14 Geophysical data acquisition, processing and interpretation

Exploration geographic locations, or different depths in a drillhole.

model Define survey objectives
(mapping vs targeting) The location assigned to the measurement is usually the
Geophysical
model sensor location but may be some point between the trans-
mitter and the sensor. The resultant measurements, i.e. the
dataset, comprise a spatial series in the spatial domain.
DATA ACQUISITION
Expected Each of the measurements may comprise a single reading,
responses Survey design
or may be a series of readings made over an interval of time
Logistical & Field surveys
cost constraints
to form a time series in the time domain, or over a range of
Additional frequencies to form a frequency series in the frequency
geophysical
surveys domain. In some geophysical methods (e.g. electrical meas-
Topography DATA PROCESSING
urements), time- and frequency-series data provide the
Instrument Reduction
information about the nature of the rocks at the measure-
responses Enhancement
Noise
ment location; and in other methods (e.g. seismic and
characteristics Multiple products some kinds of electromagnetic measurements) they are
used to infer variations in the geology with distance from
DATA DISPLAY the measurement location. This might be lateral distance
1D/2D/3D Products from a drillhole, but is most commonly depth below a
Modify
based on surface reading. The latter are then known as soundings.
Multiple products
evolving
interpretation
Series of all types of geophysical data can be conveni-
ently treated as waves, and we use wave terminology
DATA INTERPRETATION
throughout the text. It is strongly recommended that those
Qualitative
Other readers unfamiliar with waves and their properties consult
geoscientific Quantitative
data (modelling) online Appendix 2 for details.

2.2.1 Absolute and relative measurements

Most kinds of geophysical surveys make absolute
measurements of the parameter of interest. This is not
Outcomes
(pseudo-geological map) always necessary; for some kinds of survey, notably gravity
(drilling targets) and magnetic surveys, relative measurements provide suf-
ficient information. In general, relative measurements have
Figure 2.1 The principal stages of a geophysical programme in mineral
exploration: from identifying the objectives of the geophysical survey(s) the advantage of being cheaper and easier to make than
through to providing an interpretation of the subsurface geology. absolute measurements.
A survey comprising relative measurements requires one
or more reference locations, called base stations, and the
2.2 Types of geophysical measurement
measurements are said to be ‘tied’ to the base stations. The
The parameters measured in the various types of geophys- absolute value of the parameter at the base stations may be
ical surveys described in Section 1.2 are continuous, i.e. known, in which case making comparative measurements
they vary in time or space and without gaps or end. The at other locations allows the absolute values to be deter-
variations are an analogue representation of the physical mined elsewhere. For example, when we say that the
property variations that occur in the subsurface. Measuring strength of Earth’s magnetic field at a base station is
or sampling an analogue signal at discrete times or at 50,000 nanoteslas (nT), we are referring to the absolute
discrete locations is known as digitisation. The continuous value of the field. If the field strength at a second station is
variation is then represented by a series of data samples 51,000 nT, then its relative value with respect to the base
forming a digital series, a form most convenient for storage station is +1000 nT (and the base station has a relative
and processing by a computer. value of –1000 nT with respect to the second station). If the
A geophysical survey consists of a series of measure- magnetic field at a third station has a relative strength of
ments made at different locations; usually different +2000 nT with respect to the base station, then it has an
 2.2 Types of geophysical measurement 15

absolute value of 52,000 nT. In terms of relative values, the a)

Vertical Horizontal
base station is assigned a value of zero. In large surveys gradient gradient

there may be a master base station from which a series of Gradient

Gradient distance
subsidiary base stations are established. This facilitates distance
surveying by reducing the distance that needs to be trav-
Sensors
elled to the nearest base station. Note that the accuracy of
the absolute value of a parameter obtained by relative
measurement from a base station is dependent on the
accuracy of the absolute value at the base station and the
accuracy of the relative measurement itself. b)
Z

PZZ
2.2.2 Scalars and vectors PZY PYZ
PZX
Physical quantities are classified into two classes. Those PYY Y
PYX
that have magnitude only are known as scalar quantities or PZ
PY
simply scalars. Some examples include mass, time, density PXZ
and speed. Scalar quantities are described by multiples of PX PXY
X
their unit of measure. For example, the mass of a body is PXX
described by the unit of kilogram and a particular mass is
described by the number of kilograms. Scalar quantities are Figure 2.2 Gradient measurements. (a) Vertical and horizontal
manipulated by applying the rules of ordinary algebra, i.e. gradiometers. (b) The three perpendicular gradients of each of the
three perpendicular components of a vector parameter P forming the
addition, subtraction, multiplication and division. For
gradient tensor of P, shown using tensor notation; see text for details.
example, the sum of two masses is simply the addition of
the individual masses.
Some physical quantities have both magnitude and dir-
2.2.3 Gradients
ection and are known as vector quantities or simply vectors.
Some examples are velocity, acceleration and magnetism. Sometimes it is useful to measure the variation in the
They are described by multiples of their unit of measure amplitude of a physical parameter (P) over a small distance
and by a statement of their direction. For example, to at each location. The difference in the measurements from
describe the magnetism of a bar magnet requires a state- two sensors separated by a fixed distance and oriented in a
ment of how strong the magnet is (magnitude) and its particular direction is known as the spatial gradient of the
orientation (direction). The graphical presentation and parameter. It is specified as units/distance in the measure-
algebraic manipulation of vectors are described in online ment direction, and so it is a vector quantity. As the
Appendix 1. measurement distance decreases, the gradient converges
Measuring vector parameters in geophysics implies that to the exact value of the derivative of the parameter, as
the sensor must be aligned in a particular direction. Often would be obtained from calculus applied to a function
components of the vector are measured. Measurements in describing the parameter field. For the three perpendicular
perpendicular horizontal directions are designated as the X directions X, Y and Z, we refer to the gradient in the X
and Y directions, which may correspond with east and direction as the X-derivative and, using the notation of
north; or with directions defined in some other reference calculus, denote it as ∂P/∂x. Similarly, we denote the
frame, for example, relative to the survey traverse along Y-derivative as ∂P/∂y and the Z-derivative as ∂P/∂z.
which measurements are taken. Usually the X direction is Gradients may be measured directly using a gradi-
parallel to the traverse. Measurements in the vertical are ometer, which comprises two sensors positioned a short
designated as Z, although either up or down may be taken distance apart (Fig. 2.2a). Alternatively, it is usually
as the positive direction depending upon accepted stand- possible to compute gradients, commonly referred to as
ards for that particular measurement. We denote the com- derivatives, directly from the non-gradient survey meas-
ponents of a vector parameter (P) in these directions as PX, urements of the field (see Gradients and curvature in
PY and PZ, respectively. Section 2.7.4.4).
16 Geophysical data acquisition, processing and interpretation

a) measurement point. We denote the derivative of the X

Horizontal gradient of response component of P (PX) in the X direction as ∂PX/∂x, and
the derivatives of the same component in the Y and Z
0 directions are ∂PX/∂y and ∂PX/∂z, respectively; and
similarly for the Y and Z components. They form a tensor
and are displayed and manipulated in matrix form:
0 1
Response ∂PX ∂PX ∂PX
B ∂x ∂y ∂z C
B C 0 1
B C PXX PXY PXZ
B ∂PY ∂PY ∂PY C
B C or @ PYX PYY PYZ A ð2:1Þ
B ∂x ∂y ∂z C
0
B C PZX PZY PZZ
B ∂P C
@ Z ∂PZ ∂PZ A
Source ∂x ∂y ∂z

Several components of the tensor are related as follows: PXY

b) ¼ PYX, PXZ ¼ PZX and PYZ ¼ PZY, so it is not necessary to
Horizontal gradient of response measure all of them. This means that less complex sensors
are needed and measurements can be made more quickly.
The full-gradient tensor of nine components, i.e. the gra-
0
dients in the three components in all three directions, pro-
vides diagnostic information about the nature of the source
of a geophysical anomaly. Tensor measurements are made in
Response
airborne gravity surveying (see Section 3.3.2) but are other-
wise comparatively rare in other geophysical surveys at
present. It seems likely that they will become more common
0 in the future because of the extra information they provide.

Source
2.3 The nature of geophysical responses
Figure 2.3 Horizontal gradient data across (a) a localised source, As described in Section 1.1 and shown schematically in
and (b) a contact. Note how the gradient response is localised near Fig. 1.1, geophysical surveys respond to physical property
the source edges.
contrasts, so changes in the local geology can produce
changes in the geophysical response of the subsurface.
Gradient measurements have the advantage of not being When the measured property of a target zone is greater
affected by temporal changes in the parameter being meas- than that of the host rocks, the contrast is positive; when
ured; the changes affect both sensors in the same way so any lower, it is negative. Typically the changes are localised,
difference in the parameter at each sensor is maintained. arising perhaps from a body of mineralisation or a contact
Gradient data are very sensitive to the ‘edges’ of sources. of some kind. These deviations from background values are
They comprise variations that are more spatially localised called anomalies. The simplest form of anomaly is an
than non-gradient data and so have an inherently greater increase or decrease of the measured parameter as the survey
spatial resolution (Fig. 2.3). The main disadvantage of gra- traverses the source of the anomaly. Often, though, peaks in
dient measurements is that they are very sensitive to vari- the anomaly are offset from their source and/or may be more
ations in the orientation of the sensor. Also, long- complex in form; for example, the response from magnetic
wavelength variations in the parameter, which produce very sources may comprise both an increase and an adjacent
small gradients, are often not large enough to be detected. decrease in response, forming a dipole anomaly.
The derivatives in the three perpendicular directions of Although the underlying physics of each geophysical
each of the three components of a vector parameter (P) method is different, some important aspects of the measured
(Fig. 2.2b) completely describe the parameter at the responses are the same. Figure 2.4 shows some general
 2.4 Signal and noise 17

a) l The amplitude of the anomaly decreases the further the

+ source is from the transmitter (if there is one). Also,
A
0
increasing source–detector separation causes the ampli-
tude of the response to decrease and to extend over a wider
–
area, i.e. there is an accompanying increase in wavelength
(λ) of the anomaly. The increased separation could be
Depth

because the source is at a greater depth below the surface

P=Positive P=Negative
or because the sensor is at a greater height above the
surface, as in the case of airborne measurements. When
b) the source varies in shape, this variation also affects the
+ anomaly, with increasing source width producing longer
0 wavelength responses (Fig. 2.4d).
– Figures 2.4 and 2.49a (see Section 2.11.4) illustrate some
important characteristics of many kinds of geophysical
Depth

responses. In general, the deeper the anomalous body

P=Smaller P=Larger
and/or the smaller its property contrast, the larger its size
must be in order for it to be detectable against the inevit-
c) able background noise (see Section 2.4). Also, anomalies
+ with the same amplitude and wavelength can be caused by
0
various combinations of source depth, geometry and con-
– trast with the host rocks. Without additional information
about these variables, the actual nature of the source of the
anomaly is indeterminable. This problem of ambiguity is
Depth

discussed further in Section 2.11.4.

In summary, whether or not an anomalous physical
d) property distribution produces a recognisable geophysical
+ response depends on its size and the magnitude of the
contrast between it and the surrounding rocks. In addition,
0
– the physical property contrast of a geological feature can
change markedly as the properties of the surrounding
rocks and/or those of the target feature change, both lat-
Depth

erally and with depth. This can significantly change the

nature of the geophysical response; it can form multiple
Figure 2.4 The general characteristics of a geophysical response and geophysical targets related to different parts of the same
how these change with variations in (a) sign of the physical property geological feature and even change the type of geophysical
contrast (ΔP), (b) magnitude of the contrast, (c) depth of the source measurements needed to detect it.
and (d) shape of the source.

2.4 Signal and noise

characteristics of (non-seismic) geophysical responses using
a simple shaped anomaly. A negative contrast produces an A measurement of any kind, but especially one made in a
anomaly that is an inverted image of that for an identical setting as complex and unpredictable as the natural envir-
positive contrast of the same source geometry (Fig. 2.4a). The onment, will be contaminated with unwanted information.
amplitude (A) of an anomaly depends on the magnitude of This unwanted information is known as noise and is a
the physical property contrast and the physical size of the source of error in a measurement, whilst the information
anomalous distribution. In general, increasing the property being sought in the measurement is known as signal. The
contrast increases the amplitude of the anomaly proportion- relative amounts of signal and noise in a measurement are
ally. Figure 2.4b shows that for two bodies of the same size the quantified by the signal-to-noise ratio (SNR). Ideally, one
anomaly has larger amplitude when there is a larger contrast. hopes that the amplitude of the signal, the signal level, is as
18 Geophysical data acquisition, processing and interpretation

high as possible and that the amplitude of the noise, the Wind is a common source of noise from the atmos-
noise level, is as low as possible in order to obtain an phere. It causes objects attached to the ground to move,
accurate measurement of the parameter of interest. As a e.g. trees and buildings, which produces noise in seismic,
general rule, if SNR is less than one it will be very difficult electromagnetic and gravity surveys. The movement of
to extract useful information from the measurement, wires linking sensors to recording equipment may also
although data processing techniques are available to create noise because of voltages induced by their move-
improve the situation (see Section 2.7.4). ment through the Earth’s magnetic field (see Section
Suppression of noise is of utmost importance and must 5.2.2.2). Wind turbulence also causes variations in the
be considered at every stage of the geophysical programme, position and orientation of geophysical sensors during
from data acquisition through to presentation of the data airborne surveys which affect the measurements.
for interpretation. Active geophysical methods usually As well as creating noise, natural phenomena may
allow the SNR to be improved by changing the nature of reduce the amplitude of the signal: for example, radioactiv-
the output from the transmitter, e.g. increasing its ity emitted from soil is attenuated when the soil is satur-
amplitude or changing its frequency. This advantage is lost ated by rainfall. The variability and unpredictability of
with passive methods, where the geophysicist has no con- natural phenomena cause noise levels to vary during the
trol over the natural ‘transmitter’. course of a geophysical survey.
The signal depends solely on the objective of the survey, Cultural noise includes the effects of metal fences, rail-
and geological responses not associated with the objective ways, pipelines, powerlines, buildings and other
of the survey constitute noise. Of course, any response of infrastructure (see Section 2.9.1). In addition, cultural fea-
non-geological origin will always be considered noise. As tures may radiate energy that causes interference, such as
data are revisited the information required from them may electromagnetic transmissions (radio broadcasts etc.),
vary, in which case so too do the representations of signal radioactive fallout and the sound of machinery such as
and noise in the data. A useful definition of signal is then motor traffic. Mine sites are particularly noisy environ-
‘what is of interest at the time’, whilst noise would then be ments, and noise levels may be so high as to preclude
‘everything else’; just as the saying goes, ‘One man’s trash is geophysical surveying altogether.
another man’s treasure’, so also one geoscientist’s signal is The two most troublesome forms of geological environ-
another geoscientist’s noise. mental noise are those associated with the shallow subsur-
Two basic types of noise affect geophysical measure- face and with topography; the latter are known as
ments. Firstly, there are effects originating from the local topographic or terrain effects. In both cases it is possible,
environment, i.e. environmental noise. Secondly, there is in principle, to calculate their effects on the data and
methodological noise, which includes unwanted conse- correct for them. To do so requires very detailed infor-
quences of the geophysical survey itself and of the process- mation about the terrain and/or physical properties of the
ing of the geophysical data. A feature in the data that is subsurface, which is often lacking. This is an example of
caused by noise is referred to as an artefact. It goes without the geophysical paradox (see Section 1.3). To fully under-
saying that identification and ignoring of artefacts is crit- stand the geophysical signal, and the noise, requires infor-
ical if the data are to be correctly interpreted. mation about the subsurface. However, it was to acquire
such information that the geophysical survey was
undertaken.
2.4.1 Environmental noise
The main types of environmental noise affecting the 2.4.1.1 Topography-related effects
different types of geophysical survey are summarised in Some examples of topography-related noise are shown
Table 2.1. Environmental noise can be categorised by its schematically in Fig. 2.5a. In rugged terrains, topography
origin; as either geological or non-geological. Geological creates noise by causing variations in the distance between
environmental noise is produced by the geological envir- geophysical transmitters and/or sensors and features in the
onment, including topography. Non-geological environ- subsurface. This changes the amplitude and wavelength of
mental noise includes sources in the atmosphere and the responses (see Section 2.3). These effects can some-
outer space, plus cultural responses associated with human times be accounted for by modifying the measurements,
activities. during data reduction. The accuracy with which this can be
 2.4 Signal and noise 19

Table 2.1 Common sources of environmental noise and the forms in which they manifest themselves for the various geophysical
methods. Specific details are included in the relevant chapters on each geophysical method.
Electrical and
Source of noise Gravity Magnetics Radiometrics electromagnetics Seismic

Regolith Changes in Oxidation of magnetic Concealment of bedrock High conductivity Changes in

thickness and mineral species responses leading to poor signal thickness and
internal variations Formation of Mobilisation of radioactive penetration and internal changes
in density causing maghaemite causing materials causing electromagnetic in velocity
spurious spurious anomalies responses that are not coupling with affecting
anomalies indicative of bedrock measurement array responses
Internal changes in (statics)
conductivity Reduction in
(groundwater, clays) the energy
causing spurious transmitted from
anomalies source
Superparamagnetic
behaviour (maghaemite)

Glacial Changes in Magnetic detritus Concealment of bedrock Internal changes in Changes in

sediments thickness and causing spurious responses conductivity causing thickness and
internal variations anomalies Mobilisation of radioactive spurious anomalies internal changes
in density causing materials causing in velocity
spurious responses that are not affecting
anomalies indicative of bedrock responses from
below (statics)

Permafrost and Changes in ice Concealment of bedrock Internal changes in Changes in

snow cover content causing responses conductivity causing thickness and
spurious spurious anomalies internal changes
anomalies in velocity
affecting
responses from
below (statics)

Hydrological Formation of low Dissolution/precipitation Changes in groundwater Formation of

density (i.e. transportation) of salinity causing changes dissolution
dissolution soluble radioactive in conductivity features in
features in elements causing responses Movement of carbonate rocks
carbonate rocks that are not indicative of electrolytes creating that scatter
causing spurious the presence of K, spurious anomalies seismic waves
anomalies U and Th

Atmospheric Turbulence Turbulence during Turbulence during Turbulence during Wind (water
phenomena during airborne airborne surveys airborne surveys airborne and some waves,
and ground Uneven distribution of ground surveys movement of
surveys radioactive aerosols Noise spikes due to tree roots etc.)
Movement of gaseous lightning strikes (sferics) and rain-related
radioactive species noise
Moisture suppressing
responses from the
ground

Extraterrestrial Temporal changes Magnetic fields Radioactivity of cosmic

phenomena in gravity due to associated with origin
the position of the processes occurring in
Sun and Moon the ionosphere
20 Geophysical data acquisition, processing and interpretation

Table 2.1 (cont.)

Electrical and
Source of noise Gravity Magnetics Radiometrics electromagnetics Seismic

Topography Terrain-related Terrain-related Terrain-related responses Terrain-related Variable relative

responses responses Variable relative positions responses positions of
Variable relative Variable relative of sensors and anomalous Variable relative sensors and
positions of positions of sensors bodies positions of transmitters, anomalous
sensors and and anomalous bodies sensors and anomalous bodies
anomalous bodies bodies

Man-made Responses due to Responses from Responses from materials Responses from metallic Sound of motor
large buildings ferrous objects, roads, created by radioactive objects, electrical vehicles and
and excavations, pipelines, etc. fallout from nuclear powerlines and heavy
such as open pits constructed of explosions and reactors electromagnetic machinery
magnetic materials transmissions
and electrical
powerlines

a)
to this type of surveying in very steep terrain than the less-
Survey aircraft unable
to maintain manoeuvrable fixed-wing aircraft. Whether drape is main-
constant ground
clearance tained or not depends upon the width (wavelength) and
Prevention
of access slope of the topographic variation, the speed of the aircraft
Localised Responses from rocks and its climb capability. Small topographic features tend to
terrain-related not below the
anomalies sensor Unexpected be ignored (Figs. 2.6 and 2.7a) producing spurious noise in
physical-
property the measured geophysical response. Another problem for
contrasts
airborne surveys is that an aircraft can descend faster than
it can climb. When the terrain begins to rise rapidly in
b) Lateral & vertical
front of the aircraft, the survey pilot must anticipate the
Variable separation
Degenerated of sensor &
changes in
physical
need to gain height so as to pass safely over the hill, and so
input/response bedrock
properties begins ascending before the feature is reached. The result is
Altering
an increase in terrain clearance adjacent to the hill. The
Cover
Transportation
physical-
property
easier descent on the other side of the hill often means it is
of source
materials
contrast possible to re-establish the specified terrain clearance rela-
Bedrock tively quickly. The same occurs when surveying across a
valley. The result is illustrated by the actual flight paths
from an aeromagnetic survey across part of the Hamersley
Figure 2.5 Environmental noise associated with (a) topography and
(b) the near-surface environment. Cover could be unconsolidated iron-ore province in northern Western Australia (Fig. 2.6).
sediments, regolith, glacial till or ice. The terrain clearance not only varies across the hill, but
also depends on the direction the aircraft is travelling with
respect to the topography. For surveys where adjacent
achieved depends on a number of factors, notably how well traverses are flown in opposite directions (see Section
the local geology (and associated physical property vari- 2.6.3.3), the terrain clearance is fairly consistent for alter-
ations) and topography are known. nate traverses and different for adjacent traverses resulting
Historically, airborne surveys were usually conducted at in corrugations, i.e. traverse-parallel artefacts caused by
a constant barometric altitude above the undulating ter- readings being anomalously high and then low on alternat-
rain. Modern airborne surveys are draped over the terrain ing traverses.
as the aircraft attempts to maintain constant height above When measurements are made on the ground the geol-
the undulating ground surface, i.e. they are flown at ogy may be adjacent to, or even above, the sensor
constant terrain clearance. Helicopters are better adapted (Fig. 2.7b). This is also a common occurrence for
 2.4 Signal and noise 21

a) Flight direction a)

800
Nominal Sensor Sensor height Sensor height
Height (m) 600 clearance height
400
200

0 Ideal geometry Reading too low Reading too high

clearance (m)

600
Terrain

400 b)

200

0 Constant Sensor
Sensor height Sensor height
clearance height
Location on flight-line

b) Flight direction Ideal geometry Reading too high Reading too low

800 Figure 2.7 Influence of topography on geophysical measurements. (a)

Height (m)

600 Effects when sensor height varies, as with loose-draped airborne

400 surveys. (b) Effects when the sensor is maintained at constant terrain
200 clearance, as in ground and close-draped airborne surveys in rugged
0 terrain. Note how the effects of the topography on the reading are
opposite in (a) and (b).
clearance (m)

600
Terrain

400

200
0
measured geophysical response is a distortion of that for a
Location on flight-line
horizontal survey surface.
Other features of the terrain producing similar effects to
0 2000
Draped flight path
Metres
topography include open-pit quarries and mines, tall
Constant altitude flight path
obstructions such as buildings and other infrastructure,
Figure 2.6 Survey height and associated variations in terrain
and trees and thick vegetation. The availability of high
clearance for draped and constant barometric height surveys flown resolution digital terrain information from airborne and
in opposite directions across a ridge. The draped paths are actual satellite sensors, and stereo photography, is an important
flight paths from an aeromagnetic survey in Western Australia. development in terms of compensating for terrain effects.
Based on diagrams in Flis and Cowan (2000). The height of the terrain above sea level can be obtained
from airborne geophysical surveys by combining the air-
craft height above the terrain, measured with a radio
low-level helicopter-borne surveys conducted in rugged
altimeter, with the GPS-derived height. A digital elevation
terrains. Both situations can cause anomalous responses
model (DEM), also known as a digital terrain model
which might be misinterpreted (Mudge, 1998). Making a
(DTM), for the survey area can be created in this way.
measurement in a gully or adjacent to a cliff means more of
A useful source of radar-derived terrain information is the
the ‘geology’ is closer to the sensor and can create anomal-
Shuttle Radar Topography Mission (SRTM) dataset
ously high readings, with the opposite occurring when on
(Cowan and Cooper, 2005). As a general rule, it is good
the top of ridges etc. Note that this is exactly the opposite
practice to have a terrain model available when interpret-
of the effects on imperfectly draped airborne surveys in
ing geophysical data.
Fig. 2.7a.
Another problem caused by topography is the distortion
of the geophysical response. This is particularly a problem 2.4.1.2 Near-surface and deep-seated responses
for electrical and electromagnetic surveys where the near- A significant source of geological noise is the near-surface
surface flow of electrical current is strongly influenced by environment (Fig. 2.5b). Regolith can present major prob-
the shape of the (conductive) terrain (see Section 5.6.7.3). lems for geophysical surveying (Doyle and Lindeman,
Also, surveys conducted on sloping terrain need to account 1985), as can cover such as sand dunes, glacial deposits
for the terrain slope in the analysis of the data, because the and snow (Smee and Sinha, 1979), and permafrost