Seismic refraction introduction

Seismic methods involve generating acoustic energy in the form of pressure or

shear waves that propagate into the ground. Sensors (geophones) are installed in
lines on the surface, and they record the faint ground motions caused by the
acoustic energy (seismic waves) that returns from boundaries where the elastic
properties of subsurface materials change. At such boundaries there are two
fundamental physical effects. "Reflection" means some of the energy impinging
on the interface is returned or "bounced" back towards the surface - light
reflecting from a semi-transparent mirror is a good analogy. "Refraction" is less intuitive, but the
concept is recognizable in images of things under water. When light arrives at a boundary, the portion
that passes through is bent so it's direction of propagation changes. In fact, seismic refraction
involves detecting energy that has been bent so that it travels along the interface between two
subsurface materials, before returning to the surface where it's effects can be measured. The
phenomenon is not intuitive, but it is explained in the "Ray paths" section of this chapter.
There are far more professional geophysicists employed in the seismic industry than any other branch
of applied geophysics. Most are working at seismic reflection in it's many forms for the oil/gas
exploration industries. Some mineral exploration companies have recently made 3D seismic
reflection surveying common practice for helping delineate known ore bodies, but it is not
universally used. The geotechnical and environmental industries use seismic refraction regularly for
efficient assessment of thickness and mechanical properties in the top few 10s of metres. Seismic
reflection has more recently become economical for shallow work (from roughly 5 metres depth,
down to 100's of metres).

A small refraction survey on mine tailings. There are 12 geophones, a hammer is used on the side
of a metal block (the operator is standing on it) to produce shear waves.
The recording device, it's computer and battery are just left of centre.

In this chapter we will discuss only one form of seismic surveying - general purpose seismic
refraction. Seismic refraction most useful when the velocity of seismic signals in each layer increases
with depth. Therefore, where higher velocity (e.g. clay) layers may overlie lower velocity (e.g. sand
or gravel) layers, seismic refraction may yield incorrect results. In addition, seismic refraction
requires geophone arrays with lengths of approximately 4 to 5 times the depth to the layer of interest
(for example the top of bedrock). Therefore seismic refraction is commonly limited to mapping
layers to depths less than 30-50 metres. Greater depths are possible, but the required array lengths
may exceed site dimensions, and the shot energy required to transmit seismic arrivals for the required
distances may necessitate the use of large explosive charges.

Fundamental physics for seismology

If we strike the earth with a hammer, an explosive charge, or an earthquake, the material of the earth
near the source becomes compressed (or extended) and twisted. Earth materials are elastic and these
deformations propagate away from the source. The speed of propagation and the type of motion
propagated depends upon the elastic properties of the material, namely
1. density:
2. bulk modulus:
3. shear modulus:

(compressibility)
(twistability)

By knowing the time at which the source occurred, and by measuring the travel time that it takes for
the various waves to arrive at a receiver, it is possible to obtain information about the velocity and
density of the material. This information is then used to provide information about structure and rock
type.

Applications of seismic surveying

1.
2.
3.
4.
5.

delineation of sedimentary layers

determination of depth to bedrock, and topography of the bedrock horizon
locating faults
large scale structure of earth (using earthquakes)
characterization of near surface material properties.

Stress
When an external force is applied to a body, there are balanced internal forces set up within the body.
Stress is a measure of these internal forces. Units are "force per unit area" or N/m 2. There are two
types of stress:
Stress can be normal to
the surface it is applied
on, and hence
compressive or tensile
(top figures).
Stress can be parallel to
the surface. This is shear
stress (bottom figure).

shear

Strain
If a body is stressed it will undergo
a change in size and shape. The
body is strained. If the strains are
small then the body will recover its
original shape when the stress is
removed. This is elastic strain. If
the strain is too great so that the
yield strength is exceeded, then
the body deforms plastically and
may ultimately fracture. Plastic
strain is not recoverable.
The figure plots how strain
(horizontal axis) accumulates as
stress (vertical axis) is applied.
While behaviour is elastic, increasing or decreasing the stress does nothing to change the material.
Once stress is sufficiant to make material behave platically, reducing the stress rsults in reduced strain
along a different path on the graph. Once the fracture point is reached, the strain is released by
breaking.

Elastic Moduli
The linear relationship between stress and strain is known as Hooke's Law and is specified by five
elastic moduli or elastic constants which express the ratio of a particular stress to a resultant strain.
Four important constants are explained briefly next. Elastic constants are defined in Wikedia, and in
the Schlumberger "Oilfield Glossary".
Young's Modulus
Consider a rod (figure right) of length l,
cross-sectional area A, to which a
uniform force F is applied to each end (a
tensional force in this case):
Young's modulus, E, is given
by

Poisson's Ratio
The rod will also contract in radius (because it is being extended). The radial strain is
Poisson's ratio is given by
Bulk modulus
Consider an initial volume V which is subjected to a hydrostatic pressure
(force per unit area is the same in each direction). Let V be the change in
volume.
The bulk modulus is

r/r.

Shear modulus
If we apply a force parallel to a surface the force per unit area is known as the
shear stress . Consider a rectangular cube. Application of shear stresses to the
top and bottom interfaces produce a strain proportional to tan. The shear
modulus is

Elastic model for the Earth rocks

Earth materials are elastic. The easiest way to obtain intuition is to think
about rocks as being made up of a set of connected springs. Apply a force
(stress) to any part and you will eventually get motion elsewhere. There
are two general types of waves that can travel in the elastic material.
1. Body waves which travel through materials
2. Surface waves which travel only along boundaries between
materials such as the ground / air boundary.
Specific types of waves within these general types are described below.
Each wave type travels with a velocity that depends upon the elastic
properties.

Body Waves
Compressional waves (P waves)
propagate by compression and rarefaction,
and the velocity of such waves,Vpis given in
terms of elastic moduli introduced above
and density , by the equation shown here.

Shear waves (S waves)

propagate by a pure shear strain
perpendicular to the wave propagation.
There are two planes of polarization.
SV and SH waves travel at the same
speed, Vs, which is given in terms of
elastic moduli introduced above and
density , by the equation to the right.

Note that if

= 0 then Vs = 0. This tells us that shear waves do not travel in a liquid.

Seismic reflection and refraction surveying is usually carried out by observing how P-waves travel,
reflect, and refract within the geologic layers near the surface. We will see that the energy is
generated by an artificial source. S-waves can also be used, but they are a little harder to generate
artificially (compared to P-waves), and require special ground motion sensors. However, since
Vs < Vp, there are situations where it is beneficial to use S-waves instead of P-waves.

Surface waves
Rayleigh waves propagate along
a free surface or on the boundary
between two materials. Particle
motion is a retrograde ellipse,
and in the same plane as wave
energy propagation. The
amplitude of particle motion
decays exponentially with depth.
Rayleigh wave speed VR < VS.
Large earthquakes can generate Rayleigh waves that circumnavigate the globe. This provides information
about the velocity structure in the upper few hundred kilometers of the earth.
Love waves exist in a surface
layer when the shear wave
velocity of the upper layer is less
than the shear wave velocity of
the lower layer. The waves are
trapped in the upper layer and the
particle motion is parallel to the
free surface and perpendicular to
the direction of propagation.
VS1 < VLove < VS2
Both Love waves and Rayleigh waves are dispersive. That is, different frequency components travel
at different speeds. So the wave changes shape as it travels. Also, the dispersion can be used to
provide information about the velocity structure in the upper region of the earth. For shallow work, it
is possible to generate surface waves artificially, and then observe the waves at a series of locations at
increasing distances from the source. This type of field work is sometimes called multi-channel
analysis of surface waves or MASW. This is usually considered an "advanced" topic in applied
geophysics.

Waves and Rays

A wavefront indicates the
locations at which the phase of
the wave has the same value. For
example, visualize the peaks (or
troughs) of water ripples after a
rock has been thrown in. The
direction of propagation of the
energy is normal to the
wavefront. Seismic rays are
imaginary lines perpendicular to
the wavefront that indicate the
path along which the wavefront
is travelling. Rays are not
physical entities. They exist only
to illustrate where the energy
travels.

Wave Velocity and Particle Velocity

Seismic waves typically travel in the ground at
2-7 km/s. This is the velocity at which the energy
moves, not the particles themselves. For
comparison, sound travels in air at approximately
0.33 km/s. The wave energy can be recorded
many kilometers from the source even if the
source is small. The velocity and displacements
of individual particles in the rocks are however
very small; typical particle speeds are 10 -8 m/s
and typical ground displacements are 10 -10 m.

P-wave velocity of earth materials

Some characteristics of P-wave velocities are:
1. Vp increases with confining pressure;
2. sandstones and shales show a systematic
increase in Vp with depth of burial and
age (progressive compaction and
cementation);
3. For a wide range of rocks there is an
approximate relationship between density
and Vp;
4. The presence of gas in sedimentary rocks
reduces the elastic moduli, Poisson's
ratio, and the ratio Vp / Vs.
P-wave velocities of some representative
geological materials are listed in the adjacent
table.

Attenuation
The amplitude of seismic waves falls off with
distance from the source. There are two primary
reasons:
Geometrical spreading - that is, energy
falls off as 1/r2 and hence the amplitude
falls of as 1/r.
Earth materials are not perfectly elastic.
Some frictional heating occurs as the
waves propagate through the earth. This is
often described as "absorption" and the
absorption coefficient expresses the
proportion of energy lost as the wave
travels a distance of one wavelength.
The figure here shows the progressive change of
shape of an original spike pulse during its
propagation through the ground due to the effects
of absorption (After Anstey 1977.)

Vp (km/s)

The spike's shape changes as well as experiencing reduced amplited. This is because the different
frequencies making up the pulse decay at different rates - in fact, higher frequencies decay
more rapidely than lower frequencies. This is easily observed on earthquake signals that have
been recorded at different locations. As noted above in the context of surface waves, such
frequency dependent behaviour is called dispersion.

Sources and sensors

Seismic Sources
For seismic surveying purposes there are a number of generic goals. Not all of these
objectives can be met perfectly, but sources are designed to meet them as effectively
as possible.
1. High energy is required - the stronger the source, the better the signal to
noise ratio.
2. High frequency is preferred - the higher the frequency, the shorter the
seismic pulse and hence the better the resolution.
3. Repeatability is important for reliable measuremnets - we often want to stack the signals to
improve the signal to noise ratio so repeat "shots" must all have identical source energy
characteristics.
4. A well defined source signature is desirable. By this we mean that the exact pattern of the
wave energy should be known. If it is, we can use deconvolution to find the exact times of
reflections &/or refractions. Recall that deconvolution is a processing step that enhances
reflections and/or refractions.

Land Sources:
There are many ways of making the earth move suddenly so that seismic signals are generated, and
their traveltimes can be recorded.
1. Explosives: To get good coupling the explosive should be detonated in a shallow shot hole.
Explosives are generally not repeatable and may not be environmentally friendly.
2. Buffalo guns are common for small scale engineering / environmental work. See the figure
below for a diagram of this type of source.
3. A simple heavy hammer on a baseplate is inexpensive and safe for small jobs (photo above
and on the introductory page). When shear waves are required, a baseplate with ribs that
penetrate the ground is used, and it is struck on it's side. This is one of the few simple ways of
producing shear waves.
4. Finally, vibroseis is an expensive source is used mainly for reflection surveying. The input is
a frequency varying sinusoid (called a chirp signal). Multiple vibrators are often used. There
are systems ranging in size from handheld units like pavement packers through to massive
special purpose trucks that can be used singly or in coupled groups to produce signals that
penetrate beneath the crust into the mantle of the Earth.

Diagram of a "Buffalo gun" seismic source. The 3/4 inch pipe with the shell installed is lowered into
a borehole of the same size, then the steel drop rod is dropped through the tee so the firing pin strikes
the 12 guage shell. This creates the seismic energy.

Water Sources
When seismic surveying is needed to characterize ocean, lake or river bed sediments, special sources
of seismic energy are required. A few of these are listed here:
1. Air gun: Compressed air is released to the water, forming a bubble. As the bubble rises to the
surface it pulsates creating an extended source signature.
2. Water gun: Water is ejected by a piston and there is a cavity behind the water jet. There is an
implosion but no bubbles because no extra air has been introduced.
3. Sparkers: The electric charge held on a large bank of capacitors is discharged directly into the
water. This Ionizes the water and creates a plasma pulse. Voltages of ~4kV, and currents of
~200A are involved.
4. Boomers: These involve the discharge of capacitors through a coil. The changing voltage
impinges upon an aluminum plate and the interaction of the magnetic fields drives the plate
backward.

Seismic Detectors
How are the tiny ground motions (see "Wave and Particle Velocity" in
Fundamentals) detected? A seismic detector is an electromechanical device.
It responds to a mechanical input such as physical motion or pressure, and
outputs and electrical signal.
On land the instruments are called seismometers or geophones. Once the
sensor's spike (right) is planted into the ground, the geophone case moves
with the ground while a heavy magnetic inside the case stays stationary
owing to it's own inertia. The relative motion between a coil wrapped
around the magnet, and the magnetic field, sets up an voltage in the coil.
This voltage is passed along the wire to the recorder where it is converted to a digital signal and stored.
Geophones are sensitive to motion only along the axis of the coil. Vertical ground motion is best
detected by a orienting the coil vertically to build a vertical geophone. It is also possible to mount the
spring/mass system horizontally. A combination of several sensors in different orientations allows
ground motion in all three directions to be measured.
Seismic signal detectors for water-borne deployment are called hydrophones. These are generally
ceramic piezoelectric elements which produce an output voltage proportional to the pressure
variations associated with the passage of a compressional wave through the water. At sea, these are
often carried in a neutrally buoyant cable (seismic streamer). Recall that shear waves can not travel in
liquids.

Recording equipment
Seismic data consists of records of ground motion at each
geophone location. Each record shows how ground moved
as a function of time, starting at the moment the source
energy was generated. The records are usually a few
hundred milliseconds long, and are digitized (ie sampled) at
a sampling rate of perhaps several samples per millisecond.
This is not very different from the sampling rate for
digitized music. Therefore, recording equipment must have
the following capabilities:

1. One electrical signal to digital number converter for each geophone, which can work at "audio"
frequencies.
2. A synchronizing facility so that digitizing can start at
the same time as energy is initiated at the source.
3. A computer to manage the many streams of numbers coming from all the geophones.
4. The computer must be able to "replay" or plot the seismograms (signals from each geophone)
so that data quality can be checked visually. Some units will have a small strip-paper printer.
5. Since there is a computer, most systems include built in software to carry out intial
interpretations of refraction data.
The image to the right is a "SmartSeis" seismograph which can record up to 24 channels,
manufactured by Geometrics. A good place to see specifications of engineering scale seismic
equipment is http://www.terraplus.ca/products/seismic/ (Terraplus is an equipment rental company,
based out of Richmond Hill, Ontario).
Below is an image showing simulated exploration of Mars at the Haughton-Mars Project, an
international, interdisciplinary field research project. The Mars analogue field site is in the
Haughton crater, a meteorite impact crater, on Devon Island in Canada's high arctic. A line of
geophones is being planted, the seismograph is nearby. The goal is to characterize the layered
structure under the field site.

Raw refraction data

Energy in the ground
Seismic body waves travel away from the source, into the ground. The energy spreads in all possible
directions away from the source, so wavefronts expand as hemispheres under a source point, if the
ground is uniform. The energy will arrive at geophones at times depending upon the velocity of the
waves in the subsurface materials. When energy reaches a change in material, some energy will be
reflected from the interface, and some will pass through it. The geometry of this situation is shown in
the next figure.

If seismic signals travel at higher velocity in the lower layer, then some of the seismic energy travels
along the interface, returning to the surface as a "head wave" along a wave front similar to the bow
wave of a ship (figure below). These are refracted waves, and for geophones a long way from the
shot point, they represent the first arrival of seismic energy. In other words, because head waves
travel along the interface at the velocity of the "faster" material, they eventually overtake the direct
waves (green in the figure below) travelling in the slower surficial materials.

Of course energy is both reflected and refracted, so ground motion detected at a geophone is a
caused by the combination of direct, refracted and reflected energy arriving at the geophone's
location. The different types of energy are distinguishable only because they have travelled along
different pathways. Refraction surveying takes advantage of the fact that refracted waves arrive
before reflected energy, so long as the geophone is at a great enough distance from the shot point.

Signals that get recorded

Each geophone produces an electrical signal that is
proportional to ground motion (in one direction usually vertical, but horizontal with special geophones
used for shear wave work). That signal is recorded for
a short period of time starting the moment the source
of energy begins. We observe these signals by plotting
them as one wiggly line for each geophone's signal.
These signals are plotted next to each other so that
ground motion at each geophone can be seen as a
function of time. An example showing ground motion
at 24 geophones is shown to the right.
On this plot, one geophone was not working properly.
Geophones are labled with the first closest to the
source. As expected, ground motion occured earlier at
geophones closest to the source. For geophones 1
through 10, it seems as if there was no ground motion at later times, however this is an artifact of the
"gain" (amplification) applied to these traces. Gain is lower for geophone signals near the source
because signal amplitudes are larger. If the signals within the blue box are amplified and replotted,
the adjustable figure below results.

Original

Straight line segments

First breaks picked

This figure shows shorter segments of traces from 12 geophones.
The signal amplitudes have been amplified, so all ground
motions are visible. There are clear beginings of ground motion
for each trace, which appear later in time for traces farther from
the source Finding exactly what time the ground first moved at
each geophone is called first break picking. It is not difficult in
this case, but it can be challenging if signals are weak. Use radio
buttons next to the figure to see the result of picking first breaks
(also known as first arrivals) on this figure. Once the picks are
chosen, it becomes evidend that there is a definate pattern to the
arrivals - there are straight lines joining the first breaks of several
adjacent traces. The third radio button reveals straight line
patterns emphasized by drawing red lines.

We will learn that first arrivals are either direct signals (for near geophones) or refractions that have
travelled along subsurface interfaces. The objective of seismic refraction surveys is to determine the
geometry of subsurface interfaces, and this can be derived by analysis of the pattern of first arrivals.

Ray Paths in Layered Media

Reflections and refractions at a plane interface
Consider a P-wave which is incident at an
angle 1 measured with respect to the normal
of the interface. There will be a reflected wave
and a transmitted wave but the directions of the
waves are given by the diagram to the right.
Law of reflection: The angle of reflection
equals the angle of incidence. So r = 1.
Law of refraction: The angle of refraction 2
is determined through Snell's Law, which is
If the wave travels from a low velocity medium
to a high velocity medium the wave gets
refracted away from the normal. Conversely, it
gets refracted toward the normal if the wave
goes from a high velocity to a low velocity
medium.
"Critical refraction" is an important concept in refraction seismics. The maximum value of 2 is 90o.
If this is the case, then refracted waves travel horizontally in the second medium. The incident angle
that causes this, known as the critical angle, c is found using Snell's law as follows:

When the wave in the second medium is critically refracted, it travels parallel to the interface at a speed
of V2. As it travels, it radiates energy into the upper medium with the associated ray path making an
angle c with the normal. This critically refracted wave is also called a "head wave". It is somewhat
analagous to the bow wave of a moving boat.

Mode Conversion
A P-wave incident upon a boundary can
produce reflected and transmitted P-waves,
but reflected and transmitted S-waves can
also be produced. Analogous conversions
occur when there is an incident S wave on a
plane boundary. The mode conversions (P
-> S, or S -> P) can complicate
interpretation, but S-waves are always
slower than P-waves, so first arrivals will
always be P-waves unless a special S-wave
energy source is used. A valuable benefit of
using shear waves is that they provide
important information about the rigidity of
the material.

Waves for a layer over a halfspace

We do seismic refraction surveys in order to learn about the geometry of geologic layers and velocities
(ie types of) materials. To do this we must build relations relating what we know and can measure to
the things we want. In other words we must build equations that relate what we want (depths and
velocities) to what we measure (surface distance and total travel times).
Consider a layer of thickness h and velocity V1 overlying a uniform halfspace of velocity V2. A source
is detonated at time t=0. We are interested in the waves and arrival times of those waves at a receiver
which is located a distance x from the source at position D in the figure below.
There are three principle waves that will travel through the earth and arrive at position D. i) direct
waves, ii) reflected waves, and iii) critically refracted waves.

The travel time curves for these ray paths

are shown to the right, and expressions for
the ray paths and important parameters of
these travel time curves are as follows:
xcrit is the critical distance at which
the refracted arrival first arrives.
xcross is the crossover distance.
Beyond this distance the refracted
arrival is the first arrival on the
record.
Travel times of visible arrivals are
related to distance between
geophone and source (x), thickness
of the layer (h) and the velocities of
signals within the two layers (V1
and V2). Three times are of interest:
tdir is travel time of direct arrivals,
tref is arrival time of reflections and
tcr is the refraction travel time (see figure above).
These parameters are all related as follows:
The first and third of these are important for interpretation
of seismic refraction data, and the next page
explains how they arise, and how they are used with
refraction data on a T-X plot to obtain useful geological information.

Refraction Interpretation for

Horizontal Layers
The goal of a seismic refraction experiment is generally to determine depth to bedrock, depth to water table,
or to delineate major sedimentary layers. In its complete generality, the seismic refraction problem is quite
complicated. However, there are certain geometries for which formulae can be generated to predict the arrival
times. With the aid of these formulae, measurements of the refracted arrivals can be used to determine layer
parameters. We will restrict ourselves to the following situations:
1.
2.
3.
4.

The subsurface is composed of layers separated by planes and possibly dipping interfaces.
The seismic velocity in each layer is constant.
The layer velocities increase with depth.
The ray paths are in the vertical plane. That is, there is no cross-dip (which means dipping into or out of
the page in the figure to the right).

With the above assumptions, there are a number of special cases to be considered.
1. A 2-layer earth with a horizontal interface.
2. A 3-layer (and multiple-layer) earth with horizontal interfaces.
3. A 2-layer earth with a dipping interface.
In all cases, the development of the travel-time curves requires only that we know the rules of propagation, i.e.,
energy travels in straight lines in a uniform medium and refracts according to Snell's law when it enters a medium
with different velocity. We must also be able to calculate the lengths of the ray path in each layer and along the
refractor. Travel-time curves are graphs showing the travel-time, t, (length of time for a seismic signal to travel
from the source to the receiver along which ever path is being considered) versus distance between the source
and receiver, x.

One layer over basement - the horizontal interface

We want relations involving things we want in terms of things we either know of can get. We want depths,
but in fact, we can get travel times in terms of distances (known) and velocities (obtainable - see below). Depth
is obtained as a last step.
Consider an earth composed of a uniform layer with velocity V 1 and thickness z overlying a medium with velocity
V . Let be the critical angle and x denoted the distance between the source at A and a receiver at D. Let x c
2

denote the critical distance.

From elementary geometry the following relationships hold:

,

The travel time is the cumulative time for the wave to traverse the path ABCD. This is t=t AB+tBC+tCD.
Generally time = distance / velocity, so we can write t AB = L/v1 = (z/cos( )) / v1, (using L from just above).
Also, we can note that tAB = tCD and the distance BC is X-XC. So we can now state that t=2tAB+tBC , or
.
It is convenient to rearrange this slightly differently. Using the definition for critical angle sin =V1/V2, we
can make the "velocity triangle", so expressions for the angle arise directly from simple trigonometry:

Use these two relations for cos and tan in the expression for t above to obtain a useful set of relations.

You can convince yourself that x is in fact the complete distance between shot and geophone by carrying out
the intermediate analysis steps.
This simple relation says that the travel time curve is a straight line which has a slope of 1/V2 and an intercept
of ti . This intercept time is the time where the refraction line extends to intercept the Y-axis above the source
position. This is not a real "time" - it is derived from the graph.
The velocities of the seismic layers and the layer thickness are obtained in the following manner.
1. Plot the times of first arrivals on an time-offset plot
("offset" is distance between source and geophone).
2. The direct arrivals are observed to lie along a straight
line joining the origin. The slope of this line is 1/V 1 ,
giving the velocity of the upper layer.
3. The refracted arrivals appear as a straight line with
smaller slope equal to 1/V2, giving the velocity
of the lower layer.
4. For the refracted wave, this intercept time is

We therefore can obtain all three useful parameters about the earth, (V 1, z, V2).
There is another useful point that is observable from the first arrival travel-time plot. We can often discern
the crossover distance. Since this is the location where the direct wave and the refracted wave arrive at the
same time, we can write
Thus

Combine to obtain

This can be used as a consistency check, or it can be used to compute one of the variables given values
for two others.

Two Horizontal Layers Over a Halfspace

The extension to more layers is in principle straight forward. However, the algebra is slightly more
involved because we need to compute the times due to the ray path segments in the two top layers.
Consider the diagrams below:

Using arguments that are entirely analagous to the two layer case (above) the travel time for the wave
refracted at the top of layer three is given by

All quantities are defined in the diagrams, and the angles are

Note that 2 is a critical angle while 1 is not. You can prove the relation for 1 yourself by using
Snell's law at the two interfaces, and recalling that the angle of the ray coming from point B is the same
as the angle arriving at point C.
The straight line that corresponds to an individual refractor provides a velocity (from its slope) and a
thickness (from the intercept). Thus the information on the above travel-time plot allows us to recover all
three velocities and the thickness of both layers. The travel time curves for multi layers are obtained from
obvious extension of the above formulation.

Irregular Surfaces and

the Plus-Minus Method

There is one more step in

sophistication that is required so that
the refraction technique is useful.
This occurs in environments where
the interface has topography and is
no longer approximated by a plane.
If the curvature of the interface is not too great then we can still think of the refracted wave travelling along
the interface. The wave travels at the speed of the lower medium. Consider the sketch to the right. Formulae
which explicitly assume a plane interface (even dipping) do not work well enough.
We need a better
.
approximation.
We would like to be able to obtain
the depth of the interface beneath
the source and receiver. To do this
we used the concept of delay time.
Consider the diagram to the right.
The travel time is t = x/V2 +ti
where ti can be thought of as the
time delay.
The delay time can be partitioned into two parts: (1) the
"delay" associated with the source location and the (2)
"delay" associated with the receiver location. Consider
the delay at the source.
Let the delay time be denoted
by

We need to use AB=z/cos , BC=ztan , and Snell's law (sin =v1/v2, or v1=v2sin ). Then we can
write the delay time as follows, where we arrive past the second equals sign by using some
trigonometric identities.

Finally, using the definition of the velocity triangle we can write

(eqn 1).

Thus the depth can be computed if the delay time, a, can be measured.
We can write any general travel time so that it involves a delay at the source a s and a delay at the
.
receiver or detector a d. It looks like this:
.
As in the diagram at the top of this page, as and ad might be very different. To make progress we
implement Hagedoorn's "Plus-Minus" method. This method will also make it possible to estimate
depths under every geophone that has a refracted signal from both sources. This will yield several
depths under the survey line rather than just two depths under sources, thus allowing more detailed
image of the interface's shape.

Hagedoorn's Plus-Minus Method

The geometry illustrates the two shots
occurring at locations S1 and S2. The point
indicated by D lies at any location (ie any
geophone) between the two shots where
refractions can be received from both shots.
Let aS1 and aS2 denote the delay times at the
two shot locations. Although the refracting
interface is not flat beneath D, the slope is not
large and the ray path lengths AD and BD are
almost identical.
Let us write all portions of the travel times for signals that would travel between S1 and S2,
between S1 and D, and between S2 and D. These travel times are:
The "Plus" term gives "delay time" at a geophone
When the last two equations are added together,
we see that the result includes the expression for the
first. Here is what we get:

Rearranging this equation gives us the so-called "Plus" result, which is in fact the delay time beneath
location D. In involves travel times that are measured - namely the refracted arrival time to D from
both shots 1 and 2, and the time a signal would take to travel from shot 1 to shot 2. This is the
reciprocal time introduced earlier.

Referring back to the expression for z labelled (eqn 1) above, the depth to the refractor under position
D can therefore be given by

This can be evaluated if the two velocities are known.

The "Minus" term gives V2 under D.

We will assume that V1 is known from the
direct arrivals. It remains only to find V2.
Unfortunately it cannot be obtained
directly from the first arrival plot. The
refractor surface is irregular and thus the
travel time curve for the refractor is
warped. The sketch here illustrates the
difficulty.
To obtain about the velocity we look at the difference in arrival time at location D. Subtracting the
two equations yields

where the constant c is fixed for fixed S 1, S2. This suggests that v2 will be
. the slope of some straight
line. But what is plotted to get this slope? Not t vs X, but (tS1D- tS2D) vs X. The explanation follows:
Suppose we have a number
of receivers between
locations D1 and D2.
For each receiver we can measure
the refracted arrivals from
both shots. The time t =
ts1D - ts2D can be computed
.
for each detector and plotted
on a graph. According to the equation above, the slope of this graph should be 2/V2.
From the above analysis we can obtain an estimate of the depth of the refractor and the velocity
of the refractor.
Note that this "Plus-Minus" method works well if the dips
are small (< ~10o ), but begins to be less successful
for steeply dipping interfaces.

Other topics:
Hidden layers, Lateral velocity variations, Phantoming, and Statics
Hidden Layers
When acquiring refraction data we attempt to obtain first arrivals from each refractor (in fact first arrivals are
the onlyuseful data - signal energy following first arrivals can not be used unless seismic reflection is being done
- a topic covered in geop301). Sometimes it is not possible to obtain refracted arrivas from each and every
refracting horizon. There are two reasons why the second layer might go undetected.

1. Second Layer is a Low Velocity Layer: If V2 < V1

the incoming wave will refract towards the
normal and there will be no critical refraction along
the upper boundary. Assuming V3 > V1 the first
arrival travel time curve will look as shown in the
second half of the figure.
1.

2. Second Layer is Too Thin: Even though

there is a headwave on its upper surface,
the headwave travelling along the top of the
third layer arrives first. The dependence of the
travel time curves upon the thickness of the second
layer is shown in this sequence of three figures.

2.

3.

Lateral Velocity Changes

The headwave travels at the speed of the underlying medium and that velocity might vary laterally. This
produces another complication when interpreting the first arrivals. Consider the two velocity models and
traveltime curves given below. The traveltime curves are the same. We note that the crossover point between
V2a and V2b in (b) is always shifted away from the shotpoint with respect to the location of the position of
the subsurface change.

a.

b.

To determine which structure is correct we need another shot. It could be a reversed shot, or a shot that is
off the end of the profile. If the shot is offset then the travel times from the two velocities are shifted as
indicated below. In (a) the traveltime curve is shifted to the left by an amount S. In (b) a constant time
increment is added to each arrival and the location of the crossover remains the same.

Phantoming
As shown in the above diagram, refracting arrivals from two shots from the same end produce
traveltime curves which are parallel. This provides a good way to group first-break arrivals according
to specific refractors. Ideally we want to have many arrivals from a single interface. This can be
accomplished by using the parallelism noted above and phantoming the arrivals. That is, we can
determine the arrival time of the refracted event at a location where no seismometer really existed.
Phantoming can be carried out in the cases where:
1. The receiver array is fixed, and off-end
shots are recorded
2. The shot and receiver array is moved.
The diagram here shows an example where the
spread is moved.
1. For the shot at x-position 140 feet , data from
the shot at 100 feet can fill in arrival times
between positions 0 and 40 feet.
2. For the shot at 100 feet, data from 140 can phantom
in artificial arrivals that arrive before x crit . The data
can help define the intercept time because we now
have more points to help estimate a straight line.
Note that these phantomed values don't actually
exist but they are valid data for helping us estimate
the slope and intercept for the particular
refractor.

Static Data Corrections

There are two corrections that are sometimes applied to the recorded travel times. These are corrections
for elevations and similar corrections when there is a known weathered layer which has variable velocity.
Elevation Correction: The goal is to reduce the data to a datum plane, that is, to define a flat surface on which
the data might have been recorded. The importance of this lies in the fact that all of our interpretation formulae
assume that the upper surface of the earth is flat. Let hS and hDrespectively denote the height of the shot and
receiver about the datum plane (negative heights are allowed). The angle that the waves in this region are
travelling vertically is determined by Snell's law. If V n denotes the velocity of the refractor then the
elevation correction is

Advanced techniques
There are several other, more involved methods of interpretation, and modern instruments often include
one or more of these methods coded into the computers that drive the instruments. One such method is the
GRM or Generalized Reciprocal Method, which expands upon the ideas behind the Plus-Minus method
introduced above.
The figure to the right illustrates one
program that uses the GRM approach. On
top is a time distance plot of the data, with
the sympols and colours indicating which
layer each arrival is interpreted to have
refracted from. The bottom panel shows
the cross section resulting from careful
interpretation using the GRM method.
Purple is the interface below the top layer,
yellow is the second interface, and the
third interface (between the third layer,
and bedrock) is outlined with red arcs.
The arcs help identify a more accurate
horizon when the dip is interpreted to be
steeper than around 10o. Use of arcs like
this is a processing step known as
"migration", a topic more commonly covered in relation to seismic reflection.
There are also wave-propagation methods implemented by some commercial programs. These are inversion
techniques. They start with a "guess"or estimated first model, calculate how wavefronts will propagate
through a the model from the sources to the geophones of the survey, then compare the resulting calculations
to measured first arrivals. The process is repeated after adjusting the model so as to generate a data set that
more closely matches the measurements. The process is "finished" when there is a model that can cause data
that look close to measurements according to pre-defined statistical criteria - ie when "misfit" is within
pre-defined specifications.
Examples of two interpretation methods at one site are given to the right. They come from the paper
"2-D velocity structure of the buried ancient canal of Xerxes: an application of seismic methods in
archaeology", 2001, V.K. Karastathis, S. Papamarinopoulos, and R.E. Jones; Journal of Applied
Geophysics 47 2001 2943.

TOP: The refraction first arrival T-X plot (ie, raw data) for line 96-12.

TOP: Raw T-X plot from first arrivals.

MIDDLE: A GRM solution derived from the data shown. There are clearly "steeply" dipping sections
of the interface between oveburden and bedrock. The 34m wide channel represents the target for
this survey, which was trying to determine whether a canal referenced in greek legends was real or not.

MIDDLE: model produced using GRM interpretation method.

BOTTOM: A ray-tracing inversion solution derived from the same data set. The outcome is
"smoothly" varying velocities, but there is an apparent interface represenetd by tightly spaced
contours. This corresponds to the interface derived from the GRM solution. It illustrates how different
methods will produce different results, but that both are equally interpretable for geological or
geotechnical information.

BOTTOM: model produced using rat-tracing inversion.

Ray tracing ray plot, and data matching graphs. The ray tracing inversion algorithm can also produce an
image showing the pathways that energy would have travelled from sources to each geophone, using the
model that was determined at the end of the inversion process. The bottom portion of the image shows
how well the calculated data (based on the model) corresponds to real measurements. The small vertical
bars are first arrivals with error bars on them. The lines show "T-X" graphs of calculated data, and they
should be within the error bars assigned for measurements. They are sometimes a little off, but the
outcome is considered quite good given this challenging situation.

Seismic refraction examples

Here are three short examples of seismic refraction applied to engineering-scale problems.

Velocity as a guide to rock strength

(From P.V.Sharma, 1997, Environmental and engineering geophysics, Cambridge University Press, p175)
Seismic velocity and rock type (empirical, only for one the location mentioned).
Degree of separation C
No joints
0.65-1
Few joints
Jointed rock
Numerous joints
Strongly jointed
rock

RQD
Seismic velocities are dependent upon rock type
very good but also upon factors like porosity, water content,
and degrees of fracturing and jointing. In
0.45- 0.65 4000- 4500 good
environments where the rock type is uniform,
variations in the P-wave velocity may sometimes
0.3-0.45 3500- 4000 moderate be used to infer how these factors vary with
location. As an example, in studying crystalline
rocks from Sweden, there is a empirical
0.15-0.3 3000- 3500 bad
relationship between the jointing factor (C),
seismic velocity (Vp), and the rock quality
0.00- 0.15 < 3000
very bad designation (RQP). These are listed in the table
reproduced here from Sharma, 1997.
Vp (m/s)
> 4500

Rippability chart, courtesy of Caterpilar Inc. (In Sharma, 1997)

The P-wave velocities of the surface layers

can also be a useful guide to the rippability
of the material for excavation. Surface layer
materials with velocities less than 2000 m/s
may usually be ripped by a bull-dozer.
Generally, the lower the velocity of any rock
type, the more rippable it is. This is
characterized by the table shown to the right.

Mapping Sand and Gravel Deposits

(From P.V.Sharma, 1997, Environmental and engineering geophysics, Cambridge University Press,
p180-182)
Delineation of sand and gravel deposits is of special importance in connection with the planning of
quarrying operations at potential exploitation sites. These deposits, when located above the water table,
are characterized by low velocities in the range 400 -1000 m/s, in contrast to water-saturated sand and
moraine which typically have velocities in the range 1300 - 2000 m/s. The following example is from
Denmark.
The seismic refraction profile in an area of Quaternary sedimentary deposits Ourdrup Kirke, Denmark is
shown below. Interpretation of the travel time curves indicated the first layer (gravel) with velocity varying
from 330 - 500 m/s, the second layer (sand) of velocity between 560 and 1000 m/s, and the third layer
(water-saturated chalk) of velocity 1650 - 2800 m/s. The lower layer also included a few isolated zones of
high velocity (3400 - 4000) m/s as well as zones of low velocity (1000 - 1250) m/s. The interbed thickness
of the gravel deposit varies from 5 m to 14 m and the combined thickness of the gravel and sand is about
25 m. Interpretations of velocities and depths were made using the plus-minus method.

Seismic refraction profile in an area of Quaternary sedimentary deposits, Oudrup Kirke, Denmark.
Interpretation of the travel-time curves indicated the presence of three layers corresponding to the gravel,
sand and chalk formations. Note the relatively large lateral velocity changes in the chalk. This is a fairly
large data set for an engineering project.
Questions:
1. How many geophones were used for each spread, and what was their spacing?
2. How many spreads were involved in the entire survey?
3. What was the TOTAL number of shots fired into the eastern-most spread?
4. How many of the shots fired into the eastern-most spread were located further east from location 0?
5. How many shots were detected by the geophone at location 0?

Shear wave refraction

This record was recorded using geophones mounted horizontally and transverse to the profile. To generate
the shear wave signal a metal bar with fins to anchor it to the ground was placed transverse to the profile,
then struck with a hammer on either end. Records from these two shots were subtracted so as to increase
signal-to-noise ratio.
Aprominent air wave (A) can be seen, followed by a horizontally polarized shear wave (S). High amplited
late arrivals are surface waves or ground roll. S-wave refractions were used to identify the structure of glacial
till below the shallow water table. Shear waves are useful here because the water table refraction is nearly
invisible because the water does not contribute to shear wave velocity. When P-waves were recorded at this
site, the refraction at the top of the water table was the dominant feature because of the significant velocity
contrast at the top of the water table. Glacial till structure was therefore easier to interpret from shear wave
data than from P-wave data.

This is a shear wave refraction record from Baden, Ontario. P-waves are barely visible but their location is
marked with the P. Shear refractions were not readily visible on P-wave records.
From Greenhouse and Gudjurgis, 1995, Reference notes for an EEGS short course on applications of
surface geophysics to environmental investigations.

Seismic Refraction
Glossary of terms
These are terms commonly in use for seismic refraction work.
Arrival times
The time at which the direct or refracted energy first arrives at a geophone.
Approximate Velocity Depth Migration
A procedure for converting the time-depth section to depth units using refractor velocities and all but
one layer thicknesses. The one unknown layer thickness is that immediately above the target
refracting layer. This method is not highly sensitive to the proper selection of the XY value.
Average Velocity Depth Migration
A procedure for converting the time-depth section to depth units using a velocity which represents
the average of all units above the refracting layer. This method is sensitive to the proper selection of
the XY value.
Critical Distance
The distance from the shot within which no refracted energy can return to the surface. Only direct or
reflected energy can be seen within the critical distance. It is related to the critical angle, which is the
incident angle which results in refraction along the interface.
Depth Migration
The process of converting the time-depth section to units of depth. Two methods of depth migration
are available in the GRM: Average Velocity and Approximate Velocity methods. Comparison of the
results from each method can indicate if hidden layer conditions exist. If the two methods do not
approximately agree, then there is cause for concern, and the possibility of a hidden layer should be
considered.
Dipping, planar layer
A smooth (no topography) layer, that lies at some angle to the surface. It is also assumed that there is
no variation perpendicular to the survey line.
First arrivals
See Arrival times.
First break picks
See Arrival times.
Forward, Mid or Reverse shots
See Shot Direction.
Forward Modelling
The process of constructing artificial results in order to learn something about a real geophysical
situation. Interpretation may involve adjusting the artificial (usually mathematical) model until
results compare favourably to field data.
Geophone array
See Seismic Spread
Geophone Spacing
The spacing between geophones in a shot.
GRM
The Generalized Reciprocal Method, one of several methods for interpreting shallow seismic
refraction data. It provides a more complete solution than other methods, but also requires a more
complete data set.
Group interval
See Geophone Spacing

Hidden layer
A layer of material that is not visible on the seismic section, either because geophone spacing is not
adequate, or because it has a seismic velocity that is less than that of the material above it.
Layer Assignment
The process of determining the layer from which an arrival was refracted. Ideally, all arrivals
refracted from the same layer will fall on a line with a constant slope. In practice, however, irregular
layer interfaces and velocity variation tend to scatter the first break arrival times. In this case,
patterns observed between shots with different offset spacing can be used to guide layer assignment.
Phantoming
The phantoming process involves shifting a set of arrival times vertically in time, so that they link up
with arrivals from an adjacent shot. The purpose of phantoming is to create a complete set of
forward and reverse arrivals along the line, all referenced to the same shot position.
Raw trace data: The signal recorded by a geophone as a result of one shot. It consists of a set of
parameters defining the data (a header) and the digitized voltage for a fixed amount of time
following the shot time.
Reciprocal Time
The time it takes for energy to travel from the forward shot point to the reverse shot point. This is
NOT the time from one shot to the last geophone.
Seismic Line
A combination of seismic spreads collected end to end (often overlapping by a few geophones).
Some interpretation programs organize all shots and spreads together in one interpretation file.
Seismic Shot
Some programs organize arrival time information based on shots. Each shot is made relative to a
spread, and so 12 or 24 arrival times will be associated with each shot.
Seismic Spread
A seismic spread is a combination of 12 or 24 geophones in position on the ground. Various shots
may be made at different positions relative to the spread.
Shot Direction
The shot direction refers to the position of the shot relative to the spread, and relative to the order in
which spreads are collected along the seismic line. The shot direction can be either Forward, Mid or
Reverse.
Shot Offset
The relative distance between the shot position and the first or last geophone. "Zero offset" refers to
shots in which the shot was made at geophone position 1, while "Long offset" refers to shots made
some distance from geophone 1, along the seismic line, but not within the range of the spread.
Shot Position
The actual line position of the shot. (ie. the position relative to the start of the seismic line.)
Time Distance plot
A graph showing first arrival time versus distance from the shot (or distance along the line). Time
increases up on the Y-axis, and distance increases right on the X-axis.
Time-Depth Analysis
A cross section showing layer positions in units of travel time (as opposed to units of depth).
Directly analogous to the time sections created from seismic reflection recording. Time-depth
analysis curves are usually presented for a series of XY value, so as to select the optimum XY value.

T-X plot
See Time Distance plot
Velocity Analysis Curves
Curves, calculated using a GRM based equation, which are used to determine both the velocity of
the refracting layer and the optimum XY. The inverse of the slope of the velocity analysis curve is
the refractor velocity. Changes in the slope of the curve indicates changes in the velocity of the
bedrock. The curves are usually calculated for a range of XY value, so as to select the optimum XY
value.
Vertical Travel Time
The time it takes seismic energy to travel vertically from the surface to an interface. See also timedepth.
XY Value
A measure of distance, along the ground surface, between the arrival of forward and reverse
refracted energy emitted from a single location on the refracting interface. The XY value is central to
the GRM procedure, and accurate depth sections can only be determined from proper XY selection.
The optimum XY value is determined by inspection of the Velocity Analysis Curves and the TimeDepth Section.

F. Jones, UBC Earth and Ocean Sciences.