We examined the coast effect in Antarctica along the 80

Magnetospheric ultra-low-frequency waves (ULF waves, 1.7–5

In this context, it is important to exclude contaminations of the signals
such as, for example, the effects of the ground conductivity. Waves
transmitted from the magnetosphere to the ground can be partially reflected
by the ground itself, so that a ground magnetometer measures both the primary
source (the magnetospheric waves) and the variations induced by underground
electric currents, which represent the secondary source

Early investigations in the time domain conducted by

More recently, experimental observations by

Moreover, similar results are found at TNB by

The availability of simultaneous measurements from observatories at TNB and
Scott Base (SBA; data provided by INTERMAGNET database), allows us to make an
interesting comparison, in that the stations are located approximately at the
same geomagnetic latitude (

Our statistical investigation shows that, in the frequency range
1–7

Geomagnetic field variations were measured at the three sites by means of
three-axis fluxgate magnetometers along the northward (

The geomagnetic anomalies are investigated in the frequency domain. The
ground conductivity anomaly effect on vertical component can be estimated by
means of the following empirical relationship

The contribution of anomalies to the vertical and horizontal geomagnetic
field can be estimated by using the so-called interstation transfer functions

the normal field variation

the anomalous part

From

Fujiwara and Toh (1996) and Beamish (1982) further simplified the linear
system, assuming a negligible

The transfer functions

In the present study we estimated the transfer functions by means of the
least-square method over

It is worth noting that the errors in transfer functions are negligible
because they are reduced by a factor

A second set of equations (from Eq. 2) relates the fluctuations of the
horizontal field at reference and anomalous sites

At a given station, through the complex coefficients

The interstation analysis allows us to estimate the normal field
fluctuations, in the frequency domain. Assuming that interstation transfer
functions of horizontal components are known, by inverting the Eq. (4), and
taking into account that

Although useful, the above technique requires caution in computing the
transfer functions. Indeed, this method fails if the plane wave assumption is
not valid (

a high correlation of the horizontal components at different stations (i.e., between TLD and coastal stations of TNB and SBA);

a low coefficient

Moreover, the transfer function formulas have a common dependence on

Geographic coordinates, IGRF08 corrected geomagnetic coordinates and time in UT of the geomagnetic local noon for the three stations.

The position of the geomagnetic stations. Dashed lines mark geomagnetic coordinates.

We used TLD as the reference station, since our analysis demonstrated that it
is poorly affected by coast effects; the station is located in the inner
Antarctic region (

In Sect.

Average spectral density of the

Average multiple coherence between vertical and horizontal components for inland (red line) and coastal (black and blue lines) ground stations.

We found that the spectral power on the

Figure

The clear difference between spectral behaviors at coastal stations with
respect to the inland station suggests that at TNB and SBA the ULF
fluctuations, at frequencies in the range 0.3–

Evidence of a possible relationship between the vertical and horizontal field
variations at each station can be revealed by estimating the multiple
coherence

These results suggest that the geomagnetic field measured at TLD station is the normal field, unaffected by ground anomalies, so TLD can be adopted as a reference station.

It is worth noting that the differences between the coastal stations and the
inland station tend to decrease with decreasing frequency, probably due to
both the effects of a horizontally homogeneous deep lithosphere under all
geomagnetic stations (i.e., also under the sea), which responds to the low
frequencies, and the effects of uniform-inducing ULF waves at all ground
stations, since at low frequencies the wavelengths (several Earth radii) are
larger than the maximum distance between stations (

First and second row:

To apply the interstation technique, we need to satisfy two requirements, as explained in Sect. 2: the source should be a common plane wave simultaneously observed at coastal and reference stations, and the coherence between horizontal components at the reference station should be sufficiently low.

Figure

The coherence between horizontal homologous components at reference and coastal stations as a function of frequency. The horizontal bar indicates the 99 % confidence level.

In order to take into account only clear events generated during disturbed
geomagnetic conditions, we restricted our analysis to events corresponding to
AE higher than 50

For the selected events we examined the coherence between homologous
components at reference and each coastal station. Figure

Remote reference transfer functions

Single-station induction arrows at SBA, TNB and TLD

Horizontal transfer functions from interstation method at TNB

Real and imaginary parts of the estimated remote reference transfer functions

The estimated RRIAs at TNB and SBA are shown, using the stereographic
projection, in panel (b) of Fig.

Regarding the horizontal transfer functions at TNB and SBA, we show the real
and imaginary parts of each transfer function (Eq.

In order to study the polarization characteristics due to coast effects, we
estimated

Figure

This example clearly shows that the polarization characteristics (more
evident in the azimuthal angle in this case study) may be affected by ground
conductivity anomalies, probably due to the closeness of TNB to the
coastline, as suggested by

In the present work we studied the coast effect at the Antarctic geomagnetic
stations TNB and SBA during 18 January–14 March 2008, also using the inland
temporary station TLD, more than 200

In particular, the Pc5 power of the vertical component at the coastal
stations is higher than at the reference station, and the spectral ratio
between vertical and horizontal components is similarly high. Moreover, the multiple coherence
of the horizontal components with the vertical component reveals higher
values at SBA and TNB with respect to TLD. These results suggest that the
sea–land interface affects measurements at coastal stations, and confirms
TLD as a suitable reference station, since the amplitude of the induction
arrows is generally 20 times smaller than at coastal stations, in agreement
with

We proposed a method for estimating directly, in the frequency domain, the
normal field variations at coastal stations, by inverting the linear
relationship between horizontal field measurements at coastal and reference
stations. As an example, we showed the Pc5 event on 11 March 2008, for which
we observed different azimuthal angles at TNB and TLD. When corrected by
means of our method, the azimuthal angle at TNB changes, becoming similar to
the angle at TLD, while the polarization ratio and ellipticity do not change
significantly. These results indicate that the azimuthal angle of polarized
ULF waves at the coastal station of TNB is probably affected by horizontal ground
conductivity anomalies, attributable to the sea saltwater, as suggested by

Most of the geomagnetic stations installed in Antarctica are situated close to the sea, and so geomagnetic measurements could be affected by local coast effects. This work can be regarded as a useful reference in order to remove local effects: for example, without considering coast effects, a comparison between polarization parameters at coastal and inland stations could be misleading, leading to incorrect conclusions. In addition, the spectral techniques shown here could be used not only to study anomalous variations at coastal stations, where the anomaly is persistent, but also to detect possible anomalous effect due to sporadic changes in ground conductivity.

TNB data can be downloaded from the INGV web site:

MR performed the data analysis. MR, PF, MDL and SL drafted the manuscript. MR, PF, MDL and SL participated in the study design and interpretation of results. SL provided geomagnetic field measurements at TLD. AP and SU deployed the magnetometer at TLD in the 2008 temporary campaign. All authors read and approved the final paper.

The authors declare that they have no conflict of interest.

The research activity at Mario Zucchelli station and Talos Dome has been
supported by the Italian PNRA (Programma Nazionale Ricerche in Antartide).
The results presented in this paper rely on the data collected at Scott Base;
we thank the Institute of Geological & Nuclear Sciences Limited (New
Zealand) for supporting its operation and INTERMAGNET for promoting high
standards of magnetic observatory practice
(