Geophys. J. Int. (2004) 159, 565–572 doi: 10.1111/j.1365-246X.2004.02420.

Sea-surface observations of the magnetic signals of ocean swells

F. E. M. Lilley,1 A. P. Hitchman,2 P. R. Milligan3 and T. Pedersen4

1 Research School of Earth Sciences, Australian National University, Canberra, ACT 0200, Australia. E-mail: [email protected]
2 Research School of Earth Sciences, Australian National University, Canberra, ACT 0200, Australia. E-mail: [email protected]
3 Geoscience Australia, PO Box 378, Canberra, ACT 2601, Australia. E-mail: [email protected]
4 Research School of Earth Sciences, Australian National University, Canberra, ACT 0200, Australia. E-mail: Tina [email protected]

Accepted 2004 April 22. Received 2004 February 20; in original form 2003 June 3

Downloaded from https://academic.oup.com/gji/article-abstract/159/2/565/2065304 by guest on 19 January 2020

SUMMARY
Ocean swells have a magnetic signal, caused by the motional induction of sea water moving
in the steady main magnetic field of Earth. To check the character of such signals at the sea
surface, a magnetometer has been set free from a ship to float unrestricted on the surface of
the ocean for periods of several days. The path of the floating magnetometer was tracked by

GJI Marine geoscience

satellite; this procedure enabled also the eventual recovery of the magnetometer by the ship.
Superimposed upon a background of slow change of magnetic field, as the magnetometer
drifted across different patterns of crustal magnetization, are high-frequency signals generated
by the strong ocean swell present at the time. These wave-generated signals are typically up to
5 nT trough-to-peak, consistent with theory for their generation by ocean swells several metres
trough-to-peak in height.
The power spectra of the magnetic signals show a consistent peak at period 13 s, appropriate
for the known characteristics of ocean swell in the area. The power spectra then exhibit a
strong (−7 power) fall-off as period decreases below 13 s. This strong fall-off is consistent
with oceanographic observations of the spectra of surface swell, combined with motional
induction theory.
Key words: magnetic, motional induction, ocean, signals, swell, waves.

observations of the magnetic signals of ocean waves are also re-

1 I N T RO D U C T I O N
ported by Fraser (1966) and Watermann & Magunia (1997). The
A magnetic signal, generated by motional electromagnetic induc- topic is naturally closely associated with the electric signals of such
tion, accompanies ocean waves and swell. Sea water moves across waves, which are studied in papers such as Cox et al. (1978) and
the flux lines of the main magnetic field of the Earth and in so Hemer et al. (1999). More widely, the topic of electromagnetic
doing generates electric currents that flow through the sea water. induction in the ocean is reviewed by Larsen (1973) and Palshin
These electric currents have their own magnetic fields, which are (1996).
the magnetic fields generated by the wave motion.
To provide sea-surface information on such wave signals, mea-
surements were made in 1998 in the Southern ocean off South Aus-
tralia. The Southern ocean was a suitable place for the observations, 2 T H E O RY
as a result of the presence there of ocean swells, which consis-
tently are several metres in height. The observations were part of 2.1 Motional induction
the Southern Waters of Australia Geoelectric and Geomagnetic In- Reference is made here to the theory of motional induction of ocean
duction Experiment (SWAGGIE), described by Hitchman (1999) swell as developed by Weaver (1965) and Weaver (1997). An ocean
and Popkov et al. (2000). Attention had been drawn to the magnetic swell is analysed as a simple harmonic wave on the surface of sea
signals of ocean swell by recent aeromagnetic surveys offshore in water, which is deep in the sense that the depth is much greater than
Australia (Milligan & Barton 1997) and by the presence of wave sig- the horizontal wavelength of the wave. The wave propagates in the
nals (both surface and internal) in natural electromagnetic induction x direction with wavenumber m and angular frequency ω: the phase
data used to study seafloor conductivity structure. speed of the wave c is thus given by
The magnetic fields of ocean waves have been the subject of
theoretical investigation by Longuet-Higgins et al. (1954), Weaver c = ω/m. (1)
(1965), Beal & Weaver (1970), Podney (1975), Chave (1984) and
Weaver (1997). Maclure et al. (1964) and Ochadlick (1989) report The magnetic signal b(x, t) in the direction of the main field of Earth
measurements in agreement with the theory of Weaver (1965) and (the component measured by a total-field magnetometer) generated

C 2004 RAS 565
566 F. E. M. Lilley et al.

by the ocean swell is then given by Note that the value of W at the sea surface (for which h = 0
and so A = 1) is linearly proportional to each of the variables T, F
b(x, t) = b F exp(iωt − imx), (2) and σ . The reduction of the signal with height above sea surface is
where t denotes time and bF , the amplitude of the magnetic signal exponential and, of the latter three variables, is a function of T only.
of the ocean swell, is given by The height at which the signal falls to 1/e of its surface value (the
1/e folding height) is given by gT 2 /4 π 2 , that is by g/ ω2 where
T agµ0 σ F
bF = (cos2 θ cos2 I + sin2 I ) exp(−4π 2 h/gT 2 ). (3) ω is the angular frequency of the wave. For the Southern ocean off
8π South Australia, representative values for F, σ and I are 60 280 nT,
Here, T is the period of the wave (s), a is the amplitude (m), g is 4.4 S m−1 and −68.4◦ , respectively. The electrical conductivity value
the acceleration as a result of gravity, µ 0 is the permeability of free results from values for surface practical salinity of 35.5 (dimension-
space, σ is the electrical conductivity of the sea water, F is the total less units) and temperature of 16◦ C taken from maps in Tomczak &
magnetic field intensity and I the inclination of the main magnetic Godfrey (1994), combined with tables for the electrical conductivity
field of the Earth, θ is the direction of wave travel measured as an of sea water in Bullard & Parker (1970). Thus, for the sea surface
angle positive eastwards from the direction of magnetic north and h (taking g = 9.8 m s−2 )

Downloaded from https://academic.oup.com/gji/article-abstract/159/2/565/2065304 by guest on 19 January 2020

is the height of the magnetometer above the sea surface. Introducing
a constant K given by W = 0.0176(cos2 θ + 6.38)T. (9)

K = gµ0 /8π, (4) This relationship is shown in Fig. 1.

The exponential decrease with altitude for the area will be the
a geometric function G(θ , I ) given by quite general one, given for A in eq. (6) and shown as a function of
wave period in Fig. 2. Note that in the altitude range 50 to 200 m
G(θ, I ) = cos2 θ cos2 I + sin2 I, (5)
(a common height for aeromagnetic surveys), the attenuation of the
an exponential decay with height A(h, T) given by sea-surface signal is considerable and sensitive to the period of the
swell.
A(h, T ) = exp(−4π 2 h/gT 2 ) (6)
and notation W for the amplitude of the total-field magnetic signal 2.2 Spatial gradients
per unit amplitude of wave given by
In the present experiment, the sensor is in motion. The data recorded
W = b F /a, (7) may thus be affected by the spatial non-uniformity of the ambient
then W may be expressed magnetic field of the Earth, through which the sensor is moving.
Attention will now be given to three possible effects of this sensor
W = K T Fσ G(θ, I )A(h, T ). (8) motion: (i) oscillation of the sensor up and down with the waves

MAGNETIC FIELD (W) vs WAVE DIRECTION (θ)

8

60 s
7
55 s

50 s
6

45 s
5
40 s
W (nT m−1)

35 s
4

30 s

3 25 s

20 s
2
15 s

1 10 s

T=5s
0
0 50 100 150 200 250 300 350
WAVE DIRECTION θ (degrees)

Figure 1. Sea-surface values of swell magnetic signal (W ) for different periods of swell, as a function of direction of travel of swell (θ ), for the SWAGGIE
experiment off South Australia. A maximum period of 60 s has been included to illustrate the behaviour of eq. (9): in practice the local swell periods are less
(for example 13 s in the example below).

C 2004 RAS, GJI, 159, 565–572
 Magnetic signals 567

A (h,T) vs ALTITUDE (h) FOR DIFFERENT VALUES OF PERIOD (T)

1

0.9

0.8

0.7

0.6

0.5
A

Downloaded from https://academic.oup.com/gji/article-abstract/159/2/565/2065304 by guest on 19 January 2020

30 s
0.4 1/e
25 s
0.3
20 s

0.2
15 s 2
(1/e)
0.1 10 s
T=5s
0
0 50 100 150 200 250 300 350 400 450 500
ALTITUDE (m)

Figure 2. Exponential decrease of swell signal with height of detector above sea surface, for different periods of swell.

themselves; (ii) oscillation of the sensor laterally with the waves; below. The steepest horizontal gradient sampled by a floating mag-
(iii) secular movement of the sensor as a result of its free drift later- netometer in deep water will there be seen to be 0.01 nT m−1 and in
ally across the geological magnetic pattern of the seafloor. more shallow water 0.04 nT m−1 . Thus oscillations with amplitude
1 m in such gradients will produce oscillatory signals in the data
of amplitudes 0.01 and 0.04 nT, respectively (trough to peak 0.02
2.2.1 The dipole field of the Earth and 0.08 nT). While the horizontal gradient perpendicular to the
line of drift may differ from that along the line of drift, it would be
To estimate the effects of the vertical movement, a floating magnetic surprising if it were greatly different.
sensor will move vertically through a distance equal to the wave For vertical gradients, the nearest available aeromagnetic data
height (twice the amplitude of the wave). Simple dipole models of have been taken, which are for the Eyre Peninsula (see Fig. 3 below).
the main magnetic field of the Earth predict that the vertical gradient These data have been upward continued to a height of 2000 m to
of the total field will be −0.02 nT m−1 at the equator and −0.04 nT mimic the sensor–source separation for a sensor floating on water
m−1 at the poles. A magnetometer floating in the Southern ocean off 2000 m deep and a map of the vertical gradient of the total-field
South Australia in a swell with height 2 m will thus pick up a spatial calculated at that height. A histogram of the vertical gradient values
signal of order 0.06 nT (trough to peak). While not negligible, this then shows a distribution with a peak at −0.008 nT m−1 and with 95
signal is small relative to the motional induction signal, predicted per cent of the values less than 0.06 nT m−1 in amplitude. Oscillation
from Fig. 1 to be 3 or 4 nT for a typical swell period of 13 s. of the magnetometer sensor in such a vertical gradient field will give
The horizontal gradient of a dipole model for the field of the Earth signals that again are small relative to the motional induction signal
is everywhere less than the vertical gradient. Zonally the horizontal predicted.
gradient is zero generally and azimuthally the horizontal gradient Finally, under the effects of wind and currents, a drifting sensor
has maxima of some 0.006 nT m−1 in mid-latitudes between values will traverse the magnetic pattern of the geological crust. As will
of zero at the poles and equator. In the ocean off South Australia, be seen from the examples presented, the effect of this changing
a lateral oscillation of say 1-m amplitude in ocean swell will thus background signal will be to give a slowly changing baseline against
cause a spatial gradient magnetic signal of 0.01 nT trough to peak, which the swell signals are seen. Measured as a change with time,
again small relative to the motional induction signal. the geologically caused signals are very slow and will not affect
time-series analysis of the wave signals.

2.2.2 The crustal magnetic field

3 METHOD
The magnetic pattern of the crustal rocks over which the magne-
tometer is floating will also contribute horizontal and vertical gra- On two separate occasions a floating magnetometer (floater-mag)
dients in which the magnetometer is oscillated by wave motion. was released from the vessel ORV Franklin and left to float free
The horizontal gradient in the direction of drift is measured directly for approximately 4 days, while the ship carried out other work. The
by the floating magnetometer and will be evident in data presented complete floating package also carried a satellite transmitter, so that,

C 2004 RAS, GJI, 159, 565–572
568 F. E. M. Lilley et al.

135˚ 136˚ 137˚

N.T.
Eyre Peninsula
Qld
W.A.
S.A.
N.S.W.

Vic

-34˚ Tas

50 km
Spencer
Gulf

Downloaded from https://academic.oup.com/gji/article-abstract/159/2/565/2065304 by guest on 19 January 2020

OTH

Anchor mags 1
3
-35˚ 2
c

b a TW

Floater mag 2

Floater mag 1
-36˚
floating magnetometers
anchored magnetometers
10
00
20

seafloor magnetometer
00
30

land magnetometer
00
40

conductivity anomaly
5000
00

Figure 3. Paths of the two floating magnetometer deployments, floater 1 and floater 2, off South Australia. The land site used as reference, OTH, is shown.
Water depth contours are in metres.

as it drifted under the influence of winds and ocean currents, the netometer package kept it upwind of the satellite transmitter, which
position of the magnetometer could be tracked and the instrument was tethered at the end of 6 m of buoyant rope. Windage drag on a
eventually recovered. Care was taken in the package design to keep flag beacon, at the satellite buoy, helped keep the (slightly magnetic)
the magnetometer itself in a non-magnetic environment and free satellite buoy downwind and remote from the magnetometer. The
of stray magnetic fields. A description of the experiment is given magnetometer took a reading at intervals of 3 s. The magnetometer
in Hitchman et al. (2000), in which the floater-mag data are used arrangement was an adaption of, and similar to, the package used
to determine induction arrows, using a land station as reference. by Lilley et al. (2001) to make measurements vertically through the
Hitchman et al. (2000) note the presence of a swell signal on the ocean column in the East Australian current.
floater-mag records.
This paper presents an analysis of the floater-mag data for the
swell signal recorded. The paths followed by the drifting magne- 5 R E S U LT S
tometer, termed floater 1 and floater 2, are shown in Fig. 3.
Earlier measurements, testing the equipment, had been made in 5.1 Time-series
1997 in the SODA experiment (Lilley et al. 2001). The magnetome-
ter was then tethered to a stationary ship and floated several hundred Total-field variations recorded by floater-mag, and those from land
metres from it downwind. reference stations, are shown in Fig. 4. The floater-mag time-series
have been subdivided for analysis into different lengths and var-
ious reference stations have been used, depending on availability.
Floater 1 was deployed before the SWAGGIE seafloor and land ar-
4 EQUIPMENT
rays began recording, thus Canberra Magnetic Observatory is used
The equipment used was a total-field magnetometer, set up for free- as a reference station for segment floater 1a in Fig. 4. Data from a
floating marine operation. The recording electronics and power sup- seafloor instrument were available for segment floater 1b, but at a
ply were housed in one water-tight sphere and the magnetometer lower sampling rate. By the time floater 2 was deployed, the land
sensor was housed in another, separated 1.8 m from the first by a array of magnetometers had commenced recording, enabling OTH
rigid non-magnetic frame. In the water, a drogue fitted to the mag- (shown on Fig. 3) to be used as a reference station.

C 2004 RAS, GJI, 159, 565–572
 Magnetic signals 569

40 5
20
Floater 1a
0
-20 9/4/98 10/4/98

total field (nT)

15 18 21 00 03 06 09 12 15 18 21 00 0

40 Floater 1b
20
0
-20 -5
-40
-60
11/4/98 12/4/98
-80
00 03 06 09 12 15 18 21 00 03
-10
9:00 9:05 9:10
80 Floater 2a Universal time (h:mm)
60
40
20 Figure 5. Example of time-series for floater 1a over a period of 10 min,

Downloaded from https://academic.oup.com/gji/article-abstract/159/2/565/2065304 by guest on 19 January 2020

0
showing the magnetic signals of individual ocean swells. The zero level for
-20
the magnetic field is arbitrary. The significant signal height for this example
-40
-60
14/4/98 15/4/98 is 4.6 ± 0.2 nT.
21 00 03 06 09 12 15 18
total field (nT)

140
120
Floater 2b
100 seafloor, as it was carried by ocean currents and winds. However,
80 when the time-series from the floater-mag are expanded, a short-
60
40
period signal is consistently evident. The signal is variable in am-
20 plitude, typically ranging up to 5 nT in trough-to-peak value. An
0 example is shown in Fig. 5.
-20
-40
Adapting, to Fig. 5, the oceanographic concept that a significant
-60 signal height is the mean height of the one-third highest waves (Pond
-80
15/4/98 16/4/98 & Pickard 1983), a significant signal height for the data in Fig. 5 is
-100
21 00 03 06 09 12 15 18 calculated to be 4.6 nT, with a standard error of 0.2 nT.
120
100
Floater 2c
80
60
40
5.2 Power spectra
20
0
Power spectra for each of the floater-mag time-series are presented in
-20 Fig. 6, together with spectra for the same variations recorded at each
-40
of the reference stations. The important feature, for the purposes of
-60
-80
the present paper, is the consistent peak in all the floater-mag spectra
-100
Ref at approximately 13 s. It is this spectral peak that corresponds to the
16/4/98 17/4/98 Floater
-120
21 00 03 06 09 12 15 18 swell signal and that will now be examined more carefully.
Universal time (h) It should at this stage be mentioned that Pc3 pulsations occur
naturally in the geomagnetic field over a period range that includes
Figure 4. Simultaneous total-field time-series for deployments of the float- 13 s (Campbell 1997) and that circumstances could arise in which
ing magnetometer (thick trace, as a result of the swell signal) and a land
such pulsations confuse the recognition of the magnetic signal of
reference station (thin trace). The slow drifts of the time-series result from
the passage of the floating magnetometer across the crustal magnetic pattern
ocean swell. In the present case however such pulsations would not
of the seafloor. The land records indicate times of magnetic substorms: these be expected to occur continuously over days as shown in Fig. 4, or
are recorded differently by the floating instrument because of local induction have the clear spectral peak shown by the floater-mag data in Fig. 6.
effects (Hitchman et al. 2000). Further, inspection of the Canberra Magnetic Observatory record
for the time of Fig. 5 shows no particular pulsation activity and note
that the observatory records show no spectral peak at 13 s in the top
These time-series for the floater-mag deployments show large- left panel of Fig. 6. There is thus no reason to attribute the signal in
scale changes in the total-field occurring slowly, as the instrument Fig. 5 to geomagnetic pulsations.
drifts across the crustal magnetization pattern. Because floater-mag The swell signal is thus ubiquitous in the floater-mag data, and
moved with the ocean currents, its observations record both tempo- attention is now focussed particularly on the records of floater 1.
ral and spatial total-field variations. The reason is that as a result of their observation in deeper water,
Floater 1 was deployed in deep water (2000–3000 m) and (partic- these data satisfy more exactly the condition adopted in the theory
ularly in floater 1b) the crustal field is evident as a broad anomaly in of an infinitely deep ocean. In contrast, for floater 2 data, the ocean
the time-series. In contrast, floater 2 was deployed above the conti- depth of 100 m is of the same order of magnitude as the wavelength
nental shelf in water approximately 100 m deep and may therefore of the swell. Another benefit of the deep-water data is the lesser
be more strongly influenced by the crustal anomaly field. There magnetic gradients at the surface as a result of the crustal magnetic
is some evidence in Fig. 4 that the time-series from floater 2 are pattern.
characterized by larger aperiodic variations than the time-series of A power spectrum for the entire time-series of floater 1 is shown
floater 1. in Fig. 7. The power spectrum shows a clear peak in power at 13.3 s,
Thus, large and slow changes in Fig. 4 are the result of the mag- caused by the ocean swell. The period of the swell found in this
netometer moving across the crustal magnetization patterns of the analysis corresponds closely to the findings of Hemer & Bye (1999)

C 2004 RAS, GJI, 159, 565–572
570 F. E. M. Lilley et al.

Frequency (Hz) Frequency (Hz)

100 10-1 10-2 10-3 10-4 100 10-1 10-2 10-3 10-4
102 102
Floater 1a Floater 1b

100 100

Power (nT2.s)

Power (nT2.s)
10-2 10-2

10-4 10-4

10-6 10-6

Downloaded from https://academic.oup.com/gji/article-abstract/159/2/565/2065304 by guest on 19 January 2020

10-8 10-8
100 101 102 103 104 100 101 102 103 104
Period (s) Period (s)
Frequency (Hz)
100 10-1 10-2 10-3 10-4
102
Floater 2a

100
Power (nT2.s)

10-2

Ref Floater
10-4

10-6

10-8
100 101 102 103 104
Period (s)
Frequency (Hz) Frequency (Hz)
100 10-1 10-2 10-3 10-4 100 10-1 10-2 10-3 10-4
102 102
Floater 2b Floater 2c

100 100
Power (nT2.s)

Power (nT2.s)

10-2 10-2

10-4 10-4

10-6 10-6

10-8 10-8
100 101 102 103 104 100 101 102 103 104
Period (s) Period (s)

Figure 6. Power spectra for the floater-mag total-field variations, together with similar spectra for the land reference stations. Except for floater 1a, for which
the land reference is Canberra Magnetic Observatory, the land-station data have a lesser digitizing interval than the floater data, so their power spectra do not
go to periods as short as the swell period (13 s). However, the ubiquity of the peak in the floater spectra at the swell period of 13 s is clear, as is the absence of
land data signal at this period.

and Hemer et al. (1999) for the swell period in the South Australian Thus, for signals measured by total-field magnetometers, a fall-
sea in 1998 April. off with frequency with a theoretical slope of −7 is predicted for the
Under many circumstances the wave-energy spectrum of the power. A double logarithmic plot of the power versus the frequency
ocean is approximately represented by the Phillips spectrum of the magnetic signal shows this phenomenon well, see Fig. 7.
(Phillips 1977). The Phillips spectrum has a characteristic fall-off
with frequency of power −5. Inspection of eq. (3) shows that if a 2 ,
6 DISCUSSION
representing the power of the swell, has a frequency dependence of
−5, then bF 2 will have a frequency dependence of −7, as the T in It is a straightforward explanation of the fluctuations seen in Fig. 5
eq. (3), when squared, will contribute a further factor of ω−2 . that they are magnetic variations associated with ocean swell. The

C 2004 RAS, GJI, 159, 565–572
 Magnetic signals 571

2
10

1
10

POWER
0
10

Downloaded from https://academic.oup.com/gji/article-abstract/159/2/565/2065304 by guest on 19 January 2020

−1
10
−2 −1 0
10 10 10
FREQUENCY (Hz)

Figure 7. Power spectrum for the deep-water floater 1 time-series of total magnetic field, plotted with a line of slope −7 as predicted by theory.

Figure 8. Map of significant wave height (swell plus wind waves) and swell direction forecast on the basis of numerical modelling for the time of the floating
magnetometer observations shown in Fig. 5. Nominally, the time is 00 00 h UT on 1998 April 10. For the SWAGGIE area, the swell direction is 45◦ geographic
(i.e. 39◦ magnetic) and the significant swell height is in the range 2 to 3 m.

pattern in Fig. 5 shows packets of waves, of varying heights, as are of monitoring swell heights in the open ocean. Further tests where
known to occur in practice. Weather forecasts at the time by the control information on actual swell heights is available would be
Australian Bureau of Meteorology model a swell height of between valuable. Further refinements would include observations over con-
2 and 3 m for that part of the Southern ocean, as shown in Fig. 8, tinental shelves, where the nature of the waves should change (being
and a direction of swell travel of 45◦ geographic (39◦ magnetic). no longer deep-water waves) and an investigation of the effects of
Global hindcast data from the NOAA Wavewatch III internet site the electrical conductivity of the water diminishing with depth, as
(http://polar.ncep.noaa.gov/waves) give a swell period of 13 s for the temperature reduces.
area at the time in question, in agreement with the spectral peaks For some purposes, such as aeromagnetic surveying, the magnetic
observed. signal of ocean swell is a source of error and it is well appreciated
The values of 39◦ for θ and 13.3 s for T may now be used in that a way to suppress such signals is by increasing the height of
eq. (9) to give a value for W of 1.63 nT m−1 . This value, combined measurement above the sea surface. The exponential decay of the
with the result from Fig. 5 of a significant magnetic signal height of swell signal with height is then straightforward and the 1/e folding
4.6 ± 0.2 nT, gives a significant swell height of 2.8 ± 0.1 m. The height is a function of swell period only. With such a strategy, the
swell height predicted for the SWAGGIE area in Fig. 8 is seen to be benefits of suppressed swell signal are offset by reduced definition
greater than 2 m and less than 3 m, so that there is agreement within of the observed magnetic pattern, which is also a consequence of
the approximate nature of Fig. 8. increasing height.
The agreement suggests that a floating total-field magnetome- However, an important result of the present work has been to
ter, which has no moving parts, may be a simple independent way demonstrate the −7 power fall-off of the magnetic signal with

C 2004 RAS, GJI, 159, 565–572
572 F. E. M. Lilley et al.

increasing frequency. This phenomenon causes the magnetic signal Chave, A.D., 1984. On the electromagnetic fields induced by oceanic internal
to have a narrow frequency band. When the ocean swell has travelled waves, J. geophys. Res., 89, 10 519–10 528.
some distance from its place of origin, this result should apply quite Cox, C.S., Kroll, N., Pistek, P. & Watson, K., 1978. Electromagnetic fluctu-
generally. Such an implied sharp spectral peak in the swell magnetic ations induced by wind waves on the deep sea floor, J. geophys. Res., 83,
signal may aid its filtering in aeromagnetic and other applications 431–442.
Fraser, D.C., 1966. The magnetic fields of ocean swell, Geophys. J. R. astr.
of marine magnetic measurements.
Soc., 11, 507–517.
Hemer, M.A. & Bye, J.A.T., 1999. The swell climate of the South Australian
7 C O N C LU S I O N S sea, Trans. R. Soc. South Aust., 123, 107–113.
Hemer, M.A., Heinson, G.S., Bye, J.A.T. & White, A., 1999. Wave energy
A research cruise off the coast of South Australia has provided an turbulence spectra from the measurement of electric fields in the ocean,
opportunity to observe the magnetic signals of ocean swell, gener- in, The wind-driven air-sea interface, pp. 49–56, ed. Banner, M.L., Univ.
ated by motional induction. The place of the experiment has been of NSW, Sydney.
favourable, as a result of the occurrence there of swell generated by Hitchman, A.P., 1999. Interactions Between Aeromagnetic Data and Elec-
the storms of the Southern ocean. Magnetic signals measured by a tromagnetic Induction in the Earth, PhD thesis, The Australian National

Downloaded from https://academic.oup.com/gji/article-abstract/159/2/565/2065304 by guest on 19 January 2020

magnetometer floating freely on the surface of the Southern ocean University, Canberra.
Hitchman, A.P., Lilley, F.E.M. & Milligan, P.R., 2000. Induction arrows from
are of the correct form to be created by an ocean swell of period
offshore floating magnetometers using land reference data, Geophys. J.
13 s and height several metres. The constant of proportionality in the Int., 140, 442–452.
generation process as predicted by Weaver’s theory is approximately Larsen, J.C., 1973. An introduction to electromagnetic induction in the ocean,
1.6 nT of magnetic signal for every 1 m of wave motion. Phys. Earth planet. Int., 7, 389–398.
The sharp peak in the signal spectrum at 13-s period decreases Lilley, F.E.M., White, A. & Heinson, G.S., 2001. Earth’s magnetic field:
very strongly with increasing frequency, consistent with a prediction ocean current contributions to vertical profiles in deep oceans, Geophys.
from physical oceanography. The results of the experiment suggest J. Int., 147, 163–175.
that a floating total-field magnetometer, which has no moving parts, Longuet-Higgins, M.S., Stern, M.E. & Stommel, H., 1954. The electric field
may be a simple independent way of monitoring swell heights in the induced by ocean currents and waves, with applications to the method of
open ocean. towed electrodes, Pap. Phys. Oceanogr. Meteorol., 13, 1–37.
Maclure, K.C., Hafer, R.A. & Weaver, J.T., 1964. Magnetic variations pro-
duced by ocean swell, Nature, 204, 1290–1291.
AC K N OW L E D G M E N T S Milligan, P.R. & Barton, C.E., 1997. Transient and induced variations in
aeromagnetics, Records of Australian Geological Survey Organization
We thank our collaborators in the SWAGGIE project and the captain Vol., 27, AGSO, Canberra.
and crew of ORV Franklin (cruise FR04/98) for their contributions Ochadlick, A.R., 1989. Measurements of the magnetic fluctuations associ-
to the work presented in this paper. We thank John Bye for the benefit ated with ocean swell compared with Weaver’s theory, J. geophys. Res.,
of discussions during his visit to Canberra in 2000. 94, 16 237–16 242.
The magnetometer set up in the floating package was a GEM Sys- Palshin, N.A., 1996. Oceanic electromagnetic studies: a review, Surv. Geo-
tems GSM-19 Overhauser magnetometer. The satellite transmitter phys., 17, 455–491.
used to track the floating magnetometer was part of the Argos sys- Phillips, O.M., 1977. The dynamics of the upper ocean, Cambridge Univ.
Press, Cambridge.
tem. The software of Wessel & Smith (1995) has been used in the
Podney, W., 1975. Electromagnetic fields generated by ocean waves, J. geo-
preparation of this paper. APH carried out his part of the work while phys. Res., 80, 2977–2990.
supported by an ANU Research Scholarship. TP carried out her part Pond, S. & Pickard, G., 1983. Introductory Dynamical Oceanography, 2nd
of the work while a visiting scholar at ANU. PRM publishes with the edn, Pergamon Press, Oxford.
permission of the Executive Officer, Geoscience Australia. Karen Popkov, I., White, A., Heinson, G., Constable, S., Milligan, P. & Lilley,
Weitemeyer is thanked for her help in the preparation of the paper. F.E.M., 2000. Electromagnetic investigation of the Eyre Peninsula con-
We have benefitted from discussion with Graham Warren of the ductivity anomaly, Explor. Geophys., 31, 187–191.
Australian Bureau of Meteorology, who is thanked for supplying Tomczak, M. & Godfrey, J.S., 1994. Regional Oceanography: an Introduc-
Fig. 8. We thank Jean Francheteau, Rob Tyler and an anonymous tion, Pergamon Press, Oxford.
reviewer for comments that have improved this paper. Watermann, J. & Magunia, A., 1997. Propagation parameters of sea sur-
face waves inferred from observations from two closely spaced vector
magnetometers, J. Geomagn. Geoelectr., 49, 709–720.
REFERENCES Weaver, J.T., 1965. Magnetic variations associated with ocean waves and
swell, J. geophys. Res., 70, 1921–1929.
Beal, H.T. & Weaver, J.T., 1970. Calculations of magnetic variations induced Weaver, J.T., 1997. Generation of magnetic signals by waves and swells,
by internal ocean waves, J. geophys. Res., 75, 6846–6852. in, Transient and induced variations in aeromagnetics, Records of Aus-
Bullard, E.C. & Parker, R.L., 1970. Electromagnetic induction in the oceans, tralian Geological Survey Organization vol., 27, Australian Geological
in, The Sea, Vol. 4, pp. 695–730, ed. Maxwell, A.E., Wiley-Interscience, Survey Organization, pp. 15–16, eds Milligan, P.R. & Barton, C.E.,
New York. AGSO, Canberra.
Campbell, W.H., 1997. Introduction to Geomagnetic Fields, Cambridge Wessel, P. & Smith, W.H.F., 1995. New version of the Generic Mapping
Univ. Press, Cambridge. Tools released, EOS, Trans. Am. geophys. Un., 76, 329.

C 2004 RAS, GJI, 159, 565–572