G Magnetics Methods
G Magnetics Methods
G Magnetics Methods
Introduction
Introduction to Magnetic Exploration - Historical Overview
Unlike the gravitational observations described in the previous section, man has been
systematically observing the earth's magnetic field for almost 500 years. Sir William Gilbert
(left) published the first scientific treatise on the earth's magnetic field entitled De magnete.
In this work, Gilbert showed that the reason compass needles point toward the earth's north
pole is because the earth itself appears to behave as a large magnet. Gilbert also showed that
the earth's magnetic field is roughly equivalent to that which would be generated by a bar
magnet located at the center of the earth and oriented along the earth's rotational axis. During
the mid-nineteenth century, Karl Frederick Gauss confirmed Gilbert's observations and also
showed that the magnetic field observed on the surface of the earth could not be caused by
magnetic sources external to the earth, but rather had to be caused by sources within the
earth.
Geophysical exploration using measurements of the earth's magnetic field was employed
earlier than any other geophysical technique. von Werde located deposits of ore by mapping
variations in the magnetic field in 1843. In 1879, Thalen published the first geophysical
manuscript entitled The Examination of Iron Ore Deposits by Magnetic Measurements.
Even to this day, the magnetic methods are one of the most commonly used geophysical
tools. This stems from the fact that magnetic observations are obtained relatively easily and
cheaply and few corrections must be applied to the observations. Despite these obvious
advantages, like the gravitational methods, interpretations of magnetic observations suffer
from a lack of uniqueness.
• Geophysical exploration techniques that employ both gravity and magnetics are
passive. By this, we simply mean that when using these two methods we measure a
naturally occurring field of the earth: either the earth's gravitational or magnetic
fields. Collectively, the gravity and magnetics methods are often referred to as
potential methods*, and the gravitational and magnetic fields that we measure are
referred to as potential fields.
*The expression potential field refers to a mathematical property of these types of force
fields. Both gravitational and the magnetic forces are known as conservative forces. This
property relates to work being path independent. That is, it takes the same amount of work to
move a mass, in some external gravitational field, from one point to another regardless of the
path taken between the two points. Conservative forces can be represented mathematically
by simple scalar expressions known as potentials. Hence, the expression potential field.
• Unlike the gravitational field, which does not change significantly with time**, the
magnetic field is highly time dependent.
*One order of magnitude is a factor of ten. Thus, four orders of magnitude represent a
variation of 10,000.
**By this we are only referring to that portion of the gravity field produced by the internal
density distribution and not that produced by the tidal or drift components of the observed
field. That portion of the magnetic field relating to internal earth structure can vary
significantly with time.
Magnetic Monopoles
Recall that the gravitational force exerted between two point masses of mass m1 and m2
separated by a distance r is given by Newton's law of gravitation, which is written as
where G is the gravitational constant. This law, in words, simply states that the gravitational
force exerted between two bodies decreases as one over the square of the distance separating
the bodies. Since mass, distance, and the gravitational constant are always positive values,
the gravitational force is always an attractive force.
form, this expression is identical to the gravitational force expression written above. There
are, however, two important differences.
As described by Coulomb's equation, the size of the arrows should decrease as one over the
square of the distance between the two magnetic poles* and the direction of the force acting
on p2 is always in the direction toward p1 (the force is attractive)**.
If instead p1 is a positive pole located at x=1, the plot of the magnetic force acting on p2 is
the same as that shown above except that the force is always directed away from p1 (the
force is repulsive).
*For plotting purposes, the arrow lengths shown in the figures above decay proportional to
one over the distance between the two poles rather than proportional to one over the square
of the distance between the two poles. If the true distance relationship were used, the lengths
of the arrows would decrease so rapidly with distance that it would be difficult to visualize
the distance-force relationship being described.
**If we were to plot the force of gravitational attraction between two point masses, the plot
would look identical to this.
Magnetic Dipoles
So far everything seems pretty simple and directly comparable to gravitational forces, albeit
with attractive and repulsive forces existing in the magnetic case when only attractive forces
existed in the gravitational case. Now things start getting a bit more complicated. The
magnetic monopoles that we have been describing have never actually been observed!!
Rather, the fundamental magnetic element appears to consist of two magnetic monopoles,
one positive and one negative, separated by some distance. This fundamental magnetic
element consisting of two monopoles is called a magnetic dipole.
Now let's see what the force looks like from this fundamental magnetic element, the
magnetic dipole? Fortunately, we can derive the magnetic force produced by a dipole by
considering the force produced by two magnetic monopoles. In fact, this is why we began
our discussion on magnetism by looking at magnetic monopoles. If a dipole simply consists
of two magnetic monopoles, you might expect that the force generated by a dipole is simply
the force generated by one monopole added to the force generated by a second monopole.
This is exactly right!!
On the previous page, we plotted the magnetic forces associated with two magnetic
monopoles. These are reproduced below on the same figure as the red and purple arrows.
If we add these forces together using vector addition, we get the green arrows. These green
arrows now indicate the force associated with a magnetic dipole consisting of a negative
monopole at x=-1, labeled S, and a positive monopole at x=1, labeled N. Shown below are
the force arrows for this same magnetic dipole without the red and purple arrows indicating
the monopole forces.
The force associated with this fundamental element of magnetism, the magnetic dipole, now
looks more complicated than the simple force associated with gravity. Notice how the
arrows describing the magnetic force appear to come out of the monopole labeled N and into
the monopole labeled S.
You may recognize this force distribution. It is nothing more than the magnetic force
distribution observed around a simple bar magnet. In fact, a bar magnet can be thought of as
nothing more than two magnetic monopoles separated by the length of the magnet. The
magnetic force appears to originate out of the north pole,N, of the magnet and to terminate at
the south pole, S, of the magnet.
Notice that the red lines representing the field lines are always parallel to the force directions
shown by the green arrows. The number and spacing of the red lines that we have chosen to
show is arbitrary except for one factor. The position of the red lines shown has been chosen
to qualitatively indicate the relative strength of the magnetic field. Where adjacent red lines
are closely spaced, such as near the two monopoles (blue and yellow circles) comprising the
dipole, the magnetic force is large. The greater the distance between adjacent red lines, the
smaller the magnitude of the magnetic force.
*Unlike the force plots shown on the previous page, the arrows representing the force have
not been rescaled. Thus, you can now see how rapidly the size of the force decreases with
distance from the dipole. Small forces are represented only by an arrow head that is constant
in size. In addition, please note that the vertical axis in the above plot covers a distance
almost three times as large as the horizontal axis.
The magnetic field strength, H, is defined as the force per unit pole strength exerted by a
magnetic monopole, p1. H is nothing more than Coulomb's expression divided by p2. The
magnetic field strength H is the magnetic analog to the gravitational acceleration, g.
Given the units associated with force, N, and magnetic monopoles, Amp - m, the units
associated with magnetic field strength are Newtons per Ampere-
meter, N / (Amp - m). A N / (Amp - m) is referred to as a Tesla (T),
named after the renowned inventor Nikola Tesla, shown at left.
Magnetization of Materials
Magnetic Induction
When a magnetic material, say iron, is placed within a magnetic field, H, the magnetic
material will produce its own magnetization. This phenomenon is called induced
magnetization.
In practice, the induced magnetic field (that is, the one produced by the magnetic material)
will look like it is being created by a series of magnetic dipoles located within the magnetic
material and oriented parallel to the direction of the inducing field, H.
The strength of the magnetic field induced by the magnetic material due to the inducing field
is called the intensity of magnetization, I.
Magnetic Susceptibility
The intensity of magnetization, I, is related to the strength of the inducing magnetic field, H,
through a constant of proportionality,k, known as the magnetic susceptibility.
* We have not mentioned previously that the source of all magnetic fields is most probably circulating electric
currents - the field due to a current loop is identical with the field due to a dipole pair. Orbiting electrons in an
atom are therefore conceptually responsible for a magnetic field which can react with the ambient field, or
which can be acted on by the ambient field.
Unlike density, notice the large range of susceptibilities not only between varying rocks and
minerals but also within rocks of the same type. It is not uncommon to see variations in
susceptibility of several orders of magnitude for different igneous rock samples. In addition,
like density, there is considerable overlap in the measured susceptibilities. Hence, a
knowledge of susceptibility alone will not be sufficient to determine rock type, and,
alternately, a knowledge of rock type is often not sufficient to estimate the expected
susceptibility.
This wide range in susceptibilities implies that spatial variations in the observed magnetic
field may be readily related to geological structure. Because variations within any given rock
type are also large, however, it will be difficult to construct corrections to our observed
magnetic field on assumed susceptibilities as was done in constructing some of the
fundamental gravitational corrections (Bouguer slab correction and Topographic
corrections).
*Although susceptibility is unitless, its values differ depending on the unit system used to quantify H and I. The
values given here assume the use of the SI, International System of Units (Système International d'Unités)
based on the meter, kilogram, and second. Another, older, unit system, the cgs, centimeter, gram, and second
system is also used in exploration geophysics. To convert the SI units for susceptibility given above to cgs,
divide by 4¹.
Remanent Magnetization
So, as we've seen, if we have a magnetic material and place it in an external magnetic field
(one that we've called the inducing field), we can make the magnetic material produce its
own magnetic field. If we were to measure the total magnetic field near the material, that
field would be the sum of the external, or inducing field, and the induced field produced in
the material. By measuring spatial variations in the total magnetic field and by knowing what
the inducing field looks like, we can, in principle, map spatial variations in the induced field
and from this determine spatial variations in the magnetic susceptibility of the subsurface.
Although this situation is a bit more complex than the gravitational situation, it's still
manageable. There is, however, one more complication in nature concerning material
magnetism that we need to consider. In the scenario we've been discussing, the induced
magnetic field is a direct consequence of a magnetic material being surrounded by an
inducing magnetic field. If you turn off the inducing magnetic field, the induced
magnetization disappears. Or does it?
If the magnetic material has relatively large susceptibilities, or if the inducing field is strong,
the magnetic material will retain a portion of its induced magnetization even after the
induced field disappears. This remaining magnetization is called Remanent Magnetization.
The only way you can measure the remanent magnetic component of a rock is to take a
sample of the rock back to the laboratory for analysis. This is time consuming and
expensive. As a result, in exploration geophysics, we typically assume there is no remanent
magnetic component in the observed magnetic field. Clearly, however, this assumption is
wrong and could possibly bias our interpretations.
When discussing gravity, we really didn't talk much about how we describe gravitational
acceleration. To some extent, this is because such a description is almost obvious;
gravitational acceleration has some size (measured in geophysics with a gravimeter in
mGals), and it is always acting downward (in fact, it is how we define down). Because the
magnetic field does not act along any such easily definable direction, earth scientists have
developed a nomenclature to describe the magnetic field at any point on the Earth's surface.
*In this context, and throughout the remainder of these notes, F includes contributions from the Earth's main**
magnetic field (the inducing field), induced magnetization from crustal sources, and any contributions from
sources external to the Earth.
**The main magnetic field refers to that portion of the Earth's magnetic field that is believed to be generated
within the Earth's core. It constitutes the largest portion of the magnetic field and is the field that acts to induce
magnetization in crustal rocks that we are interested in for exploration applications.
The remaining 10% of the magnetic field can not be explained in terms of simple dipolar
sources. Complex models of the Earth's magnetic field have been developed and are
available. Shown below is a sample of one of these models generated by the USGS. The plot
shows a map of declinations for a model of the magnetic field as it appeared in the year
1995*.
If the Earth's field were simply dipolar with the axis of the dipole oriented along the Earth's
rotational axis, all declinations would be 0 degrees (the field would always point toward the
north). As can be seen, the observed declinations are quite complex.
As observed on the surface of the earth, the magnetic field can be broken into three separate
components.
• Main Field - This is the largest component of the magnetic field and is believed to be
caused by electrical currents in the Earth's fluid outer core. For exploration work, this
field acts as the inducing magnetic field.
• External Magnetic Field - This is a relatively small portion of the observed magnetic
field that is generated from magnetic sources external to the earth. This field is
believed to be produced by interactions of the Earth's ionosphere with the solar wind.
Hence, some temporal variations associated with the external magnetic field are
correlated to solar activity.
• Crustal Field - This is the portion of the magnetic field associated with the
magnetism of crustal rocks. This portion of the field contains both magnetism caused
by induction from the Earth's main magnetic field and from remanent magnetization.
The figure shown above was constructed to emphasize characteristics of the main magnetic
field. Although this portion of the field is in itself complex, it is understood quite well.
Models of the main field are available and can be used for data reduction.
*As we'll describe later, another potential complication in using magnetic observations is that the Earth's
magnetic field changes with time!
Therefore, if we were to measure the magnetic field along the surface of the earth, we would
record magnetization due to both the main and induced fields. The induced field is the one of
interest to us because it relates to the proximity of our measure to rocks high or low
magnetic susceptibility. If our measurements are taken near rocks of high magnetic
susceptibility, we will, in general**, record magnetic field strengths that are larger than if
our measurements were taken at a great distance from rocks of high magnetic susceptibility.
Hence, like gravity, we can potentially locate subsurface rocks having high magnetic
susceptibilities by mapping variations in the strength of the magnetic field at the Earth's
surface.
Consider the example shown above. Suppose we have a buried dyke with a susceptibility of
0.001 surrounded by sedimentary rocks with no magnetic susceptibility. The dyke in this
example is 3 meters wide, is buried 5 meters deep, and trends to the northeast. To find the
dyke, we could measure the strength of the magnetic field (in this case along an east-west
trending line). As we approach the dyke, we would begin to observe the induced magnetic
field associated with the dyke in addition to the Earth's main field. Thus, we could determine
the location of the dyke and possibly its dimensions by measuring the spatial variation in the
strength of the magnetic field.
There are several things to notice about the magnetic anomaly produced by this dyke.
*We will assume that there is no remanent magnetization throughout the remainder of this discussion.
**Unlike gravity, magnetic anomalies are rather complex in shape and making sweeping statements like this
can be very dangerous.
Like the gravitational field, the magnetic field varies with time. When describing temporal
variations of the magnetic field, it is useful to classify these variations into one of three types
depending on their rate of occurrence and source. Please note explicitly that the temporal
variations in the magnetic field that we will be discussing are those that have been observed
directly during human history. However, the best-known temporal variation, magnetic
polarity reversals, while important in the study of earth history, will not be considered in this
discussion. We will, however, consider the following three temporal variations:
• Secular Variations - These are long-term (changes in the field that occur over years)
variations in the main magnetic field that are presumably caused by fluid motion in
the Earth's Outer Core. Because these variations occur slowly with respect to the time
*In this context, "rapidly" means on the order of hundreds to tens of years, down to minutes.
Shown below is a plot of the declination and inclination of the magnetic field around Britain
from the years 1500 through 1900.
At this one location you can see that, over the past 400 years, the declination has varied by
almost 37 degrees while the inclination has varied by as much as 13 degrees. These changes
are generally assumed to be associated with the Earth's main magnetic field. That is, these
are changes associated with that portion of the magnetic field believed to be generated in the
Earth's core. Solid-earth geophysicists are very interested in studying these secular
variations, because they can be used to understand the dynamics of the Earth's core.
To understand these temporal variations and to quantify the rate of variability over time,
standard reference models are constructed from magnetic observatory observations about
every five years. One commonly used set of reference models is known as the International
Geomagnetic Reference Field. Based on these models, it is possible to predict the portion of
the observed magnetic field associated with the Earth's main magnetic field at any point on
the Earth's surface, both now and for several decades in the past.
Because the main magnetic field as described by these secular variations changes slowly
with respect to the time it takes us to complete our exploration magnetic survey, this type of
temporal variation is of little importance to us.
The plot below shows typical variations in the magnetic data recorded at a single location
(Boulder, Colorado) over a time period of two days. Although there are high-frequency
components to this variation, notice that the dominant trend is a slowly varying component
with a period of about 24 hours. In this location, at this time, the amplitude of this daily
variation is about 20 nT.
These variations are believed to be caused by electric currents induced in the Earth from an
external source. In this case, the external source is believed to be electric currents in the
upper atmosphere, or the ionosphere. These electric currents in the ionosphere are in turn
driven by solar activity.
Given the size of these variations, the size of the magnetic anomalies we would expect in a
typical geophysical survey, and the fact that surveys could take several days or weeks to
complete, it is clear that we must account for diurnal variations when interpreting our
magnetic data.
Magnetic Storms
In addition to the relatively predictable and smoothly varying diurnal variations, there can be
transient, large amplitude (up to 1000 nT) variations in the field that are referred to as
Magnetic storms.
The frequency of these storms correlates with sunspot activity. Based on this, some
prediction of magnetic storm activity is possible. The most intense storms can be observed
globally and may last for several days.
The figure to the right shows the magnetic field recorded at a single location during such a
transient event. Although the magnetic storm associated with this event is not particularly
long-lived, notice that the size of the magnetic field during this event varies by almost 100
NT in a time period shorter than 10 minutes!!
Exploration magnetic surveys should not be conducted during times of magnetic storms.
This is simply because the variations in the field that they can produce are large, rapid, and
spatially varying. Therefore, it is difficult to correct for them in acquired data.
Magnetometers
Measuring the Earth's Magnetic Field
Instruments for measuring aspects of the Earth's magnetic field are among some of the oldest
scientific instruments in existence. Magnetic instruments can be classified into two types.
• Mechanical Instruments - These are instruments that are mechanical in nature that
usually measure the attitude (its direction or a component of its direction) of the
magnetic field. The most common example of this type of instrument is the simple
Since that time, several other magnetometer designs have been developed that
include the Proton Precession and Alkali-Vapor magnetometers.
In the following discussion, we will describe only the fluxgate and the proton precession
magnetometers, because they are the most commonly used magnetometers in ground
exploration surveys. (Fluxgate magnetometers are now obsolescent as exploration
instruments, but the transducers themselves are found in other devices, ranging from
electronic compasses to some kinds of electrical exploration equipment.)
Fluxgate Magnetometer
The fluxgate magnetometer was originally designed and developed during World War II. It
was built for use from low-flying aircraft as a submarine detection device. Today it is used
for making borehole measurements, and the transducers are found in electronic compasses
and in laboratory devices for measuring remanent magnetisation. Much airborne magnetic
surveying was carried out using fluxgate detectors between 1945 and 1985, and hand-held
portable devices were used for making vertical-component ground measurements between
1965 and 1985. A schematic of the fluxgate magnetometer is shown below.
The fluxgate magnetometer is based on what is referred to as the magnetic saturation circuit.
Two parallel bars of a ferromagnetic material are placed closely together. The susceptibility
of the two bars is large enough so that even the Earth's relatively weak magnetic field can
produce near magnetic saturation* in the bars.
Each bar is wound with a primary coil, but the direction in which the coil is wrapped around
the bars is reversed. An alternating current (AC) is passed through the primary coils causing
a large, artificial, and varying magnetic field in each coil. This produces induced magnetic
fields in the two cores that have the same strengths but opposite orientations, at any given
time during the current cycle.
If the cores are in an external magnetic field, one component of the external field will be
parallel to the core axes. As the current in the primary coil increases, the magnetic field in
one core will be parallel to the external field and so reinforced by it. The other will be in
opposition to the external field and so smaller. The field will reach saturation in one core at a
time different from the other core (and fall below saturation, as the current decreases, at a
different time. This difference is sufficient to induce a measurable voltage in a secondary
coil that is proportional to the strength of the magnetic field in the direction of the cores.
The secondary coil surrounds the two ferromagnetic cores and the primary coil. The
magnetic fields induced in the cores by the primary coil produce a voltage potential in the
secondary coil. In the absence of an external field (i.e., if the earth had no magnetic field),
the voltage detected in the secondary coil would be zero because the magnetic fields
generated in the two cores have the same strength but are in opposite directions (their effects
on the secondary coil exactly cancel). In the presence of an external field component, the
behaviour in the two cores differs, by an amount which depends on the external field.
Thus, the fluxgate magnetometer is capable of measuring the strength of any component of
the Earth's magnetic field by simply reorienting the instrument so that the cores are parallel
to the desired component. Fluxgate magnetometers are capable of measuring the strength of
the magnetic field to about 0.5 to 1.0 nT. These are relatively simple instruments to
construct, hence they are relatively inexpensive ($5,000 - $10,000). Unlike the commonly
used gravimeters, fluxgate magnetometers show no appreciable instrument drift with time.
*Magnetic saturation refers to the induced magnetic field produced in the bars. In general, as the magnitude of
the inducing field increases, the magnitude of the induced field increases in the same proportion as given by
our mathematical expression relating the external to the induced magnetic fields. For large external field
strengths, however, this simple relationship between the inducing and the induced field no longer holds.
Saturation occurs when increases in the strength of the inducing field no longer produce larger induced fields.
include water, kerosene, and alcohol. The sensor is connected by a cable to a small unit in
which is housed a power supply, an electronic switch, an amplifier, and a frequency counter.
When the switch is closed, a DC current delivered by a battery is directed through the coil,
producing a relatively strong magnetic field in the fluid-filled cylinder. The hydrogen nuclei
(protons), which behave like minute spinning dipole magnets, become aligned along the
direction of the applied field (i.e., along the axis of the cylinder). Power is then cut to the
coil by opening the switch. Because the Earth's magnetic field generates a torque on the
aligned, spinning hydrogen nuclei, they begin to precess* around the direction of the Earth's
total field. This precession produces a time-varying magnetic field which induces a small
alternating current in the coil. The frequency of the AC current is equal to the frequency of
precession of the nuclei. Because the frequency of precession is proportional to the strength
of the total field and because the constant of proportionality is well known, the total field
strength can be determined quite accurately.
Like the fluxgate magnetometer, the proton precession magnetometer is relatively easy to
construct. Thus, it is also relatively inexpensive ($5,000 - $10,000). The strength of the total
field can be measured down to about 0.1 nT. Like fluxgate magnetometers, proton
precession magnetometers show no appreciable instrument drift with time.
One of the important advantages of the proton precession magnetometer is its ease of use
and reliability. Sensor orientation need only be set to a high angle with respect to the Earth's
magnetic field. No precise leveling or orientation is needed. If, however, the magnetic field
changes rapidly from place to place (larger than about 600 NT/m), different portions of the
cylindrical sensor will be influenced by magnetic fields of various magnitudes, and readings
will be seriously degraded. This may occur if the sensor is placed close to magnetically-
susceptible material, for instance. Finally, because the signal generated by precession is
small, this instrument can not be used near AC power sources.
*Precession is the motion like experienced by a top as it spins and "wobbles". Because of the Earth's
gravitational field, a top whose spin axis is not parallel to the direction of gravity not only spins about its axis
of rotation, but the axis of rotation also rotates about the vertical direction. This rotation of the top's spin axis is
referred to as precession.
Ignoring for the moment the temporally varying contribution to the recorded magnetic field
caused by the external magnetic field, the magnetic field we record with our proton
precession magnetometer has two components:
• The main magnetic field, or that part of the Earth's magnetic field generated by deep
(outer core) sources. The direction and size of this component of the magnetic field at
some point on the Earth's surface is represented by the vector labeled Fe in the
figure.
• The anomalous magnetic field, or that part of the Earth's magnetic field caused by
magnetic induction of crustal rocks or remanent magnetization of crustal rocks. The
direction and size of this component of the magnetic field is represented by the
vector labeled Fa in the figure.
The total magnetic field we record, labeled Ft in the figure, is nothing more than the sum of
Fe and Fa.
Typically, Fe is much larger than Fa, as is shown in the figure (50,000 nT versus 100 nT). If
Fe is much larger than Fa, then Ft will point almost in the same direction as Fe regardless of
the direction of Fa. That is because the anomalous field, Fa, is so much smaller than the
main field, Fe, that the total field, Ft, will be almost parallel to the main field.
Field Procedures
Modes of Acquiring Magnetic Observations
Magnetic observations are routinely collected using any one of three different field
operational strategies.
airborne surveying. When other geophysical methods are being conducted by ship,
however, it may make sense to acquire magnetic data simultaneously.
• Ground Based - Like gravity surveys, magnetic surveys are also commonly
conducted on foot or with a vehicle. Ground-based surveys may be necessary when
the target of interest requires more closely-spaced readings than are possible to
acquire from the air. In the next discussion we will concentrate on ground-based
surveys. All of this discussion, however, could be applied to air- and ship-borne
surveys also.
Because magnetic surveying is generally far cheaper than other geophysical methods,
magnetic observations are commonly used for reconnaissance. These surveys can cover large
areas and are used to identify the locations of targets for more detailed investigations.
Because of their cost effectiveness, magnetic surveys usually consist of areal distributions of
data instead of single lines of data. We will refer to the collection of geophysical
observations over a geographic area as two-dimensional surveys. Data that is collected along
a single line of observations will be referred to as one-dimensional surveys.
• Field variations can be more erratic - Unlike the gravitational field, the magnetic
field can vary quite erratically with time, as shown in the figures below. What this
means is that to adequately approximate the temporal variation in the magnetic field
by linearly interpolating between base station reoccupations, a very short
reoccupation time interval may be required. The shorter the reoccupation interval, the
more time is spent at the base station and the longer the survey will take to complete.
• Cheap Instruments - Unlike gravimeters that can cost more than $25,000,
magnetometers are relatively cheap (~$7,500).
• Instrument Drift - Unlike gravimeters, magnetometers show no appreciable
instrument drift.
With these points in mind, most investigators conduct magnetic surveys using two
magnetometers. One is used to monitor temporal variations of the magnetic field
continuously at a chosen base station, and the other is used to collect observations related to
the survey proper.
By recording the times at which each magnetic station readings are made and subtracting the
magnetic field strength at the base station recorded at that same time, temporal variations in
the magnetic field can be eliminated. The resulting field then represents relative values of the
variation in total field strength with respect to the magnetic base station. (For a large-area
survey, or an airborne survey, the magnetic field may be monitored at several locations,
continuously, to protect against spatial variations in the time-varying field.)
Variations in total field strength as a function of elevation are less than 0.015 nT per meter.
This variation is generally considered small enough to ignore. Variations in total field
strength caused by excess magnetic material in topography could, on the other hand, be quite
significant. The problem is the large variation in susceptibilities associated with earth
materials even when those materials are of the same rock type.
Recall that in applying the slab and elevation corrections to our gravitational observations,
we had to assume an average density for the rocks making up the corrections. Rock densities
do not vary much from rock type to rock type. Density variations of 0.5 gm/cm^3 are large.
Variations among different samples of the same rock type vary by even less. Therefore, we
can assume an average density for the correction and feel fairly confident that our
assumption is reasonable.
The IGRF models are regularly updated to account for secular variations. Given the latitude
and longitude of some point on the Earth's surface, the total field strength of the Earth's main
magnetic field can be calculated.
Consider a small two-dimensional survey. A plan view of such a survey is shown to the
right. One commonly used method of applying the main field correction is to linearly
interpolate the computed values of the main field at the corners of the survey throughout the
survey region. These interpolated values can then be subtracted from the field observations.
After applying this correction, you are left with that portion of the magnetic field that can not
be attributed to the Earth's main magnetic field.
This two-dimensional application of linear interpolation is only slightly more complex than
the one-dimensional linear interpolation used to reduce our gravity observations. Values of
the Earth's main magnetic field are first determined from the IGRF for each corner point of
the survey (c1, c2, c3, c4). To determine the strength of the main field at the point p, we first
perform two linear interpolations up the edges of the survey in the y direction to determine
the values of the field at the points t1 and t2. That is, first determine the value of the Earth's
main magnetic field at the point t1 by linearly interpolating between the points c1 and c4.
Then determine the value of the main field at the point t2 by linearly interpolating between
the points c2 and c3. Now, linearly interpolate in the x direction between t1 and t2. The
result is the two-dimensionally interpolated value of the field at the point p*.
*There is no reason why I have chosen to first interpolate in the y direction and then interpolate in the x
direction. I could have first interpolated in the x direction between c1 and c2 and then between c3 and c4.
You might ponder why we choose to take IGRF values for the corner points to interpolate from, rather than
going to those points and making direct measurements with our magnetometer. There is a proper answer which
explains the recommended method.
In the case of a given subsurface density distribution, the shape of the resulting gravity
anomaly is a function of the subsurface density distribution only. In fact, knowing the
gravitational anomaly produced by a simple shape such as a point mass is often enough to
guess what the shape of the gravity anomaly would be over a much more complicated
density distribution. Once you've determined the shape of the gravity anomaly that the
density distribution will produce, then you can make reasonable guesses about how the
anomaly will change as the density contrast is varied or as the depth to the density contrast is
varied. In addition, the anomaly will not change shape if the density distribution is moved to
a different location on the Earth, say from the equator to the north pole. The gravity anomaly
is a function of density distribution only.
Magnetic anomalies, on the other hand, are a function of two independent parameters: the
subsurface distribution of susceptibility and the orientation of the Earth's main magnetic
field. Change one of these parameters and you change the resulting magnetic anomaly. What
this means in practice is that magnetic anomalies over the same susceptibility distribution
will be different if the distribution is in a different location, say one located beneath the
equator versus one located beneath the north pole. Additionally, the magnetic anomaly over
a two-dimensional body such as a tunnel will look different depending on the orientation of
the tunnel, say east-west or north-south, even if the magnetic profile is always taken
perpendicular to the trend of the tunnel.
With these complexities in mind, we will not spend a great deal of time analyzing the shapes
of magnetic anomalies over simple structures; there are many computer programs available
that do this quite well. Rather, we will look at several simple examples and qualitatively
construct the magnetic anomalies over them so that you can get a better feeling for the
complexities involved and for how it might be done in the computer.
At the north (magnetic) pole, the Earth's main magnetic field, Fe, points straight down.
Because the buried sphere is composed of a material with a non-zero susceptibility, the
Earth's main magnetic field causes the sphere to produce an induced magnetic field. Field
lines associated with this induced field are shown by black lines, and the magnitude and
direction of the induced, anomalous field, Fai, at the surface of the earth are shown by the
blue arrows.
The total field, whose strength will be recorded on a proton precession magnetometer, will
be sum of the main field, Fe, and the induced, anomalous field, Fa. Notice that to either side
of the sphere, the anomalous field points in the opposite direction as the main field. Thus,
when the main field is removed from our observations we will observe negative values for
the anomalous field. Near the sphere, the anomalous field points in the same direction as the
main field. Therefore, when the main field is removed, we will observe positive values for
the anomalous field.
In this case, the anomalous magnetic field is symmetric about the center of the buried sphere,
is dominated by a central positive anomaly, and is surrounded on both sides by smaller
negative anomalies.
At the equator (magnetic), the direction of the Earth's main magnetic field is now horizontal.
It still induces an anomalous magnetic field in the metallic sphere, but the orientation of field
lines describing the magnetic field are now rotated 90 degrees. As in the previous case, these
field lines are indicated by the black lines, and the strength and direction of the anomalous
field at the surface of the earth are shown by the blue arrows. Above the sphere, the
anomalous magnetic field, Fa, now points in the opposite direction as the Earth's main
magnetic field, Fe. Therefore, the total field measured will be less than the Earth's main
field, and so upon removal of the main field, the resulting anomalous field will be negative.
On either side of the sphere, the anomalous field points in the general direction of the main
field and thus reinforces it resulting in total field measurements that are larger than the
Earth's main field. Upon removal of the main field contribution, these areas will show
positive magnetic anomalies.
As with the previous case, the resulting anomaly is again symmetrically distributed about the
center of the sphere. In this case, however, the prominent central anomaly is negative (even
though the susceptibility anomaly is positive) and is surrounded by two smaller positive
anomalies.
As in the previous examples, the Earth's main magnetic field induces an anomalous field in
surrounding the sphere. The anomalous field is now oriented at some angle, in this case 45
degrees, from the horizontal. By looking at the direction of the anomalous field, Fa, in
comparison with the Earth's main field, Fe, you can see that there will be a small negative
anomaly far to the south (to the right in the diagram) of the sphere, a large positive anomaly
just south of the sphere, and a small, broad, negative anomaly north of the sphere. Notice
that the magnetic anomaly produced is no longer symmetric about the sphere. Unless you are
working in one of those special places, like the magnetic poles or equator, this will always
be true.
We can consider the situation in the Southern hemisphere, at the same magnetic latitude, by
making some mental adjustments to the same diagram. If we take North to be to the right,
and reverse the direction of the main-field arrows, we have created the southern-hemisphere
setting. The shape of the induced, anomalous field will be the same, but the direction will
reverse as a consequence. The outcome is that the profile will have exactly the same shape!
A generalisation is that, where there is an isolated anomaly due to magnetisation by
induction, the maximum will always be on the equator side of the anomaly, and the
minimum will be on the pole side.
What does the anomaly look like if we take the profile EW rather than NS, in any of these cases?
From this simple set of examples, you now see that it is indeed more difficult to visually
interpret magnetic anomalies than gravity anomalies. These visual problems, however,
present no problem for the computer modeling algorithms used to model magnetic
anomalies. You simply need to incorporate the location of your survey into the modeling
algorithm to generate an appropriate magnetic model.