated on the surface
is in the recording duration time which is here from dynamite explosion or from an air gun
extended to allow recording reflected waves as submerged in a water lled hole. This is repeated
well as the direct arrivals. At each detector at all of the hydrophone positions, normally
stop-location (normally at 25 m spacing), the positioned at 25 m-spacing down the well.
recorded seismic trace is allowed to include With an appropriate processing sequence, the
events from up going reflected waves in addition nal corrected VSP section is produced. Process-
to the rst arrival (direct wave) event. The terms ing includes, data editing, correction to vertical
(down-going wave) and (up-going wave) are time and velocity ltering for separating unwanted
R1
down- tube up-going wave
R2 going wave
wave
R3
Ray-paths of Travel-time plot
direct and reflected waves of the tube wave,
from a three-reflector model down-going & up-going waves
1.5 Well Geophysical Logging 19
events. It is evident from the VSP section that Drilling information (drilling rate, mud den-
downward events (primary and multiples) increase sity variation)
in time with depth, and the upward events (primary Caliper and dip-meter logs.
and multiples) decrease in time with increasing
The main information which can be obtained
depth. By arranging the produced seismic traces
from well logs can be summarized as in the
(25 m spaced down the hole) a seismic VSP sec-
following table
tion is obtained. Each seismic trace of a VSP
section contains events from down-going waves
Log Type Application Comment
(direct and multiples arrival) and from up-going
Electrical Fluid-type Logging is
waves (reflections and multiples). Logs identication done in
The main application of VSP is providing Resistivity Porosity uncased
seismic section of reflectors which have not been Induction evaluation wells
reached by the drilling. In fact the VSP data are Spontaneous Boundary
Potential determination
equivalent to both the check-shot data and the Shale and sand
sonic-derived synthetic seismogram. lines
(iv) Other Logging Techniques Radioactivity Shale/sandstone Logging is
Logs Formation done in
Some logs are used to give geometrical infor-
Natural density cased wells
mation on the deviation from the vertical gamma-ray Porosity
(hole-drift angle) as well as the azimuth of the Gamma-ray Fluid type
deviation. Other logs are made to give the bed- density
Neutron
ding dip (dip-meter logs) and measurements of
gamma-ray
the well-hole diameter (caliper logs).
Acoustic Lithology types Logging is
On more limited scale of application, other Logs Formation done in
types of well logging have been used. Specially Sonic logs Boundaries cased and in
modied borehole-gravimeters, magnetometers, Well Synthetic uncased
and thermometers are examples of such tools. velocity seismograms wells
VSP Seismic-velocity
Geothermal prospecting is applied to detect functions
geological features which affect heat flow such as Reflection
shallow salt domes, faults and dykes. Also identication
ground water is investigated by this method.
1.6.1 Logging While Drilling running-in and cementing the casing is consid-
Technique ered as integral part of the completion process.
To prevent caving in of the well walls, drilling
In late 1980s a new logging procedure, logging is stopped at various depths. and a steel pipe
while drilling (LWD), was introduced. This (casing) is lowered into the well and xed in
technique is similar to the conventional wireline position by pumping cement mixture in the space
logging except that it (the LWD) uses sensors between the casing pipes and the well walls
which are incorporated in the drilling bit (cementing process). According to the planned
assembly, and thus the measurements are made drilling program, drilling after casing being
during the drilling process. This measurement cemented is continued with a different (usually
procedure coupled with specialized computer smaller) bit-size. At this stage, the drilled hole is
software provides fast geological information as lined by the appropriate casing which is likewise
lithology, porosity, fluid contents, and drill-hole cemented in. The drill-casing-and-cementing
direction. phases are repeated until the nal planned total
depth is reached. The end result of the casing
operations is a set of concentric pipes each of
1.6.2 Borehole Imaging Tools which is ending at the earth surface. When the
planned total depth is reached, all used drill pipes
Borehole imaging is a logging technique that can including the attached drilling bit, are withdrawn.
provide small-scale images of the borehole wall The next operation is perforation of the parts of
and the penetrated rock lithology. One of these the casing facing the oil bearing formations. The
methods (based on electrical micro-imaging so-produced holes are made to allow the oil to
processes) was developed by Schlumberger
under the trade-name; formation micro-imaging
(FMI). In 1986, Schlumberger developed a
dip-meter called the formation micro-scanner christmas
(FMS). Later versions of this logging tool were tree
called formation micro-imager (FMI) which is
based on micro-resistivity measurements.
The operation of the FMI tool is based on
recording changes (and not absolute values) in
electrical resistivity of rock formations. It pro-
vides, with high resolution power, identication
20-inch casing
of sedimentary characteristics and fracturing
picture in the penetrated rock column, in addition
to measuring formation dip. 13 -inch casing
seep inside the central production tubing of the equipment in addition to the mechanical system
well. that provide the blowout prevention (BOP) is
The last stage in preparing the well for the called the well head which is also equipped with
production process is xing of a group of flanges an assembly of valves that controls flow, and any
which provide a sealing structure connecting other possible interventions. This assembly is
together all the surface ends of the casing tubes. normally referred to as the Christmas tree
This part of the well which contains pressure (Fig. 1.22).
Seismic Waves
2
2.1 The Fundamental Conditions motion of elastic waves are generated and can be
recorded by the appropriate detection instru-
An application of an external force, on part of a ments. These are normally referred to as the
medium (elastic medium), leads to creation of (seismic waves).
internal opposing forces which intend to resist
the deformations caused by that external force.
Typical forms of the resulting deformations are 2.2 Theory of Elasticity
changes in volume and/or in shape which are
created at the affected location. In consequence, As it is stated above, the fundamental conditions
the medium will return to its original condition for a seismic eld to be created is that the med-
after the external force is removed. This property ium must possess the elasticity property. Two
of resisting of changes in volume and in shape main concepts are governing the propagation of
and return to original conditions after removal of seismic waves in an elastic medium: the (stress)
the external force is called (elasticity). Provided and the (strain). Stress represents the external
that the changes are small, rock media in nature force applied to the elastic medium, and strain is
are considered to be perfectly elastic in nature. the resulting changes in volume and in shape.
As a result of the elasticity property of media, The relation between stress and strain, for a
the changes (volume and shape changes) oscil- particular medium (perfectly elastic medium),
late about their neutral positions and, at the same gives evaluation expressions for the elasticity
time, propagate away from the energy property of that medium. The stress-strain pro-
source-location. Energy transfer in this manner portionality constants are the elastic coefcients
(motion that leaves out no permanent distortions) which serve as measures of the elasticity of a
is commonly referred to as (wave motion). particular medium.
The fundamental condition for the creation The principal types of changes experienced by
and propagation of seismic waves (seismic eld) a medium due to passage of a seismic wave are
is a source of mechanical energy of impulsive re-distribution of the internal forces (stress
type which is initiated within an elastic medium. changes) and modication of the volume and
The energy source may be natural (as in geometrical shape (strain changes). The theory of
earthquake-generated waves) or articial (as in elasticity deals with analysis of these principal
ring of a dynamite charge). In both cases wave effects and the related physical changes.
T
Tz
Ty
Tx
A
Fig. 2.1 Stress (T) and its components (Tx ,Ty, Tz) acting on the elementary area (DA)
2.2 Theory of Elasticity 25
C
y
z
The three stress components (Tx , Ty , Tz) per each of the three
areas (A, B, C)
Tyz B
A
T
Txz Tzz
T T
Tyx
Tyy
C
Txx Tzx
Txy Tzy
Fig. 2.2 The stress components. The mutually perpendicular planes (A, B, C) making up the area-set and the nine
components of the involved stress tensor
meter (N/m2) which is called Pascal, where one In general, when a body is subjected to elastic
Pascal is equal to 1 N/m2. stress, both of its size and shape will change. As
it is mentioned above, the resulting changes
represent elastic strains when each point of the
2.2.2 Strain stressed body experiences a displacement of its
own which is different from the displacements
In reference to Fig. 2.3, let us consider the two experienced by the other points of the body. This
points (P1 and P2) located within an unstressed implies that there are two types of strains, namely
body, where the rst point, P1 is located at (x, y, z) the volume strain and the shape strain
and the second point (P2) at (x + dx, y + dy, (Fig. 2.4).
z + dz). Now, we let this body to deform as a
result of a stress system created within it. If the two
points (P1 and P2) were displaced from their 2.2.3 Common Types of Strain
original positions by equal displacements (D,
say), then it is considered that there is no strain Mathematical analyses of strain show that the
taking place. Strain occurs only when there is total strain of a three dimensional body, depends
variation of displacement of any point, within that on only six different derivatives of displace-
medium, with respect to the others. In the lan- ments. These strain components (eab), can be
guage of mathematics, we say that strain depends written down as follows (Richter 1958, p. 236):
on the derivatives of the displacement-
components with respect to the chosen coordi- exx = Dx / x
nates (x, y, z). The concept is claried in Fig. 2.3. eyy = Dy / y
ezz = Dz / z
exy = ( Dx / y + Dy / x) / 2
exz = ( Dx / z + Dz / x) / 2
P1 P2 eyz = ( Dy / z + Dz / y) / 2
(a)
Fig. 2.4 The two types of strain; volume and shape strains
2.2 Theory of Elasticity 27
Rigid body-translation and rotation represent (or compressional) stress. The longitudinal strain
cases of no strain, since no volume and no shape (e) is dened to be the change in length in a
deformation are involved. In Fig. 2.5, an ele- certain dimension, of a body under stress relative
mentary cube (shown here in plan) is used to to its original length. For a rectangular lamina of
show simple types of elastic deformation (strain) dimensions (Dx by Dy), the longitudinal strains
and no-strain changes. (ex and ey) in the x and y directions are dened as
In general, an elastic body under stress can (Fig. 2.6):
experience two types of distortions; changes in The longitudinal strains can be extensional
volume and changes of shape. These changes, (tensile strain) or compressional (contraction
which occur as result of stress, are expressions of strain). Longitudinal strains (ex and ey) are
the physical properties of the stressed body. In its dened as:
simple form, elastic strain can be divided into
two main types. These are: the volume-changing ex Dx =Dx
strain (leading to body compression or dilatation) ey Dy =Dy
and the shape-changing strain (leading to body
shape distortion). where (Dx and Dy) are the changes in length in
x and y directions respectively.
The minus sign that appeared in the ey
2.2.4 The Volume-Changing Strain expression is entered to denote that the change
(Dy) is compression which is in opposite direc-
The familiar example on this type of strain is the tion to the dilatation change (Dx), in the x-
longitudinal strain of a body under an extensional direction.
Dy
y y
x x Dx
ex = Dx / x ey = - Dy / y
Fig. 2.6 Denition of the longitudinal strain as applied for a rectangular lamina of dimensions (Dx by Dy)
28 2 Seismic Waves
x x
2.2.5 The Shape-Changing Strain displacements. When the displacements are equal
a body may experience pure translation (rigid
As longitudinal strain gives expression for the body-translation) or pure rotation (rigid
volume changes resulting from stress application, body-rotation), as shown in Fig. 2.8.
the shear strain gives the corresponding measure Rigid-body changes which do not involve
for the shape deformation. Using the example volume or shape changes, such as these, are not
above (Dx by Dy rectangular lamina). The shear considered to be elastic strains.
strain (also called angular strain) is considered to
be the average of the two angles by which two
neighboring sides rotate as a result of the shearing 2.2.6 The Cubical Dilatation
stress. Thus, the shear strain (exy) is dened as:
exy / b=2 A parameter, closely related to longitudinal strain
and of special importance in the theory of elas-
where (/ and b) represent the angles of rotation ticity is the Cubical Dilatation (h). At a certain
of the two sides (Dx and Dy) brought about by point within a strained medium, this is dened as
the shear stress (Fig. 2.7). the fractional change in a unit volume surrounding
Since these two angles are very small (usually that point. Thus, for a three dimensional body with
so, in seismic-eld conditions), they can be rep- longitudinal strains (exx, eyy, ezz), the cubical
resented by their corresponding tangents, giving dilatation can be computed as follows:
where the angles (a & b) are in radians. For small strains exx, eyy, ezz (which is the
It should be emphasized that strain occurs case in seismic-eld conditions), the products of
only if the body particles experience unequal these terms may be neglected giving the result:
2.2 Theory of Elasticity 29
h exx eyy ezz Dx =Dx Dy =Dy Dz =Dz For an isotropic body (physical properties are
independent of direction) and for an elastic body,
under small strain, strain varies linearly with the
The sign convention of (h) is negative for applied stress. This linear stress-strain relation-
compression and positive for expansion strains. ship is governed by a well-known mathematical
equation. It is the Hookes law.
strain
30 2 Seismic Waves
The linearity property governed by Hookes law 2.2.9 The Elastic Moduli
means that there is a proportionality-constant for the
linear stress-strain relation for any particular body The elastic modulus of a body is the propor-
under stress. Mathematical studies showed that, for tionality constant of the stress-strain linear rela-
an isotropic body, two elastic coefcients are suf- tionship. It expresses an important physical
cient (Richter 1958, p. 238). These are the Lames property which is the extent of resistance of that
coefcients (k & l), which are sufcient in char- body to the applied stresses. Moduli of important
acterizing the elastic properties of a medium. practical applications are Youngs Modulus, bulk
By use of Lames coefcients (k & l), modulus, and shear modulus. These are dened
Hookes law can be presented in the following in the following discussions.
compact form:
(i) Youngs Modulus and Poissons Ratio
Let a simple tensile stress (Tx) be applied to
Tij = ij+ 2 eij an isotropic bar placed along the x-axis. This will
cause the bar to experience a longitudinal
extension (ex) in the x-direction and, at the same
where the symbols (i & j) take the values x, y,
time, it experiences lateral contractions along y-
and z, and the term dij = 1 when (i = j), and
and z-directions. Being an isotropic body, the
dij = 0 when (i 6 j). T and e are the stress and
contractions in the y- and z-directions (ey, & ez)
strain respectively.
are equal. These changes (expressed by the
This compact form of the Hookes law can be
strains ex, ey, & ez) are governed by the elastic
presented in the following explicit equations:
coefcients of the stressed body. The coefcients
which govern the stress-strain relation, in the
Txx = + 2 exx presence of the tensile stress (Tx), are Youngs
Tyy = + 2 eyy modulus (Y) and Poissons ratio (r).
Tzz = + 2 ezz For a one-dimensional stress acting on a body
obeying Hookes law, Youngs modulus (Y) is
For pure shear strain (that is with no change in the proportionality constant in the linear relation
volume, for h = 0), the Law expresses the rela- that connects stress (Tx) with strain (ex). The
tions for purely shearing strain, that is: relationship is:
V
A F
P
P x
V-V
h
P
Fig. 2.11 An elastic cube under hydrostatic compression Fig. 2.12 An elastic rectangular block under shearing
forces force (F) acting on area (DA)
32 2 Seismic Waves
k rY=1 2r 1 r
2.3 Wave Motion Equation
l Y=21 r
If two neighboring points in a stressed medium
experience the same stress, no motion of one of
Other relations are presented in Table 2.3.
them will occur with respect to the other. How-
ever, relative motion will take place when there
is a stress difference. In other words, motion
2.2.10 The Elastic Moduli
occurs when there is a stress gradient. This
Interrelationships
reminds us of an analogous case we met in the
creation of strain (see Sect. 2.2). The two cases
For a homogeneous and isotropic medium under
may be expressed as follows: Displacement
stress, the stress-strain relationship is linear
gradient is required to create strain and stress
within the elastic (Hookean) state. The propor-
gradient is required to cause motion.
tionality constants are the elastic moduli or
2.3 Wave Motion Equation 33
2.3.1 One-Dimensional Scalar Wave This strain (exx = Dx/x) is produced by the
Equation corresponding stress gradient (Txx/x).
By making use of the fact that the net force
In this section, we shall deal with the wave acting on any face is given by the stress acting on
motion equation which expresses the motion of a that face times the face area, we get the resultant
disturbance in one dimension. The disturbance in force (Fx) in the x-direction due to the stress
this particular case is the scalar quantity, the change (Txx/x) dx that occurred across the
cubical dilatation (h). distance (dx). This is computed as follows:
Let us consider an elementary parallelepiped
(of dimensions: dx, dy, Dz) located inside an Fx @Txx =@x dx dy dz
elastic isotropic medium (Fig. 2.13).
At each face of this elementary body, when it By applying Newtons second law of motion
is under elastic stress, there exist three stress we can express (Fx) in terms of mass of the
components: one is normal and two are tangen- parallelepiped (dx dy dz times density q)
tial to the particular face. The three multiplied by acceleration (2Dx/t2) in the x-
stress-components (Txx, Tyx, and Tzx) are acting direction giving:
on the face perpendicular to the x-axis. Under
dx dy dz q @ 2 Dx =@t2 @Txx =@x dx dy dz
elastic stress-strain conditions, each of these
components will have a gradient in the
x-direction (Txx/x, Txy/x, and Txz/x). For or:
a complete three dimensional state, additional
similar gradients occur in the other two directions q @ 2 Dx =@t2 @Txx =@x
(y-direction and z-direction).
In order to simplify the mathematical deriva- This is the one-dimensional (dimension, x in
tion of the equation of seismic wave motion, let a this example) wave motion equation which
plane compressional seismic wave to be describes particle motion (displacement, Dx) in
advancing in the x-direction. In this case the terms of the applied stress (Txx). However, the
three stress-components are reduced to only one motion can be expressed in terms of displace-
component (Txx) creating the corresponding ment only. This is done by using the stress-strain
strain (exx). When a seismic plane wave propa- linear relationship expressed by Hookes law
gates in the x-direction, the two faces (perpen- equation for isotropic media (Txx = k h + 2l
dicular to the x-axis) of the parallelepiped will be exx). Substituting for (Txx), the previously-
unequally displaced, and hence, it is subjected to derived wave equation becomes:
an elastic strain (exx) which is, by denition,
given by the displacement gradient (Dx/x). q @ 2 Dx =@t2 k@h=@x 2l@exx =@x
x
plane wave fronts
36 2 Seismic Waves
2.3.4 The P- and S-Waves 330 m/s in air, 1450 m/s in water, and (2000
6000) m/s in rocks.
As it is presented above, there are two types of A solid medium having its Poissons ratio
disturbance that can move in accordance with the equal to (1/4) is called Poissons solid (Sheriff
standard wave motion equation. These are the 2002, p. 266).
scalar cubic dilatation (h) and the vector shear
strain (w).
From the wave equation it can be shown that 2.4 Classification of Common
the disturbance (h) moves faster than the other Elastic Waves
disturbance (w). Thus, when the two distur-
bances are generated by a certain source, the From analyses of stress and strain, we have seen
(h-wave) arrives earlier than the (w-wave). For that strain is, in general, made up of two types of
this reason the two waves are called Primary elastic disturbance; the cubic dilatation and the
(P-wave) and Secondary (S-wave) respectively. shear strain. Solution of the equation of motion
It is to be noted that the ratio of the P-wave showed that each of these types of deformation
velocity (vp = [(k + 2l)/q]1/2) to the S-wave travels through the medium with its own veloc-
velocity (vs = [l/q]1/2) is equal to [(k + 2l)/ ity. The rst type of disturbance represents the
l]1/2. Using the relationship connecting (k) and moving volume strain and the second type
(l) in which k/l = 2r/(1 2 r), we can write: involves the shape strain. The rst type is
called Longitudinal, Compressional, or Primary
vp =vs 2 2r=1 2r1=2 wave (or just P-wave) which travels faster than
the second type which is called Transverse,
This formula clearly shows that the ratio of Shear, or Secondary wave (or just S-wave).
the P-wave velocity (vp) to the S-wave velocity These two types of waves (P- and S-waves)
(vs) is function of Poissons ratio (r) only. belong to a class of waves (called body waves)
According to (Dobrin 1960, p. 18), Poissons because they can propagate through the interior
ratio (r) generally ranges from 0.05 to 0.40, of the earth body. This group of waves is called
averaging about 0.25 for hard rocks. With this so to differentiate them from another class of
value (r = 1/4), the velocity ratio (vp/vs) waves which move on and near the free surface
becomes 31/2 (=1.732). This means that P-wave of the medium, called (surface waves) which
moves with velocity which is about 1.7 times as include Rayleigh- and Love-waves. Classica-
fast as the S-wave moving in the same medium. tion of the common elastic (seismic) waves is
It is useful to note that P-wave velocity is shown in Fig. 2.15.
2.4.1 Body Waves P-wave is the fastest wave for a given medium
and, therefore, its arrival at a certain observation
Body waves are waves that can travel through an point is the earliest among the seismic
elastic materialistic medium in any direction. As wave-types. This is a common observation of
they move, the waves may experience changes in seismologists working on analysis of earthquake
their energy level and in their travel-path geom- seismograms. Propagation velocity (vp) of
etry subject to the physical properties of the P-wave depends on the medium density (q) and
medium. There are two sub-types of these waves; elastic properties (k & l) and it is given by the
the longitudinal and the transverse waves expression vp = [(k + 2l)/q]1/2.
(Fig. 2.16).
(ii) Transverse Waves
(i) Longitudinal Waves The travelling disturbance in this case is the shear
This type of waves is also known as compressional, strain or shape deformation. The medium
Primary, or just P-wave. The travelling disturbance which is traversed by this type of waves experi-
in this case is volume deformation expressed by ences no volume changes. A consequence of the
the cubical dilatation (h) as dened above. shear strain (rotation of part of the medium) is the
The particles of the medium, traversed by a transverse displacement of the path particles
plane P-wave, vibrate about their neutral posi- relative to the propagation direction. They are
tions in the direction of the wave propagation. also called (shear waves) or (Secondary, or just
The travel path consists of a sequence of alter- S-waves).
nating zones of compressions and rarefactions A horizontally moving S-wave, which is so
(Fig. 2.16a). This is the type of waves which is polarized that the particle motion is conned to
commonly employed in seismic reflection and vertical plane, is known as SV-wave (Fig. 2.16
refraction exploration work. b). When the polarization plane is horizontal, it is
(b)
SH-
propagation
direction
(c)
propagation
direction
38 2 Seismic Waves
called SH-wave (Fig. 2.16c). The velocity of The main sub-types of surface waves are
S-waves, vs is given by vs = [l/q]1/2. In Rayleigh waves and Love waves (Fig. 2.17).
liquid-media, where (l = 0), S-waves do not
(i) Rayleigh Waves
propagate.
Rayleigh waves, which were discovered by an
English scientist, Lord Rayleigh in 1885, are
usually developing at the free surface of a
2.4.2 Surface Waves
semi-innite solid medium. Its wave amplitude
decays rapidly with increasing depth. The trav-
As it is implied by its name, surface waves are
elling disturbance in this case is a sort of com-
waves that move on the free surface of the earth.
bination of particle-motions of both P- and
The main features common among all surface
SV-waves. The particle motion, which has a
waves, observed on earthquake seismograms, are
retrograde elliptical orbit, takes place in a vertical
their relatively large amplitudes (high energy
plane parallel to propagation direction
content) and low frequencies when compared
(Fig. 2.17a). The minor axis of the elliptical orbit
with the body waves. In addition to that, they
is parallel to wave motion direction and it is
move with velocity which is generally slower
equal to two-thirds of its major axis. Rayleigh
than body waves moving in the same medium. It
waves travel on the surface of a solid medium
is a common observation that the dispersion
with velocity of 0.92 of the velocity of S-waves
phenomena are more prominent in surface waves
moving in that medium (Bullen 1965, p. 90). In a
due to dependence of the velocity on the fre-
sense, Rayleigh waves are similar to the familiar
quency of individual harmonic component.
water waves, with a fundamental difference, and
(a)
propagation
direction
(b)
surface layer
propagation
direction
Fig. 2.17 Particle displacement-mode of a medium traversed by plane surface-waves, (k) is wavelength. a Rayleigh
Wave, b Love Wave
2.4 Classification of Common Elastic Waves 39
that is the particle motion in case of Rayleigh Since they possess no vertical component,
waves describe an elliptical path whereas the Love waves are not detected by the geophone or
particle-motion path in case of water waves are by any such-like vertical-component sensing
circular in shape. instrument.
In the case where the semi-innite medium is
overlain by a low-velocity surface layer, Rayleigh
waves exhibit a phenomenon known as (disper- 2.4.3 Seismic Noise
sion). Harmonic components of longer periods
(lower frequencies) travel faster. Consequently, Broadly speaking, the term (noise) used in seis-
the Rayleigh wave seismograms would, in gen- mology, is applied to all types of disturbance
eral, show decrease in period along the wave-train. which may interfere with (and impose masking
Components of too-long wavelengths (too long effects to) the seismic signal of interest. In this
compared with the thickness of the surface layer) way, the concept of seismic noise bears a relative
penetrate deeper and travel with velocity of about implication. Thus, when the interest is focused
0.9 times the S-wave velocity in the subsurface on reflected body waves, surface waves and other
material. The short wavelengths travel mainly in non-reflection waves (as direct and refraction
the surface layer with velocity of about 0.9 times arrivals) are considered to be the unwanted
the S-wave velocity in the surface layer. troublesome noise. If the interest is in the
Surface waves, normally seen on shot records, refraction arrivals, reflection arrivals become
obtained in seismic reflection surveys, are com- the unwanted noise. In the strict sense, however,
monly called (ground roll) and these are identied the ambient seismic disturbances (usually of
to be of Rayleigh-wave type. Sometimes, these are random energy distribution which form the
called pseudo-Rayleigh waves (Sheriff 2002). background of a distinct travelling signal) are
Ground-roll waves are considered to be unwel- considered to be the seismic noise.
comed noise and efforts are usually made to get rid Seismic noise has destructive effects on the
of them or at least minimize their masking effect seismic signals of interest. A signal recorded
caused to the seismic reflection signal. amid a background of noise is distorted and
weakened because of the interfering noise. Signal
(ii) Love Waves
resolution is badly affected with noise develop-
This is the second sub-type of surface waves which
ment. A measure for the signal resolution, called
was discovered, in 1911, by another English geo-
the signal-to-noise ratio (S/N ratio) is usually
physicist named A.E.H. Love (18631940). It
applied. It is dened to be the ratio between
develops only in cases where a solid elastic
signal amplitude detectable amid a background
semi-innite medium is overlain by a horizontal
seismic noise.
low-velocity layer. Like SH-wave vibration mode,
In exploration seismology seismic noise is
the particle movement is transverse and is conned
divided into two main types; coherent and inco-
to the horizontal plane (Fig. 2.17b). Love waves
herent noise (Fig. 2.18).
travel by multiple reflections between the top and
bottom boundary-planes of the surface layer. The (i) Coherent Noise
propagation velocity approaches S-wave velocity Coherent noise is a seismic event characterized by a
in the subsurface medium for very long wave- distinct apparent velocity and well-dened onset.
lengths and to that of the surface layer for short In reflection seismology, coherent noise which
wavelengths (Dobrin 1960, p. 23). Love waves appear on shot records, are source-generated seis-
always exhibit dispersion. As in the case of Ray- mic events. They are made up mainly of surface
leigh waves, Love waves propagation-velocity waves (ground roll) and air-waves which are of
increases with the period of the harmonic compo- fairly narrow bandwidth with low frequency range.
nent. Again, the vibration amplitude decays expo- Frequency content of this type of noise is typically
nentially with depth in the lower medium. below 20 Hz (Fig. 2.18).
40 2 Seismic Waves
amplitude
incoherent noise, random
noise (c), in relation to that
of the reflection signal (b)
0 20 40 60 80 100
frequency, Hz
(ii) Incoherent Noise at, in seismic data acquisition. Several ways and
Unlike coherent noise, the incoherent noise means are followed in the eld-acquisition stage
consists of seismic events with unpredictable or in the following data-processing stage to get
amplitude and onset. This type of noise, which is enhanced S/N ratio. Suitable measures are
basically of random nature, forms the applied to the parameters of the seismic source
seismic-energy background of any seismic and detectors as well as those measures applied
shot-record. In earthquake seismology it is in processing work, in order to attenuate these
commonly known as (microseisms), and in noises and enhance the S/N ratio.
prospecting seismology it is called (incoherent
background noise), or (ambient noise) as it is
sometimes referred to. In addition to the ran- 2.5 Propagation of Seismic Waves
domness nature, the incoherent noise is charac-
terized by a broad amplitude spectrum that Seismic waves are generated from a sudden
covers a wide range of frequencies compared change in the internal strain occurring inside an
with the nearly limited bandwidth of reflection elastic medium. The generating source may be
signals or coherent noises (Fig. 2.18). In the natural as in the case of earthquakes or articial,
geophysical literature we sometimes meet terms like exploding a charge of dynamite, as nor-
like (white noise) indicating wide-band noise, mally done in seismic exploration. All parame-
and (red noise) for low-frequency random noise. ters of an advancing seismic wave (waveform,
Intensive research work has been undertaken, speed, and travel-path geometry) may change
directed towards a greater understanding of the during the wave propagation. Form and mag-
source and characteristics of the incoherent nitude of these changes depend on the physical
noise. It is now generally accepted that it is properties of the host medium. Whether the
generated as a result of external energy sources source is natural or articial, a seismic eld is
like wind movements, sea-waves collisions with created when a sudden pressure pulse is initi-
sea coasts, in addition to other various natural ated. The generated seismic energy moves away
and articial man-made activities. from the source zone in a form of a wave
Because of seismic noise which are unavoid- motion propagation. Under these conditions
able seismic events which get recorded alongside (seismic energy source within an elastic med-
the objective signal, the signal-to-noise-ratio ium), the seismic wave spreads out from the
(S/N) becomes an important parameter in signal source zone in every possible direction. A travel
detection studies. The S/N ratio is used as mea- ray-path, in a particular medium, is dened once
sure for the signal quality-level. Signal clarity the locations of both of the source-point and
(S/N enhancement) is a central objective, aimed detector-point are dened.
2.5 Propagation of Seismic Waves 41
ray
wave-front
42 2 Seismic Waves
the wave front is a line drawn normal to the 2.5.4 The Concept of the Interface
spherical wave-front (circles on the surface
plane) at that point. Concepts of the wave-front The Interface is that boundary-surface separating
and rays are shown in (Fig. 2.20) for a case of two different media. As far as the changes (chan-
dropping a pebble into a still pond. ges in spectral structure and propagation direction)
of seismic waves are concerned, two media are
considered to be different if both of the wave
2.5.3 Huygens Principle propagation velocity and the medium bulk density
are different. Since velocity is function of elastic
Huygens Principle states that each point on a coefcients, it can be said that density and elastic
wave-front acts as a source of a new wave which, properties are the factors which control the specic
in a homogeneous medium, generates a sec- characters of the media. The parameter which
ondary spherical wave-front, the envelope of expresses the combined effect of velocity and
which denes the position of a wave generated at density is called (acoustic impedance) which is
some later time. dened to be the product of velocity by the density.
Huygens model of wave propagation requires To clarify the concept of the interface and the
that the secondary wave-fronts are active only at roll of the acoustic impedance waves hitting an
the points where the envelope touches their sur- interface let us consider a two-layer model which
faces. The wave energy is spreading out from the consists of two adjacent media (M1 & M2) of
primary source-points in all directions, but their velocities and densities (v1 & q1) for medium (M1)
mutual interactions make the resultant distur- & (v2 & q2) for medium (M2). The acoustic
bance zero everywhere except at the points impedances (z1 & z2) for the two layers are
where they touch the common envelope. (z1 = q1v1) and (z2 = q2v2) as shown in Fig. 2.22.
Applying the principle on plane-wave propaga- In analogy to the role of electrical impedance
tion in a homogeneous, and in an inhomoge- in the flow of electrical current, the acoustic
neous medium, is shown in Fig. 2.21. impedance expresses the extent of resistance the
Z1 = 1 V1 M1
Interface plane
Z2 = 2 V2 M2
Fig. 2.22 Concept of the Interface and denition of the Acoustic Impedance, z (=qv)
seismic energy meets when traversing a medium. place on ray-path direction or on the waveform of
The higher the acoustic impedance, the lower the the travelling seismic pulse. In nature, however,
particle vibration-velocity will be, and vice versa. the medium is far from this idealized form. In the
Acoustic impedance is measured by (kg s1 m2) solid crust of the Earth, it is commonly made up of
or by the equivalent (Ns m3) units. rock layers of varying physical properties and
At an interface, an incident seismic wave varying geometrical forms and sizes.
(normally a P-wave in seismic exploration work) In such inhomogeneous environments a
would, under certain geometrical conditions, give moving seismic wave would suffer from a num-
rise to wave conversion in addition to reflection, ber of changes whenever it meets an interface
refraction, and diffraction. These cases shall be across which there is change in the properties of
dealt with in some details in the following the medium. In particular, changes in energy
discussions. content, waveform (spectral structure), propaga-
tion velocity, direction of motion, and new wave
generation. These changes, are generated at the
2.5.5 Changes of Propagation interface planes dening the layer bounding
Direction at Interfaces surfaces (Fig. 2.23).
The common changes in ray-path direction,
In an idealized homogenous and elastic medium, a which are of signicance to exploration seis-
seismic wave propagates with no changes taking mology, are: reflection, refracted transmission
Layer-1
Layer-2
Layer-3
Layer-4
44 2 Seismic Waves
V1 V1 V1
V2 V2 V2
(refraction), and diffraction. These shapes of the of different density and elastic properties (dif-
moving wave ray-path occur at the boundaries of ferent acoustic impedances), four new wave
media having different seismic propagation phases are generated; reflected and refracted P-
velocities (Fig. 2.24). and SV-waves. If, however the incident is
SH-wave, the generated waves are only reflected
and refracted SH-wave. The SV-waves, gener-
2.5.6 Wave Conversion at Interfaces ated from an incident P-wave, (or P-waves gen-
erated from an incident SV-wave) are called
When a seismic wave impinges on an interface (converted waves) (Fig. 2.25).
separating two media, which differ in acoustic An incident seismic wave onto an interface
impedances, the incident seismic energy is partly will be partly reflected and partly transmitted
reflected and partly transmitted with certain across the interface. In general, the interface will
waveform changes. When the ray-path of an bring about wave conversion, reflection, trans-
incident seismic wave is oblique, that is inclined mission, and diffraction. It should be noted here
with respect to an interface, new waves are that refraction is a special case of transmission.
generated. If, for example, the incident wave is Refraction (ray-path bending) occurs only in the
P-wave (or SV-wave) separating two solid media case of inclined incidence.
P SV SV SH
P SH
SV
P
M 1 , V1 , 1 M 1 , V1 , 1 M 1 , V1 , 1
M 2 , V2 , 2 M 2 , V2 , 2 M 2 , V2 , 2
P P
SV SV SH
V2 > V 1
Fig. 2.25 Wave conversion at an interface for three types of incident waves (P, SV, and SH waves). The symbol (h)
denotes angle of incidence
2.5 Propagation of Seismic Waves 45
2.5.7 Energy Partitioning incident waves (P, SV, SH) and impedances
and Zoeppritz elements (velocity and density) for each of these
Equations wave-types. A typical set of Zoeppritz curves for
the case of an obliquely incident P-wave is
The mathematical expressions describing the shown in Fig. 2.26.
energy partitioning of an obliquely incident wave Referring to Fig. 2.26, it can be seen that, for
among the created converted waves, were normal incidence (angle of incidence = 0), no
derived rst by Knot (1899) and later on by S-wave is generated and thus all the energy is
Zoeppritz (1907), but not published until (1919). shared by the reflected and transmitted (re-
Using an approach (based on displacement fracted) P-wave. At a small angle of incidence,
computations), Zoeppritz has derived the equa- the converted S-waves are of small energy level.
tions (commonly known as Zoeppritz equations) As this angle increases the generated S-waves
which express the relative energy partitioning as grow stronger at the expense of reflected and
function of angle of incidence and acoustic refracted P-waves. At the critical angle of the
impedances of the media separated by the incident P-wave, the transmitted P-wave energy
involved interface. falls to zero, and at the same time, both of
In the geophysical literature, these equations reflected P-wave and reflected S-wave grow
are presented in the form of curves for certain large. Buildup of energy of reflected P-wave, as
two-layer models with dened density and elas- the critical angle is approached, is referred to as
ticity properties (see for example, Grant and (wide-angle reflection). This phenomenon (in-
West 1965, p. 54, Telford et al. 1990, p. 157, crease of reflection coefcient near the critical
Sheriff 2002, p. 401). A complete coverage of angle) is sometimes made use of in seismic
various types of incident waves, with different reflection exploration (Sheriff 1973, p. 241).
types of media, is found in (Ewing et al. 1957). Further, as the angle of incidence approaches
Because of the numerous possible grazing incidence (angle of incidence = 90),
parameter-values required to dene the behavior energy of the reflected P-wave increases and, at
of energy-distribution as function of incidence grazing incidence (where there is no vertical
angle, many curves are required for the various component for the incident P-wave), the S-waves
cases. These cases represent selected types of disappear and no transmission process occurs
1.0
relative
energy transmitted P-wave
reflected P-wave
reflected SV-wave
transmitted SV-wave
Fig. 2.26 Typical Zoeppritz curves of energy partition- P-wave at an interface separating two media of specied
ing as function of angle of incidence. The curves are for properties (sketched, based on Dobrin and Savit 1988,
the converted waves created by an oblique incident p. 43)
46 2 Seismic Waves
and consequently all the incident energy is con- case of reflections from a given horizontal
ned to the reflected P-wave. reflector. There are, however, situations where
the angle of incidence does not vary with the
offset. Thus, in a multi-reflector case, the angle
2.5.8 Amplitude Variation of incidence (which is equal to angle of
with Angle of Incidence reflection for the same wave-type), varies with
reflector depth for a xed offset. Also, in certain
For non-normal incidence, an incident P-wave cases, it is possible to get different offsets for a
leads to wave conversion in which both reflected xed value of reflection angle. These two cases
and transmitted P- and S-waves are sobtained. which occur in multi-reflector situation are
The obliquely-incident wave energy is dis- shown in Fig. 2.27.
tributed among all the converted waves in such a As expressed by Zoeppritz equations, the
way depending on the properties of the involved reflection coefcient shows variation with
media on both sides of the interface. According increasing angle of incidence (or with increasing
to Zoeppritz equations, the reflection coefcient offset). Depending on the distribution of the
is function of rock properties (density and elastic acoustic impedance on both sides of the inter-
properties) in addition to the angle of incidence. face, the reflection coefcient can vary from
For a given reflector, the amplitude variation large-negative to large-positive values. This
with angle of incidence (AVA) of a reflected behavior (variation of reflection coefcient with
seismic wave is found to be dependent on Pois- angle of incidence) can therefore be used as an
sons ratio as well as on impedance contrast indicator to predict lithological changes or type
across the reflection interface. In this way, the of fluid deposits.
parameter (AVA) possesses the same informa-
tion contained in a combined P- and S-waves
data. 2.6 Effect of the Medium on Wave
It is important to note that Zoeppritz equa- Energy
tions give direct relation of amplitude variation
with angle of incidence (AVA) and not ampli- Due to the earth ltering effect and other causes,
tude variation with offset (AVO). However, the wavelet generated by the source energy, is
offset is proportional to angle of incidence, in changed from its initial high-energy, impulsive
(a) surface
(b) surface
i1
i1
reflector-1 reflector-1
i2 i2
reflector-2 reflector-2
i3 i3
reflector-3 reflector-3
i1 > i2 > i3 i1 = i2 = i3
Fig. 2.27 Variation of reflection angle with reflector with increase of reflector depth for xed offset, and
depth for a xed offset and variation of receiver offset for b angle of incidence is constant for varying offset
a xed reflection angle. a Angle of incidence (i) decreases
2.6 Effect of the Medium on Wave Energy 47
Table 2.4 Factors 1. In the source zone 2. In the path zone 3. In the detector zone
contributing to energy
changes of a travelling Energy-Source parameters Reflection coefcient Detector response
seismic signal Source coupling Geometrical spreading Detector coupling
Near-source geology Inelastic attenuation Near-detector geology
Source-generated noise Wave conversion Surface noises
Noises and interferences Noises and interferences Noises and interferences
reflector
medium. It is dimensionless quantity and inde- parameter is dened to be the natural logarithm
pendent of frequency. The quality factor is inver- of the ratio of two neighboring amplitudes of a
sely proportional to the attenuation factor (a). The gradually fading wave-train. This is customarily
term 1/Q is called the specic dissipation. measured by the ratio of two amplitudes sepa-
rated by one wavelength (Fig. 2.30).
By denition, the logarithmic decrement (d) is
2.6.3 Seismic Wave Energy given by:
Measurement
and the DB Unit d lna1 =a2 ln ear =ear k ln eak
function of the frequency component of the travelling wave, is moving. The wave-train or the
travelling wave. Dependence of velocity on fre- energy package (expressed by the envelope of
quency means that each frequency component of the wave train) is travelling with different
a seismic signal moves with its own velocity. velocity called (group velocity, U) as shown in
Thus a wave, composed of several Fig. 2.32.
frequency-components will experience The group velocity (U) is mathematically
component-separation, and hence, related to the phase velocity (V) and wavelength
change-of-form that occurs during travel. Dis- (k) of the frequency component, by:
tortion of the wave-form due to dependence of
the velocity on individual frequency-components U V kdV=dk
is called (wave-dispersion).
The dispersion phenomenon leads to changing where, V, k, and dV/dk are average values for
of the shape of the wave train with travelled the range of frequencies making up the principal
distance. Each frequency component (that is, part of the pulse (Telford et al. 1990, p. 154).
each wave-phase) moves with its own individual When the phase velocity (V) increases with
velocity (the phase velocity, V). This is the increase of the component period, it is termed as
velocity with which a given point, marked on the (normal dispersion), and in this case the group
2.6 Effect of the Medium on Wave Energy 51
time
0
velocity is less than the phase velocity (U < V). Dispersion phenomenon occurs in a disper-
For the opposite case (inverse dispersion), it is sive medium, as when surface waves are travel-
when phase velocity decreases with period we ling through a semi-innite medium which is
get (U > V). In the absence of dispersion, the overlain by a low velocity surface layer. Dis-
two velocities are equal (U = V) and no distor- persion of seismic body waves (P- and S-waves)
tion to the wav-form occurs. are too small to be detected in practice.
The Seismic Velocity
3
acoustic impedance for the most commonly coefcients increase in such a way that it offsets the
known rock-types (Al-Sadi 1980, p. 70). effect introduced by the density increase. An
empirical relationship between P-wave velocity
(v) and bulk density (q) for the common sedimen-
3.2.2 Elasticity and Density Effects tary rocks (sandstone, shale, limestone, anhydrite
) is given as follows (Gardner et al. 1974):
The wave motion equation of seismic waves
includes the velocity factor present in its mathe- qv kv1=4
matical expressions. For a solid homogeneous
medium, the propagation velocity for P- and where (k) is equal to 0.31 when (v) is in m/s and
S-waves (vp and vs) are functions of elastic constants equal to 0.23 when (v) is in ft/s.
of the medium (Lames elastic constants, k and l) as Density values for common rock types are
well as its bulk density (q). These functions are: found in Table 3.1 presented above.
vp k 2l=q1=2
3.2.3 Porosity and Saturation Fluid
and, Effects
1=v =vf 1 =vm involve the effects of the rock and fluid contents
individually. It has the form:
or, in terms of interval transit time, T (= 1/v), 1=v C1 C2
noted that only a small amount of gas, present in resulting in a velocity increase with both depth and
pores, produces sharp decrease in velocity, but geological age.
further increase in gas saturation produces only a Faust formula (Faust 1951), for a given geo-
minor effect. This phenomenon was applied as a logical interval, can be presented in terms of the
tool for detection of hydrocarbon accumulation interval transit time, T (=1/v). This form is pre-
in oil traps. Being more compressible than oil or sented as follows (Pennebaker 1968):
water, gas has the effect of lowering the velocity
much more than the presence of oil or water as Th Chn
interstitial fluids. Due to its effect on the elastic
properties of rocks, even a small quantity of gas or,
present in the pores, would result in large
log Th n log h log C
reduction in velocity (Sheriff and Geldart 1995,
p. 109). Clearly, this property has an important
where (C) is constant related to the rock lithol-
application in hydrocarbon exploration.
ogy, pore pressure, and geological age.
For sand-shale sequence, Pennebaker (1968)
found that the index (n) has a value of about
3.2.4 Depth and Geological Age
(1/4). This is considered to be representing a
Effects
geological section existing under normal
compaction-pressure. As it is seen above, the
With depth of burial and with geological age, in
(T-h) relation is linear when it is expressed in the
general, rock material becomes more compact.
log-domain. Thus, when a linear (log T log h)
This would result in increase in elastic constants
relation is obtained in a certain study, the geo-
as well as in the density and, consequently,
logical section is considered to be under normal
increase in velocity. Dependence of velocity (v),
compaction pressure conditions. Deviation from
of a particular geological bed, on geological age
the linear relation (straight-line relation) is nor-
(g), existing at depth (h), is expressed by the
mally considered as indicative of abnormal
following empirical formula (Faust 1951):
pressure situation. This is another useful and
effective tool that can be used to investigate
vg; h kg hn
subsurface pressure conditions and as an indica-
where (k) is constant (equal to 125.3 when depth tor for other stratigraphic features.
in feet, velocity is in ft/s, and geological age in
years). The index (n) is found in this study to be
equal to (1/6). In a later study (Gregory 1977) 3.2.5 Overburden Pressure Effect
made on sand- and shale-formations data, the
index (n) is found to be equal to (1/4). A theoretical relationship developed by Gassman
Variation of velocity with the depth of burial (1951) showed that velocity of a rock-bed model
was investigated by a number of geophysicists made up of spherical tightly-packed grains varies
(e.g. West 1950; Kaufman 1953; Acheson 1963), as (P1/6), where P is the applied compaction
conrming the general trend of velocity increase pressure. The similarity of this result to Fausts
with depth. empirical formula (Faust 1951) suggests that
As the function, v(h, g) is showing, geological increase in velocity (i.e. decrease in the interval
age bears a relationship with velocity similar to transit time) is due to increasing pressure
that of the depth factor. It shows that velocity is in imposed by the overburden rocks.
general increasing with depth and geological age. Compaction of a saturated porous layer is
The explanation is that, grain compaction increa- function of the weight of the overlying material (i.e.
ses with increase of depth and geological age. In due to the overburden pressure, Po) and the pressure
consequence, elastic constants are increased of the saturation fluid (the pore pressure, Pp). Since
3.2 Factors Influencing Seismic Velocity 57
the effect of (Pp) in compressing the porous rocks is phenomenon. In case of anisotropic medium,
in opposite direction to that created by (Po), the velocity varies with changes in propagation
compaction net effect (compaction pressure, Pc) is direction. The main factors affecting velocity can
proportional to the pressure difference (Po Pp). be summarized as follows:
That is:
(i) Lithology
Pc Po Pp In general, igneous and metamorphic rocks
have larger velocity values than sedimen-
As rock compaction process brings about a tary rocks. Sub-types of rock lithologies
corresponding change in the physical properties, it are characterized by wide ranges of
is expected that seismic velocity (or interval transit velocities which are overlapping with each
time) would vary with the applied compaction other. For example, the velocity-range for
pressure. Velocity increases with increasing sandstone (25) km/s is overlapping with
compaction pressure and decreases with increase that of shale (1.54.5) km/s, and so on.
of interstitial pore pressure. This velocity-pressure (ii) Elasticity and Density
relationship forms the basis used in predicting Theoretically, velocity is directly propor-
subsurface pressure conditions from seismic tional to elastic constants and inversely
velocity data. Use of seismic velocity as an indi- with density. However, because of the
cator of pressure conditions was used by several greater effect of elasticity, it is observed
workers like Pennebaker (1968), Reynolds (1970), that velocity has an apparent direct pro-
Louden et al. (1971), and Aud (1974). Departure portionality with density changes.
of the measured velocity of a given geological (iii) Porosity and Saturation Fluid
section from the normal-compaction trend serves Velocity is inversely proportional to
as indicator of abnormal-pressure zone. porosity, and the change in velocity
depends on the type of pore fluid. Velocity
is lowered when the pores are lled with
3.2.6 Other Velocity-Affecting water or oil, and more lowered when the
Factors pores are lled with gas.
(iv) Depth and Geological Age
In addition to the above mentioned factors, there In general, velocity increases as both of
are other less important factors which can cause the overburden depth and geological age
small effects on seismic velocity. Examples of increase. A power-law relationship is
such factors are pore shape, anisotropy, tempera- connecting the velocity with the depth and
ture, and wave-frequency. It is found that velocity geological age. The power index for
decreases slightly with increase of temperature. It sand-shale sequence is found to be around
is found that, for an increase of 100 C, velocity the value of (1/61/4).
decreases by about 5 % and that the velocity (v) Overburden Pressure
drop-rate is considerably more when the rocks are Due to compaction caused by the overbur-
saturated with heavy crude oil or tar. Velocity in den weight, elastic constants as well as
water saturated rocks experiences sudden increase density get increased. There is a direct rela-
as temperature is lowered to the freezing point tion between velocity and the net com-
(Sheriff and Geldart 1995, pp. 120121). paction pressure which is equal to the
Velocity is practically unchanged over a broad difference between pore pressure and over-
frequency range. Small change is observed in case burden pressure (geostatic pressure as it is
of body-wave dispersion (velocity variation with sometimes called). Velocity increases with
frequency). Low-frequency components are fas- increasing compaction pressure and decrea-
ter because of frequency-dependant absorption ses with increase of interstitial pore pressure.
58 3 The Seismic Velocity
vp vp vp vp
water saturated
oil
gas
oil saturated
Pore pressure fluid saturation temperature frequency
(vi) Other Minor Factors express the propagation velocity (V) as a math-
Pore shape: Velocity decreases when ematical function of the travel-distance, V(x) or
pores are elongate. function of travel-time, V(t). For practical
Anisotropy: Velocity varies with changes applications, velocity is, more often, expressed as
in propagation direction. function of vertical reflection time, that is, as
Temperature: Velocity slightly decreases function of the two-way (reflection) vertical time
with increase of temperature. V(T). In a homogeneous, uniformly changing, or
Frequency: Velocity practically unchan- layered medium, the distance-versus-time func-
ged with frequency. With normal disper- tions can be linear, curved, or segmented
sion, low-frequency components are respectively. The corresponding V(T) functions
faster because of frequency-dependant can be plotted as curved or straight lines. Typical
absorption phenomenon. velocity functions are shown in Fig. 3.3.
Due to the geological complexity of the earth,
Main factors affecting velocity are shown in it is not, in practice, possible to express seismic
Fig. 3.2. velocity-variation in the form of an explicit
mathematical function. However, under certain
conditions, a geological medium is approximated
3.3 The Velocity Function by a dened set of specications that can allow
derivation of a mathematical function describing
Seismic propagation velocity, as we have seen the variation of velocity along the travel-path.
from the previous discussion, depends on a One of these models is one in which the instan-
number of variables, like type of lithology, geo- taneous velocity (V) increases linearly with depth
logical age, density, porosity, pressure, and depth. (h). This model is expressed by the linear func-
In seismic reflection exploration work, velocity is tion; V(h) = V0 + kh, where (V0) is the velocity
usually presented as mathematical functions. The value at the surface (h = 0) and (k) is the
velocity function is taken to describe its variation velocity-depth gradient. The signicance of this
with reflection travel-distance (or, more often, relation is that it gives a close approximation to
with travel-time). the actual velocity variation observed in many
Since the real earth is not homogeneous sedimentary-basin areas (Dobrin 1960, p. 77).
(typically layered medium), seismic waves move Derivation of the ray-path geometry (which turns
with different velocities in different parts of its out to be of circular form) and travel-time func-
travel-path. In this case, it becomes very useful to tions, are given by several authors (see for
3.3 The Velocity Function 59
example, Nettleton 1940, pp. 257261 or Sheriff of the location in the travel-path of the moving
and Geldart 1995, pp. 9394). wave front. This means that at any point of the
travel-path, the velocity can be measured or
computed at that point. This is the (instantaneous
velocity) which can be dened to be the velocity
3.4 Types of Velocity Functions
of wave motion at a given point in a medium
traversed by an advancing seismic wave.
Unlike electromagnetic waves which are all of
Mathematically, the instantaneous velocity (v) at
one type (e.g. wave of visible light which is of
a point is given by the slope of the tangent to the
S-wave type), seismic waves are of many types,
distance-versus-time curve at that point (Fig. 3.4).
each of which has its own velocity-value which
The instantaneous velocity, (v) of a wave
is dependent upon the density and elastic con-
moving along a distance (x) is dened as:
stants, as mentioned above. Because of the
inhomogeneous nature of the crustal part of the v dx=dt
Earth (which is typically made up of layered rock
media), several types of velocity functions are
needed to express velocity variation as function
of travelled distance. Five velocity-types are in
common use in seismic exploration. These are:
(i) Instantaneous Velocity (V) v = dx/dt
(ii) Interval Velocity (IV)
(iii) Average Velocity (AV)
(iv) Root Mean Square (RMS) Velocity (RV)
x(t)
(v) Stacking Velocity (SV)
The closest example to the instantaneous covering that distance. In seismic exploration
velocity is the velocity function in which the work, the average velocity is computed for a set
interval transit time (which is equal to the of layers, usually starting from the datum plane
reciprocal of velocity) is measured across each down to the required reflector level.
depth-interval of one meter down a drill hole. For a vertical travel-path, the average velocity
The resulting record (the sonic log) gives detailed of a geological section made up of (n) layers will
velocity-information at points spaced by 1 m be given by dividing the total thickness hn (=
interval. This is a practical approximation of the Dh1 + Dh2 + Dh3 + + Dhn) by the travel-time
instantaneous velocity-variation with depth. tn (= Dt1 + Dt2 + Dt3 + + Dtn), that is:
AVn hn =tn
depth, h
3.4 Types of Velocity Functions 61
depth, h
of that layer respectively. Using the corre- medium (medium made up of horizontal layers of
sponding travel times, (Tn) and (Tn1), the different interval velocities). A three-layer model is
required formula is: shown in the following Fig. 3.7.
3.4.4 Root Mean Square Velocity The root mean square velocity (RV) can be
calculated from the interval velocity data (IV),
Root mean square (RMS) velocity (RV) is dened down to the nth layer, by use of the following
as the square root of the average of the squares of the formula:
weighted interval-velocities, where the weighting From the RMS velocity (RV), it is possible to
factors are the layer thicknesses or the interval transit derive the interval velocity (IV), given the two
times. In seismic exploration, the RMS velocity, like RMS velocities for the section, from the datum
other types of velocities, is usually computed for level down to the base (RVn) and down to the
vertical ray paths of waves traversing a multi-layer top (RVn1) of that layer. Using the
Layer-3
RV3 = [(IV21 t1+IV22 t2+IV23 t3)/(t1+t2+t3)]1/2
h3 , t3
depth, h
62 3 The Seismic Velocity
corresponding travel times, (Tn) and (Tn1), the gather-traces. This is one of the fundamental
required formula is (Dix 1955): processing steps (called NMO correction) which
For a given geological section, the RMS velocity bring about coherency of the reflection arrivals
is typically a few percents larger than the corre- which, on stacking, give enhanced reflection
sponding average velocity (Sheriff 1973, p. 228). event. It is called stacking velocity because of its
role in enhancing the stacked reflection signal.
Since its direct role is in the NMO correction, the
3.4.5 Stacking Velocity term (NMO velocity) will be more appropriate
than the commonly applied term (stacking
Stacking velocity is the main velocity variable velocity). The CMP concept is explained in
that enters in the NMO-correction formula (ex- Sect. 4.4.
plained in Sect. 4.3). It is applied to remove the Role of the stacking velocity in enhancing the
time-contribution of the receiver-to-source dis- reflection signal is schematically shown in
tance (called, offset) from the total reflection Fig. 3.9.
travel-time. This is claried in the following In dipping parallel reflectors (dip angle, h),
(Fig. 3.8). with parallel velocity layering, the stacking
Stacking velocity is determined by velocity velocity (calle it VS) is equal to RMS velocity
analysis technique whereby the time contribution (VRMS) divided by cosine of the dip angle (h)
of offset is removed before stacking of the CMP that is (Sheriff and Geldart 1995, p. 134):
T0 = 2h/V
Tx = (4h2 + x2 )1/2 / V
CDP-gather stack traces other than the true propagation direction. In seismic
A B C exploration, the apparent velocity is often dealt
0 x
with in connection with the movement of plane
wave-front advancing along a path inclined with
respect to the ground surface. A plane wave
A approaching a ground surface with true velocity
(V), along a ray-path making an angle (h) with the
B normal to the surface will have an apparent velocity
(Va) of its motion along the surface (Fig. 3.10).
During a time interval (t), the wave front
C moves a distance of (Vt) with the true velocity
(V), while, at the same time, the moved distance
T0 Tx at the surface (x) is covered by the apparent
velocity (Va), that is:
Fig. 3.9 Role of the stacking velocity in the NMO-
correction of a CDP trace gather. In this model, three
stacking velocities were applied (cases: A, B, and C). The Va Dx=Dt
optimum velocity (case, B) gave the highest staked signal
hence,
VS VRMS =cos h Va V=sin h
This relation between the stacking velocity This relation shows that the apparent velocity
and RMS velocity can be applied by interpreters is always greater than the true velocity by a
in their interpretation work of seismic reflection factor depending on the angle of approach (h, in
data and in velocity-changes studies. this example). The apparent velocity approaches
Stacking velocity is sometimes called RMS innity when (h = 0), that is when the wave path
velocity, because stacking velocity is the nearest direction is perpendicular to the surface plane. In
in value to RMS velocity in a multi-layer med- any case the apparent velocity is always greater
ium. Like RMS velocity, stacking velocity is than the true velocity with which the wave is
slightly greater than the average velocity. approaching the horizontal surface plane.
Another point of interest which is related to
this subject is the apparent wavelength. Suppose
3.4.6 The Apparent Velocity that the measurements were taken for time of one
period (t = s), distance measured over the sur-
The Apparent Velocity (Va) is dened as the face corresponding to time of one period will be
propagation velocity measured in the direction
x
Va t
Wave Vt
wave motion
front
direction
Fig. 3.10 Ray path geometry of a plane wave approaching the surface plane at an approach angle (h). Velocity in the
ray-path direction (V), and velocity of wave front measured on the horizontal plane surface is the apparent velocity (Va)
64 3 The Seismic Velocity
one apparent wavelength (x = ka), hence (using 3.4.8 Representations of the Velocity
the relationship; k = Vs): Functions
depth
3.5.2 The Up-Hole Velocity Survey The electronic structure of the sonde is
designed in such a way that the output is made to
This method is applied to determine velocity be the difference in the travel-times to the two
changes in relatively shallow depths, in the range receivers. The time difference, measured in
50100 m. It typical application is velocity- time-units per one-foot, called (interval transit
determination of the near-surface weathered zone. time), is plotted (normally in micro-second units)
The same principles used in well-velocity survey- against depth to give the continuous wiggly
ing are applied in this method, but with the curve known as the (sonic log). Sonde basic
source-detector conguration reversed. In the structure is shown in (Fig. 3.14a).
up-hole case, the sources (small dynamite charges) To avoid tilting and hole-irregularities effects,
are placed inside the borehole at few-meter spacing two seismic-pulse sources are used instead of one,
and sequentially red. The detection system is making what is called a (borehole-compensated
placed on the surface at a location near the sonde) as shown in (Fig. 3.14b). Seismic pulses
well-head. Velocity computation follows the same are emitted alternately from the two sources and
equations quoted above (Fig. 3.13). the transit times from the two oppositely traveling
refracted P-waves are averaged electronically.
The borehole compensated sonde (BHC) gives
an average interval transit time which is plotted
3.5.3 Continuous Velocity Logging
on a paper strip. The produced log in this case is
normally referred to as BHC sonic log.
This method uses a recording system consisting
The BHC-log is used in computing synthetic
of a borehole-logging tool, called the (sonde). In
seismograms, in identifying lithologies, and in
its standard form, it contains a seismic-pulse
determining formation boundaries (Fig. 3.15).
source and two receivers, one foot apart, and
The interval transit time can be integrated down
source placed three feet from the nearest receiver.
the well to give the total travel time of the part of
Sound pulses, emitted from the source at uniform
the well for which continuous velocity logging
time intervals, are detected by the two receivers.
was conducted.
Due to the fact that P-wave velocity in the dril-
ling fluid is lower than that of the rock-medium Seismic Data Analysis Methods
surrounding the well, the transmitted P-wave gets Reflection travel-time function, in a homoge-
refracted at the wall-side, moving into the rock neous and isotropic medium, is of a hyperbolic
medium and then recorded by the receivers.
sonic log
(a) (b) geological column transit time, sec
140 90 40
shale
S S
R limestone
R R
R
T0
T T
in the analysis is explained in more details in Since density variation is very small, com-
Sect. 10.7.1. pared with velocity variation, that is putting
(q1 = q2), the direct (A-V) relationship can be
3.5.7 Seismic Velocity Inversion obtained which is:
Table 3.2 Fields of application and precision assessment of various types of velocity (after Al-Chalabi 1979)
Velocity Main uses Precision requirements
Stacking velocity Stacking of seismic sections Modest to low
Migration processing Modest to low
Estimation of RMS velocity Dependant on situation
RMS velocity Migration-velocity estimation Generally modest
Interval-velocity estimation Dependant on situation
Average-velocity estimation Dependant on situation
Interval velocity Lithologic and stratigraphic studies High to modest
Interpretation works Modest to low
Abnormal pressure detection High to modest
Ray tracing computations Dependant on situation
Migration processing Generally modest
Average-velocity estimation Dependant on situation
Average velocity Depth conversion Generally modest
Interpretation works Modest to low
Precision requirements: high = 0.11.0 %, modest = 15 %, low >5 %
static and dynamic (NMO) corrections. Amplitude conversion and in mapping structural and strati-
compensations (as in geometrical spreading and graphic features.
inelastic absorption) and seismic migration Summary of elds of application and preci-
depend on velocity. In interpretation activities, sion assessment of the various velocity-types are
velocity has a fundamental role in time-to-depth given in the following Table 3.2.
Seismic Wave Reflection
and Diffraction 4
In the real Earth Crust, a seismic wave may meet A diffraction event bears a marked relation
a variety of geological changes. Typically, the with reflection events since both are types of
media traversed by seismic waves are made up of seismic energy generated from an intervening
layered rock formations of different physical reflector. A plane surface (surface-reflector)
properties and different geometrical shapes. In causes reflection and point obstacle (point-
such environments, some of the seismic wave reflector) causes diffraction.
energy gets reflected from interfaces or diffracted
from structural obstacles. The rest of the incident
wave is transmitted through the interfaces with
their ray-paths being bent (refracted) in case of 4.2 Wave Changes at Reflection
inclined incidence and with no bending when Interface
ray-path is perpendicular to an interface.
In this chapter, discussion shall deal with the The reflection process involves two main types
reflection and diffraction of seismic waves since of changes. These are: change in the propagation
they are physically more closely related to each direction and change in the energy content
other. Transmission with its special case, the (Fig. 4.2).
refracted transmission (refraction) shall be dealt The two types of changes occurring in the
with separately in the following chapter. reflection phenomenon involve the wave energy
content and its travel-path geometry. These are:
Fig. 4.1 The four principal types of wave arrivals (direct, reflection, refraction, and diffraction) representing the
commonly recorded seismic events
SV
amplitude) compared with the incident wave. Sometimes, the reflection coefcient is dened
The ratio of the reflected amplitude to that of the in terms of energy ratios instead of amplitude
incident represents the reflection coefcient. ratios. With this approach, the coefcients are
Zoeppritz equations provide the mathematical expressed by squares of the amplitudes.
expressions for the reflection and transmission By applying the stress and strain continuity-
coefcients of all of the produced waves (two conditions at the interface level (z = 0), it is
P-waves and two SV-waves in this example). possible to derive expressions for the reflection
The important feature here is that the reflection and transmission coefcients in terms of the
and transmission coefcients are functions of contrast in acoustic impedance (see Sheriff and
both of the angle of incidence and the contrast in Geldartl 1995, p. 76). The reflection coefcient
acoustic impedance existing across the interface. (R = Ar/Ai) and transmission coefcient
The fundamental theoretical principles upon (T = At/Ai) for the case of normal incidence, are
which solution of Zoeppritz equations is based given in terms of the contrast in the acoustic
on, is the fulllment of the conditions which impedance (Z) of the two media on either side of
require that all normal and tangential stresses and the interface. Thus, by denition, the reflection
displacements at the interface are continuous. For and transmission coefcients (R and T) are:
displacement continuity, the sum of normal and
tangential displacement-components (at the R Ar =Ai q2 v2 q1 v1 =q2 v2 q1 v1
interface) in the rst medium must be equal to Z2 Z1 =Z2 Z1
the corresponding sum in the second medium. T 1 R At =Ai C=A 2q1 v1 =q2 v2 q1 v1
Concerning stress, it is required that, the sum of 2Z1 =Z2 Z1
normal and tangential stress-components are
similarly equal. Another important note in this
If (R) is positive, compression displacement is
contest is that Zoeppritz equations are based on
reflected as compression because, by reflection in
the assumption that all the involved waves are
this case, both of the wave-motion direction and
pure sine waves of same frequency which drops
particle displacement are reversed. If, however,
out of the equations (Richter 1958, p. 670). This
(R) is negative, a compression is reflected as
means that no frequency change taking place in
dilatation, meaning that we get (180) phase
reflection and transmission processes.
change in this case.
It is evident from these equations (case of case where the acoustic impedance in one of the
normal incidence) that the reflection coefcient two adjacent media is approaching zero or innity.
depends only on the contrast in the acoustic Thus, when (Z1) is very small compared with (Z2),
impedance (Z2 Z1). The greater the contrast, that is when (Z1) approaches zero, (R) approaches
the larger reflection coefcient will be. This unity (R = 1). In this case, the interface is consid-
implies that no reflection occurs from an inter- ered to be ideal reflector since this means that all of
face across which the acoustic impedance the incident energy is reflected and no part of it is
assumes the same value, even if either velocity or transmitted through. A good approximation of this
density varies individually across the interface. kind of situation is the earth free-surface which is an
The reflection coefcient ranges in value from interface between the upper air-medium and the
(1) to (+1) depending on the acoustic impe- lower rock-medium. Because of the vast difference
dances (Z1 and Z2). When the acoustic impe- between the acoustic impedances of these two
dance (Z2) in the second medium (medium in media, the reflection coefcient approaches the
which transmission occurs) is greater than that in value of (+1) for a source located in the air and (1)
the rst medium (medium of incident wave, Z1) a for a source located inside the rock-medium. The
compression displacement is reflected as com- interface in both cases is equally efcient in
pression, and the reflection coefcient is positive. the reflecting process, but in the second case there
In the opposite case, that is when (Z1 > Z2), a occurs a polarity reversal, or phase change of (p) for
compression is reflected as rarefaction (phase an incident sine wave.
change = p) and the reflection coefcient
becomes negative. Reflection coefcient is zero
when the two acoustic impedances (Z1 and Z2) 4.2.3 Geometry of Reflection
are equal, which means that there is no interface from Dipping Reflectors
existing in the way of the incident wave.
Another useful feature which may be deduced The general case of reflection geometry is the
from the reflection coefcient expression, is the case of inclined (dipping) reflector (Fig. 4.5).
(b) x 0 x
Xm
travel-time
apex
curve
T(x)
4.2 Wave Changes at Reflection Interface 75
V1 horizontal reflector
V2
x 0 x
apex
travel-time
curve
T(x)
DT x2 =2T0 V2
4.3.1 The Normal Move-Out (NMO)
Concept This is considered as an accepted approxi-
mation since (x/VT0 1), which is usually the
The Normal Move-Out (NMO) is dened as the case in seismic reflection exploration. In this
difference (T) between reflection travel-time mathematical process, the exact form of (T) is
(Tx) and the two-way vertical travel-time (T0). transformed from its hyperbolic function to the
For a single horizontal reflector found at the base approximate form which is a parabolic function.
of a homogenous layer (of constant velocity), the Both forms of these two equations show that
NMO parameter (T) is given by Fig. 4.8. (T) is function of three variables; the receiver
4.3 The NMO and DMO Concepts 77
surface
reflector-2
reflector-3
Tx
T
z
reflector
Fig. 4.8 The NMO concept. a Two-way vertical ray-path. b Two-way slant ray-path. c Reflection arrival times
corresponding to the ray-paths of a and b. The NMO (DT) is shown as the difference, (Tx T0)
T T T
x T0 V
Fig. 4.9 NMO (DT-function) direct variation with (x) and inverse variation with both T0 and V
offset (x), velocity (V), and the two-way vertical For a given offset (x), the time (Tx) can be
time (T0). The proportionality is direct with (x2) readily measured for a certain reflection event
and inverse with both (V) and (T0). These appearing on the seismic trace recorded at that
changes are shown as follows Fig. 4.9. offset. From these data, the NMO (T) can be
78 4 Seismic Wave Reflection and Diffraction
Tx Tx
R S R R S R
surface
reflection
ray-path
S R surface S
S3 S2 S1 R R1 R2 R3
reflection ray-path
horizontal reflector
CDP
horizontal reflector CRP
DP CMP
RP
MP Fig. 4.13 Denition of the common points (CDP),
(CRP), and (CMP), which coincide on each other in case
Fig. 4.12 Denition of the depth point (DP), located of horizontal reflector
vertically below the mid-point of the source-receiver
distance. It is the reflection point (RP), or mid-point (MP)
surveying. Application of the concept usually
results in great enhancement of the reflection
These points are described as common points signal by summing together (stacking) the
(CDP, CRP, CMP) in cases where multi- CDP-Gather traces.
reflections occur for the particular point. In other
words, a point is described as common when it
becomes common to more than one reflection. 4.4.2 CDP, CRP, and CMP in Case
of Dipping Reflector
4.4.1 CDP, CRP, and CMP in Case For a horizontal plane reflector, the points (DP,
of Horizontal Reflector RP and MP) coincide at a point which is located
vertically below source-to-receiver midpoint. In
By use of certain source-receiver layout, it is the case of dipping reflector, however, the
possible to shoot a number of shots such that the reflection point gets shifted up-dip by a distance
reflection points resulting from this number of depending on the offset as well as on the dip
shots coincide on each other, that is one angle of the reflector. The up-dip shift is given
point-location will serve as common point to all by:
of the implemented shots. In this type of repeated
shooting-spread, the reflection point becomes DL x2 cos h sin h=D
known as common depth point (CDP), common
reflection point (CRP), or common mid-point where (L) is the up-dip shift of the reflection
(CMP). These points are shown in Fig. 4.13. point (RP), (x) is half the source-receiver offset,
The four reflection ray-paths, shown in (h) is reflector dip-angle, and (D) is the length of
Fig. 4.13, have one common depth point (the reflection ray-path which is normal to the
CDP), will each produce a seismic trace. These reflector (Deregowski 1986). The ray-path
traces belong to the same depth point (CDP) and geometry and location of the reflection point
thus they form a group of traces called the (RP) are shown in Fig. 4.14.
(CDP-Gather). In this case (case of horizontal In case of repeated shooting (as it is done in
reflector) the CDP and the other points (CRP and the usual seismic reflection proling surveying),
CMP) will coincide on each other. the reflection points (RPs) get dispersed
Ever since the CDP concept was introduced along the dipping reflector-plane. For a given
by Mayne (1962, 1967) it has been applied on a reflector, the amount of up-dip shift (L)
routine basis in execution of seismic reflection increases with the square of the receiver-offset
4.4 The CDP, CRP, and CMP Concepts 81
RP3
RP2
RP RP1
82 4 Seismic Wave Reflection and Diffraction
S
S3 S2 S1 R R1 R2 R3
CMP
CRP
Fig. 4.16 Denition of the common reflection point (CRP) obtained from multi-reflection experiment. For dipping
reflector, the (CRP) is located at one reflection point which is coincident with the end point of the normal from the CMP
diffracted wave
spherical wave fronts
in every possible direction giving a phenomenon this type is normally referred to as a point-
called (diffraction) and the wave which leaves the diffractor.
obstacle after incidence is called (diffracted The diffraction phenomenon can be explained
wave). by considering the diffraction point as a
A closely related term is (wave scattering) point-source which, upon being hit by an incident
which is used to describe diffracted wave-eld wave, becomes an activated source that radiates
caused by small structural irregularities, as for waves in all directions. According to Huygens
example found with seismic energy reflected principle, an obstacle hit by an incident wave
from rugged basement surfaces. becomes an energy-source from which seismic
waves are generated and transmitted in all possible
directions. If a plane seismic wave, for instance hits
4.5.1 The Point-Diffractor a point-diffractor embedded in a homogeneous
medium, a diffracted wave with spherical wave
A structural obstacle, having radius of curva- fronts will be generated. The diffracted wave will
ture which is shorter than that of the incident move away from that source-point causing inter-
wavelength, acts as a diffraction-generating ference with the incident wave-train and all other
point (or diffraction point). An obstacle of waves that may be coexisting at the time (Fig. 4.17).
4.5 The Seismic Wave Diffraction 83
reflected wave
incident wave
terminating reflector
terminating reflector
fl t point-
point-diffractor
Point
point diffractor
diffractor
diffracted
spherical
wave front
84 4 Seismic Wave Reflection and Diffraction
0 0
Tr Tr Td Td
depth
point source
x
T0
Tx
seismic image
diffraction
time hyperbola
edge, we have for the same offset (x), the travel and,
time function (TDx) for diffraction, and (TRx) for
the reflection, are given by: DTd x2 =T0 v2
h i1=2 giving,
TDx x=v2 T0 =22
h i1=2 DTd 2DTr
TRx x=v2 T20
This shows that the normal move-out of a
and, wave diffracted from a terminating reflector is
approximately equal to double that of a wave
h i1=2
DTd TDx T0 x=v2 T0 =22 T0 =2 reflected from the same interface at the same
h i1=2 offset. This important feature is used in dis-
DTr TRx T0 x=v2 T20 T0 criminating diffraction from reflection events.
where, (Tr) and (Td) are the normal move-out 4.5.5 The Diffraction Hyperbola
for the reflection and diffraction respectively, and
(v) is the propagation velocity. For small The seismic image of the diffraction arrivals is (for
offset-to-depth ratio, that is for (x/vT0 < 1), these constant velocity) a hyperbolic curve centered
equations can be approximated by: about the diffraction source-point. A diffraction
event such as this is expected to appear on a seis-
DTr x2 =2T0 v2 mic section made up of zero-offset traces. The
86 4 Seismic Wave Reflection and Diffraction
diffraction point
diffraction hyperbola
seismic stack section is effectively a section of of their interferences, these arrivals are distorting
zero-offset traces (Fig. 4.21). or masking the reflection events and causing
Diffraction hyperbolae are often observed on smearing effects at the fault-zone leading to
seismic stack sections, indicating point-source decrease in fault-resolution. Example of the
diffractions, as those generated by faults and interferences introduced as a result of fault-
pinchouts. A seismic diffraction event (diffrac- generated diffraction hyperbolae are shown in
tion hyperbola) appearing on seismic stack sec- Fig. 4.23.
tion can be considered as the seismic response of Although diffraction events (diffraction
a depth section of a homogeneous medium con- hyperbolae) are introducing masking effects to
taining a point-source diffractor (Fig. 4.22). the primarily targeted reflection events, they
(diffraction events) can help interpreters in
identifying faults and other diffraction-sources.
4.5.6 Distortion Effects of Diffraction In general, diffraction events on stack sections
are considered to be unwanted distortive events
In nature interfaces are not always plane and which must be removed or at least attenuated as
continuous surfaces. There are cases where these much as possible. The common way to remove
surfaces are irregular as the surfaces of reef the diffraction events (diffraction hyperbolae) is
bodies or discontinuous as the faulted beds or by application of the seismic migration which is
pinchouts as found with angular unconformities. one of the principal steps applied in processing of
When a seismic wave is incident on such sub- seismic reflection data.
surface features, diffracted waves are generated
and transmitted through the medium, interfering
with, and distorting other co-existing waves such 4.5.7 The Exploding-Reflector Model
as reflection events, the main objective in normal
seismic exploration. In accordance to Huygens principle, the
On a seismic stack section, which is effec- exploding reflector model is a model in which a
tively made up of zero-offset traces, the diffrac- reflection interface is considered to be formed of
tion arrivals (diffraction event) appear as a innite number of closely packed diffraction
hyperbolic curve whose apex is coincident with points and that these points explode at a given
the causing diffraction point. Faulting is one of start-time, generating seismic waves. The
the principal sources of diffraction waves seen in exploding reflector modeling process requires
stack sections. A reflector termination caused by that each point of the refelector is sending seis-
faulting, generates diffraction arrivals. Because mic rays, in every direction (exploding) at a
4.5 The Seismic Wave Diffraction 87
seismic
image
diffraction hyperbolae
common start time. The moving seismic ampli- wave fronts of waves diffracted from closely
tude is considered to be of magnitude propor- packed diffraction points (making up the reflec-
tional to the normal-incidence reflection tion surface), is called exploding reflector model.
coefcient. The resulting wave eld generated according
The generated waves are made to propagate to to the exploding-reflector model shall propagate
the surface with velocity half that of the actual along the normal-incidence ray-paths. The con-
velocity. With this velocity-value the one-way cept is schematically shown in Fig. 4.24.
travel-time to the surface becomes equal to the Compared with the conventional record sec-
two-way reflection travel-time for zero-offset tion, the section produced from the exploding
receivers placed on the surface (Sheriff 2002, reflector model, is the same as the conventional
p. 127). Consideration of the reflection process as section having propagation velocity equal to half
being formed from the interactions of seismic of the true velocity (Fig. 4.25).
88 4 Seismic Wave Reflection and Diffraction
Application of the exploding reflector concept in of migration (both of the pre- and post-stack
computing the seismic reflection section is a migration processing) using the exploding-reflector
direct-modeling (forward-modeling) process. concept. This approach proved to result in great
Algorithms have been developed for computations reduction of the migration computational cost.
Seismic Wave Transmission
and Refraction 5
The term (transmission) is customarily used to layers down to a reflector and up to the surface.
indicate the general case of wave propagation Among other factors, reflection and transmission
where the moving seismic wave crosses an coefcients cause attenuation to the wave energy
interface whether the incidence is normal or as it is experienced by the recorded function r(t).
inclined. When an obliquely incident wave is It should be remembered that there is a reflection
transmitted across an interface, it is bent at the and transmission processes taking place at each
interface towards, or away from, the normal at interface existing in the traversed geological
the point of incidence. This is the well known section (Fig. 5.1).
phenomenon of (refraction). A refracted wave is,
therefore a transmitted wave resulting from an
inclined incidence. No refraction occurs when 5.1.1 Transmission Coefficient
the wave is incident perpendicularly at an inter- at Normal Incidence
face. Thus, it can be sayed that the refraction
phenomenon is a special case of transmission. It The transmission coefcient (or transmittance,
(refraction) occurs only with normal incidence. as it is sometimes called) is dened to be the
ratio of the amplitude of the transmitted wave to
that of the incident wave. Sometimes this is
5.1 Seismic Wave Transmission expressed in terms of energy instead of ampli-
tude. At normal incidence, the transmission
When a seismic wave hits an interface, part of the coefcient (T), expressed as a ratio of the
wave energy is reflected and the rest is transmitted transmitted amplitude (At) to the incident
across that interface. The amount of energy which amplitude (Ai), thus:
is reflected from an interface is determined by the
reflection coefcient and what is left from the T At =Ai
incident energy is transmitted across the interface
into the second medium. At the interface the As it is with the reflection coefcient, the
ray-path is bent if the incidence is oblique and transmission coefcient is function of the
continues with unchanged direction when the acoustic impedances of the two media separated
incidence is normal to the interface. by the involved interface. For an interface sepa-
In the real layered Earth crust, the initiated rating two media of acoustic impedances (Z1 and
seismic energy, the source function s(t), is Z2), the transmission coefcient (T), is related to
recorded as r(t) after it penetrates all the rock the impedances by:
R4 T12 1 R 1 R 1 R2
Fig. 5.1 Wave-attenuation due to reflection and trans- And in terms of impedances it is given by:
mission processes which are normally occurring in a real
layered rock medium. s(t) and r(t) are source receiver T12 4 Z1 Z2 =Z2 Z1 2
time-functions respectively
The two-way transmission coefcient (T12) is
useful in computing the effective attenuation
T 2Z1 =Z2 Z1
factor of a wave reflected from a subsurface in-
It is apparent from this formula that as the terface after being transmitted through several
contrast in the acoustic impedances is smaller the layers during its total reflection travel path.
transmission coefcient becomes greater, and in
the limit when they are equal (Z1 = Z2), trans- 5.1.3 Attenuation Due to Reflection
mission coefcient (T) becomes unity and all of and Transmission
the seismic energy shall pass through and no
reflected energy will take place. For an n-layer medium, the wave which is
For the case of inclined incidence, the trans- reflected from the nth interface and received at
mission coefcient, like the reflection coefcient, the surface would have crossed the (n 1)
depends on both of the angle of incidence and on interfaces twice. Assuming normal incidence, the
the contrast in the acoustic impedance of the effective attenuation factor (RTn) of a wave
adjacent media. Unlike reflection, however, there reflected from the base of the nth layer (nth
is no special situation whereby phase change subsurface interface) and transmitted through
occurs. (n 1) interfaces will be given by:
RTn An =Ai
5.1.2 The Two-Way Transmission
Coefficient By use of the concept of the two-way trans-
mission coefcient, it is possible to derive a for-
The transmission coefcient (T) is related to the mula that computes the net attenuation due to the
reflection coefcient (R) by the expression (see combined effects of the reflection and transmis-
Sect. 4.2.2): sion coefcients. In a multi-layer geological model
5.1 Seismic Wave Transmission 91
R3
R4
A1 Ai R1
5.1.4 Role of Transmission in Seismic
A2 Ai R2 1 R21 Exploration
A3 Ai R3 1 R21 1 R22
... Seismic wave transmission surveys make use of
seismic waves which are directly propagating
An Ai Rn 1 R21 1 R22 1 R23 . . . 1 R2n1 from sources to receivers. The normal direct
wave, which is always recognized on seismic
A general recursive formula can be derived by reflection and refraction-shooting records, is a
dividing the amplitude of the nth reflector by that typical transmission method used to calculate the
of the (n 1) reflector, giving: velocity of the surface layer. In exploration
seismology, there are several transmission-based
An An1 Rn Rn =Rn1 1 R2n1 ; n2 methods which are applied for specic explo-
ration purposes. Up-hole surveying, well velocity
Repeating, this formula is valid for (n 2), surveying, vertical seismic proling (VSP) fan
and for (n = 1), we have A1 = Ai R1. shooting, and seismic transmission tomography,
In conclusion the effective attenuation factor are typical examples of exploration practices
RTn (=An/Ai) due to reflection-transmission employing the seismic wave transmission phe-
process for a wave reflected from the nth nomenon. The up-hole surveying and that of
reflector of a multi-layer geological section, is well-velocity are dened here-below (Fig. 5.3).
92 5 Seismic Wave Transmission and Refraction
seismic
source detector
i r
M1 (V1 ) M1 (V1 ) M1 (V1 )
M2 (V2 ) M2 (V2 ) M2 (V2 )
rr i
i<r i>r
i=r=0
Fig. 5.4 Occurrence of refraction in case of oblique direction reversed. The velocities of the two media are V1
incidence. a Normal incidence, no refraction, b oblique and V2, where V1 < V2
incidence, from medium M1 to medium M2, c incidence
i i
medium1 velocity v1
medium2 velocity v2
r
refracted
wave
SV
P
P
rsv1
Fig. 5.7 An obliquely incident P-wave and the generated four wave phases: reflected and refracted P- and SV-waves
5.2 Seismic Wave Refraction 95
an interface. If medium-2, for instance, is liquid, incidence angle (i) increases, and when it reaches
no transmitted SV-wave is formed. a value of (i = ic), refraction angle reaches the
Using the symbols shown in Fig. 5.7, the value (r = 90). The angle of incidence (ic) for
generalized Snells law is quoted as follows which the angle of refraction (r) becomes (90) is
(Sheriff 1973, p. 200):
(sin ip1)/vp1 = (sin rp1)/vp1 = (sin rsv1)/vsv1 = (sin tp2)/vp2 = (sin tsv2)/vsv2 = k
incident P reflected P&SV refracted P&SV
5.2.2 The Critical Refraction and In this state, where the incident angle is equal
Head Wave Generation to the critical angle (ic), a special refracted wave
(called the head wave) is generated. This criti-
Consider a geological model made up of a med- cally refracted wave travels along the interface
ium (M1) of velocity (v1) overlying a medium with the velocity (v2) of the second medium
(M2) of higher-velocity (v2). An obliquely- (M2). This wave is refracted back to the earths
incident wave in the rst medium shall refract surface at the same angle (ic) and with the
on hitting the separating interface. In such a propagation velocity of the upper medium (v1).
model, the refraction angle (r) increases as the Development of head waves is shown in Fig. 5.8.
90
96 5 Seismic Wave Transmission and Refraction
Head waves have great practical importance the direct wave. The sourcereceiver distance
as they form the fundamental basis of seismic (0 to xcr), at which the refraction wave overtakes
refraction exploration. the direct wave, in the same layering set-up, is
called the (cross-over distance).
The velocities of the two involved layers (v1 and
5.2.3 Ray-Path Geometry
v2) are computable from the slopes of the travel-time
and Travel-Time Curves
lines (as seen from the gure above). The velocity is
the inverse of the slope of the time-distance line.
On seismic shooting records, three main wave-
arrivals are usually observed. These are the direct,
refracted, and reflected waves. Their travel-time 5.2.4 Refraction Travel-Time
functions are linear for the direct and refracted Function
arrivals, but hyperbolic for the reflected wave. At
the critical distance (xc), where the incidence is at A travel-time function of a moving wave
the critical angle (ic = sin1(v1/v2)), a head wave is expresses the mathematical relation between the
generated which is refracted to surface at an angle time covered by the moving wave and the
equal to the critical angle. The distance (0 to xc), no source-receiver distance. This is also depending
refraction arrivals exist, hence it is a (shadow zone) on the velocity values of the two involved media.
for refracted waves (Fig. 5.9). Two cases are presented here; these are cases of
For a given interface separating two constant- horizontal and dipping plane interfaces.
velocity media, the refraction travel-time function
is a straight line tangent to the reflection hyper- (i) Case of Horizontal Refractor
bola at the point where both of these waves arrive Consider rst a simple one-refracting horizontal
at the same time. The distance at which reflection interface separating two layers of velocities (V1
time equals the refraction time, which occurs at and V2) as shown in Fig. 5.10.
the critical angle, is called the critical distance. Referring to this gure, the refraction ray-path
Due to its shorter travel-path, the direct wave, consists of three segments, AB, BC, and CD. The
at small receiver distance, arrives earlier than the refraction travel time T(x) is given by
refraction wave. However, after some time, the AB/V1 + BC/V2 + CD/V1. That is:
faster refraction wave catches up and surpasses
Tx 2z=V1 cos ic x2z tan ic =V2
reflection curve or, (using sin ic = V1/V2 from Snells law and the
Tx refraction curve sine-cosine relationship)
Tx x sin ic 2z cos ic =V1
direct curve
or,
1=2
T0 Tx x=V2 2z V22 V21 =V1 V2
x
0 xc xcr where (ic) is the critical angle, and (z) is depth of
the interface. The two velocities, (V1 and V2) are
velocities of the two layers. These are equal to
z ic ic the reciprocals of the slopes of the direct and
M1, v1
refracted travel-time curves respectively.
head wave M2, v2 It is readily noted that refraction travel-time
function is linear in (x), and its curve is straight
Fig. 5.9 Ray-paths and travel-time curves of direct,
critically-refracted, and reflected waves, with critical line of slope equal to (1/V2) and its intercept (T0)
distance (0 to xc) and cross-over distance (0 to xcr) is given by:
5.2 Seismic Wave Refraction 97
B ic z2
C V1
V2
1=2
T0 2z V22 V21 =V1 V2 : z 1=2xcr V2 V1 =V2 V1 1=2
Hence, for down-dip shooting, the refraction apparent angle of dip in the direction of the
travel-time function is given by: shooting-line (x-axis direction).
In case of one horizontal refractor, the sufciently accurate; provided that the
refraction travel-time function, is given by: refractor-dip is less than about 10 (dip < 10).
h i1=2
Ta x=V2 2z V2 2 V1 2 =V1 V2
5.2.6 Refraction in a Multi-layer
Medium
That is:
dTS dTR T0 =2
slope = 1/V1
surface
Z1
V1
Z2 V2 > V1
Z3
ic V3 > V2 ic
V4 > V3
function of depth, V(z) = a + bz, (a and b are consist of two parallel but displaced linear seg-
constants), then the travel-time function is given ments of slope equal to the inverse of the velocity of
by Nettleton (1940, pp. 257261) and Dobrin the faulted layer (slope = 1/V2). These two seg-
(1960, pp. 7778): ments correspond to the refracted arrivals from the
reflectors original surface and from the down-
Tx 2=bsinh1 bx=2a thrown surface respectively.
In reference to Fig. 5.15, the fault throw (Z)
The ray-path, in this case, is of circular shape of is calculated from the difference in intercept
radius (r = zm + a/b), where (zm) is the maximum times (T), and the angle (ic) which is equal to
depth of penetration which is given by: cos1(P/Z). Since P = T V1, and (sin
h i1=2 ic = V1/V2), we can write:
zm x=22 a=b2 a=b h i1=2
DZ DT V1 V2 = V2 2 V1 2
Geometrical shape of ray-paths and travel-time
curve are shown in Fig. 5.14. For more accurate determination of the throw
(Z) and fault location, another reversed refrac-
tion survey is needed to be conducted.
5.2.8 Refraction from a Faulted
Refractor
5.2.9 Applications of Seismic Wave
Refraction surveying can be used to detect a fault Refraction
and determine its throw. Consider a two-layer
system where the high-velocity layer (velocity, V2) Refraction eld work involves laying out a
is vertically faulted (fault throw, Z) as shown in spread made up of sources and detector, that suite
Fig. 5.15. The travel-time curve of refraction recording both of the direct and refracted
experiment across the strike of the fault plane will wave-arrivals. Because of the relatively long
5.2 Seismic Wave Refraction 101
distance
0 x
c1 c2 c3 line of centres
r1 a/b
earth surface
zm x
0
surface
Z1
ic Z2
fault throw, Z i P
c V1
V2
travel-path, refraction arrivals suffer from atten- (ii) Small-scale refraction surveys
uation in both of the energy and high frequency (iii) Seismic fan shooting
contents. To lessen these adverse effects, an (iv) Seismic tomography
adequate increase in energy source and adequate
(i) Large-Scale Refraction Proling
detector frequency response are used.
Large-scale refraction can be used in exploring
The main elds of application of refraction
geological rock layering of regional dimensions.
waves are:
Source-receiver spreads can be employed to
(i) Large-scale refraction proling
survey lines of several kilometers to several tens
102 5 Seismic Wave Transmission and Refraction
A
source
points
B
5.2 Seismic Wave Refraction 103
source-point
receiver-point
against source-receiver distances. For more the surface. Like any other inversion process,
accurate mapping of the anomalous body more seismic tomography aims at determination, as
than one source point is used for the same accurately as possible, of the earth structural
receiver array. This is similar to the scheme used model for the area under study.
by the modern tomography procedure followed Seismic transmission tomography deals with
in exploring anomalies by source-receiver seismic transmission waves which, in general,
recording systems. More detailed description of experience refraction processes at boundaries
the method is presented in the geophysical liter- where velocity changes. There are two types of
ature like Nettleton (1940), Bth (1971) and seismic tomography, reflection and transmission.
Kearey et al. (2002, p. 116). Reflection tomography deals with reflection
travel-time measurements, while transmission
(iv) Seismic Tomography
tomography deals with source-to-receiver
The word (tomography) is of Greek-language
ray-paths along which the seismic waves are
origin, meaning (cross-section drawing). In the
directly transmitted. The more commonly
seismic eld, tomography found its way as an
applied technique is the transmission,
effective exploration tool under the name of
borehole-to-borehole (also called cross-hole)
Seismic Tomography. This technique has been
technique. In this method (cross-hole method), a
applied successfully in imaging seismic velocity
set of source points are distributed down a
variation and producing subsurface geological
borehole and another set of receivers are located
models. Seismic tomography is considered to be
in the other borehole, as shown in Fig. 5.17.
one of the modern techniques used in seismic
Another form of the seismic tomography is
exploration.
the surface-to-borehole technique. In this method
Seismic tomography is a type of inverse
the source points are located on the surface and
modeling (inversion process) which uses
the receiver points are distributed down the
acquired seismic data to generate earth models
borehole, as shown in Fig. 5.18.
(usually in terms of velocity and dimension
Interpretation of the data recorded from the
information). The more applied technique in this
cross-hole surveying (or from that of the
eld is the (transmission tomography) which
surface-borehole) aims at velocity distribution
requires putting the source points in a borehole
in the medium between the source and receiver
and the receiver points at another borehole or at
104 5 Seismic Wave Transmission and Refraction
source-point
receiver-point
sets. Iterative, forward modeling is usually cess, until computed travel-times become
applied with the help of an especially-designed closest to the measured travel-times. The end
tomography-algorithm, starting with an objective of the whole process is to determine
assumed model of velocity distribution. The the subsurface geological model as accurately
model is changed after each comparison pro- as possible.
2D Seismic Reflection Surveying
6
The seismic reflection method depends on the produces a data volume in which the seismic
same principle of the well-known echo phe- amplitude is function of the variables (x, y, & z).
nomenon. A mechanical shock generates a seis- Reflected waves carry useful exploration
mic wave (normally, P-wave) which propagates information in two forms: the travel-time and
through the earth material. On hitting a surface of waveform changes. Both of these two types of
a rock layer (an interface), the incident changes are functions of the physical properties
wave-energy is partly transmitted (passing and layering geometry of the subsurface geology.
through the interface), and partly reflected back In other words, the seismic information furnished
to the earth free surface. by reflection surveying, can supply two main types
The earliest successful application of seismic of exploration knowledge which are: structural
reflection technique was made about 1930 (Net- exploration information from travel-time mea-
tleton 1940, p. 233). Since then, and due to its surements, and stratigraphic exploration informa-
great efciency in the exploration work, the tion from wave-form measurements. These two
method was intensively applied in exploring types of changes constitute the principal seismic
subsurface oil-traps. In its conventional form, the tools for exploring the subsurface structural and
technique is based on initiating of a seismic wave stratigraphic geology (Fig. 6.1).
and recording its reflection arrivals by detectors
deployed on the earth surface along a straight
line. The normally applied technique, the seismic 6.1 Reflection Surveying Concepts
reflection surveying, is based upon the use of a
linear source-receiver spread moving along a 6.1.1 The Spread Configuration
dened seismic line. To cover a complete survey
of an area, the process is repeated for all lines of The shooting spread used in seismic surveying is
a dened line-network designed for that area. dened as the geometrical relationship between
The end result of this technique is obtaining a the location of the source-point and that of the
two-dimensional seismic section for each seismic receiver-array. This shot-receiver conguration is
line, which is, in effect, expressing the variation kept unchanged during surveying of the seismic
of seismic-reflection amplitude as function of the line. The main elements of a spread are the
two variables: distance, x and depth, z. The source-point (commonly known as the
section lies in the vertical plane that coincides shot-point), and number and inter-spacing of the
with the surface survey-line. For this reason, the receiver-points. In practice shot-points and
method is called two-dimensional 2D surveying receiver-points are not points but sets of points.
to differentiate it from the 3D surveying which Thus a receiver point is usually made up of a
STRUCTURAL STRATIGRAPHIC
EXPLORATION EXPLORATION
group of detectors (geophones) and the shot-point multiple of the number (12). In 2D seismic sur-
consists of a group of source-points. That is why veying the number of spread-channels used has
these are commonly referred to as geophone increased from 48 in the early 1970s to 240; and
group and shot pattern. For computation pur- 360 channels in the following forty years. In 3D
poses, their geometrical centres are used to rep- surveying the number may exceed 1200.
resent the shot-point and receiver-points locations
respectively. The shape of the shooting spread and
6.1.2 The Seismic Trace
its main elements are shown in Fig. 6.2.
Based on the position of the shot-point in
The seismic trace, considered to be the building
relation to the receiver-point, there are three types
brick for the whole seismic survey data, is a graph
of spread conguration in common use. The
representing the recorded amplitude variation as
shot-point may be located at one end of the
function of reflection travel-time. The trace con-
receiver linear array giving a type called (end-on
tains all the outputs of an active seismic channel
spread). Another type of spread is the one in which
during the recording time. It is, in fact, a record of
the shot-point is located within the spread it is
the whole sequence of the reflection arrivals as well
called (split- or unbalanced-spread), and when it is
as other arrivals such as direct waves, refracted
located exactly at the centre of the receiver
head waves in addition to the various types of
array-line, it is called (centre spread). The three
seismic and none-seismic noises. There are three
common spread types are shown in Fig. 6.3.
common modes of trace-display. These are:
The number of the active receiver points (or
seismic channels) per spread is customarily xed (i) Wiggly-trace, when the amplitude variation
at a whole-number which is customarily taken as is represented by a wiggly line recorded.
Shot-point
(i) end-on spread
Shot-point
(ii) unbalanced spread
Shot-point
(iii) centre spread
(ii) Variable-area trace, when the peak-parts reflection point (RP) or common depth point
of the wiggly trace are blacked in. (CDP). Referring to Fig. 6.5, a seven-channel
(iii) Variable-density trace, when the ampli- spread (end-on type) produced a shot record
tude variation is represented by an appro- containing seven traces that have been reflected
priate shade-intensity. from corresponding seven CDPs.
An actual record of a 48-trace shot-point
Usually, the eld shot-records are displayed in
obtained from a centre-spread is given here-below
wiggly-trace mode (Fig. 6.4).
Fig. 6.6. It is showing fairly clearly, reflection
Seismic stack sections are displayed by super-
arrivals as well as direct and refracted arrivals.
imposing the wiggly- and variable-area modes of
display.
6.1.4 The Seismic Profiling
6.1.3 The Shotpoint Technique
A seismic survey of a certain seismic line starts Seismic surveying along a linear track is carried
with deployment of the spread elements (shot out by a technique known as (multi-channel
and receivers) along that line. The designed proling technique). The survey procedure
spread, which consists of linear array of involves the use of a xed-shape spread that
receiver-points which are co-linear with the moves along a linear course at a regular move-up
shot-point, is laid down along the line to be shift. The rst shot-record is obtained from the
surveyed. When activating the energy source (as spread which is laid out at the start-point of the
ring a dynamite charge), a seismic wave-front line. In the following step, the second shot is
advances through the medium in the form of recorded after shifting the spread location by a
spherical surfaces (spherical when the medium is certain step-up distance. This process
homogenous and isotropic). The seismic rays are (shift-and-shoot process) is repeated from start
reflected from intervening interfaces, back to the point to the end point of the line. The move-up
surface to be detected by the geophone groups distance is normally made to be integral multiple
planted on the surface. The reflection arrivals are of the receiver-point spacing. A map of all
then fed into the recording system via the elec- locations that have been occupied in surveying
trical channels which are displayed in a the entire line is called (surface-coverage map).
trace-gather (the shot-gather). For more clarication of the proling tech-
Each recorded trace (provided by an active nique, an example is given in Fig. 6.7 which
channel of the spread) represents a series of wave shows a seismic line of length of 39 receiver
arrivals of signals reflected from a subsurface stations, surveyed by an end-on spread. The
108 6 2D Seismic Reflection Surveying
Fig. 6.4 Part of 48-channel shot-record showing reflection arrivals and random seismic noise. Traces are displayed in
wiggly-trace mode
surface coverage, shown in Fig. 6.7, consists of midpoint. This means that each receiver-point
seven positions of the spread that advanced along of a spread will receive a seismic arrival reflected
the line at move-up shifts of four from a CDP located below the source-receiver
receiver-stations per shift. midpoint. In fact, shooting of one spread, the
Because this procedure gives a seismic section reflector will be sampled by a number of reflec-
representing a prole of the earth along the sur- tion points (CDPs) equal to the number of the
veyed seismic line, it is referred to as a proling spread active receivers. From the geometry of the
technique. spread ray-path, the CDP spacing will be half
that of the receiver points of the spread. Thus, the
subsurface coverage (CDP-line) covered from
6.1.5 The Fold of Coverage shooting one shot will be about half the spread
length (Fig. 6.8).
By denition the CDP is a subsurface point Seismic proling procedure is accomplished
located vertically below the source-receiver by repeating the shooting process (shot ring and
6.1 Reflection Surveying Concepts 109
R1 R2 R3 R4 R5 R6 R7
S
Reflector-1
CDP1 CDP3 CDP5 CDP7
TX
Reflector-1
Reflector-2
Reflector-3
reflection record
Fig. 6.5 Shot-point ray-paths and its corresponding seismic shot record. Spread (end-on type), made up of 7
receiver-points (R1, , R7) producing 7 seismic traces (TR1, , TR7) reflected from 7 CDPs (CDP1, , CDP7)
reflection-arrivals recording) at a series of equally it is shown that when the spread move-up is too
spaced shot-point locations along the surface line. large (larger than half the spread), the subsurface
By making the spread move-up distance equal to coverage is incomplete, with gaps in the CDP
half of the spread length, one reflection per one sequence where patches of the reflector are giv-
CDP will result for the entire line. The resulting ing no-reflection information Fig. 6.9a. When the
seismic section in this case is commonly known 12-channel spread is made to advance by
as (single-fold section). For larger move-ups, 6-receiver station move-up rate, the reflector is
gaps will result in the sampled reflector. That is, completely covered with single-fold coverage
certain CDPs will not have reflections at all. On (Fig. 6.9b). In the third case (Fig. 6.9c), the
the other hand, if the move-up distance is smaller spread is made to move with a spread move-up
than half the spread length, then CDPs will get distance of two receiver-stations, which resulted
more than one reflection, giving the case of in threefold coverage for the whole line except
multi-fold proling technique. The number of for the less-than threefold zones (called fold tails)
reflections realized per CDP is called the fold of at the beginning and at the end of the line. In this
coverage, (or just, fold). Of course the shorter the example, the fold, in the tail-zone, is building up
move-up, the greater the fold will be. from a single fold at the rst CDP to threefold
Considering an example of a seismic proling over 8 CDPs. The same thing occurs at the end of
case in which a twelve-channel spread is used. the line but in reverse direction.
The effect of the move-up distance on the fold of In a regular shooting, which is normally
coverage is claried in Fig. 6.9. In this example, applied in seismic proling, the fold of
110 6 2D Seismic Reflection Surveying
Fig. 6.6 An actual 48-trace shot record showing clear reflection arrivals with relatively high-amplitude direct and
refracted arrivals. Traces are displayed in variable-area mode of display
coverage (F) can be calculated from the fol- there is a direct relationship between fold and
lowing formula: reflection signal-quality appearing in the stack
trace. However the relation is not linear but fol-
F N=2n lows a power law of the form (F1/2) (Robinson
1983a, p. 111). Accordingly, the rate of change
where, (N) is the total number of the spread in the (S/N) enhancement diminishes with the
active receiver-channels and (n) is the move-up increase of fold (Fig. 6.10).
distance expressed in number of receiver-station
spacing covered in one move-up distance.
This equation shows that the fold is directly 6.1.6 The CDP Trace Gather
proportional to the number of channels and
inversely proportional with the move-up distance. As mentioned above, a shot-gather is the group of
The fold of coverage is a major factor in traces recorded by the spread channels of that shot.
increasing the signal-to-noise ratio (S/N). In fact Thus, a seven-channel shot has seven recorded
6.1 Reflection Surveying Concepts 111
receiver stations
1 10 20 30
source receivers
S1
S2
S3
S4
S5
S6
S7
Fig. 6.7 Surface coverage of a line made up of 7 spread move-up shifts
source
receivers
Fig. 6.8 Shooting spread surface-coverage, reflection ray-path, and CDP subsurface coverage, as realized from one
seismic shot-point
seismic traces. Likewise, the CDP-gather is a this regrouping process (trace re-sorting), a simple
group of seismic traces which belong to one proling example is given here (Fig. 6.11).
CDP. Naturally, the number of traces in a certain In this example (shown in Fig. 6.11) we have
CDP is equal to the fold of coverage in that CDP. a survey made up of ve shots (A, B, C, D, and
In a certain seismic proling process, the traces E). Shot-A of traces (a1, a2, , a6) and shot-B of
obtained for each shot (shot-gather), are re-grouped traces (b1, b2, , b6), and so on, for the rest of
into CDP gathers ready for being stacked after shots. The complete surface coverage is occu-
being subjected to certain processing steps which pying a total of 11 receiver-stations with one
aim at enhancing the reflection signal. To explain receiver-station for the spread move-up distance.
112 6 2D Seismic Reflection Surveying
(a)
move-up distance-1 move-up distance-2 move-up distance-3
S1 S2 S3 S4
(c)
Move-up distance-1
Move-up distance-2
Move-up distance-3
Move-up distance-4
S1 surface coverage
S2
S3
S4
sub surface coverage
CDPs
Fig. 6.9 Surface and subsurface coverage in seismic Spread move distance is 2 receiver-stations, resulting in
proling surveying using an end-on spread. Move-up nominal fold of 3, with tail-fold at beginning and end of
distance: greater than half spread length (a), equal to half the line
spread length (b), and more than half spread length (c).
Corresponding to this surface coverage, is a known the (Coverage Diagram). In Fig. 6.11, the
subsurface coverage made up of 14 live CDPs, shot-gather is indicated by the horizontal arrow, and
(CDP1, CDP2, , CDP14). The trace gather for the CDP-gather is indicated by the vertical arrow.
CDP1 (CDPG-1) consists of one trace (a1). For Number of traces in each gather, represents the fold
the second CDP (CDPG-2) the gather has also of coverage. In this example, the fold is equal to 3
one trace (a2). The CDPG-3 has two traces for the range (CDP5CDP10), and tail fold for the
(a3 + b1), and so on. The number of traces per ranges (CDP1CDP4), and for (CDP11CDP14).
CDP is representing the fold of coverage. The re-distribution of the traces from the
Identifying the trace gathers (for shots and shot-gathers to CDP-gathers (called CDP sorting) is
CDPs) is much simplied by drawing what is shown in the following gure (Fig. 6.12).
6.2 Seismic Reflection Data Acquisition 113
Fig. 6.10 Curve expressing type of dependence of the 6.2.1 Seismic Energy Sources
S/N ratio on fold of coverage of reflection signal. The
curve is of the power-law form (F1/2)
The traditional method applied in generating
seismic waves is exploding dynamite in shot-
holes. There are however, other methods which
6.2 Seismic Reflection Data have been introduced as alternative seismic energy
Acquisition sources. Choice of the source-type depends on the
surface conditions prevailing in the given survey
In seismic reflection data acquisition, we are, in
area. The main criteria considered in evaluating a
general, dealing with generating and detecting
particular source-type are the following:
P-waves which are propagating from the source
region and getting reflected from subsurface strong enough to generate strong seismic
interfaces. The complete data acquisition process signal that can be reflected from deep inter-
involves seismic source generation, reflected face and be detectable at the farthest receiver
wave detection, and digitally data recording. (receiver of maximum offset).
In order to record the reflection arrivals with the generated seismic signal is rich in high
minimum distortions, optimum operation parame- frequency components to be able to resolve
ters (called eld parameters) are determined and closely spaced reflectors.
applied in conducting the seismic survey. The the generated noise is of least energy level.
SHOTS receiver-stations
& 1 2 3 4 5 6 7 8 9 10 11
CDPS
FOLD 1 1 2 2 3 3 3 3 3 3 2 2 1 1
Fig. 6.11 Example of proling process, where a horizontal arrow and the CDP-gather is indicated by the
6-channel end-on spread moving at one receiver-station vertical arrow. Nominal fold is 3 over the central 6 CDPs
move-up distance. Shot-gather is indicated by the and the tail-fold zones are at beginning and end of line
114 6 2D Seismic Reflection Surveying
LVL
weathering zone
dynamite charge
seismic rays
surface (low velocity layer, LVL), the high fre- the bottom of a heavy vehicle to increase coupling
quency components are severely attenuated. This of its plate with ground surface (Fig. 6.15).
method is rarely applied at present. On detonating the gas mixture, a sudden
pressure impact occurs which is transmitted
(iii) Gas Exploding
through the base-plate to the ground surface. By
This is another impulsive source (known as
hydraulic system the plate is locked into position
Dinoseis) in which the energy is created by
after the impact, to prevent repeated impacts.
exploding a mixture of oxygen and propane gases
Being a relatively weak-energy source, more
contained in a conned chamber, the bottom of
than three units are red simultaneously by a
which is a moveable plate resting on the ground
control signal sent from the recording system,
surface. The so designed chamber is attached to
and the shot is repeated many times, then vertical
stacking is applied to get improved S/N ratio.
The general features of Dinoseis are similar to
the Weight Drop method. Both are surface
sources which are generating relatively weak
falling mass seismic energy, developing strong surface waves,
and producing low-frequency seismic pulses.
The method is rarely applied these days.
Fig. 6.14 Weight-Dropping seismic-energy source
gas chamber
116 6 2D Seismic Reflection Surveying
(iv) Dynamite Cords Exploding in laying the explosive cords. Since the method
This seismic source, well known by the name depends on trenches for the explosive cords, it
(Geoflex), consists of an explosive cord which is means that it is not possible to be used in areas of
buried in the ground at a shallow depth (about too hard surface layer or in cities or villages. It is
half a meter depth). It is laid down by a not advisable to be used in areas where the sur-
hydraulically-operated plough which is espe- face layer is too thick, since in such conditions the
cially designed for this purpose (Fig. 6.16). transmitted energy will be much weakened.
Implementation of the Geoflex method starts
(v) Vibroseis
with laying the explosive cord (about 100 m
This is a non-impulsive seismic-energy source
length) inside a trench of about a half-meter
which was introduced in early 1950s and rapidly
depth and detonated at one end. While the
gained popularity as an acquisition tool. In 1982,
explosive-generated energy is travelling very fast
over 40 % of worldwide seismic surveys were
through the cord from the detonation point,
using it (McQuillin et al. 1984, p. 38). Unlike
seismic waves are generated from each point of
impulsive sources, Vibroseis method creates
the cord at delayed times. This will cause these
mechanical energy which is continuously vibrat-
waves to have phase differences bringing about
ing for certain time duration. A mechanical, truck-
the ltering effect similar to that made by
mounted vibrator is hydraulically driven, produce
source-arrays. Points of the earth surface coin-
electronically controlled vibration. The generated
ciding with the cord, are receiving the liberated
energy is conveyed to the ground by a metal pad
energy not at the same time, but in sequencial
(about 1 m2) attached beneath the truck
manner. For this reason, the down-going
(Fig. 6.17).
wave-front becomes inclined with respect to the
When in operation, the Vibroseis system
horizontal level. Along this wave-front (shown
transmits into the earth a seismic signal vibrating
by dotted lines in Fig. 6.16), highest energy level
at frequency which is varying linearly with time.
is obtained. In effect, there will be a ltering
This electronically-controlled vibration function
effect such that the near-surface horizontally
is called the Sweep. In practical application, the
travelling energy (surface waves) will be ltered
sweep duration (called sweep length) is normally
out before reaching detectors.
within the range (1020) s. Over this time-span
The main advantages of the Geoflex method, is
the sweep-frequency applied is (1050) Hz.
the attenuation of the surface waves, in addition
Apart from the taper imposed at both ends of the
to generating wide-band seismic waves. Regard-
sweep (about half a second in length), the
ing eld operation the method is fast and simple
explosive cord
Detonation
point wave-fronts
of
generated
Seismic waves
6.2 Seismic Reflection Data Acquisition 117
vibrating
metal pad
amplitude
time
(c)
taper zone
amplitude
time
118 6 2D Seismic Reflection Surveying
In order to extract the impulsive-like record of is the (Unipulse), where the air continues
the reflections in their actual reflection-times, the flowing into the generated bubble for some
recorded traces are cross-correlated with the time after the initial discharge. This is done
applied sweep. This process can be done in the to lessen the effect of the sudden bubble
eld within the recording unit of the Vibroseis collapse which results in bubble oscillation.
system, or later by the processing software (see (ii) The Steam-Gun (also called Vaporchoc)
processing section). utilizes hot steam instead of air. It is
The main advantages of the Vibroseis method considered to be better than the Air-Gun,
are being fast, safe, and comparatively cheap to because with using hot steam, the effect of
run. It can be applied along roads and even in cities bubble oscillation is reduced.
since it causes no damaging effects on environ- (iii) The Gas-Gun is still another similar marine
ments. Technically, it has the advantage of having source which uses explosion of a mixture of
the source-function being under control. The propane and oxygen contained in a steel
sweep parameters (time duration, frequency range, chamber. The pressure pulse created by the
and taper) can be changed at will. The main explosion of the propane-oxygen mixture is
problem with the method is the low source passed to the water via a flexible seal which
energy-level. Like other weak surface sources, forms part of the chamber. The (Aquapulse)
vertical stacking is carried out at the same time as works on this principle.
the recording is going on. In normal surveying (iv) Dynamite explosion is implemented as a
work, several Vibroseis-trucks (typically 4) are marine energy source. A method called
operating at the same time, and about 2060 (Maxipulse) uses small dynamite charges
vibrations are conducted per each shot-point red directly in the water body. Another
location. The shot-gather traces per each shot, similar method (Aquaseis) uses long det-
are obtained from vertical stacking of all of the onating cord.
recorded traces involved in the one shot-point. (v) Electrical Arc Discharge is used to gen-
erate energy by electrical discharge in
Marine Seismic Sources water. Examples of this method are the
The most common types of seismic energy (Sparker) and the Wasp. This is a rela-
sources are: tively weak energy source which is more
(i) The Air-Gun is a widely used energy source suited for shallow-geology exploration.
in marine seismic surveying. It generates Out of these methods, the Air-Gun is by far the
energy by discharging highly compressed most commonly used type of energy source in
air into the water. A variation of this method
6.2 Seismic Reflection Data Acquisition 119
seismic marine surveying. In certain situations, the land and hydrophones for use in marine
Air Gun, is used as a land seismic source by gen- environments.
erating a seismic pulse from an Air-Gun sub-
(i) The Geophone
merged in a water-lled pit made on the earth
Detection of reflection arrivals in land surveying
surface. Typical example of such an application is
is done by geophones (seismometers as they are
in providing the seismic energy for VSP surveying.
sometimes called). The geophone is a
It should be noted here that the names of the
seismic-detection instrument which can trans-
energy types are trademarks given to them by
form the ground vibration motion into an elec-
their originators. The following Table 6.1 con-
trical voltage. The most commonly applied
tains a summary of the types of seismic sources,
geophones in seismic reflection surveys are the
with their trademark references.
electromagnetic type. It operates on the principle
of voltage generation in a coil moving within a
6.2.2 Seismic Detectors magnetic eld. The generated voltage is propor-
tional to the velocity of the motion of the coil
Earth movement due to the arrival of a seismic relative to the magnet. For this reason, geo-
wave is generally very small. The vibration phones of this type are normally referred to as
amplitude is of the order of 108 of an inch (velocity geophones).
(Gardner 1938). In addition to the The geophone is made up of a permanent
small-amplitude feature, the seismic waves have cylindrical magnet and a coil suspended by
two other criteria which the detection and leaf-springs inside a circular slit made in the mag-
recording equipments should be able to cope net. The slit is separating the magnetic south-pole
with. These are the wide amplitude-range and the (inner part of the magnet) from the north-pole
wide frequency-range. (outer part). The magnet is rmly attached to a case
The seismic detection process is based on which is tted with a spike for easy xing on the
conversion of the ground vibration to an electri- ground in an upright position (Fig. 6.20).
cal signal by a special transducer which can For operation, the geophone is planted verti-
respond to seismic amplitude and frequency full cally into the ground. In this position the magnet
ranges and without distortions. Seismic detectors will vibrate in vertical direction when the ground
are of two main types, the geophones for use in is seismically activated, and the coil stays
Fig. 6.20 Schematic damping maximum damping applied without stopping the
representation of the resistor oscillation. Damping of about 0.7 times the
electromagnetic geo-
phone. Symbols N and critical damping gives a response which increa-
S denote magnetic north N S N ses smoothly with increase of frequency. This
pole and south pole damping factor (D = 0.7, in Fig. 6.21) is the
respectively damping which is normally applied in seismic
surveying environments.
(ii) The Hydrophone
This is a geophone-equivalent detector used in
magnet detecting seismic waves in marine surveying
environments. The hydrophone is a
pressure-sensitive detection device that uses
spike
substances that generate electrical voltage pro-
portional to pressure changes caused by the arri-
val of a seismic signal. Such a substance (called
piezoelectric substance) has a property of gener-
ating an electric voltage when subjected to pres-
sure. Piezoelectric transducers are also called
stationary because of its inertia. The generated electrostrictive devices (Sheriff 2002, p. 263). The
voltage in the coil is function of the vibration rate pressure changes, caused in a water medium, due
and the coil parameters (number of turns, radius, to passage of a seismic wave, are proportional to
and eld intensity of its magnet). the velocity of the water particles set into motion
In order to reduce oscillation, geophones are by the signal (Dobrin and Savit 1988, p. 63).
normally provided with a damping resistor In marine surveying, hydrophones are housed
shunted across the geophone terminals. The in a special hose lled with a certain liquid of
damping device controls the frequency-response such a density that makes the net density of the
of the geophone vibration. Typical forms of the hose (containing the hydrophones and liquid)
geophone response characteristic curves for approximately equal to that of water in order to
different damping factors, are shown in facilitate controlling its depth below water sur-
Fig. 6.21. face and in reducing effects of external sudden
Too heavy damping (over-damping) reduces forces affecting the streamer (cable jerking).
detection sensitivity, while too little damping Marine data acquisition is conducted by tow-
(under-damping) will lead to continued oscilla- ing the hose, equipped with hydrophones (called
tion. Critical damping is dened to be the streamer) behind the survey boat (Fig. 6.22).
frequency
Natural
frequency
6.2 Seismic Reflection Data Acquisition 121
energy source
6.2.3 The Seismic Data Recording signal. With this unit the too-small amplitude is
increased and the too-large amplitude is reduced.
For about 30 years after seismic exploration was This operation results in compressing the signal
introduced, seismic shot-records were directly amplitude-range to that value which is matching
recorded on paper as wiggly traces. In the early the dynamic range of the recording unit. The
1950s, however, recording of seismic data on technique of amplitude matching with the
magnetic tapes was introduced in analogue recorder, for undistorted output, is an essential
mode. After about ten years, the more superior feature of all seismic recording systems. A mod-
digital recording technique was introduced. Soon ication of the AGC is the programmed gain
after mid 1965, the analogue method was com- control (PGC) with which the gain is not tied up
pletely superseded by the digital method. Despite to the input amplitude-variation. A PGC-coupled
the fact that analogue recording is now consid- amplier has its gain that varies with record time
ered to be obsolete, (that is, no more in appli- by a preset (programmed) time-dependant func-
cation), a brief description of the analogue tion. This is usually made available as an option,
system is presented here-below. This will be applied when it is required.
useful as the digital system involves some of the Next to the amplication part is the magnetic
components of the analogue system. recording system which produces a permanent
record of the data stored on magnetic tapes. This
(i) The Analogue Recording System
is accomplished either directly or after being
The analogue recording system carries out the
frequency-modulated (FM). A carrier frequency
recording in two stages: the amplication process
(typically, 4 kHz) is modulated by the signal
followed by the recording process. In the rst
received from the amplication system. In this
stage, the seismic signal, received from the
way it can produce a recorded signal of S/N ratio
geophone group, is amplied so that it will be at
up to 60 db, and frequency response which is
a level that suits the following recording stage.
practically flat over 3200 Hz (Evenden et al.
The seismic signal may be too small for the
1971).
recorder to detect, or too large for the recording
Through a playback of the recorded magnetic
unit to handle without being saturated. In order to
tape, or directly from the amplier output, a
cope with this wide range of amplitude variation
visual display may be obtained. The display unit
(estimated to be 80100 db) amplication factor
(camera unit) is usually included in the recording
is made to be variable. This is done by using a
system. This unit uses the principle of gal-
special electronic unit called the automatic gain
vanometer deflection caused by voltage variation.
control (AGC), which makes the amplier gain
By certain optical arrangement, galvanometer
to vary with the amplitude level of the incoming
122 6 2D Seismic Reflection Surveying
magnetic tape
AGC camera recorder
geophone
group
AMP FIL
AGC
AMP FIL GC
D-A
Fig. 6.24 Schematic diagram showing the signal flow in a digital recording system
d5 d4 d3 d2 d1
D
c5 c4 c3 c2 c1
C
b5 b4 b3 b2 b1 .. d3 c3 b3 a3 d2 c2 b2 a2 d1 c1 b1 a1
B Sample-1 Sample-2 Sample-1
a5 a4 a3 a2 a1
A
multiplexing
Fig. 6.25 Principle of the multiplexer operation, trace-sequential input with sample-sequential output, using 4
channels (A, B, C, D)
To conceive the multiplexing operation, let us d1) in this sequence. In the second rotation cycle
consider a mechanical device in the form of an arm the outputted amplitude-values (samples) will be
rotating with constant rotation speed where the (a2, b2, c2, and d2). With continued rotation, the
arm head will trace a circular ring. Referring to output through a single channel will continue
Fig. 6.25, consider four channels, arranged feeding the A-D conversion unit, until the end of
equally spaced on that ring. In the rst rotation the recording time. In this way, the seismic data
cycle, the arm shall make contacts with the chan- has been converted from trace-sequential
nels (A, B, C, and D) in that order. In this cycle, the amplitude-values to sample-sequential mode, as
multiplexer will output amplitudes, (a1, b1, c1, and it is shown in this gure (Fig. 6.25).
124 6 2D Seismic Reflection Surveying
twenty years ago, the tapes in common use were 21 When the cell is magnetized it represents (1) and
track-one inch tapes or 9 track-half inch tapes. In when it is not, then it is (0). This concept is claried
that recording system the rst track is usually from the following example (Fig. 6.26).
assigned for the parity check. At present narrower An analogue signal, shown at the top of
tapes (e.g. 4 mm tapes) and higher recording Fig. (6.26), is digitized and its sample values are
density tapes (in forms of small cartridges) were measured and converted into 8-bit binary words.
introduced for the seismic data recording. The eight bits of each sample (each binary word)
are recorded across the tape by magnetizing the
(iii) Concept of the Dynamic Range
cell when the bit is (one) and demagnetizing the
Any measuring or sensing instrument (geophone,
cell when it is (zero). The sign bit (marked as
amplier, recording system) has a bounded
S-bit) is a flag indicating the algebraic sign of the
capability of faithful detection. The system
sample value. It is assigned the value of (one) for
noise-level sets the detection lower limit, while
negative sign and (zero) for positive sign. For
the system technical characteristics set the
example, a sample value of minus nineteen
detection upper limit. For a given system, dy-
(19), the binary word in this example (9-track
namic range is dened to be the ratio (normally
tape) is (100010011), and for a value of (+27),
expressed in dbs) of maximum measured
say, will be (000011011), where the rst digit is
amplitude to the minimum recoverable signal,
for the sign (the sign bit).
where the minimum signal is taken to be the
In the case the recorded data is in multiplexed
noise amplitude-level.
form, the sequence followed will be in sample
The dynamic range (DR) is expressed by the
sequential, and if the data is in demultiplexed
ratio:
form, the recording will be in trace-sequential
DR A=adb mode. Normally, the magnetic tape is divided
into blocks where the rst block (called the
where, (A) is the maximum signal amplitude header block) carries general information like
which a system can measure (with tolerable record number, sampling period, record length,
distortion level) and (a) is the system noise-level. and other recording parameters. The next block is
The concept is often applied in evaluating the the seismic data which are usually in multiplexed
detection and measuring capabilities of recording form. Along with each sample value, the applied
systems. The early seismic analogue systems gain for that sample and sign-bit flag are recor-
which were recording directly on paper have ded. One track is reserved for the parity-check
dynamic range of about 25 db. The dynamic for checking the bit-column across the tape. The
range of analogue systems recording on magnetic end block contains information signaling the end
tapes is about 45 db. In case of digital recording of that set of recorded data.
systems, the dynamic range is dependent on the The Society of Exploration Geophysicists
number of bits used in expressing the sample (SEG) has developed several seismic data
value. For 14-bit recording, it is equal to 84 db. recording formats since the 1960s when the
digital recording system was introduced. In 1967,
(iv) The Magnetic Tape Recording the formats SEG-A, SEG-B, and SEG-X, were
The magnetic tape is a plastic strip, coated with published. Later on, other formats were pub-
magnetic material, is divided into tracks of cells. lished, as SEG-C in 1972, SEG-Y in 1975 and
After being digitized, each sample-value is repre- SEG-D in 1980. The formats of SEG-C and
sented by a binary word which is recorded on the SEG-Y are recording the data in de-multiplexed
magnetic tape by magnetization status of the cells. mode, whereas SEG-D can record in both
126 6 2D Seismic Reflection Surveying
analogue signal
(+)
0
(-)
sample values
6 9 22 69 66 15 0 -7 -4 30 32 11
0
2 0 1 0 1 0 1 0 1 0 0 0 1
21 1 0 1 0 1 1 0 1 0 1 0 1
22 1 0 1 1 0 1 0 1 1 1 0 0
3
2 0 1 0 0 0 1 0 0 0 1 0 1
24 0 0 1 0 0 0 0 0 0 1 0 0
5
2 0 0 0 0 0 0 0 0 0 0 1 0
6 0
2 0 0 0 1 1 0 0 0 0 0 0
2 7 0 0 0 0 0 0 0 0 0 0 0 0
S 0 0 0 0 0 0 0 1 1 0 0 0
digital magnetic tape
Fig. 6.26 Simplied sketch showing the principle of 9-track tape recording format. The binary word for each sample
value consists of 8 bits plus the sign bit (S)
multiplexed and de-multiplexed modes. At pre- (D-A) converter, demultiplexer, and the camera
sent, eld recording is mostly done using unit which displays the shot trace-gather (usually
de-multiplexed SEG-D format. in wiggly-mode traces) on the computer monitor
Developments in the tape-recording tech- or on paper. A typical shot record is presented in
niques concerned mainly the packing density and Fig. 6.27.
storage capacity. The density was developed This gure shows about one second of data of
from 800, 1600, then 6250 byte per inch. The a shot record made up of 48 seismic channels
tapes used nowadays are of 8 or 4 mm width, plus two auxiliary channels (at the left side of the
contained in cartridges or cassettes. record): the time break indicating the zero-time
of the shot, and up-hole-time. Traces are dis-
played in wiggly mode of display.
6.2.4 Data Playback and Display
For quality control (QC) purposes, the end result 6.3 Seismic Noise Characterization
of the recording station (the shot trace-gather) and Attenuation
needs to be displayed. The playback process is
the reverse of the recording process. It is done As we mentioned in a previous discussion, all
through a complete playback system which non-reflection arrivals are considered noise which
consists of a digital AGC, a digital-to-analogue cause distortions and masking effects to the
6.3 Seismic Noise Characterization and Attenuation 127
+ + +
receiver
output
receiver array
reflection wavelets
reflector
Fig. 6.28 Role of a multi-detector receiver in attenuating random (and coherent) noises and enhancing reflection signal
of a vertically arriving reflected wave
a group of shot holes of a shot-point pattern. source) is used. The array source-elements
Compared with a single-element receiver, a (shot-holes, for example) are detonated simulta-
multi-element receiver (receiver made up of neously generating body waves which reach the
many geophones), has the capability of attenu- detector array after being reflected from subsur-
ating horizontally moving waves in addition to face interfaces. At the same time, the
enhancing the signal-to-random noise ratio by a source-generated surface waves travel from the
factor depending on the square root of the source to the receiver. As it is with the receiver
number of the array elements. Efciency of the array, the nearly vertical incident body waves are
array in attenuating surface waves is expressed in phase and hence they interfere constructively
by a characteristic function called the array to give enhanced reflected signal. For attenuation
response, dened to be the ratio of the amplitude of the surface waves, the length of the source
of an array-output to that of the same number of array should be equal to the wavelength of the
the array-elements gathered together at one generated surface waves.
location (Sheriff and Geldart 1995, p. 247). This The response of any given array (having given
is expressed in the form of a mathematical number of elements with given element spacing)
function called (directional response function, or can be presented in the form of a curve known as
just directivity function). (response curve or directivity curve). The
The theory, on which the concept is based, is abscissa is the ratio of the wavelength of surface
the same whether it is applied to the shot waves (coherent noise) to element spacing, and
hole-pattern or to the receiver geophone array. the ordinate is the array output expressed as the
Instead of using a single source point, a source ratio of the array-output to that of the same
array (shot-hole pattern, in case of dynamite number of elements gathered together at one
6.3 Seismic Noise Characterization and Attenuation 129
location. The array response is normally emphasized that (k) is representing the apparent
expressed in db units. wavelength measured in the array direction.
The basic principle is that waves travelling in Typical response curves (directivity curves) of
near vertical direction are enhanced while those geophone linear-arrays are shown in Fig. 6.29.
travelling horizontally are attenuated. This tech- The characteristics of the array response are
nique has been borrowed from engineering work summarized as follows:
done in radio-antenna design (Dobrin and Savit
The response curve repeats itself at
1988, p. 99). The application of the principles in
(x/k = 1), that is at (x = k), so it is com-
geophone- and source-arrays was published as
pletely dened in the interval (x/k = 01).
early as mid-1950s (Lombardi 1955) and (Parr
The curve is symmetrical about the point
and Mayne 1955).
(x = k/2)
(ii) The Array Response Function It has (n) lobes (maxima), where the principal
Directivity curves express the array response to main lobe centered at (k = ) and another
the different wavelengths of horizontally travel- equal lobe (the rst alias lobe) centered at
ling surface waves. For a linear array consisting (k = x)
of n elements with x inter-spacing, the response It has (n 1) nodes (zero-response) at
function, R(b), is given by: k = nx, nx/2, nx/3, , n x/(n 1). At
these points the attenuation is innite, that is
Rb sin nb=sin b there is no output.
It has (n 2) side lobes found in the range
where, b = px/k, and k is the apparent wave- k = nx to nx/(n 11). This zone, called
length of the surface waves travelling in the the (reject zone) or the (effective array
direction of the linear array. length), is the zone where the most effective
The R(b) function is periodic, repeating at a noise-attenuation occurs. Sometimes the
period of (b = p). The function is fully dened in reject zone is taken to be the zone found
the range of (b = 0 to p), that is, in the range between the 3 or 6 db points. A response
(x/k = 0 to 1). The curve, within this range, is curve for an 8 elements linear array is shown
symmetrical about the point (b = p/2), that is in Fig. 6.30.
about (x/k = 1/2), and it crosses the b-axis at
This gure shows that, the attenuation-zone
the points b = p/n, 2p/n, 3p/n, , (n 1)p/n.
(reject zone) lies between (k = 8x) and
The curve has a total of (n) maxima, where the
(k = 8x/7). This zone can be made wider (ex-
principal one is located at (b = 0), and the peaks
tended) and its response level lowered further, by
of the side lobes occur at the mid points between
increasing the number of elements (n). However
the zeros of the function.
this will cause narrowing of the pass band
Mathematical derivation is based on summing
(marked by the rst lobe where k nx) and
of spatial harmonic waves moving along a linear
increasing its cutoff slope.
array made up of n elements, arranged at con-
The role of the directivity curves is in evalu-
stant interspacing x (Sheriff and Geldart 1995,
ation of array performances in attenuation of
p. 247).
horizontally moving surface waves. Given the
(iii) The Array Response Curve wavelength of the dominant coherent noise, a
The response curve is normally plotted as relative seismic detector array can be designed such that
response against (x/k), or against (k). The its response is most effective in attenuation of the
response value is plotted directly or in db units. noise having that wavelength. From the drawn
As it is mentioned above, the response, by de- curve, one can directly read the relative
nition, is expressed by the ratio between the array noise-cancelation effect corresponding to the
output and that of the same elements gathered at wavelength of the dominant surface waves (k), or
one location, that is when (x = 0). It should be to the parameter (x/k). This means that we need
130 6 2D Seismic Reflection Surveying
response
2-geophone linear array
array
0.5
0.5
0.5
Fig. 6.29 Schematic representation of response curves for different linear arrays
1.0
main alias
lobe reject zone lobe
0.5
R()
x/
0
0 1/8 2/8 3/8 4/8 5/8 6/8 7/8 1
Fig. 6.30 Schematic representation of response curve for 8-element linear array, showing the main and alias lobes in
addition to the arrays reject zone
6.3 Seismic Noise Characterization and Attenuation 131
to know the wavelength of the surface-waves in The xed spread method is done by using
order to design the geophone array which is end-on spread which is made up of geophones
appropriate for the attenuation of the surface spaced at short distances (typically, 5-m spacing).
waves. The wavelength, which is representative After the rst shot is red and the shot record is
of this type of coherent noise, is normally produced, the shot is moved along the spread line
determined by special seismic experiment called by a distance equal to the spread length, and the
the (noise test). second shot-record is obtained. This
shift-and-record action is repeated a number of
times such that the resulting combined record will
6.4 Field Measures for Signal have its maximum offset to be about equal to the
Enhancement maximum offset planned for that survey. An
alternative method (called, the xed-shot method)
Before commencing a seismic reflection survey, is conducted by xing the shot location and move
certain eld procedures are normally carried out the receiver spread (along the spread line) away,
to determine the nature of the dominant noise in hence, the term (walk-away test).
the area in order to be able to design the source In both methods, each receiver point is
and receiver parameters which can output stron- occupied by a single geophone or by a group of
gest reflection signal and least possible noise geophones gathered in one location to give
level. For noise analysis, a special seismic clearest noise possible (Fig. 6.32).
experiment is done in the eld, called the noise The more commonly applied method is the
test (called also, walk-away or micro-spread xed-spread method because it is easier to move
test). Analysis of the resulting seismic record the shot location than moving the spread.
will lead to determination of coherent noise Noise-test operation is normally including
parameters such as apparent wavelength, veloc- measures for studying the broad-side noise as
ity, periods, and frequency, which are necessary well as the inline coherent noise. For this pur-
in the design process for both of the source and pose, another spread is laid perpendicular to the
receiver arrays. normal inline spread (Fig. 6.33).
The end result of the eld operations in con-
ducting the noise test is a noise-test record in which
6.4.1 The Noise Test the 5-m-spaced traces are drawn. It is important to
note that no ltering should be applied during the
There are two types of eld procedures that can recording process. For interpretation purposes, the
be followed to conduct the noise test in the eld. noise-record display should be produced without
These are the xed-spread method and the any type of time-variant scaling. Any amplitude
xed-shot method (Fig. 6.31). distorting process should be avoided in order to be
shot
receiver points
Shotpoint
cross-line
receiver
spread
geophones
grouped in one
location
Fig. 6.33 Noise test L-shaped spread to detect both of the inline and cross-line coherent noises. A receiver point may
be occupied by a single geophone or by a group of geophones
able to measure relative amplitudes, and for easier Interpretation of the noise-test data is nor-
event identication. However, time-invariant scal- mally done visually. It starts with identication
ing can be applied. Principal seismic noise events of the types of noises and other types of seismic
(and other wave arrivals) are schematically shown events. The main events a noise-test record
in Fig. 6.34. contains are: direct waves, refracted waves,
offset
A
B F
E
C
F
C
D time
noise cone
Fig. 6.34 Schematic representations of the main events, e back-scattered waves, f reflected waves. The group of
normally found in a noise-test record. a Direct wave, events between events (a) and (d) forms the noise cone
b refracted wave, c ground roll, d air wave,
6.4 Field Measures for Signal Enhancement 133
630m/s
2270m/s
reflection event
970m/s
450m/s
350m/s
Fig. 6.35 Actual noise-test record, showing reflection, refraction, sound, and surface waves
ground roll, air-wave (sound wave travelling period (s) are manually measured from the paper
through air), back-scattered waves, and wave record. From these parameters, frequency (f), and
arrivals of energy reflected from deep interfaces. wavelengths (k) are then calculated using the
The strongest events, appearing in a noise-test relations (f = 1/s, k = v/f). Interpretation results
record, consist of surface waves, covering a are found to be as quoted in the following
range of apparent velocities and apparent wave- Table 6.2.
lengths. These events make up the coherent noise These data show that the most distortive event
forming a group of large-amplitude events is the surface wave-3 (of velocity 970 m/s and
known as the (noise cone). An actual record of wavelength 108 m). It has the largest amplitude
noise test is shown in Fig. 6.35. and largest extent, covering all the record time
In interpreting the actual noise record, given and all of the offset range in this exampler. On
in Fig. 6.35, it is found that it contains body average the wavelength (100 m) can be consid-
waves (direct, refraction, and reflection events), ered as representative of the surface waves in the
surface waves (ground roll), and air wave (sound survey area, and can be adopted in the design of
wave). The wave parameters: velocity (v) and the geophone and shot-hole arrays.
6.4.2 The Experimental Shooting done visually to determine the best parameters.
Choosing the appropriate parameter is based on
The source parameters depend on the type of the reflection-signal strength, frequency content,
source used to generate the seismic energy. For the reflection-signal resolution, noise level, and
dynamite source, there are three main parameters. extent of penetration depth.
These are: charge depth, charge weight, and
shot-hole pattern. The optimum charge depth and
charge weight are determined through direct 6.4.3 Determination of the LVL
experimental shooting. For the charge depth, Properties
several trial shots are conducted such that the
charge weight is xed while varying the depth. The earth surface is characterized by its
The same approach is followed in optimizing the non-uniform topography (variable elevation),
charge weight that is by xing the depth and and the surface layer is, in general, made up of
varying the weight of the charge Fig. 6.36. lose low-velocity material. This layer, (com-
As for the shot-hole pattern, the suitable monly referred to as the low-velocity layer, LVL)
number of shot-holes is determined by applying is made-up of one or more layers of velocity and
the same principles applied in the case of the thickness that can vary with location within the
receiver geophone array. However, very often survey area. The LVL thickness and velocity are
and for economical reasons, this is determined by typically of ranges (1050) m and (5001500)
the experimental-shooting method alongside with m/sec respectively.
the determination of optimum charge depth and Because of the large velocity contrast nor-
charge weight. Conducting multi-shot sources is mally found between the LVL material and that
another approach for increasing the of the medium below it, the base of the LVL acts
signal-to-noise ratio. This is normally done in as a strong reflector and refractor. Multiple
case of surface-sources as in the case of vibroseis reflections (reverberation and ghosts) can also
and weight dropping techniques. A typical plan develop in such environments. Effect of the LVL
for an experimental shooting for optimizing the is not restricted to the travel-time changes, but
source parameters is as shown in Table 6.3. also on the reflection waveform. In particular,
The shot records for these trial shots should be high-frequency components of the travelling
outputted with no ltering and with no seismic waves experience severe attenuation in
time-variant scaling being applied to them. the LVL due to the relatively strong absorption
A comparative study of the obtained records is phenomenon.
Fig. 6.36 Experimental shooting for determination of: a charge depth and b charge weight
6.4 Field Measures for Signal Enhancement 135
Table 6.3 A typical plan for an experimental shooting for optimizing the source parameters
No. Shot-hole depth (m) Number of holes Charge weight (kg)/hole Shot-hole Spacing (m)
1 6 2 3 25.0
2 6 3 3 12.5
3 6 4 3 8.3
4 9 2 4 25.0
5 9 3 3 12.5
6 9 4 4 8.3
7 12 2 3 25.0
8 12 3 3 12.5
9 12 4 4 8.3
An integral part of the eld activity in a points arranged at certain depths inside the
seismic reflection survey is determination of the borehole Fig. 6.37.
properties of the surface layer, the LVL. In par- These dynamite capsules are red in sequence
ticular, effort is made to determine the thickness starting at the base of the hole and continuing
and velocity which are essential information upward till the last shot which is nearest to sur-
needed in correction-computations of the reflec- face. For recording the arrivals at the surface, a
tion and refraction travel times. The commonly group of geophones are usually planted at equal
applied methods are the up-hole surveying and distances from the surface location of the hole.
specially designed refraction surveying. For analysis and interpretation of the recorded
data, travel times (reduced to vertical travel-path)
(i) LVL Characterization by Up-hole
are plotted against depth. From the slope of the
Surveying
resulting plot, and the depths of the points at
The up-hole survey involves drilling a borehole
which slopes show abrupt changes, the velocity
of depth exceeding the expected LVL thickness,
and thickness of the LVL layer (or layers) are
normally within the range (50100 m). Small
calculated.
charges (dynamite capsules) are red at a series of
(a) (b)
shot hole detector vertical time, Tcos
depth LVL
(h) thickness
slant
shot time, T
depth
(h)
Fig. 6.37 Up-hole survey: a travel-path and b plot of vertical travel time (Tcosh) against depth (h), where (T) is the
slant time and (h) is the angle between travel-path and the borehole
136 6 2D Seismic Reflection Surveying
(a) (b)
time time time
distance distance
direct arrival
refracted arrival
Fig. 6.38 Two methods of refraction survey specially designed for determination of the LVL properties (thickness and
velocity), a xed spread, and b xed shot-point method
6.5 The Seismic Field Crew 137
(ii) Topographic Surveying Section responsible for the dynamite storage and
Fixing on the ground, of the survey points transport taking all the safety and secu-
(shots and receivers) and measuring coor- rity precautions.
dinates (x, y, & z) of each of these points. (v) Mechanical Engineering Section
These data, in addition to data concerning For mechanical work needed by the crew
geographical nature and surface environ- as maintenance of drilling machines,
ments, are documented and reported. trucks, electricity generators and the
(iii) Drilling Section Vibroseis systems.
This is concerned with drilling the (vi) Administration and Finance Section
shot-holes, and other holes needed by the This section is responsible for personnel
survey, such as the deep holes needed for recruitment, living requirements, trans-
up-hole surveys. port, communications, material storage
(iv) Shooting Section and all nance affairs.
Workers in this section do all the nec-
These are the main sections of a typical crew
essary steps needed to prepare a
using dynamite for the seismic energy source. All
shot-hole. This involves preparation of
the sections are headed by the party chief who is
the right amount of charge and placing it
managing the crew work through direct contacts
at the required depth and lling the
with the crew and through the daily meetings
hole with water and mud mixture to
held every night with the sections heads.
secure coupling. The section is also
3D Seismic Reflection Surveying
7
In the years following the dramatic introduction three-dimensional shape, which is spherical, in a
of the analogue and digital data recording and homogeneous medium. In reality, the geological
processing, the main development-trends were in medium is made-up of different rock layers of
increasing recording channels and in more different physical properties and different geo-
advanced magnetic tape technology. These metrical shapes. With this realistic type of envi-
developments have led, in the early 1970s, to ronments, the advancing wave-front, is still
more effective exploration techniques including having the three dimensional form but no more
the introduction of seismic three dimensional perfectly spherical.
(3D) surveying. The rst proper 3D exploration The created seismic rays travel in every
work was conducted using xed cross spreads by direction in the space surrounding the source
Walton in the early 1970s (Walton 1972), and zone. In other words, the geological eld, as well
soon after the 3D seismic method entered appli- as the seismic eld in which it is created, is 3D in
cation on commercial-production basis. At pre- nature (Fig. 7.1).
sent the 3D seismic surveying became very In the conventional 2D shooting, the seismic
common especially for detail exploration and eld is created as three-dimensional wave-eld,
development of oil elds. but detected by two-dimensional array of
detection-points. Thus, the 2D surveying is
concerned with a limited portion of the reflected
7.1 Introduction 3D wave-energy. The rest of the reflected energy,
with all the useful information it is carrying, is
7.1.1 Nature Is Three Dimensional left to pass undetected.
In the 3D technique the source energy is uti-
The sub-surface geology targeted by seismic lized more efciently, since most of the reflected
exploration is, in essence, three-dimensional in 3D wave front will be detected. It is, in fact,
nature. The seismic eld created in seismic considered to be the more logical approach for
reflection surveys is likewise, three-dimensional seismic exploration than the conventional 2D,
(3D). In an ideal homogeneous medium, the because the 3D surveying conforms to the 3D
wave-front of the advancing seismic wave is of nature of the subsurface geology.
(a) Subsurface geology is 3D in nature of the transit-time with the depth of the well
which is also a 1D function.
In the conventional 2D seismic surveys,
where the shot-point and receiver-points are
co-linear, the resulting stack section consists of
a series of seismic traces each of which belongs to
a CMP location. The stack traces of a stack section
are uniformly spaced along the distance coordi-
nate (x). Thus the seismic amplitude in the pro-
(b) Seismic field is 3D in nature duced section is function of both trace-position
(x) on the seismic line and the two-way reflection
time (t). This means that the amplitude in the
produced seismic section, is represented by the
two-dimensional (2D) function f(x, t), or f(x, z)
when time is scaled by the propagation velocity.
With the 3D shooting technique, the
produced CMP-locations form a two-dimensional
array over the surveyed reflectors. In this case, we
Fig. 7.1 By nature, both of the subsurface geology have a seismic trace for each of these CMPs,
(a) and the seismic eld (b) created to explore it, are three
forming a data volume in which the seismic
dimensional
amplitude is represented as a function of its
7.1.2 1D-2D-3D Terminology position in space, dened by the three-
dimensions (x, y, and t) or by the dimensions (x,
A seismic trace can be viewed as a time function y, z) when the third dimension is expressed in
of the seismic amplitude which is normally rep- terms of depth, z. Thus, the amplitude is expressed
resenting variation of vibration-velocity as as three-dimensional (3D) function, f(x, y, t) or
function of recording time. If we disregard its f(x, y, z).
position information (that is neglecting the x-y In reference to (Fig. 7.2), the three terminol-
coordinates of its CMP location), the seismic ogy denitions are considering the seismic
trace is represented as a one-dimensional (1D) amplitude (a) as function of depth (z), function of
function of amplitude with time, f(t), or with (x, z), or function of (x, y, z).
depth, f(z). The synthetic seismogram is another The end product of the 3D seismic surveying
example of the 1D seismic-function. It represents is a seismic data volume which represents a
reflection-amplitude variation with the depth of three-dimensional function a(x, y, z) expressing
the drill-hole. The sonic log expresses variation the variation of seismic amplitude (a) with the
1D 2D 3D
y
x
Fig. 7.2 Seismic data represented as 1D-function (seismic trace), as 2D-function (seismic section), or as 3D-function
(seismic data volume). The corresponding functions are a(z), a(x, z), a(x, y, z)
7.1 Introduction 141
three coordinates: x, y, and z. In this way, one (i) Erroneous Image Positioning
may consider the 2D section as a special case of Geometry of the 2D seismic line, assumes that
the more general 3D data-volume, in which one the source-receiver line, the incident seismic ray,
of the dimensions (x or y) is equal to zero. In the and the reflected ray are all lying in one vertical
geophysical literature it is customary to use the plane (the ray-path plane). Further, the reflection
terminology (data-box, or data-volume) for the point is located vertically below the
whole stack-data set outputted by the 3D survey. source-receiver mid-point. This is, in fact true
For display purposes, 2D sections from the data only when the reflector is a horizontal plane
volume, in both vertical and horizontal direc- surface or the seismic line is shot along the
tions, can be obtained. non-pinging axis of an anticline for example
(Fig. 7.3a). In the case of a dipping reflector,
however, the reflection events may be received
7.1.3 Limitations of 2D Seismic from points located outside the vertical plane
Surveying which is, in 2D surveying, assumed to be the
plane of the reflection ray-path (Fig. 7.3b).
Despite its outstanding success in subsurface A dry well may be obtained due to erroneous
exploration which has been achieved throughout positioning of the seismic image that may appear
the past years, the conventional 2D seismic on the 2D seismic section.
method is facing certain difculties in retaining An example showing erroneous image-
its well-established success standard. This is due positioning is the case of a seismic line passing
to the fact that most of the large structural traps near-by a dome. The 2D line does not cross the
with clear geophysical anomalies have already structural dome and yet the produced seismic
been discovered and what remained undiscov- section shows an anticlinal image (Fig. 7.4).
ered are those which are characterized by being Appearance of the image of a subsurface
small in size, complex in structure, weak in dome, which is not crossed by the vertical plane
seismic response, and may be situated in inac- containing the seismic line, is due to reflections
cessible areas. Limitations of the 2D seismic from the sides of the offside dome.c
surveying may be summarized as follows:
(ii) Distortion due to Dipping Reflectors
Strictly speaking, the 2D surveys give true struc-
A B tural pictures only when the surveyed areas are
made up by horizontally layered geology. In
RCVR line general, there is always a certain amount of
image distortion produced by 2D method when-
ever dipping reflectors exist. In processing of 2D
data, the reflection events are assumed to have
been received from reflection points which are
CMP line
located vertically below the surface seismic-line.
When the reflectors are not horizontal, and not
continuous, planes the 2D data will give erroneous
dipping subsurface images. With 2D migration, distortions
reflectors due to reflectors, dipping in the source-receiver
direction, can be corrected, whereas distortions
Fig. 7.3 Reflection ray-paths for zero-offset receivers in due to dipping in other directions cannot be cor-
2D surveying. a Line shot along strike direction, located
directly above an anticline of non-pinging axis. b Line
rected. Regarding dip-effect on seismic results we
shot parallel to line (a), but at horizontally-shifted position can distinguish the following three cases:
142 7 3D Seismic Reflection Surveying
2D seismic line
time
2D seismic section
Fig. 7.4 2D seismic line shot and an offside dome. The produced seismic section shows a false structure due to
reflections from the side of the offside dome
R
S R S R
RP RP
RP
(a) (b) (c)
Fig. 7.5 Distortion effect of dip direction on reflection point location. a Case of no dip, b dip in the S-R direction, and
c dip is perpendicular to S-R direction. S and R are the source and receiver points respectively
(Case-1) Dip is of zero value (Fig. 7.5a) (Case-3) Dip has component perpendicular to
In this case, where the reflectors are the shot-receiver direction (Fig. 7.5c)
horizontal planes, the reflection point is The reflection point, in this case, is
located vertically below the shot- shifted outside the shot-receiver ver-
receiver midpoint, hence the reflec- tical plane. In this type of situation 3D
tion images, produced by 2D data, are survey has to be conducted and the
found in their proper positions, and no resulting data-volume must be sub-
distortions occur in this case. jected to 3D migration process in
(Case-2) Dip is in the shot-receiver direction order to get the correct subsurface
(Fig. 7.5b) structural image.
In this case, the reflection point is Distortion always occurs in case of dipping
shifted up-dip by an amount of shift reflectors and the distortion-severity depends on
which is dependent on the dip value. the dip and on dip direction. Thus, if a line in
The 2D method is adequate for this type 2D survey happened to be along the
of situation provided that the resulting dip-direction, then the reflection ray-path will be
stack section is corrected through in the source-receiver vertical plane, and in this
an appropriate 2D-migration process. case, no out-of-plane reflection points shall
7.1 Introduction 143
occur. The dip-caused distortion in this case, is On migration of the sections of these two
correctable in the processing stage (by applying intersecting lines, the strike-line section experi-
2D migration). In the general case, where the 2D ences no change whereas all events on the
line does not exactly coincide with the true dip dip-line section will be up-dip shifted. It is
direction, the reflection-events will not be cor- expected that this situation will lead to a mistie
rectly repositioned even with the process of 2D on the intersection point of the two sections. In
migration. In fact the amount of correction will fact, a stack section of a strike-line has two
be corresponding to that component of the dip problems. These are: creation of a mistie (at the
found along the line direction. To get the com- intersection-point) with the section of the dip
plete migration, the data must be subjected to line, and the misallocations of the reflection
another migration-process to be implemented in a CMPs which are not lying in the vertical plane as
direction perpendicular to the line direction the 2D stack sections are normally displaying.c
which is, of course, not possible. It can be con-
(iv) Weak Resolution of Structural Changes
cluded, therefore, that the 3D data volume,
Due to the wide spacing of the 2D lines, con-
obtained from a 3D survey, is a necessary
struction of the isochron contour maps are based
requirement in order to achieve proper migration
on interpolation of the reflection times measured
process regardless of the dip direction of the
on the survey lines. For this reason, structural
subsurface reflectors.c
details are generally not resolvable by the 2D
(iii) Creation of Misties surveying. Structural resolution gets worse as the
Let us consider the zero-offset sections (as stack line-spacing gets large compared with the sizes
sections) of two intersecting seismic lines of the structural anomalies. The isochrones map
(dip-line and strike-line) shot over a dipping of an area covered with ve 2D lines is shown in
plane reflector (Fig. 7.6). (Fig. 7.7a). In this case, the 2D survey failed to
For the strike line, the reflection ray-path resolve the three domes detected by the 3D sur-
plane is slant and not vertical, as it appears in the vey carried out in the same area (Fig. 7.7b).
seismic stack section, and the reflection events The weak resolution power of 2D surveying
obtained from the dipping reflector shows no becomes more serious problem when the struc-
dipping feature. On the other hand, the dip line tural anomalies are of dimensions smaller than
will be of slant ray-paths, all lying in the vertical the applied line-spacing.
ray-path plane.
7.1.4 Merits of the 3D Technique
dip line
surface plane
Due to economic constraints the 3D method is
strike line used only in cases when more detailed and more
zero-offset
reflection accurate exploration are aimed at. Typically, it is
ray-path
applied for restricted purposes such as develop-
ment and appraisal of already discovered
zero-offset oil-elds. When economy allows, however, 3D
reflection CMPs
ray-path P surveying is applied in the normal general-
Reflection
purpose seismic exploration.
dipping plane As we have seen above, the 2D method gives
CMPs
distorted images of the subsurface geology espe-
cially in case of dipping layers. The 2D method also
suffers from weakness in its ability of resolving
Fig. 7.6 Effect of dipping reflection plane on zero-offset
small geophysical anomalies and ne structural
reflection points in case of two 2D intersecting lines.
Strike stack section shows no dip, whereas dip line does. and stratigraphic details. Inability of surveying
On migration, a mistie occurs at intersection point (P) inaccessible areas is yet another limitation of
144 7 3D Seismic Reflection Surveying
Line-5
the 2D seismic technique. These limitations were direction, in addition to possibility of getting
removed or very much reduced by the application horizontal sections (time-slices) at any level
of the 3D method. The advantages of the 3D sur- within the data-volume (Fig. 7.8).
veying may be grouped under the following three
(ii) Capability of Surveying Inaccessible
groups of application domains:
Areas
(i) In the Geophysical Domain As far as seismic surveying is concerned, an area
3D surveying is furnishing good control on the is considered to be inaccessible, when neither
determination of processing parameters, such as seismic energy-sources nor receivers are allowed
static values, stacking velocity, and dip-vectors, to be located within the area boundary. An
and hence it allows more effective three- inaccessible area can be surveyed by deploying
dimensional migration process. Increased ef- the receivers on the boundary of the area and
ciency in the use of seismic source energy, due to shooting at source-points which are also dis-
the large number of active detection channels tributed over the area boundary. The surveying is
normally used in 3D eld acquisition. The 3D data conducted by shooting the source points in
processing provides the stack data-volume, which sequence while the receiver spread is kept xed
allows displaying vertical sections at any throughout the shooting process. For each shot,
3D data box
vertical
y-sections
vertical vertical
x-sections sections
horizontal
sections
7.1 Introduction 145
source
point
there will be a number of CMPs located at the (iii) In the Geological and Reservoir
sourceto-receiver midpoints, which will be Domains
located within the surveyed area (Fig. 7.9). The 3D data can furnish accurate information on
By this type of shooting (called loop- the subsurface geological structure of the area.
cshooting), the area will be covered with CMPs Unlike 2D data, it can resolve small and complex
without having survey points (sources or recei- structural and stratigraphic anomalies. It provides
vers) existing within the area. Special software direct information (with no interpolation proce-
can handle the recorded data and get it processed dure) of the subsurface geology expressed in
to produce a 3D data-volume. its real three-dimensional image. Due to the
Although this technique makes seismic sur- dense sampling points of the geological space,
veying of an isolated inaccessible area possible, the 3D data becomes more readily-interpretable,
it suffers from a number of weak points. The leading to increased degree of accuracy and
distribution of both of the recorded CMPs and much improved resolution-power (Fig. 7.10).
the fold of coverage are not uniform. Also, it is In the reservoir domain, 3D data provide
possible to lose shallow reflections with traces delineation of the 3D shape and spatial limits of
recorded at large-offset receivers. These limita- oil-elds as well as giving more accurate esti-
tions become less effective, the smaller the area mation of their hydrocarbon reserves. The data
and the deeper the targeted reflectors. Applica- can help in outlining oil/water contact which can
tion of the concept (loop spread surveing) was assist in more accurate emplacement of devel-
dealt with in more detail in (Alsadi 1992, 1994). opment and production wells. Reservoir studies,
based on 3D data, help in more accurate reservoir study and prepare its nal report for guiding work
characterization in terms of facies distribution in the plan implementation.
and fluid content. Such information would lead The preplanning phase is an important pre-
to more accurate oil-eld characterization and requisite needed in order to facilitate the work
enhanced production rates. execution especially in the data-acquisition
stage. Pre-knowledge of the area assist the geo-
(iv) In the Economical Domain
physicist to be prepared, and make the necessary
Despite the overall high cost and long survey
provisions for any problem that may crop up
duration compared with the 2D method, the 3D
during the work. Examples of such problems are:
method is considered to be cost-effective when the
skip shots, make-up shots, seismic-energy to be
merits of its results are taken into consideration.
used, work permit, and other administration and
From the economical point of view, 3D data would
logistical matters.
give increasing success-ratio (producer-to-total
The preplanning report of a 3D seismic survey
wells ratio) and improving well production-rates.
normally covers the following basic items:
The over-all eld evaluation-cost is reduced
through less development-wells, proper well (i) Compilation and assessment of available
emplacement and shorter development-period. technical data.
In view of its exploration efciency, the 3D (ii) Survey objectives.
method gives an enhanced success ratio of (iii) Operational constraints.
development drilling and increase in the oil (iv) Survey-design plan.
reserves. These features form the basis for the (v) Scouting and preliminary tests.
justication of conducting 3D surveys especially (vi) Work organization and scheduling.
for oil-eld appraisal and eld development. In (vii) Cost analysis.
comparison with development drilling, the justi- (viii) Health, safety, and environment
cation is even stronger in comparison with the protection.
2D method. For example, a 10 km2 3D survey
It should be noted here, that the item of the
may cost no more than 10 % of that of drilling a
contract preparation is deliberately omitted from
well of 3000 m-depth. The overall saving
the above mentioned list. The omission of this item
achieved from a 3D survey is represented by
is made on the basis that it represents an indepen-
reduced eld-evaluation costs and by earlier
dent piece of work to be carried out in the following
production due to shortened development-period.
stage, rather than being part of the preplanning
report. This is natural procedure, since the contract
items are depending on the survey parameters
7.1.5 Survey Preplanning which need to be prepared and made ready before
starting of contract drafting. Here-below we shall
Although, in principle, work preplanning is give brief explanatory notes for each of the items
essential in any seismic surveying project, it included in the preplanning report.
becomes more necessary prerequisite when deal-
ing with 3D surveying. The urge for the preplan-
ning phase stems from the complex nature of the 7.1.6 Cost Considerations of 3D
3D surveying work which involves large nancial Surveys
investments. A feasibility study should be con-
ducted as a joint effort by a team of personnel from 3D surveying is usually quoted as being a costly
all parties concerned with the 3D survey project. In process. To be more objective, however, the cost
particular, the team should include specialists in the assessment must be made in relation to the
data acquisition, data processing and interpretation, turnover benets. In fact, even 2D surveying can
as well as experts on legal and contractual matters. become very expensive when using very small
The task of this team is to carry out the preplanning line spacing, and high fold of coverage. It is
7.1 Introduction 147
generally accepted now that the 2D and 3D the region of (1015) % of the cost of the data
methods become equivalent cost-wise when the acquisition. These gures are near-realistic when
2D lines are spaced by about half a kilometer. the survey-points for both methods are spaced by
The 3D surveying is considered to be 50 m and the coverage folds are 1600 and 3200 %
cost-effective compared with the 2D method for the 3D and 2D respectively. In general, the
considering the merits of the 3D data. Compar- cost of 3D surveying depends on the surface
ison between 2D and 3D survey-cost is more environment of the survey area (Fig. 7.11).
meaningful, when it is realized that 3D data
(ii) Dependence of Cost on Survey
provide dense spatial sampling of reflectors,
Parameters
enhanced structural resolution, and undistorted
The survey parameters which are most influential
3D images.
on survey cost are the spread parameters, fold of
The 3D-Survey Cost Elements coverage, bin size, and shot density used in a given
In discussing survey costs, and 3D-to-2D cost survey. These elements are interrelated in their cost
comparison, it is of course very difcult to quote contribution. Thus, fold (F) may be statistically
precise gures. However, it is possible to give a estimated from the formula:c
fair assessment of the main elements of the cost.
These elements are important issues to be con- F Ntr =Nbin
sidered in cost analysis of any 3D surveying
project. where, Ntr and Nbin are the survey total number
The main cost-elements of the 3D survey can of traces and total number of bins respectively.
be summarized as follows: By substituting the product (shot density,
SHD) x (survey Area) x (live channels per shot,
(i) Surface Conditions NCHN), for Ntr, and (survey area/Bin Size) for
Difculty in quoting precise gures is due to the Nbin, we get the important relation:
wide variation between types of surveys. In par-
ticular, prevailing surface conditions, and types of F SHD NCHN Bin Size
the applied survey parameters differ from survey
to survey. In the present world-wide seismic Hence,
industry, the cost of the data-acquisition of
onshore surveys (on-land surveys) is approxi- SHD F=NCHN Bin Size
mately in the range of (510) k$/km for 2D
method and (2040) k$/km2 for 3D. The cost of Since total survey-work depends on shot
processing and interpretation is estimated to be in density (SHD), this formula is showing that the
total work done in conducting a 3D survey (and
cost hence total cost) is directly proportional to fold of
k$/km2 coverage and inversely proportional to bin size
and number of active channels used per shot.
100 A closely related to cost, is the (turnaround
time), or the duration of the complete 3D survey.
For a given survey (given area size and given
50 survey parameters), survey duration depends on
the work-production rate, which is, in turn,
farm, shallow Jungle dependant on the applied work parameters, like
offshore desert urban water survey parameters, type of energy source, surface
conditions, eld processing, topographic survey,
Fig. 7.11 Sketch showing dependence of survey cost permit fees, and number of work-shifts fol-
(cost of eld data acquisition) on surface environment lowed in conducting the survey.
148 7 3D Seismic Reflection Surveying
It is found that the eld data-acquisition of a subsurface circle will have a radius of half that of
land 3D survey of a 500 km2-area, say, would the surface circle, and all reflected waves received
last for about (23) months. Adding to this, time by the detectors, distributed over the surface cir-
needed for processing and interpretation, the total cle, will have arrived at the same time. However, if
duration time to complete a 3D project of such an the receiver points are laid down on a straight line
area is estimated to be in the order of about (68) passing through the source point, the reflection
months. points from which the incident energy is reflected
and detected by the co-linear receivers, will fall on
a subsurface straight-line with point-spacing equal
7.2 Definitions and Basic Principles to half that of the receive points. In this case, the
detected reflection information (seismic ampli-
7.2.1 Definition of 3D Surveying tude) is restricted to the 2D vertical plane con-
taining the reflaction ray-paths (Fig. 7.12).
Consider a case where seismic energy, generated As this gure is showing, the difference
at a surface point, is reflected from a subsurface between 2D survey and 3D survey is based on
horizontal planer reflector. The generated seismic the way the spread elements are distributed on
wave-front, which is of 3D form, spreads out the surface. In fact, what makes a survey 2D or
through the geological medium (assumed to be 3D is the way the receivers are deployed on the
homogenous) with a velocity decided by the surface. In 2D surveying, receivers are laid down
physical properties of the medium. At the instant in line with the source point, and the used spread
the wave-front hits the reflection plane, a circular is linear, made up of one-dimensional array of
zone of it (of the reflection plane) will be illu- receivers, and in this case, the reflection points
minated by the incident wave-front which is, will fall on a subsurface straight line which is the
according to reflection laws, reflected back to the projection of the receiver line onto the horizontal
surface. reflection-plane. In the case of 3D surveying, the
Using the ray-concept, detection points dis- spread consists of receivers which are distributed
tributed over a surface circle centered about the over an area, and the used spread consists of a
source point, will receive reflected energy from two-dimensional array of receivers. The reflec-
reflection points located on a corresponding sub- tion points in this case will be distributed over a
surface circle. Based on the laws of reflection, the subsurface area of the reflector-plane.
RP RP RP RP RP
subsurface reflector plane
7.2 Definitions and Basic Principles 149
The spread in the 2D geometry consists of a Considering horizontal plane reflectors, all the
receiver points co-linear with the source point, reflection ray-paths, of the CDP trace-gather in
whereas the spread in 3D survey takes the geo- 2D surveying, are coincident on one common
metrical form of a rectangular grid of receiver vertical plane (the ray-path vertical plane). In the
points. Typical 2D and 3D spreads are shown in 3D method, however, each reflection ray-path
the following Fig. 7.13. may fall in its own vertical plane. These ray-path
Whether it is 2D or 3D survey, the planes are generally not coinciding on each other,
reflection-points (RPs), created from ring a because of the varying source-receiver directions
shot-point, are spaced by half the receiver spac- (azimuths) of the CDP gather-traces. Thus, the
ing. The spread used for 3D surveying (3D receiver azimuth in 2D surveying is constant. It is
spread) is normally covering a rectangular area. in the direction of the seismic line. The corre-
From geometrical consideration, it can be seen sponding azimuth, in 3D surveying, is variable
that the subsurface area covered by the reflection because of varying receiver bearing in relation to
points is also a rectangle which is of an area the source point.
equal to quarter of the area covered by the 3D The ray-path planes of the reflection seismic
spread. This can be readily observed from the rays (of a 3D-survey shot-point) are vertical
Fig. 7.13. planes but they assume different directions. This
Shot-records, obtained from 2D and 3D is the case because the receiver points are located
spreads, are shown schematically in the follow- at different bearings in relation to the shot loca-
ing Fig. 7.14 tion. The ray-path bearing, with respect to the
RP RP RP RP RP
R R R R R
(b) S
RP RP RP RP RP
S RP
source receiver reflection point
150 7 3D Seismic Reflection Surveying
CMPs CMPs
distance, x
time line-1 line-2 line-3 line-4
T
shot point, is usually referred to as the azimuth. centered about the CDP. The width and length of
Azimuth variation is considered as an advantage this search area are equal to the CDP spacing in
of the 3D data because with this feature the the two perpendicular directions of the survey
reflection seismic waves shall sample the rock station-grid. This search area is called the (bin)
medium in three-dimensional space and not and the traces included within it form the
restricted to one plane as the 2D data is provid- bin-gather. The process of sorting of traces into
ing. The concept is shown in Fig. 7.13. the appropriate bin is called bin-gridding or
binning. These concepts are shown in Fig. 7.15.
The bin is normally specied by certain
7.2.4 The CDP Bin and Bin Attributes properties or identication criteria. The bin can be
rectangular or square in shape, having a dened
In normal proling technique followed in 2D dimensions and surface area commonly referred
surveying, the spread moves along a uniformly to as the (bin size). Other parameters of the bin
spaced station-points, at equal move-up rate. In concerns its trace gather. These parameters, called
this way, the reflection points of each CDP the (Bin Attributes), include the trace offset, az-
trace-gather are all coinciding at the imuth, and fold of coverage. Normally, the bin
CDP. Sometimes, due to a shifted placement of a attributes are displayed in a special diagram
shot-point, one (or more than one) reflection called the spider diagram as shown in Fig. 7.16.
point may fall outside the CDP location. In CDP
sorting, a search distance (usually dened to be
equal to half of the CDP-spacing) is set in pro- 7.2.5 The Surface and Sub-surface
cessing. All traces found within the search dis- Coverage
tance of a CDP are included in the trace gather of
that CDP. In the normal swath shooting, the surface of the
The same principle is applied to the CDP-grid survey area becomes covered by two sets of
in case of 3D surveying, except that in 3D, we mutually perpendicular lines (the station lines).
have a (search-area) instead of the 2D These are the receiver lines and the shot-lines.
search-distance. The search area is normally The corresponding subsurface lines are the sub-
having a rectangular shape (or square shape) lines (inlines) and cross-lines which intersect at
7.2 Definitions and Basic Principles 151
CDP line
search distance
CDP grid
search area
(BIN)
2 4
3 5 seismic surveying, the corresponding
end-product is a three-dimension space of
trace-1 offset azimuth stacked traces called the data volume or the
data box as it is sometimes called. The data
7 6 volume represents a three-dimensional function
in which the seismic amplitude varies with the
three coordinates; (x, y, t). The reflection-time
BIN CDP
(t) may be replaced by the depth dimension (z),
Fig. 7.16 Spider diagram showing the bin attributes where (z = tv/2) and v is velocity.
(offset, azimuth, and fold) of a bin having 7 traces in its The data volume (in digital form) consists of
gather (fold = 7) discrete volume-elements, called the (resolution
cells) dened as parallelepipeds of dimensions
bin centres. By denition, the bin is a subsurface (x, y, t), where (x) and (y) are the CMP
rectangular (or square) area having dimensions spacing in the inlines and cross-lines respec-
which are equal to half of the corresponding tively. The third dimension (t) is equal to the
surface receiver-station spacing (Fig. 7.17). sampling period of the stack traces. This volume
The subline and cross-line seismic sections element is considered to be the building brick of
are made up of sequence of stack traces, each of the data box, the ultimate product of the seismic
which is located at a bin centre. In this sense, 3D survey.
subline and cross-line seismic stack sections are The resolution cell is also called the (voxel),
considered as of subsurface locations and not in analogy to the term (pixel) used for the unit of
surface locations. a digital picture. The data volume is a subsurface
data-set, made up of voxels of dimensions (x,
y, t) as illustrated in Fig. 7.18.
7.2.6 The Seismic Data Volume With the help of special processing software,
it is possible to extract a variety of seismic sec-
The end-product of a 2D-data processing, is the tions from the data-volume. Normally, three
2D seismic stack section. In case of the 3D types of mutually perpendicular sections can be
152 7 3D Seismic Reflection Surveying
sub-lines
(inlines)
(SL)
cross-lines
(XL) XL
CMP bin y
SL
receiver point x
y SL2
x SL1
t
3D data box
(data volume) resolution cell
(voxel)
extracted. A vertical section may be obtained representation of the seismic amplitude values
along any subline, or along any crossline, and a which fall on the same reflection-time. In fact, it
horizontal section (called a time slice) can be is the locus of equal-time amplitudes existing in
obtained at any reflection-time. It is also possible the data volume. In color displays, the time-slice
to obtain a vertical section connecting points at plots are given in color-coded amplitude-values
arbitrary locations as for example, sections along with clear distinctions between peak- and
distances connecting several well locations. trough-values.
A section which is not a subline or a crossline is The time slice has a similar form as that pre-
normally referred to as a crooked, oblique or sented by the time contour map. The difference
diagonal section (Fig. 7.19). between the time slice and the structural contour
The horizontal section, called seiscrop sec- map is that in time slice, we display different
tion, or more often called (time slice), is a amplitude values at constant reflection time (time
7.2 Definitions and Basic Principles 153
time slice
of the time slice). The structural contour map and data analyses are applied in the eld to
shows different time values of the same reflection determine the optimum parameters applied in the
surface. Briefly stated we say that time slice is data acquisition.
showing different amplitudes measured at one Here below, are the main aspects of the
time-slice, and in contour map, we display the data-acquisition requirements involved in 3D
same reflection event (belonging to a given seismic reflection surveying. Spread parameters
reflector surface) at different time values. Several and shooting procedure is most important aspect
reflectors may be represented in a given of the 3D eld acquisition activities. Many types
time-slice, but different time-values of the same of spread are in application, which can be dis-
given reflector, are shown in a given contour map. cussed under two main headings: marine and
Derivation of structural contour maps from land 3D spreads.
time-slice sections can, in principle, be obtained
by marking the amplitudes of the same seismic
reflector, on a sequence of time-slices, and pro- 7.3.1 Types of Marine 3D Spreads
jecting the picked values for that reflector on one
map. Conventionally, marine 3D surveying is con-
ducted by using spreads similar to those used in
land 2D seismic surveying. The boat and the
7.3 3D Field Data Acquisition towed-behind streamer is recording as it moves
along straight-line courses. The technique differs
In 2D seismic surveying, the shooting spread from the normal 2D surveying in that the survey
consists of a shot point and receiver points which shooting is conducted along closely spaced par-
are all arranged in one straight line (linear allel lines. Line spacing (typically, 100200 m),
spread). In the 3D surveying, however, these is kept constant throughout the survey area.
survey-points are distributed over an area (areal With the progress that took place in the nav-
spread). Essentially the eld work consists of a igation techniques and in navigation data pro-
set of procedures which implement certain geo- cessing, certain improvements were introduced
physical parameters. The eld parameters are on the applied spreads. In cases of calm seas, the
normally adopted through compromising survey boat and its trailing streamer move along
between the ideal technical requirements and the a straight line. In other cases, and because of
overall acquisition cost. Certain test-recording local sea cross-currents, the streamer may drift
154 7 3D Seismic Reflection Surveying
source
hydrophone receiver
recorded CMPs
away from the boat linear course. This is referred reflection points varies as shooting progresses.
to as the (feathering effect), and the angle Computations of the reflection points and CMP
between the drifted cable and the linear distributions are done in the processing stage
boat-track is called the feathering angle based on the gathered boat readings of the cable
(Fig. 7.20). geometry during shooting.
The shape of the drifted cable, for each shot, To increase survey efciency, multi-source
is determined from the readings of the special and multi-streamers have been used in the marine
compasses tted along the cable. Because of surveying. Examples of marine spreads which
changes of sea conditions, the cable shape and can be congured are: single source-single
the feathering angle change from shot to shot. streamer, dual source-single streamer, single
Consequently, distribution of the subsurface source-dual streamer, and dual source-triple
streamer (Fig. 7.21).
Spread movement during recording, is done
(a) normally along parallel straight lines (parallel
linear paths). Because of the long streamer-line
towed behind, the boat needs to turn with a rel-
(b)
atively large steering radius in order to record the
following line. A variation of this method is
(c) shooting along circular paths. The boat moves
along overlapping or spiral circular paths. The
advantages of the circular type of spread, is that
(d)
the survey is completed with no times lost due to
long-time turns of the boat needed in case of
source
hydrophone receiver straight-line surveying. These types of
spread-movements are shown in Fig. 7.22.
Fig. 7.21 Types of spreads used in marine 3D survey- It should be noted here that, deviation of the
ing. a Single source-single streamer, b dual source-single streamer course-line from the linear form (caused
streamer, c single source-dual streamer, d dual by the circular spread or by the feathering effect),
source-triple streamer
Fig. 7.22 Types of spread movements used in marine 3D surveying. a Parallel linear path, b Overlapping circular
path, c Spiral circular path
7.3 3D Field Data Acquisition 155
(d) (e)
Spread
swathmovement along
first spread last spread
Fig. 7.25 Spread movement (shift-and-shoot movements) over the swath, with overlapping shooting spreads
7.3.3 The Swath Shooting Technique gives uniform bin fold and adequate offset and
azimuth coverage. On the other hand, swath
The swath is dened to be a strip of the survey shooting requires land with fully accessible space,
area of width equal to the width of the used and type of surface conditions which allow the
spread that moves along the dened strip in a appropriate freedom for the survey maneuvering.
roll-along movement similar to the A typical example, of shooting parameters
shift-and-shoot technique used in conventional applied in swath survey, is given here-below
2D proling surveying (Fig. 7.25). (Table 7.1).
In practice, when the template reaches the end
of the rst swath, the template is re-congured at
the near-by end of the adjacent swath, then car- 7.3.4 Types of Templates Used
rying out the shift-and-shoot process from end in Swath Shooting
to end. In this way all the rest of swaths are
sequentially covered. In general, swaths are not There are many types of templates that can be
laid down side by side, but laid down with used in swath shooting. A template may contain
overlap to obtain fold build-up (Fig. 7.26). one, or more than one, shot-lines which are
Greater fold coverage is obtained with bigger arranged to be in-centre, off-centre, or at ends of
overlap made between adjacent swaths. the template. The more common types are those
Swath shooting is a common acquisition where the shot-lines are perpendicular to the
technique implemented in todays land 3D sur- receiver lines. Another variation is to have the
veys. It has the advantages of being simple to shot-lines inclined with respect to the
congure, and efcient to execute in the eld. It receiver-lines (Fig. 7.27).
swath-4
swath-3
swath-2
1.4 m
array centre 10.0 m
10.0
m
48.56 m
Geophone type/polarity SM-4/SEG standard
Natural frequency 10 Hz
Critical damping 70 %
String conguration 12 geophns (2 parallel 6 series)
Receiver array parallelogram
Geophones/geophone-group 36
Geophone spacing 4.16 m
Number of strings 3
Stagger between strings 1.4
String separation 10 m
Array length 48.56 m
Array width 20 m
(continued)
158 7 3D Seismic Reflection Surveying
5m
array centre
37.5 m
Source type Vibroseis System (vibrator)
Peak force 45,000 lb
Number of vibrators 4
Sweeps per VP 1
Sweep length 16 s
Sweep frequency 864 Hz
Sweep taper 500 ms at both ends
Source array Parallelogram 2 2
Vibrator inline stagger 12.5 m
Vibrator lateral spacing 5m
Array length 37.5 m
Array width 5m
4. Data recording parameters
Recording system I/O system-II
Channels recorded 1440 data, 4 auxiliary
Record length 6s
Sampling period 4 ms
Tape recording format SEGD
Recording gain 48 db
Diversity staking Active
Stacking/correlation Correlation after vertical stacking
Fig. 7.27 Types of templates used in swath shooting. a Centre symmetric split. b Asymmetric split. c Double off-end.
d Double within. e Centre multi shot-line. f Centre of inclined shot-line
More elaborate spread designs are based on shot-lines. One of such designs is what is called
grouping the receiver points in blocks with the Checkerboard spread. An example represent-
source points distributed regularly over parallel ing this type of spread is shown in Fig. 7.28.
7.3 3D Field Data Acquisition 159
(b)
(a)
Fig. 7.29 Loop-Spread. a Survey points (source and receiver points) laid down around the irregular inaccessible area.
b Survey points laid down on the perimeter of a rectangle (or square) enclosing the survey area
160 7 3D Seismic Reflection Surveying
migration migration
aperture aperture
7.3 3D Field Data Acquisition 161
surrounding the target migrated area is required pre-dened minimum to maximum offset values.
to insure full migration of the target area. The As a rule of thumb, an offset should be roughly
width of this strip (called the migration fringe) is equal to the depth of the target reflector. Thus, by
equal to the migration aperture which is dened dening the depths of target reflectors, it is possible
to be the horizontal distance a reflection event is to determine a range of offsets to suit both of the
moved by a migration process. The width of the targeted shallow reflectors and deep reflectors.
migration fringe is function of the dip measured For a given bin-size, geophone group spacing
in the direction of the fringe width. The migra- becomes known, and both total active channels per
tion fringe is shown in Fig. 7.30. receiver-line and total number of receiver-lines per
The survey area should be large enough to template can be decided, based on the maximum
include the fully migrated zone, migration fringe, offset required. Maximum offset (Xmax) and mini-
tail-fold zone, and the required surface- mum offset (Xmin) can be calculated for the largest
coverage zone. For economic reasons, and when and smallest source-to-receiver points in the
the S/N ratio is sufciently high, the migration designed template (Fig. 7.31).
fringe may be extended in such a way as to include
(iii) The Swath Parameters
part of the tail zone. This step can result in
The main parameters of the swath are shooting
reduction of the total survey area to the area which
direction (survey orientation) and roll-aside
is sufcient for achieving the targeted
overlap. Swath length and swath width are con-
fully-migrated data volume.
trolled by the survey-area boundaries and tem-
Estimation of the fringe area is considered to
plate width respectively. Based on the
be an important issue in 3D surveys, since it is
fold-computation formula, the fold is function
directly connected with the overall survey cost.
of both of the template movement-overlap and
The percentage cost-increase due to addition of
swath role-aside overlap. Hence the overlap
migration fringe depends on the size of
(measured in terms of number of receiver-stations
survey-area as well as on the fringe width used.
for the template movement and in terms of
Cost saving, as we mentioned above, can be
receiver-lines for swath movement) can be com-
effected by allowing the fringe to be extended
puted once the fold is dened (Fig. 7.32).
into part of the tail-fold zone. According to
The other important parameter related to the
(Sheriff and Geldart 1995, p. 452), the fringe area
swath is the shooting direction, or survey orien-
should be equal to the migration aperture plus the
tation. When there is a distinct dip-trend in the
radius of the rst Fresnel zone.
area, receiver lines are oriented in the direction of
(ii) Template Parameters dominant dip. In practice, this constraint is
The template is made up of a number of relaxed for two reasons: rst, dip value and dip
receiver-lines, shot-lines, number of active chan- direction are generally not constant for the same
nels per receiver-line, and number of source-points area, and second, the dip move-out effect can be
per shot line. A template must be so designed that it taken care of, by dip move-out (DMO) correction
will provide a range of offsets covering a or by pre-stack migration in the processing stage.
162 7 3D Seismic Reflection Surveying
(iv) Bin Dimensions and Bin Attributes formula (Liner 2004, p. 272; Sheriff and Geldart
By denition, the bin is a subsurface rectangular 1995, p. 452):
(or square) area having length and width
Dx v=4f sinh
dimensions. These dimensions are equal to half
of the corresponding surface receiver-station
where (v = f k) is velocity, (f) is the highest
spacing. In processing, seismic traces are sorted
frequency-component, and (h) is the angle of dip
into bin gathers. This means that the bin will
component in the x-direction.
contain a number of traces (equal to the fold of
Derivation of the formula can be done with
coverage) of different offsets, and different az-
the help of Fig. 7.33.
imuths. The centre of the bin marks the CMP at
This formula expresses the relation between
which the stack-trace of the bin-gather is located.
the CMP trace spacing (x) and the three vari-
Thus, the bin area (normally referred to as the bin
ables (v, f, h). It has shown that (x), which is
size) should be small enough to achieve appro-
representing the bin dimension in the x-direction,
priate resolution but not too small to create
is directly proportional to velocity (v) and
empty or low-fold bins. An optimum bin size
inversely proportional to wave frequency (f) and
must, therefore, be determined.
reflector dip (h).
The theoretical basis used for computing the
Typical bin dimensions used in land 3D surveys
optimum bin size is the same as that used for
are (12.525) m in the inline (sub-line) direction
avoiding spatial aliasing in case of 2D common
and (2550) m in the cross-line direction. These
mid-point (CMP) spacing computations. Thus, in
ranges are corresponding to (2550) m and
order to avoid spatial aliasing, at least two sur-
(50100) m of surface sampling intervals (receiver
face samples are required to be present for each
spacing). However, when square and small enough
apparent wavelength. For a given reflector dip,
bins are used, no spatial aliasing occurs. Small bins
the apparent wavelength should be as small as
mean ne subsurface sampling which is necessary
possible in comparison with the CMP spacing.
for increasing structural resolving power.c
From the geometry of the normal incidence
reflection from a dipping reflector and using Bin Attributes
(x = k/2 and t = s/2), a quantitative evalua- The group of traces included in a bin gather, are
tion is obtained for the maximum spatial sam- traces of different offsets recorded by receivers
pling interval (x), expressed by the following positioned at different azimuths. The number of
7.3 3D Field Data Acquisition 163
wave-front at
time, t
Sin =(t v/2) / x dipping
t = /2, = 1/f reflector
hence, x = v/4fsin
traces in the bin-gather represents the fold of Nyquist frequency is 250 Hz. which is expected
coverage for that bin. The parameters (offset, to be far higher than the highest frequency of
azimuth, and fold) form the bin attributes. recorded reflection signals. For economic rea-
In the design process, it is aimed at sons, the recorded data is often re-sampled to
attribute-values that cover as wide coverage-range 4 ms before being inputted to processing. To
as possible, considering the economic constraints. make sure that no aliasing occurs, an anti-alias
Bin attributes are normally displayed in a special lter is applied in the acquisition recording stage
diagram (called spider diagram) in which offsets prior to AD conversion and it is also applied
are drawn as straight lines radiating from the bin prior to any re-sampling process carried out
centre. The line length is proportional to offset during processing work.
value and line bearing represents the azimuth. The sampling period, together with the bin
Number of lines is equal to the fold of coverage dimensions, set the three dimensions of the res-
(Fig. 7.34). olution cell (the voxel) which has the general
In addition to the desired large coverage-ranges form of a parallelepiped or a cube. Its volume is
of attributes, the bin should be small and square to dened by the two horizontal dimensions (x
insure equally-ne sampling in both of the and y), and the vertical dimension (vt/2),
sub-line (inline) and cross-line-directions. where (x and y) are equal to the bin dimen-
sions in the sub-line and cross-line directions,
(v) Sampling Period and Record Length
(t) is the sampling period, and (v) the seismic
The sampling period normally used in the seis-
velocity (Fig. 7.35).
mic recording is 2 ms. With this value, the
XL1 XL2
(a) (b) (c)
SL2
y
x SL1
t
m
apex
circular m
wave-front
diffraction
s hyperbola
7.4 Processing of 3D Data 165
BIN-G.
3. PARAMETER OPTIMIZATION
- Velocity analysis
- Residual-static analysis
- Deconvolution analysis
- Filter analysis
4. POST-STACK PROCESSING
- Zero-phase deconvolution
- S/N Enhancement (f-xy deconvolution)
- Band-pass filtering
- Equalization
- QC displaying
STK-SEGY
166 7 3D Seismic Reflection Surveying
Fig. 7.38 The 3D data volume obtained from processing of a 3D data-set of an area of (10 by 20) km2
Fig. 7.39 Vertical sections (in-line and cross-line) obtained from a 3D data volume
168 7 3D Seismic Reflection Surveying
A seismic eld is created as a result of a sudden nding a way to apply statistical communication
mechanical impact that occurs at a point inside a theory to the seismic problem (Enders Robinson
medium. The created packet of energy (called the 1983, p. 221).
seismic energy pulse) spreads out from the The GAG research activities have resulted in
source zone in all directions. It travels through furnishing a way for applying the principles of
the medium in such a way that energy moves the communication theory in the analysis of
through the particles of the medium without seismic wavelets. In essence, the seismic wavelet
creating any permanent changes to the medium. is considered as a propagating signal which can
Propagation of the generated seismic energy in be processed using principles of the communi-
this way is expressible in terms of a mathematical cation theory in the same way as processing of
equation, the (wave-motion equation). the electromagnetic signal.
In a seismic reflection exploration that uses The GAG scientic ndings in the late 1950s
impulsive energy-sources, the generated seismic have set the cornerstone of modern digital pro-
pulse is generally made up of an oscillating cessing of the seismic signal (the seismic
energy level that diminishes with time and with wavelet).
travel distance, forming what is normally called
the (seismic wavelet). This travelling pulse (the
wavelet) can be recorded by placing a seismic 8.1 Definition of the Seismic Signal
detector (seismometer) in, or on the surface of,
the medium. The seismic wavelet detected near In general, the travelling seismic wavelet (de-
the source location is normally referred to as the tected directly, or after being reflected or refrac-
source wavelet (or source signature), and that ted) is considered to be a travelling signal that
detected after being reflected from an interface, it bears useful information. The seismic reflection
is called the reflection wavelet. signal, normally representing the particle-
In 1952, Massachusetts Institute of Technol- vibration velocity, is a function of space and
ogy, MIT (USA) set up and sponsored a research time. The value of the function at any time is
group, called the Geophysical Analysis Group dened to be the amplitude (or energy) of the
(GAG) which was in 1953 taken over by a signal. Usually, the seismic signal is presented as
consortium of oil and geophysical companies. a mathematical function that expresses amplitude
This group of scientists had the responsibility of variation with time (Fig. 8.1).
Table 8.1 Changing factors of the seismic reflection signal during its sourceto receiver travel-path
Changes in source zone Changes in ray-path zone Changes in receiver zone
Source type of energy source Inelastic attenuation Receiver detection response
Source coupling Geometrical spreading Detector coupling
Near-source geology Reflection coefcient Near-receiver geology
Source-generated noise Wave conversion Noises and interferences
Noises and interferences
8.1 Definition of the Seismic Signal 171
f(t)
F()
t
1
g(t)
G()
t
2
t 1 2
Fig. 8.2 Concept of the frequency spectrum. The mono-frequency functions, f(t) and g(t) have one line spectrum each.
Their sum, f(t) + g(t), have two-line spectrum
A0 time (t)
A2
A( )
a0 a1 a2 a3 a4 a5
A3
1 2 3 4 5
A4
phase spectrum
()
A5
Fig. 8.3 Pictorial representation of spectrum analysis. The signal (at top of gure) is Fourier-transformed into its
frequency components shown as amplitude and phase spectra, A(x) and (x)
This simplied presentation represents a syn- present in the analyzed signal. Two spectral parts
thesis process by summing two sine functions. are necessary to completely dene the complete
The reverse of this process is the spectrum anal- signal characteristics. These are the frequency
ysis which involves (using Fourier transform) spectrum and the phase spectrum. The two con-
determination of all of the frequency components cepts are schematically shown in Fig. 8.3.
of a given signal. The complete analysis is to get The frequency range of the seismic reflection
amplitudes and phase shifts of all frequencies signal is typically (2060) Hz.
172 8 The Seismic Reflection Signal
8.1.3 Comparison with the Radar conditions. The manner and severity of these
Signal changes depend largely on:
Geometrical shape of the traversed layers
Radar signal and seismic signal are similar, in that (structural changes)
they both provide the source-to-receiver distance Physical properties of the layers (stratigraphic
of a far-away object, using the wave-reflection changes)
phenomenon. The source signal used in the case of Fluid content like water, oil, and gas (reser-
radar is typically a short pulse (of a wide voir conditions)
frequency-band), a chirp signal (linearly-changing The seismic wavelet is therefore considered as
frequency signal), or a frequency-modulated a message-bearing record of the changes which
(FM) signal. The radar function is based on two are coded in the wavelet spectrum-parameters
main types of measurements; the travel-time and (amplitude, frequency, and phase). In the
Doppler-effect measurements. The Doppler Effect time-domain language we say that we have
(variation of frequency with target motion) is waveform-deformation whereas in the frequency
made use of, in velocity measurement. The com- domain this is described as spectral-deformation.
mon feature between radar operation and seismic Compared with the source function, and due
reflection technique is that both of them use the to these affecting factors, the detected reflection
reflection travel-time measurement. Naturally, signal becomes lower amplitude, lower fre-
Doppler Effect has no application in seismic quency, extended in shape, noisy, and
reflection operation. time-shifted with respect to its initial origin time.
The main differences between the seismic Mathematically, these changes can be classied
signal and the electromagnetic signal used by into the following four types of processes:
radar are summarized in Table 8.2. superposition, scaling, ltering, and time
This table shows that the parameters of the shifting.
seismic signal are dispersed over a wide range of (i) Superposition
values, compared with those of the radar signal in Interaction of the reflection seismic signal
which the parameter has only one value. This is with noises and other interfering events
because the radar signal is travelling through a (such as coherent and incoherent noise,
medium which is far more homogeneous than refractions and diffractions) is a simple
that traversed by the seismic signal. superposition process. The received signal
is resulting from additions of all of these
interferences, bringing about distortions to
8.1.4 The Seismic Reflection-Signal the wavelet geometrical shape.
Changes (ii) Scaling
This is an arithmetic multiplication pro-
The seismic reflection wavelet is, in fact, the cess, where the reflection wavelet is mul-
source pulse after being modied and distorted as tiplied by a constant scaling factor, similar
a result of the interactions with the travel-path to the effect of gain-application. The effect
8.1 Definition of the Seismic Signal 173
Fig. 8.4 Summary of the - high amplitude (i) superposition - lower amplitude
physical changes of the - high frequency content (ii) scaling - lower frequency content
reflection signal as it travels - sharp shape (iii) filtering - extended
from the source point (S) to - with no noise (iv) time shifting, to - noisy
the detector point (R), via - Onset-time, t=0
- time Shifted, t=t0
the reflection point (RP) S R
source impulse reflection wavelet
V1
s(t) r(t)
t V2 t
0 0 t0
V3
V4
RP
does not deform the signal but only detector offset from the source.
change its size. Examples of this type of Time-shifting becomes a deformation
effect are the effect of the geometrical factor when reflectors are so close to each
spreading and the reflection coefcient other that wavelet-arrivals overlap causing
which are independent of frequency. This distortions to the wavelets and loss of the
means that they incur no wavelet shape resolution of the reflection events. Phase
deformation. A scaling effect can, under shifts are the time shifts of the individual
certain conditions, reverse the wavelet frequency components which, in general,
polarity. This happens when the scalar is take place in frequency-ltering actions.
of negative value, as in negative reflection These changes are summarized in Fig. 8.4.
coefcient.
(iii) Filtering
Frequency ltering can be considered as a 8.2 The Seismic Trace
frequency selective scaling. A typical
example of this type of effect is the earth A seismic pulse incident at an interface is partly
high-cut lter, which is operating on the reflected and partly transmitted into the following
seismic wavelet attenuation throughout its layer. In a multi-layer medium this is repeated at
travel-path. Any effect that is each interface present in the way of the advanc-
frequency-dependent (as the earth lter- ing wave. A detector placed on the surface
ing) incurs shape-deformation to the receives sequentially, the reflected wavelets, at
propagating seismic wavelet. The defor- time-intervals depending on the depths of the
mation severity depends on the type of reflectors. The source pulse reaches the detector
lithology and fluid-content of the geolog- after being affected by the various types of
ical formations traversed by the seismic modications (as discussed above) including the
signal. Being an analogue-type, the earth process of reflection in which amplitude and
ltering effect introduces phase shift to the polarity are governed by the reflection coefcient
frequency components of the seismic of the reflector. From each reflector a wavelet
signal. representing the source wavelet, scaled by the
(iv) Time-Shifting reflection coefcient, will arrive at the detection
The source impulse takes time to reach the point. Thus, a series of such wavelets, each of
detector. This is the total reflection which is shifted in time with respect to the pre-
travel-time which is depending upon the ceding wavelet, are superimposed on each other
depth of the reflector as well as on the to form a record which is the (seismic trace).
174 8 The Seismic Reflection Signal
Reflector-1
1
Reflector-2
2
Reflector-3
3
Reflector-4
4
Reflector-5
5
Fig. 8.5 Sketch of the convolutional model of the seismic trace (Alsadi 1980, p. 192)
Seismic trace
Source function
Earth filter
Reflectivity
Instrument. Resp.
Geom. Spreading.
Noise
Fig. 8.7 The seismic
reflection trace. a Pictorial (a) (b)
representation. b Actual amplitude amplitude
seismic trace from a
0
dynamite source-energy
1
reflection time
2
in
3
seconds
8.2.2 The Synthetic Seismogram The earth layering model is normally obtained
from a drilled exploration-well. The well logging
A direct application of the concept of the con- data furnish both of the velocity (from the sonic
volutional model is in constructing an articial log) and density (from density log). From these
seismic trace, commonly known as the (synthetic two logs, the acoustic impedance of each layer is
seismogram). The basic requirements for this calculated, and hence the reflection coefcients
process are the source wavelet and the reflectivity of the layer boundaries are determined. This is
series, which in turn, require the distribution with the reflectivity series which will be in the form of
depth of the layers acoustic impedances. values, ranging between (1) and (+1),
176 8 The Seismic Reflection Signal
(x) (*)
representing the reflection coefcients of all of with the real geological section. It is normally
the interfaces (layer-boundaries) penetrated by displayed (using several duplicated synthetic
the well. Convolution of the source wavelet traces) superimposed on the seismic stack section
(assumed or obtained from a near-by shot-record) at the well-location. Main purpose is to attach the
with the reflectivity series, gives the synthetic stratigraphic identities for the reflection events
seismogram. This computation method expresses diagnosed on the seismic stack section (Fig. 8.9).
the convolutional model according to which the
real seismic trace is assumed to have taken place
in real time seismic reflection recording. 8.2.3 The Digital Seismic Trace
In display of the computed synthetic seismo-
gram, the time scale must be doubled in order to The seismic trace received at the detection point
match the seismic section in which the time is a is represented as a continuous amplitude varia-
two-way vertical time. A schematic representa- tion with recording time. At this point it is an
tion of the process is shown by Fig. 8.8. analogue function expressing the seismic ampli-
Interpreters are usually provided with syn- tude as function of time. The recording-system
thetic seismograms both with and without multi- transforms the amplitude-versus-time function
ples to help in sorting out the multiple-reflections, from its analogue form to digital form, and gets it
if any. stored on the magnetic tape as a digital function
The synthetic seismogram is now considered (sequence of sample-values at regular time
as an indispensable tool in the hand of the intervals, t), where the sampling period (t) is
interpreters to tie the seismic reflection images normally chosen to be 2 ms (Fig. 8.10).
8.2 The Seismic Trace 177
TWT
With the digital mode, the seismic trace stor- 8.3.1 Definition of the Seismic
age and processing become feasible to carry out Wavelet
by electronic digital computation systems.
A wavelet is dened as a transient signal with a
denite time origin. It is characterized by two
8.3 The Wavelet Concept properties; it has a denite onset time and it has a
nite energy (Robinson 1983, p. 128). These two
The seismic pulse created by a mechanical shock characters implies that the wavelet is one-sided
(seismic source) is a wavelet that is transmitted entity (wavelet values are zero before onset time),
through the earth medium and detected by the and it is transient with its energy diminishing after
receiver geophone-group and then recorded by onset time. Ricker wavelet (named after Norman
the recording system. At the time it is recorded, Ricker 18971980) is a special type of a
the wavelet is no longer a short pulse (source theoretically-computed zero-phase wavelet com-
impulse) as it was created at the source location, monly used in seismic modeling studies. In par-
but an extended wavelet made-up of few cycles, ticular, Ricker wavelet is used in the seismogram
weakened pulse having energy diminishing to synthesis, by convolving it with the reflectivity
zero level within a short time (tens of millisec- function, to produce the now-accepted convolu-
onds) (Fig. 8.11). tional seismic trace model (Robinson 1983, p. 224).
Functions of geophysical signals can be repre- The process of converting an analogue function
sented either as continuous function (analogue into digital form (A-D conversion) is normally
function) in which the function is dened at all of referred to as (digitization or sampling) process.
the points along the abscissa or dened at discrete It involves two-steps; sample denition followed
points, normally regularly spaced values (digital by quantization and coding in which the sample
function). Considering time-domain functions, value is determined and converted into digital
the analogue function is normally quoted as s(t), number. The output of the sampling process is
and the digital function as s(n t), where t is the then stored on a suitable digital storage device or
sample spacing along the time-axis, normally entered in a certain digital data processing
called the sampling period. The two forms of operation. In the language of mathematics, the
functions are shown in Fig. 8.13. A-D process is represented as follows:
t t
0 0
analogue Function, s(t) digital Function, s(n. t)
s(n. t) = 5 3 1 5 3 2 0 7 12
n = 3 2 1 0 1 2 3 4 5
s(n. t) = { 2 5 3 1 3 0 2 5 7 }
s(n. t) = { 2 5 3 1 3 0 2 5 7 }
s(n. t) = { 2 5 3 1 3 0 2 5 7 }
-3 -2 -1 0 1 2 3 4 5 6 7 8
8.5 The Sampling Theorem that sampling which produces at least two sam-
and Aliasing ples per period of the period-component having
the shortest period in that function. The concept
Digitization of a signal implies reading the signal
is shown in Fig. 8.15.
at isolated points which means that the infor-
In the frequency-domain language the theo-
mation of the signal in between these points is
rem statement can be rephrased as, it is possible
permanently lost. The amount of information lost
to completely recover the original analogue
is naturally greater as the sampling period is
function from its digital form if the sampling
larger. In the sampling process, the shortest the
frequency (fs) is greater than twice of the highest
sampling period, the better is the recovery of the
(maximum) frequency (fm) present in the original
details of the function. In fact, the analogue
182 8 The Seismic Reflection Signal
t t
t = /2
fs 2 fm
8.5.2 The Aliasing Phenomenon
There is another term which is intensively There are three categories of sampling; ne
used in connection with signals sampling, called sampling, critical sampling, and coarse sampling,
Folding frequency or Nyquist Frequency, (fN), depending on the magnitude of the Nyquist fre-
which is dened to be half the sampling fre- quency (fN) relative to the cut-off frequency
quency, that is: (maximum frequency component, fm) present in
the original analogue signal. The categories are:
f N f S =2 1=2Dt (i) Fine Sampling (fN > fm)
In this case high sampling frequency is
With the introduction of the Nyquist Fre-
applied where (fN > fm) complying with
quency (also called folding frequency), the con-
the sampling theorem and complete
dition in the sampling theorem can be re-stated as
recovery of the original analogue signal
(For complete recovery of the original analogue
can be obtained, with no distortions.
function from its digital form, the Nyquist fre-
(ii) Critical Sampling (fN = fm)
quency must be greater than the highest fre-
As in the ne-sampling case, no distor-
quency present in the original signal).
tions occur in this case. It is critical, in the
The important conclusion extracted from the
sense that the applied sampling frequency
Sampling Theorem is that, Nyquist frequency,
sets the marginal limit of the Nyquist fre-
used in sampling of an analogue signal, must be
quency below which distortion (frequency
equal or greater than the highest frequency pre-
aliasing) shall take place.
sent in that signal, that is:
8.5 The Sampling Theorem and Aliasing 183
fm
fA
Fig. 8.16 The aliasing phenomenon. Input frequency (fin) = 100 Hz, Sampling frequency = 120 Hz, Nyquist
Frequency = 60 Hz, Aliasing frequency (fA) = 20 Hz
(iii) Coarse Sampling (fN < fm) frequency (fN f). This means that if the input
With coarse sampling (or under-sampling, to the sampling process is (fN + f), the output
as it is sometimes called), those compo- frequency will be (fN f). The two frequencies
nents which are of greater frequencies than (fN + f) and (fN f) are called aliases of each
Nyquist frequency will appear as lower other (Sheriff 1973, p. 4). Thus, for example, a
frequencies in the sampling output. This 100-Hz signal is sampled at sampling frequency
phenomenon, which occurs only when of 120 Hz (fN = 60 Hz), then the output will be:
coarse sampling is applied, that is when 60 (100 60) = 20 Hz, which is the aliased
(fN < fm), is called (aliasing). An example frequency (Fig. 8.16).
of an aliasing case is shown in Fig. 8.16. The general rule governing any sampler
input-output relationship is that any frequency
component higher than the Nyquist frequency
8.5.3 The Aliasing Frequency (fN), present in the signal prior to sampling, is
Computation outputted as alias frequency. The sampling fre-
quency (fS), or the Nyquist frequency fN (=fS/2),
In any sampling process, the frequency compo- is the deciding factor for the possibility of
nents in the input analogue function which are occurrence of aliasing. Cases for the ne, critical,
higher than Nyquist frequency will appear at the and coarse sampling of an input frequency of
sampler output as false components (aliased (125 Hz) are presented in Table 8.3.
components) of frequencies lower than input In Table 8.3 one can notice that, in the rst
frequency. An input frequency component (fin) two cases (belonging to the case of
higher than Nyquist frequency by (f) will ne-sampling), the Nyquist frequency is higher
appear to the sampling system, as the lower than the input frequency and hence no aliasing is
Table 8.3 Cases for ne, critical, and coarse sampling of an input frequency of (125 Hz)
Input fm (Hz) Sampling freq. (Period) Nyquist freq. (Hz) Output freq. (Hz) Output status
1 125 500 Hz (2 ms) 250 125 no aliasing
2 125 333 Hz (3 ms) 166.7 125 no aliasing
3 125 250 Hz (4 ms) 125 125 no aliasing
4 125 200 Hz (5 ms) 100 75 with aliasing
5 125 166.6 Hz (6 ms) 83.3 41.6 with aliasing
6 125 142.8 Hz (7 ms) 71.4 17.8 with aliasing
7 125 125 Hz (8 ms) 62.5 0.0 with aliasing
184 8 The Seismic Reflection Signal
0
input frequency, fin
Fig. 8.17 Input-output relationship, in a sampling process. The zone up to the Nyquist frequency (fN) is the
principal-aliases zone (PA) in which no aliasing occurs
taking place. In the third case (the applied with no aliasing effect. An input fre-
critical-sampling case), Nyquist frequency is quency, lower than the Nyquist frequency, will
equal to the input frequency and again no alias- be outputted with no aliasing effect.
ing occurs. In cases (4, 5, and 6) Nyquist fre- The periodic linear relation of the sampler
quency is less than the input frequency, hence input-output function, shown in Fig. 8.17, can be
aliasing takes place. In the last case (case 7) the used in deriving a general formula that can be
difference (Df) has reached its maximum value used to calculate the output aliased frequency,
which is equal to the Nyquist value giving an given the input frequency and the sampling (or
output of zero-frequency. Nyquist) frequency. With reference to this gure
To give a quantitative measure for the output it can be shown that the output aliased frequency
frequency, use is made of the periodic linear (fout) is related to the input frequency (fin) by the
relation connecting the input frequency (fin) and following relationship:
output frequency (fout), as shown in Kanasewich The integer (k) is calculated from floor
(1973, p. 89). Because of the feature that the truncation of (fin/fN), which is symbolically
spectrum-part above the Nyquist frequency can represented as shown above, that is,
be folded back about the Nyquist frequency, k bf in =f N c
aliasing is sometimes called (folding) and the For application, let us nd the output aliased
aliased frequency, called folding frequency. The frequency for the input frequency, (fin = 782
(fin fout) relation is re-produced in Fig. 8.17. Hz), sampled at 2-ms sampling period, i.e. at
This gure is helpful in determining the sampling frequeny (fS = 1/0.002 = 500 Hz). The
inputoutput relation for any sampling process, Nyquist frequency, fN = 500/2 = 250 Hz, and
once the sampling frequency (fS), and hence, the the integer (k) is obtained from truncation of
Nyquist frequency (fN) is known. In the zone (0 (fin/fN = 782/250 = 3), hence, fout = |782
fN), the output frequency (fout) is the same as [3 + {1 (1)3} 250| = 218 Hz. Similarly,
the input frequency. This region of the spectrum, it is possible to calculate the aliased frequency of
called the principal aliases section, marks the any other input sampled signal having frequency
limits of the sampling frequencies which can be exceeding the Nyquist frequency. As examples
8.5 The Sampling Theorem and Aliasing 185
for this method of calculation, let us work out the 8.5.5 Aliasing in the Frequency
aliased frequencies for the input frequencies Domain
(fin = 457, 557, 957, 1957) Hz, each of which is
sampled at 2 ms sampling period. The output The analogue function can be looked upon as a
aliased frequencies for any one of these fre- digital function sampled at zero-sampling period,
quencies will be (fout = 43) Hz. which will have an innitely-long spectral repe-
tition cycle. This is equivalent to saying that its
spectrum is of non-repetitive nature in case of an
8.5.4 Effect of Sampling on Signal analogue signal. When the analogue is converted
Spectrum into digital form the amplitude spectrum is
repeated over the frequency axis at
The direct effect of the sampling process on the repetition-spacing equal to the sampling fre-
frequency spectrum of a signal is generation of quency (fS). The aliasing problem occurs when
repeated spectrum-replicas of the original ana- the cut-off frequency (fm) of the analogue signal
logue signal. Theoretical computations have exceeds that of the Nyquist frequency (fN), that is
proved that Fourier transform of a digital function when (fm > fN). In this case, the time-domain
is itself periodic with repetition interval equal to aliasing becomes spectrum overlapping in the
the sampling frequency used in the sampling frequency domain with the consequence of
process (Kansewich 1973; Kulhanek 1976). Thus, spectrum distortion due to the resulting
after sampling of an analogue signal, its spectrum spectrum-interference. Aliasing, in the frequency
gets repeated over the frequency axis at regular domain, is, for this reason, called spectrum
spacing which is equal to the sampling frequency contamination (Bth 1974, pp. 148151).
(1/t), where (t) is the sampling period. The important point here is that when the
The lower the sampling frequency the smaller input analogue signal contains frequency com-
the spacing becomes. With too-low sampling ponents higher than the Nyquist frequency
frequency (too large t), the spectrum spacing (fm > fN), these high frequencies will fold over
gets smaller and when Nyquist frequency and add to those frequencies existing below the
becomes less than the cut-off frequency of the Nyquist frequency. The outcome of the process
signal (fN < fm), the repeated spectra overlap and in this case is spectrum distortion in the overlap
spectrum distortion occurs. zone as shown in Fig. 8.19.
The spectrum behavior caused as a result of
digitization is shown in Fig. 8.18.
1/t
digital function repeated spectrum
186 8 The Seismic Reflection Signal
freq. freq.
fN fN
8.5.6 The Remedy for the Aliasing than the Nyquist value, and hence, aliasing is
Effect avoided.
In summary, in order to avoid aliasing, there
To avoid aliasing and its distortion consequences are two alternative methods:
a procedure should be taken such that the con-
(i) Applying higher sampling frequency (small
dition (fN > fm) is restored. To achieve this state,
sampling period) up to the value with which
one can either delete those frequencies which are
the Nyquist frequency becomes greater than
greater than Nyquist frequency prior to the dig-
the cut-off frequency of the input analogue
itization process, or increase the sampling fre-
signal.
quency in such a way as to get a Nyquist
(ii) Applying a high cut lter (the anti-alias
frequency of greater value than the cut-off fre-
lter) to remove those high frequencies
quency of the original analogue signal.
(from the original analogue signal) which
In practice this is done by application of a
are higher than the Nyquist frequency
suitable high-cut lter (usually referred to as
before starting the sampling process. The
anti-alias lter), to the input seismic data prior
two methods are explained in (Fig. 8.20).
to the sampling process. In effect, the applica-
tion of the anti-alias lter makes the cut-off In seismic reflection data-recording, the sam-
frequency of the input analogue signal, lower pling period is normally set at 2 ms value. This
makes the Nyquist frequency to be 250 Hz. This vertical resolution and horizontal resolution (ex-
means that with this Nyquist value, no aliasing plained here-below).
effect is expected since the maximum frequency
(signal cut-off frequency) is expected to be far
below (250 Hz). However, all recording systems 8.6.1 Vertical Resolution of Seismic
are equipped with the anti-aliasing lter that can Signals
be applied when it is required. The anti-aliasing
lter is normally applied whenever re-sampling Vertical resolution of seismic reflection events, is
of the data is carried out. Very often, in seismic dened as the minimum vertical distance
data processing, the recorded digital data is between two interfaces that give two distict
re-sampled from 2 to 4 ms sampling period for reflection events on a seismic section. It is basi-
economic motives. cally governed by the wavelength of the seismic
signal. The shorter the wavelength (i.e. the higher
the frequency) the greater the vertical resolution.
8.6 Signal Resolution In addition to the frequency factor, depth and
and Resolution Power reflector spacing have signicant effects on the
resolution. The vertical resolution is governed by
Resolution is dened as the ability of distin- the ratio of the depth separation-distance of the
guishing individual objects gathered to gather in reflectors (z) to the wavelength (k) of the
one group, or the details of shape changes of an incident seismic signal. The lowest limit (reso-
irregularly shaped object. In the eld of seismic lution limit) of this ratio; (z/ k) is found to be
exploration, the seismic resolution is the ability (1/4) (Sheriff and Geldart 1995, p. 174). This
of recognizing two adjacent seismic events as means that the reflector separation must be more
distinct two events and not as one blurred event. than quarter of the wavelength (z > k/4). This
Resolution power can be measured by the also means that the image separation (events
minimum separation distance between two seis- separation measured on a seismic stack section)
mic events that can be resolved as two distinct should be more than half a period (t > s/2) in
features on the seismic section. Obviously the order to be distinctly resolved (Fig. 8.21).
sharper the reflection wavelet, and higher Adoption of the (k/4) criterion for the resolu-
signal-to-noise ratio, the better the resolution tion limit, implies that the two events reflected
power will be. The term (resolution power) is from two neighboring interfaces are separated by a
used to imply ability of detecting and bringing to half cycle, which means that depth separation (z)
vision a certain seismic event between two neighboring interfaces greater than
In seismic work, we are concerned with two (k/4) will lead to minimum destructive
types of resolutions of seismic reflection events:
z = v. t/2 = v. (/2)/2 = /4
188 8 The Seismic Reflection Signal
interference between the reflected waves from the earth high-cut lter with depth, both are
two interfaces, causing resolution deterioration. leading to increase of the wavelength and
hence decreases the resolution power.
(i) Factors Affecting Resolution
Application of high-cut lters (as
In general the higher the frequency content
application of the anti-aliasing lter),
of the seismic trace the better is the reso-
which lead to attenuation of the high
lution power. Well logs (wireline logs)
frequencies of the signal and hence
have greater resolution power than seismic
result in lowering the resolution power.
traces since well logs are generated by high
(ii) The problem of Thin Beds
frequency sources. These logs can resolve
A special case, related to the subject of
beds on centimeter-meter scale while seis-
resolution which brought appreciable
mic reflection records cannot resolve
attention by geophysicists, is the problem
so-much detailed variations. Reflection
of resolving thin beds. Two reflectors
survey data can resolve reflectors at
spaced by less than quarter of a wave-
depth-separation of about 10 m at its best.
length, have reflection responses depend-
The main factors affecting resolution are
ing on the layering model. A layer is
reflector spacing, reflector depth, and
regarded as a thin layer when its thickness
reflection signal frequency. Closely spaced
is less than a quarter of the dominant
reflectors cause interferences of reflected
wavelength (Sheriff 2002, p. 353).
waves which lead to loss of resolution. The
Consider a thin bed of thickness of (k/4)
resolution-power is generally decreasing
and of velocity (V2), sandwiched between
with depth for the following reasons:
two layers of velocities (V1 and V3) where
Earth lter which is cutting high fre-
(V1 = V3 < V2), the wave reflected from
quencies, that is cutting short wave-
its top and that from its base will interfere
lengths. Thus for depth of 800 m, say,
constructively producing a high-amplitude
velocity of 1000 m/s, and frequency of
reflection, forming what is normally
100 Hz, the wavelength will be 10 m
referred to as the thin-bed effect or tuning
and the resolution becomes only 2.5 m.
effect. If, on the other hand, the thin-bed of
However, when depth is 3000 m,
velocity (V2) found between two layers of
velocity of 4000 m/s and frequency of
velocities (V1 and V3) where (V1 < V2 <
25 Hz, the wavelength will be 160 m
V3), destructive interference will result
and the resolution becomes 40 m. In
(Sheriff and Geldart 1995, p. 174). These
general resolution gets less (poorer
two models are shown in Fig. 8.22.
resolution) with increasing depth due
Vertical resolution is always improved
to the effect of the earth high-cut lter.
with higher seismic frequencies. But, due
Increase of velocity due to compaction
to the earth ltering effect, frequencies get
and decrease of frequency due to the
lower with increase of reflector depth.
(a) (b)
V1 = V3 < V2 V1 < V2 < V3
V1 V1
V2 V2
V3 V3
Fig. 8.22 Two models of a thin bed, having interval velocity (V2). Model A (V1 = V3 < V2) and model B
(V1 < V2 < V3)
8.6 Signal Resolution and Resolution Power 189
Table 8.4 Estimates of the minimum depth interval resolved, corresponding to thre values of sampling period
Sampling period (ms) Highest frequency (Hz) Wavelength (k) (m) Min. depth-Interval (z = k/4) (m)
1 500 5 1.25
2 250 10 2.50
4 125 20 5.00
This table shows that it is always possible to Decimal 105 104 103 102 101 100
devise a new numbering system, but the most system
suitable for us is the decimal system and most Binary system 25 24 23 22 21 20
suitable for computers (computations and stor- Octal system 85 84 83 82 81 80
age) is the binary system. Hexa-decimal 16 5
16 4
16 3
16 2
16 1
160
system
where R is the radix, and the letters (A, B, C, ) 8.7.2.1 Conversion to Decimal System
are the digits of the number. Conversion is done by applying the mathematical
The following table shows the concept sequence that denes the numbering system
applied to the four most commonly applied concerned. Examples of the conversion process
numbering systems: are given in the following examples:
192 8 The Seismic Reflection Signal
47 D 101111 B 57 Oct
and
Examples:
[ 001 001 010 001 . 100 101 ]B = [ 1121.45 ]O
It is to be noted here that there is yet another less-common numbering system called the
binary-coded decimal (BCD) numbering system. In this system, each of the decimal digits is sub-
stituted by the corresponding group of the binary digits. Each of these groups consists of 4 binary
digits, as in the following example:
There is another method to do the subtraction To clarify the process let us take the example of
process, and that is by use of what is called subtraction of the number [+53]D from the number
(Twos Compliment method). The subtraction [+67]D, assuming that the subtraction process is
process starts with converting the number to be done by an 8-bit computer. This is done as follows:
8.7 The Common Numbering Systems 195
[ 67 ]D = [0 1 0 0 0 0 1 1 ]B
[ 53 ]D = [0 0 1 1 0 1 0 1]B
For an 8-bit computer, the digit (1) appearing the decimal numbering (1, 2, 3, ,). Note that after
after the 8th digit will be outside the computer the digit (9) in the decimal system comes (10) and
binary-word limit and so will be neglected and the after the digit (F) in the hexa-decimal system
result will be the 8-digit word (1110) which is [14]D. comes (10), and after the digit (7) in the octal
system comes (10), and after the digit (1) in the
binary system comes (10). This means that, after
8.7.3 Counting in the Different the complete set of digits of each system comes
Numbering Systems (10), then counting continues. The following table
shows this method applied for the hexa-decimal,
Here is a table of different numbering systems decimal, octal, and binary systems.
(hexa-decimal, decimal, octal, and binary). Each
column contains the numbering corresponding to
0 10 20 30 . 0 10 20 30 . 0 10 20 30 . 0 10 100 110 .
1 11 21 31 . 1 11 21 31 . 1 11 21 31 . 1 11 101 111 .
2 12 22 32 . 2 12 22 32 . 2 12 22 32 . .
3 13 23 33 3 13 23 33 3 13 23 33
4 14 24 34 4 14 24 34 4 14 24 34
5 15 25 35 5 15 25 35 5 15 25 35
6 16 26 36 6 16 26 36 6 16 26 36
7 17 27 37 7 17 27 37 7 17 27 37
8 18 28 38 8 18 28 38
9 19 29 39 9 19 29 39
A 1A 2A 3A
B 1B 2B 3B
C 1C 2C 3C
D 1D 2D 3D
E 1E 2E 3E
F 1F 2F 3F
The Seismic Processing Tools
9
From mid 1950s till mid 1960s a jump in the In order to extract the useful message (geological
seismic data processing took place. Within this information) from these recorded data, these
interval, the analogue magnetic-tape recording seismic traces are subjected to a set of analysis
and processing, followed by the introduction of procedures, using certain processing tools. These
digital technique, were introduced. A third and tools are numerical analysis techniques applied
equally important factor that boosted the seismic on the input seismic traces. The most common
data processing is the adoption of concepts bor- tools applied in seismic data processing are:
rowed from the communication theory which
Fourier Analysis
considered the seismic wavelet as a travelling
Correlation functions computations
signal similar to the electromagnetic signal.
Convolution (frequency ltering)
The input data to the processing system con-
Deconvolution (inverse frequency ltering)
sists mainly of the digital seismic traces, recorded
Frequency- and velocity ltering
by the eld recording system. The fundamental
Equalization (trace scaling)
principle on which processing is based upon, is
Sample Editing (time-shifting, sign-changing,
that the seismic trace is considered to be as a
sample zeroing)
digital signal of value (amplitude) which is
function of the reflection travel time. In order to These processes form the basic processing
extract the useful information the seismic signal tools in the hand of the geophysicist, ready to be
(seismic reflection wavelet), a number of math- used in processing of seismic reflection data.
ematical and statistical processes, are applied.
These processes are considered to be the tools 9.1.1 The Sine Function
employed to do the required analyses.
The building brick of any physically realizable
function is the sine function. According to Fourier
9.1 The Seismic Processing Tools theorem, a time- or distance-function, is made up
of a sum of innite number of sine-functions. Due
The raw seismic data acquired in the eld, which to its important role in processing in general, and
form the input to the processing system, are in Fourier analysis in special, an introductory
normally digitally recorded traces on special treatment of the sine function shall be given
magnetic tapes. The reflection wavelets received here-below, preceding the discussions of the
by a surface-positioned detection-system (the processing tools listed above.
geophone group) are digitally recorded as func- One way to generate a sine function is by using
tion of recording time forming the seismic trace. a geometric approach. Let us consider a unit circle
P1 P1 360
P
P2 P0 P0 P2
O R 90 180 270 360 450
P3 P3
of radius equal to unity, in which the radius is relation between the two unit systems; degrees
rotating with constant rotation speed about the and radians:
circle centre. At a given time, the radius (OP in
Fig. 9.1) makes an angle (h) with the x-axis, and 360 2p rad
its tip, point (P) has its projection on the x-axis
represented by point (R). By denition, the sine Given (p = 3.1415927), the relation between
function (sin h) has a value given by the length the two systems is that 1 rad = 57.2958, or
(PR). Now, as the radius rotates about the centre 1 = 0.017453 rad. The argument of sine func-
point (O) the length, PR (=sin h) varies with the tions (h, in this example) is normally quoted in
angle (h). By drawing the value of the PR-length radians.
as function of (h), we obtain the curve for the sine
function (sin h), as it is shown in Fig. 9.1.
9.1.3 Parameters of the Sine
Function
9.1.2 The Degrees-Radians
Relationship The general form of the sine function is that it is
function of angle (h). Other common forms of
Referring to Fig. 9.1, the point (P) is rotating at this function are: function of time s(t) and
constant speed and the generated curve is made function of distance s(x), thus:
up of repeated cycles corresponding to the
repeated rotation cycles of the point P around the sh a sinh
circle circumference. The generated (sin h curve)
repeats regularly every 360. The complete angle or,
swept by the rotating arm is from (h = 0) when P
st a sin2pf t t tp ; f t 1=s
is at location (P0) to (h = 360) when P is back at
location (P0) passing through all of the points or,
(P0, P1, P2 and P3).
It should be noted here that angles can be sx a sin2pf x x xp ; f x 1=k
expressed in radians rather than in degree-units.
An angle (h) in radians is dened as the ratio where, (a) is amplitude, () is phase angle, (ft) is
between the subtending arc divided by the radius temporal frequency, (fx) is spatial frequency, (s)
of the circle in which the angle is central. When is period, (k) is wavelength, (tp) is phase shift in
the arc length is equal to the radius, the angle time units, and (tx) is phase shift in distance
value is 1 rad. Since circumference of a circle is units.
equal to 2p times its radius, the value of the total The maximum value of the sine function is
central angle is 2p. This leads to the important equal to the length of the rotating arm applied for
9.1 The Seismic Processing Tools 199
generating the sine function considered above. arm sweeps a complete central angle (360),
A sine function, represented by a sine curve which is (2p) when expressed in radian units.
(sometimes called sinusoid) is completely When frequency is expressed in terms of rate of
dened by three parameters: amplitude (a), fre- change of the angle (dh/dt), measured in radians
quency (ft or fx) and phase (, tp, or xp). The per second, it is usually called, radian frequency.
amplitude is of positive value (+a) represents a Thus, for constant rate, we can write:
peak and negative value (a) represents a trough
of the given sinusoid. x h=t 2p=s 2pf
The most common form of the sine function is
the form a sin(2pft + tp) quoted as function of where, (x) is the radian frequency (rad/s) and
time (t). The frequency, (f) is expressed in cycles (f) is the cyclic frequency (cycle/s).
per second. The sine function is of innite length The frequency (f) is normally referred to as
extending along the time axis in both of the cyclic frequency, measured in cycles per second, to
negative and positive directions. Part of a sine differentiate it from the radian or angular frequency
function and its three characteristic parameters (x) which is measured in radiance per second.
are shown in Fig. 9.2. Frequency Units
The unit normally used in measuring frequency
is the Hertz (Hz), where 1 Hz is 1 cycle/s.
9.1.4 The Frequency Concept Another closely related measurement-unit
applied in measurement of frequencies is the
Any periodic motion, as the rotation-arm model Octave. Instead of quoting the frequency in its
shown in Fig. 9.1, is specied by two periodicity absolute units (Hertz), it is quoted in frequency
parameters: the frequency and the period of the ratios measured in Octave-units. It is dened as
motion. Frequency is dened to be the number of the number of times (n) with which a frequency
revolutions (cycles) made in one second. The (f2) value gets doubled over a reference fre-
length of the time interval (normally measured in quency (f1). A frequency ratio (f2/f1) is one
seconds or milliseconds) for one cycle, is called Octave if (f2/f1 = 2). In general, the number of
the period. The relation between frequency octaves (n) is related to the ratio (fn/f1) by:
(f) and period (s) is:
f n =f 1 2n
f 1=s n 1=log 2 logf n =f 1
n is number of Octaves
For the rotation arm, shown in Fig. 9.1, the
period (s) is the time interval during which the
slope slope
(in db/octave) (in db/octave)
filter frequency bandwidth
frequency
f1 f2 f3 f4
Fig. 9.3 Slopes of a four-corner spectrum quoted in decibel (db) per octave units, where f2/f1 = f4/f3 = 2
9.1 The Seismic Processing Tools 201
s(t) s(t)
tp tp
t t
0 0
phase lag = tp phase lead, tp
s(t)
s(t)
tp tp
t t
0 0
Phase - 90 + /4
lag
Zero 0 0
phase
Phase + - /4
lead
90
+
180 + /2
Out-of -
phase 180 - /2
202 9 The Seismic Processing Tools
s(t)
s(x)
a a
t x
Fig. 9.6 Denition of the period (s) and wavelength (k) of a sine function in the two domains, s(t) and s(x)
peaks, in a time-domain sine function. Likewise, simplest form, this type of function has the form
the distance separating two peaks in x-domain (Richardson 1953, p. 38):
sine function represents the wavelength (k) as
shown in Fig. 9.6. sx; t a sin 2pt=s x=k
In radians units, we have the temporal relation
(x = 2pft = 2p/s) and the corresponding spatial Using the relationships (x = 2pft = 2p/s) and
relation (k = 2pfx = 2p/k). Another important (k = 2pfx = 2p/k), the function can be written
relation is the one connecting the temporal as:
parameters with the spatial parameters, which is:
sx; t a sinxt k x
v k=s
where k is constant known as the wave number,
It should be noted that when frequency is v is propagation velocity, and x is the radian
quoted as (f) without the subscript (t), it is nor- frequency measured in radians per second.
mally meant to be the temporal frequency (ft). It is important to note that the wave propa-
gation velocity (v) is connected with both of the
temporal frequency (ft) and wavelength (k) by
9.1.7 Propagating Sine Wave the simple relationship:
v f tk
The rotating arm model (shown above), helped in
generating a one dimensional sine function that
or,
depends on one variable which is angle (h) or
time (t). The time-dependent sine function v x=k
describes the case of stationary (non-propagating)
periodic motion. However, we may have a Using a simple pictorial method, the velocity
two-dimensional sine function s(x, t) describing formula (v = ft k) can be derived. Referring to
the case of a propagating sinusoidal wave. In its Fig. 9.7, the number of cycles (ft) of a
9.1 The Seismic Processing Tools 203
v = ft
propagating sine wave measured in 1 s is mul- (iv) The integral of f(t)dt over a complete
tiplied by the cycle length (wavelength, k) gives period is convergent.
the total distance covered in that 1 s, which is the
It so happened that almost all geophysical
propagation velocity (v). Thus (v = ft k).
phenomena, including the seismic signal, obey
Dirichlet conditions and hence can be analyzed
by Fourier Theorem.
9.2 Fourier Analysis and Concept
of Spectra
9.2.1 Fourier Series
In 1807, Joseph Fourier (17681830) presented
his theorem (now identied by his name) which According to Fourier Theorem, any periodic
states that any function such as f(t), satisfying function, f(t) of period (s), satisfying Dirichlet
certain restrictions can be expressed as a sum of conditions, can be represented by the sum of
an innite number of sine waves. In 1809 sinusoidal functions of frequencies which are
Dirichlet formulated the mathematical restrictions multiples of the fundamental frequency of that
(normally referred to as Dirichlet conditions) function. Mathematically, the periodic function
under which the theorem is mathematically valid. f(t) is expressed by the following innite series
The restrictions are: (called Fourier series):
a0 =
an =
bn =
These three expressions are derived by mul- of frequencies (xn) and amplitudes (an and bn).
tiplying both sides of the Fourier series in turn by The term (a0) represents the DC level of the
(1), (cos nxt), and (sin xt), and integrating with function f(t).
respect to (t) over the period length (s). The
Odd and Even Functions
derivation is based on use of the orthogonality
For an odd function fo(t) in which fo(t) = fo(t)
properties of the sine and cosine functions, which
are (for m and n being integers):
The frequencies (xn) that are multiples of the the coefcients (an) become all equal to zero,
fundamental frequency (x), are called harmonic since in this case:
an = = +
=- +
=0
frequencies or just harmonics. Thus a periodic Similarly, it can be shown that, for an even
function f(t) is made up of the sum of an innite function fe(t) in which fe(t) = fe(t), the coef-
number of harmonics (sine and cosine functions) cients (bn) become equal to zero for all (n).
9.2 Fourier Analysis and Concept of Spectra 205
Thus, when the function f(t) is odd, the approximation for the continuous parts of the
Fourier series will consist of sine terms only and function. This kind of distortions (synthesized-
when it is an even function, the series will consist function overshooting- and oscillation-behavior)
of cosine terms only. is normally referred to as Gibbs Phenomenon
(Fig. 9.8).
Application of the Fourier series analysis of
9.2.2 Gibbs Phenomenon periodic functions is found in many standard
texts on the subject, see for example (Alsadi
As we have stated above, a periodic signal (pe- 1980, p. 111).
riod, s), satisfying other Dirichlet conditions, can
be expressed as a sum of sinusoidal components
of frequencies; n/s, (n = 0, 1, 2, 3, ). The 9.2.3 Fourier Transform
zero-frequency component (n = 0) represents the
DC term. This implies that if we have a signal, As it is presented above, Fourier series is
f(t) dened over the time interval (t = 0 to restricted to periodic functions, but theory can be
t = T), then, according to Fourier Theorem, the extended to cover non-periodic functions. Func-
function within this interval can be expressed as tions of nite length are called transient func-
the sum of the frequency components (n/T) tions. With special integrals, it is possible to
where (n = 0, 1, 2, 3, , ). transform a transient function f(t) to another form
Fourier Theorem states that the signal f(t) will in which it becomes function of frequency F(x),
be recovered exactly from the Fourier series only with no loss of information. The process of
when an innite number of terms are included in converting f(t) into F(x) is called (Fourier
the summation, that is when the integer (n) runs transform) and the reverse process, F(x) into f(t),
from (n = 1) to (n = ). In practice, only a is called (inverse Fourier transform).
nite number (n = N, say) of terms is used in the The transient function can be considered as a
summation. As a result of this truncation of the periodic function of innitely long period. The
series (i.e. when n is nite number), distortion process of transforming a transient function from
shall occur in the recovered (synthesized) signal. its time domain f(t) to the frequency domain
In this case, the sum will shoot beyond the value function F(x) and vice versa, are done through a
of f(t) in the neighborhood of the discontinuities set of integral equations called (Fourier Integral
found in f(t). The overshoot oscillates about the equations) which are more commonly known as
function value with decreasing amplitude as we Fourier Transform equations. The two equations
move away from the discontinuity. Increasing (Transform and Inverse Fourier transform equa-
the number of terms leads to a better tions) are Bth (1974, pp. 3337):
f(t)
f(t)
t t
(few-terms) approximation (more-terms)
Fig. 9.8 Gibbs Phenomenon computed for a rectangular pulse for few- and more-terms truncation of Fourier series
206 9 The Seismic Processing Tools
(Fourier Transform)
In general, the function F(x) is complex, similar to Light spectrum in which the Light
consisting of a real part a(x) and imaginary part frequency-components appear when analyzed as
b(x). Thus, F(x) can be expressed as: it passes through a transparent prism. In the
The two functions, a(x) and b(x), known mathematical sense, F(x) is complex function
respectively as cosine and sine transforms, are consisting of the modulus; jFxj called (am-
dened as: plitude spectrum) and the argument (x), called
or.
F() = A()ei()
where,
Z1 p
jFxj ax2 bx2
bx f t sin xt dt
x tan1 bx=ax
1
represents the Fourier spectrum of the analyzed phase-spectra. Through another form of integra-
function. tion (Inverse Fourier transform), it is possible to
recover the original time-domain function from
9.3.1 The Line Spectrum the transformed frequency-domain function. For
complete recovery, this process requires both of
Based on the nature of the input signal (periodic or the amplitude spectrum and phase spectrum. The
non periodic), the computed spectrum is obtained concept of frequency spectra (amplitude- and
as discrete points or as continuous curve. Fre- phase-spectra) computed by Fourier transform
quency analysis of a periodic function involves equations are illustrated pictorially in Fig. 9.10.
determination of the Fourier series coefcients of In summary, it can be stated that
the input function. In this case the output Fourier series analysis of periodic functions
frequency-domain function is plotted as discrete yields line spectra and Fourier transform analysis
lines along the frequency axis forming what is of non-periodic functions yields continuous
commonly referred to as the (line spectrum). spectra.
The line spectrum S1(x) of a sine function
s1(t) = a sinx1t, is represented by one line of
height proportional to the amplitude (a) of the 9.3.3 The Fourier Power Spectrum
sine function, located at frequency (x1). For an
another sine function, s2(t) = a sinx2t of fre- The total energy (E) of a real function f(t), such
quency (x2), will, likewise, be of one line placed as a seismic signal, is generally taken to be
at frequency (x2). Further, the sum of the two proportional to the integral of the square of its
sine functions s1(t) and s2(t) will have a line amplitude. That is:
spectrum made up of two lines placed at (x1) and
Z1
(x2). The principle is shown in Fig. 9.9.
E ff tg2 dt
1
9.3.2 The Continuous Spectrum
It can be shown (Bth 1974, p. 82) that this
Analysis of a non-periodic signal, by the Fourier expression is related to the power spectrum as
transform, gives continuous amplitude- and follows:
s2(t)
S2( )
t
2
s1(t) S1( )
+ +
s2(t) S2( )
t
1 2
208 9 The Seismic Processing Tools
a1
a2 A()
a3 a0 a1 a2 a3 a4 a5
1 2 3 5
a4
() phase spectrum, ()
a5
Fourier Pair
f(t) F()
p(t) 2a
2a
t
-a 0 +a -2 /a - /a 0 /a 2 /a
Fig. 9.12 The window pair of the rectangular-pulse, box-car time function, pa(t) and its spectrum, sinc-type function,
P(x) = (2sin ax)/x
210 9 The Seismic Processing Tools
In the geophysical literature, this function is pulse rectangle gets narrower, the corresponding
known as the (sinc) function. The function (sinc amplitude spectrum gets wider, and in the lim-
x) is dened as (sinx)/x or (sinpx)/px (Sheriff iting state, the pulse becomes spike-function
2002, p. 320). Fourier transform P(x) of the (impulse function) with innitely wide spectrum
rectangular function (normally called box-car (Fig. 9.13).
function, p(t) is the sinc function, 2a sinc (ax/p). Another important conclusion can be drawn
from this behavior, and that is the spike time
function contains innite number of equal-
9.3.6 Spectrum of the Spike Pulse amplitude frequency components. For this rea-
son the spectrum of a spike signal is often
Mathematical analysis proved that there is an described as being white spectrum.
inverse relationship between the width of the
time-domain function (as a seismic pulse) and its
frequency domain amplitude spectrum. Thus, as 9.3.7 The Dirac-Delta Function
the pulse gets narrower, the corresponding
frequency-domain spectrum becomes wider, and The zero-width pulse (impulse) is called (Dirac
vice versa. This behavior can be readily seen in delta function) and given the symbol d(t). This
case of the rectangular pulse-spectrum Fourier special function is dened as being a compressed
pair. The main lobe of its spectrum extends from rectangular pulse dened to be of width,
(x = p/a) to (x = +p/a). This clearly indicates approaching zero and having a unit area (Bth
that the spectrum has inverse relation to the 1974, p. 52). It is considered as being even
original pulse width (2a). In fact, this is one of function and hence it has a real spectrum func-
the important properties that hold for all Fourier tion which is constant at the value of unity for all
pairs of pulse-shaped functions. frequencies. This means that the Delta function
According to this principle (called the can be synthesized by superposing an innite
reciprocity property) it is readily seen that as the number of sine functions which are in phase at
f(t) F()
t
0 0
f(t)
F()
)
0 t 0
9.3 Concept of the Frequency Spectrum 211
+
+
t=0
one point (at t = 0) where the components will The signal length (T, say) denes the lowest, or
add up constructively, and destructively else- fundamental frequency (fL = 1/T) in the spectrum.
where as shown in Fig. 9.14. Also, it denes the frequency increment, f (= 1/T)
Dirac delta function, d(t) is not a proper of the spectrum computation. The digitization
mathematical function. It is usually considered as interval, or sampling period (t), on the other hand,
a mathematical concept which possesses its own sets the upper frequency limit of the spectrum. In
mathematical properties. It is dened as: fact, the upper limit is the Nyquist frequency (fN)
which is equal to (1/2t). With consideration of
Z1 these limits, spectrum of an observed signal, of
dt 0 for t 6 0 and dtdt 1 nite length (T) and digitized at (t) interval, is
1 usually computed for the frequencies:
The delta function is even, d(t) = d(t), and it 1=T; 2=T; 3=T; . . .; 1=2Dt
forms the Fourier pair (d(t) $ 1). Further, we have
another important property which states that its Thus, a digital signal made up of (N + 1)
convolution with a function leaves the convolved samples, will be of length (T) given by
function unchanged, that is: d(t) * f(t) = f(t). (T = Nt). By substituting (T/N) for (t), com-
putation frequencies will be:
Finally, it should be emphasized that the length These are expressed in the frequency domain by
(T), controls the (resolution power) of the com- the amplitude- and phase-spectra as shown in
puted spectrum, since the frequency interval f Fig. 9.15.
(= 1/T) is inversely proportional to the length (T). In Fig. 9.15, signal-1 has negative time-phase
(positive phase-angle), while signal-2 is of
zero-phase. The rest of frequency components
9.4 The Phase Spectrum are of negative phase angles. These amplitude
and phase characteristics are shown in the cor-
From the previous paragraph, we learned that the responding frequency domain spectra.
frequency spectrum F(x), of any real time
function is generally a complex function, con-
sisting of the modulus; jFxj called (amplitude 9.4.1 The Zero-Phase Spectrum
spectrum) and the argument (x), called (phase
spectrum). The importance of the phase charac- If all the frequency components are of zero phase
teristics of a seismic signal is that it has direct (symmetrical about time zero), their sum will
influence on the signal shape. Signal amplitude, give a zero-phase wavelet, symmetrical about
on the other hand has a direct effect on the signal zero time. The phase spectrum, in this case, is a
energy. The phase of a frequency component is straight line drawn along the frequency axis at
measured by the time shift of a peak of a cycle zero phase-value. Amplitude and the phase
with respect to time zero, considered to be the spectrum for this case are shown in Fig. 9.16.
start time. Thus, the start time is the time relative Note that the amplitude spectra in the two
to which all phases of the frequency components Figs. 9.15 and 9.16 are the same, but their
are referred to. In general, the phase value may respective time domain signal-shapes are differ-
be negative, zero or positive. ent. This difference in shape is due only to their
Seismic signals, like seismic reflection wave- different phase spectra. Another useful note is
lets, are all of the type of signals which are that the recorded seismic reflection wavelet is
existing only in the positive side of the origin one-sided signal. It cannot be zero-phase since,
time. Such functions (called one-sided functions) by denition, it has no energy before time zero as
consist of the sum of frequency components of one-sided signal demands. However, conversion
different amplitudes and different phase shifts. into zero-phase wavelet can be done by
amplitude spectrum
1 t
2
1 2 3 4
3
+
amplitude spectrum
1 t
2
1 2 3 4
3
+
4
-
zero-phase spectrum
sum=
t=0
Fig. 9.16 Set of frequency components (sine functions) of a zero-phase signal, with its amplitude and phase spectra
application of special computer programs. This frequency. In this case, each frequency compo-
can be done in processing stage to convert the nent will be shifted by the same time shift
one-sided wavelet to a zero-phase wavelet of (Sheriff and Geldart 1995, p. 533). The sum, in
symmetrical shape. This is normally asked for by this case, will be like the zero phase spectrum
interpreters for more clear denition of reflection (symmetrical wavelet), but the symmetry will be
events and easier to follow in the process of about a time later than zero-time. The phase
interpretation especially for structural purposes. spectrum will be a straight line inclined to the
frequency axis (Fig. 9.17).
A closely associated case is the linear phase The constant phase spectrum is a special case of
spectrum, in which the phase is linear function of linear phase spectrum. The shift is not function of
amplitude spectrum
1 t
2 1 2 3 4
+
3
-
sum = linear-phase spectrum
t=0
Fig. 9.17 Set of frequency components (sine functions) of a linear-phase signal, with amplitude and phase spectra
214 9 The Seismic Processing Tools
frequency but having a constant value. Thus spectrum in this case is straight line parallel to
when the phase value is constant at zero-value, the frequency axis and located at (p) below (or
we get a zero-phase spectrum as shown in above) the frequency axis. The (p) or (+p)
Fig. 9.16. If the constant shift is a non-zero phase of a component implies that it has a trough
value, the phase spectrum is also a straight line, at time zero. The wavelet resulting from sum-
but it is parallel to the frequency axis. For ming such components will be symmetrical and
example if the constant phase shift is at later-than will have a trough at time zero as shown in
the time zero by quarter a period (s/4, say), Fig. 9.19.
which is equivalent to (p/2), the phase spectrum
will be a straight line located at (p/2) below the
frequency axis. In this case, the frequency com- 9.5 Spectra of Observational Data
ponents will all have the constant phase of (p/2)
and their sum will give an inverted-symmetry As far as Fourier analysis is concerned, an
wavelet that has its main peak immediately fol- observational function differs from an analytic
lowing the time zero (Fig. 9.18). function in two main aspects. The observed
If the phases of the frequency components are function is not innite and not continuous as an
all equal to (+p/2) instead of (p/2), the phase analytic function would be. Thus, a signal,
spectrum will be a straight line at (p/2) above the (normally, time function), inputted to the trans-
frequency axis. In this case, all the components formation process, is nite in length and digital
will have troughs immediately following the time in form. Observational functions, which are input
zero, and the wavelet resulting from summing the for spectral analysis, are usually not continuous
components will, likewise, have a trough fol- and not innite in length as Fourier Transform
lowing time zero. integral demands. For this reason, observational
The last interesting constant-phase case is a spectra suffer from the two types of distortion
wavelet having all its component frequencies of which are resulting from signal-truncation and
equal phases, at the constant phase of (p), signal-digitization. These effects shall be dis-
which is the same as (+p). The constant-phase cussed as follows:
amplitude spectrum
1 t
2 1 2 3 4
3
+
4
-
constant-phase spectrum
sum =
t=0
Fig. 9.18 Set of frequency components (sine functions) of a constant-phase (at phase = p/2), with amplitude and
phase spectra
9.5 Spectra of Observational Data 215
amplitude spectrum
1 t
2
1 2 3 4
3
+
4
-
sum =
constant-phase spectrum
t=0
Fig. 9.19 Set of frequency components (sine functions) of a constant-phase (at phase = p, or at phase = +p), with
amplitude and phase spectra
This shows that signal truncation incurs a 9.5.2 The Rectangular Window
smoothing effect due to convolution of the true (Box-Car)
spectrum F(x) with the spectrum of the window
function, W(x). In addition of the smoothing In order to get the least possible spectrum dis-
effect side lobes are also created. So, the spectrum tortion, the time interval of the window function
of a truncated signal is distorted, and the nature of must be as long and smooth (free of sharp cor-
distortion is the smoothing effect and develop- ners) as it can be (Bth 1974, pp. 155171).
ment of side lobes as shown by Fig. 9.20. Many types of window functions have been
Since an innitely long rectangular pulse has developed for application in this eld (8 window
a spectrum approaching an impulse function, it types were presented by Sheriff 2002, p. 397).
time-domain frequency-domain
truncation process convolution process
f(t) . w(t) F( ) * W( )
- R( ) * F( )
Fig. 9.20 Effect of varying the truncation window width (T) on smoothing of the spectrum F(x) of the truncated signal
f(t). Smoothing is more severe with shorter length, T
9.5 Spectra of Observational Data 217
t - /T /T
-T 0 T - 0
t
-T T - 2 /T 2 /T
Fig. 9.21 The two time windows in common use in seismic data processing and their spectra. a The triangular
time-function, or Box-car window and b the triangular time-function, or Bartlett window
218 9 The Seismic Processing Tools
This denition can be represented in the fol- where, () represents correlation process, (s) is
lowing symbolical way: called the time lag.
f1 ( i ) = 2 2 1 -1
f2 ( i ) = 2 1 0 -1
2 1 0 -1
2 1 0 -1
2 1 0 -1
2 1 0 -1
2 1 0 -1
2 1 0 -1
C12 ( j ) = -2 -2 1 7 5 1 -2
f1 f2 C12
f1 ( i ) = 2 2 1 -1
2 2 1 -1
2 2 1 -1
2 2 1 -1
2 2 1 -1
2 2 1 -1
2 2 1 -1
C11 ( j ) = -2 0 5 10 5 0 -2
f1 f1 C11
In practice, it is a shift, multiply, and sum maxima at a period equal to the periodicity of the
process as it is done in computing the hidden periodic signal if it is existing in the orig-
cross-correlation function (Fig. 9.23). inal function. Most intensive use of the autocorre-
The function C11(s) is even function, lation functions is in computing power spectra and
C11(s) = C11(s), that is, symmetrical about the in the deconvolution computations.
point (s = 0).
It is useful to note here that the autocorrelation function in the time domain is transformed into
power spectrum in the frequency domain. It is also important to note that the phase spectrum, 11(x),
is equal to zero. This provides an alternative method which allows computing power spectra from the
autocorrelation function instead of squaring the amplitude spectrum.
9.7 Convolution
Like correlation functions, Convolution is a mathematical process taking place between two functions
to output a new function the convolution output. Convolution of the two functions; f(t) and h(t) is
dened by the integral:
yj = fi * hi = fi . hj-i
Computation of the convolution process is done in the same way as in the correlation process
(described above) except for the need to reverse one of the two functions involved in the process.
Here is an example to clarify the process of convolving fi with hi where fi = { 2 1 3 0 1} and
hi = { 3 1 2}. The convolution process starts with reversing one of the two functions and then
continues with shifting, multiplication, and summing processes, shown as follows:
f( i ) = 2 1 3 0 -1
h(i)= 2 1 3
2 1 3
2 1 3
2 1 3
2 1 3
2 1 3
2 1 3
y (i) = 6 5 14 5 3 -1 -2
There are several other methods to compute convolution. These methods involve use of: special
tables, special matrix, and by z-transform. Methods of computation and mathematical properties of
the convolution process are summarized in Alsadi (1980, p. 8890).
The process involved in convolution (shift-multiply-and-sum process) is shown graphically in
Fig. 9.24, in which the function f(t) is convolved with the function h(t).
9.7 Convolution 221
t t
0 0
f(t) F()
h(t) H()
y(t) Y()
process that can result in computing one of two this reason a special statistical method is
convolved functions, given the convolution out- employed.
put together with one of the two input functions. In the frequency domain, convolution
Thus, for a convolution process f(t) * h(t) = y(t), between two time functions, become multiplica-
deconvolution is the process of nding the tion of their respective spectra, and using the
function f(t), say, given the output function y(t) Fourier pair dt $ 1, we get:
and the second input function h(t), that is:
Hx Gx 1
yt ft ht Gx 1=Hx
Deconvolution is the process of nding a Thus, in frequency domain, the relation between
function, g(t), called deconvolution operator convolution operator h(t) and deconvolution-
which, by convolution with the output function y operator g(t), becomes simple arithmetic multipli-
(t), will recover the original input function f(t), cation process.
that is:
RC-2
reflection
time
RC-3
repeated reflections), which take place during the Derivation of the spiking deconvolution
seismic reflection journey. These changes are operator is based on a statistical approach. It
produced by convolution processes. Deconvolu- depends on a principle called Least-Square
tion is, therefore, the process that can unravel Wieners Optimality (Robinson and Treitel
these convolution processes and recover the 1967). The operator wavelet (inverse lter) is
undistorted seismic trace. After deconvolution, derived from the autocorrelation function of the
the recorded reflection wavelets become well input trace. By convolving the derived operator
separated spike-shaped wavelets and free of with the input trace, the spiking deconvolution is
reverberations. obtained.
Referring to Fig. 9.26, consider an input
seismic-trace record containing primary reflec-
9.8.2 Spiking (Whitening) tions (a1, a2, a3). From the autocorrelation
Deconvolution function the deconvolution operator is derived
and, on application (convolving it with the input
Based on the type of wavelet-distortion resulting trace), will output all the reflection events col-
from convolution processes affecting the source lapsed into spike-shaped wavelets. Application
wavelet during its travel, there are two most of this type of operator (spiking deconvolution
commonly used types of deconvolution. These operator), will output the spiky primary
are the spiking deconvolution and the gapped reflections.
deconvolution (explained as follows). In the frequency domain (lower part of
In the time-domain, this type of deconvolution Fig. 9.26), the whitening deconvolution operator
is described, as spiking deconvolution, because it is designed such that at each frequency, the value
produces compressed spike-like reflection of the operator spectrum is the inverse of the
wavelets. In the frequency domain, it is whiten- value of the trace spectrum, at that frequency. In
ing because its output spectrum is flat (white other words, the product of the two spectrum-
spectrum) over wide frequency range of the data. values (input spectrum and its inverse) at each
Spiking deconvolution counteracts the high-cut frequency is equal to unity. This mathematical
ltering effect (convolution process) that the process produces a frequency-spectrum of con-
source wavelet experiences due to inelastic stant value along the frequency axis, i.e. pro-
attenuation. ducing white spectrum output.
224 9 The Seismic Processing Tools
a3 a3
A closely related type of deconvolution, is the whitening deconvolution. The derived operator,
so-called (signature deconvolution) or (designa- in this case, consists of an isolated spike at its
ture). This is derived by a statistical method. The rst point followed by a series of zero
procedure is computing the average spectrum of sample-values (the gap) and then the operator
the shot-gather traces, then, from the inverse of series of samples.
this average, the time-domain deconvolution- Referring to Fig. 9.27, consider an input
operator is derived and convolved with the shot seismic-trace record containing primary reflec-
traces. This approach gives better estimate of the tions (a1, a2, a3) and reverberation reflections
source wavelet, because of the averaging (b1, b2, b3). From the autocorrelation function
process. the deconvolution operator is computed and then
applied (convolved with) to produce the output
trace with the reverberations removed.
9.8.3 Gapped Deconvolution With this form of deconvolution, the operator
(called Prediction Error operator or Prediction
As we have mentioned before, the main distor- Error lter) will effectively predict the reverber-
tions that occur to the seismic trace is signal ation onsets on the seismic trace and subtract
deformation (inelastic attenuation effect) and them leaving only the unpredictable events of
wavelet repetition (reverberation). Whitening random distribution which are the primary
deconvolution restores the wavelet original reflections and random noise.
shape, but the reverberation is removed by the In the frequency domain reverberations,
other non-whitening type of deconvolution, which are appearing on the input trace, cause
known as (Gapped or Predictive Deconvolution). corresponding undulations to its amplitude
Reverberation reflections appearing on a spectrum. The aim of the gapped deconvolution
seismic trace are exhibited on the autocorrelation is to remove reverberation events. This is
function as side-lobes separated by an interval achieved through removing these spectrum-
equal to the reverberation period. The deconvo- undulations by smoothing of the input spec-
lution operator is derived from the autocorrela- trum. The deconvolution operator in this case
tion function of the trace just as in the case of (lter having a gap-length equal to the
9.8 Deconvolution 225
a3 a3
b2 b2
b3 b3
(x) (=)
reverberation period) does not whiten the wavelet deconvolution, the operator in the frequency
spectrum (unless the gap is very short). It domain, is the inverse of the total spectrum of the
smoothes the spectrum by removing the ripples data which includes both of the signal and noise.
caused by the reverberation. This is shown at the Ideally, the spiking deconvolution output is a
bottom of Fig. 9.27. spike-shaped reflection signals in the time
Spiking deconvolution is considered as a domain and flat spectrum (white spectrum) in the
special case of the gapped deconvolution in frequency domain (Fig. 9.28).
which the gap-length is only one sample-period. Considering the frequency-domain deconvo-
Because the noise existing within the spectrum lution, noise gets amplied at the spectrum parts
width was not spectrum-equalized (i.e. not whi- where the S/N ratio is low. The noise amplitude is
tened), the output trace is not as noisy as the thus, amplied in the frequencies outside the sig-
spiking deconvolution. nal main frequency-band. It is usually observed
that seismic data, outputted from whitening
9.8.4 Noise Behaviour Under deconvolution, contain increased noise level.
Deconvolution In order to avoid the phenomenon of noise
development, there are two alternative options.
(i) Case of Whitening Deconvolution The rst option is to remove noise before
A normal seismic trace contains coherent events derivation of the operator as by using the shaped
such as primary reflections, multi-reflections, deconvolution, in which the autocorrelation
coherent and random noise. The general charac- function is ltered with a band-reject lter before
teristics of random noise (also called white noise) being used in the derivation of the operator. In
is that it has a white spectrum, which implies that this way, the whitening process is conned to
all its frequency components are of equal that part of the spectrum where the
amplitude. In the case of whitening signal-to-noise ratio is high, and hence less noisy
226 9 The Seismic Processing Tools
output trace will be obtained. This type of A closely related issue is the often needed
deconvolution is called shaped (or band-limited) practice of adding a small percentage of white
deconvolution. noise to the input amplitude spectrum before
The second option of avoiding noise devel- computing of the spiking deconvolution opera-
opment is application of a band-pass lter to the tor. In the frequency domain, the operator is
deconvolved data. This option is the more often derived from multiplication of the input spectrum
followed in normal processing-work. by its inverse. This is equivalent to dividing
spectrum by itself for each frequency component.
(ii) Case of Gapped Deconvolution
If a frequency component happened to be of
The process of smoothing out the ripples caused
near-zero amplitude, instability problem occurs.
by the reverberations in the amplitude spectrum
Thus, to avoid division by near-zero value, a
would theoretically have no effect on the S/N
constant value is usually added to the input
ratio (Fig. 9.28). The signal-to-noise character-
amplitude spectrum. In the time domain, this is
istics of the input spectrum remained unchanged
equivalent to adding a constant to the autocor-
under gapped deconvolution. In general,
relation function at zero-lag. The addition pro-
gapped-deconvolution output appears to be less
cess is usually done by multiplying the zero-lag
noisy than the data outputted from whitening
autocorrelation value by a chosen scaling con-
deconvolution.
stant. Typical values of this factor is within a
The role of the gap-length in gapped decon-
small range of 1.011.05.
volution is that the shorter the gap, the more
whitening effect occurs but with more develop-
ment of whitened random noise. The opposite 9.9 Frequency Filtering
effect happens when the gap-length increases.
Thus, with increased gap length, noise level is 9.9.1 Definition of the Filtering
reduced but whitening effect is decreased, lead- Process
ing to deterioration of resolution of the reflection
wavelets. In view of this discussion, it appears Coloured glass objects are lters which pass
that a compromise is needed to be made between certain components of the spectrum of white
resolution and the noise level resulting from light on basis of wavelength. Likewise, seismic
whitening deconvolution. This can be controlled detectors (as geophones) are lters which pass
by the choice of the optimum gap-length of the components of a seismic signal on basis of fre-
applied gapped deconvolution. quency. Another type of ltering is velocity
signal signal
noise t noise
t
freq. freq.
ltering, which does the separation of compo- dependant scaling process. The principle of l-
nents on the basis of propagation phase velocity. tering of a seismic signal is shown in Fig. 9.30.
The frequency-ltering is a sorting process The ltering system may be of analogue
that separates frequency components on basis of mode, as in the earth ltering of seismic signals,
frequency. These lters may be natural or arti- or digital mode as is done in digital recording and
cial. A propagating signal in nature (electrical, processing of the digitally recorded signals.
electromagnetic, or seismic) suffers from the l-
tering effect. The earth, for example, acts as a
lter that passes the low-frequency components 9.9.2 Role and Objectives
of the propagating seismic wave. Articial fre- of Filtering in Seismic
quency ltering, by analogue or by digital Exploration
instruments, can be made to separate selected
frequency components of the signal. Frequency ltering plays an important role in the
Frequency ltering is a convolution process eld of seismic exploration. To start with, the
that can be done by mathematical computation. It source wavelet experiences low-pass ltering by
is considered to be an input-output system from the earth during its source-to-receiver propaga-
which a ltered function, g(t), is obtained from tion. In normal seismic exploration, it is noticed
an input function, f(t) as shown in Fig. 9.29. that the recorded reflection wavelets, and due to
The frequency-ltering of seismic signals is a earth ltering, gets lower in frequency content
process that allows certain frequency components with increase of reflector depth. At about
of the input signal to pass and masking or 4 km-depth, the amplitude of a 40 Hz
weakening rest of components. In this way, fre- reflection-wavelet is one-third, of that of the
quency ltering can be viewed as frequency- 20 Hz, while that for a 100 Hz wavelet is about
input output
frequency components frequency components
1
1
2 2
3 3
4 4
228 9 The Seismic Processing Tools
one fortieth (McQulillin et al. 1984, p. 38). For in order to avoid the occurrence of aliasing in the
this reason, seismic resolution deteriorates with digitized data.
increasing reflector-depth. The Notch lter is another reject-lter which
Seismic detectors impose ltering effects on can reject a dened narrow band of frequencies.
the received seismic energy (signals and noises), Recording systems normally include an analogue
causing changes of form and severity decided by Notch lter for the special purpose of removing
their own response characteristics. electrically-induced mono-frequency interfering
The recording system is normally equipped signals as those emitted by high-tension power
with a high-cut type of lter (the anti-alias lter) to lines. Typically, the power-line frequency is
cut out those frequencies higher than the Nyquist 50 Hz, in certain countries and 60 Hz in others.
frequency to avoid aliasing. All these lters are of Both of these lters (anti-alias lter and Notch
analogue type applied to the seismic wavelet lter) are also applicable in digital mode. Thus,
before being converted to the digital mode. One in processing, a digital anti-alias lter is applied
property which is common to all analogue lters is when re-sampling of the data is done. Likewise, a
that they all have non-zero phase spectra. This digital notch lter may be applied to the digital
implies that each frequency component of the data if a mono-frequency noise is detected.
signal which is inputted to an analogue lter shall
experience its own time-delay. As a consequence
of this phenomenon, all analogue lters are 9.9.3 Concept of the Linear
expected to incur waveform distortions of the l- System
tered seismic signal. The role offrequency ltering
in the seismic reflection exploration is shown A linear ltering system is dened as that system
(Fig. 9.31). whose output is linearly dependent on its input.
Recording systems are usually equipped with Linearity is also called superposition principle.
special analogue lters to be applied to seismic Thus, for a linear system, the relation between
data before being converted to digital mode. The input function f(t) and output, g(t) is that,
two main reject analogue lters are the anti-alias for inputs [c1 f1(t) + c2 f2(t)] we get an output
lter and the Notch lter. The anti-alias lter is [c1 g1(t) + c2 g2(t)]. The linearity property allows
of high-cut type of lter that can remove those flexibility of positioning of the ltering process
frequencies which are higher than the Nyquist in the seismic processing sequence. Thus, in
frequency. This lter is normally put in operation principle, the net effect is the same whether the
f1
travel path
reflector
9.9 Frequency Filtering 229
ltering is done on seismic traces before being capability is function of the physical properties of
stacked or on the stacked trace after stacking the ltering device, if it is analogue. It is function
(Fig. 9.32). of the parameters of the designed ltering func-
There is another closely related concept, tion if it is digital. The efciency of a lter in
which is, in fact another property of ltering ltering action is expressed by the lter response
systems. This is the timeinvariance property. characteristics for the particular ltering system.
A system is time invariant (stationary) when One of the more common ways to determine
its response is independent of time. In other the lter response is done by inputting an im-
words, for a certain input f(t ts) we get an pulse function. The output of the ltering system
output g(t ts) for any time shift (ts). The two from inputting an impulse function is considered
properties, linearity and being stationary, are as measure for the ltering capability of that
independent of each other. system. This is a logical decision, because the
It is important to note that all frequency lters, impulse function contains the innite frequency
including natural analogue lters and the range. In fact the spectrum of the impulse is real
articially-applied digital lters, applied in data and innite in extent, and thus the output gives
processing, are of linear, time invariant systems. an ideal expression for the system ltering ef-
ciency. The impulse response of a ltering sys-
tem (Fig. 9.33) is the output of the system for an
9.9.4 The Filter Impulse Response impulse-function input.
The impulse response function of a linear
Filters differ in their ltering capabilities. Theo- system is proportional to the amplitude of the
retically, the ltering capability of a ltering input impulse function. When the input impulse
system ranges from the extreme case where it is is delayed with respect to the origin time, the
passing the input signal with no change (all-pass response function is delayed by the same
ltering) to all-cut or no-pass case. The ltering amount. Furthermore, the response function is
Case-A Case-B
Fig. 9.32 Superposition principles of linear systems. Summing of ltered signals (case-A) is equivalent to ltering of
their sum (case-B)
t t
FILTER
0 0
impulse impulse
function response
230 9 The Seismic Processing Tools
T2 t T2 t
0 T1 0
T1
cr(t-T2)
c (t-T2)
scaled by the input impulse amplitude. Referring Suppose the input to a linear system is made
to Fig. 9.34, we have three impulses; rst up of the series of impulses, a0d(t), a1d(t t),
impulse is at time zero, second impulse at time a2d(t t), , aMd(t Mt). Each of these
T1, and third impulse has negative value (in- impulses will make the lter to output the
verted polarity) and located at time T2. The impulse response function scaled (weighted) by
response functions are scaled and shifted the impulse amplitude and shifted by the impulse
accordingly, as seen in Fig. 9.34. time-shift. The output of the system due to the
input series of impulses will be the algebraic sum
9.9.5 Mechanism of the Filtering of the overlapping impulse response functions
Process (Fig. 9.35). That is:
a 1 (t-t) a 1 r(t- t)
a 0 (t) a 0 r(t) filter output
sum of responses
t t
filte r
0 t 2t
-a 2 (t-2 t) -a 2 r(t-2 t)
Fig. 9.35 Three-sample input function to a linear ltering system. The output is the sum of weighted and shifted
impulse response functions of the lter
9.9 Frequency Filtering 231
where r(t) is the impulse response function of the where the function R(x), usually referred to as
lter. the (Transfer Function, or system function), is the
The impulse response of several lters con- Fourier transform of the lter impulse-response
nected in tandem is equal to the convolution of function, r(t). It is, in general, complex quantity
their impulse responses. Thus, the impulse made up of real part |R(x)|, called the amplitude
response of the three lters R1(t), R2(t), and characteristics of the lter and imaginary part
R3(t), connected in tandem, is R1(t) * R2(t) * (x), the phase characteristics. That is:
R3(t), since we consider the output from the rst
lter as input to the second lter and the output Rx jRxjeix
of these two lter as input to the third lter.
Further, we may note that, according to the
convolution properties, the output is the same 9.9.7 Types of Filters
regardless of the order of the lters in the
sequence. Depending on the applied basis of classication,
In conclusion, an input function can be there are several ways used in lter classication.
approximated by a series of uniformly spaced Since the response function is the principal
impulses, each of which has the amplitude equal parameter that denes the lter, the main classi-
to the value of the function at the time-location of cation scheme is dependent on the response
the impulse. The impulse amplitude will act as characteristics of the lter. Based on the response
the scaling factor (weight factor) to the impulse characteristics, lters are classied into the three
response function of the lter. The lter output main types: low-pass, high-pass, and band-pass
will be the algebraic sum of all the weighted and lters. The low-pass lter passes the range (0
shifted impulse response functions. An analogue xc), the high-pass lter passes all frequencies
function can be considered as a digital function above a dened lower limit (xc), and the band
for which the sampling period is zero. This pass lter passes a dened frequency range (x1
means that ltering mechanism we have just x2). The frequency response functions of these
described is applicable to the analogue function lters are sketched in Fig. 9.36.
as well.
In the frequency domain, ltering becomes
multiplication of the amplitude spectrum of the 9.9.8 Application of the Band-Pass
input function with the transfer function of the Filter
lter, and addition of the corresponding phase
spectra. We recall from the convolution theorem In order to decide on the type of lter to use in l-
that convolution in the time domain, becomes tering of a certain time function (such as a seismic
232 9 The Seismic Processing Tools
0 c 0 c 0 1 2
Fig. 9.36 Types of lters. Transfer functions of low-pass, high-pass, and band-pass lters
t t
freq. freq.
freq.
shot record), the time function is rst Fourier and then converted by Fourier inverse-transform
transformed into frequency domain function, pro- to time domain (Fig. 9.37).
ducing its frequency spectrum. The input to the
lter is the time-domain function prepared in dig-
ital mode. Measures should have been taken to 9.9.9 Filter Shape (Side-Lobes
avoid the occurrence of aliasing prior to the sam- Problem)
pling (or re-sampling) process. Since digital lters
separate components on the basis of frequency, In practice, lter design is usually made in the
signal and noise frequency-spectra must be com- frequency domain, since in this way, the fre-
puted and their frequency bands are determined. quency spectrum becomes easier to visualize and
A band-pass lter is dened so that it will pass easier for doing the design manipulation. The
the signal, and lter out the noise. The ideal case is band-pass lter, which is the most commonly
that the transfer function is a rectangular-pulse applied lter in seismic data processing, has the
(box-car) type of function which has the constant shape of rectangular pulse in its ideal form. It
value of unity over the frequency-band to be passed passes a dened frequency band and cuts out all
and zero outside that band. In the frequency domain, other frequencies present in the signal under the
ltering is effected by multiplying the transfer ltering process. In the frequency domain, l-
function of the lter with the amplitude spectrum of tering is done by multiplying the box-shaped
the input function. At the same time the phase of the transfer function with the amplitude spectrum of
lter shall be added to that of the input frequency the input signal function. This is equivalent to
components causing time shifting (and certain convolution of the lter impulse response with
amount of shape-distortion) to the ltered wavelet. the input signal function in the time domain.
To avoid phase changing of the input frequency- Theory of the ltering in the time domain
components, zero-phase-lter is normally designed. (convolution), demands that the impulse
The ltering output is the result of multiplying response must be of innite extent, which is not
the lter-amplitude spectrum by signal spectrum, the case in practice. In other words, the impulse
9.9 Frequency Filtering 233
freq.
- +
f1 f2 0 t
finite length
pass-band filter impulse response
with side lobes
freq.
f1 f2 0 t
response function is always truncated to a nite 9.9.10 Effect of the Filter Length
length. This would result in distortion of the ideal
box-shape transfer function. The nature of dis- Side lobes development depends on two main
tortion is expressed in development of side lobes factors: the truncation-length of the lter
which, effectively, results in adding extra response function and the slopes of the derived
band-pass extensions to the originally dened transfer function. In general, it is found that as
box-type transfer function (Fig. 9.38). the length of the lter response function is
Effectively, truncation of the innitely-long increased, the side-lobes amplitude decreases and
impulse response function becomes convolution side lobes number is increased (Fig. 9.40).
operation between the spectrum of the impulse It is also found that amplitude of the side
response and the spectrum of the truncation lobes gets larger when the truncation point is
box-car type function. As a result of truncation, located at a high-part of the response function,
development of the side lobes and smoothing and gets smaller when the truncation point is
effect are obtained. Both of these distortions get nearer to a zero-crossing. In order to minimize
less severe with longer truncation window, For these effects, lter length is increased and trun-
more details of this phenomenon see for example cation point is positioned at a point where the
(Bth 1974, pp. 9597). response value is close to zero value.
The ltering process will be multiplication of
the transfer function that has the side lobes with
the input spectrum. The result of this process is 9.9.11 Effect of the Filter Spectrum
passing additional parts of the input spectrum Shape
which are present outside the required band (f1
f2) as well as the (f1 f2) frequency range. Another important effect on the side lobes is the
Passing additional unwanted parts of the input effect of the shape of the lter transfer function.
spectrum (called spectral leakage) is an The more gently sloping lter margins the less
unavoidable phenomenon that occurs merely the side lobes amplitude will be. In practice the,
because of truncation of the impulse response. lter is made of a trapezoidal- rather than of
The ideal lter, designed to pass only a dened rectangular-shape, where the lter margins are
part within (f1 f2) range, cannot avoid passing sloping lines rather than vertical lines. The lter
other parts outside this range because of the side is dened by four frequency points, called corner
lobes (Fig. 9.39). frequencies (f1, f2, f3, f4) (Fig. 9.41).
234 9 The Seismic Processing Tools
freq freq
f1 f2 f1 f2
t freq.
0 f1 f2
pass-band filter
finite length short
with side lobes
impulse response
t freq.
0 f1 f2
The frequency separations, (f1 f2) and of four corner frequencies, where (f2) is at
(f3 f4), are used as measures for the slopes of zero-separation from (f1) and (f4) is at
the lter edges. The effect of the lter shape on zero-separation from (f3). The following two
the side lobes is shown in Fig. 9.41. In this cases (cases B and C) are with different non-zero
gure, three cases are shown. In the rst case separations. Comparing case B with case C, it is
(case-A) the ideal box-shaped lter is of vertical seen that the larger the separation (the more
edges, which can be considered as being formed gentle slopes) the smaller the side lobes.
9.9 Frequency Filtering 235
freq.
f1 f2 f3 f 4
freq.
f1 f2 f3 f4
t fre q
0 0
t fre q
0 f1 f2
1
fre q t t
f1 f 2 f3 f4
before normalization after normalization
maxima are of maxima are equalized to
different values unity valu
these events differ according to their arrival knee time to the end of the trace), the scalar is
times. In general, high frequency components are kept at unity-value. The nal output is the sum of
dominant at the shallow parts of the trace. Lower all of these separately ltered and scaled traces.
frequencies are more dominant with increasing The sum of the scalar values at any time of the
record time. For this reason different lters are trace is equal to unity and the applied band-pass
needed to lter the different parts of the trace. lter, by this method, is made to vary smoothly
The normal practice is to apply lters of band- with trace-time. Effectively, the seismic trace is
widths that vary with the time of the reflection subdivided into a number of overlapping seg-
arrivals. This process is called time-variant l- ments (time gates) which secure smooth and
tering (TVF). continuously changing ltering down the whole
In practical applications, a number of lters trace. The technique is shown in Fig. 9.44.
are designed to be applied at a sequence of
time-points (called knee times) specied along
the given seismic trace (stacked or un-stacked 9.9.15 The Filter Tests
trace). The input trace is ltered separately with
each of the specied lters, and each of these Filter tests (or lter analysis processing) is a
output traces is scaled with an amplitude scaling series of processing tests aiming at determination
function such that the scalar has a value of unity of the optimum time-variant lter which gives
at the relevant knee time and linearly decreases to the best possible signal-to-noise ratio of reflec-
zero-value at the two adjacent time knees (except tion events. Filter test is carried out by passing a
for the rst and last scalars). For the rst scalar, dened part of a seismic section (50100 traces,
the scalar value keeps at the value of unity for say) through a series of trial band-pass lters.
times from the rst knee time to the beginning of This is commonly conducted in two stages. First,
trace. Similarly, for the last scalar (from the last the chosen section is passed through band widths
t3 F3
t3 F4
in which the high-cut frequency is kept xed reversing the impulse response in time causes the
while the low-cut frequency is varied. A typical phase to reverse in the algebraic sign, while
lter test involves using the trial frequency keeping the amplitude spectrum unchanged.
bands: (645, 845, 1045, 1245, 1445) Hz.
The outputted sections are then interpreted
and the optimum lter is determined. In the 9.10 Frequency-Wave Number (FK)
second stage, this process is repeated with the Filtering
trial bandwidths that have their low-cut fre-
quency is xed at the value determined in the Fourier transform is applied to transform a
rst stage and the high-cut frequency is varied. time-domain function f(t) into frequency domain
Supposing that the optimum bandwidth deter- F(x). In just the same way a distance-domain
mined from the rst stage is (1245) Hz, the trial function f(x) can be transformed into spatial
lters for the second stage will be something frequency-domain or wave-number domain, F
like (1225, 1230, 1235, 1240, 1245, (k). In the geophysical literature, F-K ltering
1250) Hz. comes under different names, such as dip ltering,
In both of these two stages, choosing the fan ltering, pie-slice ltering, and velocity l-
optimum lter is based on certain criteria tering. The more commonly known term is (F-K
examined for the reflection wavelet and reflection ltering) which expresses the role of the two-
horizon appearing in the chosen section. These dimensional Fourier transform of two-
are mainly the reflection energy (wavelet ampli- dimensional t-x functions such as seismic
tude), continuity, resolution, and noise level. As records and seismic stack sections.
a guide that can help in the determination, is that
it is expected that high frequencies are present in
the shallow parts of the section, while lower 9.10.1 The t-x and F-K Domains
frequencies are more dominant in the deeper
reflection horizons. Another tool used in giving In case of two-dimensional functions, Fourier
more quantitative assessment of frequencies is by transforms are similarly applicable, using the
computing Fourier amplitude spectra. This can two-dimensional Fourier Transform. A seismic
be done for time-gates selected by the processing record (or seismic section) is effectively a
geophysicist at the zones of interest. two-dimensional function. It is a function of both
time and distance f(t, x) and can be Fourier-
transformed into the combined frequency (F) and
9.9.16 Phase Removing Filtering wave-number (K).
A normal shot record shows, in general, the
Analogue lters have non-zero phase spectra, seismic events which are mainly the direct arri-
and hence different frequency components are vals, refraction arrivals, and reflection arrivals in
time-shifted by different amounts, giving dis- addition to coherent noise (events with apparent
torted wavelet shapes. This is expected to happen velocities) like ground roll and sound waves.
since, under ltering the phase of the lter is Reflection (and diffraction) events are curved and
added to that of the data. In cases where different the rest of events are all linear events of dips
analogue lters have been applied, it becomes governed by their propagation velocities
desirable to remove the phase shifts (which may (Fig. 9.45).
be variable), and get a zero-phase output data. On the F-K plane (Fig. 9.45), the linear t-x
The process of removing phase from a set of events are mapped as linear F-K events with their
data, involves getting the impulse response of the appropriate F-K slopes which are in inverse
analogue lter, reversing it in time, and relationship to the t-x dips. The events of the
re-ltering that set of data. The principle, on back-scatter noise (dipping opposite to the rest of
which this process depends upon, is that linear events), is mapped on the negative side of
9.10 Frequency-Wave Number (FK) Filtering 239
-k 0 +k
the K-axis. Due to their curved shapes, reflection increasing (or decreasing) dip values,
events (showing relatively wide dip range) are which transform into the F-K plane as a
mapped to their appropriate positions on the F-K range of dipping lines covering the
plane. dip-range of the curved t-x event. An
The main features of the F-K transform are the event with dip approaching zero-value
following (Fig. 9.45): would be transformed on to a vertical
line along the F-axis on the F-K plane.
(i) A linear event of dip (d) in the t-x domain
will map as a linear event in the FK map
passing through the origin and having a
9.10.2 Application of the F-K
slope (D) equal to (1/d). The FK-lter,
Filtering
which discriminate seismic events on basis
of dip, is called dip lter, and because of
An FK lter can be designed such that it can pass
the velocity-dip relationship (va = 1/d), it
part of the FK amplitude spectrum and reject the
is also called velocity lter.
rest of the transformed energy. In this way, the
(ii) All linear parallel events (events with the
transfer function of the FK lter, R(F, K), can be
same dip) in the t-x domain, regardless of
dened as follows:
their positions in the t-x plane, will Such a lter can be made to pass, or cut out,
transform to the one-dip line through the all FK energy found within the two straight lines
origin on the F-K plane. Different dips of slopes (D1 and D2) as shown in Fig. 9.46.
(d) in the t-x plane, transform into the Usually it is made to act as a reject-lter.
corresponding dips (D) in the F-K plane. In practice, FK ltering is carried out through
(iii) A curved event in the t-x plane transforms a sequence of steps. First, the seismic data is
into a wedge-shaped zone in the F-K transformed into FK domain by application of
plane. A curved event is considered as a Fourier two-dimensional transform. On the pro-
series of short linear segments with duced FK plane, the noise energy is dened by
240 9 The Seismic Processing Tools
d1
D1
t F
d2 D2
-K 0 +K
Fig. 9.46 Method of dening the FK lter. It will pass (or reject) FK energy bounded by the two lines of slopes D1 and
D2 in the FK plane. Symbols (d and D) are the dips of the event in t-x plane and F-K plane respectively
drawing a polygon containing the unwanted parts stratigraphic changes. For this reason, severe
(reject zones) of the FK spectrum. The ltering equalization is not recommended especially
process is done by setting to zero all amplitude when the data processing is directed towards
values in the dened reject zones. By inverse stratigraphic interpretation.
Fourier transform, the t-x domain, with the noise
removed, of the ltered data is recovered.
FK ltering is normally done by application 9.11.1 The Need for Trace
of special computer programs as part of the Equalization
seismic data processing software-package. Plot
of the FK domain (the two-dimensional spec- Sometimes, seismic traces show abnormal vari-
trum), is normally presented as contour-values of ation in the energy level (amplitude values) tak-
the FK amplitude spectrum or using color-coded ing place in certain parts of the one trace. The
plot. Actual examples of the FK ltering are whole trace, or group of traces in a given data set
shown in Fig. 9.47. (shot record or stack section), may show abnor-
mally low (or abnormally high) energy level
compared with the rest of traces in that set.
9.11 Seismic Trace Equalization Factors which can lead to unbalanced energy
level, may originate in the eld because of
In an equalization process of a seismic trace, the
inadequate detection and recording conditions or
trace amplitudes are so adjusted that their aver-
at a certain stage in the processing sequence.
age is constant over a window of a dened
Factors affecting amplitude value can be
length, sliding down the trace. In this process the
divided into two groups:
mean of trace amplitudes, (e.g. RMS amplitudes)
over a sliding window, are equalized. With this (i) Time-invariant factors:
process, abnormal variations in amplitude level These are factors which are of constant value (not
are minimized. Weak reflection events are rela- changing with time) affecting the energy level of
tively boosted while abnormally high amplitudes the whole seismic trace. Typical examples of this
are reduced. The net effect of equalization is to type of factors are: the effect due to the
produce more homogeneous amplitude level trace-offset, source and receiver detection con-
within the one trace and for all traces of a given ditions, or due to channel recording conditions.
data set. In principle, amplitude equalization can This can occur in processing as in case of vari-
conceal amplitude variations caused by able fold applied in CMP stacking.
9.11 Seismic Trace Equalization 241
(a) (b)
Fig. 9.47 Actual examples of FK ltering. a t-x domain, shot records before and after FK ltering, b FK domain, the
corresponding FK spectra
SF constant=p1=2
P
where, P = (1/N) (an)2, n = 1, 2, 3, , N.
The average power (P) is computed from the
square root of the sum of the amplitude squared,
P
(an)2, divided by the total number (N) of trace
samples (an).
part of traces whole traces
weak zones weak traces
9.11.3 Time-Variant Equalization
Fig. 9.48 Cases of abnormal energy levels requiring
equalization processing. a Part of the trace is abnormally Time-variant equalization or time-variant scaling
weak (weak zones), b Group of traces in a data set
showing low energy level
(TVS) is, in principle, the same as time invariant
equalization except that several scaling factors
are systematic (uniform) type of changes and are used per trace instead of one scalar used in
they are compensated by their relevant process- the time-invariant scaling.
ing procedures. These changes are not included One form of time-variant equalization is car-
in the data-equalization treatment. ried out by dividing the input trace into a number
Common cases of abnormal seismic energy of time-gates which may be of different lengths.
(trace amplitudes) are shown in Fig. 9.48. As we have done in the invariant equalization
In view of these two groups of amplitude- computation, the scaling factor, which is inver-
changing factors, two types of trace equalization sely proportional to the average power, is cal-
are in common processing applications, these culated per each gate. The scaling factor of each
are: time-invariant and time variant equalization. gate is applied at the center of the gate, and linear
interpolation of the factors is performed between
gate centers. Gates may be designed to be over-
9.11.2 Time-Invariant Equalization lapping with user-specied overlap-percentages.
The process is multiplying each sample value of
Time-invariant equalization is a scaling process the trace by the corresponding scaling value.
whereby the scaling factor is kept constant for Time-variant equalization is shown schemati-
the whole seismic trace, or for the whole of cally in Fig. 9.49.
group of traces, in a shot record or in a stack The scaling factor (SF) is in inverse relation-
section regardless of time of application. This ship with the square root of the average power
type of equalization is normally applied in cases (p1/2) of the gate.
where traces are weak due to weak energy- This type of trace scaling is called (Unity
sources or due to source and receiver poor cou- Scaling), and there is another type of time-variant
pling conditions (Fig. 9.48). scaling called (Square Root Scaling) which uses
One common method used in time-invariant (p1/4) instead of (p1/2). This means that scaling
equalization, is done by scaling of each trace of factor in unity scaling is proportional to ampli-
the data-set, using a scaling factor which is tude, whereas the scaling factor in the square-root
constant for the whole trace length and its value scaling is proportional to the square root of
is inversely proportional to the square root of the amplitude. The Square Root Scaling preserves
average power of that trace. Thus, one form amplitude variation more than Unity Scaling.
9.11 Seismic Trace Equalization 243
Gate-1
Gate-2
Gate-3
Gate-4
Gate-5
For structural interpretation the attention is on signs of certain samples of a trace which is also
the continuity of the reflection horizon and, called (polarity reversing). In certain casess in the
therefore, amplitude preservation is not that processing work, there may be needs to set to
important. Traces are equalized to help in picking zero of the values of some selected samples.
and following up the reflection horizon. In Such practices are implemented in processes
stratigraphic interpretation where amplitude such as static correction, normal move-out
preservation is needed, equalization which can (NMO) correction, trace editing, trace muting,
remove relative differences in amplitude levels is and in re-formatting of the seismic data before
not applied, or its implementation is done with being recorded on a certain storage medium.
large-gate equalization. These activities are considered to be important
and, sometimes, mandatory operations normally
followed in processing of seismic reflection data.
9.12 Trace-Samples Manipulations For these reasons sample manipulations are
considered as important tools used in the seismic
In the course of processing we may need to move data processing.
samples of the digital seismic trace from place to
place (time-shifting), or to change the algebraic
Processing of Seismic Reflection Data
10
With the introduction of the digital recording in elds), seismic signal processing was developed
the mid 1960s, a fundamental change in pro- into the now-known highly effective structural and
cessing of the seismic data began. Application of stratigraphic exploration tool.
the Information Theory to seismic exploration
was made possible by the Geophysical Analysis
Group (GAG), 10 years before, coupled with the 10.1 General Overview
great advances in the computer industry, laid
down the basis of the modern seismic data Seismic data processing is a sequence of math-
processing. ematical and statistical processes carried out on
The Information Theory has been developed the acquired raw seismic data in order to extract
during the Second World War to enhance S/N more readily interpretable data conveying useful
ratio of the radar signal. The outcome of the geological information. A processing activity can
GAG research activities is that new effective be represented as an input-output system in
mathematical analyses were introduced to be which the seismic recorded data, as the input, and
applied on the seismic-wavelet. Under these the nal stack data as the output. The processing
analyses, the seismic wavelet is treated as a system involved in this operation is made up of
propagating signal, like that generated, trans- two main parts; the processing computer system
mitted, and received electromagnetic signal. (the hardware) and the computer operation and
With the application of the principles of Infor- application programs (the software). This is
mation theory (in electrical and electromagnetic summarized as follows:
The processing operation normally includes a To achieve these three main objectives, the
group of steps (processing tools) which are input data provided by the eld crew, is sub-
mathematical and statistical manipulations aim- jected to a series of processes, called the pro-
ing at clarifying the seismic reflection signal and cessing sequence. In more details, the main
other required objectives. processes carried out by a processing system, can
be summarized as follows:
1. Reformatting of the recorded data
10.1.1 Main Processing Objectives 2. Removal of noises and interferences
3. True amplitude recovery
Main objectives of the data processing, is to 4. True spectrum recovery
achieve corrected form of the seismic reflection 5. Equalization of the seismic traces
signal through: 6. Correction of the travel time of shot traces to
Enhancement of signal-to-noise ratio. a dened datum level (static correction)
Correction of the reflection travel-time to a 7. Correction of the travel time of shot-
dened datum plane. traces to the zero-offset positions (NMO
Correction of shape and position of the correction)
reflector seismic image 8. CDP-gather sorting and trace-stacking
datum
CRP CRP
SEISMIC IMAGE
zero-time
processed
data
raw
data
processing
system
PROCESSING ACTIONS
before processing : after processing :
1. reflection travel time from surface 1. reflection travel-time from datum
2. distorted spectrum structure 2. restored spectrum structure
3. poor S/N ratio and poor resolution 3. enhanced S/N, better resolution
4. reflection event, incorrect position 4. event put in correct position
10.1 General Overview 247
Fig. 10.2 Elements of seismic exploration: Data acquisition, processing, and interpretation. Processing occupies a
central position between eld acquisition and interpretation stages
248 10 Processing of Seismic Reflection Data
(iii) Signal Enhancement important processing step in which the nal pro-
The most effective process in signal enhancement is cessing product is recorded on a permanent stor-
what is called the common depth pint (CDP) stack- age medium (usually on magnetic tapes) ready to
ing. In this process several seismic traces generated be delivered to the interpretation geophysicist
from the same reflection point (i.e. same CDP who will load it in his interpretation system.
trace-gather) are algebraically summed-up. The In addition to the conventional seismic stack
result of this sum is a new trace (the stack-trace) is sections, other types of sections may be produced
representing the mean of the traces in the according to special requests from the interpreter.
CDP-gather. In this process, reflection signals on the Examples of such data are seismic-attribute sec-
individual traces are summed constructively pro- tions in which parameters other than amplitudes
ducing stronger (larger amplitude) reflection signal. (such as frequency, phase or propagation veloc-
The seismic noise, being of random nature, is ity) are displayed in the form of seismic sections
destructively summed, and hence gets reduced. (seismic attribute sections). Such sections are
Application of digital lters is another option additional tools helping in the interpretation
for reducing noises and other unwanted inter- process, especially in stratigraphic interpretation.
ferences and enhancing the reflection signal.
There are several types of lters. The most
commonly applied type is the frequency lter 10.1.4 The Processing Sequence
which discriminates seismic events on the basis
of frequency. Another well known lter is the In order to get a seismic section made up of traces
velocity lter (also called frequency-wave num- containing reflection wavelets which are cor-
ber (Fk) lter which discriminates events on the rected with respect to shape and corrected with
basis of their apparent velocity. Frequency lters respect to travel-time, a set of processing-steps is
and velocity lters are powerful tools in seismic normally performed. These steps are taken in
signal enhancement. sequence, where some are mandatory (compul-
sory) and others are optional, decided by the
(iv) Structural Amendment
processing geophysicist according to need.
In certain cases, where there are structural com-
There is no xed sequence for the processing
plexities (folds and faults) in the subsurface
steps. Details of seismic data processing differ
geological structure, the seismic image of the
with different processing centers, and may vary
corresponding seismic stack section will have
from data-set to data-set in the same center.
structural distortions of magnitude depending on
However, there is a kind of a standard form for
the degree of the structural complexities. This
the basic processing steps normally followed in
type of distortion occurs because of shifting of
all processing activities.
the location of the reflection point due to dipping
To simplify discussions, processing activities
of the reflectors. Another type of distortion
are divided into four stages of processing. These
results from the diffraction phenomenon which is
are: data re-organization, pre-stack processing,
associated with terminating reflectors like faults
parameter optimization, and post-stack process-
and pinch-outs. Restoration of seismic sections
ing (Fig. 10.3).
affected by such distortions is made by a special
The raw data provided by the eld crew,
process (the seismic migration).
which form the input data to processing, consist
(v) Data Storage and Display of two sets of data: the seismic shot records
For quality control purposes, paper displays (or (normally recorded on magnetic tapes) and the
on-screen displays), after each processing step is supporting statistics data (normally recorded
normally produced. The nal product, which on CDs or magnetic tapes). The shooting statis-
consists of stack sections of seismic lines (or tics include the observer report in which com-
stack data-volume, in case of 3D surveying) is plete descriptions and notes on the recorded shots
recorded on a certain storage medium. This is an are documented. The report also includes the
10.1 General Overview 249
DATA RE-ORGANIZATION
PRE-STACK PROCESSING
PARAMETER
CMP-STACKING
OPTIMIZATION
POST-STACK PROCESSING
OUTPUT PROCESSED-DATA
SHOT GATHER
10.2 Data Re-organization 251
trace-sequential mode. The demultiplexed data is a convolution of the rst function, f1(t) with the
series of traces each of which consists of its own reversed second function, f2(t), that is:
sequence of samples. To further clarify the con-
cept, suppose we have four-trace record, where, C12 t f 1 t f 2 t
Trace-A is made-up of the samples a0, a1, a2,
a3, , aN, and
Trace-B is made-up of the samples b0, b1, b2,
C11 t f 1 t f 1 t:
b3, , bN,
Trace-C is made-up of the samples c0, c1, c2,
By applying this property on the equation of
c3, , cN,
the trace convolutional model, yc(t), we get:
Trace-D is made-up of the samples d0, d1, d2,
d3, , dN, yc t st ht nt st
When samples are organized in groups
sequenced in the order of sample number, hence,
(sample sequential mode), it is said that the data
is in multiplexed form, and when the samples are yc t at ht nt st
organized in the sequence of traces (trace
sequential mode), the data is said to be in where, s(t) * s(t) is the autocorrelation a(t) of
demultiplexed form. Thus, re-arrangement of the the sweep function.
sample values in the order of sample number (a0, Thus, in order to extract the impulsive-like
b0, c0, d0, a1, b1, c1, d1, a2, b2, c2, d2, , aN, bN, record of the reflections in their actual
cN, dN) gives the multiplexed mode of data reflection-times, the recorded traces are cross-
storage and when the arrangement is in the order correlated with the applied sweep, which is
of trace number (a0, a1, a2, , aN, b0, b1, b2, , equivalent to convolution with the reversed
bN, c0, c1, c2, , cN, d0, d1, d2, , dN), the sweep. This process can be done in the eld by
demultiplexed mode of storage is obtained. the Vibroseis recording system, or later by the
The necessity of demultiplexing comes from processing software (Fig. 10.5).
the fact that computer programs deal with seis- In this process, it has been shown that
mic traces in the form of time-series of sample cross-correlation of the Vibroseis seismogram
values, sequenced per each trace (trace- with the sweep function resulted in convolution
sequential mode (demultiplexed mode).
yt st rt nt
where, s(t), r(t), and n(t) are the source function Vibroseis trace after cross-correlation
sample n+1
sample n+2
timing line
sample n+3
10.2 Data Re-organization 253
on CD or on magnetic tape. For QC purposes, a for example, the header (fold) is created after
special plot (called the coverage diagram) can be CMP-stacking step.
produced to display the distribution of the sur-
face and subsurface geometry-data.
10.2.7 Data Re-sampling
10.2.6 Trace Header Assignment In its digital form, the trace consists of a series of
regularly-spaced sample-values. Normally, the
The seismic trace is considered as a record of eld trace is sampled at 2 ms sampling period and
seismic amplitude variation with reflection time, re-sampled into 4 ms sampling period, before
in addition to various types of noise. In short, it is being input to processing. In order to avoid pos-
amplitude time-function. In order to be able to sibility of aliasing, a digital anti-alias lter is
direct the processing system to a particular trace, applied to the data prior to the re-sampling pro-
each trace is provided with identication infor- cess. The type of the anti-alias lter is high-cut
mation called the (Trace Header). Among other lter that removes frequency components higher
things, trace header of a particular trace will than the Nyquist frequency. Since Nyquist fre-
specify to what shot-point or to what CMP a quency (fN) is tied up to the sampling frequency
particular trace belongs. (fs), by the relationship (fN = fs/2), there will be
The trace header consists of two parts, the no need to specify the anti-alias lter. The soft-
header name and the header value. For example, a ware is designed to identify the lter
trace belonging to shot number (17), recorded by system-function, once the sampling period (or the
channel (123), whose offset is (2450 m), and sampling frequency) is specied by the user.
included in the gather of CMP having x-coordinate
(1273.54 m), will have the following headers:
(a) (b)
(c) (d)
tS = (e-h)/v1 tS = 0
tS = + (h-e)/v1
tS = (e-h)/v0 tS = 0
tS = + (h-e)/v1
(b)
e h
z
v0
v1
symbol definitions
tS=- (z-h)/v0+(z-e)/v1 tS = (z-e)/v0 + (z-e)/v1
(c)
Fig. 10.15 Effect of earth ltering is convolution of the input source impulse with the earth response function. The
original source impulse is restored by spiking (whitening) deconvolution
260 10 Processing of Seismic Reflection Data
The term (spiking) refers to the action made inverse of the amplitude spectrum of the wavelet
by the spiking deconvolution, which is the to be deconvolved. Since deconvolution, in the
retrieval of the spike-shape (impulsive form) of frequency domain, is a division-process (or mul-
the source function. In the frequency domain, the tiplication by the inverse of operator spectrum),
spike-shaped function becomes broad and flat the output becomes nearly flat (constant value)
spectrum (making all frequency components of over wide frequency bandwidth. This is called
same amplitude). Amplitude spectrum of this whitening deconvolution. Whitening deconvolu-
form is described as white spectrum, hence the tion has the side effect of strengthening noise that
term (whitening deconvolution). happens to be within the spectrum frequency-
range of the signal. Boosting of random noise is a
(ii) The Spiking Deconvolution Operator
common observation on seismic sections which
The mathematical derivation of the deconvolu-
have been undergone whitening deconvolution.
tion lter (called the deconvolution operator) is
based on Wiener least-squares optimality prin- (iii) The Time-variant Deconvolution (TVD)
ciple, in which, an operator (shaping lter), is Usually, spiking deconvolution is applied in
computed from an input function and then, time-variant manner (time-variant deconvolution,
convolved with that function and shape it into TVD). The trace is divided into gates (normally
another desired output. Spiking deconvolution with 50 % overlapping), and an operator is
can be considered as a special case of a shaping derived from the autocorrelation for each gate,
lter where the desired shape is a spike. This then these gates are scaled and summed to pro-
process begins with computing the autocorrela- duce the nal deconvolved trace. The process is
tion function (ACF) of the gate dened for the repeated for each of the input traces (Fig. 10.17).
process. From the ACF, a deconvolution operator The technique is the same as that used in
is derived which, on convolution with the input time-variant ltering (TVF).
gate, will give the desired output of spike-shaped A closely related type of deconvolution is
reflection wavelet. For the random distribution of called (signature deconvolution) or (designature).
reflection events (an assumption required by the The procedure is to use the average spectrum of
underlying theory), the ACF is approaching a the shot-gather traces then from the inverse of the
spike at zero-lag and zero at all other lags calculated average the deconvolution operator is
(Fig. 10.16a). derived and convolved with the individual traces
It is easier to visualize the process in the fre- of that shot. This approach gives better estimate
quency domain (Fig. 10.16b), where the ampli- of the source wavelet, because of the averaging
tude spectrum of the deconvolution operator is an process.
freq. freq.
freq.
10.3 Pre-stack Processing 261
reverber -
ations
seismic deep reflector
trace
(b)
ampl. ampl. ampl
Fig. 10.19 Design of gapped (predictive) deconvolution operator for removing reverberations. a time-domain and
b frequency domain
CMP-Gathers
R4 T4
Normal moveout (NMO) is dened to be the
difference (T) between the slant reflection time Fig. 10.24 Variation of the NMO parameter (DT) as
(Tx) and the vertical reflection time (T0) for the shown in a seismic record. (T) increases with trace offset
same CMP (Fig. 10.23), that is: and decreases with depth of reflector
Fig. 10.25 Concepts of (a) Shooting raypath Static correction NMO correction CMP stacking
static and NMO
corrections. a Ray-path of
datum
CMP reflections, b seismic
image of seismic traces of a
CMP gather, after static
reflector
correction, NMO correction CMP CMP CMP CMP
and CMP stacking
(b)
plane, the adopted datum plane. Likewise, NMO used in NMO correction as (stacking velocity).
correction, effectively, places the source and This is because the velocity, when it is optimum,
receiver on a common position on the datum gives an accurate DT-correction, and hence, best
plane. With the static and NMO corrections, stack response. Since RMS-velocity is the near-
reflection events from a horizontal reflector (on a est to stacking velocity, RMS-velocity is often
CMP-gather), will be aligned on the same timing used as an alternative terminology for the
line, giving enhanced reflection amplitude on stacking velocity.
stacking of the CMP trace-gather. These con-
NMO Residual
cepts are shown in (Fig. 10.25).
The difference between the correct DT and the
Role of Velocity in NMO Correction erroneous DT is called NMO-residual (or resid-
To compute the NMO correction (DT), we need ual NMO), and may be given the symbol (DDT)
to have the offset (x) and (T0) which are known or D2T, dened as:
parameters. What is left to be determined is the
correct velocity (v) which is usually not readily D2 T DTe DT
available. The usual way to determine velocity is
through special velocity analysis processing. where DT corresponds to the correct velocity
In the derivation of the DT expression, (v) and DTe corresponds to the applied (erro-
velocity is assumed to be constant for the med- neous) velocity (ve).
ium. In this case the ray-path is straight line. The Using the approximate DT-formula, we can
more realistic model for the subsurface geology write:
is a layered medium with velocity increasing h i
D2 T = x2 =2T 1=ve 2 1=v2
with the layer depth. This model would give a
curved ray-path instead of straight path as it
would be in case the medium is homogeneous This formula shows that we get positive
having one constant velocity. NMO-residual (D2T) called over-correction
An effective velocity for a multi-layer medium when (ve < v), or negative (under-correction)
is an average velocity or RMS velocity. It has when (ve > v). This is explained pictorially in
been shown (Dix 1955) that, use of the RMS Fig. 10.26.
velocity (instead of the average velocity) for the NMODMO Correction
multi-layer medium will lead to a ray-path which So far, discussion was concerned with horizontal
is closer to the actual curved ray-path and give multi-layer model where NMO is computed
optimum NMO correction. Processing geo- using RMS velocity which can be derived from
physicists are used to call the velocity which is the interval velocities of the different layers by
266 10 Processing of Seismic Reflection Data
Fig. 10.26 Schematic uncorrected for NMO Under-correction optimum correction over-correction
representation of the effect input CMP-gather -ve NMO-residual zero-residual NMO +ve NMO-residual
of the applied stacking
velocity on the
residual-NMO
R1
R2
R3
R4
applying Dix formula. In case of complex geo- horizontal reflector. For this type of
logical structures and dipping reflectors, there is source-receiver set-up, there is one point that is
no simple formula for the NMO correction. In common to all of these ray-paths, which is the
case of a dipping reflector, the increase in source-receiver midpoint (the CMP). The sub-
reflection time for a given reflector is function of surface position of the CMP is considered to be
dip as well as for offset. Change in reflection the reflection point (RP) for the reflection
travel-time due to dip is called dip move-out ray-path perpendicular to the dipping reflector.
(DMO). This means that in case of dipping The group of traces which belong to the same
reflectors, we need to correct for the dip CMP is referred to as the (CMP-Gather).
move-out (DMO) as well as for NMO. The CMP concept is more general than that of
When the reflector is dipping, there cannot be the CDP. In fact the CDP is a special case of
a common reflection point (CRP) nor common CMP which is applicable when the dip is zero.
depth point (CDP). In fact, for the commonly Thus, when the reflector is horizontal (dip is
used shooting method, reflection points get dis- zero) all the points CRP, CDP, and CMP will all
persed over the reflection plane. They move coincide with each other.
up-dip with increasing offset (Fig. 10.27). In processing, DMO correction effectively
In Fig. 10.27, the four reflection ray-paths, moves the reflection point of each trace in the
having one common point (the CMP), will each CMP-gather to the point which is the projection
produce a seismic trace. These traces contain of the CMP on to the dipping reflector that is to
reflection events which are reflected from the point CRP in Fig. 10.28.
dispersed reflection points (RP, RP1, RP2, RP3,) It is important to note that, with this type of
and not from one point as in the case of source-receiver layout (equal offset step-ups), we
S
S3 S2 S1 R R1 R2 R3
S
CMP S3 S2 S1 R R1 R2 R3
RP3 RP3
RP RP2
RP1 2
RP RP1
CRP
Fig. 10.27 Denition of the common midpoint (CMP) Fig. 10.28 Principle of DMO correction, moving the
in case of dipping reflector. Its subsurface position is the reflection points of the CMP gather-traces to coincide
reflection point (RP) of reflection path is perpendicular to with the CRP which is the projection of the CMP onto the
the dipping reflector dipping reflector
10.3 Pre-stack Processing 267
cannot have a common reflection point (CRP). zero to unity is usually made gradually and not
However, in processing (as in DMO, and in abruptly. Thus, in order to avoid creating spikes
pre-stack migration), reflection arrivals of a or sudden breaks in the trace at the extremities of
CMP-gather traces, from a dipping reflector, are the mute-zone the scaling function takes the
corrected such that the end result will be corre- ramp shape. To secure gradual change, a
sponding to rays being reflected from a common ramp-type of function is used in muting rather
reflection point (CRP) which is the projection of than a step-type of function. The time of the ramp
the CMP onto the dipping reflector. This is, part of the scaling function is called (ramp
effectively, moving back of the RPs to one length). Trace muting is sometimes called
common point, the CRP (Fig. 10.28). ramping because of the use of ramp scaling
DMO is effectively removing, the dispersal of function. The concept is shown in Fig. 10.29.
the RPs of non-zero offset ray-paths, where both The muted zone can be at the beginning of the
of the reflection-point location and reflection trace, at end of the trace, or located within the
arrival time are changed. This implies that DMO trace. These are called on-muting, off-muting,
has the action of migration and that is why it is and surgical muting, respectively (Fig. 10.30).
sometimes called (pre-stack partial migration). In With the muting step, the pre-processing stage
recent years, full pre-stack migration is carried is completed. The corrections made in this stage,
out instead of DMO, especially in cases of consist of two types of corrections: signal form
complex geological structures which involve restoration and signal reflection time correction.
severely dipping strata. The rst group of corrections is directed to
enhancing the reflection signal by spectrum
restoration, amplitude corrections, and noises
10.3.9 Trace Muting removal. The second group of the pre-processing
steps is directed towards correction of the
Ideally, the traces, corrected for NMO, should be reflection travel-time (to vertical time measured
freed from non-reflection signals before being from a dened datum plane) by static and NMO
input to the following processing step; the CMP corrections. Input to this stage of processing is
stacking. Normally, the early parts of a seismic the eld shot-records and the output is the same
trace contain large amplitude arrivals of the shot records but corrected and re-sorted into
direct and refracted waves. Another important CMP trace gathers.
factor that makes the early parts of the trace As an example for the main processes done in
undesired is the effect of NMO correction on the pre-processing, a centre-spread shot (48 chan-
early arriving non reflection events. It introduces nels) was subjected to the following
abnormal amplitude stretching of these early
arrivals. Also, during seismic recording, a noise
before scaling after
burst (noise with abnormally high amplitude) function
muting muting
may be also recorded. A typical example of noise 0 1
burst is that caused by cable jerk in marine
shooting, or by sudden mechanical agitation in
land shooting. Such events could be recorded at
any time during sensing or during recording. ramp
length
In processing, these abnormal events are
removed by a process called (muting) which is
carried out simply by zeroing the undesired parts
of the trace. The muting process is done by
multiplying the trace by a scaling function which
is of zero-vale over the parts to be muted and Fig. 10.29 Concept of trace muting, using mute
unity over the rest of the trace. The change from ramp-function
268 10 Processing of Seismic Reflection Data
time
10.4 Stack and Post-stack
Processing
(c) (d)
(e) (f)
10.4 Stack and Post-stack Processing 269
3. CMP-STACKING
4. POST-STACK PROCESSING 10.4.2 CMP Stacking
- Post-stack migration
- Post-stack deconvolution The process of CMP-stacking, introduced by
- Random noise attenuation (Maine 1962), is summing the trace-gather of a
- Frequency filtering
- Equalization CMP. This is done after both static and NMO
- Data reformatting, displays & storage corrections are applied to the gather-traces in
order to bring reflection events in phase. The
stack trace is normally subjected to a certain
scaling process as nding the arithmetic mean of
FINAL STACK the summed traces with or without weighting.
Thus, each CMP-gather will, after stacking, give
Fig. 10.32 Main steps normally taken in stack and one stack trace showing enhanced reflection
post-stack processing stage wavelet with marked (S/N) enhancement of the
reflection events (Fig. 10.33).
process are collectively referred to as post-stack The CMP-stack can be considered as a pro-
processing, or just, post-processing. The main cess that has its own characteristic response
steps included in the pos-processing stage, are which is mainly dependent on the velocity
summarized in Fig. 10.32. (stacking velocity) used in the NMO correction.
The stacking-response curve has a maximum
value corresponding to the optimum stacking
10.4.1 CMP Stacking of Seismic velocity. The response falls from the maximum
Traces value when velocity deviates from the optimum.
Too-low velocity (positive NMO residual) and
Stacking of a group of seismic traces is a process too-high velocity (negative NMO residual), make
by which seismic traces are summed up together the stacking response to fall from its maximum
to form one output trace, called the stack trace. value (Fig. 10.34).
The process involves nding the algebraic sum The main advantage of the CMP stack is
of the values of trace-samples which have a enhancing the seismic reflection signal. This is
common recording time. The output is normally because reflection events are summed in phase
divided by the number of traces entering in the resulting in constructive interference whereas the
summation to get the arithmetic mean of the background noise, being of random nature, is
traces. In seismic data processing, there are three destructively interfering with each other resulting
with noise-attenuation. The increase in S/N ratio done normally with weak surface-sources (such
due to CMP stacking is estimated to be of the as weight dropping and Vibroseis sources) to get
order of square root of the number of the traces in enhanced reflection signals. Vertical stacking is
the CMP-gather (that is square root of the fold of applied without doing any time-corrections to the
coverage). traces before being summed together.
Another important advantage of the CMP
stack is attenuating multiple reflections. The
attenuation is done on the basis that a multiple 10.4.4 Horizontal Stacking
would have a greater NMO than a primary
reflection that occurs at the same arrival time. As The term (horizontal stacking) is used to refer to
velocity is generally increasing with depth, a stacking of a number of traces belonging to the
multiple is recognized from its apparent same shot-point, or to the same CMP. This type
abnormally-low stacking velocity. Seismic of stacking is rarely applied in normal processing
events such as multiple reflections and diffraction activities. However, it is applied in certain cases
arrivals (when appearing on a CMP-gather) show of too-weak reflection signals recorded amid a
larger NMO value than the primary reflection at highly noisy background. Although horizontal
the same time. Thus, when NMO-correction, stacking results in loosing reflection character, it
with optimum stacking velocity, will align the will help in improving reflection strength and
primary arrivals but leave the non-primary arri- reduce noise level. The computation applied in
vals with large NMO residual. On stacking, pri- horizontal stacking involves a running-average
mary reflection will add up constructively while type of procedure. The method can be explained
multiples and other non primary reflections will as follows:
be attenuated because of the residual NMO. Referring to Fig. 10.35, we have a shot record
made up of 6 traces (TR1TR6). Using
three-trace averaging, the rst three traces (TR1,
10.4.3 Vertical Stacking TR2, TR3), NMO-like correction is applied to the
traces with respect to the rst trace (TR1) in this
Vertical stacking, involves summing corre- group, and then, stacked to produce the stack
sponding traces from several shot-point records trace (STR2) which will replace the group
which are executed in the same location. This is mid-trace (TR2) in the original record. The
10.4 Stack and Post-stack Processing 271
Shot record After horizontal stacking usual CMP stack as well as in vertical stack, to
TR1 TR2 TR3 TR4 TR5 TR6 STR2 attenuate noise bursts and abnormally-high
R1 R1 amplitude arrivals (such as ground roll), and to
prevent them from appearing in the stacked trace.
R2 The term is also applied to stacking in which
R2
amplitudes which are exceeding some specied
R3 R3 threshold are completely excluded instead of
being given the appropriate weights (Sheriff
1973, p. 60).
TR1 TR2 TR3 TR1 TR2 TR3 STR2
R1
10.4.6 The Stack Section
R2 + +
The stack trace is obtained from summing of all
R3 NMO-corrected traces in the CMP-gather, a
process which is repeated for all the
Differential stacking CMP-gathers after being CMP sorted. From
NMO
displaying these stack traces in the order of CMP
Fig. 10.35 Concept of horizontal stacking. The rst 3 sequence-number, the stack section is obtained.
traces (TR1, TR2, TR3) are subjected to differential NMO A stack section represents a subsurface seis-
and stacked to produce stack trace (STR2) replacing the mic image of the subsurface geological structure.
group mid-trace (TR2) in the original record
Normally, the stack section is provided with
top-label showing the CMP sequence numbers,
process is repeated with the second three-trace
elevation, and static correction drawn as function
group (TR2, TR3, TR4) to produce (STR3)
of distance along the seismic line. Another
replacing (TR3) in the original record. This pro-
important label (normally drawn below the sec-
cess is repeated until the whole record is com-
tion) is the fold of coverage drawn as function of
pleted. In this example we chose the group to
the CMP sequence number.
consist of three traces. The number of traces
A seismic stack section can be looked upon as
taken in the group is optional, but, usually not
a two dimensional function representing the
too-large number.
variation of amplitude with respect to both of the
The important point in this process is the
distance (x), measured along the seismic line,
application of NMO-like correction to the traces
and two-way vertical time (t), that is f(x,t).
prior to stacking. This correction (called differ-
A typical stack section is shown in Fig. 10.36.
ential NMO) is done with respect to the rst trace
(trace of the smallest offset) in the group.
top label of section for CMP sequence numbers and other data
10.4.5 Diversity Stacking
geological reflectors
(a) (b)
PS PM
TS TM
reflection
M dipping segment
reflector record
S
surface
Fig. 10.42 Nature of distortion in unmigrated stack b reflection segment distortion includes position shifting
sections in case of dipping reflector, a reflection event and changes in both of its dimensions and dip-angle
shifting from its true position at M to false position at S,
10.5 Seismic Migration 275
vertically below position (PS) at the same where, (Xs, Zs, Ts, hs) refer to the stacked section
reflection time. That is travel time (PS-M) is geometry before migration, and (Xm, Zm, Tm, hm)
equal to travel time (PS-S). In this way the event refer to the stacked section geometry after
is placed in the wrong position as it has been migration. The terms (xm, Tm) represent the
shifted both horizontally and vertically from its restoration shifts realized in the migration process.
true position. These relations tell that migration results in
For the same reason, a dipping reflection seg- the following restorations:
ment (Fig. 10.42b), will occupy a wrongly-placed
(i) Segment length decreased (M < S)
position relative to its true position. In addition to
(ii) Segment dip increased (hm > hs)
its wrong placement, the reflection segment will
(iii) Shifting the segment horizontally and
experience changes in its dimensions and in its
vertically by (xm) and (Tm) respec-
dip-angle. Migration process is the process that
tively in up-dip direction (S moved to M).
can remove all these distortions from the stack
Both shifts are functions of the reflector
section and bring about the required restorations
dip as well as of its depth (reflection time
to recover the true reflection surface.
and velocity of the medium).
10.5.5 Types of Migration starting with the early manual and semi-manual
Techniques methods for historical interest.
(i) Dip-Bar Migration Method
The history of migration principles started with
This method is historically the earliest migration
the history of the exploration seismic method
technique (after Karcher primitive migration)
(Robinson 1983b, p. 95). In 1921, J.C. Karcher
applied to migrate dipping reflection events.
established reflecting interfaces by manually-
Prior to computer time, seismic migration was
drawn envelops to the circular arcs centered at
performed by this graphical method called
the surface observation points. This processing
(Dip-Bar Method).
step (applied by Karcher in 1921) is historically
This method is based on calculating the real
considered to be the rst migration process that
dip-angle (h) from the equation:
was applied for seismic reflection data. In the
1950s a step forward in the development of Sin h Dt v=2 Dx
migration computations was taken by Hagedoorn
who introduced a graphical method based on where Dt is the difference in reflection times
drawing the wavefont-maximum convexity observed at adjacent zero-offset traces, Dx apart,
curves (Hagedoorn 1954). assuming constant average velocity, v (Dobrin
With the introduction of the digital processing 1960, p. 136).
systems (including special computer programs) in Once the velocity (v) is known or assumed,
the mid 1960s, migration lengthy computations the true dip angle (h) can be calculated and each
were made practical. During the 1970s1980s reflector segment is shifted up-dip by the deter-
interval, three main methods dominated the mined dip-value (h) to the new migrated position.
migration computation processing. These are; This process is repeated with all of the reflector
Diffraction stack (Kirchhoff Stack), Finite- segments producing the nal migrated section
Difference, and FK migration. In recent years, (Fig. 10.44).
the modern the Reverse Time Migration It should be noted that the dip measured on a
(RTM) was introduced. stack section is the true geological dip if the
Development trend of migration techniques prole is in the dip direction, that is, the line is
seems to be towards building more sophisticated perpendicular to strike. If, however, the line is at
algorithms for computing pre-stack depth azimuth (a) with the true-dip direction, then
migration, based on wave equation modeling. migration requires adjustment of the migration
Here, a short denition noteis given to the velocity. For a single dipping reflector with uni-
common methods used in seismic migration, form overburden velocity (v), the correct
original
position Reflection
migrated segments
position
10.5 Seismic Migration 277
circular
wave-front CMP
corrected, TS
record
surface
reflector circular
surface wave-front
Two facts are employed in the process of Like the wave front graphical method, this
migration. First, the diffraction point is located at method is now obsolete.
the apex of maximum convexity curve (the
diffraction curve), and secondly, the wave front (v) Diffraction Stack (Kirchhoff) Method
(semi-circle for constant velocity) of a reflection Diffraction stack migration is based on the
event intersects the maximum convexity curve at assumption that a reflector is made up of closely
two points. The rst point (unmigrated-position spaced diffraction points each of which produces
point), marks the point at which the reflection a diffraction-arrivals event (hyperbolic curve for
segment is tangent to the diffraction curve. The constant velocity) as seen by surface receivers.
second point (the migrated-position point), is the The true position of the diffraction point is
diffraction point (the curve apex). The migration located at the apex of that hyperbolic curve. All
process is done by drawing appropriate maxi- diffraction arrivals from the point-diffraction
mum convexity and wave-fronts curves and source will fall on that hyperbola. For a contin-
moving of reflection events to the apex of the uous reflector (reflector with no termination
appropriate diffraction curve (maximum con- points), the reflectors unmigrated image, is
vexity curve). considered to be formed from the superposition
To explain the method, the steps needed to of the diffraction hyperbolae. The reflection
carry out the migration can be presented in the event appears as a result of constructive inter-
form of a manual procedure. Based on the velocity ference at those diffraction points is considered to
eld of the section, a wave-front curve and a be making up the reflection surface (Fig. 10.49).
diffraction curve (maximum convexity curve) are If, however, the reflector is terminating, the
constructed. The diffraction curve is slid laterally termination point, acting as a diffraction point,
until a part of it becomes tangent to the unmi- will create a diffraction hyperbola that appears
grated segment at point PS (Fig. 10.48a). The superposed on the stack section. This hyperbola
wave-front curve is placed with its central axis can be collapsed to the hyperbola-apex, which is
passing through the reflection point to be migra- located at the diffraction-point. One of the
ted. The position of the migrated point is located migration methods used in collapsing diffraction
at the apex of the diffraction curve (Fig. 10.48b). arrivals is based on diffraction-stack (also called
The true after-migration position of the segment Kirchhoff-stack) migration.
will be the segment which is passing through the Based on this approach, migration of a point in a
point (PM) and tangent to the wave-front, having record section is achieved by summing all data
its central axis through point (PS). values (sample values) falling along the diffraction-
This process results in migrating the segment hyperbola that is centered at that point. Result of
at point (PS) into its reflector surface segment summation along the hyperbola is assigned to the
passing through the apex of the diffraction curve, apex of that hyperbola. By considering the record
point (PM). section to be the outcome of superposition of all the
hyperbolae generated by the diffraction points,
(a) (b) each point on the section is replaced by the sum of
amplitudes falling at the intersections of the
apex established hyperbola (centered at that point) with
circular
PM the seismic traces in that section. When the sum-
PM wave-front
PS med amplitude-values belong to the diffraction
PS arrival they will be in phase and constructive
curve of max.
convexity interference will result. If, on the other hand, the
values are not diffraction arrivals, a destructive
Fig. 10.48 Principle of migration by maximum convex- interference will result in cancellation, leaving only
ity curve method: a record surface of reflection segment,
and curve of maximum convexity, and b wave-front the diffraction amplitudes to be summed and placed
intersecting the maximum convexity curve at the apex of the hyperbola.
10.5 Seismic Migration 279
surface surface
Fig. 10.49 Diffraction stack migration is based on considering the reflector is made-up of closely spaced diffraction
points, each of which generates a diffraction hyperbola
To further clarify the method, the diffraction (vi) Finite Difference Method
hyperbola (for the given velocity eld) is drawn This method is based on the principle of the
on the section, then sum (stack) all the downward continuation principles of the wave
trace-amplitudes (sample values) found at the eld observed at the surface level of the seismic
points of intersection of the hyperbola with the section (Claerbout 1976). Migration computation
traces of the section. The summation result is is carried out by solving the wave-equation based
plotted at the apex of the hyperbola. The process on nite-difference approximation. At the time
is repeated for every sample of that trace. By when this method was introduced (mid 1970s), it
repeating this process for all possible apexes and was called the Wave-Equation Migration. Later,
for all traces in the given section, the whole Larner and Hatton (1976) showed that both of the
section will be migrated. The method is diffraction stack migration and the wave equation
schematically presented in Fig. 10.50. migration were in fact based on the same
Diffraction stack (Kirchhoff) migration incor- model, but with different computation approa-
porates scaling and other processes before sum- ches. Diffraction stack method solves the wave -
mation in order to reconstruct the amplitude and equation by integration and the wave equation
phase of the migrated traces properly. This migration method by nite difference computa-
migration method can be performed as time tions. Hence the term (wave equation migra-
migration (using RMS velocities) or as depth tion) was replaced by the term (nite-difference
migration (using interval velocities). migration), leaving the term (wave equation
(a) (b)
input traces
P
diffraction
hyperbola
diffraction point
Seismic trace
migration output = an at P
Fig. 10.50 Principle of diffraction-stack migration. by placing at the apex (P) the sum of amplitudes found at
a The seismic trace is the result of summations done for intersection points of the hyperbola with the traces in the
all the hyperbolic arrivals. b Diffraction stack migration is section
280 10 Processing of Seismic Reflection Data
migration) to be applied for all methods which until depth (z) of the Nth reflector (at z = NDz,
are based on wave-equation solution. Likewise, say) is reached.
the term (diffraction stack) was also replaced by To clarify the concept let us consider the effect
the term (Kirchhoff migration) since it has been on a record section containing diffraction arrivals
shown later in 1970s, that diffraction stack is an created by a point diffractor. The input section
expression of the Kirchhoff integral solution of shows the hyperbolic diffraction event. By use of
the wave equation (Robinson 1983b, p. 99). wave equation computations, the section is sim-
The Finite Difference method is based on a ulated at a lower level, at depth (z = Dz). At this
reflector model called (exploding-reflector level, the process results in showing how the
model). According to this model, the reflector section looks when the zero-offset receivers were
is made up of closely spaced diffraction points, placed at this depth. With progressive downward
and the record section is created from rays which shifting of the receivers, the diffraction hyperbola
are normal to the reflector surface and received diminishes as the computation level approaches
by the zero-offset receivers. The exploding- the true location of the diffraction point-source.
reflector travel times must be doubled (or using When the receivers were at the level of the
half of the medium velocity) in order to get a diffraction point, the diffraction hyperbola would
record section having the same ray-path geome- have been completely collapsed. At this stage, the
try as that obtained with the normal stack sec- section, down to that level (z = NDz), is com-
tions (zero-offset traces). pletely migrated. The process is schematically
The record section (the input to migration shown in Fig. 10.51.
process) represents the seismic amplitude as According to this model, migration becomes a
function of distance (x) and time (t) recorded at process which is reversing the wave propagation,
surface plane (depth = 0), that is f(x, t, z = 0). which constructs the reflector states that existed
Given the velocity eld and, with the aid of at time (t = 0) from the input time section which
wave-motion equation, the function f(x, t, is the zero-offset traces recorded at time (t = total
z = Dz) is reconstructed at depth of (Dz) below reflection travel time).
the surface. In the next step, the output of the
(vii) Frequency-Wave Number (FK)
previous step f(x, t, z = Dz) is inputted to the
Migration
following step to work out the wave eld (seis-
This method of migration (sometimes called Stolt
mic amplitude) at depth (2Dz). For an N-reflector
Migration) was developed in 1978 by Stolt
section, the downward computation is repeated
(1978), The technique is based on
surface z=0
z z z
z z
Fig. 10.51 Principle of the downward continuation method. By successive downward shifting of the zeo-offset
receivers, the hyperbola shrinks and completely collapsed at the true position of the diffraction source
10.5 Seismic Migration 281
two-dimensional Fourier transform of the stack only, where the depth (z) is equal time (t) multi-
section, which is essentially a two dimensional plied by the velocity scaling-factor (v/2), that is
function of time and distance domain, f(t, x), into (z = tv/2). The crucial parameter in differentiat-
FK domain. Migration is carried out in the FK ing between time-migration and depth-migration
domain and then apply the 2D Fourier inverse is velocity. Time-migration assumes velocity to
transform to transform the data back into the change only in vertical direction, whereas depth
original time-distance, f(t, x) domain. The tech- migration assumes velocity variation in both
nique was extended to cover variable velocity by vertical and horizontal directions (Sheriff 2002,
(Gazdag 1978). p. 84 and p. 359). These two cases are explained
as follows.
(viii) Hybrid Migration Method
The method is based on frequency-space nite (i) Case-1 Velocity varies vertically only
difference migration. This method (called This is the case of a geological section made up of
Omega-x migration) can handle both steep dips horizontal layers with interval velocities increas-
and lateral-velocity variation (Sheriff and Geldart ing with depth. Geometry of the true-diffraction
1995, p. 334). ray-paths, in this case, shows that the diffraction
curve is generated in accordance with Snells law
which governs refractions at interfaces. The
10.5.6 Time Versus Depth Migration resulting diffraction curve for the layered medium
(Fig. 10.52a) is not that smooth hyperbolic curve
Effectively, migration is an inversion process produced by a constant-velocity (RMS velocity)
transforming seismic data from time domain medium (Fig. 10.52b). In both cases, the curve
(seismic time image) to depth domain (seismic apex (the diffraction point) and the emergence
depth image). The input to migration is the point of the vertical diffraction ray, fall on the
seismic record which can be non-zero offset vertical line passing through the diffraction point.
(pre-stack, CMP-gathers) or zero-offset traces Depth migration computes curve-shape
(stacked CMP-gathers). The amplitude in the exactly, taken into consideration the propaga-
input set of traces, can be considered as a tion ray as it refracts according to Snells law,
two-dimensional function of distance (x) and whether the reflection interfaces are horizontal or
time (t) as seen by receivers on the adopted inclined. Time migration on the other hand,
datum plane, that is at zero-depth level (z = 0). approximates the curve by an idealized hyper-
After migration, the time image is transformed bolic curve by using a constant value for the
into the corresponding true image position and velocity. The value normally-used is the RMS
true geometrical shape. Migration can be con- velocity derived (by Dix-formula) from the
sidered as re-mapping of the seismic amplitude interval velocities of the horizontal layers.
as function of distance (x) and depth (z). The Time migration uses RMS velocity which is
migration output is the true geological not exact velocity, but approximation of the
depth-picture seen at the start time (t = 0). This velocity eld. In depth migration, which uses
is summarized as:
(ii) Case-2 Velocity varies vertically and terms of fault resolution and in enhancing of the
horizontally S/N ratio (Lorentz and Bradley 2012).
By ray tracing method, it is possible to determine
the position of the diffracting point accurately, if
the diffracted ray is followed as it refracts at 10.5.7 Migration Parameters
interfaces according to Snells law. It is observed
that the apex of a diffraction curve is located on Migration restoration efciency is dependent on
the diffraction ray that approaches the surface quality of the input seismic data, degree of
plane at right angles. This is considered as the complexity of the subsurface geology, the
basic concept upon which migration process is
considered to be depth migration. Essentially, the
(a)
process is depth migration when it observes v1 90 diffraction
horizontal changes in velocity (allowing for the ray path
v2
ray bending at interfaces) even if interfaces are
diffraction
not horizontal (Sheriff and Geldart 1995, p. 333). point
Referring to Fig. 10.53a, the apex of the
diffraction curve coincides with the emergence (b)
point of the vertical diffraction ray. The apex apex point time image
(un-migrated
point is shifted laterally with respect to the section)
diffraction point (Fig. 10.53b). Depth migration
positions diffraction point correctly whereas time
migration is incomplete (diffused shape) and (c)
placed at wrong position, as it is shifted with
respect to the true position of the diffraction depth
time migration
point. Thus, depth migration results in accurate migration
imaging and accurate positioning, whereas time
migration results in wrong positioning and
incomplete imaging (Fig. 10.53c). Fig. 10.53 Difference between depth- and time migra-
The usual options decided upon in choosing tion, case of velocity variation in both horizontal and
of migration type are post-stack or pre-stack time vertical directions. a Diffraction ray-path, b Diffraction
curve (record surface) having shifted apex, c Migration
(or depth) migration. The normal trend now is in output showing focused and properly placed diffraction
the direction of using pre-stack time migration. point by depth migration compared with incomplete
Pre-stack depth migration proved superior in imaging by time migration
10.5 Seismic Migration 283
appropriate migration software, and the appro- the up-dip shifting (xm) and the upward shifting
priate migration parameters which secure full (Tm), where:
migration efciency in positioning of the reflec-
tion events and removing of the diffraction arri- Dxm vTs=2 sin hm
vals. The main migration parameters which the
processing geophysicist must optimize are the and
following: DTm Ts Ts cos hm
Migration Aperture
These equations show that migration is shift-
Reflector Dip
ing (by xm and Tm) of the reflection event to
Velocity-eld
its correct position. The apparent dip (dip com-
Target Depth
ponent in the direction of the seismic line) shall
Trace Spacing
change to the higher value (hm) after migration.
Layer Thickness
Data quality (iii) The Velocity Field
As it is with dip, inaccurate velocity leads to
(i) Migration Aperture incomplete migration process. Since the width of
The migration aperture is dened to be the lateral the migration aperture is function of velocity
displacement of a reflection event moved when a (and dip) it is expected that we get under, or over,
section is properly migrated. In Kirchhoff method it migration depending on whether the inputted
is expressed by the number of traces (included in velocity-values are too low or too high respec-
the one hyperbola) taken in the summation process. tively. Because of the direct relationship between
In case of dipping reflectors, the aperture the aperture and the velocity, incorrect velocity
width (xm) is given by: will lead to improper aperture. For optimum
migration, the velocity eld should be as close as
Dxm vTs =2 sin hm possible to the real geological.
The effect on images of synclines and anticlines
This formula shows that the aperture width is that, too-slow velocity causes under-migration,
(xm), is directly proportional to the three factors: leading to too-narrow syncline-images and lead-
velocity (v), dip (hm), and reflection time (Ts). ing to too-wide anticline images, whereas too-fast
This implies that the aperture is larger with deeper velocity causes over-migration, leading to
and steeper reflectors. Usually, the aperture is too-wide synclines and too-narrow anticlines. In
calculated such that it corresponds to the steepest practice, however, the applied velocity-functions
dipping event that is required to be fully migrated. are smoothed and scaled before they are inputted
For practical reasons, it is not always possible in the migration process. In case of diffraction
to use the full aperture width. There are cases in arrivals, too-slow velocity causes under migration,
which an incomplete (i.e. truncated) aperture is with which the diffraction hyperbola is only par-
used. Thus, for example, migration is incomplete tially collapsed, and too-fast velocity causes
near the edges of migrated seismic sections. This over-migration where artifacts (called migration
adverse edge-effect is inherent in migration smiles) are created.
implementations whether we have 2D- or
3D-data. Due to aperture truncation that occurs at (iv) Target Depth
the boundaries of sections, incomplete migration From the ray-path geometry of dipping reflector
will result near section boundaries. appearing in a stack section, it is evident that the
migration time (Tm), required to move a segment
(ii) The Reflector Dip of a diffraction hyperbola to its apex, is function
Referring to the equations expressing the geom- of the segment record-time (Ts), velocity (v), and
etry of migrating dipping events, the position its true dip (hm). The equation connecting these
correction is expressed by the two parameters: variables is (Tm = Ts cos hm).
284 10 Processing of Seismic Reflection Data
limited scope of application. For comparison of interpretation purposes. Examples of such pro-
the RTM migration with the conventional cessing are attributes and impedance inversion
wave-equation migration, migrated sections are computations.
shown in Fig. 10.54.
The main features of the reverse time migra-
tion technique can be summarized by the fol- 10.6.1 Frequency Filtering
lowing points:
In general, a seismic trace shows reflection
It is a pre-stack reverse-time, depth migration,
wavelets of frequency spectra that vary with the
based on wave equation solution.
depth of the reflector. The frequency tends to
It is based on reverse-time, forward modeling,
decrease as the reflection travel-time increases. It
of wave eld of the shot-gathers.
is therefore, expected that the signal-noise inter-
It requires much more computational power
action varies with reflection time. This phe-
and it is estimated that its cost is about
nomenon is the basis for the application of a
10-times that of the conventional wave-
time-variant ltering (TVF). If the area is large
equation migration.
and showing complex geology, there may be need
It has no dip limitation and thus it is used in
for a distance-variant ltering as well. However,
migration of data from areas having structural
spatial variation of the lter is rarely applied.
and velocity complexities, giving accurate
Specication of the frequency band-pass lter
positioning and signal clear focusing.
is normally found through a special lter-test
processing. With the aid of Fourier-spectrum
10.6 Application of Filtering analysis, the frequency changes with
and Equalization reflection-time can be investigated. Based on the
spectrum analysis and the lter test, a time
A good practice is to provide the interpreter with variant ltering system is designed and applied to
two nal migrated sections: one data-set with, the stacked and migrated data-set.
and one without the application of the ltering
and without equalization. Since interpretation
software is usually equipped with ltering and 10.6.2 Equalization
equalization modules, the interpreter can use the
option that suits his purpose best. This will give In order to preserve reflection amplitude charac-
the interpreter the freedom for preparing the data ters, the time-variant scaling is applied with rel-
for other specialized processing that suits specic atively long gates (23 s). This is taken in
286 10 Processing of Seismic Reflection Data
pretation, however, shorter gates may be used. As PARAM ET ER OPT I M I Z AT I ON AN ALY SES
in the case of ltering, equalization parameters are
normally determined by special processing tests. - Velocity analysis
- Residual-static analysis
There are few more optional steps which can - Deconvolution analysis
be applied in addition to the basic steps done in the - Filter analysis
post-processing stage. Examples of these options - Equalization analysis
- Mute analysis
are processes directed towards random-noise
attenuation and enhancement of the continuity
Output:
of the reflection horizons. Processing software, optimum parameters
usually contain modules (under the names like for
coherency ltering and trace mixing) which can use in corresponding processes
From inspection of the available stack section, NMO then a preliminary stack section is
one can observe the need for the correction. produced.
Presence of residual statics is expressed as una- (ii) The time difference due to dip is eliminated
ligned and poor-continuity reflection events. dening gates enclosing the reflection
Computation of the short-wave residual statics event. With this process (removing of dip
is based on use of statistical principles. The effect and correcting for datum statics and
corrections are assumed to be time-invariant and NMO), the reflection time differences
surfaceconsistent. The surface-consistency remaining on events in the selected gates
concept (one correction value for all traces are due to residual statics which are com-
occupying one surface-location) is honored in the puted by the following statistical technique.
computation process. The input to this process is (iii) A model trace (also called the pilot trace),
the CMP gathers provided by the last step (the from the traces of each CMP-gather traces,
CMP-sorting step) in the pre-processing stage. is computed. The model trace is computed
The input gather traces should possess header by stacking (summing) all traces in the
values for the numbers and location-coordinates gather except the one to be corrected. It is
of the input shots and receiver points). important to note here that the CMP traces
Referring to Fig. 10.57, the applied compu- have no common-shot, and no common-
tation processes can be summarized by the fol- receiver locations. This situation supports
lowing steps: the statistically-random nature of the
existing residual statics. The pilot trace is
(i) The CMP-gather traces, for each CMP, are
considered to have error-free static cor-
corrected for the eld-statics and for the
rection, since the sum of the random
residuals is approaching zero-value.
CMP
(iv) The residual-static value for each trace of
gathers the CMP-gather, contributing to the pilot
trace, is then found by cross-correlating the
gate-part of the trace with the pilot trace.
stack Measurement of the correlation lag-time of
section
the peak of the cross-correlation function,
gives the residual-static value for that trace.
removal of
structural This process is repeated with all of the
dips CMP-gathers of the seismic line under
analysis.
pilot trace (v) The last step in the residual statics com-
putations is determination of the residual
cross correlation
of pilot trace with
statics for the source and receiver points
CMP-gather traces separately. This is achieved by working
out the arithmetic mean of the residuals for
the common-source, and for the common-
cross correlation
functions receiver respectively.
The cross-correlation functions are normally
shot receiver
displayed for inspection and editing when
OUTPUT required.
residual statics residual statics
Adjustment of the eld statics is then made by
Fig. 10.57 Principle of residual statics analysis. Com-
adding algebraically the residual value to the
putation is based on cross-correlation of the pilot trace eld static value of the particular trace. As a
with each of the input CMP-traces common practice, several iterations (normally 2
10.7 Parameter Optimization 289
of a seismic Section (50100 traces, say) through length, and gate time-shifting along the trace. In
a series of trial band-pass lters. This is com- practice, no intensive effort is spent on nding
monly conducted in two stages. First, the chosen optimum equalization parameters. The type of
section is passed through band widths in which scaling and time-variance scheme, are fairly
the high-cut frequency is kept xed while the known, if the purpose of the processing is plan-
low-cut frequency is varied. For example, the trial ned for structural or for stratigraphic type of
bands are (645, 845, 1045, 1245, 14 interpretation. As a general rule, the equalization
45) Hz. The outputted sections are then inter- should not be so severe that amplitude characters
preted and the optimum lter is determined. In the are removed. Moderate scaling effect is preferred
second stage, this process is repeated with the trial to give a balanced section and at the same time
bandwidths have their low-cut frequency is xed keeping the amplitude-characters for the indi-
at the value determined in the rst stage and the vidual reflection events are preserved.
high-cut frequency is varied. Supposing that the
optimum bandwidth determined from the rst
stage is (1245) Hz, the trial lters for the second 10.7.6 Mute Analysis
stage will be something like (1225, 1230, 12
35, 1240, 1245, 1250) Hz. Out of the three types of muting (on-mute,
In both of these two stages, choosing the off-mute, and surgical mute), the on-mute is the
optimum lter is based on certain criteria most usually applied mute. The other two types
examined for the reflection wavelet and reflection are rarely applied. The on-mute is applied to
horizon appearing in the chosen section. These cut-out the rst arrivals and the early refractions
are mainly the reflection energy (wavelet ampli- which are highly affected by the NMO correction.
tude), continuity, resolution, and noise level. As This is necessary as these events are stretched by
a guiding tool that can be used in giving more NMO to an extent that they can, on stacking,
quantitative assessment of frequencies, is by obscure the shallow reflection events if any.
computing Fourier amplitude spectra. In aiding Two main parameters concern the muting
of the lter-test interpretation, it is useful to ramp function. These are the ramp start-time
remember that the natural trend of frequency (onset-time) and the ramp length. By the same
changes is the decrease of frequency content as usual trialanderror method, A data-set
depth increases. (NMO-corrected CMP-gather traces) of part of
the seismic line, is subjected to ramp functions of
different onset times. A panel output (stack sec-
10.7.5 Equalization Analysis tions) is produced for the applied different
ramp-lengths. From comparative study of these
Equalization analysis aims at determination of stack sections, the optimum mute parameters
the optimum equalization-parameters which are (mute onset time and ramp length) are
mainly: the type of scaling function, time-gate determined.
Extra Exploration Tools
11
The principal techniques used in seismic explo- Like P-wave, the S-wave gets reflected and
ration are based on reflection or refraction of refracted at interfaces, and hence can also be used
P-waves generated articially from seismic as an exploration tool. Unlike SV-waves, SH-
energy sources. These techniques are mainly waves incident on interfaces do not cause wave
directed towards exploring petroleum deposits at conversion. For this reason, SH-waves are more
depths reaching thousands of meters below sur- commonly used in S-wave exploration than
face. The same types of waves, with certain SV-wave. Both of the energy source and detectors
modications, are also applicable in exploring are designed in such a way as to generate and
shallower geological and archaeological features. detect the SH-components.
In support to the usual seismic tools (reflec- The limitation in the use of S-wave is that it
tion and refraction P-waves), there are other less does not propagate through liquid media.
commonly applied techniques used in the seismic
exploration. These techniques are all depending
on generating, recording and processing of seis- 11.1.1 Shear-Waves Generation
mic waves. The groups of methods (called here
extra or specialized exploration tools) are basi- To generate an S-wave, we need an energy
cally seismic exploration tools, but differ in the source that can create horizontal vibration creat-
source function of generated waves and in the ing a seismic wave that penetrates the earth in
sourcereceiver layout geometry. These sup- nearly vertical direction. The energy-source is so
porting exploration tools, referred to here as designed that it will generate stronger SH-waves
(extra exploration tools) shall be briefly pre- than SV-waves. The SV-wave may generate
sented in this chapter. P-wave when hitting an interface at an oblique
incidence, whereas SH does not. For this reason
SH-wave is more commonly used in S-wave
11.1 Seismic Exploration surveying.
with S-waves There are several types of S-wave energy
sources, all of which have the common feature
An S-wave moving in an arbitrary direction, the which is creating horizontal particle-vibration for
particle motion can be resolved into horizontal the nearly vertically advancing S-wave. In the
and vertical components, called the SV and SH eld of engineering geophysics, a hammer is
waves. In certain cases, when the particle motion blowing horizontally a metal plate which is ver-
of an S-wave is conned to one plane, it is known tically planted in the earth. For larger energy-
as plane-polarized wave, as SV- or SH-wave. penetration, horizontal vibration systems have
from the well-head, and the (offset-VSP) where Recording is usually started with the detector
the source is placed at long distance (of about placed at the bottom of the well, and repeated at
one-to-two kilometers) from the well-head. The each stop-location upwards. To cut on noise level,
zero-offset VSP, which is the more commonly vertical stacking of several records made for each
applied type, can help in identifying and recording-stop. For verication of the constancy
removing multiple reflections, whereas the offset- of the source waveform, a monitor detector is
VSP can help in obtaining information about located near the source for recording the source
structural and stratigraphic features existing in wavelet. An important type of coherent noise
the areas around the well-location. Such infor- occurring in VSP recording is the tube-waves.
mation can be used in studying faults, facies These waves are much reduced by the more ef-
changes, and fracturing systems. cient detector coupling and can be completely
removed in the processing stage. The survey can
be done in cased (and in uncased) wells, with the
11.2.2 The VSP Field Operation detector, well-coupled to the well-wall.
R3
294 11 Extra Exploration Tools
(a) depth (b) (c) on rays passing through that object. The X-ray
tomography is a technique applied in medical
examination and in certain other elds of appli-
cation. The computer assisted tomography (ter-
time DG med CAT-scan) technique uses X-rays that have
penetrated a body along many ray-paths in many
UG directions. In medical examination, tomography
is used to explain the loss in intensity of the
X-rays because of the absorptive properties of
the VSP aligned aligned different parts of the body (Sheriff and Geldart
raw Data down-going waves up-going waves
1995, p. 492).
Fig. 11.4 Sketch showing VSP record, a raw-data show- Seismic tomography is a type of inverse
ing the two sets of events (down-going and up-going modeling (inversion process) carried out for the
waves and their multiples), b the down-going arrivals are purpose of determining a subsurface geological
aligned after the traces being up-shifted, and c record with model based on travel-time measurements. Usu-
the down-going waves subtracted leaving the up-going
arrivals aligned after traces were down-shifted ally the changes found in the measured travel
times are interpreted in terms of velocity chan-
ges. Other geophysical parameters, such as the
same shifts have been applied to the original tra- attenuation factor, may be investigated by am-
ces, but in opposite direction, bringing about plitude modeling.
horizontal alignment to the up-going events. With
this process, primary reflection events are placed at
time equal to that shown by a normal seismic 11.3.1 The Forward and Inverse
reflection section passing through that VSP- Modeling Concepts
surveyed well. Multiples are removed by apply-
ing predictive deconvolution. A common geophysical phenomenon is that a
The main application of the VSP record is in geological anomaly produces a corresponding
calibration of seismic stack sections, in the same anomaly in one of the natural geophysical elds.
way as it is done with application of synthetic An anomalous high-density body, for example,
seismograms. Since the reflection wavelets are creates a corresponding high-gravity anomaly in
preserved, VSP-data provide useful information the Earth gravity eld. The change produced by a
on acoustic impedances and other petrophysical geological anomaly to the geophysical eld, may
properties of geological formations not reached be referred to as the (geophysical response) of the
by the drill hole. These are the advantages of the causing geological feature.
near-vertical (or short-offset) VSP recording. For Modeling processes are usually divided into
the long-offset VSP, there are other additional two types. These are the (direct modeling) and
advantages like furnishing lateral subsurface the (inverse modeling). The direct modeling, also
coverage, and the development of converted called forward modeling, is based on assuming
shear waves which give shear wave information. the geological anomaly (model shape and phys-
ical properties), and followed by computing its
geophysical response. The inverse modeling, on
11.3 Seismic Tomography the other hand is computing a model that can
give a geophysical response which is closest to
The term tomography has a Greek origin mean- that observed. The two types of modeling pro-
ing cross section drawing. In general, tomogra- cess are explained in Fig. 11.5.
phy denotes determination of a certain physical Geophysical exploration is invariably based on
property of an object from measurements made measuring geophysical elds (natural as gravity
296 11 Extra Exploration Tools
inverse modeling
which the process is repeated several times until In order to accurately measure the changes,
observed and computed travel-times become data acquisition and data processing are kept the
closest to the measured travel-times. At this point same in all of the executed surveys. Geophones
the produced velocity distribution is considered are usualy cemented at the bottom of shallow
to be the realistic velocity-distribution in the boreholes, in order to x the receiver coupling for
studied area. Computations involved in tomog- the repeated surveys. Seismic sections from the
raphy interpretation are usually carried out with rst survey (base survey), can be subtracted from
specially-designed tomography-algorithm. those made at later times, to give (difference-
Detailed reviews of the seismic tomography sections) to clearly display the changes (Sheriff
technique are presented in some geophysical lit- and Geldart 1995, p. 499).
erature as in Gadallah and Fisher (2009) and The more commonly studied features of a
Sheriff and Geldart (1995). reservoir by means of time-lapse surveying are
behavior of the gas cap, gas-water contacts, and
oil-water contacts. These investigations are based
11.4 4D (Time-Lapse) Surveying on interpreting the changes observed on the
reflection-wavelet shape and on reflection events
The term 4D surveying denotes a 3D survey geometry. Interpretation of the time-lapse data
repeated over the same area at specied can also give useful information on reservoir
time-interval. For this reason it is also called parameters, such as pressure of pore-fluid, type
time-lapse surveying. The produced data volume of pore-fluid, and temperature changes.
is a three dimensional function of seismic ampli-
tude value, expressed as function of (x, y, and z).
The 4D survey gives repeated 3D data volumes at 11.5 Multi-component (3C and 4C)
a sequence of time-intervals. The concept can be Recording
envisaged pictorially as presented in Fig. 11.8.
Time-lapse measurements refer to geophysical In normal seismic reflection surveys, the P-wave
measurements repeated after a certain time is recorded by geophones designed to record
interval (time-lapse) in order to study changes in only the vertical component of the arriving
certain rock properties which may have occurred seismic wave. In fact, the arriving seismic wave
during that interval. A typical time-lapse opera- includes other components of particle vibrations.
tion is repeating 3D seismic surveying over an The arriving wave, at a certain point, on the
oil-eld to monitor changes which may take surface (or on the sea-floor), can be fully mea-
place in some of reservoir parameters. All of sured by recording its three mutually perpendic-
these changes in reservoir parameters, or some of ular components. This can be accomplished by
them, may take place as a result of accumulating using three orthogonal detectors, where each
production or due to fluid injections. detector is designed to measure one component,
hence the name three component (3C) method.
The set of the orthogonal 3C system is often
T0
T1 T2
placed at the observation point in such a way that
y
it will measure the three components, the verti-
xx time, T cal, the East-West, and the North-South compo-
z
z nents (Keary et al. 2002, p. 76).
The main advantage of the 3C technique is
that it can detect S-waves as well as P-wave,
Fig. 11.8 Sketch-representation of the 4D concept. The separately at the same time. Since, with P-wave
3D survey at time T0 is repeated at later times, T1 and T2 inclined incidence, wave-conversion occurs, it is
298 11 Extra Exploration Tools
always expected that an S-wave is arriving at way of application of this tool is briefly pre-
detectors as well as the reflected P-wave. In the sented, as follows:
processing stage the recorded waves are sepa- A record of a shot-gather (or CMP-gather) is
rately processed. Studies of fracturing and considered as 2D function, f(x, t), which
porosity are typical applications of the 3C sur- expresses the variation of seismic amplitude in
veying method. the x-t domain. The main events observed in an
Another closely related detection system, x-t record can be surface waves or body waves
called (4C) recording system, in which another which, in general, include direct, refracted,
detector which is a hydrophone-type, is added to reflected, and diffracted waves. Some of these
the 3C system. The 4C recording seismometer is events appear as linear arrivals (as surface waves,
used in marine environment using ocean bottom direct, and refracted waves) and other events (as
seismometers consisting of three orthogonal geo- reflections and diffractions) take the form of
phones and a hydrophone (Sherrif 2002, p. 147). hyperbolic curves.
Seismic data (shot gather, or CMP gather) can
be transformed into another domain without loss
11.6 Passive Seismic Surveying of information. By Fourier transform, it is possi-
ble to transform the f(x, t) into the corresponding
Seismic exploration techniques which involve spectrum-domain (the FK domain) in which the
measurements of the Earth seismic eld without seismic amplitude is expressed as function of
using articially created energy sources are ter- frequency and wave-number. Another transform,
med (passive seismic techniques). Typical Radon transform (also called slant stack), re-maps
example of application of this technique is the f(x, t) function of the seismic amplitude from
measuring microseisms and the natural seismic the (x-t) domain into the (s-p) domain, where (s)
noises. Passive seismic is sometimes called (low- is intercept time of the event at (x = 0), and (p) is
frequency seismology) because it is based on the apparent dip of the event (dt /dx), measured as
detection of low-frequency seismic waves. Usu- the slope of the event in its (x-t) domain. This
ally the survey is done using 3C detection sys- measured dip represents the reciprocal of the
tems which are setup to record data over long apparent velocity (temed, slowness). On the (s-p)
listening times (several hours to several days). domain, linear events in the (x-t) domain, trans-
In petroleum seismic exploration, passive form into points, and the hyperbolic events into
seismic is applied in studying fluid behavior in ellipses (Sheriff 2002, p. 348).
oil-reservoirs. Fluid movements due to injection A slant stack is obtained by applying a linear
or due to production may cause changes in the move-out (LMO) process followed by summing
micro-seismicity pattern. Recently, sophisticated the amplitudes over the offset axis. This process
detection-systems have been applied in the eld is accomplished by applying a time shift, to each
data acquisition. The detection system is usually trace, proportional to its distance from a dened
coupled to the earth rmly and made to record reference point, then stacking the so-shifted tra-
the micro-seismic events for long recording-time ces. Radon transform is sometimes called (slant
(several hours per one reading). stack) for the similarity of the end results in these
two processes.
One important application is removing co-
11.7 Tau-P Transform herent noise and refractions. Such events can be
zeroed in the (s-p) domain, then, by inverse s-p
The Tau-P Transform is a processing tool, which transformation, the noise-free record can be
helps in ltering coherent events that are inter- recovered. Special algorithms can be used in
fering with the reflection signals. Concept and (s-p) transform implementation.
11.8 Amplitude Variation with Offset (AVO) 299
section, a seismic attribute-section (or attribute attribute domain, we get an attribute section.
volume, in case of 3D data) is obtained. Attribute sections (like frequency sections, or
Any of the wavelet attributes can be plotted as a velocity sections), are common tools applied by
function of the reflection time producing the cor- interpreters in their seismo-stratigraphic interpre-
responding seismic-attribute trace. When all of the tation. More information on this important
traces of a seismic section are transformed into the exploration tool is given in Chap. 13.
Interpretation of Seismic Reflection
Data 12
Seismic interpretation is the third and last stage eld, sonic logs, and density logs which are
in the seismic exploration project. In the two normally obtained from well logging and VSP
preceding stages (data acquisition and process- surveys.
ing), reflected waves were rst generated then The interpretation process aims at extraction
processed to give the stacked seismic sections (or of geological information from the seismic stack
data volumes) in a form ready for the concluding data. The scope of interpretation is now extended
exploration stage which is the data interpretation. to include exploring stratigraphic changes and
This involves transition of these seismic data into direct hydrocarbon detection. In addition to iso-
the corresponding geological information. chrones and structural depth maps, modern
Success of this process (geological interpreta- interpretation activities include modeling (direct
tion of seismic data) depends largely on the coor- and inverse modeling) and computing of seismic
dination between the geophysical and geological attributes.
knowledge. Those techniques which were applied
in pre-computer times were dependent on manual
12.2 The Seismic Interpretation
and semi-manual procedures with emphasis on
Tools
establishment of the structural changes of the
subsurface reflectors. In recent years, with the
The geological nature of an area can be divided
advances witnessed by computers and specialized
into two main types of geological changes: the
software, the interpretation process became faster
structural and the stratigraphic changes. The
and more efcient in exploring the geological
seismic stack section (or 3D stack-data volume)
structural as well as stratigraphic features.
is a picture (seismic image) that carries infor-
mation about the geological features (structural
and stratigraphic) with their mineral and fluid
12.1 Scope and Objectives
contents. On this basis, it is possible to divide the
seismic interpretation techniques into two main
Seismic interpretation is basically an inverse
approaches (interpretation tools). These are: the
problem solution. Given the seismic data, in the
travel-time tool and the waveform tool, each of
form of seismic stack sections, interpretation
which is aided by the seismic wave propagation
procedure is carried out to extract the corre-
velocity (Fig. 12.1), are dened as follows:
sponding subsurface geological picture. To aid in
the interpretation process, additional geological (i) The travel-time tool:
and geophysical data are made available to the Reflection travel-time computations lead to
interpreter to use. These data include the velocity determination of the variation of the reflector
Time-Depth Curve
Fig. 12.3 An actual time-depth curve used in the depth conversion process
However, sometimes diagnosis of faults is not not decisive, the interpreter can investigate the
that easy. Difculty in identication of faults is continuity of the horizon between intersecting
mainly because of small fault throw and poor seismic lines. Existence of a fault may lead to
break-sharpness. Very often the fault feature is creation of a mistie-phenomenon if the fault was
not a clear-cut break of the reflection horizon, but not indorsed in the interpretation of that horizon.
it is exhibited as a limited non-reflection zone. Another indication of faulting is nding a sudden
The appearance of a fault zone, rather than a change in reflector dip. Care should be taken in
sharp fault plane, may be due to presence of this case as the sudden change in dip may be due
fractured and crushed rocks in that part of the to intensive folding and not due to faulting.
reflector that is affected by faulting. Under certain geological conditions, the fault
In cases where the faults are not clear, inter- zone suffers from some local deformations (as
preters usually seek help by other side features. layer dragging or sagging) in addition to rock
Reflection wavelet characteristics used in the fracturing. These fault side-effects may lead to
picking process across the fault zone. If this was reflection-wavelet weakening, or complete
306 12 Interpretation of Seismic Reflection Data
fault graben
zone zone
between the acoustic impedances of these for- sedimentary basins which are largely controlled by
mations and the wavelet amplitude and polarity. sea-level changes. One or more contemporaneous
To sum up, wavelet energy content and fre- depositional systems (fluvial, delta, continental
quency content (including wavelet polarity) are shelf or slope) compromise the principal elements
all parameters carrying messages on the strati- of basin ll. These deposition systems constitute the
graphic changes of the subsurface geological seismic depositional sequences. Diagnosis and
formations. The same principle is applied in the delineation of the fundamental basin-ll sedimen-
more specialized technique in exploring the type tary units from seismic sections is referred to as
of hydrocarbon, the method commonly referred seismic sequence analysis. In essence, seismic
to as direct hydrocarbon detection, DHD. stratigraphy principal concepts are developed to
describe precipitation processes and environments
involved in creation of the sedimentary basins.
12.4.2 Basic Stratigraphic Concepts
Due to sea-level changes, the coastline changes
position accordingly. The coast line may advance
Stratigraphy (geology of sedimentary rocks), is a
land-ward (sea transgression) or retreat sea-ward
branch of geology which deals with the study of
(sea regression). With the sea-rise, for example,
the rock-layers including, origin, lithology,
sedimentation area advances land-ward making a
composition, chronologic succession, and corre-
depositional pattern called (Coastal Onlap) and the
lation. Seismic stratigraphy is concerned with the
vertical component of the sedimentation zone is
study of sedimentary formations (stratigraphic
called (Coastal Aggradation). In general, a sedi-
changes) based on seismic data. The basic con-
mentary system made up of several sedimentary
cept underlying the seismic stratigraphic analyses
layers, which were contemporaneously-deposited
is that a sedimentary sequence is expressed in
under the same deposition-environment and
terms of corresponding seismic reflection events
bounded by unconformity subsurface, is called a
(seismic sequence).
(depositional sequence).
In studying a stratigraphic phenomenon, two
A depositional sequence (time-stratigraphic
closely related aspects need to be considered,
unit) is therefore dened as a set (package) of
geological and geophysical aspects. The geo-
sedimentary layers deposited contemporaneously
logical aspect is that a sedimentary section was
as components of the same depositional system
formed as a result of precipitation under conti-
created under the same depositional environ-
nental and marine environments. The layering
ments. The unit-package (sequence) is usually
geometry and rock composition of the large
bounded by an unconformity surface at both its
sedimentary basins are largely controlled by the
top and its bottom.
sea-level changes. The other aspect is of geo-
physical nature in which the seismic (ii) The Seismic Sequence
reflection-events (representing the stratigraphic This is the expression of a depositional sequence
features appearing on seismic sections), have as a seismic image appearing on a seismic sec-
geometric forms and wavelet characters which tion. It is a group or package of seismic reflection
are directly related to the causing geological events which are bounded by unconformity sur-
(stratigraphic) features. With these two aspects in faces at both of its top and base.
mind, seismic interpretation is implemented. (iii) The Seismic Facies
A summary of the basic denitions and con- This is a genetically related set of seismic
cepts of seismic stratigraphy is given as follows: reflection events appearing on a seismic section,
(i) Seismic Stratigraphy and the Depositional corresponding to a given set of sedimentary
Sequence rock-types (lithofacies). A (seismic facies unit)
Seismic stratigraphy, that includes sequence anal- possesses its own distinctive seismic character-
yses and reflection events conguration patterns, istics which are expressing its sedimentary
are based mainly on development mechanism of properties and depositional environments.
308 12 Interpretation of Seismic Reflection Data
sed. basin
salt dome
which occur as a result of either structural uplift is indicative of growth structures. A possi-
activities, as in case of growth of salt bodies, or ble cause of growth may be shale-doming or salt
because of sratigraphic activities as in case of diaperism. In such cases the uplift-structure is
sediment influx into ocean depressions. In these subsiding at a slower rate than the region sur-
cases, thickness shall experience changes which rounding the uplift. This means that the
are expressed into the corresponding conver- uplift-structure is growing up in relation to the
gence or divergence in the corresponding seismic surrounding area resulting in thinning of the
reflection events. overlying formations.
Typical cases of such events are the con- Concerning interpretation of reflection cong-
verging reflection events observed over salt uration patterns of the thickening and thinning
domes, where the layers are thinning over the salt phenomena, it is useful to note that, in seismic
body. In contrast to this case is divergence of sections, the convergence and divergence phe-
reflection events which usually represent thick- nomena may result because of velocity changes
ening of basinal formations, normally seen dip- and not due to real thickness changes. This must be
ping in the direction of a basin centre (Fig. 12.8). taken into consideration before giving nal con-
Sometimes these two types of congurations clusion as regards thickness-change interpretation.
are called inward convergence (for convergence)
(iv) Reflection Complex Congurations
and outward convergence (for divergence).
Complex patterns indicate prograding deposits
(iii) Sedimentary Basins Conguration formed by two building components: out-building
Patterns and up-building of sediments. The phenomenon
With sufcient sediment influx, the rate of sub- occurs over continental shelves and slopes and
sidence is greater at the centre of the basin, also in association with deltaic-environments
leading to thickening into-basin patterns. Thus, deposition. These patterns (known as prograda-
thickening of a formation into a basin indicates tion deposits) are subdivided into sigmoidal and
both subsidence and abundance of sediments fed oblique conguration patterns (Fig. 12.9).
into that basin. A thick, reflection-free, image in Sigmoidal reflection patterns are gentle
the deep part of the basin indicates rapid depo- S-shaped events with the top of the pattern being
sition and rapid subsidence. parallel to the upper boundary of the sequence
Thinning of prograding seismic sequence into unit. It corresponds to sedimentary growth cre-
basin may be due to greater compaction at the ated by out-building with signicant up-building
basin deeper parts, indicating compactable type processes. This is accomplished through smooth
of lithology (as clay and sand deposits, for transition from top coastal formation towards
example). Thinning due to scarce sediments can basin centre. The process is accompanied with
be used as guide for determination of the thinning due to increased compaction and general
sediment-source location. Further, thinning over decrease of the deposited material. Sigmoidal
310 12 Interpretation of Seismic Reflection Data
basin
centre
water as shelf surfaces and shelf margins. apparent seismic anomaly may be created for the
Favorable environment for the growth of coral waves reflected from reflectors below the reef
reef is calm, clear, and warm water. Reef deposits, body. For the same high-velocity property, reflec-
made up of porous limestone rocks, are normally tion events, from deep reflectors (reflectors below
found within consolidated mudstone or shale the reef-body), passing through the reef, may cause
The seismic event representing the reflection up-curving of those events, creating what is called a
from the upper surface of the reef body differs from (velocity pull-up) phenomenon (Fig. 12.11).
that of the sand lens in two ways: rst reef Other additional indicators helping in diag-
boundaries are usually rugged surfaces which nosis of reef seismic events are the appearance of
result in creating diffractions and wave scattering diffraction events and the abnormally weak
and weak and unclear coherent reflections. In this reflection events coming from reflectors found
case, identication of reefs is achieved with the below the reef body. The diffraction events are
help of the reflection events produced by those created from the jagged and terminating nature of
overlying formations which are draping across the the reef surface. The weakening effect is attrib-
reef top. The second difference between reef and uted to the absorption and ltering effects of the
sand lens seismic events is that the velocity of the reef rock-matter.
reef material is higher than that of the host medium.
This property makes the reef body to act as a (iii) Salt Domes
converging lens for the waves penetrating the reef. Within the zone occupied by the salt-body no
As a result of the converging effect a false or reflection events are seen. This reflection-free zone
312 12 Interpretation of Seismic Reflection Data
shale
salt dome
is a common feature that serves as a characteristic By virtue of the vertical and horizontal
property aiding in salt dome identication. pushing forces resulting from the growing salt
Another characteristic feature is the thinning of the body, folding and faulting of the overlying sed-
up-pushed formations which are overlying the salt imentary cover take place. The formations,
dome. Diffraction arrivals may develop from the pierced by the migrating salt, will appear as
termination points of the reflectors that are pierced reflection events terminating against the bound-
by the growing salt material. As it is the case with aries of the salt body. Also, because of the dif-
reef bodies, salt domes affect the seismic images of ferential pressure effect, it is expected to see
the deep reflectors in which the reflection thinning of those parts of the formation found
ray-paths pass through the salt deposits. Since the above the salt dome. An actual seismic image of
salt velocity is higher than that of the surrounding a salt dome is shown in Fig. 12.13.
medium, a false structural feature (the pull-up
phenomenon) is generated. (iv) Unconformity
Most prominent features used in identifying a The unconformity feature is of fundamental
salt dome is the termination of reflection images importance in the seismo-stratigraphic studies as
at the salt dome surfaces, and absence of reflec- it acts as the basis of dening seismic sequences.
tion events within the salt body. In other words, The unconformity surface is marking a
the region dening the salt dome becomes deposition-break geological episode. It is repre-
reflection-free zone (Fig. 12.12). senting a period of non-deposition or a period of
12.4 The Seismic Stratigraphic Interpretation 313
Fig. 12.14 Sketch representation of the angular unconformity showing the angular relation between dipping layers
and a overlying horizontal layer, with pinch-out feature
erosion, followed by resumption of deposition. the reflection wavelet characteristics may vary
When the unconformity plane is at an angle with along the reflection event representing the
the overlying sediments, it shows clearly on unconformity surface.
seismic sections exhibiting an up-dip wedge In cases where the unconformity surface is
(pinch-out zone) against the overlying layers. irregular or rugged, diffraction and scattering of
Because of the angular relation between the two the incident waves, my occur, leading to quality
sets of formations involved in the unconformity, deterioration of the unconformity-generated
this feature is normally called angular uncon- reflection horizon. The effect may be so severe
formity (Fig. 12.14). that it will be difcult to delineate the uncon-
On seismic sections, the angular unconformity formity extensions.
is readily recognizable from the angular relation The most common example of this case is the
and pinch-out phenomena. A typical seismic basement surface which is considered as an
image of the unconformity is shown in Fig. 12.15. unconformity feature. Due to the jagged or
The unconformity surface usually acts as a undulation nature of the basement surface,
normal reflector, but it is not expected to be of diffraction events may be generated. The base-
constant wavelet amplitude because of the vari- ment surface unconformity (better called dis-
ation of the reflection coefcient as the surface conformities rather than unconformity) is readily
crosses different formations. For the same reason, recognizable on the seismic section, as an
314 12 Interpretation of Seismic Reflection Data
sedimentary
basin
basement
rocks
Fig. 12.16 Seismic image of a sedimentary basin having reflection events terminating against an irregular basement
surface (discontinuty)
irregular geometrical shape surface at which seismic facies units associated with the deltaic
reflection events of sedimentary formations, ter- deposits (Fig. 12.17).
minate. Below the basement surface, coherent As regard deep marine conditions, two types
reflections events disappear (Fig. 12.16). of mechanism are suggested on submarine ero-
sion and deposition. A submarine channel may be
(v) Facies Changes created as a result of river erosion or by turbidity
Within a basin, sedimentary facies which are of currents. River action may lead to formation of
coarse-grained deposits are found near basin sediment fans on the continental shelves, whereas
margins, while more ne (silt or shale) sediments turbidity currents can give rise to deep submarine
are found near basin centre. Usually, marked channels which are lled with relatively high
changes in rock lithology are found in the areas density sediments. On seismic sections, turbidites
where deltaic deposition environments prevail. are expected to be expressed by locally-dened
On seismic stack sections, the stratigraphic chaotic reflections. The other major group of
facies changes give corresponding anomalous marine deposits includes the whole range of car-
seismic events. More often such events indicate bonate rocks. These are the chemical deposits
12.4 The Seismic Stratigraphic Interpretation 315
Seismic Velocities
determination of velocity and other
reflection wavelet attributes
(from aqueous solutions) and the organic deposits As in structural interpretation procedure,
(from parts of living organisms). Typical isolated stratigraphic interpretation is carried out through
bodies of carbonate nature are the reef deposits. a sequence of steps; the stratigraphic interpreta-
tion sequence (Fig. 12.18).
12.4.5 Stratigraphic Interpretation The input to the interpretation process is the
Sequence stack section which has been subjected to a
sequence of processing of the type that preserve
Three main types of analyses are normally applied the waveform unreformed. The stratigraphic
in seismic stratigraphic interpretation. These are: interpretation starts with the identication of the
major seismic sequences by marking the
i. Identication of the major seismic bounding unconformity surfaces of the seismic
sequences. sequences. From the event conguration pat-
ii. Delineation of seismic facies found within terns, the stratigraphic units and sedimentary
the seismic sequences. environments are determined. In the following
iii. Computations of waveform parameters. steps, wavelet attributes, velocity and acoustic
The seismic sequence analyses would estab- impedance inversion are done. The whole inter-
lish the major depositional units (stratigraphic pretation work is concluded by determination of
sequences), while seismic facies analyses are the subsurface stratigraphic model including
directed towards determination of lithofacies and lithological changes, deposition environments,
depositional environments. and hydrocarbon contents.
316 12 Interpretation of Seismic Reflection Data
The main parameters commonly applied in the velocity. These indicators and their exploration
interpretation process are grouped in the following potentialities are briefly reviewed in this section.
table (Vail et al. 1977).
12.5.1 Hydrocarbon Indicators
Seismic reflection parameter Geological
interpretation
Reflection conguration Bedding patterns
Geophysical studies proved that presence of
Depositional hydrocarbon matters in the pores o a sedimentary
processes formation leads to lowering of both of the seis-
Erosion mic velocity and bulk density of those formations
Palaeogeography
Fluid contacts
(Sheriff 1980, p. 186). This means that we get a
corresponding change in the reflection coefcient
Reflection continuity Bedding continuity
Depositional (and hence in reflection amplitude) of the inter-
processes face separating the hydrocarbon-bearing rocks
Reflection amplitude Velocity-density and the surrounding medium. The direct depen-
contrast dence of these parameters on fluid contents of
Bed spacing rock formations, make them active indicators in
Fluid content
hydrocarbon detection.
Reflection frequency Bed thickness These indicators are explained as follows.
Fluid content
Areal association of seismic Depositional (i) Amplitude
facies units environment Amplitude of the reflection wavelet is an expres-
Sediment source sion of the wavelet energy-level which is con-
Geologic setting
trolled by the overall physical properties of the
Interval velocity Estimation of subsurface geology. In particular, types and
lithology
Estimation of quantities of fluid content (water, oil, and gas).
porosity A common observation is that presence of
Fluid content hydrocarbon gas in a rock formation results in
lowering of the seismic velocity. As a result of this
phenomenon we get increase of the reflection
In the 1980s, these parameters were grouped
coefcient, and hence large amplitudes over the
under one general term; the (seismic attributes)
gas-reservoir. Normally, the reflection-event rep-
which were developed extensively as strati-
resenting the top boundary of the gas-saturated
graphic interpretation tools (Chen and Sidney
formation exhibits polarity reversal. This abnor-
1997; Barnes 2001).
mal seismic feature was recognized in the early
1970s and was given the name (Bright Spot).
In presence of gas, it is found that amplitude
12.5 Direct Hydrocarbon Detection of the reflected wavelet changes with offset. The
and Seismic Attributes study tool that expresses this phenomenon is
known in the geophysical literature by the term
In the same way as the stratigraphic changes (amplitude versus offset, AVO). In general, for
affect the reflection wavelet, presence of hydro- low impedance contrast, as in the case of gas
carbon deposits causes waveform changes. reservoir, AVO shows, in general, increase of
Interpretation tools and procedures employed in reflection amplitude with the receiver-offset.
hydrocarbon detection are commonly referred to Amplitudes of reflections from top of gas reser-
as (direct hydrocarbon detection). All known voir (negative reflection), or from its base (pos-
hydrocarbon indication-tools (hydrocarbon indi- itive reflection), will both show increase with
cators) are related to waveform parameters, as offset, causing an AVO anomaly over the reser-
amplitude, frequency, phase, and propagation voir (Sheriff and Geldart 1995).
12.5 Direct Hydrocarbon Detection and Seismic 317
In order to keep amplitude changes restricted formation. In addition to that the velocity factor
to the stratigraphic nature of the medium and to can be used in exploring lithological variations and
its hydrocarbon contents, amplitude-preservation pressure conditions since velocity is function of
processing must be applied. the rock physical properties (elasticity and den-
sity). One of the common usages of the velocity is
(ii) Frequency detection of reservoir gas-pressure conditions.
An abnormal lowering of frequencies of the Normally, overpressure reduces velocity.
reflection events from reflectors beneath the
hydrocarbon reservoirs is brought about by the 12.5.2 Seismic Attributes
strong ltering effects of the hydrocarbon-bearing
formations. For this reason (ltering effect), a In the geophysical literature, the term (attribute)
low-amplitude shadow is produced by a hydro- is used to mean a property or a parameter
carbon reservoir. attributed to a certain seismic element. It serves
Frequency, as an important attribute of the as a specic character related to a geophysical
reflection wavelet, is another active indication element. For example, it is said that the fold of
tool for presence of hydrocarbon deposits. coverage is a CDP attribute, and the azimuth (or,
(iii) Phase offset) is bin attribute, because fold and azimuth
The attribute (phase) is generally useful in the are parameters of the CDP and bin respectively.
study of reflector termination and polarity Being dependant on the wavelet parameters,
reversals. In this way, phase is considered as an seismic attributes carry useful information on
effective indicator of faults, pinch-outs, and an- certain physical properties of the medium tra-
gular unconformities. verse by the reflection wavelet. An attribute can
help in clearer identication of stratigraphic
(iv) Velocity and Acoustic Impedance features, patterns, and oil-reservoir characteris-
In the mid 1970s, geophysical researches were tics. The seismic attribute is more sensitive to
directed towards employing the interval velocity one of the wavelet parameters than the qualitative
and acoustic impedance in stratigraphic explo- examination of the wavelet overall changes.
ration and in direct hydrocarbon detection. Because of this property (direct link with the
Examples of these pioneering publications are; wavelet parameters), seismic attributes lend
Sheriff (1975), Taner and Sheriff (1977), Dobrin themselves as active indicators for stratigraphic
(1977), Nestvold (1987). In this regards, the changes and for direct hydrocarbon detection.
procedure is built on computing interval velocity
(i) Seismic Attribute Denition
sections (or acoustic impedance sections) derived
In the eld of the seismic stratigraphy, the term
from the normal stack sections. This process is an
(Seismic Attribute) is used for calling one of the
inverse modeling which is normally referred to as
parameters of the seismic reflection wavelet. The
(inversion process).
attribute in this sense is a measured or computed
Two main changes take place from hydrocar-
value of a seismic parameter of a (or, related to)
bon presence in a reservoir formation; increase in
the reflection wavelet. Typical seismic attributes
the reflection coefcients (with possible polarity
are wavelet amplitude, frequency, and phase.
reversal), and increase in the travel-time
Many other attributes have been developed since
(time-delay) of the reflection wavelet passing
early 1980s, when the basic attributes were put to
through that formation. Due to the time-delay
application as stratigraphic exploration tools.
effect, it is expected to see an abnormal sagging of
Tens of attributes have been devised so far, as
the reflection events from reflectors found below
quoted in the geophysical literature.
the hydrocarbon reservoirs.
The typical role of the parameter (velocity) is (ii) Seismic Attribute Section
the velocity lowering-effect resulting from pres- A normal seismic trace is considered as a time
ence of hydrocarbon (gas, in particular) in a porous function of the parameter (amplitude). After
318 12 Interpretation of Seismic Reflection Data
being converted into the attribute-domain, we get authors as Taner et al. (1994), Chen and Sydney
an attribute trace which is a function of the attri- (1997), Taner (2001), Liner (2004), Chopra and
bute value that varies with the reflection time. Any Marfurt (2005), and Brown (2011). These clas-
of the wavelet attributes (amplitude, frequency, or sication schemes differ in the basis applied for
phase) can be plotted as a function of the reflection the classication process. Based on the compu-
time producing the corresponding seismic- tational approach applied in the derivation of the
attribute trace. attribute, the seismic attributes can be divided
By converting all of the seismic traces of a into two main groups (Taner 2001). These are the
stack section into the attribute domain, a seismic instantaneous attributes and the wavelet attri-
attribute-section (or attribute volume, in case of butes. The rst group consists of the instanta-
3D data) is obtained. Attribute sections, (like neous amplitude, frequency, and phase, while the
frequency sections, or velocity sections), are second group includes wavelet parameters which
common tools applied by interpreters in their are sensitive to physical properties and geomet-
seismo-stratigraphic interpretation. rical shapes of the reflection interfaces.
(iii) Types of Seismic Attributes The Instantaneous Attributes
Seismic attributes constitute an open set of values of The three basic instantaneous attributes are usually
any physical, mathematical, or statistical measurable derived from the seismic stack trace, using the
quantity related to seismic reflection wavelets. time-domain Hilbert Transform (Taner et al.
Typical attributes are the wavelet-parameters 1979). The instantaneous amplitude is related to
(amplitude, frequency, and phase). Other parame- the energy level of the reflection wavelets and
ters, like direct amplitude measurements, gradients, hence it can give information on the lithology and
arithmetic-mean, RMS-mean, median, cross- porosity changes. The instantaneous frequency is
correlation, and parameters computed from more sensitive to changes in the inelastic absorp-
frequency-spectra, are included in the family of tion property of the rock medium and in studying
seismic attributes. Sometimes, velocity and sedimentary facies. The third and last attribute in
velocity-dependent tools (as acoustic impedance), is this group, is the instantaneous phase which is
considered as seismic attribute on the basis that it is a considered to be very active tool in showing
parameter related to the reflection wavelet. reflection-termination phenomena as they appear
All these attributes are derived from stacked with faults, unconformities, and pinch-outs.
data, but there is an attribute derived from pre-
stack data (from gather traces), and that is the The Wavelet Attributes
(amplitude-versus offset, AVO) which is found These attributes are determined by frequency-
to be an effective tool in exploring gas-saturated domain computations (using Fourier Transform)
formations. of the seismic reflection wavelet. According to
Within two decades, after the concept of (Taner 2001), these attributes are subdivided into
attributes was introduced in 1979 (Taner et al. two sub-groups: the physical attributes and the
1979), hundreds of attributes were dened (Chen geometrical attributes. The Physical Attributes
and Sidney 1997; Barnes 2001). Thus, in addition are sensitive to the physical properties of the rock
to the basic attributes (instantaneous amplitude, medium as lithology, porosity, acoustic impe-
frequency, and phase), published in 1979, other dance, and fluid contents. They also act as indi-
attributes were devised. Examples of these are: cators for the changes in the inelastic absorption,
polarity, interval velocity, absorption coefcient, propagation velocity, acoustic impedance, and
the quality factor (Q factor), acoustic impedance, AVO parameters (gradient and intercept of the
amplitude-versus-offset (AVO), and many more. AVO plot). The Geometrical Attributes are sen-
Classication and historical development of sitive to variation of the geometrical shapes of
seismic attributes were published by many the reflection interfaces. This means that these
12.5 Direct Hydrocarbon Detection and Seismic 319
attributes are suitable for exploring structural et al. 1999). The spectral attribute is derived
changes as folding, faulting, and geological dip. directly from the amplitude spectrum (or from
In order to get better statistical support in the the power spectrum) of the seismic wavelet. Any
seismic attribute interpretation, several attributes parameter of the spectrum function can be com-
can be used in the interpretation process. One puted to be a seismic attribute. Examples of these
way of the use of several attributes is computing parameters are: the spectrum peak (maximum
a new attribute from combination of several other value), spectrum gradient, area under the
attributes. As an example on using multi-attribute spectrum-curve, spectrum width, and spectrum
approach is the use of three wavelet-attributes centroid.
(Krehbiel 1999). In this study a (trivariate attri- One of the spectral attributes (the Spectrum
bute) was derived from wavelet amplitude, Centroid) was introduced by (Al-Sadi 1973), and
amplitude derivative, and autocorrelation. was later applied to seismic reflection wavelet
(Barnes 1993b, p. 420). This attribute is consid-
Spectral Attributes ered to be one of the most efcient seismic attri-
The basic concept underlying attribute derivation butes, because each of the Centroid coordinates is
based on Fourier-spectra is not new. Spectral function of the distribution manner of the fre-
attributes were introduced in early 1970s (Bth quency components of the analyzed wavelet.
1974, pp. 203218). In a study based on earth- Computation of this attribute is done by a sliding
quake P-waves spectra, several spectral attributes window with length and shift-step which are
(called then, spectrum parameters), were intro- decided upon by the processing geophysicist. The
duced to explore properties of the Earth Upper principle of deriving of the spectrum-centroid
mantle (Al-Sadi 1973). In later years, spectral coordinates, is shown in Fig. 12.19.
attributes were made use of in the eld of seismic In 1995, the Spectrum-Centroid attribute sec-
reflection exploration. During the 1990s, tion was computed for a seismic stack section
more emphasis was directed to application of (Fig. 12.20). In this gure both of the attribute
attributes based on Fourier spectral analysis of sections (instantaneous amplitude, and Spectrum-
the seismic reflection wavelet (Barnes 1993a; Centoid sections) are presented for comparison
Trappe and Hellmich 1998; Addy 1998; Partyka purposes.
frequency
time, t
320 12 Interpretation of Seismic Reflection Data
(a)
(b)
Fig. 12.20 Seismic-Attribute colour-coded sections. a Spectrum centroid section and b Conventional instantaneous
amplitude section (Alsadi 1998)
In this Fig. 12.20, the section of the Centroid From comparison of the two attribute-sections, it
attribute represents variation of one of the cen- can be noticed that the Centroid Attribute section
troid coordinates (the amplitude coordinate). is showing better resolution.
References
Acheson, C. H., 1963, Time depth and velocity depth Bth, M., 1973, Introduction to Seismology.
relations in Western Canada. Geophysics, vol. 28. BirkhuserVerlag, Basel, 395p.
Addy S.K., 1998, Neural Network Classication Method Bth, M., 1974, Spectral Analysis in Geophysics. 563p.,
Helps Seismic Interval, Interpretation, Oil & Gas Elsevier Scientic Publishing Company, Amsterdam.
Journal, No. 38, pp 4759. Bth, M., 1979, Introduction to Seismology,
Al-Chalabi, M., 1979, Velocity determination from BirkhauserVerlag, 428p, Basel, Switzerland.
seismic reflection data: In Developments in Geophys- Birch, F., 1966, Compressibility, Elastic Constants, in
ical Exploration methods, Vol. 1, Fitch, ed., pp 168, Handbook of Physical Constants, by S. P. Clark,
Amsterdam: Elsevier. Geological Society of America, Memoir 97,
Al-Sadi, H. N., 1973, Dependence of the P-wave p. 97173.
amplitude spectrum on focal depth. Pure and Applied Bracewell, R., 1965, The Fourier Transform and its
Geophysics, vol. 104, p 439452. Applications. McGraw Hill Book Company,
Al-Sadi, H. N., 1980, Siesmic Exploration, Technique 381p. New York.
and Processing. Birkhuser Verlag, Basel, 215p. Brown, A. R., 2011, Interpretation of three dimensional
Al-Sadi, H. N., 1992, 3D surveying of inaccessible areas, seismic data, 7th. edition, AAPG Memoir 42 (9),
presented to the Iraqi tenth geological conference (in 646p.
Arabic Language), 17p. Bullen, K. E., 1965, An Introduction to the Theory of
Al-Sadi, H. N., 1994, Use of the loop spread in seismic Seismology, The University Press, 381p, Cambridge,
surveying of a rugged mountain, presented to the England.
twelfth Iraqi geological conference (in Arabic Lan- Chen, Q. and Sidney, S., 1997, Seismic attribute
guage), 12p. technology for reservoir forecasting and monitoring:
Al-Sadi, H. N., 1998, New Seismic Attributes for The Leading Edge, 16, no. 05, 445,447448,
Stratigraphic Exploration, presented to the thirteenth 450,453456.
Iraqi geological conference, (in Arabic Language), 9p. Chopra, S., and Marfurt. K.J., 2005, Seismic attributes-a
Al-Sadi, H. N., and Baban, N., 2014, Introduction to historical perspective, Geophysics. Vol 70(3)
Gravity Exploration Method. 203p, Sulaymaniyah p.3SO28SO.
University, Kurdistan Region, Iraq, Sulaymaniyah. Claerbout, J. F., 1976, Fundamentals of Geophysical Data
Araya-Polo, M. et. al., 2009, 3D Seismic Imaging through Processing, McGraw-Hill, New York, 274p.
Reverse-Time Migration on Homogeneous and Deregowsky, S. M. 1986, What is DMO?, First Break,
Heterogeneous Multi-Core Processors, Scientic Pro- vol. 4, No. 7, P. 624.
gramming, vol. 17 no. 12, pp. 185198. IOS, Press Dix, C., 1955, Seismic Velocities from Surface Measure-
publications. ments. Geophysics, vol. 20, p. 6886.
Aud, B. W., 1974, Abnormal pressure zones can be Dobrin, M. B., 1960, Introduction to Geophysical
predicted by seismic data. World Oil, Aug-74, pp. 37 Prospecting. McGraw Hill Book Company,
39. 446p. New York.
Barnes, A., 1993a, When the Concepts of Spectral Dobrin, M. B., and Savit, C. H., 1988, Introduction to
Frequency and Instantaneous Frequency Converge, Geophysical Prospecting. McGraw Hill Book Com-
Leading Edge, Oct-1993, P.10201023. pany, New York, 867p.
Barnes, A., 1993b, Instantaneous spectral bandwidth abd Dobrin, M., 1977, Seismic Exploration for Stratigraphic
dominant frequency with applications to seismic Traps, AAPG Memoir, P329352.
reflection data, Geophysics, vol. 58, P.119428. Evenden, B. S., and Stone, D. R., 1971, Instrument
Barnes, A., 2001, Seismic attributes in your facies: CSEG Performance and Testing. in Seismic Prospecting
Recorder, 26, Sept., pp. 4147. Instruments, 195p, GebruderBorntraeger, Berlin,
Bth, M., 1971, Average crustal structure of Sweden., Germany.
Pur. Appl. Geophys, vol. 88, p. 7591.
Ewing, M., Jardetzky, W., and Press, F., 1957, Elastic Liner, C. L., 2004, Elements of 3D Seismology,
Waves in Layered Media,. McGraw-Hill, 380p., PennWell, Tulsa, Oklahoma, 526p.
New York. Lombardi, L. V., 1955, Notes on the Use of Multiple
Faust, L. Y., 1951, Seismic Velocity as a Function of Geophones, Geophysics, vol. 20, p.
Depth and Geologic Time, Geophysics, vol. 16, Lorentz, M. and Bradley, R, An Introduction to Migra-
pp. 192206. tion, on the link: http://www.geol.lsu.edu/jlorenzo/
Gadalla, M. R. and Fisher, R., 2009, Exploration ReectSeismol97/rcbradley/WWW/rcbradley1.html
Geophysics, Springer-Verlag, Berlin, 262p. retrieved on 10 December 2012.
Galperin, E. I., 1973, Vertical seismic proling, SEG, Louden, L. R., Mathews, W. R., McClendon, R. T.,
Tulsa, Oklahoma, USA. Rehm, W. A., Aud, B. W., 1971, Ultra deep drilling
Gardner, D. H., 1938, Measurement of relative ground guided by seismic data: Part 2. World Oil, 171, June,
motion in reflection recording. Geophysics, vol. 3, pp. 97105.
pp. 4045. Maton, A., Hopkins, J, Johnson, S., LaHart, D., Warner,
Gardner, G. H., Gardner, L. W., and Gregory, A. R. 1974, M., Wright, J., 1995, Exploring Physical Science,
Formation velocity and density-the diagnostic basics Prentice Hall, 818p, Englewood Cliffs New Jersey,
for stratigraphic traps, Geophysics, 39, 770780. USA.
Gardner, L. W., 1939, An areal plan of mapping Mayne, W. H., 1962, Common reflection point horizontal
subsurface structure by refraction shooting. Geo- stacking techniques, Geophysics, vol. 27, p.927938.
physics, vol. 4, pp. 247259. Mayne, W. H., 1967, Practical considerations in the use of
Gassman, F., 1951 Elastic waves through a packing of common reflection point technique, Geophysics, vol.
spheres. Geophysics, vol. 16, p. 673685. 32, p.225229.
Gazdag, J., 1978, Wave equation migration by phase McQuillin, R., Bacon, M., Barclay, W., 1984, An
shift, Geophysics, vol. 43, p.13421351. Introduction to Seismic Interpretation . Graham &
Grant, F. S., and West, G. F., 1965, Interpretation Theory Trotman, 287p., London.
in Applied Geophysics. McGraw-Hill, 584 p., Mitchum, R. M. & Vail, P. R., 1977, Seismic Stratigraphy
New York. Interpretation Procedure, AAPG Memoir 26,
Green, C. H. 1938, Velocity determination by means of p. 135144.
reflection proles. Geophysics, vol. 3, p: 295305. Nafe, J. E. and Drake, C. L. 1963, Physical properties of
Gregory, A. R., 1977, Aspects of rock physics from marine sediments, The Sea, Interscience, New York,
laboratory and log data that are important to seismic 794815.
interpretation, AAPG memoir 26, p. 1546. Nestvold, E. O., 1987, The use of 3D seismic in
Gurvich, I., 1972, Seismic Prospecting (translated from exploration, appraisal and eld development, Proc.
Russian by Jean Agrell). 463p., Mir publications., 12th World Petroleum Congress, Houston, 1987, 2,
Moscow. 125132, Wiley, Chichester.
Hagedoorn, J. G., 1954, A process of seismic reflection Nettleton, L. L., 1940, Geophysical Prospecting for Oil.,
interpretation. Geophysical Prospecting, vol. 2, McGraw Hill Book Company, 440 p., New York.
p. 85127. Newman, P., 1973, Divergence effects in a layered earth,
Hardage, B. A., 1983, Vertical Seismic Proling. Hand- Geophysics, vol. 38, p.481488.
book of Geophysical Exploration. Geophysical Press, Papaulis, A., 1962, The Fourier Integral and its Applica-
London. tions. 318p., McGraw-Hill New York.
Kanasewich, E. R., 1973, Time Sequence Analysis in Parasnis, D. S., 1971, Physical Property Guide for rocks
Geophysics. 352p., The University of Alberta Press, and minerals. Geoph. Mem. 4/71, (ABEM,) Stock-
Edmonton. holm, Sweden.
Kaufman, H., 1953, Velocity functions in seismic Parr, J. O., Jr., and Mayne, W. H., 1955 A New Method
prospecting. Geophysics, vol. 18, pp. 289297. of Pattern Shooting, Geophysics, vol. 20, p. 539565.
Kearey, P., Brooks, M., and Hill, I., 2002, An Introduc- Partyka, G, et al., 1999, Interpretational, Applications of
tion to Geophysical Exploration, Blackwell Science Spectral Decomposition in Reservoir Characterization.
Ltd., Oxford, 262p. Leading Edge, March-1999, P.353.
Krehbiel, S. C., 1999, Trivariate Attribute Displays, Pennebaker, E. S., 1968, Seismic data indicate depth
Leading Edge, Nov.-1999, pp. 12481257. magnitude of abnormal pressures. World Oil, June 68,
Kulhanek, O., 1976, Introduction to Digital Filtering in pp. 7378.
Geophysics. Elsevier, Amsterdam, 169 p. Pickett, G. R., 1969, Principles for Application of
Larner, K. and Hatton, L., 1976, Wave equation migra- Borehole Measurements in Petroleum Engineering,
tion. Two approaches. 8th Annual Offshore Technol- Log Anal., May-June, p. 2233.
ogy Conference, Houston, Texas. Poulter, T. C., 1950, The Poulter seismic method of
Lavergne, M., 1989, Seismic Methods., 172 p., Editions geophysical exploration, Geophysics, vol. 15, p.181
Technip, Paris. 207.
Levin, F. K., 1971, Apparent velocity from dipping Reynolds, E. B., 1970, Predicting overpressured zones
interface reflections, Geophysics, vol. 36, p.510516. with seismic data. World Oil, Oct- 70, pp. 7882.
References 323
Richardson, E. G., 1953, Sound, Edward Arnold & Co., Sheriff, R. E., and Geldart, L. P., 1995, Exploration
352p, London, England. Seismology. Cambridge University Press., 592p., UK.
Richter, C. F., 1958, Elementary Seismology, Freeman Stolt, R. H., 1978, Migration by Fourier transform,
and Company Inc., 768p. Geophysics, vol. 43, p.2348.
Robinson, E. A., 1954, Predictive deconvolutionof time Taner, M. T., 2001, Seismic attributes: CSEG, Recorder,
series with applications to seismic exploration. Cam- 26, 4856.
bridge, MIT, Geophysical Analysis Report no. 7, Taner, M. T., and Sheriff, R. E., 1977, Application of
(reprinted in Geophysics, v32, (pp. 418484, 1967). amplitude, frequency and other attributes to strati-
Robinson, E. A., 1967, Statistical Communication and graphic and hydrocarbon determination. AAPG
Detection with Special Reference to Digital Data Memoir 26. P301328.
Processing of Radar and Seismic Signals. Hafner, Taner, M. T., J. S. Schuelke, R. ODoherty, and E.
New York, USA. 362p. Baysal, 1994, Seismic attributes revisited: 64th
Robinson, E. A., 1983a, Seismic Velocity Analysis and Annual International Meeting, SEG, Expanded
the Convolutional Model. IHRDC, Boston, USA. Abstracts, 11041106.
290p. Taner, M. T., Koehler, F., and Sheriff, R. E., 1979,
Robinson, E. A., 1983b, Migration of Geophysical Data. Complex seismic trace analysis. Geophysics, vol. 44,
IHRDC, Boston, USA., 208p. P10411063.
Robinson, E. A., and Treitel, S., 1967, Principles of Telford, W. M., Geldart, L. P., Sheriff, R. F., 1990,
Digital Wiener Filtering. Geophys. Prospect., vol. 15, Applied Geophysics. Cambridge University Press.,
p. 311333. Cambridge,770p.
Robinson, E. A., and Treitel, S., 1980, Geophysical signal Trappe, H. and Hellmich, C., 1998 A real Predication of
analysis, Prentice Hall Inc., Englewood Cliffs, N. J., Porosity thickness from 3D seismic Data by means of
466p. neural Networks, EAGE, 59th. Cong. And Exhibition
Sharma, P., 1976, Geophysical Methods in Geology. Extended Abstracts.
Elsevier, Amsterdam. Vail, P. R., Mitchum, R. M., and Thompson, S., 1977,
Sheriff, R. E., 1980, Seismic Stratigraphy. IHRDC, 227p., Seismic Stratigraphy and global changes of sea level,
Boston, USA. Application to hydrocarbon exploration, Payton, C.,
Sheriff, R. E., 1969, Addendum to Glossary of Terms E., ed., AAPG Mem. 26, Tulsa, Oklahoma, 516p.
used in Geophysical Exploration Geophysics, vol. 34, Walton, G. C. 1972, Three Dimensional Seismic Method.
p. 255270. Geophysics, vol. 37, No. 3, p. 417.
Sheriff, R. E., 1973, Encyclopedic Dictionary of Explo- West, S. S., 1950, Dependence of seismic wave velocity
ration Geophysics, Society of Exploration Geophysi- upon depth and lithology. Geophysics, vol. 15,
cists (SEG), Tulsa, Oklahoma, USA. pp. 653662.
Sheriff, R. E., 1975, Factors affecting amplitudes. Geo- Wuenschel, P. C., 1976, The vertical array in reflection
physical Prospecting,Vol. 23, P.125138. seismology, Geophysics, vol. 41, p.219232.
Sheriff, R. E., 1977 Limitations on resolution of seismic Wyllie, M. R. J., A. R. Gregory, and G. H. F. Gardner,
reflections and geologic detail derivable from them. In 1958, An Experimental Investigation of Factors
seismic Stratigraphy-Applications to Hydrocarbon Affecting Elastic Wave Velocities in Porous Media,
Exploration, AAPG, Memoir 26, pp. 314. Geophysics, vol. 23, pp. 459493.
Sheriff, R. E., 2002, Encyclopedic Dictionary of Explo-
ration Geophysics, Society of Exploration Geophysi-
cists (SEG), Tulsa, Oklahoma, USA.
Index
A
Absorption coefcient, 48, 318 B
Acoustic logging, 16 Band-pass lter, 226, 231, 232, 237, 285, 289
Air-gun, 65, 118, 293 Bartlett window, 217, 235
Aliasing, 162, 163, 181185 Baselap reflections, 308
Aliasing effect remedy, 184, 186, 189 Bin, 150, 151, 159, 162, 163, 317
Aliasing frequency, 122, 183, 185 Bin attributes, 150, 162, 317
Aliasing in the frequency domain, 185 Bin dimensions, 160, 162164
Amplitude, 35, 38, 39, 48, 67, 68, 71, 89, 295, 298, 299, Bin size, 147, 150, 162
302, 316, 318, 319 Binary numbering system, 194
Amplitude of reflection wavelet, 299, 302, 307, 313, 317, Binning, 150
319 Body waves, 36, 3739, 51, 63, 127, 128, 133, 298
Amplitude of sine function, 197, 198, 200, 210 Borehole compensated (bhc) sonic log, 17, 19, 66
Amplitude spectrum, 40, 185, 200, 206208, 210, 220, Bow-tie event, 272, 304
224, 226, 231, 232, 235, 238240, 260, 262, 319 Bright spot seismic image, 310
Amplitude variation with angle, 46, 299 Bulk modulus, 3032
Amplitude variation with offset, 46, 299 Buried focus event, 272
C De-reverberation, 261
CDP concept, 80 Diffraction from terminating reflectors, 83, 306
CDP trace gather, 110 Diffraction hyperbola, 8386, 164, 273, 278, 280, 281,
Centre spread, 106 283, 284
Checkerboard spread, 159 Diffraction seismic image, 83
CMP concept, 62, 79, 81, 266 Diffraction-stack migration, 279
CMP-gather, 67, 81, 263, 266271, 281, 287289, 298 Diffraction travel-time function, 84
CMP sorting, 249, 263, 288 Digital function, 176, 178181, 185, 231
CMP stacking, 240, 250, 254, 267, 269, 270, 272, 302 Digital function functional representation, 180
Coastal aggradation, 307 Digital function graphical representation, 181
Coastal onlap, 307, 308 Digital function tabular representation, 180
Coherent noise, 39, 40, 127, 131, 238, 254, 293, 294, 298 Digital function time-series representation, 180
Coherent noise attenuation, 127 Digital function z-transform representation, 180
Concept of spectra, 203 Digital recording system, 122, 125
Concept of the linear system, 228 Digital seismic trace, 176, 197, 243
Concordant reflections, 308 Dim spot seismic image, 310
Constant-phase spectrum, 213, 214 Dinoseis, 115, 119
Continuous spectrum, 207 Dip-bar migration, 276
Continuous velocity logging, 65, 66 Dip move-out (DMO), 76, 79, 81, 161, 266, 267
Continuous velocity survey, 64 Dip move-out (DMO) correction, 161, 266
Convergent reflection patterns, 309 Dirac-delta function, 210
Convolution, 117, 174, 176, 197, 208, 211, 215, 216, Direct hydrocarbon detection, 3, 299, 301, 307, 316, 317
220223, 227, 230232, 251, 260, 306 Distortions in stack section, 275
Convolutional model, 174, 175, 251 Divergent reflection patterns, 309
Convolution in frequency domain, 208, 215, 221 Diversity stacking, 273
Corner frequencies, 200, 233235 DMO correction, 161, 265, 266
Correlation functions, 197, 218220 Downlap reflections, 308
Correlation functions in frequency domain, 219 Drill-hole casing, 11
Cosine transform, 206 Dynamic correction, 256, 264
Critical refraction, 95 Dynamic range, 121, 124, 125, 254
Cross correlation function, 218 Dynamite, 18, 23, 40, 65, 102, 107, 113, 114, 117, 118,
Cross-line, 150, 151, 162, 163, 166 128, 135, 137, 259
CRP-Concept, 79 Dynamite cords exploding, 116
Cubical dilatation, 28, 31, 33, 37
Cyclic frequency, 199
E
Elasticity, 1, 23, 28, 53, 54, 57, 317
D Elastic moduli, 30, 32, 292
Data box, 151, 164, 166 Electrical arc, 118
Data display, 166 Electrical induction logging, 13
Data playback, 126 Electrical logging, 12, 15
Data preconditioning, 247 Electrical resistivity logging, 12
Data re-organization, 166, 248250 Electrical resistivity surveying, 7
Data storage, 3, 248, 251, 286 Electromagnetic surveying, 8
Data storage and display, 248, 286 End-on spread, 106, 107, 131
Data volume, 105, 140, 151, 152, 161, 166 Equalization, 197, 236, 240, 242, 243, 249, 262, 263,
DB unit, 49, 129 285, 289, 299
Decibel, 49, 200 Equalization analysis, 249, 289
Decimal numbering system, 190 Equipotential surveying, 8
Deconvolution, 197, 208, 221226, 247, 256, 259262, Exploding-reflector model, 86, 87, 280
264, 268, 289 Exploration seismology history, 3
Deconvolution, analysis, 249, 289
Deconvolution in frequency domain, 208, 222
Deconvolution objectives, 223 F
Deconvolution parameters, 262, 289 Facies changes seismic image, 314
Delay-time concept, 98 Faust formula, 56
Demultiplexing, 249251 Feathering effect, 154
Depth map, 301303 Field static correction, 256
Index 327
Marine data acquisition, 120 Planet Earth, 2, 4, 6, 8, 9, 20, 36, 43, 46, 53, 58, 59, 74,
Marine seismic sources, 118 89, 103, 105, 114, 116, 119, 134, 170, 173175,
Maximum delay wavelet, 178 177, 188, 222, 227, 259, 261, 291, 295, 298, 306,
Maxipulse, 118, 119 319
Migration aperture, 161, 164, 283 Point-diffractor, 82
Migration fringe, 160, 161, 164 Poissons ratio, 30, 32, 36, 46, 53, 292, 299
Migration parameters, 273, 282 Post-stack migration, 88, 272
Migration techniques, 272, 276, 284 Post-stack processing, 166, 248, 250, 268
Minimum delay wavelet, 178 Power spectrum, 206208, 220, 319
Multi-component recording, 297 Prediction error operator, 224, 261
Multiple reflections, 39, 75, 134, 155, 270, 293 Predictive deconvolution, 224, 261, 262, 289, 294, 295
Multiple reflections types, 293 Pre-stack migration, 81, 161, 249, 267, 272
Multiplexing unit, 122 Pre-stack processing, 166, 248, 249, 253
Mute analysis, 289 Processing objectives, 246
Muting, 243, 249, 267, 289 Processing of seismic reflection data, 86, 197, 243, 247
Processing sequence, 18, 166, 228, 240, 246, 247, 248,
286
N Progradation deposits, 309
Natural gamma-ray logging, 15 Propagating sine wave, 202, 203
Neutron gamma-ray logging, 15, 16 P-waves, 17, 44, 45, 53, 66, 73, 113, 291, 292, 319
NMO correction, 62, 67, 246, 247, 249, 256, 264267,
269, 287, 289
NMO-DMO correction, 165 Q
NMO in case of a dipping reflector, 78 Quality factor (Q), 48, 318
NMO in case of a horizontal reflector, 76
NMO in case of multiple horizontal reflectors, 78
NMO residual, 265, 269, 270, 287 R
Noise attenuation, 127, 249, 254, 286 Radioactivity logging, 14, 15
Noise test, 131, 133 Radioactivity method, 4, 9
Normal dispersion, 50, 58 Radon transform, 298
Normal move-out (NMO), 76, 84, 85, 243, 247 Ramp function, 289
Numbering system concept, 191 Ramp length, 267, 289
Nyquist frequency, 122, 163, 182187, 211, 228, 253 Random noise, 40, 127, 219, 224, 225, 254, 261, 286,
293, 294
Random noise attenuation, 127, 269
O Ray, 9, 12, 14, 15, 18, 41, 42, 61, 75, 79, 81, 86, 92, 107,
Oblique reflection pattern, 310 139, 141, 149, 267, 280, 281, 284, 295, 296
Octal numbering system, 191, 195 Rayleigh waves, 38, 39, 53
Octave, 199, 200 Reciprocity principle, 210
Oil-well drilling, 10 Rectangular (box-car) window, 216
One-sided function, 212 Red noise, 40
Onlap reflections, 308 Reef seismic image, 312
Reflection coefcient at inclined incidence, 72
Reflection coefcient at normal incidence, 73
P Reflection complex congurations, 309
Parallel reflection patterns, 308 Reflection conguration patterns, 308, 309
Parameter optimization, 166, 248, 249, 286, 289 Reflection from dipping reflectors, 74
Passive seismic surveying, 298 Reflection from horizontal reflectors, 75
Period of sine function, 198, 199 Reflection from multiple reflectors, 75
Periodic function, 35, 203205, 207 Reflection identication, 19, 302
Phase concept, 200 Reflection interface, 46, 71, 81, 86, 281, 299, 318
Phase of sine function, 198, 199 Reflection picking, 303
Phase removing ltering, 238 Refraction from a faulted refractor, 100
Phase spectrum, 171, 206, 207, 212214, 220 Refraction in a medium of linear velocity variation, 99
Phase velocity, 50, 51, 63, 227 Refraction in a multi-layer medium, 99
Piezoelectric substance, 120 Refraction proling, 101
Pinchout seismic image, 86 Refraction surveys, 101, 102, 257
Plane waves, 35 Refraction travel-time curve, 100
Index 329
Stack section, 69, 83, 86, 140, 142, 143, 151, 176, 187, Two-domain concept, 208
238, 240, 247, 248, 271278, 280, 281, 284, 286, Two-way transmission coefcient, 90, 91
289, 295, 300302, 304, 315, 317, 319
Stack-section distortion, 273
Static correction, 219, 243, 246, 247, 249, 256, 257, 261, U
271, 287, 288, 292 Unbalanced split spread, 106
Station lines, 150 Unconformity seismic image, 313
Steam-Gun, 118 Unipulse, 118, 119
Strain, 23, 2634, 36, 37, 73 Up-hole surveying, 91, 92, 135, 257
Stratigraphic interpretation sequence, 315
Streamer, 120, 153, 154
Stress, 1, 23, 24, 26, 2830, 32, 33, 73 V
Structural interpretation sequence, 302 Vaporchoc, 118, 119
Sublines, 150 Velocity analysis, 62, 64, 67, 78, 265, 287, 289
Subsurface coverage, 108, 109, 112, 295 Velocity analysis Method, 67
Surface consistent amplitude compensation, 256 Velocity by T DT Method, 67
Surface coverage, 108, 111 Velocity by X2 T2 method, 66
Surface waves, 36, 38, 39, 69, 115, 127129, 131, 254, Velocity-density relationship, 54
298 Velocity-depth relationship, 58
SV-waves, 38, 44, 73, 291 Velocity-elasticity relationship, 54
Swath parameters, 160, 161 Velocity function, 58-60, 64, 65, 67, 256, 287, 293, 303
Swath shooting, 150, 156, 157 Velocity-geological age relationship, 56
S-waves, 3638, 45, 46, 53, 54, 291, 292, 297, 299 Velocity map, 302, 303
Sweep length, 116 Velocity-overburden pressure relationship, 57
Sweep removal, 117, 249, 251 Velocity-porosity relationship, 55
Sweep taper, 158 Velocity pull-up, 311
Synthetic seismogram, 17, 19, 66, 174176, 295, 302 Vertical resolution, 187189
Vertical seismic proling (VSP), 18, 19, 91, 119,
292295, 303
T Vertical stacking, 114, 115, 118, 252, 254, 269, 270,
Tau-P transform, 298 292294
Telluric currents surveying, 9 Vibroseis, 3, 116119, 134, 178, 219, 251, 270
Template parameters, 160, 161 Voxel, 151, 163, 164
Temporal frequency, 35, 198, 202 VSP data processing, 294
Thumper, 114 VSP eld operation, 293
Time average equation, 54 VSP seismic record, 293
Time-depth curve, 303
Time-invariant equalization, 242
Time-lapse surveying, 297 W
Time slice, 152, 153 Wasp, 118, 119
Time-variant deconvolution (TVD), 260 Wave conversion, 43, 44, 46, 47, 73, 95, 170, 291
Time-variant equalization, 236, 242, 263 Wave dispersion, 49, 63
Time-variant ltering (TVF), 236, 241, 260, 285 Wave-front, 41, 42, 47, 63, 81, 116, 139, 148, 189, 277,
Time versus depth migration, 281 278
Toplap reflections, 308 Wave-front common tangent method, 278
Trace editing, 243, 249, 253 Wavefront-maximum convexity method, 277
Trace equalization, 240, 262 Wave generation, 43
Trace header assignment, 249, 253 Wavelength, 15, 35, 39, 49, 50, 63, 127129, 133, 178,
Trace interpolation, 284 187, 188, 198, 203, 284
Trace muting, 243, 267 Wave motion equation, 3234, 36, 54
Transient function, 205, 209 Wave number, 35, 63, 202, 238, 248
Transmission coefcient at normal incidence, 89 Wave scattering, 82, 311
Transmittance, 89 Wavelet attributes, 300, 315, 318
Transverse waves, 1, 37 Wavelet concept, 177
Travel-time corrections, 69, 247 Wavelet delay properties, 178
Triangular (Bartlett) window, 217 Wavelet energy, 178, 302, 307
True amplitude recovery (TAR), 246, 254 Weight dropping, 114, 134, 270, 293
Truncation effect, 215, 235 Well completion, 20
T-spread, 155 Well velocity surveying, 16, 17, 64, 91, 92, 292, 293
Index 331