Seismic Acquisition

and Processing

Petroleum Geoscience
Lecture Notes

Tutor: Ted Cox

Introduction to Exploration Seismic
___________________________________________________________________________________

INTRODUCTION TO
EXPLORATION SEISMIC

CONTENTS

INTRODUCTION Page No:

BASIC OBJECTIVES AND PRINCIPLES 1
STRUCTURE OF THE COURSE 2

SECTION 1: BASIC REFLECTION THEORY 3

1.1 THE BASIC SYSTEM 3

1.1.1 Types of Waves
1.1.2 Wavefronts and Raypaths
1.1.3 Elastic Deformation in rocks
Bulk Modulus, Shear Modulus, Young's modulus

1.2 THE REFLECTION SEISMIC METHOD 7

1.2.1 Polarity
1.2.2 Polarity Convention
1.2.3 Acoustic Impedance
1.2.4 Reflection geometry
1.2.5 Multiplicity
1.2.6 Stacking Diagrams
1.2.7 Digital Data Sampling
1.2.8 Aliasing

___________________________________________________________________________________
Ted Cox – April 2002 ii
Introduction to Exploration Seismic
___________________________________________________________________________________

SECTION 2: SEISMIC DATA ACQUISITION 20

2.0 SEISMIC SURVEY 20

2.0.1 Types of Survey
2.0.2 Costs

2.1 LAND OPERATIONS 21

2.1.1 People
2.1.2 Survey planning and design
2.1.3 Field equipment and layout
2.1.4 Vehicles
2.1.5 Surveying
2.1.6 Data Corrections
2.1.7 Energy Sources
2.1.8 Signal Recording
2.1.9 Receivers
2.1.10 Noise Analysis and Pattern Design
2.1.11 Cables
2.1.12 Recording Instruments

2.2 REMOTE OPERATIONS 35

2.2.1 Mountainous Regions
2.2.2 Jungles

2.3 MARINE OPERATIONS 37

2.3.1 Planning and design
2.3.2 People
2.3.3 Vessels
2.3.4 Positioning
2.3.5 Sources
2.3.6 Signal Recording
2.3.7 Receivers
2.3.8 Cables
2.3.9 Recording Instruments

2.4 SPECIAL MARINE APPLICATIONS 45

2.4.1 Dual Vessel Operations
2.4.2 Deep Water Acquisition

2.5 TRANSITION ZONE ACQUISITION 47

2.5.1 Survey planning and design
– Special Considerations
2.5.2 Techniques – Examples

___________________________________________________________________________________
Ted Cox – April 2002 iii
Introduction to Exploration Seismic
___________________________________________________________________________________

2.6 3D-ACQUISITION 48
2.6.1 Survey planning and design
2.6.2 Special Considerations
Line Orientation, Cost considerations
2.6.3 Resolution
Vertical and Lateral resolution
2.6.4 Survey Size
2.6.5 Review of existing data
2.6.6 The ‘Good’ Seismic Cycle
2.6.7 Land methods
2.6.8 Marine methods
2.6.9 Binning

2.7 HIGH RESOLUTION SURVEYS 52

2.8 4D SURVEYS 53

2.9 HEALTH & SAFETY 54

2.9.1 Procedures
2.9.2 Compliance with Standards
2.9.3 Safety Management
2.9.4 Reporting
2.9.5 Training

2.10 QUALITY CONTROL 56

2.10.1 Environmental Considerations
2.10.2 Pre-survey Testing
2.10.3 Operating Standards
2.10.4 Quality Assurance
2.10.5 Monitoring Operations
2.10.6 Field Processing
2.10.7 Reporting

2.11 LOGISTICS 59
2.11.1 Supply of Crews and Equipment
2.11.2 Contract Mobilisation / Demob.
2.11.3 Vital Back-up, Spares & Repairs
2.11.4 Electronic and Technical Support
2.11.5 Design, fabrication & testing
2.11.6 Training

___________________________________________________________________________________
Ted Cox – April 2002 iv
Introduction to Exploration Seismic
___________________________________________________________________________________

SECTION 3: SEISMIC DATA PROCESSING 61

3.0 PROCESSING OBJECTIVES 61

3.0.1 The Basic Processing Sequence

3.1 DATA PREPARATION 64

3.1.1 Initial Processing
3.1.2 Data Media
3.1.3 Demultiplexing
3.1.4 Editing
3.1.5 F-K Filtering
3.1.6 Digital Fourier Transform

3.2 DATA CORRECTION 70

3.2.1 Amplitude corrections
3.2.2 Overall Gain Correction (Gain Recovery)
3.2.3 Static and Dynamic Corrections
3.2.4 Static Corrections – Field Statics
3.2.5 Residual Statics
3.2.6 Long Wavelength Statics
3.2.7 Undesirable Effects
3.2.8 Sorting
3.2.9 Binning
3.2.10 Seismic Resolution - Vertical resolution
3.2.11 Horizontal (or Lateral) Resolution
3.2.12 Velocity Analysis and Stacking
3.2.13 Velocity Analysis Methods
3.2.14 Velocities for 3D Surveys
3.2.15 Dix's formula
3.2.16 Muting

3.3 DATA REDUCTION 93

3.3.1 CMP Stacking
3.3.2 NMO Multiple Removal

3.4 DATA ENHANCEMENT 94

3.4.1 Deconvolution
3.4.2 Convolution
3.4.3 Spiking Deconvolution
3.4.4 Predictive Deconvolution
3.4.5 Wavelet Deconvolution
3.4.6 Seismic Inversion
3.4.7 DMO and Migration
3.4.8 Dip Moveout (DMO)
3.4.9 Migration
3.4.10 Pre-Stack Depth Migration (PSDM)
3.4.11 Frequency filtering
3.4.12 Display parameters
3.4.13 Section Label
3.4.14 Data Formats
3.4.15 3D Data Processing
___________________________________________________________________________________
Ted Cox – April 2002 v
Introduction to Exploration Seismic
___________________________________________________________________________________

SECTION 4: SEISMIC INTERPRETATION 126

4.0 OVERVIEW 126

4.1 SEISMIC DATA INPUT 128

4.1.1 Display Technique
4.1.2 Stacked/Migrated Data
4.1.3 Pitfalls
4.1.4 Migration
4.1.5 Other problems
4.1.6 Multiples and Noise

4.2 SEISMIC CHARACTER 136

4.2.1 Reflection Character, Acoustic Impedance
4.2.2 Geological Controls on Velocity

4.3 INTERPRETATION METHODOLOGY 139

4.3.1 Basic Techniques
4.3.2 Seismic Reflectors
4.3.3 Reflection Polarity (again . . . )
4.3.4 Continuity
4.3.5 Seismic Sequences
4.3.6 Terminology
4.3.7 Types of Truncation
4.3.8 Delta and Basin Sands
4.3.9 Carbonates
4.3.10 Salt Structures
4.3.11 Structural Movement

4.4 INTERPRETATION OF 2D SEISMIC DATA 154

4.4.1 Reading the Seismic Label
4.4.2 Correlation with Geology
4.4.3 Well Log Data
4.4.4 Check Shot Surveys
4.4.5 Reflection Coefficient Log
4.4.6 Synthetic Seismograms
4.4.7 VSP Surveys
4.4.8 Where to Pick the Event
4.4.9 Looping Data, Common Problems
4.4.10 Picking Procedure
4.4.11 Misties

4.5 INTERPRETATION OF 3D SEISMIC DATA 166

4.5.1 Overview
4.5.2 General Approach
4.5.3 A Recommended 3D Interpretation Methodology
4.5.4 Attribute Displays

___________________________________________________________________________________
Ted Cox – April 2002 vi
Introduction to Exploration Seismic
___________________________________________________________________________________

4.6 CONTOURING AND MAPPING 175

4.6.1 Structure Maps
4.6.2 Isochron Maps
4.6.3 Hand Timing Seismic Sections
4.6.4 Manual Contouring
4.6.5 Computer Contouring

4.7 DEPTH CONVERSION 178

4.7.1 The Problem
4.7.2 Sources of Velocity Information
4.7.3 Vertical Depth Conversion Methods
4.7.4 Average Velocity
4.7.5 Interval Velocity
4.7.6 Velocity Function

4.8 CONCLUSIONS 185

COLOUR PLATES 182

BIBLIOGRAPHY

___________________________________________________________________________________
Ted Cox – April 2002 vii
Introduction to Exploration Seismic
___________________________________________________________________________________

INTRODUCTION TO
EXPLORATION SEISMIC

INTRODUCTION

Basic Objectives and Principles

Geophysical methods are used to look beneath the earth’s surface for geological structures
that may contain oil or gas reserves hidden from sight by buildings, vegetation, water and
by non-productive rocks.

Comparatively basic gravity and geomagnetic surveys (known as ‘potential fields’

methods) can identify potentially oil bearing sedimentary basins and major structures, but
more complex and costly seismic surveys are essential to discover and resolve hydrocarbon
prospects. Sedimentary rocks generally have low density and poor magnetic response, and
are often underlain by strongly magnetic and denser basement rocks. By measuring
anomalies, or variations from the regional average, a three-dimensional picture can be
calculated. In recent years, high-resolution aeromagnetic (HRA) surveys flown by
specially equipped aircraft at 50-150 m altitude have been used to delineate faults and near
surface volcanic rocks.

Detailed information about the sub-surface rock layers is obtained by deep ‘echo-sounding’
techniques (seismic reflection surveys). Offshore, these surveys are conducted by
specially-equipped vessels (PLATE 1A) towing both a submerged energy source to
produce short bursts of sound energy and one or more streamers several kilometres long
containing hundreds of detectors known as hydrophones. These register the echoes from
reflecting layers and the information is recorded digitally. The depths of the reflecting
layers can be calculated from the two-way travel time taken for the echoes to arrive. The
operation is repeated along a line many kilometres long. The amplitude of the echoes is
plotted against two-way time and displayed as ‘wiggle traces’.

The digital recordings are passed through many stages of signal processing and computer
summing to improve the quality of the reflections and bring out subtle geological features.
First, wiggle traces are sorted into ordered groups and, after various corrections, the
separate traces are added together, or stacked. This process allows reflections to enhance
one another and noise to be reduced. The stacked traces are displayed side by side to
produce a seismic section, normally displayed with a time scale, but sometimes converted
to depth in order to aid the visualisation of geology.

The final processed seismic sections provide 2-dimensional (2D) views of sub-surface
structure. Special shooting techniques that involve recording data closer together allow the
production of 3-dimensional (3D) seismic images of vertical sections and horizontal
sections (‘time-slices’).

___________________________________________________________________________________
Ted Cox – April 2002 Section 1: Theory Page 1
Introduction to Exploration Seismic
___________________________________________________________________________________

The interpreter then analyses these 3D sections to show geological structure down to fine
sedimentary details. Reflections are ‘picked’ and their depths plotted on maps and
contoured. Sections picked manually can be entered into a computer gridding and
contouring program. Contour maps are then generated to illustrate the subsurface structure.

Rock layers may be identified by comparison with log data from existing wells in which
the rock sequence is known. The seismic maps are used to identify structures that have
hydrocarbon potential; ‘high’ areas forming anticlinal domes or upthrown ‘horst’ fault
blocks. These are possible prospects into which oil or gas may migrated and been trapped.
In some instances subtle sedimentary traps can be identified involving changes in porosity.
High amplitude ‘bright-spots’ may indicate the presence of gas and a gas-fluid interface
may stand out as a distinctive ‘flat spot’ reflection amongst dipping reflections.

In the manner outlined above, the reflection seismic method has become the main
technique used by the oil industry to identify and delineate hydrocarbon prospects.

STRUCTURE OF THE COURSE

This course is divided into four sections, and sets out to explain, with the use of talks and
workshops, the following topics:

1. BASIC REFLECTION THEORY:

The physical principles of the seismic reflection method,

2. SEISMIC DATA ACQUISITION:

How to obtain data in the field (often in difficult circumstances),

3. SEISMIC DATA PROCESSING:

The steps necessary to convert field data into a meaningful format,

4. SEISMIC INTERPRETATION:

The understanding of geology from the seismic data.

___________________________________________________________________________________
Ted Cox – April 2002 Section 1: Theory Page 2
Introduction to Exploration Seismic
___________________________________________________________________________________

SECTION 1:

BASIC REFLECTION THEORY

The physical principles of the seismic reflection method.

1.1 THE BASIC SYSTEM

In the words of Mike Badley in his book Practical Seismic Interpretation, “Usually, it all
begins with a bang . . .which sends a short, sharp pulse of sound into the ground”.

This seismic energy makes rock particles move in a sinusoidal manner about their
undisturbed positions. The energy radiates out from this point source through the earth as
an expanding wavefront, which will be spherical or ellipsoidal in shape depending on
whether the propagating medium is homogeneous or heterogeneous.

Simplistically, one can picture the ‘shock wave’ generated as being similar to the ripples
spreading on the surface of a pond into which a stone has been thrown, or the blasts
spreading outwards as seen on aerial photographic films of bombing raids.

1.1.1 Types of Waves

Three types of seismic wave can be generated:

i) P (longitudinal) waves - particle motion is backwards and

forwards, parallel to the direction
of wave propagation (like the
shunting of railway trucks). (Fig 1.1)

ii) S (transverse) waves - particle motion is from side to side,

perpendicular to the direction of wave
propagation (like the shaking of a
skipping rope).

iii) Boundary waves - low velocity, low frequency surface

waves where particle motion is complex
(e.g. Rayleigh waves also known as
ground roll).

Collectively P and S waves are called body waves because they travel through the body of
the earth. S waves travel at approximately half the speed of P waves. Their behaviour can
be simulated by the use of a long coil spring. Creating a compressional or transverse
displacement at one end, the different modes of propagation can be seen to travel along it,
and even ‘reflect’ back if the far end is fixed.

___________________________________________________________________________________
Ted Cox – April 2002 Section 1: Theory Page 3
Introduction to Exploration Seismic
___________________________________________________________________________________

Figure 1.1 – Propagation of compressional (P-wave) energy

If we were to measure the disturbance of the earth as this shock wave passes, i.e., against
time, we would see the response shown in Figure 1.2.

Figure 1.2 – The earth’s ideal response to a P wave

1.1.2 Wavefronts and Raypaths

Depending on the objective, we can view seismic waves in two different ways, as
wavefronts and raypaths (Fig. 1.3).

___________________________________________________________________________________
Ted Cox – April 2002 Section 1: Theory Page 4
Introduction to Exploration Seismic
___________________________________________________________________________________

S Æ R = Direct (surface) ray

S Æ B’ Æ R = Reflected ray
S Æ B Æ B” Æ R = Refracted ray
S Æ B Æ O = Transmitted ray

Figure 1.3 – Wavefronts and Raypaths

• Raypaths - Raypaths are nothing more than lines that show the direction
that the seismic wave is propagating. For any given wave, there is an
infinite set of raypaths that could be used. In the example shown above, for
instance, a valid raypath could be any radial line drawn from the source.

• Wavefronts - Wavefronts connect positions of the seismic wave that are

doing the same thing at the same time. In the example shown above, the
wavefronts are spherical in shape. One such wavefront would be the result
of an airgun array firing in deep water. The wavefront connects all portions
of the wave that have equal amplitude at a particular time.

Wavefronts are generally perpendicular to raypaths. Given either a set of wavefronts or a

set of raypaths, we can construct the other.

___________________________________________________________________________________
Ted Cox – April 2002 Section 1: Theory Page 5
Introduction to Exploration Seismic
___________________________________________________________________________________

1.1.3 Elastic Deformation in rocks

In simple terms, if we apply a stress to a rock, it suffers strain. It will also deform
elastically up to the elastic limit, above which it fractures.

It can be shown that in homogeneous, isotropic media the velocities of P and S waves
through the media are given by the expressions shown below.

Vp = (4µ/3 + k ) / ρ

Vs = µ /ρ

Where Vp and Vs are the P and S wave velocities of the medium, ρ is the density of the
medium, and µ and k are referred to as the shear and bulk moduli of the media. Taken
together, µ and k are also known as elastic parameters. The elastic parameters
quantitatively describe the following physical characteristics of the medium:

• Bulk Modulus (k) is also known as the incompressibility of the medium, or

its resistance to squeezing from all sides. If k is very large, e.g., solids and
liquids, then the material is very stiff, meaning that it only compresses
slightly even under large pressures. If k is small, like a gas, then a small
pressure can compress the material by large amounts.

• Shear Modulus (µ) describes how difficult it is to deform a cube of

material under an applied shearing force. If the base of the cube is firmly
fixed, sideways pressure on one of its top edges will cause deformation.
Material with a small shear modulus will easily deform in the direction of
pressure. Material with a large shear modulus will take a larger force to do
this. Gases and fluids cannot support shear forces, i.e., µ = 0. From the
equations above, you will notice that fluids and gases do not allow the
propagation of S waves.

Shear Modulus, µ = ρ * Vs2

• Young's modulus E and Poisson's ratio φ are also related to Bulk Modulus
by the formula :

k = E/(3(1-2φ))

• Prior to the elastic limit, the relationship between stress and strain is linear,
and defined by:
σ = Eε ,

where E = Young's modulus (the constant of proportionality) and σ = stress

___________________________________________________________________________________
Ted Cox – April 2002 Section 1: Theory Page 6
Introduction to Exploration Seismic
___________________________________________________________________________________

1.2 THE REFLECTION SEISMIC METHOD

In order to determine the distribution of different rock layers beneath the earth’s surface,
we can use the rock properties described above. A series of impulses of energy (or ‘shots’)
are put into the ground at different locations. Some of this energy is reflected back from the
rock interfaces beneath. Detectors, or receivers, located at regular intervals on the surface
will receive this energy and allow us to record and measure changes in the response
characteristics of different rock types (Figures 1.4 and 1.5). These recordings are
conducted along a series of seismic lines (or profiles) arranged in a grid. After some
intensive data processing, they will allow us to estimate the nature and distribution of the
geological formations.

Figure 1.4 – Configurations of shots and detectors

(a) Split spread or straddle spread
(b) Single-ended or end-on spread

___________________________________________________________________________________
Ted Cox – April 2002 Section 1: Theory Page 7
Introduction to Exploration Seismic
___________________________________________________________________________________

Figure 1.5 – Example Shot Record with detectors arranged in split spread
configuration.

1.2.1 Polarity

A seismic (P) wavelet received by a detector at the surface represents either a

compressional or a dilational reflection. The wavelet is part of a sub-circular wavefront. A
seismic ray is the locus of the wavefront and is thus perpendicular to it. A compressional
(or positive polarity) reflection is one where push (compression) occurs before pull
(dilation or rarefaction); a dilational (or negative polarity) reflection is where pull occurs
before push. The surface and hence the detector will continue to oscillate after the initial
reflection is received because of a train of wobbles following on at the tail of the seismic
wavelet (Figure 1.6). The energy of the wavelet is proportional to the area, either positive
or negative, defined by the amplitude versus time curve, and will thus be approximately
proportional to the square of the amplitude.

___________________________________________________________________________________
Ted Cox – April 2002 Section 1: Theory Page 8
Introduction to Exploration Seismic
___________________________________________________________________________________

Figure 1.6 – Positive polarity wavelet generated in the earth by a

compressional wave.

(Note that one half of the oscillation is coloured in black to make it easier to
see)

1.2.2 Polarity Convention

Figure 1.6 shows a typical wavelet for the earth’s response to a P-wave reflected back to
the surface. We have chosen to colour half of the wavelet oscillations in black, to make
them easier to see. As we shall see in the next section, the type of wavelet that we record
will depend on our energy source and the types of rock the shock wave passes through.

In order to derive full benefit from this information, it is important to have some standard
convention by which the same parts of the wavelet are always coloured consistently.

The Society of Exploration Geophysicists (SEG) has set a standard for the industry to
follow. This states that “the onset of a compression from an explosive source is represented
by… a downward deflection when displayed graphically” (see Fig. 1.6). This means that
for a positive polarity wavelet, the first break from zero amplitude will be displayed as a
‘white trough’, although the strongest part of the wavelet will appear as a black peak.

As you might expect, the earth is not as simple as this, and we shall be revisiting the
concept of polarity a number of times during our studies.

1.2.3 Acoustic Impedance

The acoustic impedance (AI) of a rock is defined as the product of density (ρ) and P-wave
velocity (V),
i.e., AI = (ρ x V).

A compressional wavelet is reflected back to surface with no polarity change where a layer
of low acoustic impedance overlies a layer of high acoustic impedance. This boundary is
said to have a positive reflection coefficient 1 and is sometimes called a ‘hard kick’.

The wavelet reflected back to surface where a layer of high acoustic impedance overlies a
layer of low acoustic impedance is dilatational. This boundary is said to have a negative
reflection coefficient and is sometimes called a ‘soft kick’. This reflection exhibits a
polarity reversal of the downgoing wavelet (Figures 1.7).

___________________________________________________________________________________
Ted Cox – April 2002 Section 1: Theory Page 9
Introduction to Exploration Seismic
___________________________________________________________________________________

Figure 1.7 – Effects of Reflection Coefficient, R

(a) Reflection from an interface with positive reflection

coefficient.
(b) Reflection from an interface with negative reflection
coefficient –
Note how the latter produces a reflected wave of reversed
polarity.

The proportion of the energy reflected from a boundary depends on the value of the
reflection coefficient. For example, a boundary with a reflection coefficient of +0.1 will
reflect 10% of the seismic energy as a compressional wavelet; a boundary with a reflection
coefficient of -0.2 will reflect 20% of the seismic energy as a dilational wavelet.

In both cases, for any ray, the angle of incidence to the boundary is equal to the angle of
reflection and both rays are symmetric about a line perpendicular to the boundary. Most of
the remaining seismic energy is transmitted downwards into the underlying layer. It will
be refracted according to Snell's Law 2 as it crosses the boundary (Figure 1.8).

___________________________________________________________________________________
Ted Cox – April 2002 Section 1: Theory Page 10
Introduction to Exploration Seismic
___________________________________________________________________________________

Figure 1.8 - Angles of Incidence, Reflection and Refraction.

1
The full mathematical expressions governing the reflection amplitudes are
called Zoeppritz's equations. These can be simplified for normal incidence:

Reflection Coefficient:
R = (ρ2V2) – (ρ1V1)
(ρ2V2) + (ρ1V1)

where layer 1 overlies layer 2, ρ is density and V is P-wave velocity.

2
Snell’s Law of Refraction:
sin θ2 = V2
sin θ1 V1

where
θ1 = angle of incident ray with the normal through boundary 1--2;
θ2 = angle of refracted ray with the normal below the same boundary.

Part of the seismic energy returning to surface as deep reflections will be reflected back
into the earth from acoustic impedance boundaries at shallower levels. However, the
polarity of these downward-reflected wavelets is opposite to that of the upward-reflected
(primary) wavelets. If the downward-reflected wavelets are once more reflected upwards
from a deeper acoustic impedance boundary they are known as multiples (Figure 1.9).

Figure 1.9 – Different Multiple Paths.

___________________________________________________________________________________
Ted Cox – April 2002 Section 1: Theory Page 11
Introduction to Exploration Seismic
___________________________________________________________________________________

A seismic wavelet measured at surface consists of a spectrum of different frequencies.

One cycle consists of a peak and a trough and the time taken for the complete cycle is
known as the period of the oscillation (Figure 1.10). Period (τ) is measured in fractions
of a second. The inverse of period is frequency (F), which is measured in Hertz (the
number of cycles per second). Period and frequency (measured in the time domain) are
analogous respectively to wavelength (λ) and wave number (K) measured in the space
domain.

Figure 1.10 – Comparison of the Space and Time Domains.

One periodic oscillation can be described mathematically in terms of a cosine and/or a sine
curve. A cycle which begins with a peak amplitude is known as zero phase and is a pure
cosine function. A cycle which begins at zero amplitude and rises to a peak is known as
minimum phase and is a pure sine function (Figure 1.11).

Figure 1.11 – SEG Standard Polarity positive wavelets.

Minimum Phase (top) begins with a down kick,
Zero Phase (bottom) shows symmetrical black peak.

Any cycle which does not begin at maximum, minimum or zero amplitude can be shown to
consist of a combination of one cosine curve and one sine curve having the same period.

___________________________________________________________________________________
Ted Cox – April 2002 Section 1: Theory Page 12
Introduction to Exploration Seismic
___________________________________________________________________________________

Another way of representing these combined curves is as a shifted (or delayed) cosine
function. The shift is usually measured as an angle in radians and is know as the phase
angle. The phase varies between -π rad (-180 o ) and π (180 o ) (Figure 1.12). A zero phase
cycle has a phase angle of 0, i.e. no shift. A minimum phase reflection has a phase angle
of π/2.

Figure 1.12 – Frequency Spectra for a waveform.

(a) Amplitude Spectrum
(b) Phase Spectrum

A composite wavelet can be broken down into component frequencies with each frequency
having a specific phase angle and amplitude (Figure 1.13). This process, which is known
as frequency analysis, is done during seismic processing using a mathematical technique
known as the Fourier transform. In theory, frequency analysis can also be done manually
by cross-correlating the wavelet with the component frequencies, but it takes a long time!

___________________________________________________________________________________
Ted Cox – April 2002 Section 1: Theory Page 13
Introduction to Exploration Seismic
___________________________________________________________________________________

(a) (b)

Figure 1.13 – Composition of a seismic wavelet.

(a) Zero phase wavelet with max. amplitude of all frequencies at time zero (dashed line),
(b) Mixed Phase wavelet with frequencies having different amplitudes at time zero.

A mixed phase wavelet can be converted to an approximate zero phase wavelet by a

process known as deconvolution. Deconvolution involves cross-correlating an
approximation of the input wavelet with the actual wavelet. If deconvolution works well
then a cross-correlation spike or narrow zero phase wavelet will be produced that
represents the reflection coefficient of the originating reflection boundary. This process
will be discussed more fully in Section 3: Seismic Data Processing.

1.2.4 Reflection geometry

It is normal to record the response to each shot at a large number of regularly spaced
receiver locations. Ray paths can be constructed to predict the part of the subsurface that
will be sampled by a particular combination of source and receiver location. In order to do
this, and in the absence of any better information, it will be assumed that the reflector is
horizontal and hence the angles of incidence and reflection are equal.

The point on the surface equidistant between source and receiver is termed the ‘common
mid point’ (CMP). This can be related to the point on the reflector vertically beneath
which is known as the ‘common depth point’ (CDP). For each source position there will
therefore be a large number of CDP’s sampled, as defined by the different receiver
positions (Figure 1.14).

___________________________________________________________________________________
Ted Cox – April 2002 Section 1: Theory Page 14
Introduction to Exploration Seismic
___________________________________________________________________________________

(a) (b) (c)

Figure 1.14 – The Common Mid Point Technique.

(a) Single Source and Receiver
(b) Two Sources and Receivers
(c) Six Source - Receiver Pairs

1.2.5 Multiplicity

As the seismic survey operation proceeds along a line, the source and receiver positions are
varied by regular short distances. These intervals are chosen so that different CDP’s will be
sampled more than once, by different source and receiver configurations. The number of
times a particular CDP is sampled is known as the multiplicity, or ‘fold of coverage’
(Figure 1.15).

Figure 1.15 – Multiplicity (300% or 3 Fold Coverage).

This multiplicity is a fundamental factor in data processing to aid noise cancellation. After
appropriate corrections have been made, traces from the same CDP can be added together,
or stacked, to enhance the true reflection information at the expense of background noise.

The multiplicity or fold of the data can easily be calculated by the formula:

___________________________________________________________________________________
Ted Cox – April 2002 Section 1: Theory Page 15
Introduction to Exploration Seismic
___________________________________________________________________________________

M = N.S
2.n

where M = multiplicity (e.g. 120)

N = number of detector stations in spread
S = number of shots per spread (normally 1)
n = number of stations by which spread is advanced

1.2.6 Stacking Diagrams

The CMP is a point of reference used in defining all field geometries. It is useful to be able
to portray visually the build-up of CMP coverage both for 2D and 3D surveys. This can be
done by means of ‘Subsurface Coverage Charts’ or ‘Stacking Diagrams’.

There are various versions of the diagram, but each shows graphically the relative build up
of CMP fold relative to shot and detector positions (Figure 1.16). For 3D surveys, the
display is usually in the form of a map with a colour display to indicate the degree of
multiplicity.

Figure 1.16 – Subsurface Coverage Chart (3-fold Coverage).

___________________________________________________________________________________
Ted Cox – April 2002 Section 1: Theory Page 16
Introduction to Exploration Seismic
___________________________________________________________________________________

1.2.7 Digital Data Sampling

One further topic of theory needs to be introduced here, although it will be discussed again
in Sections 2 and 3. This is the concept of Digital Data Sampling, which is the
fundamental process of converting the earth's response to our impulse sources as they are
received at the geophones into digital data values for computer analysis.

In the early days of seismic prospecting, the response at each geophone was observed by
plotting the results against elapsed time by means of an analogue pen trace on paper, or by
the oscillation of a mirror galvanometer reflecting light onto photographic film to produce
a 'monitor record'. These records were produced in real time and hence could be studied
directly in the field.

Whilst monitor records are still important today for quality control purposes, technology
has advanced to the point where a far greater amount of data will be recorded and the
results stored digitally on magnetic tape cartridge for detailed computer processing 'on the
fly' and also at a later date. In order to achieve this, a technique known as sampling is
performed by means of a special device that converts the analogue signals from the earth
into discreet digital data values. This device is known as a 'multiplexer' and behaves very
much like a 'rotating switch' (Figure 1.17).

after Brian J Evans

Figure 1.17 – The Multiplexing System

___________________________________________________________________________________
Ted Cox – April 2002 Section 1: Theory Page 17
Introduction to Exploration Seismic
___________________________________________________________________________________

Basically, each input channel of data representing a geophone (or group of geophones) is
sampled in turn at a regular short time interval such that the oscillating analogue waveform
can later be reconstructed from the sample values. The process can be pictured as the
switch rapidly rotates making contact (thus sampling the analogue value) with the first and
then each successive channel before returning to make the second sample from the first
channel. The whole process thus continues until all the channels have been sampled for
the required length of time (i.e., until all the reflected energy has been received).

Each analogue sample value is converted into a digital number, which is in turn written to a
magnetic tape or cartridge in the data recorder. The order in which the values are written
onto the seismic 'field tape' will be in one of a number of standard formats defined by the
Society of Exploration Geophysicists (SEG) and adopted universally, with minor but often
annoying modifications. (See Section 2.1.12 for more details). The format information
must be noted with the magnetic tape, as it will be required when the tape is read for Data
Processing (see Section 3.1.3).

1.2.8 Aliasing

As so often happens with a partial solution, there are disadvantages in the use of this
technique. Sampling of the waveform must be adequate (i.e., at a high enough rate) to
allow reconstruction of the original seismic frequencies from the sampled values.

The sampling theorem states that if f(t) the time series of values (our sampled waveform)
is band limited (to a certain range of frequencies = fm Hertz, then f(t) will be completely
characterised by samples taken at uniform intervals less than 1/(2fm) seconds apart.

For example, if the highest expected frequency were 100 Hz (i.e., 100 cycles per second),
sampling should be at 1/200 or 0.005 seconds (every 5 milliseconds). For practical
reasons, sampling rates of 1, 2, 4 (and rarely 8) ms are used in seismic recording, so in
this case a rate of 4ms should be used. This sampling rate is sometimes called the
sampling interval.

If a waveform is not sampled at a rapid enough rate, the phenomenon known as aliasing
will occur, and a lower frequency (hence incorrect) waveform will be recovered. This is
illustrated in Figure 1.18.

___________________________________________________________________________________
Ted Cox – April 2002 Section 1: Theory Page 18
Introduction to Exploration Seismic
___________________________________________________________________________________

after Dale G Stone

Figure 1.18 – Two Examples of Waveform Aliasing

The maximum defined rate for a particular frequency is known as the "Nyquist" rate and
similarly, the maximum defined frequency for a particular sampling rate is known as the
Nyquist frequency. It can be demonstrated, however, that under some circumstances a
waveform of the maximum frequency will not be completely restored by using the Nyquist
rate. In practice, therefore, sampling will be performed at twice the Nyquist rate to ensure
faithful recovery of the maximum required frequency.

We shall encounter other instances of aliasing in the following sections of the course.

___________________________________________________________________________________
Ted Cox – April 2002 Section 1: Theory Page 19
Introduction to Exploration Seismic
___________________________________________________________________________________

SECTION 2:

SEISMIC DATA ACQUISITION

How to obtain data in the field (often in difficult circumstances).

The theory of the reflection seismic method has been outlined in the previous section. The
aim of seismic acquisition is to record these reflection data digitally in the field at optimal
cost and quality.

2.0 SEISMIC SURVEY

A seismic survey crew acquires the data. A crew consists of seven components:

• people
• vessels or vehicles
• sources
• receivers
• cables
• recording instruments
• positioning systems

The design of the survey is constrained by geography, timescale, weather, economics,

politics, permits and geological and geophysical parameters. Seismic data can be acquired
on land or at sea. The most expensive and difficult surveys are often those in transition
zones such as marshes or surf zones where a combination of land and marine techniques
are used.

2.0.1 Types of Survey

As the logistics and composition of field crews vary greatly depending on the location and
nature of the survey, we shall examine the components and requirements of each type of
survey in turn:

• Land Surveys
• Remote Operations
• Marine Surveys
• Deep Water Surveys
• Transition Zone Crews
• 3D Surveys – Special Requirements
• High-Resolution Surveys
• 4D Surveys

___________________________________________________________________________________
Ted Cox – April 2002 Section 2: Acquisition Page 20
Introduction to Exploration Seismic
___________________________________________________________________________________

2D seismic surveys are normally shot in a grid pattern to allow reconstruction of the 3-
dimensional geological structure in interpretation. The line spacing will normally be of the
order of kilometres.

3D seismic surveys are generally acquired as a series of ‘swaths’ or ‘patches’ of more

closely spaced points to allow a greater resolution of the subsurface. Typical line spacings
will be of the order of tens of metres. This has inevitable cost implications, and 3D has
traditionally become the accepted method of survey for reservoir and field definition.

2.0.2 Costs

The costs of data acquisition represent approximately 70% of the total costs worldwide
spent on the reflection seismic technique by the oil industry. A further 10-15% is spent on
data processing and less than 10% is invested in interpretation.

At present it costs about $500 per line km of 2D marine seismic and about $4000 per line
km or land seismic. However, the latter depends very much on the accessibility and type
of terrain being covered. A typical marine 3D seismic survey may cost in the order of
£10,000 per km 2 in the North Sea. Other geophysical information such as gravity and
aeromagnetic data might be acquired at the same time as the seismic survey.

2.1 LAND OPERATIONS

2.1.1 People

Wherever a survey is being conducted, seismic field crew personnel have to be rugged
individuals who are capable of working as part of a team having to produce high-quality
results, safely, often in extremely uncomfortable conditions, under pressure of deadlines,
for long periods cut off from civilisation. It’s little wonder they often have difficulty in
rehabilitating when their work in the field ends!

The composition of a typical crew is shown in Figure 2.1.

Figure 2.1 – Field Crew Organisation Chart

___________________________________________________________________________________
Ted Cox – April 2002 Section 2: Acquisition Page 21
Introduction to Exploration Seismic
___________________________________________________________________________________

2.1.2 Survey planning and design

Before a survey can begin, it is necessary to define the geological objectives and establish
whether there will be any operational problems such as gaining access to the area,
topographical obstacles, zones of particular ecological or political sensitivity. It is therefore
necessary to undertake a thorough research of the area over which the survey will be
conducted. It will be the job of the contractor’s permit man to travel the area to prepare a
way for the seismic crew. He must ensure that local people understand what is about to
happen, and that permission is granted for the crew to cross private land; a process often
achieved by making the appropriate payments.

2D seismic surveys are generally acquired as a regular orthogonal grid of lines, parallel to
and perpendicular to the general direction of structural dip. The spacing between the
seismic lines depends on the approximate size of the structures that are to be mapped. The
line spacing must be less than half the estimated wavelength of the folds or fault blocks.
The wavelength of the structures can be estimated from existing seismic data or, where
these are not available, from gravity or aeromagnetic data. It is also essential to intersect
any wells that have been drilled in an area, both in the dip direction and the strike direction.
This will allow the interpreter to correlate the reflections with the geology encountered in
the wells.

The orientation of the grid of dip and strike lines will normally depend on three factors:
• the geological ‘grain’ and perceived trend of structures (more lines are
normally programmed in the dip direction),
• the spacing and size of the structures (dictating the dimensions of the grid),
• and last, but by no means least, cost constraints.

Wherever possible the designer should try to:

• Keep line spacing equal
• Keep lines straight
• Intersect any wells in two directions
• Avoid restricted areas
• Leave out areas of no importance (if costs are critical)
• Allow lines to extend approx. 2 km. beyond area of interest to allow
migration of dipping reflectors during processing (see Section 3.4.8).

Most of this work is necessary to define the basic survey specifications before a seismic
company will be asked to quote for the work. It is quite common for the exploration
company to hire acquisition design specialists to perform this task.

___________________________________________________________________________________
Ted Cox – April 2002 Section 2: Acquisition Page 22
Introduction to Exploration Seismic
___________________________________________________________________________________

2.1.3 Field equipment and layout

Following the recommendations of the survey design team, a seismic crew will be
commissioned to perform the survey, and will mobilise into the area.

The equipment inventory for a 3D land crew is shown in Figure 2.2. It will comprise
various quantities of all of the following elements.

• positioning systems
• sources
• receivers
• cables
• recording instruments

Figure 2.2 – Land Crew Field Equipment Inventory.

___________________________________________________________________________________
Ted Cox – April 2002 Section 2: Acquisition Page 23
Introduction to Exploration Seismic
___________________________________________________________________________________

2.1.4 Vehicles

The range of vehicles used on a crew will vary depending on the nature of the survey. In
areas of extreme vegetation (forests and jungles) or rugged topography (mountain ranges) a
minimum number of vehicles will be used, and all equipment made as portable as possible.

Where access is moderate to good, the crew will comprise a larger number of vehicles
including specially designed units to house the recording instruments, the source
equipment, the distribution and collection of detectors, as well as safety equipment and
other personnel transport. A typical inventory for a 3D desert crew is shown in Figure 2.3.

Figure 2.3 – Transport Inventory for a Field Crew.

___________________________________________________________________________________
Ted Cox – April 2002 Section 2: Acquisition Page 24
Introduction to Exploration Seismic
___________________________________________________________________________________

2.1.5 Positioning Systems

The seismic contractor is responsible for obtaining all the topographic data needed for the
survey. Most of the positioning used on today’s seismic crews is obtained using GPS – the
Global Positioning System of Earth orbiting satellites (Figure 2.4). Aerial and satellite
photographs will be used digitally to compile a map of the area containing all prospect
boundaries, exclusion zones and geo-hazards. Comparison of this with the desired
programme map will allow adjustment of the line positions to avoid field constraints.

Figure 2.4 – Differential Global Positioning System (DGPS).

Before detailed surveying can take place, it is necessary to set up at least 3 base stations,
the positions of which are very accurately fixed by monitoring GPS signals over a period
of several days. The base lines between these stations are calculated and a secondary infill
network of control reference points generated. These secondary control points are placed
close to the start and end of lines. Local readings are compared with a global geoid (record
of the earth’s surface elevations) in order to derive elevations relative to Mean Sea Level
(MSL) within the local network. Calibration of the system is checked monthly while the
survey continues.

Typical equipment used are 8 channel dual frequency receivers capable of logging GPS
data. An example is shown in Plate 2A.

The staking out (or marking) of each source and detector position on the lines is then done
using portable GPS receivers. UHF band radios are used to communicate between the
mobile units and the base stations. This method is rapid and 95% accurate to within 0.2
metres. Each day’s survey data is downloaded into specialist positioning software for QC
purposes.

In areas where the use of RTK GPS is not possible (radio-masked or urban areas), all
source and detector points must be staked out and surveyed using a traditional optical plane

___________________________________________________________________________________
Ted Cox – April 2002 Section 2: Acquisition Page 25
Introduction to Exploration Seismic
___________________________________________________________________________________

table triangulation method. Field data are loaded into a data-logger (e.g., PCMCIA card)
and downloaded each evening. Azimuth control is provided by sun-shots and by sighting
known permanent markers.

During acquisition, the source positions of mobile source units are achieved by providing
the Differential GPS (DGPS) receiver in each unit with a pre-programmed SPS file, and
the actual positions during recording compared with the programme in real time by radio
link with a reference station. This normally allows operations to proceed with positional
accuracy within 5 metres. All point co-ordinates are downloaded into field management
software and used to QC actual against theoretical positions. Once, verified, the co-
ordinates can be output as positional files (in UKOOA or SPS formats) used to prepare
shot point maps for field QC and processing.

2.1.6 Data Corrections

Certain corrections must be applied during processing to compensate for some field effects.

One correction needed is static correction which will allow all shots and receivers to be
related to a common time reference datum. Variations in the relative height of different
source and receiver locations and the effects of shallow weathering are corrected for by
means of a ‘static’ or fixed time-shift.

The other correction needed is the ‘Normal Move-Out’ correction (or NMO). This
compensates for the time delay resulting from the varying horizontal separation (or
‘offset’) of the source and receiver, since we only wish to compare the vertical component
of travel times.

These topics will be discussed more fully in the Processing Section, but it is necessary for
the survey crew to record and communicate details of the field geometry in order that these
corrections can be calculated and applied at the processing centre.

2.1.7 Energy Sources

Signal Generation

Seismic sources are required to produce a high output of elastic energy containing a broad
band of frequencies that is stable and repeatable.

Two main sources are used in land acquisition, namely dynamite and Vibroseis. Large
explosions produce the best energy output but dynamite is only used in certain closely
controlled circumstances nowadays because of environmental and security restrictions.

The use of dynamite involves drilling a shot hole to below the weathered layer (usually
10-50m deep) and letting off a controlled explosion. A good signal is usually produced
and the thickness and seismic velocity of the weathered layer can be measured directly for
static corrections. However, it is quite time consuming and therefore expensive. Dynamite

___________________________________________________________________________________
Ted Cox – April 2002 Section 2: Acquisition Page 26
Introduction to Exploration Seismic
___________________________________________________________________________________

is usually not feasible where buildings, warring people or fragile ecosystems exist. Also,
the signal generated may not be consistent between shots.

Vibroseis is the main alternative to dynamite. The Vibroseis technique employs 4 or 5

trucks with heavy metal plates, or pads, beneath them that can be lowered onto the ground
surface (Figure 2.5 and Plate 2B). These pads are used to generate a known seismic
signal. This signal, which is known as a sweep, lasts typically for 20 seconds and begins at
a low frequency and continues through to a high frequency. The trucks line up and
generate the sweep together before moving to the next source position. In this way a
significant amount of seismic energy is generated, albeit over a much longer period of time
than an explosive source. Each reflection will consist of a delayed, low amplitude copy or
inverted copy of the original sweep (Figure 2.6). Therefore, the raw reflection data
recorded during a vibroseis survey is unintelligible until it has been digitally correlated
with the initial signal (Figure 2.7). The thickness and velocity of the weathered layer is
calculated from the time of refraction arrivals.

Vibroseis surveys can only be carried out where trucks can gain access. However, they
cause a minimum of disturbance in built up areas because the survey can be carried out
along roads. It has even been reported that the vibrations generated during a vibroseis
survey have an aphrodisiac effect!

Figure 2.5 – Vibroseis Truck with pad lowered and ready to vibrate.

___________________________________________________________________________________
Ted Cox – April 2002 Section 2: Acquisition Page 27
Introduction to Exploration Seismic
___________________________________________________________________________________

Figure 2.6 – Vibroseis Method – Basic Principles.

Figure 2.7 – Vibroseis Correlation in detail.

___________________________________________________________________________________
Ted Cox – April 2002 Section 2: Acquisition Page 28
Introduction to Exploration Seismic
___________________________________________________________________________________

2.1.8 Signal Recording

Recording the reflected energy is a more complicated it process which requires three
elements:
- receivers
- cables
- recording instruments

These will be dealt with in the following paragraphs. The overall field layout is illustrated
diagramatically in Figure 2.8.

Figure 2.8 – Schematic Field Layout.

2.1.9 Receivers

Geophones are the detectors used on land to measure seismic response. Geophones
convert ground motion into electrical voltage by the relative motion between a sprung coil
and a magnet (Figure 2.9). In standard seismic acquisition it is only P (compressional)
waves that are of interest. Geophones are therefore designed that theoretically only allow
for vertical particle motion, i.e. P waves that are coming from depth. S waves and complex
waves are may be recorded for special applications using 3-component geophones.

___________________________________________________________________________________
Ted Cox – April 2002 Section 2: Acquisition Page 29
Introduction to Exploration Seismic
___________________________________________________________________________________

Figure 2.9 – Geophone cut away to show internal moving coil.

A seismic survey comprises a line or spread of recording stations set at regular distances
away from the source. Each recording station may consist of 12-120 geophones, normally
with a 5 – 10 cm spike on the base, which must be ‘planted’ in good contact with the earth
in a regular pattern. The number and spacing of geophones in the pattern will have been
designed to help suppress horizontally travelling ground roll and random noise (Figure
2.10).

___________________________________________________________________________________
Ted Cox – April 2002 Section 2: Acquisition Page 30
Introduction to Exploration Seismic
___________________________________________________________________________________

Figure 2.10 – The use of a geophone pattern to attenuate noise.

As can be seen from Figure 2.10, vertical reflection signals will arrive at each detector at
the same time, so the response of the pattern is proportional to the number of detectors.
Organised noise, such as ground roll, arrives at different times and suffers destructive
interference. Non-coherent (random) noise is also reduced by interference.

2.1.10 Noise Analysis and Pattern Design

A seismic wavelet suffers many different types of degradation as it travels downwards into
the earth and is reflected back again (Figure 2.11). We shall be discussing many of these
later in the Processing Section. However, it is good field practice to ensure that as much
noise is eliminated from the record as possible rather than rely on the processing centre to
come to the rescue!

Figure 2.11 – Degradation of the Seismic Wavelet.

For this reason, before recording a survey, a field test known as a noise spread will be
performed to establish the dominant noise problems in the area. Single geophones (or
‘bunches’ of phones planted together) are arranged at close intervals (say 5 metres). Each
geophone response will be recorded separately to produce a very detailed field record with
no noise cancellation. An example is shown in Figure 2.12. From this display predominant
noise features can be identified and a surface technique (maybe a combination of source
and receiver patterns) developed with element spacings designed to cancel as much as
possible of the noise.

___________________________________________________________________________________
Ted Cox – April 2002 Section 2: Acquisition Page 31
Introduction to Exploration Seismic
___________________________________________________________________________________

Fig 2.12 – Example of a Noise Spread showing dominant noise trends.

Figure 2.13 - Noise Attenuation using Geophone Arrays.

___________________________________________________________________________________
Ted Cox – April 2002 Section 2: Acquisition Page 32
Introduction to Exploration Seismic
___________________________________________________________________________________

Element spacing is the critical factor in array design. As can be seen from Figure
2.13(top), the response of a pattern of geophones spaced at 10 m intervals to a wave with
infinite wavelength (K=0) is proportional to the number of geophones (Σ = 6). In Fig. 2.13
(centre), for waves with smaller wavelength (λ = 60, 40, 30) some destructive interference
will occur and the response is lower (Σ = 1 or 0). However, when the wavelength has
shortened to 10 m Figure 2.13 (bottom), the response is again 6 units. In this case the
wave is referred to as being ‘aliased’, and the array is giving an apparent response of an
infinite wavelength. From the noise spread, the velocity and frequency of the noise can be
measured, the wavelength calculated and an appropriate element spacing chosen to
attenuate the noise.

With today’s multi-channel recording techniques and their ability to record vast quantities
of data, some will argue that it is better to record every geophone separately and apply
summation techniques in processing that will have a similar effect. However, analysis of
the noise by someone will still be necessary. We shall discuss some of these techniques in
the Processing Section.

2.1.11 Cables

The individual geophones are grouped together into ‘strings’ of 6 detectors (Figure 2.14),
and geophone patterns at each recording point can be made up in multiples of 6 with each
string connected to the same point in the main cable. This main cable runs the entire
length of the active spread, which may include up to 120 geophone stations. It is made up
of a series of sections so that as the recording effort passes the active stations they may be
disconnected and the geophones moved forward along the line to be planted again ready
for re-use.

___________________________________________________________________________________
Ted Cox – April 2002 Section 2: Acquisition Page 33
Introduction to Exploration Seismic
___________________________________________________________________________________

Figure 2.14 – Geophone Strings and Cable Connectors.

The electrical data generated by the movements of the earth at the geophone stations is
transferred to the recording truck by wires, optical fibres or radio waves (Plate 2C).

2.1.12 Recording Instruments

The data are recorded digitally on magnetic Field Tapes or Exabyte cartridges. Each
receiver channel is sampled in sequential order at set intervals of time (typically every 2
milliseconds) and these values written to tape in what is known as multiplex format.
Special high-quality tape recording instruments are used to ensure faithful recovery of a
vast amount of seismic information in a very short space of time. One of several ‘standard’
format configurations may be used depending on the age and manufacturer of the
equipment (e.g., SEG-A - old, SEG-D - current). These are agreed industry standard
formats as defined and published by the Society of Exploration Geophysicists (Figure
2.15).

Figure 2.15 – Example of Multiplex Field Tape Format (SEG-D).

In addition to the digital recorders, the air-conditioned recording truck contains a host of
other electronic equipment. This allows the operator, or ‘observer’, to control the
recording sequence of every shot, monitor the progress of the survey and communicate
with the Vibroseis truck drivers (or dynamite ‘blasters’, the geophone laying team (or
‘juggies’), as well as the Party Manager at base camp.

___________________________________________________________________________________
Ted Cox – April 2002 Section 2: Acquisition Page 34
Introduction to Exploration Seismic
___________________________________________________________________________________

2.2 REMOTE OPERATIONS

The method described above is applicable over large parts of the continents where trucks
are able to move, from the scorching deserts of Arabia to the frozen wastes of Alaska.

Problems arise, however, where access is difficult or impossible, and new techniques have
evolved to enable acquisition to take place in these environments.

2.2.1 Mountainous Regions

Because the terrain is very rugged and few roads exist, there is often no access for heavy
equipment in mountain regions. In these circumstances it is necessary to use a scaled
down operation comprising hand-carried or Heli-portable equipment.

Vibroseis is no longer a viable source, and small charges (1-5 Kg) dynamite must be used.
A special type of dynamite, known as seismic gelatine, available in relatively safe forms,
allows charges of suitable strength to be detonated on the surface or lowered into shallow
pre-drilled holes (~5 m deep) produced using a pneumatic process with a portable rig
(Figure 2.16).

Figure 2.16 – Different Forms of Seismic Gelatine and Detonators.

The geophone system is standard, but will probably employ fewer geophones, and
difficulty may be encountered in planting their spikes into the rocky ground. For this
reason a special type may be used having the spike replaced by a simple base plate which
can be placed on the surface.

In order to minimise the extent of connecting cables, signals received by the geophones are
sent to the recording instruments by radio transmission or ‘telemetry’. (See Figure 2.17
and Plate 2C).

___________________________________________________________________________________
Ted Cox – April 2002 Section 2: Acquisition Page 35
Introduction to Exploration Seismic
___________________________________________________________________________________

Figure 2.17 – Telemetry Transmitters and Receivers.

2.2.2 Jungles

Jungles are also inaccessible areas, but bring their own problems and solutions. Very often
it is possible to use a bulldozer and a team of local labour to clear the survey lines in
advance of conventional shooting techniques. Even so, it is usual to use a generally more
lightweight system.

Advance camps within the survey area may be set up in clearings in order to provide rapid
access to nearby lines. Such camps will be supplied by helicopter, and will be moved as
soon as those lines have been shot (Figure 2.18).

Dynamite sources may be used in shallow holes, rather than Vibroseis, and recording will
often be by telemetry.

___________________________________________________________________________________
Ted Cox – April 2002 Section 2: Acquisition Page 36
Introduction to Exploration Seismic
___________________________________________________________________________________

Figure 2.18 – Heli-portable Advance Field Camp in the jungle.

2.3 MARINE OPERATIONS

2.3.1 Planning and design

Planning for a marine survey is usually less arduous than for one on land. It will still be
vital to define the geological objectives, as these will dictate the orientation and spacing of
the lines, the length of the streamer cable and recording time.

It will also be necessary to ensure that any potential obstacles on the sea surface or seabed
are defined and avoided. Examples of potential problems are naval restricted zones,
constructions such as oil production platforms, areas containing fish traps, wrecks, shoals
and reefs. It is important to give adequate advance warning to the shipping authorities so
that fishing fleets, and other vessels which use the area, will stay clear of the long cables,
and to divers for whom the noise of the seismic source could be hazardous.

Again it is vital to ensure that the survey is designed to intersect as closely as possible any
exploration wells that have been drilled in the area. This will ensure that suitable
correlation with the geology as possible at the interpretation stage.

___________________________________________________________________________________
Ted Cox – April 2002 Section 2: Acquisition Page 37
Introduction to Exploration Seismic
___________________________________________________________________________________

2.3.2 People

On whatever type of vessel a survey is being conducted, seismic boat crew personnel, like
land men, have to be rugged individuals who are capable of working as part of a team
having to produce high-quality results, safely, often in extremely uncomfortable
conditions, under pressure of deadlines, for long periods cut off from civilisation. Except
in addition, these guys (and some gals) must not get seasick!

The composition of a typical crew will be different from a land crew in that, since the
seismic vessel is towing the cable, no teams of ‘juggies’ will be needed. Likewise, the
source is moving along with the vessel and needs no independent drivers. The basic
navigation and running of the vessel will be the responsibility of the ship’s crew, and that
will include the messing facilities in the ship’s galley. A vessel will usually accommodate
between 30 and 60 crew; still allowing space for an on-board client representative and even
the occasional visit of the client in person!

The seismic survey crew will comprise a party chief, observer and assistant, and a team of
perhaps a dozen others to ensure the safe deployment and recovery of the cable containing
the detectors and the reliable operation of the source equipment. A small on-board
processing facility will also be present, staffed by a handful of processing geophysicists
whose prime role is to monitor the navigation and acquisition to ensure specifications are
being met, and provide early warning if problems occur.

2.3.3 Vessels

The range of vessels used on a crew will vary depending on the nature of the survey.

There is usually one main multi-purpose vessel (Figure 2.19) which has been specially
designed for the seismic work. It serves as the recording and positioning unit, and also
houses all the crew and the equipment used for the sources and the receivers. As part of the
team there will also be one or two small safety and chase boats used for emergency rescue
craft and also to warn off and keep other vessels away from the survey operation.

Figure 2.19 – Typical Marine Acquisition Configuration.

___________________________________________________________________________________
Ted Cox – April 2002 Section 2: Acquisition Page 38
Introduction to Exploration Seismic
___________________________________________________________________________________

Basic, shallow draught boats will be used for near-shore surveys, whilst sturdier ocean-
going vessels will be used for deep-water work. Custom-built vessels, often equipped with
helipads, and capable of long operations with multiple streamers are used to acquire 3D
surveys. Such vessels are designed to offer a balance of towing power, fuel endurance and
carrying capacity. Their well-designed sea-keeping characteristics enable them to reduce
the impact of bad weather on operations (Figure 2.20 and Plate 2E).

Figure 2.20 – Seismic Vessel M.V. Discoverer (1600 tonnes, 72m long, 16m beam).

2.3.4 Positioning

Navigation information is measured using magnetic compasses and satellite positioning. A

fully integrated system is adopted, where DGPS, RGPS, radio-navigation, Laser trackers,
compasses and bathymetric acoustics are combined to link the vessel’s position with the
shot points of the survey. The seismic and navigation data is usually quality controlled in
the field and some processing may be carried out. The data is then transferred to the
processing contractor where it is once more QC'd. In large 3D seismic surveys processing
of some of the data may begin before all the data has been acquired.

The Global Positioning System (GPS) has been introduced during the discussion on land
surveying (Section 2.1.5). The system, put in place by the US Department of Defense,
operates using 24 Satellites in 6 orbital planes approximately 20,000 km above the Earth
and orbiting every 12 hours. Each satellite constantly transmits its position in space and
time using ultra-accurate Rubidium/Caesium clocks. Positional accuracy is achieved by
calculation from position fixes received from 4 or more satellites at any one time. To
determine a position fix, expressed in Lat./Long. or UTM co-ordinates, the receiver
compares the time that a signal is received from a satellite against the time it was
transmitted. This allows the user’s distance from the satellites to be computed and a
position fix arrived at by use of a complex set of equations.

___________________________________________________________________________________
Ted Cox – April 2002 Section 2: Acquisition Page 39
Introduction to Exploration Seismic
___________________________________________________________________________________

The GPS system replaces the traditional marine navigation systems for shipping, such as
Syledis, Pulse 8, Loran, Hyper Fix and sextant.

Press Update:
<The National Aeronautics and Space Administration (NASA) is implementing a new
Global Differential GPS (GDGPS) system, developed by Jet Propulsion Laboratory (JPL),
that provides seamless global real-time positioning at sub-20cm vertical accuracy and sub-
10cm horizontal accuracy for dual frequency GPS receivers. Such accuracy has previously
been available only on a selected local basis, through real-time kinematic (RTK)
techniques, which have enabled decimetre-level (and better) real-time GPS-based
positioning on a few kilometres local scale. Now even areas without RTK capability could
access real-time positioning at the decimetre level, without performance degradation.
Many applications require precise real-time, onboard knowledge of a platform's position
and velocity as a critical component. For regional applications, wide-area differential
systems such as the U.S. Coast Guard's Continuously Operating Reference Station (CORS)
and various commercial differential systems have provided meter-level positioning.
However, global real-time positioning at the decimetre level has never been attempted nor
achieved because of perceived technical and cost challenges.>

2.3.5 Sources

Signal Generation

As for land work, marine seismic sources are required to produce stable, high energy,
broad band impulses. However, because the vessel is moving whilst the recording takes
place, a rapidly repeatable source is essential. Large dynamite explosions are generally
unacceptable because of their adverse effect on marine life.

Consequently, airguns are the most commonly used sources (Figure 2.21(a)). Compressed
air is shot rapidly from a cylinder into the sea when a trigger is released. The frequency
content of the signal depends on the volume of the cylinder. In order to ensure that the
seismic wave that is generated is less noisy, a set of differently sized airguns is usually
fired together. The total volume of the airguns is usually 2000-3000 cubic inches. A
major problem with airguns is that bubbles are formed which produce pulses of noise after
the shot has been fired. However, the bubble noise generated by airguns is fairly regular
and predictable and most of the noise can be processed out by deconvolution at a later
stage. Also, if an array of different sized airguns is used the bubbles tend to destructively
interfere and therefore the noise is reduced (Figure 2.22).

___________________________________________________________________________________
Ted Cox – April 2002 Section 2: Acquisition Page 40
Introduction to Exploration Seismic
___________________________________________________________________________________

Figure 2.21 – Marine Seismic Sources: (a) Bolt Air Gun, (b) Sodera Water
Gun.

Water guns (Figure 2.21(b)) operate on a similar principle, by releasing a rapid discharge
of high-pressure water, but do not produce bubbles. They are used less nowadays because
the low frequency content of the signal is limited.

Occasionally Vibroseis can also be used as a marine or shallow-marine source (Figure

2.23).

___________________________________________________________________________________
Ted Cox – April 2002 Section 2: Acquisition Page 41
Introduction to Exploration Seismic
___________________________________________________________________________________

Figure 2.22 – Airgun Signatures: (a) Single Gun, (b) Multiple Gun Array.

Figure 2.23 – Marine Vibroseis source (not saucer!).

___________________________________________________________________________________
Ted Cox – April 2002 Section 2: Acquisition Page 42
Introduction to Exploration Seismic
___________________________________________________________________________________

2.3.6 Signal Recording

Again, recording reflected energy requires these three elements:

- receivers
- cables
- recording instruments

Marine seismic recording, like the process on land, uses the common depth point (CDP)
method. A series of regularly spaced shots is made as the survey vessel progresses along
pre-programmed lines. Unlike the land system, however, the vessel tows the receivers
along with it. Shots will be timed to occur at regular intervals normally equivalent to half
the separation of the receiver groups. In this way up to 1000 separate channels from as
many CDPs can be collected.

2.3.7 Receivers

Hydrophones are used in marine acquisition (Figure 2.24). These measure variations in
water pressure as changes in voltage by means of two matched piezo-electric crystals. S
waves and complex waves are not transmitted through water so this is less of a problem in
marine acquisition. Again, multi-component hydrophones can be used for special
applications.

Figure 2.24 – Cross Section of Marine Hydrophone.

2.3.8 Cables

In marine surveys, hydrophones are arranged in groups, say every 12.5m, within a cable or
streamer (Figure 2.25) several kilometres in length. The streamer has an oil-filled neutral
buoyancy and is kept at the correct depth (e.g. 7.5m) below the water surface by means of
vaned depth controllers called ‘birds’. The timing and orientation of marine surveys must
take account of currents and tide which are likely to deflect the streamer. The position of

___________________________________________________________________________________
Ted Cox – April 2002 Section 2: Acquisition Page 43
Introduction to Exploration Seismic
___________________________________________________________________________________

the end of the streamer is usually tagged by a radar reflector and GPS receiver. A large
part of the time in marine surveys is spent turning the boat and cables around in a large
semi-circle, so that the cable is straight for the next line, and waiting on weather.

The latest technology in use is a replacement for the oil-filled streamer called the solid
streamer. This streamer, made of new synthetic materials, is more robust and unaffected
by noise from towing which is especially critical on the near-traces. Its reduced noise
capability makes it more versatile in acquiring data under adverse weather conditions.

The electrical data from the hydrophone stations is transferred to the recording station
along the streamer wires, optical fibres or by telemetry (radio waves).

Figure 2.25 – Typical Marine Seismic Streamer Organisation.

2.3.9 Recording Instruments

Recording will be on similar equipment and formats to that described in the Land
Acquisition Section, except that, due to the greater rate of data recording, a number of
recorders will be necessary to allow for tape changes and malfunctions.

For example, a typical recording will last for 6 seconds – from just before the time of
impulse to a time beyond which no reflections are likely to return. In order to reconstruct
the earth’s response accurately, the signal received from each detector station has its
amplitude sampled every 2 milliseconds. Thus, each station contributes 6000/2 samples
and for every shot there may be 240 stations connected. In all therefore for each shot there
may be 3000 x 240 samples, i.e., 720,000 samples – all accurately recorded in 6 seconds!

Furthermore, with a typical station spacing of 12.5 metres, 80 separate records per
kilometre will be required, i.e., 57,600,000 samples. For a seismic line 20 kilometres long,
this means that 1.152 Billion samples will be recorded, and for a modest 2D survey of
2000 line-km there’ll be over 115 Billion samples.

That’s a lot of data !

(and for 3D surveys there will be more again by at least a further order of magnitude).

___________________________________________________________________________________
Ted Cox – April 2002 Section 2: Acquisition Page 44
Introduction to Exploration Seismic
___________________________________________________________________________________

One new piece of equipment on the market is Fairfield’s BOXÒ seismic data acquisition
system, capable of recording 8,000 channels of seismic data via radio telemetry and/or
cable, or stored data mode. Other systems use new high density video/data multiplexers as
remote miniaturised units which enable installation in very limited spaces while supporting
transmission of high quality video/ bi-directional data channels over a single optical fibre.

We shall further discuss dealing with large data volumes in Section 3, Data Processing.

2.4 SPECIAL MARINE APPLICATIONS

2.4.1 Dual Vessel Operations

Dual vessel techniques are employed in certain circumstances, such as:

• Undershooting man-made obstructions, e.g., an oil platform, which cannot
be approached too closely by conventional means.
• Making up coverage across a patch of shallow water.
• Split-spreads for very long offsets (>6 km.) to undershoot salt domes.
• Wide swathe surveying with two vessels running side-by-side.

Figure 2.26 – Undershooting an Oil Platform.

One vessel will be designated the ‘Master’ vessel, controlling line selection and shot
numbering, with the other vessel acting as ‘Slave’, transferring all positioning data to the
Master. Depending on the requirements, the Slave vessel may be responsible for the
source only and not be encumbered by cables. It is thus able to move in much closer to a
platform without risk of snagging (Figure 2.26).

2.4.2 Deep Water Acquisition

As technology permits drilling and hydrocarbon production in ever deeper water, our
exploration efforts move off the world’s continental shelves down the slope and onto the
ocean floors to perform Ocean Bottom Seismic (OBS).

___________________________________________________________________________________
Ted Cox – April 2002 Section 2: Acquisition Page 45
Introduction to Exploration Seismic
___________________________________________________________________________________

Special techniques and equipment are being developed for this purpose. One such
technique is the use of Ocean Bottom Cables (OBC) which are laid on the sea bed, in a
similar fashion to land techniques, and stay in position while a seismic vessel travels
overhead completing a series of shots (Figure 2.27). Each source array is equipped with its
own GPS and each receiver array has its own high frequency acoustic positioning system.
Information from the cables is transmitted back to the surface by telemetry. The system is
designed to be accurate to <1% of the water depth.

Figure 2.27 – Ocean Bottom Cables (OBC).

The cable is stationary and so there is no towing noise, and it is well below the zone of
wave action and almost totally weather independent. Since the cable is in contact with the
sea floor, it is possible to use both pressure sensitive hydrophones and motion sensitive
geophones (matched in sensitivity and phase) in the same cable. This dual information
provides additional flexibility in processing to remove multiple noise effects.

The same technique can be used as a high resolution option in more mature shelf areas
where the existing production infrastructure creates many obstacles to traditional methods.

___________________________________________________________________________________
Ted Cox – April 2002 Section 2: Acquisition Page 46
Introduction to Exploration Seismic
___________________________________________________________________________________

2.5 TRANSITION ZONE ACQUISITION

2.5.1 Survey Planning and Design – Special Considerations

Transition zones are the areas between land and sea, such as tidal areas, lagoons, sabkhas,
marshes and estuaries. In these zones, conventional land and marine survey methods are
not always possible.

The standard elements of survey design must still be considered, but on top of these is the
over-riding objective of obtaining data in an operationally complex area which are of
sufficient quality to allow the continuous interpretation of the land seismic with the marine.
A programme must be designed to link existing land and marine survey lines, and methods
must be devised to obtain the necessary fold of coverage along them.

2.5.2 Techniques – Examples

The methods used in T-Z acquisition usually involve a combination of land and marine
equipment, sometimes used in different ways (Figure 2.28).

Figure 2.28 – Transition Zone: Geophones in the Surf.

For example, a series of geophone patterns may be laid out along the beach and in shallow
water and recorded into from an air-gun source located in the shallow water. This could be
combined with a stationary marine cable, moored in the shallow water, being recorded into
by Vibroseis trucks on the beach. (Plate 2F)

The use of a ‘bay cable’ is sometimes adopted in shallow water areas. This cable is similar
to a marine OBS cable in that it contains hydrophones, and is designed to rest on the sea
bed. This technique has the benefit of keeping the detectors as far below wave base as
possible, hence reducing the noise of wave activity.

In areas unsuitable for vehicles and cable-laying, e.g., muddy estuaries, remote detectors
can be laid out on the surface by using inflatable craft at high water and their data
transferred to the recorders by telemetry (Plate 2G).

___________________________________________________________________________________
Ted Cox – April 2002 Section 2: Acquisition Page 47
Introduction to Exploration Seismic
___________________________________________________________________________________

2.6 3D-ACQUISITION

3D seismic surveys are generally acquired as a series of ‘swaths’ or ‘patches’ of more

closely spaced points to allow a greater resolution of the subsurface. Because of this
requirement to sample data more closely, increased resources and field effort are required,
especially on land. These have inevitable cost implications, and 3D has traditionally
become the accepted method of survey for reservoir and field definition.

It is becoming more common nowadays in established hydrocarbon basins to acquire 3D

seismic at the exploration stage. This is seen as a long-term investment and, provided that
enough time is available to properly interpret the data, then structural uncertainties should
be greatly reduced. A greater understanding of ancient ‘palaeogeographical’ environments
can also result from this more detailed way of examining seismic response.

In the past it could take up to 24 months to process the recordings from a 3D survey.
Acquisition time has been cut by specially designed survey vessels deploying up to ten
multiple streamers at a time (Plate 2H), or by using multiple vessels. These techniques
allow a swath of seismic data to be acquired in the same time it would take to record a
single 2-dimensional line. Specially designed ‘paravanes’ steer the cables away from each
other. Their design reduces the drag of the streamer array, which ordinarily would be
sufficient to stop even quite a powerful vessel. Modern streamers have multiple global
positioning system (GPS) sensors that constantly record the position of the streamers
relative to the vessel and the earth.

New techniques of data compression are being tried to allow the transmission of the raw
seismic records from the acquisition vessel to the shore for immediate processing, in an
effort to get the data to the interpreters faster.

2.6.1 Survey planning and design

Planning a 3D survey requires, but does not always get, extreme care and consideration.
Without these, all of the field effort and, consequently, much money could be wasted.
Design is therefore much more complex than for a 2D survey. It will be even more vital to
define the geological objectives and plan the field parameters.

The aims of a 3D survey will usually be to build a geological model of the prospect, assist
the exploration team in drilling and provide political/commercial advantage. Sometimes a
survey may be recorded because it was promised in licence application. Finally, it must be
completed as cheaply as reasonably possible.

2.6.2 Special Considerations

Planning should take into account the following list of considerations:

Line Orientation with respect to geological response, physical obstructions, water

currents (marine) and access (land). Azimuthal Variation (travel paths oblique to the line).

___________________________________________________________________________________
Ted Cox – April 2002 Section 2: Acquisition Page 48
Introduction to Exploration Seismic
___________________________________________________________________________________

Cost considerations – Logistics of survey equipment and layout (land), boat availability
and number of cables (marine). The processing requirements.(See Section 3.4.8).

Survey layout and acquisition method

• Marine - Number of Sources / Cables / OBS Cable Option
• Land - Applicability of layout to environment / Crew experience

Tendering - Getting good cost estimates for many layouts before choosing a contractor.

Survey Design Software

+
A land 3-D survey cannot be planned without the use of special planning software,
especially to evaluate the effects of dipping reflectors on fold of coverage.

Areas with Particular Problems

• Obstructions and Delimited Areas
• Marine: Water velocity and depth variations, noise from nearby operations.
• Land: Statics requirements, security, safety, access and environment.

Taking Specialist Advice

• A 3-D survey costs $Millions and good planning can ensure successful
result.
• Poor planning will result in poor results.
• Planning costs little BUT takes longer than any other part of the seismic
cycle.

2.6.3 Resolution

It will be important to consider the required seismic resolution:

• Vertical resolution will be governed by bandwidth, hence the source and receiver
response. Marine sources are largely determined by contractors and should be
predictable but array design and vessel movement may affect the resolution. Whilst on
land there is greater scope to vary the survey parameters but less control over their
earth coupling. (See Section 3.2.10)

• Lateral (horizontal) resolution – The ‘Fresnel Zone’, the area from which reflections
will contribute to one CDP, will be governed by the effective CDP spacing, or ‘bin’
size. (A bin is the area within which all common depth points, irrespective of their
travel azimuth, may be considered to represent the same CDP). It is advisable to test
previous data in this respect for spatial aliasing. (See Section 3.2.11)

2.6.4 Survey Size

The overall survey size will not just depend on the area required for the geological model,
but it will be necessary to consider also:
• the length of lead-ins to each line (until full fold of coverage is obtained)
• the migration aperture (the extra traces needed for use when processing dips)
• the relevance of the effect of the Fresnel Zone (lateral resolution)

___________________________________________________________________________________
Ted Cox – April 2002 Section 2: Acquisition Page 49
Introduction to Exploration Seismic
___________________________________________________________________________________

2.6.5 Review of existing data

Above all, it is important to have realistic expectations of what the survey will be able to
achieve. For this reason it is necessary to review any available old data from the area:

• Previous Seismic for comparison to find:

- The smallest CDP/Bin size and longest offsets previously used
- The parameters giving the best quality data
• Well Data – yields velocity information to predict the reflection coefficient
series
• 2-D Processing Tests to assist planning

2.6.6 The ‘Good’ Seismic Cycle

Planning 6 months, cost $ 50,000 (less than 10% of total cost)

Acquisition 2 months, cost $ 500,000
Processing 2 months, cost $ 50,000
Interpretation 3 months, cost $ 50,000

The Survey Plan should take a long time if it is done properly. If not, don’t shoot it!
It is important not to be pressurised by contractors’ availability.

2.6.7 Land methods

Some aspects of land 3D have already been discussed. Essentially the operation is the
same as for 2D, except that the geophones will be laid out in one direction (maybe in a
series of parallel lines, and the shots (or vibration points) will be at right angles, to record a
‘patch’ of data (Figure 2.29(a)). The whole system will then be moved along to overlap
slightly to record the next patch, making up coverage and moving over the area. (Plate 2J)

(a) (b)

Figure 2.29 – 3D Coverage: (a) Patches and (b) Swaths

___________________________________________________________________________________
Ted Cox – April 2002 Section 2: Acquisition Page 50
Introduction to Exploration Seismic
___________________________________________________________________________________

2.6.8 Marine methods

3D surveys will usually collect a swath of seismic data a few hundred metres wide parallel
to the direction of travel (Figure 2.29(b)). This will be achieved either by using multi-
vessel operations, as described for undershooting, but recording into multiple streamers
(Plate 2I), or the new purpose designed Ramform-type vessels which are able to deploy up
to 20 streamers at one time (Plate 2H).

2.6.9 Binning

Exactly where the reflection points are sampled in the subsurface depends on how the
streamers are affected by currents and tides. With accurate positioning (typically <5m)
seismic traces can be assigned to common mid-point ‘bins’ (typically 25m x 25m) onboard
the survey ship (Figure 2.30). The CDPs which make up each bin will have travel paths at
various ‘azimuths’ or horizontal angles to the direction of the line, depending on the
positions of the source and receiver. Infill lines may have to be shot where the bins contain
less than the minimum specified coverage.

Figure 2.30 – 3D Coverage: CDP collection into Bins.

___________________________________________________________________________________
Ted Cox – April 2002 Section 2: Acquisition Page 51
Introduction to Exploration Seismic
___________________________________________________________________________________

2.7 HIGH RESOLUTION SURVEYS

In order to complete an appraisal of the wide variety of types of seismic survey, it should
be observed that the method can also be used at a later stage of exploration, that of
exploration drilling. It is good practice to perform a high-resolution seismic survey in
association with side-scan sonar of a proposed drilling location, prior to moving the rig on
to location. This ensures that there will be no unpleasant or dangerous surprises, such as
irregular sea floor features or obstacles to obstruct the locating of the rig, and no gas
trapped in the shallow rock layers, which could cause a blow-out in the early drilling
phases.

Such surveys are usually recorded over very small areas, but at a close line-spacing, using
a high frequency electrical source (‘Sparker’) and a 1 millisecond sample rate down to 1 or
2 seconds two-way time.

Associated multi-beam bathymetry and imagery surveys can be performed on the little
known continental slope areas, ranging in depth from 600 to 3500 metres. These can
contribute directly to issues relating to hydrocarbon prospectivity, surficial geohazards, and
environmental conditions at the seabed. The multi-beam bathymetry data, with a pixel size
of approximately 30 metres, reveals the complex seabed morphology of the continental
slope, such as submarine canyons and evidence of recent surficial sediment failure. Many
of these failures in deeper water have linear head-scarps, developed along faults which can
be correlated with high-resolution seismic profiles. Such faults are targets for investigation
of surface hydrocarbon seeps using remotely operated vehicle (ROV) technology.

(Seabed classification is a growing need in ocean research, extending into many new
application areas, including high-resolution surveys in search of diamonds.)

___________________________________________________________________________________
Ted Cox – April 2002 Section 2: Acquisition Page 52
Introduction to Exploration Seismic
___________________________________________________________________________________

2.8 4D SURVEYS

A 4D survey is an extension of the concept of a 3D survey, where the fourth dimension is

one of elapsed time. This elapsed time is on a scale of months or years. Sometimes this
technique is referred to as Time-Lapse 3D (TL3D).

The principle behind the technique is that, if a high-resolution survey samples the seismic
response from a productive area periodically, changes in the fluid content within the field
that bring about changes in the seismic response will be seen by comparing successive
surveys. (Figure 2. 31).

Figure 2.31 – 4D Time-Lapse Survey over Producing Reservoir (1994 – 1996).

Recently BP, in the development of the Foinaven field, has installed an array of
hydrophones on cables buried just beneath the seafloor across the oil field. Their aim is to
carry out a 4D survey by shooting over the array with a seismic vessel once a year.
Changes in the reservoir reflection character will be studied to indicate the areas of the
field where oil is not being produced efficiently.

___________________________________________________________________________________
Ted Cox – April 2002 Section 2: Acquisition Page 53
Introduction to Exploration Seismic
___________________________________________________________________________________

2.9 HEALTH & SAFETY

2.9.1 Procedures

In the seismic exploration industry as much as any today, attention to Health and Safety is
VITAL. The early days of exploration are full of stories of ‘near misses’ and old surviving
‘doodlebuggers’ will all have a story to tell about a colleague or some appendages lost
through operational accident. It is partly through these bitter experiences and partly from
modern awareness that a system of safety procedures has evolved.

2.9.2 Compliance with Standards

Most seismic companies operate a policy of mandatory Health and Safety procedures
which aim to prevent injury or illness to their employees or harm to the community and the
environment. Failure to comply with these policies may well result in dismissal.

The International Association of Geophysical Contractors (IAGC) has made safety

guidance available via the IAGC Safety Manual for Geophysical Field Operations (Figure
2.32). It is updated regularly to catch up with changes in operational techniques.

Figure 2.32 – IAGC Safety Manual.

Standards not only refer to operational safety, but to maintenance of vessels, vehicles,
electrical wiring and field camps, including personal and food hygiene and waste &
sewage disposal.

2.9.3 Safety Management

It is important to have a company structure whereby managers take full responsibility for
safety in their respective areas of operation. This can be achieved by setting up appropriate
quality assurance procedures regarding safe practices.

___________________________________________________________________________________
Ted Cox – April 2002 Section 2: Acquisition Page 54
Introduction to Exploration Seismic
___________________________________________________________________________________

Most senior personnel should be qualified Outward Bound First Aiders. Land crews should
be equipped with fire-fighting apparatus and an ambulance. Those operating in certain
remote conditions should even have full time qualified doctors on site.

All crew members on board a seismic vessel should be in possession of offshore

certificates demonstrating their proficiency in safety procedures, including fire fighting,
use of life rafts and escape from an inverted submerged helicopter.

2.9.4 Reporting

The only way that lessons can be learnt if accidents occur, is to have a rigorous accident
reporting procedure which ensures that any and all accidents are reported to senior
management and are investigated. Regular safety meetings involving all crew members
can encourage this and also ensure that previous incidents are passed on.

2.9.5 Training

All new crew members should undergo a safety induction session where the safety
procedures and all work hazards are reviewed. On boarding a seismic vessel, a new crew
member should be shown the emergency evacuation procedures, especially where to find
personal survival gear and how to use it.

Party chiefs, party managers and their assistants should attend IAGC or equivalent Safety
Courses. These will include Medical and First Aid, Fire Fighting, Defensive and All-
Terrain Driving Instruction, Survival at Sea and Helicopter Escape Drill (Figure 2.33).

Figure 2.33 – Aspects of Safety Training.

___________________________________________________________________________________
Ted Cox – April 2002 Section 2: Acquisition Page 55
Introduction to Exploration Seismic
___________________________________________________________________________________

One classic saying is that you should never enter an unknown situation, unprepared, with
insufficient information and take chances that pose unnecessary risks. We must
continuously aim to minimise risks by doing research, gathering appropriate information,
watching out for others and putting rules and guidelines into place to reduce risks to a
minimum.
Remember – Safety is NO ACCIDENT!

2.10 QUALITY CONTROL

Following closely on our concerns for Health and Safety must be our constant monitoring
of the seismic survey. If we fail to make the most of our time in the field, money will have
been wasted and the final product handed to the interpreters will be inferior.

2.10.1 Environmental considerations

Historically, the oil industry has earned itself a bad reputation. Today, that reputation is
largely unjustified, but it is still hard to live down in the public eye. We must do our best
not to make matters worse.

As has been discussed in the planning sections for land surveys (Section 2.1.2) and marine
surveys (Section 2.3.1), the environmental effects of conducting a seismic survey must be
considered from the start.

2.10.2 Pre-survey testing

On land, the procedure for running Noise Tests has been explained in Section 2.1.10.
These should be conducted at a few locations chosen as representative parts of the survey
area. It is vital that these tests are run before the crew sets off to record the survey. Field
parameters can then be confirmed and operations proceed.

For marine surveys, source and cable configurations should have been considered before
the vessel arrives at the survey site. Often the vessel will have been operating in nearby
waters, and experience of the optimum parameters has already been gained. However, it is
necessary to ensure that the system is operating correctly, as it is pointless to shoot the
survey if all is not well.

2.10.3 Operating standards

For any project, the minimum operating specifications will have been stipulated in the
contract. As a survey progresses, Instrument Tests must be conducted at regular intervals
to ensure compliance.

On a land survey, it is customary to check the recording instruments, the receivers, the
vibrators and other peripheral equipment every morning before shooting commences.
Other more comprehensive equipment tests should be conducted on a weekly basis.

At sea, recording operations are usually continuous and so the emphasis must be on
continuous monitoring of the survey data and ensuring operational standards are

___________________________________________________________________________________
Ted Cox – April 2002 Section 2: Acquisition Page 56
Introduction to Exploration Seismic
___________________________________________________________________________________

maintained during the survey. The opportunity to perform tests and maintenance will be
during line changes and down time due to bad weather.

2.10.4 Quality Assurance (QA)

Most contractors operate to accepted industry standards (e.g., ISO 2000) in that they have
designed QA procedures for each operational step. These will prescribe the actions
necessary to perform each step in the survey process. All crew members involved in these
activities will be required to follow the steps and complete a form indicating they have
done so. These procedures will generally define the specifics of good industry practice and
technical excellence (Figure 2.34).

Figure 2.34 – QA Procedures for Vibroseis Testing from CGG Manual.

2.10.5 Monitoring operations

It should be clear by now how important it is to be aware of the whole operation and alert
to problems.

On board a seismic vessel, monitoring of the survey takes place in real time. Any
deficiencies in the operation, such as malfunctioning air guns or bad transmission from the
cable, or problems with noise due to other vessels or the sea state, must be identified
immediately and a decision taken whether or not to abort the line. If shooting is suspended,
it will be necessary for the vessel to sail around in a big circle and resume the course of the

___________________________________________________________________________________
Ted Cox – April 2002 Section 2: Acquisition Page 57
Introduction to Exploration Seismic
___________________________________________________________________________________

line several kilometres away so that the streamer(s) will be straight at the point where
shooting should recommence.

On land, where progress is much slower, the observer can monitor the status of the survey
equipment from hid recording cabin (Plate 2K). He must check that all is well with cable
connections, radio links, source locations and operations, and the background noise is as
low as possible, before starting each recording sequence.

It has become common practice for the client to hire Independent QC Representatives,
known as Client Reps., or sometimes ‘Bird-Dogs’. It is their job to monitor the QC
procedures conducted by the field crew and ensure that contract specifications are being
met. They will report directly to the client on a daily basis, allowing a measurement of
performance to schedule and budget.

Whether it is they or the field operators who spot a problem, it is important to take swift
corrective action otherwise data will have to be repeated and money and time wasted.

2.10.6 Field Processing

At sea, the newly recorded data are collected in the on-board operations room(s) (shown in
Plate 2L) where it is possible to perform real time processing. The basic steps are taken to
produce a provisional (‘brute’) stacked seismic section, as well as displays of the near and
far traces of every shot. These are all good checks on data quality. For 3D surveys, a plot
of fold of coverage (CDPs per bin) will also be made in the form of a map.

On land, records from the field are processed on compact in-field processing units, such as
MicroMAX to check that the desired results are being achieved. Brute stacks can only be
produced when sufficient data has been shot, possibly every 2-3 days. This work is often
done at base camp in the evenings or overnight, so that the results can be reviewed and
adjustments made for the next day’s efforts.

2.10.7 Reporting

Reporting procedures are a vital part of the whole operation. QA programmes are designed
to ensure that these reports take place.

For the Observer on land, reports (or logs) must be completed for every record, as well as
for the daily tests. At sea the job is more automated, and notes should be made at the start
of each shift and every new line as well as where problems occur. Weekly and monthly
progress reports will also be required.

For the Client Rep., daily and weekly progress reports must be made to the client as well as
special communications regarding problems and performance. The client’s views can then
be communicated back to the crew.

Everyone must write their reports at the end of the contract!

___________________________________________________________________________________
Ted Cox – April 2002 Section 2: Acquisition Page 58
Introduction to Exploration Seismic
___________________________________________________________________________________

2.11 LOGISTICS

The network of support required to keep a seismic crew operating must not be
underestimated. Most of the contractors operating today are large international companies
with local offices in many countries.

2.11.1 Supply of crews and equipment

These companies usually have a number of operational field crews and vessels, as well as
storage depots and dock facilities at strategic points around the world. From these it will be
possible to supply the crews and equipment needed (Figure 2.35).

2.11.2 Contract Mobilisation / Demob.

When a new contract is awarded a crew or vessel must be mobilised. This will involve
gathering together and shipping all the equipment and supplies from the storage depots,
and the manpower needed from standby or leave. In some instances a vessel or a field
crew will have been operating in a nearby area and can be moved on to the new location
with a minimum of effort. However, it will be important to ensure that all equipment is
fully functional and to supply a replacement work force.

Figure 2.35 – Logistical Support is Vital.

___________________________________________________________________________________
Ted Cox – April 2002 Section 2: Acquisition Page 59
Introduction to Exploration Seismic
___________________________________________________________________________________

2.11.3 Vital back-up, spares & repairs

Most crews will carry a supply of commonly needed spares. The regional storage depots
will also carry a wide range of additional spare parts that can be flown to the crew rapidly
if needed. They will also offer repair facilities for equipment that cannot be fixed on the
spot.

2.11.4 Electronic and technical support

With the large amount of electronic equipment and computer software used on a survey
today, most of these regional offices will also employ a number of electronic and technical
support personnel on continuous standby ready to fly to a crew experiencing technical
difficulties. Head office staff will also be available to travel if necessary.

2.11.5 Design, fabrication & testing

Technology advances rapidly, and each company competes to use the latest modifications
to assist in the operation. Most large companies will support their own research and design
facility where new ideas and improvements can be made and tested.

2.11.6 Training

The importance of safety and QC training has already been discussed. Such facilities are
normally available at head office. To meet the advances in technology, most companies
offer a programme of training for new recruits and also as a refresher for hardened field
crew – a task that must not be under-rated!

___________________________________________________________________________________
Ted Cox – April 2002 Section 2: Acquisition Page 60
Introduction to Exploration Seismic
___________________________________________________________________________________

SECTION 3:

SEISMIC DATA PROCESSING

The steps necessary to convert field data into a meaningful format.

This section of the course deals with the steps necessary to convert field data into a
meaningful format for the interpreter. Data Processing will involve many stages of signal
processing and computer summing – the major number crunching jobs that computers love.
A new vocabulary will be introduced – the ‘buzz words’ of the DP Community.

3.0 PROCESSING OBJECTIVES

The objectives for processing seismic data are to recover and transform the field data,
reduce noise and improve resolution so that the confidence in and precision of
interpretation are optimised.

The processing steps may be divided into four main functions:

• Data Preparation
• Data Correction
• Data Reduction
• Data Enhancement

3.0.1 The Basic Processing Sequence

A standard sequence for processing seismic data is shown on the next page.

___________________________________________________________________________________ 61
Ted Cox – April 2002 Section 3: Processing Page
Introduction to Exploration Seismic
___________________________________________________________________________________

Basic Processing Sequence:

START
INPUT FIELD DATA
+
TOPOGRAPHIC/NAVIGATION DATA
↓
• 3.1 Data Preparation
DEMULTIPLEXING
EDITING
↓
• 3.2 Data Correction
AMPLITUDE CORRECTIONS
↓
STATIC CORRECTIONS
↓
NOISE REDUCTION (F-K)
↓
SORTING/BINNING
CDP GATHERS
↓
VELOCITY ANALYSES
NMO CORRECTIONS
↓
• 3.3 Data Reduction
CDP STACKING
MULTIPLE REMOVAL
↓
• 3.4 Data Enhancement
DECONVOLUTION
↓
DMO & MIGRATION
↓
FREQUENCY FILTERING
↓
DISPLAY PARAMETERS

Each of these steps will be discussed in the numbered sections that follow.

The information regarding processing any seismic section can be found on the side label.

An example for a marine seismic line is shown in Figure 3.1.

___________________________________________________________________________________ 62
Ted Cox – April 2002 Section 3: Processing Page
Introduction to Exploration Seismic
___________________________________________________________________________________

Figure 3.1 – Actual Processing Sequence: Marine Data Example.

___________________________________________________________________________________ 63
Ted Cox – April 2002 Section 3: Processing Page
Introduction to Exploration Seismic
___________________________________________________________________________________

3.1 DATA PREPARATION

3.1.1 Initial Processing

Data recorded on field tapes are often initially processed in the field for quality control
purposes, to check that the field operation is recording data satisfactorily. Depending on
the rate of data gathering, further initial processing steps may be performed in the field to
save time at the processing centre (see Section 2.10).

At some point, however, the field tapes will be sent to a processing centre where they will
be processed through to final display prior to delivery to the interpreters. Many contractors
now offer a network of small processing offices in areas of operation. These offices are
capable of conducting basic processing steps and testing, and a dedicated high-speed data
transfer link via satellite to their mainframe computing centre where bulk processing is
done (Figure 3.2). They also offer the important facility of providing client and field
contractor liaison.

Figure 3.2 – CGG’s Remote Data Transfer System.

3.1.2 Data Media

Data recorded in the field will normally be in digital magnetic tape format or on Exabyte-
type cartridges (which are more compact). Some contractors offer the facility to transmit
data by telemetry direct to their processing centres, although, due to severe operational
conditions, this is not always reliable.

Supporting data that accompanies the tapes, or follows shortly after, will include
Observers’ Reports, Stacking Diagrams, Field Test Data, Navigation / Topographic Data.

___________________________________________________________________________________ 64
Ted Cox – April 2002 Section 3: Processing Page
Introduction to Exploration Seismic
___________________________________________________________________________________

The field tapes will be loaded onto a seismic computer such as a VAX, SUN or CRAY.
The multiplexed field tapes are read and reformatted by the central processing unit (CPU)
and stored as a database in the main memory in the form of new digital data. The database
is updated throughout the processing sequence.

3.1.3 Demultiplexing

The data are sorted so that a time series, or seismic trace, is obtained for each channel for
every shot. This sorting process is known as demultiplexing. The samples are then
normally arranged in a standard format known as SEG-X (Figures 3.3 and 2.15).

Each seismic trace contains the sampled seismic data for one source-receiver pair. The
demultiplex process groups all channels recorded for one shot together as adjacent traces
known as a shot record. The output from demultiplex will be one shot record for each
shot position along the seismic line.

Figure 3.3 – Demultiplexing

3.1.4 Editing

Editing involves correcting for problems in the data that are due to errors introduced during
recording, storage and retrieval.

This process is essentially manual, and relies on a quality assessment of the shot records.
For land data, every shot record will be displayed and examined for noisy geophone
stations, bad electrical connections, static noise bursts (such as lightning discharges or
short-circuits caused by damp), electrical interference from AC mains frequencies and
animal, industrial and wind noise (Figure 2.11). Any offensive traces will be omitted from
further processing.

___________________________________________________________________________________ 65
Ted Cox – April 2002 Section 3: Processing Page
Introduction to Exploration Seismic
___________________________________________________________________________________

Occasionally a trace will be detected whose geophone station has been connected with the
wrong polarity in the field. Such a trace can be used for processing, provided a polarity
reversal is applied to correct the fault.

Some noise bursts occur over only small parts of seismic traces, the remaining samples
being unaffected. Such noise bursts can be eliminated from further processing by zeroing
the offensive sample values whilst allowing the remainder of the trace to be used, although
frequently these noisy traces are also rejected (Figure 3.4).

Figure 3.4 – Editing Noisy Traces.

The degree to which this editing is performed depends largely on the field design and the
level of multiplicity. In areas of high multiplicity (or coverage), salvaging suspect traces is
less important, as there is a high level of redundancy (i.e., many traces for each sub-surface
point).

For marine data, this is often the case, and such labour-intensive editing is not normally
performed. Also, since the cable is being towed behind the recording vessel, there are
fewer parameters to check. Here, if a hydrophone station becomes unserviceable during
shooting, the same trace will be invalid on every record, and can thus be zeroed. Polarity
reversals can be dealt with similarly, as the same correction will apply to each record.

Noise problems that require special care are those caused by loud bursts generated when
the cable-steering ‘birds’ break the surface or generate mechanical sounds. Automatic
zeroing programs can be run to eliminate these zones.

___________________________________________________________________________________ 66
Ted Cox – April 2002 Section 3: Processing Page
Introduction to Exploration Seismic
___________________________________________________________________________________

3.1.5 F-K Filtering

Another editing process is the F-K Filter. The aim of F-K filtering is to remove from the
record any organised (coherent) noise, such as ground roll that has survived attempts to
cancel it in the field. It is introduced here, although it is usually not applied until after
statics in the correction phase of processing (see section 3.2.4).

The seismic signal produced from a shot is sampled both in time (e.g., every 4 ms.) and
distance (e.g., detector stations every 12.5m). In our familiar time-distance (t-x ) domain,
signal and noise interfere with each other because they are happening at the same places at
the same time. However, coherent noise and deep reflections can be distinguished more
easily from each other by considering the period (τ), or the frequency (F) of the sampled
time versus the wavelength (λ) or the wave number (K) of the sampled distance.

The frequency and wave number content of the shot record can be analysed by a
mathematical process known as a two dimensional digital Fourier Transform (see section
3.1.6). This process can transform a field record in the t-x domain into an F-K record.

Because the upward travelling ray paths from deep reflections tend to be near vertical, they
will arrive at all points on the surface at approximately the same time. They will thus have
relatively high F values and relatively low K values. Horizontally travelling ground roll
will conversely have low F and high K values. Therefore, coherent noise and signal can be
easily separated. (Figure 3.5)

Figure 3.5 – The F-K Filtering Process.

Top Left: 1. Field record Top Right: 3. After Filtering

Bottom Left: 2. Raw F-K record Bottom Right: 4. Filtered record

___________________________________________________________________________________ 67
Ted Cox – April 2002 Section 3: Processing Page
Introduction to Exploration Seismic
___________________________________________________________________________________

A filter can be applied in the F-K domain that suppresses all amplitudes with non-signal
characteristics (sometimes called a ‘pie-slice filter’). The filtering is essentially done by
multiplying the wanted signal by 1 and the unwanted noise by 0, although generally the
transition between the two is gradational (tapered) to avoid side effects.

After filtering an inverse Fourier Transform is performed to produce a field record with
significantly reduced coherent noise levels.

The F-K domain transform of the Noise Spread shown in Figure 2.12 is shown here for
comparison (Figure 3.6).

Figure 3.6 – Field Noise Record Transformed into F-K Domain.

___________________________________________________________________________________ 68
Ted Cox – April 2002 Section 3: Processing Page
Introduction to Exploration Seismic
___________________________________________________________________________________

One problem can arise when using this process, the phenomenon of aliasing. This is an
interference effect which is a function of the trace spacing and can result in artificial dips
being transformed. If the trace spacing is too large, the aliased noise will interfere with the
signal making it impossible to separate the two by this method. Care should be taken in
survey design to ensure this effect is minimised.

Another example of the F-K filter in practice is shown step by step in Figures 3.7 and 3.8.

Figure 3.7 – Example Field Record and its F-K Transform.

Note that the ground

roll and coherent
noise have been
effectively removed
by the filter.

Reflection
h b l b

Figure 3.8 – Example Field Record after F-K Filtering.

___________________________________________________________________________________ 69
Ted Cox – April 2002 Section 3: Processing Page
Introduction to Exploration Seismic
___________________________________________________________________________________

3.1.6 Digital Fourier Transform

A ‘simple’ mathematical expression for a digital Fourier Transform (FT) can be written
as follows:

n= N / 2
S = ∑ n =o gn.Cos(2π.t.n/T + Φn )

An alternative way of expressing this relationship is:

n= N / 2 n= N / 2
S = ∑ n =o an.Cos(2π.t.n/T) + ∑ n =o bn.Sin(2π.t.n/T)

Where S = seismic data point (sampled amplitude)

N = number of data points within time gate
n = series of integers increasing by 1
gn = amplitude of periodic oscillation
t = time when sample X was taken
T = fundamental period (width of time gate)
n/T = frequency (F) of oscillation
Φn = phase angle (-π to + π) of oscillation

gn.Cos (2 π.t.n/T + Φn) can be thought of as a rotating vector. All the values of gn and Φn
can be solved for each frequency, provided that there are enough data points. Hence, only
a maximum of N/2 frequencies can be analysed.

If the sampling is done in distance (geophone stations), then T is replaced by λ and n/λ is
the wave number (K).

Therefore, if two Fourier Transforms are carried out on a shot point record, one in the t
direction and the other in the x direction, then each value of F or K will have a calculated
amplitude value (gn). The values of gn can be plotted as F or K amplitude spectra. The
amplitude of any particular periodic oscillation will be the same in either the t or x
direction. Therefore, it is possible to determine the wave number that corresponds to any
particular frequency of periodic oscillation on the basis of its amplitude. This relationship
can be shown on an F-K plot.

The CPU of a seismic processing computer does all these calculations in a matrix, known
as a Fast Fourier Transform (FFT), but it is useful to understand the fundamentals.

It should be remembered that frequency analysis is only a different mathematical

method of describing seismic data. The FFT will be used in a number of data
enhancement operations to be discussed later.

___________________________________________________________________________________ 70
Ted Cox – April 2002 Section 3: Processing Page
Introduction to Exploration Seismic
___________________________________________________________________________________

3.2 DATA CORRECTION

As can be seen from Figure 2.11 (repeated here as Figure 3.9), a raw field record will
show a marked decease in amplitude with time. The wavelet suffers degradation due to:

• Spherical divergence
• Absorption and scattering
• Interface transmission losses
• Interface refection coefficients
• Multiple reflection effects
• Refractions and surface waves

We must apply some form of compensation for these effects if we are to see the seismic
reflections at all.

Figure 3.9 – Degradation of the Seismic Wavelet.

3.2.1 Amplitude corrections

As the source-generated wavefront expands, its energy is reduced due to certain predictable
factors outlined here. This reduction in energy means that reflections generated deeper in
the section will be weaker than those generated where the wavefront is strong. To
compensate for these effects, several forms of mathematical amplitude scaling must be
applied.

___________________________________________________________________________________ 71
Ted Cox – April 2002 Section 3: Processing Page
Introduction to Exploration Seismic
___________________________________________________________________________________

i) Spherical Divergence

The energy produced by a seismic source is spread over a wavefront with an area of
2 π d 2 , where d is the distance travelled by the wavefront (Figure 3.10).
Therefore, the seismic energy at any point along that wavefront is proportional to
1/d 2 . The amplitude of a seismic signal is proportional to the square root of its
energy i.e. 1/d or approximately 1/t, where t is the two-way time (TWT) of the
seismic data on the trace.

Figure 3.10 – Spherical Divergence.

ii) Absorption and Scattering

Much of the energy produced by a seismic source causes elastic deformations

within the upper part of the earth’s crust. This energy is conserved. However,
some of the energy is lost due to non-elastic deformations (such as friction where
the energy is converted to heat) and internal reflections which never return to the
surface. These cause seismic amplitude to decrease exponentially with time
b.t
(Figure 3.11). To compensate for this effect the data are multiplied by a factor e ,
where b is the absorption coefficient and t is the TWT.

Figure 3.11 – Amplitude decrease by Absorption.

___________________________________________________________________________________ 72
Ted Cox – April 2002 Section 3: Processing Page
Introduction to Exploration Seismic
___________________________________________________________________________________

High frequencies are absorbed more quickly than low frequencies (Figure 3.12).
Therefore, a composite seismic signal will tend to have a narrower, lower
frequency bandwidth on the deeper part of the trace (at large TWT's) than a similar
seismic signal in the shallow part (at small TWT's).

Figure 3.12 – Absorption of High Frequencies.

iii) Transmission Losses

The transmitted amplitude for a seismic signal is equal to (1- ∑ R 2 ) N , where there
are N reflectors and R is the reflection coefficient (Figure 3.13). In practice the
amplitude decreases in an approximately exponential manner. This can be
a.t
corrected for by multiplying by a factor e , similar to compensating for absorption
losses.

Figure 3.13 – Transmission Losses at each interface.

___________________________________________________________________________________ 73
Ted Cox – April 2002 Section 3: Processing Page
Introduction to Exploration Seismic
___________________________________________________________________________________

3.2.2 Overall Gain Correction (Gain Recovery)

-(a+b).t
The amplitude of the seismic trace will be expected to decay with a slope e /t due to
the various effects outlined above. The actual slope can be found by best fitting a curve
through the observed peaks (Figure 3.14). The overall gain correction required is
therefore applied by multiplying each amplitude (sample value) of the trace by the inverse
of this slope in order to pull all amplitudes up to a normalised value.

Figure 3.14 – Total Gain Correction.

In practice, a series of standard correction curves are applied to specimen traces, and the
results compared to establish the most appropriate for a particular survey.

___________________________________________________________________________________ 74
Ted Cox – April 2002 Section 3: Processing Page
Introduction to Exploration Seismic
___________________________________________________________________________________

3.2.3 Static and Dynamic Corrections

Before we can discuss the further corrections that must be applied to our seismic data, we
must recall the CMP.

The principle of common mid-point coverage has been introduced in sections 1.2.4 and
1.2.5. If all the seismic traces contributing information for one CMP (or CDP) are
collected together and arranged in order of increasing offset distance between source and
detector, they are said to form a CMP Gather (Figure 3.15).

Figure 3.15 – Common Mid Point (CMP) Gather.

Reflections from any one reflector form a hyperbolic curve on the gather because the
sound takes longer to travel to the more distant detectors. In theory, a CMP gather will
consist of perfectly hyperbolic reflections which are symmetric around the zero offset
seismic trace. The hyperbola has the form:

Tx2 = To2 + X2/V2

(Tx - To) is known as the Normal Move-Out (NMO) and V is the stacking velocity.

Correction is needed to bring the pulses to a horizontal alignment, as if they all came from
vertically below the sound source. This type of correction is known as Dynamic, or NMO,
Correction and will be discussed in more detail in Section 3.2.12. However, in the real
case the reflections rarely form on a perfect hyperbola until corrections are made for near
surface delays. These corrections are known as short wavelength statics.

3.2.4 Static Corrections – Field Statics

A static correction is a constant time-shift applied to a seismic trace to allow it to be

compared and combined with others within the survey.

Based on information gathered during acquisition, some simple static corrections are
needed:

i) Elevation Correction: Adjust all traces for elevation variations to a common

surface Reference Datum (Figure 3.16). If the survey is being conducted in a
low relief area, elevations corrections will normally be made to mean sea level
(MSL). However, in higher relief areas away from the coast, a higher reference

___________________________________________________________________________________ 75
Ted Cox – April 2002 Section 3: Processing Page
Introduction to Exploration Seismic
___________________________________________________________________________________

datum may be chosen, e.g., 500 or 1000 metres above mean sea level (AMSL).
The reason for this is to minimise the amount of correction necessary and hence
reduce the effect of any errors in its calculation.

Figure 3.16 – Elevation Datum Corrections.

where es - ds + eg is subtracted from the observed TWTs.

Vc

ii) Gun and Cable Correction (Marine): to adjust all traces to Mean Sea Level
(MSL) to compensate for the depth of operation of the gun arrays and the average
depth of the hydrophone stations in the cable (Figure 3.17).

Figure 3.17 – Marine Static Corrections.

where dg + d s is added to the observed TWTs.

Vw

iii) Weathering Correction : to remove the effects of a Weathering Layer. On most

land surveys it is found that one or more shallow unconsolidated layers exist
before true bedrock is reached. These may be due to very recent deposits (drift)
or to weathering of the rocks beneath. These unconsolidated layers are usually
very variable in character and thickness, and are also susceptible to variations in
the water table. Their anomalous properties will clearly affect travel times
measured at the surface and must be corrected. Observations made in the field,

___________________________________________________________________________________ 76
Ted Cox – April 2002 Section 3: Processing Page
Introduction to Exploration Seismic
___________________________________________________________________________________

often by a small dedicated survey team, will be used to estimate the nature of this
weathering layer and allow the correction to be calculated.

Figure 3.18 – Weathering Corrections.

where 2.Z _ 2.Z is subtracted from the observed TWTs.

Vw VC
The thickness Z and the velocity Vw for the weathered layer can be measured directly
and/or by correcting for the delay in arrival time for the direct wave (first arrival).

3.2.5 Residual Statics

After applying field static corrections, small delays ( ∆ tr) may still be present (Figure
3.19). These may be due to:
• using incorrect parameters for field static corrections;
• lateral velocity variations;
• non-ideal (complex) reflection ray paths.

Figure 3.19 – Residual Static Corrections to Hyperbola.

___________________________________________________________________________________ 77
Ted Cox – April 2002 Section 3: Processing Page
Introduction to Exploration Seismic
___________________________________________________________________________________

Residual static corrections try to remove these delays by applying small shifts (Figure
3.20). This is usually done automatically (autostatics) by cross-correlating a pilot trace
(the stacked trace of the CMP gather) with the NMO-corrected individual traces in a CMP.
A specific window length or gate is used to allow for ‘stretch’. Automatic statics only
work where the signal/noise ratio is good and for variations that occur across the width of a
CMP gather. Variations that have a longer wavelength than the gather width will not be
detected.

Figure 3.20 – Example of Residual Statics Application.

3.2.6 Long Wavelength Statics

Long wavelength statics occur where relatively large scale but subtle variations in near
surface seismic velocities are present. These may be due to mud-filled channels, laterally
discontinuous, low velocity horizons such as peat or ice, biogenic gas, etc. On seismic
sections they may be the cause of artificial structures such as pull-ups in underlying layers.
They usually show up as anomalous stacking velocities along a specific horizon. The
anomaly can be corrected for but great care must be used that genuine deep geological
structures are not removed.

___________________________________________________________________________________ 78
Ted Cox – April 2002 Section 3: Processing Page
Introduction to Exploration Seismic
___________________________________________________________________________________

3.2.7 Undesirable Effects

Data processing lessens the impact of various undesirable effects that obscure the reflected
signals; it can also compensate for some intrinsic deficiencies of the CDP method.

Some of the undesirable effects we need to remove are:

Multiples - where the sound is reflected repeatedly within a rock formation and, because
this takes time, registers as a deeper reflector;

Reverberations - reflections between the water surface and the sea-bed are a similar
phenomenon also known as ‘ringing’;

Diffractions - hyperbolic reflections generated from sharp discontinuities in reflectors, e.g.

as caused by faults, which appear like anticlinal formations.

Random noise - mainly unwanted reflections from within rock layers, horizontally
propagated and refracted sound (surface and near-surface effects), bubble pulsations from
the airguns and other effects also need to be reduced.

Stacking will reduce multiples and random noise, but the main data enhancement steps are
deconvolution, migration and filtering.

3.2.8 Sorting

We will recall that field data are acquired as a sequence of separate shot records, each one
sampling a series of adjacent CDP’s. These must be demultiplexed and edited before they
can be of use to us. In order for the processing to be completed, another large data sorting
process has to be performed.

As has been discussed in the preceding section, all the seismic traces contributing
information for one CMP (or CDP) must be collected together and arranged in order of
increasing offset distance between source and detector to form a CMP Gather. This
process must be repeated for every CMP along each seismic line in the survey.

In Figure 3.21 (next page), a seismic ray from shot S1 reflects from CDP position B and is
recorded by detector D2. From shot S2 the same CDP point is recorded at D1. To produce
a CMP record, all seismic traces containing reflections from sub-surface point B (and
hence all other reflections vertically above and below B) are selected from different shot
records by the computer. This rearrangement is known as sorting in 2D seismic (Figure
3.22).

___________________________________________________________________________________ 79
Ted Cox – April 2002 Section 3: Processing Page
Introduction to Exploration Seismic
___________________________________________________________________________________

Figure 3.21 – Two Field Records, 3 CDPs.

Figure 3.22 – Component Gather Traces for one CDP.

Once this sorting process has been performed, the data can be corrected for NMO effects
and all the traces for each CMP summed together or stacked.

Stacking causes true reflection pulses to enhance one another and random noise will tend
to be cancelled out. This process is repeated for all the CDPs on the survey line to create a
stacked seismic section, or profile.

3.2.9 Binning

The process of sorting 3D marine data is much more complex due to the fact that the
streamer cables cannot be towed exactly straight behind the acquisition vessel. Therefore,
navigation data have first to be merged with the seismic data to calculate the exact position
of the sources and recorders. The traces are then gridded and ‘binned’ into their respective
CMP groups. The size of the bins depends on the lateral resolution and multiplicity
required and these will determine the parameters used for acquisition. A typical bin is in
the order of 25 x 25m. The number of traces per bin will vary and therefore traces may be
weighted or duplicated within a bin in order to reduce or increase the coverage to a
standard figure for the survey.

___________________________________________________________________________________ 80
Ted Cox – April 2002 Section 3: Processing Page
Introduction to Exploration Seismic
___________________________________________________________________________________

3.2.10 Seismic Resolution – Vertical Resolution

Vertical resolution, or the ability to see closely spaced reflectors, is ultimately dependent
upon wavelength, which is a function of the formation velocity V, and frequency f.

λ=V/f
There are various ideas of what length of wavelet is required to resolve the top and bottom
of a reservoir.

The Rayleigh Criteria state that two events are resolved when the peak of one lines up with
the nearest trough of another, that is, one half of the wavelength. However, Widness
(1977) suggests that the thinnest resolvable bed in noise free data is one eighth of the
dominant wavelength. Figure 3.23 (after Badley, 1985) demonstrates the former case.
Maximum interference can be seen to occur at a thickness of one quarter wavelength or
tuning thickness.

Figure 3.23 – Vertical Resolution of a Thinning Limestone Wedge.

3.2.11 Horizontal (or Lateral) Resolution

In reality, any subsurface point that is sampled is actually a zone of sub-surface reflection
points (area of illumination) rather than a single point. This area is known as the Fresnel
Zone and its width depends on the wavelength of the seismic wavelet. . The radius of the
Fresnel Zone (rf ) is given by:

rf = ( λd + λ 2 )
2 16

___________________________________________________________________________________ 81
Ted Cox – April 2002 Section 3: Processing Page
Introduction to Exploration Seismic
___________________________________________________________________________________

Knowing that λ = V/f and d = ½Va.t where Va = the average velocity, we can get an
approximate idea of the Fresnel Zone width from:

rf = ( Va2t )
4f

A ray that diverges due to diffraction effects from the minimum time (or Fermat) path
begins to destructively interfere with the Fermat path ray due to a delay between the two
rays. Therefore, the maximum contribution to the amplitude of a seismic reflection at a
depth z occurs from the Fermat path ray. A reflection which is delayed by a distance of λ/2
relative to the Fermat path ray will contribute no energy to the amplitude of the reflection.
This delayed ray originates from a sub-surface point which is at a distance of
approximately (λ.z) from the reflection point of the Fermat ray in any direction.
Therefore, the Fresnel zone has approximately (2.λ.z) diameter. Deconvolution and
migration reduce the size and importance of the Fresnel zone, typically by two orders of
magnitude, because they help to focus the reflection.

Figure 3.24 – Horizontal Resolution: The Fresnel Zone.

Thus, the ability to resolve details along a reflector (Horizontal Resolution) is governed
by this relationship, and the separation between CMP’s (or bins) is therefore dictated by
the resolution required.

___________________________________________________________________________________ 82
Ted Cox – April 2002 Section 3: Processing Page
Introduction to Exploration Seismic
___________________________________________________________________________________

3.2.12 Velocity analysis and stacking

As discussed in section 3.2.3, reflections from any one reflector form a hyperbolic curve on
the gather because the sound takes longer to travel to the more distant detectors. In theory,
a CMP gather will consist of perfectly hyperbolic reflections that are symmetric around the
zero offset seismic trace. The hyperbola has the form:

Tx2 = To2 + x2/ Vs 2

Where Tx is the two-way travel time of a reflection recorded at a horizontal distance x

from the zero offset reflection, To. Vs is called the Stacking Velocity because, for any
time and offset, it will allow the appropriate amount of shift to correct the reflection
hyperbola and allow the reflector to stack correctly. The stacking velocity is approximately
equal to the effective velocity (or RMS Velocity) Vrms1 in horizontal layers or to Vrms /Cos
θ where the layers dip at an angle θ to the horizontal.

(Tx - To) is known as the Normal Move-Out (NMO).

After sorting, the traces may be displayed in order of increasing offset from the zero offset
or minimum time trace. This is known as an uncorrected CMP gather.

Figure 3.25 – Common Mid Point (CMP) Gather and Stack Trace.

Once NMO corrections have been applied, reflections will appear horizontal on the
corrected gather (Figure 3.25). Adding the traces in the corrected gather produces a
stacked trace. This process greatly improves the signal to noise ratio of the data since
stacking causes true reflection pulses to enhance one another, whilst random noise will
tend to be cancelled out. For example, the hyperbolic shape of a multiple or a diffraction is

___________________________________________________________________________________ 83
Ted Cox – April 2002 Section 3: Processing Page
Introduction to Exploration Seismic
___________________________________________________________________________________

more exaggerated than that of a primary reflection and they will therefore not line up
properly after good NMO correction2. Neither will direct waves such as ground roll, which
have linear NMO behaviour.

Due to the fact that amount of shift necessary for NMO correction is time and offset
dependent, the largest amounts of shift are needed in the shallow section at long offset.
This can create an effect known as NMO Stretch, or smear, which creates artificially low
frequency reflections on the outer traces (Figure 3.26). If the problem is severe, these low
frequencies must be muted out prior to stacking.

Figure 3.26 – Corrected Gather: Muting of NMO Stretch.

The correction process is repeated for all the CMPs on the survey line to create a stacked
seismic section, or profile. Most seismic sections used in hydrocarbon exploration are
time sections that have been processed to improve the quality of the reflections and bring
out subtle geological features, although these sections may sometimes be converted to a
vertical depth scale for reservoir development purposes. At this time, stacking velocities
become very important.
_________________________ __________________________

Note 1:Vrms is the root mean square (RMS) velocity, or effective velocity:

Vrms = Vm.Va = ∑ (Vi2. ∆ ti)/ ∑ ∆ ti

where Vi is the interval velocity, ∆ ti is the TWT across the interval, Va is the
average velocity (z/t) and Vm is the mean (weighted) velocity ( ∑ (Vi2. ∆ zi)/ ∑ ∆ zi).

2. The hyperbolic shape of a multiple has the form:

tx2 = k.to2 + x2/Ve2

The hyperbolic shape of a diffraction has the form:

tx2 = to2 + 4x2/Ve2

___________________________________________________________________________________ 84
Ted Cox – April 2002 Section 3: Processing Page
Introduction to Exploration Seismic
___________________________________________________________________________________

3.2.13 Velocity Analysis Methods

As can be seen from the formula above, in order to determine the necessary NMO
corrections before stacking, the velocities of propagation in the rock layers traversed by the
section need to be known. Analysis of traces with varying NMO corrections yields
velocity values, but more accurate data, which will be required during interpretation,
comes from special velocity surveys carried out in wells in conjunction with sonic logging.

There are four commonly used methods of analysing gather traces.

i) Constant Velocity Gather (CVG)

A CMP gather is corrected using the shifts calculated for a series of constant velocities.
Each version is displayed side by side. At the TWT at which the correct velocity applies
the display will show the reflector to be horizontal, i.e., perfectly corrected.

At different times down the gather, different velocities will cause reflections to become
horizontal. If the constant velocity is too high, the reflection hyperbola will not be fully
corrected. Conversely, if the constant velocity is too high, the hyperbola will be inverted,
hence over-corrected (Figure 3.27).

Figure 3.27 – Corrected Gather: Under- and Over- Correction.

A table of TWT versus Stacking Velocity can thus be compiled of the values that cause
observed reflections to be horizontal. The process can be repeated at regular intervals (1-2
km) along each line, and a velocity field constructed to check consistency over the survey
area. (See example, Figure 3.28).

___________________________________________________________________________________ 85
Ted Cox – April 2002 Section 3: Processing Page
Introduction to Exploration Seismic
___________________________________________________________________________________

Figure 3.28 – Constant Velocity Gather (CVG).

The velocity function will increase gradually with increasing travel time. However, it is
often difficult to tell multiple velocities from primary velocities. This process requires
manual interaction and is difficult in areas of noisy data and weak reflections.

Therefore, it is quite usual to also plot a panel of stacked traces at a range of velocity
functions around the optimum velocity function. These panels are known as constant
velocity stacks.

___________________________________________________________________________________ 86
Ted Cox – April 2002 Section 3: Processing Page
Introduction to Exploration Seismic
___________________________________________________________________________________

ii) Constant Velocity Stack (CVS)

In practice, the function used for the NMO correction relies heavily on the use of automatic
velocity analyses. The computer will stack the seismic traces in a CMP gather using a
range of possible stacking velocities. The idea is that any particular primary reflection will
show the highest amplitude after stacking when the NMO has best been corrected (see
example, Figure 3.29).

Figure 3.29 – Constant Velocity Stack (CVS).

In a similar way to the CVG, a table of TWT v Vs can be compiled by picking the features
that show the highest amplitudes.

During the velocity analysis stage, tables of Time and Stacking Velocity pairs usually have
two additional columns added. These are Interval Velocity (Vi) and Depth (z). These
values are calculated automatically using Dix’s Formula (see Section 3.2.15).

___________________________________________________________________________________ 87
Ted Cox – April 2002 Section 3: Processing Page
Introduction to Exploration Seismic
___________________________________________________________________________________

iii) Velocity Function Stack (VFS)

A variation on this theme is to estimate the trend of velocity against time and correct the
data with these values. Since there will be an element of error in these estimates, a series of
panels will be displayed using this trend and versions with velocities +/– a range of
percentages (Figure 3.30).

Figure 3.30 – Velocity Function Stack (VFS).

___________________________________________________________________________________ 88
Ted Cox – April 2002 Section 3: Processing Page
Introduction to Exploration Seismic
___________________________________________________________________________________

An interpreter who is familiar with the seismic expression of the survey area should be able
to tell from these velocity function stacks at what velocities the primary reflections have
best been stacked and the stacking velocities can be adjusted accordingly (Figure 3.31).

It is also essential to check that the picked velocities are reasonable on the basis of
borehole or lithological information. Below 2-3 seconds TWT it is often difficult to
distinguish primary reflections from noise and therefore the stacking velocities are usually
extrapolated from the reliable picks higher up.

Figure 3.31 – Velocity Functions used for Velocity Function Stack.

___________________________________________________________________________________ 89
Ted Cox – April 2002 Section 3: Processing Page
Introduction to Exploration Seismic
___________________________________________________________________________________

iv) Velocity Spectra

A more automated extension of this method involves producing a display of stacking

velocity versus TWT, contoured for reflection amplitude. It should then be possible to pick
an optimum velocity function from the graph (Figures 3.32 and 3.33).

Figure 3.32 – Amplitude Contours on Velocity Spectrum.

___________________________________________________________________________________ 90
Ted Cox – April 2002 Section 3: Processing Page
Introduction to Exploration Seismic
___________________________________________________________________________________

Figure 3.33 – Constant Velocity Gather Before and After Multiple Attenuation.

In fact, many software packages, e.g., GeoCentre’s SeisUp2000 program, allow the
computer to pick the high amplitudes automatically. The processing geophysicist is then
able interactively to examine these amplitude spectra together with the equivalent CVG
and CVS in independent windows on the same screen. This approach allows a rapid check
on its success and facilitates editing as necessary.

3.2.14 Velocities for 3D Surveys

In large 3D surveys initially a coarse grid of lines will be stacked. Sometimes the velocity
functions used for the initial stacks are derived from depth conversion functions used for
older 2D data. The interpreter will then interpret the important seismic markers on these
lines. The interpreted horizons will be digitised and contoured. These horizons then serve
as markers in the velocity analysis and the complete survey can be stacked with velocities
that concentrate on the important primary events.

In areas where little is known about the seismic expression of the strata, it may be worth
initially plotting individual CMP gathers at a range of possible constant stacking velocities
and picking the velocities which seem to best flatten important seismic reflections.
Multiples will have anomalously slow stacking velocities.

___________________________________________________________________________________ 91
Ted Cox – April 2002 Section 3: Processing Page
Introduction to Exploration Seismic
___________________________________________________________________________________

3.2.15 Dix's formula

During the velocity analysis stage, tables of Time and Stacking Velocity pairs usually have
two additional columns added. These are Interval Velocity (Vi) and Depth (z). These
values are calculated automatically the Vrms (~ Ve) values using Dix’s Formula:

Vi2 = (V22.t2) - (V12.t1)

(t2 - t1)
Dix’s formula can be used to calculate interval velocities from stacking velocities in areas
of simple geology. It is only applicable for:

i) approximately horizontal layers

ii) small offsets (Ve = Vs)

iii) no lateral velocity changes

3.2.16 Muting

In some cases F-K filtering is not sufficient to remove the effects of ground roll and
refracted arrivals. Also NMO stretching of the signal at far offsets is most exaggerated in
shallow reflections. These effects may be removed by blanking or muting, systematically
zeroing the early sample values on each trace and increasing this blanking zone with
increasing offset. The multiplicity of stack will therefore vary with depth (Figure 3.34).

Figure 3.34 – Blanking applied to remove First Arrivals and SURGICAL

BLANKING used to remove Ground Roll (usually removed by F-K filter).

___________________________________________________________________________________ 92
Ted Cox – April 2002 Section 3: Processing Page
Introduction to Exploration Seismic
___________________________________________________________________________________

3.3 DATA REDUCTION

3.3.1 CMP Stacking

Stacking will cause true reflections to be enhanced while random noise tend to cancel out.
This process is repeated for all the CDPs on the survey line. The fully corrected and
stacked traces will the be displayed side by side to create a seismic section (or profile).

Most seismic sections used by the oil industry are time sections that have undergone a long
sequence of data-processing steps designed to improve the quality of the reflections and
bring out subtle geological features. For particular purposes, after the principal reflectors
have been identified or 'picked', a time-section may be converted to a depth section.

3.3.2 NMO Multiple Removal

Normal stacking will attenuate multiples because they have a different NMO behaviour
from primary reflections. Further reduction of multiples can be achieved by producing a
multiple stack. This is done by first finding the stacking velocities of the multiples. A
multiple moveout correction is then applied and the gather is stacked to produce a trace
where the multiples are greatly enhanced. This trace is then scaled and subtracted from the
CMP record. The multiple moveout is removed and the NMO correction for the primary
reflections is applied and re-stacked (Figure 3.35).

Figure 3.35 – Differential NMO Multiple Removal: Step A) Æ Step G).

___________________________________________________________________________________ 93
Ted Cox – April 2002 Section 3: Processing Page
Introduction to Exploration Seismic
___________________________________________________________________________________

3.4 DATA ENHANCEMENT

Once the data corrections have been applied and every sample has been sorted to its correct
location, we can stack our CMP traces. This process will reduce drastically the data
volume (by a factor equal to the fold of stack), and improve the signal to noise ratio by the
square root of the fold.

However, we are still not finished with our processing task. We can now investigate what
other steps we can take to improve the appearance of our data to make interpretation of the
geology easier.

Deconvolution (‘decon’) aims to counteract the blurring of reflected wavelets by

‘recompressing’ them to the clean ‘spike’ emitted from the source. The result is clearer
reflections and the suppression of multiples.

Migration corrects distortions caused by plotting inclined reflectors as if they were

horizontal and vertically below the CMP; it also collapses diffractions.

Filtering removes undesirable noise to enhance the best reflections.

3.4.1 Deconvolution

A stacked seismic trace represents a signal (S(t)) that has been filtered by (convolved with)
the reflection response of the earth (E(t)).

Figure 3.36 – Deconvolving the Earth’s Filtering Effects.

If we know the original signal, then we can theoretically reproduce the earth’s filter by
cross-correlating the inverse of the signal (the deconvolution filter, 1/S(t)) with the stacked
traces. This mathematical technique is known as deconvolution (Figure 3.36 and 3.37).

___________________________________________________________________________________ 94
Ted Cox – April 2002 Section 3: Processing Page
Introduction to Exploration Seismic
___________________________________________________________________________________

Figure 3.37 – Schematic Illustration of Deconvolution Operator.

Figure 3.38 – Example of Deconvolution applied to Reverberations on Seismic Data.

___________________________________________________________________________________ 95
Ted Cox – April 2002 Section 3: Processing Page
Introduction to Exploration Seismic
___________________________________________________________________________________

Two or three types of deconvolution filter may be applied at various times during
processing. The functions of these different filters are:

i) Spiking - to shape the signal so that in theory it will then contain only
sharp, minimum phase wavelets;

ii) Predictive - to remove multiples (which are predictable), particularly short

period reverberations (Figure 3.39);

iii) Wavelet shaping - to convert minimum phase data to zero phase data.

The latter is not always feasible as it depends on being able to synthesise the minimum
phase wavelet by mathematical modelling in the frequency domain. However, zero phase
wavelets have the advantage that:

• for a given amplitude spectrum, the zero phase wavelet has the largest
amplitude and shortest time duration;

• the peak amplitude occurs at the impedance break (the position in time of
the reflection coefficient).

Both these properties are useful for auto-tracking data on interpretation workstations and
for lateral prediction studies.

Figure 3.39 – Comparison of Predictive and Wavelet Deconvolution.

A further processing technique is sometimes applied to zero phase data which involves
integration to produce a seismic impedance trace. The technique is known as inversion
and theoretically represents the acoustic properties of the geological strata. This technique
may be used for predicting lateral variations in porosity and/or fluid content away from a
well provided it can be assumed that the density and seismic velocity of the overlying and
underlying rock remain constant.

___________________________________________________________________________________ 96
Ted Cox – April 2002 Section 3: Processing Page
Introduction to Exploration Seismic
___________________________________________________________________________________

Deconvolution should sharpen reflections by undoing the effects of periodic oscillations

which follow on from the primary signal. The deconvolution filter is usually calculated
from the seismic data itself, except for Vibroseis where the filter can be constructed on the
basis of knowing the original (repeatable) signal. Deconvolution may be applied before
stacking (DBS) as well as after stacking (DAS).

3.4.2 Convolution

Convolution (and for that matter deconvolution) is a similar process to cross correlation.
For example, considering the sampled time series values:

E(t): 0 0 2 0 0 3 0 0 -1 0
* *
S(t): 3 -2
= =
T(t): 0 0 6 -4 0 9 -6 0 -3 2
(convolved trace)

The convolution (or deconvolution) can be done by shifting S(t) (or 1/S(t)) and
multiplying with the numbers adjacent to it in E(t) and summing the results (T(t)). This
can be written mathematically as:

T(t) = E(t) * S(t) or Tt = ∑ s −m

s −o Es.S. (t-s)

where E(t) is the filter which begins at Eo and ends at Em; S(t) is the input signal and T(t)
is the convolved result.

This same process of convolution (or deconvolution) can be done more easily by a
computer by using the Z transform. This involves converting E(t) and S(t) (or 1/S(t)) into Z
polynomials, then multiplying the two polynomials together and then converting back.

In theory we will only get an exact deconvolution of seismic data if we use an infinitely
long filter. In practice we have to truncate the filter. Provided that our data is minimum
phase, as is typical for most seismic energy sources, then the truncated filter is
approximately similar to the infinite filter. Otherwise we have to convert the data into
minimum phase. This phase conversion is done in the Z domain.

In most cases, the wavelet for deconvolution cannot be measured and must be estimated
from the data. An optimum finite filter for shaping deconvolution is derived as follows:

1. Correlate the seismic trace with itself (auto-correlate) to produce A(t);

2. Calculate by cross correlation the least squares filter (X(t)) necessary to convolve
the seismic trace into the desired output;

3. Deconvolve the auto-correlation (A(t)) with the least squares filter to derive the
deconvolution filter (1/S(t)).

___________________________________________________________________________________ 97
Ted Cox – April 2002 Section 3: Processing Page
Introduction to Exploration Seismic
___________________________________________________________________________________

The desired output after shaping deconvolution is a spike at the reflection boundary
(Figure 3.40).

Figure 3.40 – Derivation of Optimum Spiking Operator.

3.4.3 Spiking Deconvolution

The longer the filter length, the nearer a reflection will be to a spike. However, long filters
require more computation time and are liable to become unstable. To produce a close
approximation to a spike, the frequency spectrum must be flat (i.e., equal proportions of all
frequencies). Therefore, the high and low frequency part of a signal often has to be boosted
(frequency filtered) in the FK domain (Figure 3.41). This is called spectral whitening and
it is used in order to improve the vertical resolution of the seismic data. This type of
shaping deconvolution is known as spike deconvolution.

Figure 3.41 – Spiking Deconvolution operation.

___________________________________________________________________________________ 98
Ted Cox – April 2002 Section 3: Processing Page
Introduction to Exploration Seismic
___________________________________________________________________________________

Figure 3.42 – Noise introduced during Spike Deconvolution.

The signal/noise ratio is worsened by spike deconvolution but most of the introduced noise
can be cut out later by reducing the frequency range during final filtering.

3.4.4 Predictive Deconvolution

The reflection response of the earth is not only filtered by the wavelet but also by short
period multiples (reverberations). Due to the fact that these reverberations are predictable,
a deconvolution filter can be designed that will remove them (Figures 3.43 – 3.45).

Figure 3.43 – Dereverberation using Predictive Deconvolution.

___________________________________________________________________________________ 99
Ted Cox – April 2002 Section 3: Processing Page
Introduction to Exploration Seismic
___________________________________________________________________________________

Figure 3.44 – Dereverberation Operator.

Figure 3.45 – Reverberations removed by Predictive Deconvolution.

___________________________________________________________________________________ 100
Ted Cox – April 2002 Section 3: Processing Page
Introduction to Exploration Seismic
___________________________________________________________________________________

The dereverberation filter is usually combined with the shaping filter in one deconvolution
filter (Figure 3.46).

Figure 3.46 – Combined Dereverberation and Shaping Filter.

Other types of predictive deconvolution filters can also be designed. The time delay, T,
and the amplitude of the reverberation or multiple are estimated by auto-correlating the
seismic trace. The filter will then remove the specific amplitude from the seismic trace at a
predictable time after a primary reflection. Primary reflections are not predictable, i.e. they
do not show a high degree of auto-correlation, and the filter will not remove them.

Figure 3.47 – Predictive Delay Deconvolution.

___________________________________________________________________________________ 101
Ted Cox – April 2002 Section 3: Processing Page
Introduction to Exploration Seismic
___________________________________________________________________________________

A gap is required at the start of the predictive filter so that the filter does not try to cut out
very early reverberations which may coincide with primary reflection energy (Figure
3.48). This predictive gap will be determined by trial and error from a series of test panels.

Figure 3.48 – Predictive Gap Deconvolution.

This type of deconvolution is known as gap deconvolution and is different from spiking
deconvolution in that the wavelet is not shaped in the F-K domain.

3.4.5 Wavelet Deconvolution

This is applied at the end of the process in order to change the signal from minimum to
zero phase. Zero phase wavelets have the advantage that:

• for given amplitude spectrum, the zero phase wavelet has the largest
amplitude and shortest time duration;

• the peak amplitude occurs at the impedance break.

Both these properties are useful for auto-tracking data on interpretation workstations and
for lateral prediction studies.

To convert from minimum phase to zero phase requires that we know the wavelet. The
wavelet can be determined by:

• wavelet synthesis, i.e. measuring the source signature and removing ghost,
absorption and instrument effects;

• wavelet extraction, i.e. using statistical techniques on an isolated known

reflector.

Wavelet extraction is performed in the frequency domain. The amplitude is obtained for a
known reflector. This spectrum is edited to produce a smoothed spectrum. Then,
assuming that this is the spectrum for a minimum phase wavelet, the phase spectrum can
be calculated using the Hilbert Transform. The wavelet can then be reconstructed in the

___________________________________________________________________________________ 102
Ted Cox – April 2002 Section 3: Processing Page
Introduction to Exploration Seismic
___________________________________________________________________________________

time domain (Figure 3.49). Alternatively, adjacent traces can be cross-correlated and
auto-correlated in order to calculate a mean signal spectrum (as well as a mean trace
spectrum and a mean noise spectrum).

Figure 3.49 – Wavelet Deconvolution.

Once the minimum phase wavelet has been obtained, it can then be deconvolved with the
seismic trace to produce a zero phase wavelet. The wavelet deconvolution can be split into
two parts - a phase deconvolution and an amplitude deconvolution. The phase
deconvolution is intended to convert the wavelets on the seismic trace to zero phase
wavelets. This is done by subtracting the phase spectrum of the wavelet from the phase
spectrum of the seismic traces. The amplitude deconvolution consists of whitening the
spectrum. This is normally followed by zero phase frequency band-limiting using a user
defined filter.

___________________________________________________________________________________ 103
Ted Cox – April 2002 Section 3: Processing Page
Introduction to Exploration Seismic
___________________________________________________________________________________

Figure 3.50 shows and example of wavelet deconvolution, Figure 3.51 shows part of a
deconvolution gap test, and Figure 3.52 illustrates a sample seismic line before and after
predictive deconvolution, showing the autocorrelogram on which the operator is designed,
and the equivalent correlation after deconvolution.

Figure 3.50 – Example of Wavelet Deconvolution.

___________________________________________________________________________________ 104
Ted Cox – April 2002 Section 3: Processing Page
Introduction to Exploration Seismic
___________________________________________________________________________________

Figure 3.51 – Deconvolution Gap Test.

___________________________________________________________________________________ 105
Ted Cox – April 2002 Section 3: Processing Page
Introduction to Exploration Seismic
___________________________________________________________________________________

Figure 3.52– Use of Predictive Deconvolution for Multiple Removal.

___________________________________________________________________________________ 106
Ted Cox – April 2002 Section 3: Processing Page
Introduction to Exploration Seismic
___________________________________________________________________________________

Deconvolution (Summary)

Purpose:
• Attenuate short period multiples (0 - 250ms)
• Increase temporal resolution
Method:
• Estimate the amplitude spectrum from autocorrelations
• Derive an inverse filter to broaden the amplitude spectrum
• Apply filter to the data
Parameters:
• Design gate for autocorrelation
• Filter length
• Pre-whitening percentage
• Predictive length

3.4.6 Seismic Inversion

A zero phase processed filter can be integrated to give a seismic impedance trace (Figure
3.53). This is a mathematical representation of the acoustic properties of the geological
strata. The technique is sometimes known as seismic inversion and has applications when
evaluating area with many wells. With suitable correlation, the inverted traces of a seismic
line can be interpreted to identify stratigraphic changes which may assist in reservoir
characterisation.

Figure 3.53– Wavelet Inversion.

___________________________________________________________________________________ 107
Ted Cox – April 2002 Section 3: Processing Page
Introduction to Exploration Seismic
___________________________________________________________________________________

3.4.7 DMO and Migration

We have oversimplified our consideration of the earth by assuming that reflections are flat,
such that non-zero offsets can be corrected to zero-offset using the basic NMO equation
(Section 3.2.12). Now we need to consider the real earth, and make some adjustments to
our approach.

The position of the reflection point on a dipping event does not in reality lie vertically
beneath the CMP (Figure 3.54). The objective of migration is to bring seismic waves into
focus by moving reflections to their correct position and amplitude. The Fermat (minimum
time) raypath is the raypath that is normal to the reflection surface.

The zero offset ray has

the shortest travel time.
On the CMP record the
hyperbola is always
symmetrical, however
the effective velocity
depends on the dip of the
reflector.

DMO = Dip Move Out

correction.

Figure 3.54– Common Mid Point Geometry for a Dipping Reflector.

3.4.8 Dip Moveout (DMO)

The dip moveout process moves reflection energy for non-zero offset data to the CMP,
where the normal incidence ray is located (Figure 3.55).

As stacking velocity is dip dependent, we must apply a correction to allow for dip:

Vstack = Vrms/cos θ

After DMO the stacking velocity is dip independent, and a better stack will result.

___________________________________________________________________________________ 108
Ted Cox – April 2002 Section 3: Processing Page
Introduction to Exploration Seismic
___________________________________________________________________________________

Figure 3.55– Reflection Point Displacement in CMP Gathers.

___________________________________________________________________________________ 109
Ted Cox – April 2002 Section 3: Processing Page
Introduction to Exploration Seismic
___________________________________________________________________________________

Method

DMO is a ‘migration’ process in the sense that data are moved spatially across traces to
their correct positions. The process is applied in common offset sections. DMO smears
the energy of the sample at the apex of the ellipse along the elliptic trajectory (Figures
3.56 and 3.57).

Figure 3.56– DMO Impulse Response Hyperbolae.

Before DMO After DMO

Figure 3.57– Effect of DMO on Horizontal (Top) and Dipping (Bottom) Reflectors.

___________________________________________________________________________________ 110
Ted Cox – April 2002 Section 3: Processing Page
Introduction to Exploration Seismic
___________________________________________________________________________________

Benefits of DMO

DMO migrates the traces to zero offset so that:

• Constant-offset sections become identical to hypothetical zero-offset section

• CMP gathers become more accurate common reflector point (CRP) gathers.

Resultant data improvement

• Stacking of conflicting dips
• Continuity of reflections improved
• Attenuation of back-scattered energy
• Signal to noise ratio improved
• Dead traces interpolated
• Dip independent stacking velocities
• Dipping events are correctly positioned
• Reduction of CMP smear

Note: DMO does not create data! – It just focuses the image better.

3.4.9 Migration

DMO is a costly pre-stack process that is not always applied. If we have continued
processing through stacking using the assumption that the minimum time reflection is
derived from directly below the receiver, a stacked or unmigrated reflection point will
appear down-dip of the real (migrated) reflection and be less steeply dipping (Figure 3.58).

Figure 3.58– Erroneous Positioning of Reflector due to CDP Assumptions.

___________________________________________________________________________________ 111
Ted Cox – April 2002 Section 3: Processing Page
Introduction to Exploration Seismic
___________________________________________________________________________________

On stacked sections anticlines look larger than they really are and synclines look narrower.
This often means that folds have a bow-tie like appearance (Figure 3.59).

Figure 3.59 – Appearance of Curved Reflectors on Seismic Sections:

(a) a gentle syncline;

(b) a steeper syncline, producing the ‘bow-tie’ reflection;
(c) an anticline.

___________________________________________________________________________________ 112
Ted Cox – April 2002 Section 3: Processing Page
Introduction to Exploration Seismic
___________________________________________________________________________________

Examples of data before and after migration are shown in Figures 3.60 and 3.61.

Figure 3.60 – Effects of Migration: putting reflectors in their correct position.

___________________________________________________________________________________ 113
Ted Cox – April 2002 Section 3: Processing Page
Introduction to Exploration Seismic
___________________________________________________________________________________

Figure 3.60 – Characteristic ‘Bow-Tie’ and its Migration.

___________________________________________________________________________________ 114
Ted Cox – April 2002 Section 3: Processing Page
Introduction to Exploration Seismic
___________________________________________________________________________________

Migration Methods

The migration of a stacked reflection point to its true position on a seismic section can
either be cone by ray tracing or by moving the dipping event along a wavefront until it
coincides with normal incidence.

Manual Migration Using A Wavefront Chart

Figure 3.63 – A Wavefront Chart.

The shapes of the wavefronts depend on the change of velocity with depth. One can
perform migration manually by calculating the angle of emergence of a seismic reflection
and then tracing the ray path along a wavefront chart (Figure 3.63). Any ray is normal to
the wavefronts.

Manual Migration using Diffraction Curves

Another way of migrating seismic data is by moving reflection points along diffraction
curves (Figure 3.64). This is valid because reflections in the sub-surface can be regarded
as a set of many individual diffractions (Figure 3.65). The true (migrated) position of a
reflection lies at the apex of the diffraction curve (Figure 3.66).

Figure 3.64 – Migration along Diffraction Curves.

___________________________________________________________________________________ 115
Ted Cox – April 2002 Section 3: Processing Page
Introduction to Exploration Seismic
___________________________________________________________________________________

Figure 3.65 – A Reflector is made up of many Diffraction Points.

It can be seen from the set of pictures above that

.
reflections are generated from the envelope of many
diffraction curves. The true (migrated) position of the
reflector lies at the apex of the diffraction curves

___________________________________________________________________________________ 116
Ted Cox – April 2002 Section 3: Processing Page
Introduction to Exploration Seismic
___________________________________________________________________________________

Figure 3.66 – Migration along the Diffraction Curve.

Computerised Migration Methods

The methods above demonstrate two laborious manual migration techniques. Fortunately
however, any such repetitive process is ideally suited for a computer. An artificial gather
of stacked seismic traces is made for each trace. The gather is then corrected for the
normal moveout of diffractions. This is done for all traces and finally all the migrated
traces are re-stacked (Figure 3.67). Only at the true mid-point position of a diffraction will
the data stack up properly. At all other positions the data will destructively interfere.

Figure 3.67 – Computerised Migration Steps.

___________________________________________________________________________________ 117
Ted Cox – April 2002 Section 3: Processing Page
Introduction to Exploration Seismic
___________________________________________________________________________________

The three main methods of computer migration are listed below with their relative benefits.

METHOD STRENGTHS WEAKNESSES

Kirchhoff steep dips migration noise

complex velocities operator aliasing
slow

Finite Difference complex velocities steep dips (different

algorithms, 15o & 30o)

F–K steep dips complex velocities

very fast

Figures 3.70 to 3.73 show comparisons of different methods from GSI publicity material.

Choice of Parameters

Generally the velocities that are used for the migration operator are the stacking velocities
or the stacking velocities adjusted by a fixed percentage. For example, 90%, 95%, 100%,
105% and 110% of the stacking velocities can be used for a migration test and the best
apparent migration is chosen manually by the interpreter. The operator should be long
enough to migrate the steepest dip and to make sure that all diffraction curves are
completely collapsed, thus giving optimum fault definition (Figures 3.68 and 3.69).
However aliasing can occur within long migration operators.

Figure 3.68 – Required Migration Operator.

___________________________________________________________________________________ 118
Ted Cox – April 2002 Section 3: Processing Page
Introduction to Exploration Seismic
___________________________________________________________________________________

Figure 3.69 – Migration Operator.

To avoid aliasing due to migration:

• a smaller trace operator can be used (by decreasing station spacing during
acquisition or interpolating extra traces during processing);

• limiting the maximum frequency of the data (by using a high-cut filter);

• by limiting operator length (although this may prevent proper migration in

areas of steep dips).

Migration ‘smiles’ arise where anomalously high amplitudes are smeared out by the
migration operator and are often seen in the deeper part of the section where reflections are
weak and noise predominates.

___________________________________________________________________________________ 119
Ted Cox – April 2002 Section 3: Processing Page
Introduction to Exploration Seismic
___________________________________________________________________________________

Figure 3.70 – CDP Stacked Section.

Figure 3.71 – Kirchhoff Migration.

___________________________________________________________________________________ 120
Ted Cox – April 2002 Section 3: Processing Page
Introduction to Exploration Seismic
___________________________________________________________________________________

Figure 3.72 – Finite-Difference Migration.

Figure 3.73 – F - K Migration.

___________________________________________________________________________________ 121
Ted Cox – April 2002 Section 3: Processing Page
Introduction to Exploration Seismic
___________________________________________________________________________________

3.4.10 Pre-Stack Depth Migration (PSDM)

In cases where migration was not wholly successful due to complex geology, a pre-stack
depth migration will tend to yield a better result. This involves:

1. interpreting the original migrated section

2. developing a depth model for the interpretation

3. de-migrating the original section

4. de-stacking the CMP gathers and removing the NMO correction

5. correcting the CMP records for diffraction NMO as calculated by ray

tracing from the depth model (this may involve several iterations)

6. re-stacking the diffraction NMO corrected gathers

The main advantage of this method is that because a depth model is used, the pre-stack
migrated seismic data can be displayed in either TWT or depth (Figure 3.74). In areas of
complex geology, e.g., around a salt dome, this approach can yield a far more meaningful
result. It will however be directly related to the quality of the original depth model.

In order to carry out 3) and 4) one needs to know the stacking and migration velocities that
were used for the original section.

Figure 3.74 – Pre-Stack Depth Migration.

___________________________________________________________________________________ 122
Ted Cox – April 2002 Section 3: Processing Page
Introduction to Exploration Seismic
___________________________________________________________________________________

3.4.11 Frequency filtering

Prior to final display, a band pass filter is normally applied to the seismic data in order to
lessen the effects of white noise on the signal. The filter is normally time variant and will
only cut out a proportion of the amplitudes at the lower and higher ends of the frequency
range. An example is shown in Figure 3.75.

It is customary to perform filter trials using a selection of high and low filter cuts on a
small test panel of data from a line within the survey and choose the time variance based
on the appearance.

___________________________________________________________________________________ 123
Ted Cox – April 2002 Section 3: Processing Page
Introduction to Exploration Seismic
___________________________________________________________________________________

3.4.12 Display parameters

Paper copies of both stacked and migrated seismic profiles are normally produced on at
least two different scales (e.g. 1:50000, 5cm per second and 1:25000, 10cm per second).

A gain correction may be applied in order to equalise the shallow and deep amplitudes.

The amount of overlap of traces will also be defined by the interpreter. The overlap
influences the visual continuity of the reflections and the variations in amplitude that can
be detected.

3.4.13 Section Label

One should always check the processing parameters that were used on the header of the
seismic section or in the processing report. It is most important that the polarity of the data
is known. It is normal convention that a compressional wave is recorded as a negative
number and displayed as a kick to the left (white). Therefore, the top of thick hydrocarbon-
filled porous sands (soft kick) will tend to stand out as black reflections. Other parameters
that should be checked as a matter of course are the phase of the data; what sort of
migration, if any, has been applied to the data; the fold; and the frequency content at the
level(s) of interest.

3.4.14 Data Formats

Normally nowadays seismic data is delivered in a digital format on magnetic tape (e.g.
Exabyte cassette) for loading onto an interpretation workstation. The industry standard
format for processed digital seismic data is known as SEG-Y. Most of the interpretation
software packages use their own data format within the workstation environment.
Therefore, SEG-Y has to be converted, for example, to Charisma (GeoQuest) or Landmark
format. This can either be done when loading the data onto the workstation or prior to
loading by the processing contractor. Display parameters are changed interactively by the
interpreter on the workstation.

___________________________________________________________________________________ 124
Ted Cox – April 2002 Section 3: Processing Page
Introduction to Exploration Seismic
___________________________________________________________________________________

3.4.15 3D Data Processing (Additional Comments)

Because of the greatly improved seismic resolution of 3D seismic imaging, there has been
an effort to reduce the cost of 3D data acquisition and shorten the time it takes to acquire
and process the large volumes of data acquired.

In the past it could take up to 24 months to process the recordings from a 3D survey.
Acquisition time has been cut by specially designed survey vessels deploying up to ten
multiple streamers at a time, or by using multiple vessels. These techniques allow a swath
of seismic data to be acquired in the same time it would take to record a single 2-
dimensional line.

New techniques of data compression are being tried to allow the transmission of the raw
seismic records from the acquisition vessel to the shore for immediate processing, in an
effort to get the data to the interpreters faster.

The fundamentals behind the 3D method are that since all subsurface structures are 3-
Dimensional any 2D migration cannot image out of the plane events. The cost of a 3D
seismic survey is usually a fraction of the cost of wells to develop a field.

3D surveys are currently not only for exploration but also to monitor the progress of field
developments (4D).

3D Velocity Analysis
• Velocities are picked on a regular grid of the order of 500m x 500m
• Velocities normally use inline data only
• Velocity time slices are produced for a number of constant times
• Horizon velocity surfaces can be produced
• Velocity maps can be used to extract velocity anomalies

3D DMO
• 3D DMO can handle non-uniformity of azimuth and offset distribution
• 3D DMO improves signal to noise
• 3D DMO still remains a 2-Dimensional operation

The Benefits of 3D
• Increased density of spatial coverage
• Enhanced subsurface imaging
• Improved Signal to Noise ratio
• Improved resolution
• Field Delineation
• Field Management
• Field Monitoring
• 3D Migration

The greatest benefit of 3D surveys is 3-Dimensional Migration as this has the ability
to image a 3-Dimensional subsurface.

___________________________________________________________________________________ 125
Ted Cox – April 2002 Section 3: Processing Page
Introduction to Exploration Seismic
___________________________________________________________________________________

SECTION 4:

SEISMIC INTERPRETATION
The understanding of geology from the seismic data.

4.0 OVERVIEW

The seismic method, particularly reflection seismic, has been the most successful tool used
for oil and gas exploration and production, so much so that the terms ‘seismic’ and
‘geophysics’ as used by the oil industry are virtually synonymous. In that way the
‘geophysicist’, unless the term is immediately qualified with such additions as ‘gravity and
magnetics’ or ‘processing’, is generally taken to be a seismic interpreter. As the seismic
industry has progressed from providing data that have chiefly been used to evaluate
geological structure to data that can be used to give information about the rock and fluid
properties so the skills of a seismic interpreter have had to increase from geology to rock
mechanics.

Seismic sections provide 2-dimensional views of the subsurface structure. We have

explained the special shooting and processing techniques needed to produce 3-dimensional
seismic images. The concept is of a 3D ‘data volume’ which can be viewed in many
different ways including vertical sections and horizontal ‘time-slices’. Originally 3D
seismic surveys were focussed over oil and gas fields where a more detailed subsurface
picture was needed to help confirm the efficient location of production wells. In the last
decade exploration has shifted to smaller and more subtle traps and 3D seismic has become
more widely used for exploration work. The large volumes of data generated by even a
small 3-D survey mean that computer workstations have become an essential tool for
interpreting data quickly.

The interpreter’s job is to extract geological meaning, both structural and stratigraphic,
from these geophysical data. This process will require the best possible acquisition and
processing to have been performed, and also a knowledge of local geology from outcrop
and pre-existing wells. The interpreter will pick strong reflectors and correlate these picks
around the survey grid to produce sections showing geological structure and sedimentary
details. The interpretation may be done by hand on paper sections, but is more normally
done interactively on an interpretation workstation.

Structural contour maps of major rock units will be produced by correlation with well data
and ‘isopach’ maps will be constructed showing the thickness of these units. The digital
horizon picks are entered into a gridding and contouring program. Contour maps can be
plotted, or 3D colour and ‘illuminated’ images can be generated, to illustrate the subsurface
structure. The seismic maps are used to identify structures that would either require more
detailed seismic work or would warrant wildcat drilling.

The interpreter unravels the structural history of the sedimentary basin and correlates the
seismic with the stratigraphic succession. Sequence stratigraphy is the study of the
depositional interrelationships of sedimentary rocks deduced from an interpretation of

___________________________________________________________________________________ 126
Ted Cox – April 2002 Section 4: Interpretation Page
Introduction to Exploration Seismic
___________________________________________________________________________________

seismic and well data. This approach is used for finding subtle sedimentary traps involving
an understanding of palaeo-environments and changes in porosity. High amplitude ‘bright-
spots’ may indicate gas and a gas-fluid interface may be recognised by ‘flat’ spot against
the dipping reflectors.

A workstation, or PC, allows the interpreter to map specific reflectors interactively by

picking on screen or using an ‘auto-picking’ routine to follow strong horizons over the
whole 3-D survey. Digital picks, including fault planes, will be transferred directly into the
mapping software. Visualisation software is an additional tool that allows the interpreter to
view the whole 3D-data set as a cube and rotate or cut it at any angle, allowing a picture of
the subsurface geometry to be quickly seen.

Figure 4.1 shows how seismic data provides the basic working material for the
explorationist who is evaluating a whole basin or geological province through to the
hydrocarbon asset manager who is trying to produce his oil or gas field most efficiently.

Figure 4.1 – A summary of seismic data and its uses.

___________________________________________________________________________________ 127
Ted Cox – April 2002 Section 4: Interpretation Page
Introduction to Exploration Seismic
___________________________________________________________________________________

4.1 SEISMIC DATA INPUT

The raw materials of the interpreter are the seismic sections (either on paper or loaded on a
workstation), a shot point basemap showing specific shot (or CDP) and well locations and
block boundaries, well logs and any pre-existing geological and geophysical maps and
reports. There is a wealth of information held within a seismic section and the task of
seismic interpretation quite definitely requires a multi-disciplined team.

4.1.1 Display Technique

Figure 4.2 demonstrates a historical progression in the way that seismic wiggles have been
displayed. Although modern, high quality data is being displayed you can see the
improvement in interpretability as progressively more sophisticated methods are employed.
Display technique will always remain an important part of improving the interpretability of
seismic data.

Figure 4.2 A: Wiggle Trace

Whatever they mean to the person in

the street, wiggles are bread and butter
to the seismic interpreter. With modern
displays it is easy to forget that this is
our original data. Electrical signals
whose shapes mirror the response of
the Earth to our acoustic pulse. The
shape and strength of each wiggle
holds the geological information in
which we are interested.

Figure 4.2 B: Variable Area Section

By filling in some part of the area under

the wiggle the structural shape
becomes clearer. The proportion that
is blacked in, the bias, can be an
important parameter for the section’s
clarity.

___________________________________________________________________________________ 128
Ted Cox – April 2002 Section 4: Interpretation Page
Introduction to Exploration Seismic
___________________________________________________________________________________

Figure 4.2 C: Variable Area Section

The troughs are now filled in by red

colour and the colour’s intensity
increases with reflection power. Seismic
character comes through.

Figure 4.2 D: Variable Density Section

Instead of having a trace form, the

signal’s variations are represented by
intensity of colour. Seismic character
starts to become a strong feature of the
section as well as the structure it
represents.

Figure 4.2 – Seismic Display.

4.1.2 Stacked/Migrated Data

Today seismic processing power is so great that when we look at the modern seismic
profile, most likely on the workstation, it is hard to believe that we are not looking at a
perfect cross section through the earth. We must always remember this!

This hasn’t always been the case. Figure 4.3 shows an early single fold cover section from
the Netherlands. Figures 4.4 and 4.5 show the benefit of improved data acquisition and
processing, from the Brent Field, comparing a 2D section recorded in 1966 with a 3D line
recorded 20 years later. Further improvement would be seen if we recorded the line today.

We have already discussed the need for migration in the processing sequence. Throughout
the 1980’s post-stack time migrated data was the norm and it was very rare if a survey, or
section, was processed without being migrated. Now, as the 21st century begins it is
increasingly common for interpreters to work with pre-stack depth migrated time or depth
sections.

___________________________________________________________________________________ 129
Ted Cox – April 2002 Section 4: Interpretation Page
Introduction to Exploration Seismic
___________________________________________________________________________________

Figure 4.3 – Single Fold Section from The Netherlands.

Figure 4.4 – A 2D Line over the Brent Field recorded in 1966.

Figure 4.5 – A 3D Line over the Brent Field recorded in 1986.

___________________________________________________________________________________ 130
Ted Cox – April 2002 Section 4: Interpretation Page
Introduction to Exploration Seismic
___________________________________________________________________________________

4.1.3 Pitfalls

A pitfall is an error waiting to happen; the term comes from the covered pit dug by
primitive tribes to catch their unsuspecting prey.

We must always remember that our seismic section is not a geological cross-section.
Failure to do this will mean that we miss some significant differences between the geology
in depth and the geophysical representation of it in two-way travel time. We will have
fallen into the trap!

4.1.4 Migration

If the reflecting strata are dipping more than 10 o a stack section will show them in the
wrong position. Geological structures will also be mis-represented. We have seen some
examples of the problem and the processing solution in Section 3.4.8.

Figure 4.6 shows the effect of migration on simple geological structures.

• An anticlinal structure on a migrated section looks smaller than on a stack section.

• Conversely, a synclinal structure on a migrated section looks larger than on a stack
section.

Figure 4.6 A

Stack sections show

anticlines as broader
structures.

Broadening increases
with depth to an extent
that synclines are
‘lost’.

Figure 4.6 B

Stack sections show

synclines as tighter
structures.

Figure 4.6 – The Structural Distortion shown by Stacked Sections.

___________________________________________________________________________________ 131
Ted Cox – April 2002 Section 4: Interpretation Page
Introduction to Exploration Seismic
___________________________________________________________________________________

Figures 4.7 to 4.9 illustrate some problems encountered on stacked sections that can be
tackled by migration.

Figure 4.7 - A deeply buried syncline appears as a ‘bow-tie’ on a stacked

section.

___________________________________________________________________________________ 132
Ted Cox – April 2002 Section 4: Interpretation Page
Introduction to Exploration Seismic
___________________________________________________________________________________

Figure 4.8 - Stacked sections from structurally complex areas will be difficult
to interpret.

Figure 4.9 – Sideswipe.

Because of spurious ray path effects, reflections originating from out of the plane of the 2D
seismic section can cause confusing structural line-ups known as ‘sideswipe’.

___________________________________________________________________________________ 133
Ted Cox – April 2002 Section 4: Interpretation Page
Introduction to Exploration Seismic
___________________________________________________________________________________

Figure 4.10 Common problem caused by crossing an abrupt velocity

discontinuity.

___________________________________________________________________________________ 134
Ted Cox – April 2002 Section 4: Interpretation Page
Introduction to Exploration Seismic
___________________________________________________________________________________

4.1.5 Other problems

Other problems of interpretation could be:

• Apparent thinning down dip due to interval velocity increases with depth.
• Apparent rollover into faults caused by diffractions at the fault edge.
• Apparent rollover due to velocity contrasts across the fault.

4.1.6 Multiples and Noise

A multiple reflection is a reflection event that is produced by two or more successive

reflections from a combination of layer boundaries, including the Earth's surface. The
creation and suppression of multiples and other noise has been discussed in the acquisition
and processing sections. (See Figure 4.11).

Figure

Short delay multiples

• follow primaries at short time intervals,
• are mostly not distinguishable as a separate event,
• may interfere with the primary reflection, lengthening its tail and impairing
vertical resolution.

Long delay multiples

• could have twice the arrival time of the primary reflection from the given
interface,
• can often be mistaken for a primary event when it is conformable to primary
reflections.

___________________________________________________________________________________ 135
Ted Cox – April 2002 Section 4: Interpretation Page
Introduction to Exploration Seismic
___________________________________________________________________________________

Seismic noise can be

• coherent, that is traceable across several adjacent sections. Examples are

surface waves, reflections or reflected refractions from near surface
structures such as fault planes, channels, or human activity, such as traffic
or drilling rigs. Most of this noise, except multiples travel horizontally, so
beware of some coherency filtering that can accentuate the noise.

• incoherent or random, that is dissimilar and unpredictable.

4.2 SEISMIC CHARACTER

Seismic character is a mostly subjective definition of the recognisable aspect of the seismic
event. It is usually an amplitude or frequency effect and can also be described in terms of
the structural relationship with other seismic events. The interpreter will usually have
some input into the parameter choices made in processing, and so has some control over
personal preferences to help recognise different reflectors in the interpretation process.
This section deals with some of the controls that give the seismic wavelet its character.

The basic wiggle trace is derived from an electrical signal from either a geophone which
measures ground acceleration in response to a seismic disturbance or a hydrophone, which
measures pressure differences. Figure 4.12 shows some variations in reflection character.

4.2.1 Reflection Character, Acoustic Impedance

We recall that the strength of a seismic reflection from a lithological interface is controlled
by the angle of incidence and by the differences in acoustic properties in the materials
above and below the reflecting surface and are defined by Zoeppritz’s equations (Section
1.2.3)

Amplitudes from a reflecting surface along a seismic gather as opposed to a zero offset
seismic section might vary because of the increasing offset, i.e., the angle of incidence
increases with offset. This is the basis of Amplitude Versus Offset (AVO) studies.
Energy that passes through the reflecting medium is transmitted energy. T is the
Transmission Coefficient and is defined as the ratio of incident to transmitted amplitudes.

R+T=1
or
T = (1 - R)

For a Reflection Coefficient > 0 a compression front of an incident wave will be reflected
as a compression, that is, there is no polarity reversal. If the reflection coefficient is < 0 a
compression front of an incident wave will be reflected as a rarefaction, that is there will be
a polarity reversal, or phase inversion. Thus, whether a reflection is a trough or a peak will
depend on the fundamental polarity convention adhered to and whether the Acoustic
Impedance of the overlying unit is greater or smaller than the deeper, underlying unit.

___________________________________________________________________________________ 136
Ted Cox – April 2002 Section 4: Interpretation Page
Introduction to Exploration Seismic
___________________________________________________________________________________

It is vital for the interpreter that the polarity convention used in acquisition and processing
is clearly known from the section labels and supporting data. Polarity information,
particularly changes in polarity, along a reflector can have great geological significance.

Figure 4.12 – A Variable Area and Wiggle Display showing a Variety of

Wavelet Character.

___________________________________________________________________________________ 137
Ted Cox – April 2002 Section 4: Interpretation Page
Introduction to Exploration Seismic
___________________________________________________________________________________

4.2.2 Geological Controls on Velocity

The velocity of propagation of seismic waves depends on a number of factors that are
related to each other.

• The lithology: mineral composition, texture etc.

• The porosity and nature of fluids contained in the rock.
• The state of fracturing of the rock.
• The total depth of burial and the differential pressure of the formation under
consideration.

Lithology: The following table gives a guide to the Range of velocities you might expect.

_________________________________________________________________________
MATERIAL VELOCITIES MATERIAL VELOCITIES
(km/s) (km/s)
Vp Vp Vs
Surface alluvium 0.5 – 2.0 Gneiss 3.1 – 5.4
2000 m deep 3.0 – 3.5 Basalt 3.0 – 6.4
Weathered soil 0.3 – 0.9 Granite 4.0 – 5.7 2.1 –3.3
Marls 1.8 – 3.8 Fresh water 5oC 1.44
Sands 0.2 – 2.0 15oC 1.48
Limestone 1.7 – 6.4 25oC 1.50
o
Sandstone 1.4 – 6.0 35 C 1.53
Dolomite 3.3 – 7.6 Sea water 1.48 – 1.53
Salt (NaCl) 4.6 Ice 3.2 – 3.7 1.6 – 1.85
Anhydrite (CaSO4) 6.1 Oil 1.3 – 1.4
Coal 2.0 – 3.4 Methane gas 0.4
_________________________________________________________________________

Porosity and Nature of the Fluids Contained in the Rock:

The velocity in porous rocks is governed by the BIOT-GEERSTMA formula which allows
us to distinguish between two main types of rock.

Unconsolidated or very slightly consolidated rocks.

Their densities increase with depth as a result in the decrease of porosity. Density
ρ and porosity φ of these types of rocks at a depth z are given by equations of the
type
ρ = A + B ecz

φ = a ecz

For the shallow deposit A a slight increase in gas saturation rapidly reduces the
velocity but any further gas saturation will have less of an effect.

___________________________________________________________________________________ 138
Ted Cox – April 2002 Section 4: Interpretation Page
Introduction to Exploration Seismic
___________________________________________________________________________________

This explains the Bright Spot phenomenon and also illustrates that a direct
hydrocarbon indication doesn’t necessarily indicate a commercial find, it can also
indicate a small percentage of gas saturation.

Consolidated Rocks

Their physical characteristics are less affected by their depth below the surface and
the nature of the fluids contained in the reservoir has little effect on the velocity.

Influence of Fractures:

Micro-fractures reduce the movement of a rock's elementary particles in response to a

seismic disturbance which leads to a reduction in the seismic velocity.

Depth Below Surface:

In general seismic velocity increases with depth of burial due to reduction of porosity and
increase in density. This will be looked at more closely under Depth Conversion.

This is not always true, due to Differential Pressure ∆P.

∆P = Pr – Pf

Pr is the pressure due to the overlying layers

Pf is the pressure of the fluids contained in the formation

High interstitial pressure is normally referred to as sub-compaction or overpressure.

4.3 INTERPRETATION METHODOLOGY

4.3.1 Basic Techniques

There are so many problems, distortions and pitfalls that are associated with seismic
sections that it is essential to have a geological model in mind when you start to correlate
the seismic reflections.

The fundamental assumption made is that there is a one to one correspondence with the
seismic reflection and the geological horizon you are trying to map. Accepting that, the
process is simple. Tie the ‘geological top’ you are interested in with available well data.
Where continuity can be recognised, the horizon is picked using a coloured pencil on a
paper section or digitally on a workstation (Figure 4.13). Where continuity fails, leave a
gap until the outcome of correlation tests allow you to carry on with the mapping of the
seismic reflection. The reflection can be ‘picked’ closing loops of an intersecting network
of lines.

___________________________________________________________________________________ 139
Ted Cox – April 2002 Section 4: Interpretation Page
Introduction to Exploration Seismic
___________________________________________________________________________________

Figure 4.13 – Picking and Tying a Loop from Well Correlation.

When planning regional interpretations it is as well to start your interpretation in the centre
of the basin, where the geological units are likely to be thickest. It is mostly a matter of
iterative geological model testing to match these basinal reflectors with the often eroded or
reduced interval sections penetrated by the wells drilled on the structural highs.

4.3.2 Seismic Reflectors

Reflections are due to changes in acoustic impedance, the product of velocity and density.
The implication is that reflections are lithostratigraphic in nature and can transgress
geological time.

The seismic stratigrapher however holds the view that seismic reflections are generated by
stratal surfaces that are chronstratigraphic. Over the scale of the basin, this is the case.
There is a high degree of correlation implicit in identifying a stratal reflection since litho-
facies change with distance and hence a reflection will juxtapose a range of lithofacies with
their own impedance contrasts. The amplitude and frequency content of the reflection will
change along the event. Thus, stratal reflections may be thought of as correlations of
shorter reflection segments.

___________________________________________________________________________________ 140
Ted Cox – April 2002 Section 4: Interpretation Page
Introduction to Exploration Seismic
___________________________________________________________________________________

___________________________________________________________________________________ 141
Ted Cox – April 2002 Section 4: Interpretation Page
Introduction to Exploration Seismic
___________________________________________________________________________________

4.4 INTERPRETATION OF 2D SEISMIC DATA

4.4.1 Reading the Seismic Label

There are numerous advantages to seismic interpretation on a workstation but there is one
unfortunate disadvantage. Where the workstation geophysicist will switch on and
immediately start picking without knowing how one section might differ from another, the
first point of reference for the worker with paper sections is the Side Label (Figure 4.28).

Figure 4.28 – Seismic Paper Section Side Label.

___________________________________________________________________________________ 154
Ted Cox – April 2002 Section 4: Interpretation Page
Introduction to Exploration Seismic
___________________________________________________________________________________

The side label gives a lot of information that is significant for the interpreter. Some
examples are

Acquisition Parameters

• The age of the survey.

• The fold of coverage.
• The type of source; time delays associated with air gun delays.
• The navigation system; was it reliable?
• The polarity.

Processing Parameters
are most usefully split into:
• Preprocessing (pre stack). Getting the most noise-free data possible before
they are used to estimate velocities and then stacked.
• Stack.
• Post stack processing and display.

Pre-Stack Processing:
• Observe what static shifts have been applied.
• Trace summing might change trace spacing and therefore impact the
horizontal scale.

Stack:
• What was the frequency of velocity analyses, was the coverage detailed or
regional?

Post-Stack Processing:
• What was the migration algorithm and how were the velocities handled?
• What is the frequency content of the data? What is the broad resolution
limit of the data?

The display scale gives the approximate vertical exaggeration and workstation sections
should not exceed this dramatically.

Shotpoint Reference - Where is the shotpoint on the section referenced? It is often

obviously located at the source position, but it could alternatively be located at the CDP
position or the antennae position.

It is always well worth while being aware of the acquisition and processing parameters
before you start interpreting seismic sections, especially when dealing with surveys from
several vintages.

___________________________________________________________________________________ 155
Ted Cox – April 2002 Section 4: Interpretation Page
Introduction to Exploration Seismic
___________________________________________________________________________________

4.4.2 Correlation with Geology

Before we actually discuss interpreting 2D seismic sections it is best to go briefly through

the main methods used for tying the geological tops measured in depth to the two-way
time scale of our seismic sections.

We know that the amplitude of a reflection is the product of the Velocity and Density of
the reflecting strata. As these are measured by electrical logging methods it is a fairly
simple task to calculate the sort of section we might expect from the geology encountered
in the well.

4.4.3 Well Log Data

Measurements of velocity are taken in a well by lowering a ‘sonic tool’ down the well
before the casing is set. Figure 4.29 shows the basic elements of the sonic tool.

Figure 4.29 – The Sonic Tool (Sonde).

The velocity of sound propagation is higher in the rocks than in the drilling mud so the first
arrivals at the receivers will have travelled through the rocks.

Calculating the mean travel time between R2 and R4 and R4 and R1 and R3 reduces the
error due to any diameter changes in the hole and tilting of the sonic tool.

The sonic log gives the time ∆t for the compression wave to pass through one foot of the
formation as a function of the depth, usually quoted as measured depth. The units are
microseconds per foot (µs/ft) so the instantaneous velocity at a given point on the log is
given by
V = 106/ ∆t

Figure 4.30 shows a typical sonic log.

The total travel times are integrated automatically and displayed on the left-hand axis of
the log. The small ticks represent 1 millisecond (One Way Time) with a longer tick mark
every 10 milliseconds.

___________________________________________________________________________________ 156
Ted Cox – April 2002 Section 4: Interpretation Page
Introduction to Exploration Seismic
___________________________________________________________________________________

Figure 4.30 – A Section of a Calibrated Sonic Log.

Sources of errors:

• Large cavities in the well can cause the measured transit time to be that of sound in
the drilling fluid.
• In the immediate neighbourhood of the hole, the drilling fluid usually invades the
formations particularly when they are porous. The transit times can be affected.
• ‘Cycle Skipping’. If the first energy to arrive at the nearest receiver is highly
attenuated by the time it reaches the farthest receiver, the first arrival will be missed
and a later arrival with higher energy recorded instead. The transit time measured
is consequently too long. This sort of attenuation of the signal usually occurs in
unconsolidated or highly fractured formations. It can also be associated with the
present of hydrocarbons.
• Peaks of small ∆t. If the signal is recorded by the farthest receiver but not by the
first, which records a later arrival, the result is a peak of too high velocity.

___________________________________________________________________________________ 157
Ted Cox – April 2002 Section 4: Interpretation Page
Introduction to Exploration Seismic
___________________________________________________________________________________

4.4.4 Check Shot Surveys

The continuous recording of sonic logs rarely starts until the first casing has been set and
regrettably is more often only acquired towards the zone of interest. To tie the sonic log to
the surface we need to know the average velocity down to the start of the log. To correct
for the errors in the log itself we also need to have independent measurements of travel
time to known depths. This is achieved with a Check Shot Survey. Figure 4.31

Figure 4.31 – Principle of Measurement by Check

Shot
The measured times are corrected geometrically to the vertical and the time depth pairs
form a checkshot survey. Caution: The depths quoted with the checkshots are often
relative to seismic datum as opposed to measured drill depths.

Having fixed the top of the sonic log in time and depth relative to the datum the differences
between the total travel time of the log to the other checkshot points forms a drift curve.
This drift curve is used to calibrate the sonic log.

4.4.5 Reflection Coefficient Log

In order to complete the correlation of well data with seismic, we need to consider the rock
densities. The other information needed therefore is the Density Log.

With sonic and density logs corrected for poor data records, and aligned in depth it is a
matter of sampling them at discreet points (usually determined in time for example, 2 or 4
ms) to determine the acoustic impedance log.

Acoustic Impedance = velocity (v) x density (ρ)

___________________________________________________________________________________ 158
Ted Cox – April 2002 Section 4: Interpretation Page
Introduction to Exploration Seismic
___________________________________________________________________________________

4.4.6 Synthetic Seismograms

This log is then convolved with a suitable seismic wavelet to produce a synthetic
seismogram.

Figure 4.32, from LandMark's Syn Tool manual, shows the basic elements of the synthetic
seismogram display.

Figure 4.32 – Basic Elements of the Synthetic Seismogram Display.

___________________________________________________________________________________ 159
Ted Cox – April 2002 Section 4: Interpretation Page
Introduction to Exploration Seismic
___________________________________________________________________________________

4.4.7 VSP Surveys

In general seismic reflection surveys have both their source and receivers at the surface.
Seismic waves have to travel through the near-surface weathered zone twice, which leads
to an attenuation of the high frequencies. Thus the advantage of a Vertical Seismic
Profile (VSP), where usually the receiver, but sometimes the source, is placed in the well,
is that its vertical resolving power is greater than that of conventional seismic. The
limiting resonant frequency is of the order of 150 Hz for vertical geophones. However, the
frequency content of the received signal is limited by the coupling of the geophone tool to
the walls of the well.

The possible types of borehole surveys are summarised in Figure 4.33.

1. Surface to surface seismic raypath

2. VSP surface shots with well geophones.
3. Well shots with surface geophones.
4. Well to well surveying.

Figure 4.33 – Different Types of Borehole Survey.

___________________________________________________________________________________ 160
Ted Cox – April 2002 Section 4: Interpretation Page
Introduction to Exploration Seismic
___________________________________________________________________________________

Figure 4.34 shows a final, processed VSP.

Figure 4.34 – A Final Processed VSP.

Warning: it is common practice to use the closely spaced Time-Depth recordings from a
VSP survey as checkshots for a sonic log calibration prior to synthetic seismogram
generation. Slight errors in the closely spaced recordings can cause serious distortions in
the drift curve and lead to wild distortions in the calibrated time-depth relationships used to
tie the synthetic seismogram with the seismic section.

___________________________________________________________________________________ 161
Ted Cox – April 2002 Section 4: Interpretation Page
Introduction to Exploration Seismic
___________________________________________________________________________________

4.4.8 Where to Pick the Event

From Alistair Brown (1998)

Firstly we have to consider whether the interpretation is going to be used for structural
mapping or for stratigraphic work.

For accurate structural work a synthetic seismogram produced to the same phase and
frequency content of the seismic data will show us where the geological top lies on the
seismic waveform. This can be picked consistently to give a reliable structural map. But
in the case of an event picked on the flank of the waveform the amplitude extraction will
not be optimal and will probably be unusable.

We have seen that all the reflecting interfaces within the interval of ¼ of the seismic
wavelength will contribute to the seismic reflection. Therefore, having identified the
relevant reflection and assuming our wavelets have zero phase, we can pick the crests.
(Figure 4.35)

Figure 4.35 – Aspects of Picking the Wavelet.

___________________________________________________________________________________ 162
Ted Cox – April 2002 Section 4: Interpretation Page
Introduction to Exploration Seismic
___________________________________________________________________________________

Unless we are dealing with an ideal situation, rarely found in practice, our reflections of
interest will be contaminated by adjacent acoustic contrasts.

For detailed reservoir work we should try to identify the top and base reflectors. The top
and base amplitudes both contain information about the reservoir so compositing these
amplitudes, by simple arithmetical subtraction, will emphasise the effect of the reservoir
and de-emphasise the effect of the surroundings beds.

4.4.9 Looping Data, Common Problems

As interpretation progresses, each intersection has to be tied. It is important to do this

systematically by working outwards from a well tie, if one is available. If you are working
on paper sections, they must be folded vertically at the line or intersection and the mapped
seismic events and fault planes tied on each section. When all the intersections have been
tied without error you are ready to time your sections and map the horizons.

Section 4.4.6 gives a range of reasons why sections mistie and it is recommended that you
don’t attempt to follow a particular horizon by ‘dead reckoning’, i.e., trying to follow a
continuous amplitude line-up as far as it will go from each intersection.

4.4.10 Picking Procedure

Every interpreter will develop their own preferred method of working, but for novices,
some advice is offered here.

Start by applying your Seismic Stratigraphic techniques to get a basic understanding of the
geological history of the basin. Before committing colour to paper/workstation, lay out as
many parallel dip-lines as possible. Try to:

• identify the key erosive unconformities

• correlate the major faults on your base map, as well as the key geological contact
points such as onlaps, subcrops and fold axes.
• work through the dip lines without worrying too much about what is happening in
the strike direction. (We know now that migrated strike lines could have many
structural artifacts on them.)

You should have a basic understanding of the structural framework of the area you are
working in before attempting to tie the loops in detail.

It would be almost heretical nowadays to suggest that seismic interpretation can be more
effectively achieved on paper sections rather than on a seismic workstation. For 2D data
however, this can often be the case. While the workstation project once interpreted, is
already digital and therefore ready for the mapping phase, the slow speed of the
workstation in retrieving large amounts of data can restrict the steady flow of ideas.
Handling complex intersections can be easier on paper.

___________________________________________________________________________________ 163
Ted Cox – April 2002 Section 4: Interpretation Page
Introduction to Exploration Seismic
___________________________________________________________________________________

Differences in the amplitude ranges of successive vintages of 2D data make it very difficult
to use mixed data for seismic attribute work. This is the realm of the 3D survey and the
seismic workstation. Our 2D data are mainly for structural interpretation, and may not
always benefit from the workstation approach.

4.4.11 Misties

Unfortunately most misties are of our own making. When our interpretation of a seismic
event doesn’t agree around the loops, there is only one way to handle this sort of problem
and that is to go back to correct the mismatch.

As if the problem of picking consistent geological markers isn’t enough, there are some
external reasons why our sections won’t tie. Most of them are specifically related to land
data rather than marine.

a) Choice of recording datum

Different surveys may have been corrected to different reference datums. It is
therefore imperative to check the section labels to ensure this is not the case.

b) Method of static correction determination

Recording datum and static correction method are related. The datum is chosen to be a
flat surface below the weathered zone and which minimises the static corrections to be
determined. In areas where the topography varies widely different datums can be
chosen. Intersecting sections with different datums will not tie but can be unified by
the application of a bulk shift.

c) Recording system time delays or phase delays

Different recording systems will apply different time and phase shifts. The most
important delay is that associated with the firing system. If they haven’t been properly
accounted for in processing the sections will mis-tie with other surveys.

d) Recording system polarity

A change in polarity will make a strong peak on one section plot as a trough on
another. With a very easily identifiable reflector on your section this might be
detectable. In most cases it is not obvious and leads to mis-correlation at intersections.

e) Processing route
Small time shifts due to different processing software are usually within the error limits
of the seismic method. Some subtle mis-ties can occur where near traces are muted out
to help with multiple attenuation. The timing of an event is then vulnerable to mis-
picked stacking velocities.

f) Display polarity
As d) – Always check the section label!

g) Navigation errors

___________________________________________________________________________________ 164
Ted Cox – April 2002 Section 4: Interpretation Page
Introduction to Exploration Seismic
___________________________________________________________________________________

For 2D data and some early 3D data these are quite common. It is not uncommon for
2D data to be loaded onto a workstation with the wrong trace data attached to
navigation data.

h) Out of plane 2D migration

Unless all the dips are truly in the plane of the section we can expect migration to mis-
position events. We can also expect true strike lines with very little dip to mis-tie with
migrated dip lines. This is most severe when the geological dip is greater than 10 o .

___________________________________________________________________________________ 165
Ted Cox – April 2002 Section 4: Interpretation Page
Introduction to Exploration Seismic
___________________________________________________________________________________

4.5 INTERPRETATION OF 3D SEISMIC DATA

4.5.1 Overview

The basic principles for interpreting 3D data are the same as for 2D. As seismic
interpreters we are interested in unravelling the geological history from the inter-
relationship of reflection events and reflection packages. We are also trying to interpret the
complex changes in amplitudes of our key reflections to see if those changes have some
physical cause that might tell us something about the geological properties of the units
between our seismic events. What the 3D data cube gives us are exciting possibilities of
ways to go about our interpretations.

What we must always keep in mind when sat at our workstations is that we are trying to
interpret seismic reflections and not playing a video game!

4.5.2 General Approach

This will vary with whatever workstation you are using, but some basic and innovative
methods of interpretation are offered by most workstations:

• Ease of line selection from map (point and click)

• Rapid movement through 3D volume
• Easy scale and clip adjustment of section display
• Auto-tracking of reflection events along line
• Correlation polygon (window) for picking across faults
• Horizon flattening to help correlation and structural evaluation
• Time slices and chair displays
• Seismic attribute calculations and extraction
• Instantaneous phase section
• Auto-tracking (Zap!) across data volume

Figures 4.36 to 4.40 illustrate some views from the LandMark SeisWorks workstation.

Figure 4.36 – LandMark SeisWorks Dual Screen workstation allowing

seismic lines to be selected from base map on second screen.

___________________________________________________________________________________ 166
Ted Cox – April 2002 Section 4: Interpretation Page
Introduction to Exploration Seismic
___________________________________________________________________________________

Figure 4.37 – LandMark SeisWorks: Comparison of Automatic Tracking

Procedures.

___________________________________________________________________________________ 167
Ted Cox – April 2002 Section 4: Interpretation Page
Introduction to Exploration Seismic
___________________________________________________________________________________

Figure 4.38 – LandMark SeisWorks Screen showing Loop of Seismic Lines

Figure 4.39 – LandMark SeisWorks: Fault Interpretation by Triangulation.

___________________________________________________________________________________ 168
Ted Cox – April 2002 Section 4: Interpretation Page
Introduction to Exploration Seismic
___________________________________________________________________________________

Figure 4.40 – LandMark SeisWorks: Horizon Flattening to check Auto

Tracking.

Figure 4.41 – LandMark SeisWorks: Fault Picking on Time Slice Display

___________________________________________________________________________________ 169
Ted Cox – April 2002 Section 4: Interpretation Page
Introduction to Exploration Seismic
___________________________________________________________________________________

4.5.3 A Recommended 3D Interpretation Methodology

Every interpreter has their own preferences for going about an interpretation. Each
interpretation software system will work slightly differently and the methodology will have
to be varied slightly to accommodate the software. The following is some gentle advice
that is based on experience.

Give horizons names that can be readily understood by everybody. Make use of the full
number of character spaces available. Most software has the ability to order horizons
alphanumerically so use that to have the horizons in stratigraphic and workflow order. It
will really help colleagues to understand your project’s structure easily. Scratch (or
temporary) horizons are always needed for quick-look tests, intermediate calculations, etc.
So, rather than having horizons called ‘scratch’ or ‘temp’ languishing somewhere in a long
list of events, why not give them appropriate names that will allow them to split your work
flow into understandable chunks? An example is produced below for the LandMark
workstation.

Sfa----Stuart_Fletcher---Cormorant-East-----Dec1999-----Purple--+-zero
Sfa1_Base_Cretaceous_Unconformity--WELL-TIE--------Blue----peak
Sfa2_Top_Brent-------------------------------WELL-TIE--------Red-----trough
Sfa3_Intra_Triassic_Unconformity-------WELL-TIE---------Green-+/-zero
Sfb0---------------------------------------------------------------------Purple-peak
Sfb1_Base_Cretaceous_Unconformity--SEED--------------Blue---peak
Sfb2_Top_Brent-------------------------------SEED--------------Red----trough
Sfb3_Intra_Triassic_Unconformity-------SEED---------------Green-+/-zero
Sfc0---------------------------------------------------------------------Purple-trough
Sfc1_Base_Cretaceous_Unconformity--ZAP-----------------LBlue-peak
Sfc2_Top_Brent-------------------------------ZAP-----------------LBlue---trough
Sfc3_Intra_Triassic_Unconformity--------ZAP-----------------LBlue--+/-zero
Sfd0----------------------------------------------------------------------Purple--trough
Sfd1_X_Unconformity-----------------------MAP-12Dec99----Blue---peak
Sfd2_Top_Brent------------------------------MAP-15Jan00-----Red-----trough
Sfd3_Intra_Triassic_Unconformity-- -- -MAP-----------------Green---+/-zero

The interpretation can proceed as follows:

a) Tie the wells with a series of random lines.

b) Develop a regular seed grid.

The line spacing of the seed grid will depend on the complexity of the area. Some
horizons will only need a sparse grid while others will need to have a detailed,
finely spaced grid. A common starting interval might be a 128 x 128 line by trace
grid. 128 may seem a strange number, but the rationale is that it can be halved until
you reach the number 1. Having a well-organised seed dataset will help you
identify and retrieve interpreted lines from a map much more easily than scattered
lines.

___________________________________________________________________________________ 170
Ted Cox – April 2002 Section 4: Interpretation Page
Introduction to Exploration Seismic
___________________________________________________________________________________

c) Once the broadest seed has been interpreted, the volume tracker should be used
where at all possible to do the hard work. The volume tracker in LandMark is
called ZAP (in Charisma it’s called ASAP). The early ZAP runs help to determine
the structural trends and form the basis for revising the seed.

d) Resample the ZAP output to extract a more detailed seed, for example 64 lines by
128 traces. Those lines not yet confirmed by manual interpretation can be checked
and the basic fault pattern entered.

The stages c) and d) are repeated until a fit-for-purpose interpretation has been achieved.

Volume trackers can finish off the good signal/noise booming horizons very quickly. But
don’t been fooled by the pace of early progress. There is some truth in the axiom that 80%
of the interpretation can be completed in 20% of the time!

There will always be holes of very poor S/N through which a volume tracker will not be
able to correlate. Don’t worry too much about these. They can always be filled in by
means of the interpolation provided by a mapping package.

The mapping package is an essential 3D tool. Interpretation can be exported, gridded

and contoured and re-imported to see how the interpolated values sit on the seismic data.
Always check a ‘final’ interpretation in this way.

Figure 4.42 gives a flow chart summary of the above recommendations.

___________________________________________________________________________________ 171
Ted Cox – April 2002 Section 4: Interpretation Page
Introduction to Exploration Seismic
___________________________________________________________________________________

Figure 4.42 – LandMark SeisWorks 3D Interpretation Workflow Summary.

___________________________________________________________________________________ 172
Ted Cox – April 2002 Section 4: Interpretation Page
Introduction to Exploration Seismic
___________________________________________________________________________________

4.5.4 Attribute Displays

Within seismic processing it has long been understood that by displaying another Seismic
Attribute, that is, other than amplitude, we can often obtain a different perspective which
may aid our interpretation, e.g., correlation across faults and stratigraphic changes. This
process is called Complex Trace Analysis.

With the 3D data and the power of the seismic workstation at hand, it is easy for the
interpreter to calculate another seismic attribute on a line-by-line basis. Typical attributes
on offer for calculation are Quadrature Trace, Reflection Strength, Polarity,
Instantaneous Phase and Instantaneous Frequency. These are calculated using the
Hilbert Transform discussed in Section 3.4.5. For a fuller account of what these different
attributes can offer the interpreter, White (1991) is recommended.

The Quadrature Trace attribute provides a 90 o phase shift to the data such that the
centres of the peak or trough are displayed as zero crossings. This is often combined with
a 6dB/octave high cut filter to produce a Seismic Approximate Impedance Log (SAIL).
The SAIL is a very approximate seismic inversion to acoustic impedance. It is only
approximate because the seismic data does not retain the very low frequency content that
you see as huge log breaks on your sonic logs and which are clearly important in terms of
modelling the acoustic impedance log.

Reflection Strength, as the name implies, shows variations of reflection amplitude in

decibel steps of graded colour (red is high, blue and white are low). The positions of the
colours will be independent of phase. Bright Spots will be accentuated, but caution should
be exercised since not all amplitude anomalies are due to the presence of gas.

Polarity displays show events according to the polarity of the strongest part of the
reflection envelope, with colour intensity related to strength. Positive reflections are
coloured red, and negative blue. Polarity reversals along a reflector, such as would occur in
the presence of gas, will be evident.

Instantaneous Phase splits each oscillation of the waveform into discrete colour bands;
different systems offer different standard schemes, most of them are variable by the user.
The method is independent of amplitude effects and therefore helps to enhance the
continuity of weak reflectors. Reflection alignments are more evident in areas of poor
signal to noise, and fault planes, stratigraphic anomalies and gas-liquid interfaces may be
interpreted with more confidence.

Instantaneous Frequency is an accurate representation of frequency content on a sampl-

by-sample basis. Colours typically range from reds and oranges for low frequencies
through the spectrum to blues and purples for high frequencies. Frequency sensitive
stratigraphic changes are readily apparent. Some software packages offer a further
Average Frequency display which smoothes the instantaneous frequencies with an
amplitude-weighted response to portray the characteristic reflection frequency.

Some examples are shown in Plates 4A – G.

___________________________________________________________________________________ 173
Ted Cox – April 2002 Section 4: Interpretation Page
Introduction to Exploration Seismic
___________________________________________________________________________________

4.6 CONTOURING AND MAPPING

4.6.1 Structure Maps

Contour maps of horizon structure can be in TWT or depth. It is best to make clear what
the units are on the map label, and also to state the map scale, the contour interval, the
horizon name (and colour used), the date produced (to distinguish multiple versions) and
last, but not least, the interpreter!

4.6.2 Isochron Maps

Maps of equal time difference between two picked reflectors are referred to under many
names:
Isochron, Isochore and Isochronochore.

Labelling of these should also be clear and comprehensive.

4.6.3 Hand Timing Seismic Sections

The hand timing of sections has been discussed in the section on 2D interpretation.
(Section 4.3.1).

4.6.4 Manual Contouring

Contours will be drawn by hand, usually in pencil first, fitting the contours between posted
map values. Smoothing of the lines will be an iterative process.

The contours should be :

• Close enough to define details,

• Sparse enough to maintain simplicity,
• Thicker line every 5th.
• Clearly labelled,
• Marked outside with ‘ticks’ on highs,
• Marked inside with ‘ticks’ in lows.

A convention and symbols for labelling your structure contour map and isochore map is
given for guidance (Figure 4.43).

___________________________________________________________________________________ 175
Ted Cox – April 2002 Section 4: Interpretation Page
Introduction to Exploration Seismic
___________________________________________________________________________________

Figure 4.43 – Structural Mapping Symbols

___________________________________________________________________________________ 176
Ted Cox – April 2002 Section 4: Interpretation Page
Introduction to Exploration Seismic
___________________________________________________________________________________

The procedures for producing a contour map are listed below :

• Begin by finding the highest and lowest posted values,

• Decide on suitable contour interval.
• Starting with highest values
• Work outwards joining like values.
• Revise first draft by smoothing.
• Check all posted values are honoured.
• Contour steps across a fault should match the throw.

Nigel Anstey (1977) compiled a set of Contouring Rules:

• Recognize structural trends, regional dips and reversals, subcrops & onlaps.
• Look for geological anomalies (folds, faults, reefs, channels, etc.)
• Work from simple, good data areas towards sparse data and complications.
• Contours should be approximately equidistant.
• Reduce the contour interval in steeply dipping areas or areas of complex
geology.

Be suspicious of these problems:

• A closed high within a low

• Closed lows on top of highs
• A low trend running towards a high
• Like contours running parallel for some distance
• Contours that follow the seismic grid – these could be due to mis-ties.

Finally, always check your contours back against the sections. You don’t want to give
someone else the opportunity to find any errors before you!

4.6.5 Computer Contouring

A review of computer contouring procedures is included here:

• Data Input.
• Surface, Fault Polygons, Map.
• Adjust gridding and contouring options.
• Determine the closing contour.
• Test the effect of different grid spacings.
• Produce structural trend grid.
• Define fault polygons.
• Grid and contour with fault polygons.
• Output illustrative cross sections.
• Output a faulted contour map.
• Display shaded-relief contoured surface with ortho-contours.
(Good management presentation tool)

___________________________________________________________________________________ 177
Ted Cox – April 2002 Section 4: Interpretation Page
Introduction to Exploration Seismic
___________________________________________________________________________________

4.7 DEPTH CONVERSION

4.7.1 The Problem

Since our seismic data are events referenced in two way time there is obviously going to be
a distortion in portraying geological structure whenever the velocity within any layer
changes laterally. In severe tectonic areas, or areas where this lateral change in velocity is
large, e.g., in areas of salt domes, the distorting effect of the lateral velocity changes is so
great that the seismic traces have to be migrated into their correct position before being
stacked. If the normal time-migrated data can be interpreted but their lateral mis-
positioning is a concern, image ray tracing can indicate the amount of lateral displacement
that can be expected due to the rays bending at the velocity interface.

As computing costs come down it is becoming more common to have Pre Stack Depth
Migrated (PSDM) sections and in these cases the velocity studies have to be carried out
before processing begins.

However, at present the vast majority of depth conversion carried out is a simple vertical
depth conversion of our interpreted times from the seismic section or computer map
(Figure 4.44). It is sometimes referred to as ‘vertical stretching and squeezing’, to convey
the point that ray-trace bending has not been taken into account.

Depth Conversion is important to the geochemist trying to establish the source kitchens
and burial history for the generations of hydrocarbons, and to the production geologist
interested in porosity mapping. Structures can be made or lost by Depth Conversion so it
is of paramount importance.

Figure 4.44 – Time – Depth Conversion Methods.

___________________________________________________________________________________ 178
Ted Cox – April 2002 Section 4: Interpretation Page
Introduction to Exploration Seismic
___________________________________________________________________________________

4.7.2 Sources of Velocity Information

Velocity information for depth conversion (Figure 4.45) must be gathered from all
available sources:
• The picked well tops and
• Seismic times picked from the section, (as long as the correct seismic
horizon is tied to the well top)
• Well times derived from well surveys such as checkshot survey, integrated
sonic log times or VSP survey

Figure 4.45 – Determination of Average and Interval Velocity.

Normal Move-Out velocities or ‘stacking velocities’ (it would be preferable to call them
stacking functions) can also provide valuable velocity information. These velocities are
printed as VT pairs at the top of paper sections, the ‘stacking velocity boxes’ and are
supplied digitally with the sections from the processing centre (Figure 4.46).

The RMS velocities are rarely picked at exactly the same time as our interpreted seismic
horizons so the stacking velocity function has to be interpolated at the appropriate times
and a new interval velocity calculated via Dix's Formula (see Section 3.2.15). This
velocity is referred to as being ‘horizon keyed’.

Figure 4.46 – Stacking Velocity Boxes.

___________________________________________________________________________________ 179
Ted Cox – April 2002 Section 4: Interpretation Page
Introduction to Exploration Seismic
___________________________________________________________________________________

Dix's formula only strictly holds for flat layers. Also the stacking RMS velocity values are
subject to large variations, so in practice the interval velocities derived in this way have to
be smoothed in some way.

A comparison of the velocity data available to the interpreter is given in Figure 4.47 –
Seismic Velocities.

Figure 4.47 – Comparison of Seismic Velocities.

___________________________________________________________________________________ 180
Ted Cox – April 2002 Section 4: Interpretation Page
Introduction to Exploration Seismic
___________________________________________________________________________________

4.7.3 Vertical Depth Conversion Methods

Three different approaches can be used to convert time to depth, each based on different
velocity information.

4.7.4 Average Velocity

• The model is provided as a horizon map or grid.

• This is a simple form of depth conversion entailing only one map per
horizon.
• It will be very useful in areas of simple structure and geology.
• It is difficult to model the variation of average velocity over an area because
it encompasses the variations of all the shallower layers.

4.7.5 Interval Velocity

• A model for variations in each interval velocity is usually given as a

map/grid. Constant interval velocity is also possible.
• The method can produce better accuracy in areas of complex geology and
structure.
• The depth conversion is done in a layer-cake fashion, working downwards.
Hence, errors arising from the shallow intervals will propagate down to
deeper levels.
• This method demands more effort than average velocity mapping as it
involves producing one model per interval down to target horizon.

4.7.6 Velocity Function

• Models are based on a simple mathematical description of the variations of

instantaneous velocity with depth or interval velocity with depth.
• Constant parameter values often provide an accurate tie but in complex
areas each parameter can be varied as a map.
• The method can produce accurate depth estimates and be simpler to produce
than average velocity maps.
• This approach can be single or multi-layer. For multi-layer approach, the
errors from the shallower levels can propagate down to deeper units.

Some velocity function methods are described below:

Vo,K Method

In its basic form a Vo, K representation (V=Vo+KZ) assumes that the instantaneous
velocity varies linearly with depth (Figure 4.48). However in T-D conversion we are not
so much interested in the behaviour of the instantaneous velocity but in the behaviour of
the interval (layer average) velocity.

___________________________________________________________________________________ 181
Ted Cox – April 2002 Section 4: Interpretation Page
Introduction to Exploration Seismic
___________________________________________________________________________________

Figure 4.48 – The Linear Instantaneous Velocity Function

In cases where the interval velocity increases with layer depth, and the depth of the layer
changes over the area of interest, a Vo, K representation is often appropriate.

This representation cannot be used with constant parameters in areas where there has been
differential uplift. In this sort of area a Vo in map form is often appropriate.

Before invoking variable parameters, however, it is worth considering the non-uniqueness

of the parameters you are estimating.

Al Chalabi (1997) has also shown that the Vo, K parameters have a non-uniqueness if you
accept a margin of error inherent in the measurements (Figure 4.49).

This non-uniqueness leads to a solution trough where other Vo, K combinations can tie the
well equally well. The overlap of several solution troughs could provide a Vo, K
combination that would tie all the wells without having to invoke a variable Vo or K
parameter (Figure 4.50).

___________________________________________________________________________________ 182
Ted Cox – April 2002 Section 4: Interpretation Page
Introduction to Exploration Seismic
___________________________________________________________________________________

Figure 4.49 – The Non-Uniqueness of Parameter Determination.

Figure 4.50 – The Overlap of Solution Troughs.

___________________________________________________________________________________ 183
Ted Cox – April 2002 Section 4: Interpretation Page
Introduction to Exploration Seismic
___________________________________________________________________________________

There are two types of Vo, K function which, though related, are different.

Vins = Vo + KZ describing the instantaneous velocity behaviour as a function of

depth. The function can be derived from one or more wells (sonic
data) (see Al Chalabi, 1997) and does generally not only describe
burial effects. Vertical lithological changes within the layer
influence this function strongly. Although the K in this function is
sometimes referred to as ‘local K’, a more appropriate name would
be ‘instantaneous K’.

The formula to use for a multi-layer linear instantaneous velocity

with depth function is:

Vint=Vo +KZ describing the interval or layer velocity as a function of depth. This
function can only be derived using data from several wells sampling
a range of depths. Only the burial effect is modelled by this function
and not the vertical lithological changes within the layer. The
parameter K in this formula is sometimes referred to as ‘burial K’ or
‘regional K’.

The formula to use for a multi-layer linear increase of interval

velocity with depth is given in two parts:

Having calculated Vint the depth is calculated as below:

In practice Vint is often smoothed and tied to the wells before the
final stage is carried out.

The formula for numerous relationships that can be applied in a multi-layer depth
conversion is given by Al Chalabi (1997).

Functions derived from a minimum-error fit of well points will mistie the well points.
Error corrections can be calculated at the well points and mapped to provide an error
surface. This surface should match the causes of the errors wherever possible.

___________________________________________________________________________________ 184
Ted Cox – April 2002 Section 4: Interpretation Page
Introduction to Exploration Seismic
___________________________________________________________________________________

4.8 CONCLUSIONS

Having obtained as much information as possible from others working on the project area
and having considered all the seismic parameters and effects, the final depth maps should
be as close as it is possible to get to the actual geology. Only drilling will confirm the
interpretation.

A good interpreter should always be prepared to stand up and talk about his or her work.
Talk about it proudly and succinctly – busy managers don’t want to listen to waffle!

It is only by producing maps and talking about them that others can begin to appreciate the
work you have done. Don’t be afraid of criticism – every map you produce will just be
another step towards final understanding. It is part of the exploration process that as more
information becomes available, our understanding will change, and so will our maps.

Always be aware of the cost implications of what you are doing and be prepared to put in
that extra bit of effort when it’s needed. Happy hunting!

GOOD LUCK ! ! !

___________________________________________________________________________________ 185
Ted Cox – April 2002 Section 4: Interpretation Page
Introduction to Exploration Seismic
___________________________________________________________________________________

BIBLIOGRAPHY

SECTION 1 – BASIC THEORY

Britain’s Offshore Oil & Gas: BMNH/UKOOA: Fred Dunning et al (Revised Edition)
1997. ISBN 0 565 09027 5

SECTIONS 2 & 3 – ACQUISITION and PROCESSING

A number of books cover applied geophysics at an introductory level. Several of

these contain sections or chapters that are pertinent Seismic Acquisition and
Processing.

Al-Sadi, H.N. 1980. Seismic Exploration: Technique and processing, Birkhauser,

Basel. Chapter 6 is a terse but well-ordered treatment of the principal steps in
processing. Some of the mathematics may be daunting.

Anstey, N.A., 1982. Simple Seismics. Boston: IHRDC (copyright transferred to

Prentice-Hall, Inc., Englewood Cliffs, NJ)

Dobrin, M.B., 1976. Introduction to Geophysical Prospecting, McGraw-Hill, New

York. Chapter 6 deals with digital processing of seismic data. Various other
corrections are treated in Chapter 7.

Fitch, A.A. (ed), 1979. Developments in Geophysical Exploration Methods-1,

Applied Science Publishers, Ltd., London.

Robinson, E.A. and Treitel, S. 1980. Geophysical Signal Analysis: Englewood

Cliffs, NJ: Prentice-Hall, Inc.

Sheriff, R.E. and Geldart, L.P., 1983. Exploration Seismology, Volume 2: Data
Processing and Interpretation: Cambridge University Press.

Telford, W.M., Geldart, L.P., Sheriff, R.E. and Keys, D.A., 1976. Applied
Geophysics, Cambridge University Press, Cambridge. Chapter 4.7 deals with
data processing, but is restricted to those steps that involve filtering.

Waters, K.H., 1981. Reflection Seismology, New York: John Wiley and Sons.

Yilmaz, O., 1987. Seismic Data Processing. Tulsa, OK, Society of Exploration
Geophysicists.

Ted Cox – April 2002 Bibliography i

Introduction to Exploration Seismic
___________________________________________________________________________________

SECTION 4 – INTERPRETATION

4.1 Recommended Reading

‘An Introduction to Seismic Interpretation’, 2nd Edition (1984) written by McQuillin,

R., Bacon, M & Barclay, W. Published by Graham & Trotman.

‘Seismic Reflection Interpretation’, 2nd Edition (1984) written by Kleyn, A.H.

Published by Elsevier Applied Science Publishers.

Sheriff, R.E. (1977), Limitations on Resolution of Seismic Reflections and Geologic

Detail Derivable from them: In: Seismic Stratigraphy - Applications of Exploration,
AAPG Memoir 26, pp 3-14, C.E. Payton, Editor.

Widness, M.B. (1973), How Thin is a Bed? Geophysics, V.38, pp 1176-1180.

4.2 Recommended Reading

‘Velocities in Reflection Seismology’, (1985) written by Cordier, Jean-Pierre.

Published by D. Reidel Publishing Company.

‘Seismic Reflection Interpretation’, 2nd Edition (1984) written by Kleyn, A.H.

Published by Elsevier Applied Science Publishers.

Domenico, S.N. (1974): Effects of water saturation on seismic reflectivity of sand

reservoirs encased in shale. Geophysical Exploration Vol. 39.

Domenico, S.N. (1977): Elastic properties of unconsolidated porous sand reservoirs.

Geophysical Exploration Vol. 42.

4.3 Recommended Reading

Seismic Stratigraphy - Applications to Hydrocarbon Exploration. (1977) AM. Assoc.

Petrol. Geol. Mem. 26, pp 213-248.
Trusheim, F (1960) Mechanism of salt migration in Northern Germany. Bull. AAPG
XLIV, p 1519.

4.4 Recommended Reading

Brown, A.R. (1998), Picking Philosophy for 3D Stratigraphic Interpretation: The

Leading Edge Vol. 17, No. 9.
Mari, J.-L. and Coppens, F. (1991), Seismic Well Surveying: published by Editions
Technip.

Ted Cox – April 2002 Bibliography ii

Introduction to Exploration Seismic
___________________________________________________________________________________

4.5 Recommended Reading

Brown, A.R. (1988) Interpretation of three-dimensional seismic data. APG Memoir

42.

Budney, M. (1991) Seismic reservoir interpretation of deep Permian carbonates and

sandstones (NW German gas province). First Break Vol 9, p55.

White, R.E. (1991) Properties of instantaneous seismic attributes. The Leading Edge,
July.

4.6 Recommended Reading

‘Velocities in Reflection Seismology’ (1985) written by Cordier, Jean-Pierre.

Published by D. Reidel Publishing Company.

‘Seismic Reflection Interpretation’, 2nd Edition (1984) written by Kleyn, A.H.

Published by Elsevier Applied Science Publishers.

Al Chalabi, M.A. (1974) Analysis of stacking, RMS, average and interval velocity,
Geophysical Prospecting 1974. Vol 22, pp 458-475.

Al Chalabi, M.A. (1997) Time-Depth relationships for multilayer depth conversion,

Geophysical Prospecting 1997. Vol 45, pp 715-720.

Al Chalabi, M.A. (1997) Time-Depth relationships for multilayer depth conversion,

Geophysics 1997. Vol 62, No. 3, pp 970-979.

Duin, E.J.T. and Stavenga, T. (1999) Modelling Zechstein velocities in haokinetically

disturbed areas using a 3D seismic amplitude attribute.

Marsden, D. (1992) V0-K method of depth conversion, The Leading Edge, August.

Ted Cox – April 2002 Bibliography iii