Professional Documents
Culture Documents
Firdaus Sulaiman
Fisika-Universitas Haluoleo
2008
Geophysics 224 - Geophysical exploration techniques
Winter 2006
Instructor Professor Martyn Unsworth
Offices : Physics 545B and Physics 605
e-mail : unsworth@phys.UAlberta.ca
Tel : 492-3041
Course Schedule
Objectives of class
To learn the basic principles of three geophysical exploration techniques. This will
include both the physical basis of the method and data interpretation. Simple
mathematics will be used where needed, but the emphasis will be on understanding
concepts and methods. Where appropriate, PC based software will be used in the lectures
and the labs.
Course outline
A. Introduction
B. Gravity exploration ~ 2 weeks
C. Seismic exploration ~ 6 weeks
D. Magnetic exploration ~ 2 weeks
Office hours
If you have questions about class material or assignments, I will happy to arrange a
convenient time to meet with you. To do this, please see me after class, e-mail me or
drop by my office.
Assignments
There will be a total of four assignments in this class. These will be due at 5 pm on the
appropriate day, and should be handed in at my office (Physics 545B)
Recommended text
An introduction to geophysical exploration, by P. Keary, M. Brooks and I. Hill, 3rd
Edition, Blackwell, 2002. ISBN 0-632-04929-4
Course weighting
Labs 20%
Assignments 10%
Mid-term exam 25%
Final exam 45%
Policy about course outlines can be found in Section 23.4(2) of the University Calendar
(CGF 29 SEP 2003)
MJU 2006
Geophysics 224 - Geophysical exploration techniques
A: Introduction
Why?
● oil and gas exploration
● mineral exploration
● hydrogeology
● monitoring contamination and remediation
● tectonic studies
● earthquake hazard studies
●All geophysical methods remotely sense a material property of the Earth (e.g. seismic
velocity, rock density, electrical resistivity, magnetization etc)
●Geophysical and geological studies complement one another. Geologists are more
effective with a basic knowledge of what geophysics can and cannot resolve.
(Similarly, many geophysicists would benefit from a basic knowledge of geology).
●Geophysical imaging does not always give a unique answer! Additional information is
often needed to discriminate between possible solutions (e.g. other geophysical
surveys, knowledge of local geology, well log information in the study area).
●Please do not be intimidated by equations! Mathematics will be used in this class where
needed, and I hope that this will provide a review of math classes you have taken in the
recent past. I do not expect students to memorize equations. My expectation is that
students will be able to perform simple rearrangement of equations, and use a
calculator to evaluate an equation for a given set of values.
Paleo-meanders imaged in 3-D seismic data, Gulf of Thailand (AAPG memoir 42)
Plate tectonics in action. Seismic reflection and electromagnetic images of India underthusting the
southern edge of the Tibetan Plateau (Unsworth et al., 2005)
Unsworth, M.J. A.G. Jones, W. Wei, G Marquis, S. Gokarn, J. Spratt, Crustal rheology of the
Himalaya and Southern Tibet inferred from magnetotelluric data, 438, 78-81,
doi:10.1038/nature04154, Nature 2005.
B. GRAVITY EXPLORATION
Consider two point masses that are a distance r apart. Newton’s theory of gravitation
predicts that they will attract each other with a force F that is given by:
Gm1 m2
F=
r2
The quantity G is called the gravitational constant (or “big G”) but is actually a small
number. Newton deduced this equation from observing the motion of planets and moons
in the solar system. The units are as follows:
F Newton (N)
m kg
r metres (m)
2 -2
G = 6.67 x 10-11 N m kg
It can be shown that if one of the masses is finite in size (e.g. a planet or the sun), then F due to
will be the same as if all the mass were concentrated at the centre.
If m1 = m2 = 1,000,000 kg (1000 tonnes) then the variation of F as r increases will show the
“inverse square law”.
Consider the mass m2. Newton’s Third Law of Motion predicts that this mass will accelerate
with an acceleration a, where:
F = m2 a
F
a=
m2
Gm1 m2 1 Gm
a=g= 2
= 21
r m2 r
The Greek letter rho (ρ) is used to represent density. Rocks and minerals found on Earth have
densities that range from 1000-7000 kg m-3. Often densities are quoted in g cm-3.
Pure minerals can exhibit a high density since the atoms are closely packed together.
Sedimentary rocks generally have lower densities since the atoms are not as closely packed
together and pore space is filled with lower density materials.
The range of density values reflects the degree of weathering and the porosity. Why?
The composition is secondary, but explains why dolomite is more dense than limestone.
Most sedimentary rocks increase in density with depth, owing to increased compaction.
This reduces the pore space available for low density materials such as air, water and
hydrocarbons.
The density depends primarily on the rock composition. Mafic rocks are generally
more dense than felsic rocks owing to increased proportion of heavier elements
such as Fe and Mg.
The gravitational acceleration at the Earths surface is 9.8 ms-2. Subsurface variations in rock
density produce very small changes in this value, so it is more convenient to use a smaller unit.
Consider a spherical ore body with density ρ and radius a that is buried at a depth z below the
surface. The extra pull of gravity will be greatest at a point P directly above the ore body.
Remember that the pull of gravity (g) for a sphere is the same as if all the mass were
concentrated at the centre.
4 3
Total mass of ore body = volume x density = πa ρ
3
Excess mass , mE = mass of ore body – mass of rock that was already there
4 3 4
= π a ρ - πa 3 ρ 0
3 3
4
= πa 3 ( ρ − ρ 0 )
3
Above ore body, the change in gravity due to the ore body (gravity anomaly) is given by
GmE 4Gπa 3 ( ρ − ρ 0 )
Δg = =
z2 3z 2
a = 30 m
z = 40 m
ρ = 4000 kg m-3
ρ0 = 2000 kg m-3
MJU 2006
Geophysics 224
B2. Gravity anomalies of some simple shapes
Gravity measurements are made on a surface profile across a buried sphere. The sphere
has an excess mass MS and the centre is at a depth z. At a distance x, the vertical
component of g is given by
GM S z
gz = 3
(x2 + z 2 ) 2
Note that::
Question: What will the map look like for the buried sphere
When gravity measurements are made across a buried cylinder, it can be shown that the
variation in gz will be.
2Gπa 2 zΔρ
gz =
(x 2 + z 2 )
Note that :
● the maximum value of gz is located directly above the axis of the cylinder.
2Gπa 2 Δρ
gzmax =
z
● From the plots, we can see that this value is larger than gzmax for a sphere? Why?
Question: Compare the profiles across the sphere and a cylinder. Would this
information allow you to decide if the buried object was a sphere or a cylinder?
Question: If gz is measured on a grid of points, what will the resulting map look like?
Would this be a better way to distinguish between a sphere and cylinder?
Forward and inverse problems in geophysics
B2.1 and B2.2 illustrate the gravity anomaly that we would expect to observe above a
known geological target. This is called a forward problem in geophysics, and is a useful
exercise in understanding if measurements would be able to detect a particular structure.
However, we are usually more interested in solving the opposite problem. When gravity
data has been collected in a field survey, we want to find out the depth and size of the
target. This is called an inverse problem in geophysics.
Consider some gravity data collected on a profile crossing a spherical iron ore body.
● The depth of the sphere can be derived using the equation x½ = 0.766 z
z = ______metres
GM S
● To determine the excess mass, we use the equation gzmax =
z2
We know z and have measured gzmax so we need to rearrange this equation to find Ms
g zmax z 2
This gives Ms =
G
● Can the radius (a) and density contrast (Δρ) of the sphere be determined?
A layer has a infinite extent, thickness ∆z and a density ρ. The gravitational attraction of
this slab at P is:
g z = 2πGρΔz
Note that gz does not depend on the distance from the layer to the point P. Why?
What can we say about the gravitational acceleration (gz) of the two models?
This is an example of non-uniqueness in geophysics, and occurs when more than one
Earth model can explain the same set of geophysical data.
B2.4 General polygon – simple computer modelling
Simple density models such as a sphere and a cylinder can help us get an idea of what
gravity anomaly (gz) is produced by simple density models. However, to represent the
real Earth, more complicated model geometries are clearly needed. The development of
powerful, portable computers has made this possible.
In the lectures and labs we will use a very simple MATLAB program that computes gz
for a N-sided prism. This is called 2-D modelling, since the polygon is assumed to be
infinite in length in the strike direction (out of the plane of the page). Fully 3-D gravity
modelling is also widely used, but will not be discussed in detail in Geophysics 224.
Consider using the program to compute the gravity anomaly of a cylinder. The cylinder
can be approximated by an 8-sided polygon, as shown below. The dots show the exact
(known) answer for a cylinder.
An over simplified sedimentary “basin” is shown below sheet. Note that the “basin” is
less dense than the basement rocks, so the pull of gravity is slightly weaker over the
basin. The horizontal line shows the expected value of gz for an infinite layer with the
same thickness and density as the basin.
Now consider a similar basin, that is twice as deep and with a lower density contrast.
● The minimum value of gz is the same as for the first basin. Why?
● The edge effects extend further into the basin than in the first case.
In the lab we will use this program to analyse some real gravity data collected across a
sedimentary basin in Washington and an impact crater in Mexico.
MJU 2006
224B3 Other factors that cause changes in g and need to be corrected
Note that gravity exploration is different to seismic exploration in the following way:
● In a seismic survey, the travel time depends on just the velocity of the material on a path
between source and receiver.
● The acceleration of gravity at the Equator, g E = 978,033 mgal and at the poles g P = 983,219
mgal (Hammer, Geophysics, 8, 57, 1943). This difference is 5186 mgal, which is much
bigger than the anomalies we have considered, and needs to be accounted for in field
measurements.
R E = 6378 km and R P =6357 km. The ratio of flattening is approximately 1/298. Since a
point on the Equator is further from the centre of the Earth than the poles, gravity will be
slightly weaker at the Equator.
GM E
We previously showed that for a sphere g (r) =
r2
where the mass of the Earth, ME = 5.957 1024 kg.
At the North Pole, r = 6357 km and g P = 983,219 mgal.
If we move up 21 km to the radius of the equator, the decrease in gravity will be 6467 mgal
Thus g E = g P - 6467 mgal, which is too much to explain the observed difference between
the Equator and the Poles.
(B) Centrifugal forces vary with latitude
The rotation of the Earth also causes gravity to vary. Imagine you are standing at the North
Pole. The rotation of the Earth will not change g, all that will happen is that you rotate once a
day.
Now imagine you are at the Equator. If we could increase the rotation rate of the Earth
enough, you would be ultimately be thrown into space (i.e. become weightless). Thus rotation
makes gravity weaker at the equator.
With the 1 rotation per day, and RE = 6378 km, can show that g P = g E +3370 mgal
Question : How fast would the Earth need to rotate to throw you into space?
The change in shape from a sphere to an ellipsoid redistributes the Earth’s mass. Thus results
in more mass between points on the Equator and the centre of the Earth, than between the
poles and the centre of the Earth. This effect will make g E > g P and is analogous the Bouguer
correction we will discuss in section B3.2. Calculations show that g E ~ g P + 4800 mgal
where θ is the latitude in degrees. This equation is called the Geodetic Reference System for
1967. More recent revisions are essentially the same, but with more decimal places.
Example : In Edmonton θ = 53˚ 30’ 25” N and the GRS67 equations gives
g = 9.78031846 ( 1+0.003417902-0.000005395) m s −2
= 9.81369388 m s −2
The variation of g with latitude is important when a survey extends over a significant north-
south distance. Differentiating the GRS67 equation with respect to θ yields
dg
= 9.78031846 (0.0053024 x 2 sin θ cos θ - 0.0000058 x 4sin 2θ cos 2 θ)
dθ
All the these equations define the expected value of theoretical gravity (or normal gravity) at
latitude θ. Differences between this value and what is actually measured are anomalies that we
will analyse for information about subsurface density structure.
224 B3.2 How does elevation effect gravity measurements?
GM
g(r) =
r2
This means that as you move away from the centre of the Earth, the acceleration of gravity (g)
decreases. In Edmonton, g = 9.81 ms −2 and if you move up a distance, Δh, the acceleration of
gravity will decrease by
Δg = 3.086 Δh x 10 −6 m s −2
= 0.3086 Δh mgal
Gravity measurements are made at points A and B. The difference in elevation means that gB willB
When collecting gravity data, our real interest is to determine the density of the rocks below
ground. The change in elevation from ‘A’ to ‘B’ will thus contaminate the data. The Free Air
correction is a mathematical way of undoing the effect of elevation. It allows us to correct the
data collected at ‘B’ in order to make it equivalent to data collected at the same elevation as ‘A’.
In gravity surveys, we always define a reference level for the survey. Free Air corrections are
made relative to this level. In general, any reference level could be chosen, but sea level is
commonly chosen in coastal areas. In Alberta, the average level of the prairies would be a good
choice.
If a gravity measurement was made Δh above the reference level, we must add
CFA is called the Free Air correction for a given gravity measurement.
Similarly, if a gravity measurement was made Δh below the reference level, we must subtract
Question : to keep data accurate to 0.1 mgal, how accurately must we know the elevation?
(ii) The Bouguer correction
Unfortunately, this is not the end of story! Compare the gravity measurements at ‘A’ and ‘B’. At
point A, the gravity measurement is solely due to structure below the reference level (blue). At
‘B’ the gravity measurement is due to structure below the reference level, plus the gravitational
pull of the 100 metres of mountain (red). The net result is that gB > gA
B
From section B2.3 the magnitude of this extra gravitational attraction is approximately
gB - gA = 2πG ρ Δh
Thus to remove this effect we need to subtract CB = 2πG from the observed gravity
measurement at ‘B’. This is called the Bouguer correction and
CB = 0.00004193 ρ Δh mgal
Note that to apply the Bouguer correction we need to estimate ρ, the density that lies between ‘B’
and the reference level. Using the value ρ = 2670 kg m −3 this gives
CB = -0.1119 Δh mgal
This value represents an average density for crustal rocks. Other information (e.g borehole gravity
data or Nettleton’s method) may be used to give a better estimate of the density.
Pierre Bouguer
Summary
MJU 2006
224 B4. Instrumentation and field procedures
A range of devices can be used including reversible pendulum and free fall devices. They
are generally more expensive and slower to operate than a mass-on-a-spring gravimeter.
Since we have seen that it is the differences in gravity that is generally more important
than absolute values, we do not need absolute measurements of gravity at every survey
location. Often relative gravity measurements can be made over a survey area, and then
tied to an absolute value by using the relative gravimeter at a location that was
previously surveyed with an absolute gravimeter.
F = ks = mg
k(s+δs) = m(g+δg)
kδs
δg =
m
δs can be very small, so various engineering features are used to amplify the movement.
Typical accuracy is 0.01 mgal. What change in elevation produces 0.01 mgal?
Two widely used gravimeters are:
These were initially made with torsion balances. In a torsion balance, two masses are
suspended from a beam. The force required to twist the thread is determined by
calibration. The horizontal pull of gravity will tend to orient the beam so it points
towards or away from concentrations /deficits of mass. The Cavendish torsion
balance allowed measurements of G to be made. An alternative design was
developed by Baron von Eötvös in Hungary in the later 19th century and was used in
the first geophysical discovery of oil in the early years of the 20th century.
The torsion balance was quite effective at locating salt domes, owing to the lower
density of the salt compared to the surrounding sedimentary rocks. The torsion
balance required a light-proof and wind proof chamber to house the instrument.
Figures from Applied Geophysics, Eve and Keys, Cambridge University Press, 1928
4.1.4 Marine gravity surveys
The problems encountered in marine gravity surveys are even more severe in airborne
exploration (turbulence can produce an acceleration greater than g) and velocities are
higher. However, airborne measurements are excellent at covering large, roadless areas
very easily. In contrast to marine surveys, the acceleration of the aircraft can be
independently measured through accurate measurements of aircraft elevation as a
function of time. The acceleration of aircraft can be computed and then removed from
the acceleration recorded on board.
From, Sigmund Hammer, Airborne gravity is here, Geophysics, 1983. Note that:
● Upward continuation is a mathematical method that uses gravity data at one elevation
to compute the gravity data that would be measured at a higher elevation.
● Gravity measured at 300 m elevation in a helicopter agrees with the upward
continuation of the surface gravity measurements (land Bouguer).
●By moving further away from the target, small features such as the (higher density) cap
rock are not detected. This is a consequence of the inverse square law.
224 B.4.2 Instrument drift and tidal effects
Instrument drift
If repeated gravity measurements are made at the same location with a relative
gravimeter, the answer will change over time.
Occasionally, this is due to changes in the density of the Earth. However, it is usually due
to the spring in the gravimeter stretching over time. This effect can be removed by
establishing a base station that is visited repeatedly during the survey.
This problem arises because the gravimeter cannot be in two places at once. For example,
when the first measurement is made at Base 1, we would like to know the gravity
measurement that would be made at the same time at IGSN71 base. This cannot be done,
so linear time variation is assumed between the two measurements at the IGSN71 base.
This allows us to calculate the difference in gravity measurements (dial reading) between
Base 1 and IGSN71 base. Note that the Dial reading is a relative measure of g (mgal).
● Set up a series of base stations to remove instrument drift and tidal effects from data.
These should be visited several times a day.
● Details of a base station at the International Airport are listed on the following page and
can be found at:
http://www.geod.nrcan.gc.ca/database7/grav/sd/stage/95001993EN.html
Here the value is 981117.890 mgal Next time you have time to spare waiting for a
flight, see if you can find the benchmark.
Canadian Gravity Standardization Network (CGSN)
CLOSE-UP VIEW:
● Marine and airborne surveys. Ground is covered much more quickly than with land-
based methods, but in general observations made further away from potential targets.
●Satellite gravity: Cannot actually measure acceleration while the satellite is in free fall.
However, from detailed observations of the orbit, or by measuring the altitude of the
satellite above the Earth’s surface, subsurface density structure can be inferred.
−3
e.g. change in elevation = 100 m, density error 0.1 g cm gives an error of 0.42 mgal
● Once the Free Air and Bouguer corrections have been made, the Bouguer anomaly
should contain information about the subsurface density alone. The effect of latitude and
elevation should have been removed. A map of the Bouguer anomaly gives a good
impression of subsurface density. Low (negative) values of Bouguer anomaly indicate
lower density beneath the measurement point. High (positive) values of Bouguer
anomaly indicate higher density beneath the measurement point.
From M. Pilkington et al, Can. J. Earth Sciences, 37, 1453-1471, 2000. Note decrease
in Bouguer gravity anomaly towards the southwest. This is due to (a) thickening of the
Alberta Basin and (b) thicker crust beneath Rocky Mountains.
B.5.2 Regional trends and residuals
● Measurements of gravity at the Earth’s surface depend on Earth structures ranging from
spatial scales of 1m to 10,000 km.
● Deeper density structures produce gravity anomalies with a long spatial wavelength. For
a body such as a sphere, this would be expressed as a large value for the half-width (x½) .
● Similarly shallow density structures produce gravity anomalies with short spatial
wavelengths or small half-widths.
● Often the longer wavelength effects are called regional trends, while the shorter
wavelength features are called residuals or anomalies. The distinction between the two
is somewhat arbitrary.
● To emphasize shallow structure, the regional trend can be computed and then subtracted
from the measured Bouguer anomaly data. However care is needed not to eliminate
useful data during this process.
● Sometimes the regional trend can be approximated by a straight line. In other case it
should be approximated with a more complex mathematical form.
● Example 1 : This shows two density structures (a) higher density bedrock that dips to
the right and which produces a slow decrease in gB from left to right. (b) a shallow
buried cylinder. It is difficult to determine the half-width of the anomaly due to the
cylinder in the presence of the linear regional trend.
The dotted line denotes the regional trend, and is obtained by finding the best-fitting
straight line to the data. When the regional trend is subtracted from gB the anomaly of
B
● Example 2: This features the same dipping bedrock layer and a shallow, low density
river channel. Again, removal of the regional trend makes the effect of the river channel
much clearer.
● The above examples are for a single profile. However, regional trends are often removed
from Bouguer anomaly data defined on a grid (i.e. a Bouguer anomaly map).
● Must be careful not to remove useful data when removing regional trends from the data.
B.5.3 Half width techniques
● For a number of shapes the width of the anomaly can give an approximate depth to
the density anomaly. e.g. sphere : depth of centre = half-width
cylinder : depth of centre = 1.304 x half-width
● This allows the user to handle a more realistic density model. The data is usually fit with a trial
and error approach. This is illustrated with the simple MATLAB program used in Laboratory 2
to model a couple of real datasets.
●Taking the horizontal derivatives of the Bouguer anomaly emphasizes changes in the horizontal
gradient. This is effectively another way of removing (or suppressing) the regional trends in the
data.
● Consider the application to the basin example shown below. The gradient is very effective at
defining the edge of the basin. The unit E is the Eotvos. 1 E = 10-9 s-2
● In the example below, note that the second derivative has a different sign depending on
the geometry of the edge of the basin..
Gravity gradients : Alberta basement
The primary domains of basement geology in Alberta are shown on the left. Note that the
several domain boundaries are coincident with high values of the horizontal gravity
gradient (e.g. Vulcan structure in Southern Alberta and Snowbird Tectonic Zone). These
variations are clearer than in the Bouguer anomaly map shown earlier in this handout.
The horizontal gradients of the Bouguer gravity anomaly data from the Chicxulub impact
structure defines the edges of the structure very clearly. Data from Alan Hildebrand, University
of Calgary. Noise in field data can make this procedure unstable, so smoothing may be needed.
● Size and geometry of landfills. The contents of the landfill are usually lower in
density than the rock that was removed. Burger Figures 6-38 and 6-39 show a
study of a known landfill that was used to see what gravity anomaly was
associated with a small landfill. Before and after air photos were used to estimate
the depth of the landfill.
● How accurate must the vertical surveying be for data that is accurate to 0.1 mgal?
Ore bodies are often higher density than the host rock and can produce positive
gravity Bouguer anomalies. Gravity can be used to estimate the excess mass of an
ore deposit, using Gauss’s theorem. While non-uniqueness prevents the spatial
distribution being uniquely determined, the total excess mass can be estimated
reliably (see Kearey 6.10.3).
The example below comes from Voisey’s Bay, Labrador. This massive sulphide
deposit has a pronounced positive gravity anomaly. Note that gravity inversion is
an automated procedure that determines a density model that fits the measured
gravity data. An inversion is a solution of the inverse problem, and non-uniqueness
must be taken into account.
In addition to surface and airborne gravity data, satellite gravity data has
contributed to exploration in offshore areas.
Figures from: http://topex.ucsd.edu/marine_grav/mar_grav.html
● Salt structures. Salt has a lower density than surrounding sedimentary rocks,
and gravity data can be used to characterize the geometry of salt layers and
domes. Locating the base of a salt body is difficult with seismic reflection data,
and alternative geophysical methods are being used. In combination with seismic
reflection data, gravity can be used to give a joint interpretation. This can include
hypothesis testing of a velocity model. Since gravity data is usually available on
a grid, it can be used to extrapolate features between, or beyond, seismic
coverage.
●Continental scale data gravity databases can also be used. More effective when
regional features due to isostatic effects are removed (Chapin, 1998 in volume
edited by Gibson and Millegan, 1998).
B6.5 Regional scale studies of lithospheric structure
● Isostacy : Gravity measurements have also led to the concept of isostacy. Based
on observations by Pierre Bouguer in South America (1735), and George Airy
and John Pratt in India (1800’s) who noted that pendulums were not deflected as
much as expected by large mountain ranges. These observations can be
interpreted to show that these mountains and plateau are supported by a low
density root. This is very similar to the way that an iceberg floats in seawater,
where most of the ice is below sea level.
● Basin structure and earthquake hazards: The Seattle Fault forms the Southern
edge of the Seattle Basin, and may have caused M=7 earthquakes in the recent
geological past. In Lab 2 we will model gravity data that have been used to study
the geometry of this fault. In this study data were collected along the shore of
Lake Washington. This allows measurements to be made at a constant elevation
and removes the need for Bouguer or Free Air corrections to be made.
Modelling of these gravity data also provides a useful illustration of the inherent
non-uniqueness associated the interpretation of gravity data.
B6.6 Temporal variations in gravity
The data above were collected at Long Valley Caldera, California. It is not clear if
this will provide a useful tool for predicting eruptions. Typical changes are of the
order of 50 μgal.
Data analysis requires that time variations in elevation must also be accounted for.
Why is this important?
References
Gravity and magnetic methods at the turn of the millennium, Geophysics, 66, 36-
37, 2001.
Won, I., et al., Geophysics and weapons inspection, The Leading Edge, 658-662,
July, 2004
6. Applications
6.1 Locating caves and voids. Geotechnical, archaeological and military applications.
6.4 Hydrocarbon exploration. Eötvös torsion balance used in early 20th century to locate
salt domes. Airborne and marine data used in reconnaissance of sedimentary basins.
Gradient techniques for imaging base of salt layer.
6.5 Regional scale studies of crustal and upper mantle structure. e.g. Western Canada,
Chicxulub impact crater in Mexico.
MJU 2006
224 C1 Basic principles of seismology
A wave can be defined as a periodic disturbance that transmits energy through a medium,
without the permanent displacement of the medium. Also required that energy is converted
back and forward between two different types.
Before considering the details, it is useful to review some basic definitions. Consider the two
waves shown in the MATLAB movie waves.m
Frequency (f) : The number of cycles a given point moves through in 1 second. Frequency is
measured in Hertz (Hz). If the frequency is very low, then it is common to refer to the period (T)
of the signal in seconds. T = 1/f
Angular frequency (ω): Frequency is the number of rotations per second. The angular frequency
is the number of radians per second and given by ω=2πf
Wavelength (λ) : Distance in metres between two points of the wave having the same phase (e.g.
two crests or two troughs).
● Can you check that this relationship is correct for the figure above?
● Note that points on the wave move up and down, they are not translated to the right.
● In this case, the velocity is independent of frequency. This type of wave behaviour is called
non-dispersive. If velocity varies with frequency, the wave is said to be dispersive
1
In seismology, we need to understand how waves will travel in the Earth. For example, how fast
will they go, which direction, how will amplitude vary with distance etc. In general this requires
the solution of some complicated differential equations.
In Geophysics 224 we will approach this subject through visualization. Wave propagation can be
considered in two ways, by considering either wavefronts or rays. These are complementary
ways of talking about waves:
One way to visualize wave propagation over time is through Huyghens Principle. This states
that all points on a wavefront can be considered secondary sources of wavelets. These secondary
wavelets propagate outwards and at a time later, the overall wavefront is the envelope of
secondary wavelets. Examples for a point source is shown below.
2
C1.2 Wave propagation in the Earth
Having considered some general aspects of wave propagation we now need to consider how
waves will propagate in Earth. Waves in the Earth can be divided into two main categories:
(b) Surface waves are confined to interfaces, primarily the Earth-Air interface.
P-waves : Particle motion is in the same direction as the wave propagation. They are
also called compressional or longitudinal waves
and
S-waves: Particle motion is at right angles to the wave propagation. Also called shear
waves or transverse waves)
The velocity of body waves can be calculated from the properties of the material, as outlined
below. More details are given by Kearey pages 22-23, 3rd Edition. The mathematics is simplest
for the geometry shown below for a column of rock with cross-sectional area (A).
3
If a force (F) is applied at the left end, this can be considered an applied longitudinal stress =
F/A. The rock deforms elastically in response to the applied stress. The deformation of the
ΔL
leftmost disk can be quantified as the longitudinal strain =
L
which is a measure the stiffness of the rock (i.e. how much stress is needed to produce a given
strain).
This strain produces a force that will cause the shaded section of the rock to accelerate to the
right. This lowers the stress to the left, but increases it to the right. This causes the next section of
the rock to move and so on. Can show that a wave motion will move down the column at a
velocity
ψ
v=
ρ
where ρ is the density of the material. Note that the stiffer the medium (larger ψ) the greater the
force on the shaded cylinder, thus acceleration is higher and wave velocity is greater. Similarly,
as density increases, the shaded section becomes heavier and it’s acceleration (and wave
velocity) for a given force will decrease.
In general, the calculation of velocity is more complicated as the deformation will involve both
compression and shearing. Two other modulii must be defined to fully understand these effects.
4
4
It can be shown, that these modulii are related as ψ = K + μ
3
1
⎡K + 4 μ ⎤2
vP = ⎢ 3 ⎥
⎢ ρ ⎥
⎣ ⎦
1
⎡μ ⎤2
vS = ⎢ ⎥
⎣ρ⎦
● In a liquid μ=0 while K is always non-zero. Thus only P-waves can travel in a liquid, since
shear stresses cannot exist.
● These expressions for vP and vS do not depend on frequency, thus body waves (both P-waves
and S-waves) are non-dispersive.
● Note that as the rock cylinder is stretched, it will get longer and thinner. This effect can be
quantified through Poissons ratio. This is defined as:
σ= lateral strain
longitudinal strain
● Note that individual values of vP and vS depend on several modulii and density. This can make
it difficult to compare the velocities of similar rocks.
For typical consolidated crustal rocks, σ ~ 0.25 and vp/vs ~ 1.7. An increase in vp/vs and/or
Poisson’s ratio can be indicative of the presence of fluids.
5
C1.2.2 Surface waves
These waves are localized at the Earth’s surface and can be divided into two types. The particle
motions are illustrated in Kearey Figure 3.4
Rayleigh Waves are like ripples on a lake. Particle motion is in a retrograde ellipse.
Note that:
●The velocity of a Rayleigh wave does not vary with frequency when travelling in a uniform
medium and it is slower than an S-wave. In a layered Earth the velocity of a Rayleigh wave
varies with frequency (it is dispersive) and can be used to infer velocity variation with depth.
●Love waves only exist if the Earth is layered and are always dispersive.
●Large earthquakes can generate surface waves that travel around the globe. They can be large in
amplitude and cause a lot of damage during earthquakes.
●In exploration seismology, ground roll is a Rayleigh wave that travels across the geophone
array.
6
C1.3 Typical seismic velocities for Earth materials
Granite 5.5-6.0
Gabbro 6.5-7.0
Ultramafics 7.5-8.5
Air 0.3
Water 1.4-1.5
Ice 3.4
Petroleum 1.3-1.4
● why does vp apparently increase with density? e.g. for the sequence granite-ultramafics.
1
⎡K + 4 μ ⎤2
The equation v P = ⎢ 3 ⎥ suggests that vp should decrease as density increases.
⎢ ρ ⎥
⎣ ⎦
ρ = ρf Φ + (1- Φ) ρm
1 Φ (1 − Φ)
= +
v vf vm
7
Example 1 : A rock has 30% porosity and the P-wave velocities in the rock matrix and pore fluid
are 1.5 and 2.9 km s-1 respectively. What is the overall seismic velocity?
Example 2 : Consider the gas reservoir shown below. The reservoir sand has 10% porosity. The
following velocity data was obtained from well-log information
What is the average P-wave velocity in the water-saturated reservoir? What about in the gas
saturated reservoir?
Example 3 : In a similar reservoir, the P-wave velocity in the gas filled horizon was measured to
be 2.2 km s-1. What porosity does this imply?
With increasing depth, compaction increase the density of a rock through reduction of pore
space. The rigidity of the rock also increases with depth. The net effect is that velocity will
increase with depth, even if the lithology does not change.
The increase in velocity with burial depth is quite well defined, and deviations from the expected
velocity at a given depth are sometimes used to infer uplift histories in basins. This technique is
based on the observation that the compaction is essentially an irreversible process (Telford
Figure 4.21)
8
Rippability
particular rock type can be ripped with a given tractor. This also gives a useful
compilation of the range of velocities for a given material. (Kearey Figure 5.25)
As a seismic wave travels through the Earth, several factors will change the direction and
amplitude of the waves. When detected at the surface, an understanding of these factors can tell
us about sub-surface structure.
9
C1.4 Reflection coefficients at normal incidence
Consider a seismic wave that is travelling vertically downwards, as shown in the figure below. If
the wave reaches an interface, some of the energy will be transmitted, and some will be reflected.
Note that the waves travels from medium 1 into medium 2.
For each medium, the impedance is defined as the product of density (ρ) and seismic velocity (v),
Z = ρv
Ar v ρ −v ρ Z − Z1
=R= 2 2 1 1 = 2
Ai v 2 ρ 2 + v1 ρ1 Z 2 + Z 1
At 2v1 ρ1 2Z1
=T = =
Ai v 2 ρ 2 + v1 ρ1 Z 2 + Z1
The reflection coefficient is a measure of the change in impedance across the interface. These
equations are called the Zoeppritz equations. If the wave is incident at an angle the become more
complicated.
Sometimes reflection co-efficients are expressed in terms of energy, not amplitude. The equations
for this situation are listed in Kearey, chapter 3.
10
Example 1
The MATLAB script reflect_v1.m generates a movie showing how a seismic pulse propagates in
the Earth. For simplicity, density is constant and only velocity varies with depth. In this example
there is an increase in seismic velocity (and impedance) at 500 m. Note that the reflection has the
same polarity as the down going pulse.
What are the values of R and T? Is the largest amplitude in the reflected or transmitted wave?
Example 2
Same geometry as Example 1, but a decrease in velocity (impedance) occurs at 500 m depth.
Note that the reflected pulse has the opposite (negative) polarity to the down going pulse. Is this
what the Zoeppritz equations predict?
What are the values of R and T? Is the largest amplitude in the reflected or transmitted wave?
11
Example 3
This has two interfaces, each associated with an increase in velocity. Both arrivals have a
positive polarity, showing that they originate at interfaces where the velocity (impedance)
increases.
What other factor makes the amplitude of the second reflection smaller than the amplitude of the
first?
12
Example 4
This has a low velocity layer (LVZ). The first reflections has negative polarity show a decrease
in velocity, while the second has positive polarity, showing an increase in seismic velocity. Note
that a pulse reverberates within the low velocity layer producing multiple later arrivals, with
ever decreasing amplitude.
Example 5
Consider a gas reservoir with 10% porosity where the rock matrix has a vp = 3000 ms-1. The
overlying rock has vp = 2800 ms-1.
●What value of R is expected for a reflection from the top of the reservoir?
●What value of R is expected where the 2800 ms-1 and 3000 ms-1 layers are in direct contact?
This is an example of a bright spot, which is a high amplitude reflection. In the Earth, typical
reflection coefficients are ±0.2 with the maximum values ±0.5
13
C1.5 Reflection and refraction at non-normal incidence
In a more general case, the seismic wave will be incident on an interface at some angle of
incidence, θi. Note that the angle of incidence is measured from the normal to the ray. Snell’s
Law was developed for optics, but can equally be applied to the seismic case. Consider a P-wave
that strikes the interface shown below. In this case v2 > v1
Reflected and refracted P-waves are generated from the incident P-wave. For the reflected P-
wave, θr = θi . Snell’s Law states that the ray parameter, p, for the incident and refracted waves
will be constant.
Thus with a velocity increase (v2 > v1) the ray is refracted away from the normal.
If the velocity decreases (v2 < v1) then the ray is refracted towards the normal.
Note that if v2 > v1 then there will be a value of θi which results in sin θt = 1. This gives a value
of θt = 90˚ and the refracted waves travels horizontally. In this configuration θi = θc and is called
the critical angle.
sin 90 o sin θ c ⎛ v1 ⎞
= and by rearranging we find that θ c = sin −1 ⎜⎜ ⎟⎟
v2 v1 v
⎝ 2⎠
The wave travelling horizontally is called a head wave. For a head wave to develop, we must
have v2 > v1. Using Huyghen’s Principle, it can be shown that the head wave will generate
upward propagating wave at an angle θc to the normal. When these waves reach the geophones
they are called refracted arrivals. When θi > θc the wave is totally reflected.
14
In general we must also consider that the incident P-wave will generate a reflected S-wave and a
refracted S-wave through a process called mode conversion. Again Snell’s Law can be used to
calculate the angles of reflection and refraction.
The amplitudes of the transmitted and reflected waves can be calculated from Zoeppritz’s
equations, which have a more complicated form than those listed in C1.4 (see Applied
Geophysics, Telford, 2nd Edition, page 157 for details and some examples). The
amplitude of the reflected wave varies with angle, and is sensitive to the change in
impedance across the boundary. This is the physical basis of the amplitude versus offset
(AVO) technique that we will discuss later on.
C1.6 Diffraction
Seismic energy can sometimes travel in regions where ray theory (Snell’s Law) does not predict
that it will go. This is typified by diffraction which occurs when a wave strikes an object that is
significantly smaller than a wavelength. A diffractor effectively radiates seismic energy in all
directions.
When a wave strikes a corner, Huyghens Principle shows that the corner will generate waves that
propagate in all directions. We will see in real seismic reflection data that these waves can have
significant amplitudes and are detected over a wide area at the surface. The mathematical process
of migration is used to trace the pattern of diffracted waves back the point source where they
originated. This allows subsurface geometry to be correctly imaged in 2-D and 3-D.
15
C1.7 Factors that cause the amplitude to change as wave propagates
Conservation of energy requires that this quantity remain constant as r increases. Thus
1 1
X2 α and Xα
r2 r
As a seismic wave travels in the Earth the amplitude will decrease as 1/r, even if no energy is
lost. This phenomena is known as spherical divergence or geometric spreading.
(b) Attenuation
As a wave passes through the Earth, the particle motion causes the material to be distorted and
the wave energy is converted in heat. This results in an additional loss of energy, that is described
by an exponential decay:
X = Xo e-kr
16
Where e = 2.718, Xo is the amplitude at r=0 and k is a constant. If k is small, the attenuation will
be small, as k increases, the attenuation becomes stronger. In a distance 1/k the amplitude falls
1
from Xo to Xo .
e
● The pulse can be decomposed into a set of sine or cosine curves with different wavelengths
(Fourier’s Theorem and illustrated in MATLAB script fourier_square_wave.m)
● As the pulse travels the short wavelength signals attenuate more quickly.
● The long wavelengths dominate, giving the pulse a smoother shape and longer duration
(example in Kearey, Figure 3.7)
(c) Scattering
Suppose a medium is inhomogeneous and contains some grains with a different seismic velocity
to the host rock. Seismic waves will be diffracted / scatteredfrom these grains and energy will be
lost from the coherent wavefronts and turned into random seismic energy. The net result is that
energy will be lost.
Footnote : Decibels
A seismic wave changes in amplitude from A1 to A2 as it travels from point 1 to point 2. The
corresponding intensity changes from I1 to I2. Note that I1 = A12. This change in decibels can be
expressed as :
⎛I ⎞ ⎛A ⎞
dB = 10 log10 ⎜ 2 ⎟ = 20 log10 ⎜ 2 ⎟
⎜I ⎟ ⎜A ⎟
⎝ 1⎠ ⎝ 1⎠
17
C1.8 Seismic energy sources
● A high frequency seismic source is needed to generate short wavelength signals (the
wavelength should be less than thickness of layers to be studied). However, if the frequency is
too high, then attenuation will become a significant problem.
● Firearms give higher frequencies, but are not always permitted in civilized countries (e.g.
shotgun, semi-automatic etc.)
● Buffalo guns and Betsy guns allow a blank cartridge to be fired without using ammunition.
● Weight drops are also used and are commonly mounted on the back of a small truck. Larger
weight drops have included the use of helicopters.
Explosives
Explosives were once widely used, but this has declined with the development of the Vibroseis
method. Typically rotary drilling is used to place the shot in an 8-10 cm diameter hole at a depth
of 6-30 m. Drills can be mounted on a truck, or slung beneath a large helicopter. While
spectacular, blasting rocks and soil into the air represents a loss of energy and a safety hazard.
Advantages
Sharp impulsive signal. Strong signals can be generated if needed for imaging the lower crust and
upper mantle.
18
Disadvantages
Not repeatable, and quite slow. Most energy is in P-waves (why?).
Many environmental concerns and cannot be used in urban areas.
Vibroseis
For deeper exploration on land, the Vibroseis method is now the most widely used technique.
This does not generate a sharp impulsive signal, but generates a longer waveform that typically
lasts 10-20 seconds with frequencies 10-100 Hz.
Data processing requires that the field recording is analysed for correlation with the known
source signal. This can allow signal detection in areas with significant seismic noise (e.g. vehicle
traffic, wind etc)
Vibroseis trucks can generate forces up to 100,000 N and generally need a hard surface upon
which to operate. It is debatable whether they cause no damage to a road. For extra signal
strength, multiple trucks are used and they vibrate in phase. Both P-waves and S-waves can be
generated.
19
Example 0 : Two separate reflections
20
Example 2 : Two overlapping reflections and moderate noise levels
21
Stacking is used to increase the signal-to-noise ratio. The inherent assumption of stacking is that
the signal is the same each time (coherent) and the noise is different (incoherent). Coherent
noise can not be removed by stacking.
22
1.8.2 Marine seismic sources
Air guns
An airgun works by releasing a bubble of high pressure air into the water. The rapid expansion of
the bubble generates seismic energy with a frequency content around 10-100 Hz.
Advantages
● Very repeatable, reliable source.
Disadvantages
●The bubble pulse oscillates, generating a relatively long wave train. However, by using an
array of air guns at differing depths, the combined waveform can be made shorter in duration
(figure from Kearey).
● In water only P-waves are generated, but S-waves can be generated by mode conversion at
the seafloor.
● In recent years there has been a lot of concern about how marines seismic exploration affects
marine mammals.
23
Explosives
Once widely used in marine seismics, but now largely replaced by air guns. Bubble pulse problem
is difficult to overcome (Kearey figure 3.15), but can be minimized by detonating near the surface
(with some loss of energy). Occasionally large shots are used in deep studies but many restrictions
are in place to protect oil production facilities and fisheries.
24
C1.9 Seismic detectors
Geophones
On land, the surface moves as a P-wave or S-wave arrives. Generally reflected signals arrive at
steep angles of incidence. Thus P-waves produce surface motion that is dominantly vertical.
Geophones measure ground motion by converting motion into electrical signals. Most geophones
measure a single component (vertical), but multiple component ones are sometimes used.
Geophones are manufactured to detect a particular frequency band. This should match the seismic
source being used in a particular survey.
Geophones are connected to telemetry cables that transmit the recorded signals back to a
recording unit.
Seismometers
For earthquake studies a more permanent installation is usually required. Three components are
usually recorded and the sensor is tuned to detect lower frequencies. Often the seismometer is
placed in a shallow vault to minimize wind and other forms of noise.
25
1.9.2 Offshore seismic exploration
In offshore exploration, many hydrophones are placed in neutrally buoyant, oil-filled streamers
that can be in excess of 6 km long. The depth is controlled by movable fins and a tail buoy is used
to determine the direction of the streamer relative to the ship.
Feathering occurs when ocean currents push the streamer at an angle to the direction in which the
survey ship is travelling. Modern ships can pull multiple streamers, and this overcomes feathering
and also provides 3-D coverage.
26
C2.1 Travel time curve for a single horizontal interface
Seismic energy can travel from the shot (*) to the receivers (geophones) by the 3 distinct
routes shown in the upper panel. The lower panel is a travel-time curve and is a plot of
travel time as a function of distance from the shot (x).
1
(3) Reflection
A P-wave reflects from the interface between layer 1 and layer 2. The angle of incidence
and reflection are equal (see C1.5). Using Pythagoras’ Theorem, the distance travelled by
the seismic signal on the downward leg of the journey is :
x2
d = z2 +
4
From symmetry, the total distance travelled is 2d. The whole journey is travelled at
velocity v1, so the travel time is given by
x2
z2 +
4 = 4z 2 + x 2
t ref = 2
v1 v1
Note that:
(1) When x = 0, tref is not zero, in contrast to the direct wave and ground roll.
(2) tref has a minimum value when x = 0. In this situation, the seismic signal travels
vertically and makes an angle of 90° with the interface. This geometry is called
normal incidence and the travel time is tref = t0 = 2 z / v1
(4) At normal incidence, two unknown model parameters (z and v1) determine t0. The
solution of this inverse problem (to find the model parameters z and v1) is non-
unique. This is because we have two unknowns and just one equation.
(5) As x gets large the path taken by the reflection becomes very close to that taken by
the direct wave. This means that the travel times for the reflection and direct wave
become very close and tref → x / v1
Can you sketch the travel-time curve for the reflection on the figure above?
To address the problem of non-uniqueness in (3) and find the depth (z) and velocity (v1)
we need to consider values of x > 0.
2
4z 2 x 2 x2
2
t ref = + = t 2
0 +
v12 v12 v12
Note that this equation is for a parabola (tref varies as x squared). Simple re-arrangement
gives :
4z 2 + x 2 2z x x
t ref = = 1 + ( ) 2 = t0 1 + ( ) 2
v1 v1 2z 2z
1
1/ 2
⎡ x ⎤ ⎡ x 2⎤2
t ref = t 0 ⎢1 + ( ) 2 ⎥ = t 0 ⎢1 + ( ) ⎥
⎣ 2z ⎦ ⎣ v1t 0 ⎦
We can simplify the equation for tref by using a power series expansion (Taylor’s
Theorem) and assuming that x / v1t0 is relatively small.
1 1
⎡ x 2⎤2 ⎡ x 2⎤2 ⎡ 1 x 2 ⎤
t ref = t 0 ⎢1 + ( ) ⎥ = t 0 ⎢1 + ( ) ⎥ = t 0 ⎢1 + ( ) + ...... ⎥
⎣ v1t 0 ⎦ ⎣ v1t 0 ⎦ ⎣ 2 v1t 0 ⎦
If the higher order terms are ignored, then we can write that:
⎡ 1 x 2⎤ x2
t ref = t 0 ⎢1 + ( ) ⎥ = t0 + 2
⎣ 2 v1t 0 ⎦ 2v1 t 0
Re-arranging gives an expression for tref - t0 which is termed the normal moveout.
x2
t ref − t 0 =
2v12 t 0
Normal moveout (NMO) is a measure of the extra time taken for seismic signal to travel
on a non-vertical path, compared to the time for a signal travelling vertically. The word
normal refers to the seismic energy travelling at right angles to the interface.
A graph of NMO versus x2 will be a straight line, passing through the origin. The slope of
the line will be 1 / 2v12 t 0 . As x increases, the approximation made above becomes invalid
and a deviation from the straight line will be observed.
Example 1
In this example, we will look at the synthetic seismic reflection data and compute the
depth of the interface and estimate the velocity of the layer. These data are shown in
graphical and numerical form on pages 4 and 5.
3
(3) At this geophone, compute the NMO NMO = ________________
(5) Using the values for v1 and t0 find the depth of the interface. This can be done with the
equation for travel time at normal incidence t0 = 2 z / v1 . Rearrange this equation to
find the depth to the interface.
z = __________________
Does your answer agree with that shown in the plot below? If not, can you suggest a
reason for the discrepancy?
(6) What can we determine from the seismic data about the seismic velocity below the
interface (v2)?
4
Note : To be more rigorous when computing v1 it is best to plot a graph of NMO vs. x2
which should be a straight line with slope 1 / 2v12t0 . A non-straight line may indicate that
the simple form of the NMO equation is invalid.
5
(1) Look at the seismic reflection data presented above. How are they different from
those in Example 1?
6
7
C2.2 Travel time curves for multiple reflectors
● The ray paths and travel times for multiple layers can be computed using ray-tracing,
as demonstrated in Lab 3.
1
● The two examples both show three reflections. Note that:
(b) The polarity and amplitude of a reflection is due to the impedance change at
the interface
(c) Deeper reflections are flatter (i.e. they exhibit less normal moveout)
(d) The high amplitude reflection from a shallow interface can weaken later
reflections (from deeper structure)
(e)For a shot gather, the ray paths reflect (bounce) at different points on the
reflector surface.
(f)This geometry, with geophones on each side of the shot is called a split-spread
2
● In the real shot gather note that:
(b) Positive arrivals have the trace shaded black. This helps the eye to correlate traces
when noise is present. Beer sometime helps with this process, but too much has
been known to have the reverse effect. Calibration is needed to optimize this task.
(d) The ground roll is long lasting and obscures reflections at small offset (A).
(e) Slower arrivals plot at a steeper angle on the seismogram (more time needed to
travel the same distance)
(f) Reflections can be recognized by their parabolic shape and are labelled B,C,D,
and E. Note that they are never the first arrival. This means that reflections are
recorded when the ground is still moving from earlier arrivals, and reflections can
be obscured by noise.
3
Interval velocity and average velocity
Consider the multiplayer Earth model shown above. The ith layer has a thickness zi and a
velocity vi . The ray travels through the layers 1 to n and is refracted at each interface. It is
then reflected by the nth interface.
The interval velocity is the actual velocity in a specific layer and is defined as vi = zi / τi
V =
∑z i
=
∑v τ i i
∑τ i ∑τ i
Another way of averaging the velocity is to use the root-mean-square average. This is
defined as:
1
⎡ n
⎤ 2
⎢ ∑ vi τ i ⎥
2
Vrms ,n = ⎢ i =1n ⎥
⎢ τ i ⎥⎥
⎢⎣ ∑ i =1 ⎦
and is needed to compute interval velocities, as described below. For the case of n = 2
1
⎡ v 2τ + v 22τ 2 ⎤ 2
Vrms , 2 =⎢ 1 1 ⎥
⎣ τ1 +τ 2 ⎦
4
Normal move-out for multiple layers and the Dix equation
To determine the depth of the interfaces and the interval velocities, an approach similar to
that used in C2.1 is needed. Note that we need more than just the travel-time at normal
incidence, and should use a plot of NMO vs. offset (x). When the seismic signals travel
close to the vertical direction, can show that the normal moveout for the nth reflection is:
x2
Δt = t − t n = 2
2t nVrms ,n
This is very similar to the equation in C2.1 for the normal moveout above a single
interface. The only change is that the interval velocity (v1) has been replaced by the r.m.s.
velocity (i.e. the average velocity down to the nth reflector). It can be shown the root-
mean-square velocity, gives the correct average in this context.
For each reflection we can measure the zero-offset travel times (t1 , t2 , t3). If necessary a
plot of t2 versus x2 may be needed (see C2.1)
Next the r.m.s. velocities are computed from the normal moveout. Consider the 4th
geophone at distance x. The normal moveout at this value of x are Δt1, Δt2 and Δt3. From
these NMO values, the average (r.m.s.) velocities can be computed.
x2
NMO = Δt1 ( x) = t ( x) − t1 = 2
2t1Vrms ,1
x2
Vrms ,1 =
2t1 Δt1 ( x)
5
For the other reflections it is obvious that
x2 x2
Vrms , 2 = and Vrms ,3 =
2t 2 Δt 2 ( x) 2t 3 Δt 3 ( x)
The next stage of the data analysis is to compute the interval (true) velocities (v1, v2, v3)
from the average (r.m.s.) velocities (Vrms,1, V rms,2, V rms,3)
1
⎡Vrms
2
, n t n − Vrms , n −1t n −1
2
⎤2
vn = ⎢ ⎥
⎣⎢ t n − t n −1 ⎦⎥
While this appears complicated, it is really a simple recursion equation. This means that
if we know the velocity in one layer, the equation tells us what the velocity will be in the
layer below.
To start this process, we begin at the surface. For the first reflection, the seismic signal
travels only in Layer 1. Thus v1= Vrms,1
Now Vrms,2 represents an average velocity of layers 1 and 2. Thus to find v2 we can use the
Dix equation with n = 2
1
⎡Vrms
2
, 2 t 2 − Vrms ,1t1
2
⎤2
v2 = ⎢ ⎥
⎣⎢ t 2 − t1 ⎦⎥
This can then be repeated as many times as necessary. For the next layer down we can
write
1
⎡Vrms
2
, 3 t 3 − Vrms , 2 t 2
2
⎤2
v3 = ⎢ ⎥
⎣⎢ t3 − t 2 ⎦⎥
Example : Look at the data in the table above for the three reflection seismogram in
Example 1. The travel times for the first reflection are the same as for the example in
C2.1
6
Consider the second reflection
Use the Dix equation to compute the interval velocity for layer 2:
1
⎡V 2
t −V
rms , 2 2
2
t ⎤
rms ,1 1
2
v2 = ⎢ ⎥ =_______________
⎣⎢ t 2 − t1 ⎦⎥
Now compute the thickness of the second layer. To do this note that the reflection spends
a time t2-t1 in the second layer. In this time, it travels a distance 2z2 . Thus we can write
2z2
v2 =
t 2 − t1
v 2 (t 2 − t1 )
z2 = = _________________
2
Vrms,3 =______________
1
⎡V 2
t −V
rms , 3 3
2
t ⎤
rms , 2 2
2
v3 = ⎢ ⎥ = _______________
⎢⎣ t3 − t 2 ⎥⎦
v3 (t 3 − t 2 )
z3 = = _________________
2
7
Multiple reflections
(a) The second reflection is a multiple bounce from the interface at 0.4 km depth.
(b) The travel time of the multiple at zero-offset is exactly double that of the single
reflection. This can help identify a multiple.
(c) Note that when the seismic energy bounces at the Earth-Air interface, the
reflection coefficient can be large. Remember that
ρ 2 v 2 − ρ1v1
R=
ρ 2 v 2 + ρ1v1
Consider the upward travelling wave as it approaches the surface. ρ2 is now the
density of the air which is essentially zero. This gives R = -1 and almost all
energy is reflected downwards.
(d) For an example of multiples in single channel seismic data, see Kearey Figure
4.56 and the discussion in C2.3
8
C2.3 Strategies for collecting seismic data
• In many seismic surveys, we are most interested in how structure varies with
horizontal position and may be less interested in actual depths.
• Note that the seismic section is composed of many individual traces. A trace is
the time variation of the ground motion (or water pressure in a marine survey)
recorded by the receiver for a single shot.
• In this type of seismic survey the absolute depth cannot be determined. The travel
time = t = depth/ velocity. With just one measurement of travel time, t, it is
1
impossible to determine two unknowns. Only relative changes in depth can be
determined.
• Seismic reflections are generally quite weak (low amplitude) and this technique
does not generally work well on land where cultural noise levels are higher and
the coupling between the geophone and the ground is variable.
• Example above is from marine seismic exploration in the Gulf of Corinth (Kearey
Figure 4.56). Note that D is a direct wave where a signal travelling horizontally
passes the receiver. All other arrivals have travelled vertically downwards in the
Earth and back again. The seabed (SB) gives a strong reflection, and multiple
reflections are quite strong (SBM1 and SBM2) and obscure other reflections from
greater depth. RH is a reflection from bedrock that is exposed above the seafloor
on the right side of thee profile.
2
C2.3.3 Multi-channel seismic profiling
• By combining the two techniques described above, we can measure both the
velocity variation with depth and horizontal variations in structure on a
profile.
• One simple acquisition strategy is sketched above. Note that at location ‘A’ the
reflector is quite shallow. At location ‘B’ the reflector is deeper and the reflector
shows a larger zero-offset travel time and less normal moveout. Analysis of these
data will give estimates of both the interface depth and velocity.
• However this has the disadvantage that the energy received at each receiver has
been reflected at different points. The common mid-point technique can overcome
this problem. Seismic data is grouped into shot-receiver paths that share a
common reflection point. This technique is discussed in detail in C2.5
3
C2.4 Travel time curve for a dipping reflector
x 2 + 4 z 2 + 4 xz sin θ
t=
v1
● t(x) is greater than t(-x) and the travel time curve is asymmetric about x = 0
● also note that the reflection received at x = 0 m did not originate beneath x = 0.
● To account for this effect a technique called migration is used. More on this later!
Using the same approach as in C2.1, we can show that
( x 2 + 4 xz sin θ )
t ≈ t0 +
2v12 t 0
2 x sin θ
The dip moveout is defined as ΔTd = t x − t − x =
v1
1
C2.5 Common mid-point profiling
● The shot gather has the disadvantage that each ray has been reflected from a different
point on the interface. If there are lateral variations in structure, this will result in errors.
● This can be overcome by using multiple shot points and multiple receivers. From the
total dataset, a set of rays are then chosen that have a common reflection point.
● This technique is called both common-mid point (CMP) or common depth point (CDP)
profiling. The shot and receiver are located equal distances from the CMP.
Matlab script cmp_v1.m shows the path geometry for a CMP gather.
●Note that the offset is the distance from shot (*) to receiver (v). Thus with
geophones from –200 m to + 200 m, the maximum offset between shot and
receiver is 400 m.
●The parabolic travel time curve derived for the shot gather can also be used to
analyse a CMP gather.
1
●The technique for collecting CMP data is illustrated above. A seismic source is located
at the front of a 6-channel array.
●After each shot is fired, the array is moved forward a distance nΔx.
●In a marine survey this would consist of the whole airgun and hydrophone streamer
being towed forward by the survey ship. On land the last geophone would be brought
up to the front of the array, a new hole drilled, loaded with explosives and fired.
● If we have N geophones and that array moves a distance n Δx between shots, then you
can show that the number of rays that share the same common mid-points = N/2n. This
quantity is also called the fold or the coverage (in percent).
● This survey will give 3 rays for each mid-point. This is called 3-fold CMP coverage or
300% coverage.
2
Stacking
● Once we have collected a number of rays that share the same mid-point, we can use
stacking to improve signal to noise ratio in the data. This is needed because reflections
are generally weak. Remember that stacking adds a number of traces. If the signal is the
same in each, but the (random) noise is different, then the noise will cancel and the
signal be enhanced.
● Prior to stacking, the individual traces must be shifted vertically (i.e. in time) to allow
the reflection to occur at the same time.
● Traces at non-zero offset are shifted in time to be equivalent to the zero offset trace.
● This requires that we compute the expected NMO for a given reflection and offset.
However we don’t know the velocity (this is the goal of the survey!) so a trial and error
approach is used.
● This is illustrated in cmp_v2.m. If the stacking velocity is too high, then the curve
remains a downward parabola. If the stacking velocity is too low, then the curve is an
upwards parabola.
● Each reflection is analysed and the stacking velocity systematically varied. The velocity
that gives the best stack (flattest reflection) is then taken to be the r.m.s. velocity from
the surface to the reflector. This is equivalent to plotting NMO as a function of x2.
● Note that the direct wave and ground roll do not stack coherently, and this process tends
to remove them from the data.
3
4
5
6
Geophysics 224 MJU 2006
7
C2.6 Resolution in seismic reflection surveys
When interpreting geophysical data, it is very important to be aware of the resolution that
is possible with the method being used.
Consider the thin layer shown in the Figure C2.6 with the upper surface at a depth of
0.2 km. Seismic reflection data is collected at normal incidence. The reflections from
the upper and lower surfaces were computed with MATLAB script
vert_resn_reflection.m
t = 200 m
t = 20 m
1
t = 50 m
This is produced by
the second reflection
travelling a distance 2t
greater than the first
reflection.
t = 60 m
Two waves can be distinguished from each other if they are separated by distance λ/2.
Now the reflection from the lower interface has travelled an additional distance of 2t.
Thus the best resolution occurs when 2t = λ/2 which gives t = λ/4
Example: A 40 Hz seismic signal reflects from the top and bottom of a layer in a medium
with v ~ 3000 m/s. What is the minimum layer thickness that could be detected?
2
How does this compare with features seen in an outcrop?
Two factors can limit horizontal resolution in a reflection survey. One is due to the
physics of seismic wave propagation (Fresnel zones), and the other is due to the
geophone spacing. The resolution will be which ever number is larger.
Fresnel zones
It is sometime tempting to imagine that a seismic reflection occurs only at the point
where the ray hits an interface. However it is important to remember that a complte
understanding of seismic wave propagation requires us to consider both wavefronts and
rays.
It can shown that the reflection essentially occurs from a circular disk, over which in
phase reflections occur. This is called the Fresnel zone and has a width w = 2 zλ
3
The width of this zone is the smallest feature that can be detected with the surface. Note
again that higher resolution will be achieved by using short wavelengths. However this
will result in more attenuation, and the signals will have limited penetration.
Example: A 40 Hz seismic signal reflects from the base of a 2000 m thick layer with v
~ 3000 m/s. What is the minimum size of horizontal feature that could be detected with
this configuration?
Detector spacing
The reflection points on an interface will have a horizontal spacing equal to half the
geophone/hydrophone spacing. By making this spacing less than the Fresnel zone, the
survey resolution will not be limited by the layout, but by the physics of the wave
propagation.
4
C2.7 Statics, noise reduction and filtering of seismic reflection data
So far we have assumed that onshore seismic data is being collected on a flat surface
with uniform near surface geology. However this assumption is not often valid, and two
effects will cause a variation in arrival times between closely spaced traces. These time
delays are called statics because they delay (or advance) all events in the seismic trace
recorded by a particular shot-geophone combination.
Elevation statics: geophones at an elevation above a datum (or reference level) will
detect the incoming signal later than a geophone on the datum.
Weathering statics : weathering of near surface rocks produces a zone of low velocities
that is variable in thickness
1
Why are statics a problem in seismic reflection data?
Statics produce a vertical offset in seismic traces, as shown in the CMP section
generated by cmp_v3.m. Thus when the traces are stacked, they will not stack
coherently.
2
3
Example 1
Statics are present in 50 Hz seismic reflection data. How large can the statics be
and still permit coherent CMP stacking?
Example 2
If there is no weathered layer, and the bedrock has v = 3000 m/s, what elevation
change will produce a static of 5 ms?
Example 3
The weathered layer has v=1000 m/s and the bedrock has v = 2500 m/s. What
change in the thickness of the weathered layer will produce a delay of 5 ms?
Correction of statics
● In residual static analysis, traces are automatically aligned to produce the most
continuous seismic event. Example below is from Kearey Figure 4.16
4
C2.7.2 Frequency domain filtering
When seismic reflection data is recorded, the individual traces data are contaminated by
signals other than seismic reflections. These signals include ground motion due to wind
noise, vehicles etc. There are also other seismic phases that can obscure the reflections
(ground roll). Electrical powerline noise (50 Hz and 60 Hz and harmonics) can also be
picked up by the telemetry cables. In marine surveys, noise signals originate in the motion
of the ship and streamer through the water.
Some of this noise has a frequency content that is quite distinct from the seismic signals
being recorded. These can be removed by frequency domain filtering techniques.
Notch filter : removes a very narrow band of signals (e.g. 60 Hz powerline noise)
High pass filter: All frequencies above the filter frequency are transmitted. This
an remove ground roll and certain ship generated noise (also called a low-cut
filter)
Low pass filter: all frequencies below the filter frequency are transmitted. Can be
used to remove high frequency noise in some circumstances.
Example: Kearey figure 4.19. Note that the frequency content in the seismic section
changes with time. Low frequencies dominate at later times, as the high frequencies are
more rapidly attenuated than lower frequencies.
However, many artefacts in a seismic section have a frequency content that is very close
to those present in the reflections. Thus they cannot be removed by conventional
frequency domain filtering.
5
C2.7.3 Convolution
y(t) = f(t)*x(t)
= ∫ f (τ ) x(τ − t )dτ
Graphically, convolution is illustrated in the attached figure, where the function x(t) is a
series of spikes (delta functions) and f(t) is a sinusoidal wavelet (MATLAB script
convolution_C27.m).
This mathematical procedure may seem quite abstract, but it occurs in reflection
seismology, as illustrated in Kearey Figure 4.6.
6
The sharp changes in acoustic impedance (Z) in the geological section are called the
reflectivity function. The locations where this function has spikes are the locations that
reflect seismic energy back to the surface.
However, to measure the reflectivity function we need to send a seismic signal through
the Earth. This seismic pulse has a finite length (in both space and time) and the seismic
trace (measured by the geophone) can be shown to be the convolution of the reflectivity
function and the seismic pulse.
This has the effect to spread out the sharp response in the reflectivity function associated
with each interface. As a consequence, the measured seismic trace is lower in resolution
and we cannot accurately locate the time at which a particular reflection arrives. This will
limit resolution, since we showed in section C2.6 that layers can only be resolved if they
are separated by more than λ/4.
Thus a sharper response will be obtained by making the seismic pulse as short as possible
in time. However, this requires an increase in frequency, and this will have the side effect
of increasing attenuation.
7
C2.7.4 Deconvolution
The process of deconvolution can be used to undo the effect of other filtering operations.
This includes the effect of a finite length seismic pulse, as illustrated above.
Often used to shorten the length of the seismic pulse. A seismic pulse can be lengthened
by a number of processes:
To overcome these, and other effects, a range of deconvolution techniques may be used.
More details and examples can be found in Kearey (p. 62-65).
Before After
● remember that in Vibroseis ® surveys, the seismic pulse is very long. However, the
fact that the shape of the pulse is known allows us to correlate the seismic trace with
the known waveform during processing. However, in surveys using explosives are
airguns, the source function is generally not well enough known for this to be applied.
8
C2.8 Migration of seismic reflection data
Consider a shot (S) and receiver (R) that are located close together. A reflector has a true
dip angle = αt and is at a depth z below the shot point. Seismic energy that returns to R
will reflect at ‘A’ where the ray path is at 90º to the reflector.
However, in a seismic section, a reflection in plotted as it was directly below the shot-
receiver point , which in this case is ‘B’. Note the line A-B is the arc of a circle, centered
at RS. The result is the reflector is imaged by the seismic data with an apparent dip of αs
that is less than the true dip.
The result is that reflectors appear to have an apparent dip that is less than the true dip.
The effect is small at shallow angles, but becomes significant when the true dips are
greater than 20 º. Note that a vertical reflector will appear at 45 º on a seismic section.
1
C2.8.1.2 Multiple reflections in same zero offset trace
● When a syncline has very limited topography, only one reflection is observed in
each zero offset trace (i.e. shot and receiver are placed very close together and
moved along the profile).
As shown in the previous section, the dip will not be measured correctly. However,
the overall geometry in the seismic record section will be very similar to the
reflector geometry. Note that reflections occur at right angles to the surface.
● With more rugged topography, a normal reflection will occur at three locations
when the source and receiver are above the syncline. This produces the
characteristic “bow tie” in the travel times.
2
3
C2.8.1.3 Diffractions
• Diffractions occur from point reflectors and corners. The figure above shows a
small sphere that is diffracting seismic waves. The diffractor has the property that
it scatters energy in all directions. Some of this energy returns to the location of
the shot (*) along the same path (like a boomerang).
• When the shot-receiver array is moved along the profile with zero-offset, the
2 x2 + h2
travel time curve above has a travel time t diff = which plots as a
v1
hyperbola.
• A hyperbola will also result when the shot-receiver offset is varied (e.g. a shot
gather or CMP gather). In this case the move out rate will be half that observed
from a plane reflector
4
C2.8.2 Migration methods
The simplest technique works by assuming that from the travel time, we know that the
reflection point must lie somewhere on a hemispherical surface. By migrating several
points this allows us reconstruct the reflector surface.
If applied to a diffraction, this will collapse the hyperbola to a point. The diffraction
hyperbola is the steepest dipping even that can be recorded on a seismic record section.
● diffraction migration
● Kirchoff migration
More details and references for these methods are given by Kearey in chapter 4. Each
migration technique can be further divided into time and depth migration
●time migration: the vertical axis on the migrated section is still time.
●depth migration: estimates of velocity are used to convert time to depth during
migration. Even rough estimates of velocity (e.g. from computing the stacking
velocity for each CDP gather) can greatly improve the quality of the migration.
Migration can also be applied before or after stacking the CMP gathers:
●Post-stack depth migration : This was the original migration strategy. The zero
offset traces derived from the CMP stacks are migrated. This has the advantage of
lower computer time and the higher signal-to-noise ratio can make the migration
more stable. However the process of stacking assumes a relatively flat interfaces.
This is valid in geological structures that have sub-horizontal stratigraphy, but is not
valid in complex, highly 3-D environments.
5
C2.8.3 Examples of migration
The Sigsbee model was developed to simulate sub-salt seismic imaging in deeper water
exploration areas such as the Gulf of Mexico. More details can be found in Integrated
approach to sub-salt depth imaging: synthetic case study, The Leading Edge, 21, 1217-
1223, December 2002.
6
Left : Kirchoff depth migration. Reflectors to the left and above the salt body are well
imaged. Right : Finite difference Pre-stack depth migration. Coherent reflectors are
imaged in the sedimentary sequence beneath the salt body.
Multiples do not migrate correctly and along with other errors in migration this can
produce “smiles” which are upward directed hyperbolae. This is illustrated in Figure 4.84
from Telford.
Example below is from a PGS advert in The Leading Edge. Note that structures with
steeper dip are imaged more reliably with pre-stack depth migration.
7
C2.9 Hydrocarbon exploration with seismic reflection
Question : This sequence is for post-stack depth migration. How will it change for pre-
stack depth migration?
●Seismic data can give direct evidence of the presence of hydrocarbons (e.g.
bright spots, oil-water contact, amplitude-versus-offset anomalies).
1
C2.9.1 Direct indicators of hydrocarbons
● However, not all bright spots are hydrocarbons. They can be caused by sills of igneous
rocks or other lithological contrasts. In areas of active tectonics, ponded partial melt can
produce a bright spot (e.g. Southern Tibet).
● It should also be noted that true amplitudes are not always preserved in seismic data
recording and processing. When data is recorded, the amplification of the signal varies
from trace to trace. It can also vary with time in a given trace (later signals have
travelled deeper and have a smaller amplitude so the amplification is increased). Thus
variations from trace to trace in a seismic section do not necessarily imply a change in a
sub-surface property.
●Focussing and defocusing effects on an undulating interface can also change the
measured amplitude. Beware!
Example 2 : The Alba Field, North Sea Details in Macleod et al., (1999)
2
Example 3 : Alexis Field, Alberta
●The seismic data use the reflection from the Viking Fm as a datum. This requires all the
seismic traces to be aligned on this reflector. This removes any residual statics and the
velocity effects of shallower structure. This is necessary because variations in the
reflection from the reservoir are quite subtle.
●To interpret the seismic reflections, velocity measurements are made in wells (sonic
logs). From the variation of velocity with depth, a synthetic seismogram can be
computed. This is the predicted seismic trace that would be measured at this location. It
allows us to determine which velocity and lithology changes will be detectable in
surface seismic data.
●Remember that the finite wavelengths of seismic data do not allow features less than λ/4
in thickness to be imaged (section C2.6).
●The reservoir in this case is quite thin, and completely separate reflectors are not seen
from the upper and lower interfaces. However, the shape of the reflector changes
distinctly between a wells in the reservoir (15-36) and a well outside the reservoir (3-1).
3
2.9.1.3 Amplitude versus offset (AVO)
Example 1
Both P-wave and S-wave velocities increase across the interface. Remember that at non-
normal incidence, an incident P-wave will generate 4 new waves. These are the reflected
and transmitted P-waves and the reflected and transmitted S-waves. The reflection
coefficient of the P-wave decreases with angle in this example.
4
Example 2
This example simulates the effect of shale overlying a gas saturated sand reservoir. The P-
wave velocity decreases, giving a negative reflection coefficient at normal incidence.
The S-wave velocity increases from the shale into the gas sand. Why?
(v P / v S ) 2 − 2
σ=
2((v P / v S ) 2 − 1)
Poisson’s ratio is usually around 0.5 for incompressible materials, and softer material will
have values in the range 0 to 0.2 In this case, the P-wave reflection coefficient becomes
larger as the angle of incidence increases. Since an increase in angle corresponds to a
greater source-receiver offset, this is called an amplitude-versus offset (AVO) anomaly.
The effect of AVO effects on a synthetic CMP gather are shown with cmp_v4.m
5
6
Data examples of AVO
7
C2.9.2 Images of hydrocarbon reservoirs on seismic sections
Viking gas field, North Sea, Kearey 4.61. Gas reservoir is located in a faulted anticline.
The most prominent reflections in a seismic section are called markers. These are
correlated across the section though a comparison of their character and sequence. They
are tied to lithologic units through measuring well logs and computing synthetic
seismograms. Vertical seismic profiles can also be used in this respect.
8
2.9.2.2 Fold and Thrust belts
9
Post–stack depth migration Pre-stack depth migration
2.9.2.3 Reefs
Leduc reefs
● Seismic characteristics of the reef are draping of overlying sediments and upper surface.
Note also that the base of the reef does not appear as a flat feature. It is pulled up
because of the higher velocity within the reef.
10
● Seismic section above is from the Redwater Leduc reef 40 km northeast of Edmonton.
The reservoir is Devonian age carbonates, overlain by Ireton formation shales that
provide a seal.
●Also note the base of the upper Leduc and Duvernay appears to rise under the reef. This
is an example of pull up. The base is flat and appears at earlier TWTT since the
velocities in the reef are higher than at the same depth east of the reef. This is also
illustrated in the synthetic example on the left (Andersen et al., 1989, Chapter 1.)
11
● Again, well logs are used to generate synthetic seismograms, which allow the
interpretation of the seismic section. This will show which lithologic contrasts will
dominate the seismic response. Allows the structure mapped in a well to be extrapolated
away from the well in places where seismic data are available.
12
Salt creeps under mechanical loading and salt sheets can flow horizontally. Salt diapers
can rise in a column to the surface. The figure above shows 4 second seismic reflection
data that reveals the evolution of a salt sheet, progressing from a salt swell, to a diapir that
penetrates the sediments, and finally to the surface (North Sea, Telford figure 4.102)
Hydrocarbon reservoirs may be located in the sedimentary sequence beneath a salt layer.
However, the high velocity of a salt body makes it difficult to image these reservoirs.
The examples above show the Enchilada, Hickory and Tanzanite sub-salt discoveries in
the Gulf of Mexico, and are taken from an advert by Diamond Geophysical Service
Corporation (The Leading Edge, November 1999) Brown denotes the salt body and
yellow the reservoir. Imaging techniques such as pre-stack depth migration have made a
big improvement in this type of exploration problem, as described in C2.8 for the Sigsbee
dataset.
In marine seismic surveys, an array of airguns and multiple streamers give a swarth of 3-
D coverage (Kearey 4-36). Once 3-D data has been collected, it may displayed as a
seismic data volume (seismic cube). A vertical slice corresponds to the depth sections we
have discussed.
13
Three-dimensional seismic data volume
(seismic data cube) collected over a salt
dome in the Gulf of Mexico. Kearey
Plate 4.1. Colour corresponds to
amplitude of reflection.
14
Sometimes 3-D surveys can show paleo
river channels in time slices. In this
location a particular reflector is deeper
over the river channel and at a given
time, this will give a different reflection
amplitude on CDP’s that are
within/outside the channel. Kearey Plate
4.4
● Now possible with the increased computing power that is available for processing
seismic data. With steeply dipping structures, this technique is much better at imaging
than post-stack depth migration schemes.
The Leading Edge, 21, Special section PSDM 2, The Sequel, December 2002.
● Concerns about the possible effects of air gun noise on marine mammals has
intensified in recent years. Details described by Gausland, 2000.
Multiples can obscure later arrivals in reflection data. Various processing techniques can
be used to remove them. Matson et al., (1999) give a comparison of various methods.
Left - Multiples in data from the Canarvon basin (Australia) from Matson et al., (2001).
Right – section after the most convincing multiple suppression technique was applied.
Vertical scale is 0 to 1.2 seconds
15
Pre-stack depth migrated images
before and after multiple removal.
Vertical scale is 1200 to 6800 m.
Subsalt geology is definitely
visible after multiple removal.
From Guitton (1999).
● what will be the repeat times of multiples in 1 km of water? In 2 km? vw= 1500 m/s
● S-wave data can have less problems with multiples than P-wave data? Why?
● State-of-the-art :The Leading Edge, 19, Special section AVO, November 2000.
● AVO doesn’t always work! Some cautionary tales from Allen et al, (1993)
● Use either ocean bottom seismometer (OBS) or ocean bottom cables (OBC). These
can be laid in a fixed location on the seafloor, or dragged by the survey ship.
● S-waves are generated by P-to-S conversion at the reflector. This requires non-
normal incidence, so large source-receiver offsets are used.
16
Imaging through gas clouds
In the Alba Field (North Sea) the reservoir sands are overlain by a shale cap rock. The
change in P-wave velocity is small at the top of the reservoir, and it is not imaged on the
P-wave section. However there is an increase in S-wave velocity that produces a strong
reflection (Caldwell, 1999).
17
Better illumination of high angle interfaces
● Seafloor recording of P-waves and S-waves can also give wider angle data
coverage. This can improve illumination of targets. Example from Mississippi
canyon, Gulf of Mexico, PGS advert, The Leading Edge, 20, 28, January 2001
The Leading Edge, 18, Special section : Multi-component offshore, November 1999
The Leading Edge, 20, Special section : Advances in shear wave technology, September
2001.
● Repeat seismic survey at same location to determine time variations during production
(time lapse or 4-D seismics)
18
● What other factors than the reservoir could change seismic data recorded at the same
location in different surveys?
● Differencing of seismic sections. Kearey 4.41 shows time sections before and after gas
was injected into a reservoir for storage.
● Kearey 4.42. Steam injection at a well in Duri oilfield, Indonesia. Base of reservoir
initially shows pull-up and then push-down (sag). What does this tell us about the
velocity in the reservoir during the injection?
References
Allen, J., C. Peddy and T. Fasnacht, Some AVO failures and what (we think) we have
learned, The Leading Edge, 163-167, March 1993.
Caldwell, J., Marine multi-component seismology, The Leading Edge, 18, 1274-1282,
1999.
Gausland, I., Impact of seismic surveys on marine life, The Leading Edge, 19, 903, 2000.
Guitton, A., Multiple elimination using a pattern recognition technique, The Leading
Edge, 18, 92-99, 1999.
Li, G. et al, Effective processing of non-repeatable 4-D seismic data to monitor heavy oil
SAGD steam flood at East Senlac, Saskatchewan, Canada, The Leading Edge, 20, 54-
59, 2001.
Macleod, M.K., et al, The Alba Field ocean bottom cable survey: Impact on development,
The Leading Edge, 18, 1306-1312, November 1999.
Ostrander, W.J., Plane-wave reflection coefficients for gas sands at non-normal angles of
incidence, Geophysics, 49, 1637-1648, 1984.
Telford, W.M., L.P Geldart, R.E. Sheriff, Applied Geophysics, Cambridge University
Press, 1990.
19
C2.10 Deep seismic reflection studies
In addition to their use in hydrocarbon exploration, seismic reflection studies can also be
used to image the entire crust and upper mantle. A number of government and university
research groups have made a number of deep seismic studies in recent years and
improved our knowledge of the structure and evolution of the Earth.
The first example is from the British Institutions Reflection Profiling Syndicate (BIRPS)
and is shown in Kearey Figure 4.63. BIRPS collected most of their data with marine
seismic surveys around Great Britain where the continental crust is present on the
continental shelf. This reflection section is just like the sections collected for hydrocarbon
exploration, except the recording time is longer (16 seconds). Collection of these data
require a strong seismic source, typically a very large airgun array. Often more energy
was used by the compressors than the ships engines. A very long seismic streamer was
used to detect the normal move out in deep reflections. Some BIRPS surveys used to 2
ships to create a longer offset.
This BIRP clearly shows the Moho at about 8 seconds and a major thrust fault that
apparently extends into the upper mantle. A common feature of many BIRPS profiles is
significant seismic reflectivity in the lower crust. The origin of this feature is still not
defined, but lenses of fluids, or mafic underplating are two possible explanations.
1
2.10.2 COCORP (Consortium for Continental Reflection Profiling)
The Consortium for Continental Reflection Profiling (COCORP) used land seismic data
to study deep crustal structure in the United States.
The figure above shows a seismic reflection image of the Wind River Thrust in
Wyoming. This is a major feature associated with the Laramide orogeny (Brewer et al,
1980). Typical acquisition parameters for COCORP data were VIBROSEIS with a 96-
fold CMP stack, recorder spacing of 100 m and spread length of 9.6 km. The
VIBROSEIS signal was 8-32 Hz and at each location was stacked up to 80 times.
By the early 1990’s COCORP had examined the crustal structure of many regions of the
United States. This included old orogenic belts such as the study above, and active
features such as the San Andreas Fault and Basin and Range province. However not all
tectonic processes are currently active in the United States and they began to look
overseas to places where certain tectonic processes are active today.
Interest in continent-continent collisions led to the INDEPTH project in Tibet. This used a
Chinese seismic crew to collect seismic reflection and refraction data on a profile that
crossed the Himalaya at 27.5º N and extended north onto the Tibetan Plateau. The
primary energy source was explosives.
2
These seismic data imaged the Indian Plate being thrust beneath the Himalaya. The Main
Himalayan Thrust (MHT) is the top of the Indian Plate. Note also that the Moho in the
Indian Plate can be seen at a two-way travel time of 25 seconds. This observation showed
that the 80 km thick crust was made by stacking Asian and Indian crust on top of each
other (Nelson et al, 1996).
Another key observation from INDEPTH was that bright spots were observed beneath
the Tibetan Plateau extending north from 29.75º N. These have a high amplitude and
negative polarity, and may represent a thin layer of fluids.
3
Panel on the below shows how the polarity of the reflection (YBS) is determined by
comparison with the first break (FB) of the direct P-wave. Since explosives were used as
a source, it is known that the signal has a compression as the first arrival (Brown et al.,
1996).
Combined with other seismic and petrologic data, many researchers believe they
represent a zone of partial melt, overlain by aqueous fluids. Additional evidence for
widespread melting comes from magnetotelluric data that show that low electrical
resistivity is coincident with the seismic reflections (Unsworth et al., 2005).
4
2.10.4 LITHOPROBE
Lithoprobe was a federally funded geoscience project that used combined geology and
geophysics to study the structure and tectonic evolution of a number of regions of Canada
between 1980 and 2000. Deep seismic reflection was an integral part of several transects.
This included marine seismic data collection on the west coast and in the Great Lakes.
Onshore seismic data was collected in other transects, including the Alberta Basement
Transect.
The map shows the location of seismic reflection lines in Alberta (black lines). A typical
depth section is shown below that extends north-south across the Vulcan structure. The
Moho is visible on the right side at a depth of 35-40 km. More details of interpretation are
described by Bouzidi et al., (2002).
5
References
Nelson, K.D., Zhao, W., Brown, L.D., Kuo, J., Che, J., Liu, X., Klemperer, S.L.,
Makovsky, Y., Meissner, R., Mechie, J., Kind, R., Wenzel, F., Ni, J., Nablek, J.,
Leshou, C., Tan, H., Wei, W., Jones, A.G., Booker, J.R., Unsworth, M., Kidd,
W.S.F., Hauck, M., Alsdorf, D., Ross, A., Cogan, M., Wu, C. Sandvol, E., Edwards,
M., Partially molten Middle Crust Beneath Southern Tibet : Synthesis of Project
INDEPTH results, Science, 274, 1684-1686, 1996.
Unsworth, M.J. A.G. Jones, W. Wei, G Marquis, S. Gokarn, J. Spratt, Crustal rheology of
the Himalaya and Southern Tibet inferred from magnetotelluric data, Nature, 438, 78-
81, doi:10.1038/nature04154, 2005.
6
C2.11 Vertical seismic profiling
1
Applications of vertical seismic profiling
1. Time-depth calibration
2
C3.1 Seismic refraction – single horizontal interface
The P-wave is refracted at the interface between the two layers. Since v1 > v2 the wave is
refracted towards the horizontal. As the angle of incidence is increased, the geometry
results in a head wave travelling horizontally in layer 2. From Snell’s Law we can
write:
sin θ c sin 90 o
=
v1 v2
v1
Thus sin θ c = and from geometry we can show that AB = CD = z / cos θ c
v2
Can also show that BC = x - 2z tan θ c
z ( x − 2 z tan θ c ) z
t= + +
v1 cos θ c v2 v1 cos θ c
x 2z 2 z tan θ c
t= + −
v2 v1 cos θ c v2
x 2 zv2 − 2 zv1 sin θ c
t= +
v2 v1v2 cos θ c
x 2 z − 2 z sin 2 θ c
t= +
v2 v1v2 cos θ c
x 2 z cos θ c
t= +
v2 v1
x 2 z v2 − v1
2 2
t= +
v2 v1v2
x constant
t= +
v2
1
• The travel time curve for the refracted wave is a straight line with slope = 1 / v2
• The refracted wave is the first arrival at values of x greater than the cross over
distance (xcross)
• When x = xcrit the refracted and reflected waves are the same
• v2 can be calculated from the slope of the refracted wave on the t-x plot
• The depth of the interface (z) can be found by extrapolating the travel time of the
refracted wave to x = 0 where the travel time is
2 z v 22 − v12
ti =
v1v 2
v1v 2 t i
Rearranging gives z=
2 v 22 − v12
2
C3.2 Seismic refraction – multiple horizontal layers
Direct wave
x
t=
v1
Data analysis
First refraction
x 2 z1 cosθ 12 x
t= + = + t1
v2 v1 v2
Data analysis
● Compute z1 from the intercept time (t1) when x =0 and v1 and v2 are already known
2 z cosθ 12
t1 = 1
v1
Second refraction
v2
Critical refraction occurs as the wave travels from 2 > 3 giving sin θ 23 =
v3
v1
Applying Snells Law at the interface between 1 and 2 gives sin θ 13 = sin θ 23
v2
3
v1 v
Thus sin θ 13 = sin θ 23 = 1
v2 v3
Data analysis
2 z 2 cosθ 23
● From t 2 = we can then calculate z2
v2
4
Multiple layers with velocity increasing with depth
The results for two layers can be generalized for an N-layer model and written as a series.
See the text book for details. Consider the case where there is a uniform increase in
velocity with depth. What will be the form of the ray paths and travel time curves?
Low-velocity layers
● for a head wave to propagate, an increase in velocity from one layer to the next is
needed. If a decrease in velocity occurs, there will be no head wave and refraction will
fail to detect the layer.
Example : If a 3-layer model has a LVZ in the second layer, then interpretation in terms
of 2-layers will give a wrong answer e.g. soil-peat-bedrock. The peat is a low velocity
layer.
Hidden layers
Even if a layer has an increase in velocity, then it possible for the head wave on the upper
surface to never be the first arrival. Again if a 3-layer velocity model has a hidden
second layer, then interpretation in terms of 2-layers will give a wrong answer
5
6
7
C3.3 Seismic refraction – dipping interface
Direct wave
x
t=
v1
● same as in horizontal case considered in C3.2
v1
● Apparent velocity down dip = v 2 d = < v2
sin(θ 12 + γ 1 )
● Note that the apparent velocity down dip is slower than v2 which results in a steeper
slope on the travel time curve.
● Example of layer with γ1 = 5˚ v1 = 1000 m/s and v2 = 3000 m/s is shown on page 9.
8
9
● Note that several velocity models can be found that all fit the data. v1 is uniquely
determined, but there is a trade-off between γ1 and v2 (non-uniqueness again)
10
● On the basis of a single shot recorded in one direction, we cannot determine which
model is the correct one.
● Suppose we fire a second shot at the far end of the geophone array, then these waves
will travel up dip. As the refraction travels to larger offsets (x) the final leg in the upper
layer will become shorter, and the refraction will arrive earlier. This effectively
increases the apparent velocity and reduces the slope of the travel time curve.
t rev = +
v1 v1
v1
● Apparent velocity up dip = v 2u =
sin(θ 12 − γ 1 )
For with γ1 = 5˚ v1 =1000 m/s and v2 = 3000 m/s, what value of v2u is predicted?
● Note that the travel times tAD (forward direction) and tDA (reverse direction) are the
same. This is a phenomena known as reciprocity. Why is this true?
● When forward and reverse profiles are recorded, the correct model (#2) can be seen to
be the only one that fits the travel time data (attached handout)
1 v1 v
θ 12 = [sin −1 ( ) + sin −1 ( 1 )]
2 v2d v 2u
and
1 v1 v
γ 1 = [sin −1 ( ) − sin −1 ( 1 )]
2 v2d v 2u
11
12
C3.3 Seismic refraction – dipping interface - Example
Look at the travel time curve. Which way does the interface dip?
You will need this information to decide which direction is up and which is down!
13
From the travel times compute the following.
v1 = __________ m/s
γ1 = __________degrees
14
C3.4 Seismic refraction – non planar interfaces
● the delay time at the shot is the extra time needed for the wave to travel AB, compared
to the time to travel CB.
● delay time at detector is the extra time needed for wave to travel DE, compared to DF
● can see this from the travel time curve t = x/v2 + δts+ δtd
● with just a single shot, we cannot separately determine δts and δtd
● for a non horizontal interface, with many geophones δtd contains information about
how the depth of the interface (z) varies with position (x’)
● two common methods of finding δtd are the plus-minus method and the generalized
reciprocal method. More details shortly.
15
C3.4.2 Plus-minus interpretation method
Consider the model with two layers and an undulating interface. The refraction profile is
reversed with two shots (S1 and S2) fired into each detector (D). Consider the following
three travel times:
l
(a) The reciprocal time is the time from S1 to S2 t S1S 2 = + δ S1 + δ S 2 = t S 2 S1
v2
x
(b) Forward shot into the detector t S1D = + δ S1 + δ D
v2
(l − x)
(c) Reverse shot into the detector t S2 D = + δ S2 + δ D
v2
Our goal is to find v2 and the delay time at the detector, δD. From the delay time, δD , we
can find the depth of the interface.
16
Minus term to estimate velocity (v2)
(2 x − l )
(b)-(c) will eliminate δD t S1D − t S 2 D = + δ S1 − δ S 2
v2
2x
t S1D − t S 2 D = +C
v2
where C is a constant. A plot of t S D − t S D versus 2x will give a line with slope = 1/v2
1 2
1
Re-arranging to get an equation for δD δ D = (t S D + t S D − t S S )
2 1 2 1 2
17
Example 1 : Synthetic refraction data were generated for a dipping flat interface with the
MATLAB script refract_v3_data_plus_minus.m
The plus-minus technique was then applied using the MATLAB program
plus_minus_ex1.m
The interface dips to the right at 8˚, and is correctly imaged in this example.
18
Example 2: Field data from Kearey and Brooks, p.123
19
C3.4.3 Generalized reciprocal method
The plus-minus method assumes a linear interface between points where the ray leaves
the interface. A more powerful technique if the Generalized reciprocal method in which
pairs of rays are chosen that leave the interface at the same location. More details can be
found in Kearey page 109.
20
Example from Northern Alberta
Seismic refraction was used to determine depth to bedrock at the location where a
pipeline was planned to cross a creek. DC resistivity data were also collected at this
location. Data courtesy of Greg Kovacs and Komex
Note that the direct wave is the only the first arrival at the first 2 geophones. This is
because of a very high velocity contrast between the upper and lower layers. The model
below was derived from the seismic data using the general reciprocal method.
21
3.5.2 Locating water table
3.5.3 Rippability
22
C3.6 Crustal scale seismic refraction surveys
●What crustal velocity and crustal thickness is implied by the attached figure?
●How do the travel time curves differ from those we have considered with a hammer or
explosive source?
●Why can the S-waves be detected, since they are not first arrivals?
23
24
● the direct wave is referred to as Pg and
often travels as a parabola. Why?
t' = t – x/vred
25
● Can you locate the direct water wave
on this section? (P-wave velocity is
1500 m/s)
Technique first used in the 1920’s in the search for salt domes. The higher velocity of the
salt causes earlier arrivals for signals that travel though the salt.
26
3.6.4 Tectonic studies of the continental lithosphere with refraction
● used explosive shots up to 2400 kg with seismic recorders deployed on a profile from
60˚N to 43˚N
● ray tracing used to model the data. Measures the variation in Moho depth and crustal
structure. Figure above shows ray tracing for SAREX shot 5. Note that with a reduction
velocity of 8 km/s, Pn plots as a horizontal line, while the slower Pg has a positive slope.
Gorman, A.R. et al, Deep probe: imaging the roots of western North America, Canadian
Journal of Earth Sciences, 39, 375-398, 2002.
Clowes, R. et al, Crustal velocity structure from SAREX, the Southern Alberta seismic
experiment, Canadian Journal of Earth Sciences, 39, 351-373, 2002.
27
Geophysics 224
2v1 ρ1
T=
v 2 ρ 2 + v1 ρ1
Energy partitioning
1.6 Diffraction
Migration is the solution.
2. Seismic reflection
2.2 Travel time curves for multiple horizontal layers. Dix equation. Multiples.
2.4 Displaying real seismic data. Traces and seismic sections. Reduction velocity.
2.6 Resolution. Vertical limitation from wavelength. Horizontal limitation from size
of Fresnel zones.
2.7 Static corrections and sources of noise. Filtering strategies to reduce noise.
Convolution and deconvolution.
3. Seismic refraction
3.2 Travel time curves for a three layer Earth. Hidden layers (thin or low velocity)
3.3 Travel time curve for a dipping layer. Reversed refraction profile.
MJU 2003
Geophysics 224 D1: Basics of geomagnetism
D1.1 Definitions
μ = magnetic permeability
B=μH
Despite extensive searches, magnetic monopoles do not appear to exist in isolation. They
always occur in pairs of positive and negative monopoles that form dipoles (see D1.4).
http://www.oberlin.edu/physics/catalog/demonstrations/em/magneticmonopole.html
Some theories in particle physics predict that monopoles may be observed in high energy
collisions between subatomic particles:
http://www.aip.org/png/html/monopole.htm
On the atomic level, magnetic behaviour is due to atoms behaving as small magnetic
dipoles. If a uniform magnetic field (H) is applied, there are two possible responses.
Diamagnetic behaviour
Paramagnetic behaviour
B = μ (H+M)
= μ (H+kH)
= μ (1+k) H
= μ μr H
Haematite: anti-ferromagnetic
Magnetite: ferrimagnetic
Fe2O3 893 K
(1) Since temperature in the Earth increases with depth, there exists a depth below which
materials cannot behave as ferromagnetic. Thus only rocks at shallow depths in the
Earth can exhibit remnant magnetization.
(2) As a rock cools from above the Curie temperature, it will acquire a magnetic field that
records the strength and direction of Earth’s field at that time. By using radioactive
dating to find the age of the rock, this gives us a powerful tool to determine how the
Earth’s magnetic field has varied over time (paleomagnetism).
● Measurements of the inclination of the Earth’s magnetic field were described in 1576
by Robert Norman.
1
● William Gilbert developed these ideas and was the first to propose that the Earth itself
was acting as a large magnet with a dipolar magnetic field (i.e. the magnetism did not
originate in the stars or features on the surface at high latitude). He defined the
magnetic poles as places where a dip needle would point vertically downwards.
● Maps of declination were refined, including the 1702 map drawn by Sir Edmund
Halley, who proposed that two north magnetic poles and two south magnetic poles
were needed to produce the observed field. Halley also observed that the declination
was varying over time.
2
D2.1.2 Spatial variations
● The dipole accounts for about 80% of the observed magnetic field. The axis of the
dipole is inclined at 11.5 degrees to the Earth’s rotation axis. The magnetic dipole
moment, P = 7.94 x 1022 Am2 in 2000.
● At any point the magnetic field is defined by the magnetic field elements
Z = vertical component of F
H = horizontal component of F
3
D2.1.3 Temporal variations in the internal magnetic field
● The compass was invented in China, and ancient texts record the variation of
declination from at least AD720. It took a while for Europeans to catch on, and the first
record of temporal changes was in 1635 when Gellibrand noted that declination varies
with time. The declination in London was found to vary significantly over the period
1600-2000.
● Secular variation in the Earth’s magnetic field occurs on many timescales including:
(a) Westward drift: features can be seen to move west over the last century.
4
(b) Investigations of historical records from early navigators and explorers has
extended these records back to the 1600’s (Jackson et al., 2000) and are
displayed as movies at http://geomag.usgs.gov/movies/
(d) Continuous reduction of dipole field since 1600 ( P = 9.4 x 1022 Am2 in
1600, P = 7.94 x 1022 Am2 today)
(e) Complex sequence of magnetic field reversals over the observed geological
record. During a reversal the whole field switches north and south poles.
Between reversals there is evidence that the magnetic dipole axis and the
Earth’s rotation axis are approximately parallel.
The present normal polarity (Brunhes chron) has lasted for 780,000 years.
5
Periods without a reversal for 107-108 years are called superchrons.
Cretaceous normal superchron 118-83 Ma
Permo-Carboniferous (Kiaman) reverse superchron 312-262 Ma
The reduction in the main field over the last 400 years could indicate we are
approaching a reversal. How might reversals affect life on Earth?
● High temperatures inside the Earth (well above Curie temperature) exclude the
possibility of remnant magnetization generating the magnetic field. There is no large
bar magnet inside the Earth!
● The secular variation, and alignment of dipole with rotation axis, suggest that the
magnetic field originates in the relatively rapid fluid motion in a part of the Earth with a
6
high electrical conductivity. This only leaves the outer core (composed of liquid iron)
as the place where the magnetic field is generated.
The magnetic field measured at the surface of the Earth is due to sources inside and
outside the Earth. The external component is generated in the atmosphere and
magnetosphere.
●The solar wind (a stream of H and He ions) is deflected by the Earths internal magnetic
field to create the magnetosphere.
7
●The interactions between the solar wind and the Earth’s magnetic field are very
complex. Temporal changes in the solar wind, due to sunspots, solar flares and coronal
mass ejections can produce a change in the magnetic field at the surface of the Earth.
● From 50-1500 km above the Earth’s surface is the ionosphere, a region of plasma with
high electrical conductivity. Changing magnetic fields from the magnetosphere can
induce large electric currents in the ionosphere. Changes in these currents produce large
changes in the magnetic field measured at the Earth’s surface.
● When the solar wind is in a steady state, the Earth’s magnetic field shows a daily
variation that is due to the Earth turning within the current systems of the
magnetosphere and ionosphere. The typical variation is called the solar quiet day
8
variation (Sq). The amplitude is typical 10-20 nT and varies with latitude. Clearly seen
in time series above.
● A much smaller variation is seen every 25 days and is caused by the orbit of the moon.
● When the solar wind is active, the Earth’s magnetic field is said to be disturbed.
Magnetic storms occur when the current systems change over a period of several days
and the field at the Earth’s surface can change by 100’s of nanotesla. These changes are
largest beneath major ionospheric current systems. A small substorm can be seen in the
middle of the time series plotted above.
● Smaller magnetic field disturbances are classified as substorms and bays and have
timescales of several hours.
● Solar activity is characterized by an 11 year cycle and we have just passed the
maximum. Maximum solar activity results in high levels of activity in the Earth’s
external magnetic field and frequent magnetic storms and strong auroral displays.
References
Jackson, A., Jonkers, A. R. T. & Walker, M. R., 2000. Four centuries of geomagnetic
secular variation from historical records, Phil. Trans. R. Soc. London, A 358, 957-990.
9
D2.3 Comparison of the Earth’s gravitational and magnetic fields
Temporal variations with Tidal signals (< 0.5 mgal) Diurnal Sq variation (50 nT)
external origin Magnetic storms (100-1000nT)
11 year sunspot cycle
10
224D3: Magnetic anomalies produced by simple geological structures
Remember that objects can acquire both induced and remnant magnetization.
Induced magnetization will disappear when the applied magnetic field is removed.
In the following examples, we will consider just induced magnetization. This will be in a
direction parallel to the Earth’s magnetic field. Remnant magnetization can be in any
direction.
● in the presence of the Earth’s magnetic field, the dike develops an induced magnetic
moment
● if the structure extends to depth, then the lower monopole can be ignored since for a
monopole, Br = μm/r2
● The cylinder has an induced magnetization with negative monopoles on upper surface and
positive monopoles on the lower surface. One can show that net effect is a dipole at centre
of cylinder
● for a dipole, the magnetic field falls away as 1/r3
● note that the magnetic field anomaly is different at the magnetic north pole and equator
3.3 Sphere
● As for cylinder, the sphere develops an induced magnetic moment, which is equivalent to a
dipole at the centre.
From http://www.gif.ubc.ca
3.4 Thin sheet or slab
● an infinite sheet develops negative poles (South poles) on the upper surface and positive
poles (North) poles on the lower surface. If the sheet is thin (vertical direction), then these
cancel to give no net magnetic field at the surface.
● when finite in horizontal distance, the poles near the edge do not cancel and a positive-
negative anomaly is observed.
D4.1 Instrumentation
● Measures the component of magnetic field along the axis of the ferrite cores
● this frequency is measured by the weak radio signal emitted by the precessing protons
● These magnetometers use changes in the frequency splitting of optical spectral lines
of elements such as rubidium, caesium or potassium to measure the magnetic field. This
is governed by the Zeeman effect.
● Errors in survey limited by positioning errors rather than magnetic field measurements.
● By placing one magnetometer above another, the vertical gradient of the magnetic field
(dF/dz) can be measured. The high precision of modern instruments means the spacing
only needs to be 1-2 m. This has several advantages
● By computing (dF/dz)/F, temporal variations in the external magnetic field can be
effectively removed from the data. This is illustrated in the data example below (panel
f), and this provides an alternative way to remove time variations (if the base station
fails!)
● dF/dz can be more sensitive to dipole anomalies in the Earth than long wavelength
features in the regional magnetic field.
● operator must be liberated of all metal objects (especially their keys and money!)
● locate profiles with surveying, or with an instrument with integrated GPS (above)
Spatial corrections
● It is possible to correct magnetic data for latitude and elevation but these corrections
are generally much smaller than magnetic field anomalies. They are only important if
the survey has a large spatial extent.
● Topography is important, such as when measurements are made in a narrow valley or
canyon.
Time corrections
● A simple magnetic survey could use a single magnetometer that is carried along a
profile. Since the magnetic field of Earth changes with time, it is unclear if changes in
magnetic field are due to (1) temporal changes or (2) subsurface structure.
●To separate effects (1) and (2) two magnetometers must be used in fieldwork. An
example of this procedure is illustrated below. One magnetometer is placed at a fixed
location (the base station) and the other is carried along the profiles (b). Data recorded
here are shown in red and measurements at the moving station are shown in blue (a).
● Variations at the base station instrument are subtracted from those on the
moving instrument to determine variations along the profiles (c)
● Note that the three crossings of the target are virtually identical when the time
variations are removed (d,f)
● Profiles flown with line spacing 100 m to several kilometers. Elevation 100-
1000 m.
● Cross lines used as internal check that temporal variations have been
consistently removed.
● Flight path can be located with radar, altimeter or by taking aerial photos. GPS
now makes this much more accurate and simpler.
● Draped surveys flown at constant elevation above the surface (need helicopter)
● With total magnetic field measurements the motion of the aircraft or bird is not
serious.
● High resolution aeromagnetic (HRAM) surveys are now the standard field
technique and they use a dense line spacing and low flight elevation to give
detailed information about subtle, sub-nanotesla magnetic field anomalies. See
examples from Geological Survey of Finland at
http://www.gsf.fi/aerogeo/eng0.htm
D4.2.3 Marine magnetic surveys
● Marine surveying is slower and more expensive than airborne surveying, so not
widely used in commercial offshore exploration.
● However, magnetometers are routinely towed behind survey ships and have
revealed many details of the structure of ocean basins (e.g. magnetic stripes
produced by seafloor spreading)
x½ = 0.766 d
Other approximate rules can be derived but they must be used with considerable
caution (see Kearey p. 107 Interpretation of magnetic anomalies).
● The measured data is shown in (b) with both positive and negative anomalies.
Reduction to the pole (c) shows a simpler set of anomalies that reveal the true shape
of the magnetized body.
5.3 Upward continuation
● mathematical method for predicting magnetic field at a different elevation from that at
which the data were collected.
● can allow data to be filtered, since at higher elevation the short wavelength features will be
weaker than at ground level.
● Aeromagnetic data is often used for finding the lateral extent of structures or for
locating anomalies that are then investigated in more detail during follow up on the
ground.
● Anomalies are smaller when measured at flight elevation than at ground level. This
can be both good and bad.
● Generally the user alters the model until the predicted and measured field data agree
● The air in a cave has k =0. If the host rock has a non-zero magnetic susceptibility, then
a magnetic anomaly will be observed at the surface. What will be the sign and shape of
the anomaly over a cylindrical lava tube in a basalt lava flow?
● Ground-based magnetic surveys have been used in Hawaii to locate lava tubes prior to
developing agriculture or construction.
D6.2 Archaeology
● ore bodies can have a high magnetic susceptibility and may exhibit both induced and
remnant magnetization. These anomalies can easily be detected at aircraft elevations
and thus aeromagnetic exploration is a good reconnaissance tool for minerals.
● Example from Quebec (Kearey figure
7-25). The ore body is also electrically
conductive, and is well defined by two
geophysical surveys (electrical and
magnetic)
Kimberlite usually contains more magnetite than the host rock in the Slave Province and
produces a positive magnetic anomaly. However a negative magnetic anomaly occurs if:
(a) the host rock is more magnetic (contains more magnetite) or (b) the pipe has a
(reversed) remnant magnetization. If these effects cancel, the pipe can produce a weak, or
near zero, anomaly.
Extra information can be gained from airborne electromagnetic (EM) surveys (more
details in Geophysics 424 next year). This is essentially a way of measuring shallow
electrical resistivity from a moving aircraft or helicopter. Two factors make this a viable
exploration technique:
(1) Kimberlite has a lower electrical resistivity than the host rock
(2) Kimberlites weather and produce clay, with a low resistivity. If eroded, a lake
forms, also producing a lower resistivity.
“Because kimberlite pipes exhibit variable anomalies on both electromagnetic and magnetic
data, the best approach to mapping them is to simultaneously collect EM and magnetic data
from a low flying platform. The DIGHEM system collects both data sets from a sensor at 30m
altitude, sampling about every 3m.” http://www.fugroairborne.com.au
(a) Point Lake Kimberlite, NWT
The Ekati Diamond Mine is exploiting five economic kimberlite pipes in the Lac de Gras
region of the NWT. The pipes are named Panda, Koala, Misery, Fox and Leslie. BHP
Billiton mining operations at the Koala Pipe are shown on the right.
Magnetic data: Left panel shows that the Grizzly pipe has a negative magnetic anomaly
and Panda has a small positive anomaly. Koala and Fox pipes show weak anomalies.
Airborne EM data: The apparent resistivity map (centre), calculated from the 7200Hz
coplanar data of the DIGHEM survey, clearly shows the economic pipes in this data
block. The Fox pipe (south west corner) has the most distinct anomaly, and coincides
almost exactly to the overlying lake. The Koala and Panda pipes give clear anomalies,
and are also underneath lakes.
http://www.fugroairborne.com.au
http://www.mining-technology.com/projects/ekati
http://ekati.bhpbilliton.com
(c) Fort à la Corne kimberlites, Saskatchewan
www.fugroairborne.com
80 m elevation 30 m elevation
D6.6 Regional crustal structure
Even if the origin of the magnetization is not resolved, the character of the
aeromagnetic anomalies can be used to determine the extent of a geological province.
The direction of the magnetic anomalies can also reveal the geological strike of these
rocks.
Tibetan Plateau: Several geophysical techniques have suggested that unusually high
crustal temperatures exist beneath the Tibetan Plateau. How will this alter the
susceptibility of the crustal rocks? Good coverage with magnetic data in Tibet is hard to
obtain on the ground (no roads) and aeromagnetic data coverage is not widely
available.
In the 1990’s a low orbit satellite (MAGSAT) was used to map the Earth’s magnetic
field. (Why in low orbit?) Analysis of these data by Alsdorf and Nelson (1999) reveal a
pronounced magnetic low over Tibet? Can this magnetic low be explained on the basis
of high crustal temperatures and partial melting?
D6.7 Dike location
Southern Alberta
The goal of the magnetic survey was to determine if a dike connects Mt. Katmai and
Novarupta. Both erupted in the 1912 eruption, but the plumbing of this volcanic system is still
unknown.
● A summary of the state-of-the-art in oil and gas exploration can be found in Gibson and
Millegan, 1998. While oil and gas are not magnetic, useful information can be obtained
from magnetic exploration, since it can define the local geology.
● Magnetic data can be used to define the depth of magnetic basement. The magnetic
basement is usually the crystalline basement and knowledge of the depth is useful in
hydrocarbon exploration since it gives information about the overlying sedimentary
rocks (depth, location of faults etc).
Example: East China Sea, Okuma et al, in Gibson and Millegan, 1998, pages 59-62
● The Bowser Basin in British Columbia is currently being evaluated for hydrocarbon
potential. Note that spatial extent is defined from the magnetic data. Magnetic anomalies
have longer wavelengths within the basin (why?). Magnetic field data courtesy of Carmel
Lowe, Natural Resources Canada.
● As oil seeps to the surface from a trap, it can alter the rocks through which it flows.
This can change the near-surface magnetic susceptibility. Attempts to locate these
altered regions have been made with magnetics, and other airborne geophysical data.
Example: ALTREX method, J. Rowe et al, Gibson and Millegan, 1998, pages 124-129
● Generally sedimentary rocks have low magnetic susceptibilities and do not exhibit a
strong induced or remnant magnetization. Sedimentary rocks can develop a weak
remnant magnetization during deposition.
Example: ANWR, North Slope of Alaska, Phillips et al, in Gibson and Millegan, 1998,
pages 130-134
Example : Northern Canadian cordillera, shown in Nabighian et al., (2005)
In this case, the magnetic data can be used to map faults, in between the widely spaced
seismic reflection lines.
●As new oceanic crust is formed at mid-ocean ridges, it is magnetized in the direction of
the Earth’s magnetic field. For basalt, this remnant magnetization is usually stronger
than the induced magnetization.
●As the plates move away from the ridge, they carry a record of the magnetic field
polarity at the time the crust was formed.
●These so-called magnetic stripes were a key piece of evidence in the development of
plate tectonic theory. By combining the history of the magnetic field reversals with
radioactive dating, the age of oceanic crust can be determined from the magnetic field
anomalies.
References
Arzate, J.A., L. Flores, R. Chavez, L. Barba and L. Manzanilla, Magnetic prospecting for
tunnels and caves in Teotihuacan, Mexico, in Geotechnical and Environmental
Geophysics, Volume 3, SEG Investigations in Geophysics, No. 5, p. 1-30, 1990.
Gibson, R.I., and P.S. Millegan, Geologic applications of gravity and magnetics: case
histories, Society of Exploration Geophysics, 1998.
Hogg, H., Munro, S., The aeromagnetic discovery of kimberlites and sulphides at depths
up to 200m, Expanded Abstract SEG Annual Meeting, Calgary, 2000.
Pilkington, M, W.F. Miles, G.M. Ross and W.R. Roest, Potential field signature of buried
PreCambrian basement in the Western Canada Sedimentary Basin, Canadian Journal
of Earth Sciences, 37, 1453-1471, 2000.
Roberts, R.L., W.J. Hinze and D.I. Leap, Data enhancement procedures on magnetic data
from landfill investigations, in Geotechnical and environmental geophysics, Volume 2,
Environmental and groundwater, SEG Investigations in Geophysics, No. 5, 261-266,
1990.
(a) Sketch the variation in gz that would be observed. Calculate the minimum
value of gz in mgal. Also calculate the half-width of the curve.
(c) The gravity data has standard errors of 0.01 mgal. What is the maximum
depth at which the tunnel will be detectable on the basis of surface
gravity measurements? Explain your answer.
2. Gravity data was collected at three stations (A,B and C) in Southern Alberta.
These stations are on an East-West profile with latitude of 50 ˚N. Station A is on
the prairies, while B and C are further west in the Rocky Mountains.
On the first day of the survey, the gravity crew make the measurements listed in
the table below.
(a) Compute the difference in g between the CGSN station and station A. This
will require that you correct for the drift of the gravimeter. Plotting a graph of
gravimeter reading as a function of time will help you do this. You may
assume that the drift is linear with time.
(b) Use your answer to compute the absolute value of gravity at station A.
(c) What two factors cause the reading of a gravimeter to change when it is kept
in a fixed location?
3. Gravity measurements were also made at stations B and C. The table below
shows these measurements. Compute the Bouguer anomaly using the following
steps. Please write your answers in the Table below.
(a) Use the GRS67 formula from your notes to compute the predicted value
of gravity (gnorm) at each station. Compute the gravity anomaly (ganom) by
subtracting gnorm from gabs
(b) Compute the Free air correction (CFA) for each station.
(c) Apply the Free air correction to ganom to obtain the Free Air Anomaly at
each station
(d) Compute the Bouguer correction (CB) at each station, assuming a density
of 2670 kg m-3.
(e) Apply the Bouguer correction to obtain the Bouguer anomaly (ΔgB) at
each station.
(f) What does the change in Bouguer anomaly from A to C tell us about the
structure of the Rocky Mountains?
You may assume that gravitational attraction of an infinite slab with thickness Δz,
density ρ is given by gz = 2πGρΔz
and that the change in gravity due to a change in elevation of Δz metres is given by
Δg = 0.3 Δz (mgals)
1. The figure shows a shot gather that was recorded in a location where the subsurface
consists of a series of horizontal layers. The travel times are listed on the attached table.
You may answer the following by using graph paper or Excel. A copy of the travel times
will be distributed by e-mail.
(a) For the first reflection plot a graph of normal moveout (NMO) as a function of
offset squared (x2). Remember that the equation for normal moveout is an
approximation that is valid for x << z1. Up to what value of x is the approximation
valid?
(b) Compute values for the thickness (z1) and velocity (v1) of the first layer.
(c) To confirm your answer, plot t2 versus x2 and compute the thickness (z1) and
velocity (v1) of the first layer.
(d) For reflections 2,3 and 4, list the travel times at zero offset (t2, t3 and t4 ).
(e) Plot normal moveout as a function of x2 for reflections 2,3 and 4. Using the
normal moveout equation
x2
Δt = t − t n = 2
2t nVrms ,n
compute the r.m.s. velocities (Vrms, 2, Vrms, 3 and Vrms, 4). Be careful to only use travel
times for which the equation is valid.
(f) Use the Dix equation to compute v2, v3 and v4
1
⎡Vrms
2
, n t n − Vrms , n −1t n −1
2
⎤2
vn = ⎢ ⎥
⎣⎢ t n − t n −1 ⎦⎥
(g) From the travel times at normal incidence (t2, t3 and t4) compute the depths of the
interfaces below the surface.
(h) What does the shot gather tell us about the properties of the 5th layer?
______________________________________
x(m) ref1(ms) ref2(ms) ref3(ms) ref4(ms)
______________________________________
-200.000 235.36 477.26 723.17 966.07
-190.000 233.08 476.55 722.84 965.89
-180.000 230.87 475.86 722.53 965.72
-170.000 228.76 475.21 722.24 965.56
-160.000 226.75 474.59 721.96 965.41
-150.000 224.85 474.01 721.69 965.27
-140.000 223.06 473.47 721.45 965.13
-130.000 221.37 472.96 721.22 965.01
-120.000 219.80 472.50 721.01 964.89
-110.000 218.35 472.06 720.81 964.79
-100.000 217.01 471.67 720.63 964.69
-90.000 215.80 471.31 720.47 964.60
-80.000 214.70 470.99 720.33 964.53
-70.000 213.73 470.71 720.20 964.46
-60.000 212.89 470.47 720.09 964.40
-50.000 212.17 470.26 720.00 964.35
-40.000 211.58 470.09 719.92 964.30
-30.000 211.12 469.96 719.86 964.27
-20.000 210.79 469.86 719.82 964.25
-10.000 210.59 469.80 719.79 964.23
0.000 210.53 469.79 719.79 964.23
10.000 210.59 469.80 719.79 964.23
20.000 210.79 469.86 719.82 964.25
30.000 211.12 469.95 719.86 964.27
40.000 211.57 470.09 719.92 964.30
50.000 212.16 470.26 720.00 964.35
60.000 212.88 470.46 720.09 964.40
70.000 213.72 470.71 720.20 964.46
80.000 214.69 470.99 720.33 964.52
90.000 215.78 471.31 720.47 964.60
100.000 217.00 471.67 720.63 964.69
110.000 218.33 472.06 720.81 964.79
120.000 219.79 472.49 721.01 964.89
130.000 221.36 472.96 721.22 965.01
140.000 223.04 473.47 721.45 965.13
150.000 224.83 474.01 721.69 965.27
160.000 226.73 474.59 721.95 965.41
170.000 228.74 475.20 722.23 965.56
180.000 230.85 475.85 722.53 965.72
190.000 233.06 476.54 722.84 965.89
200.000 235.36 477.26 723.17 966.07
_______________________________________________________
Please contact Ersan or myself if you have questions or simply need some inspiration to
get started.
Geophysics 224 Assignment 3 – Seismic reflection and refraction
Question 1
(a) Compute the velocity at which the direct wave travels. Compute maximum and
minimum velocities consistent with the measured data (i.e. estimate an
uncertainty in v1).
(b) Compute the velocity (v2) for the refracted arrival. Again, compute maximum
and minimum velocities consistent with data (i.e. estimate an uncertainty in v2).
(c) What is the depth of the layer with velocity v2? What are the maximum and
minimum values that are possible?
(d) A third layer is present in the study area with v3 = 3000 m/s, but is not observed
on the seismogram above. What is the minimum depth at which it could be
located?
Question 2
A 48-channel streamer has hydrophones spaced 10 m apart. As the survey ship moves
forward, a shot is fired every 40 m.
(b) To obtain 24-fold coverage, what must the new shot interval be?
Question 4
The figure above shows two depth sections through the subsurface. Sketch the shape of
the zero-offset time section that would be recorded by a profile that crosses the area.
There is no need to compute exact travel times, but include the effects of diffractions etc
in a qualitative manner. Velocities are in m/s and show relative velocities.
Question 5