three years continuous record of the earth’s magnetic

15

ANNALS OF GEOPHYSICS, VOL. 52, N. 1, February 2009

Key words Antarctica – DomeC – magnetic obser-vatory – continuos records

1. Introduction

The DomeC site was selected in 1974 tostart research activities under the framework ofthe International Glaciology Project (IAGP). Ashallow coring, joint French/ American pro-gram was conducted at the beginning of the1980’s with the support of the National ScienceFoundation. The camp was abandoned after theend of the coring (to a depth of about 900m),which was hampered by three aircraft crashes.

At the beginning of the 1990’s a French Ital-ian venture planned to build a permanent scien-tific base. The DomeC site was chosen again,

mainly because of the thickness of the ice capand the low atmospheric water vapour content.The first remarkable activity was a long ice coredrilling, which should help deciphering the pastclimate of our planet over a large time span.The coring lasted from 1997 to 2005 with theaim at drilling through the whole ice cap.

Several kinds of scientific activities havenow started beside the glaciology project: at-mosphere analysis, astronomy, geophysics(seismology and geomagnetism).

There are currently a dozen magnetic obser-vatories running on the Antarctic continent(Schott and Rasson, 2007), if we term observa-tories those providing base line control. Thequality of this control is variable, ranging fromepisodic measurements performed during sum-mer campaigns to measurements made regular-ly throughout the year at the staffed bases host-ing qualified observers. Thus, as shown in fig.1, only four observatories – Dumont d’Urville(DRV), Scott Base (SBA), Mawson (MAW)and Argentine Island (AIA) – provide absolutevalues according to modern standards like those

Three years continuous record of the Earth’s magnetic field at Concordia

Station (DomeC, Antarctica)

Aude Chambodut (1), Domenico Di Mauro (2), Jean-Jacques Schott (1), Pascal Bordais (3), Lucia Agnoletto (4) and Pietro di Felice (4)

(1) Ecole et Observatoire des Sciences de la Terre, Strasbourg, France(2) Istituto Nazionale di Geofisica e Vulcanologia, Roma, Italy

(3) Institut Polaire Français Paul Emile Victor, Plouzané, France(4) C.R. ENEA-Casaccia, S.M. di Galeria (RM,) Italy

AbstractThe magnetic observatory deployed at DomeC, Antarctica, in the French-Italian base known as Concordia hasnow been permanently running for more than three years. This paper focuses on these long-term results whichare more relevant for an observatory intended to provide absolute values of the field. The problems whichemerged in this fairly long record are discussed and solutions suggested to upgrade the observatory to the stan-dards of an absolute one (i.e. Intermagnet standards).

Mailing address: Dr. Aude Chambodut, Ecole et Obser-vatoire des Sciences de la Terre 5, rue Descartes 67084, Stra-sbourg Cedex, France; e-mail: [email protected]

Vol52,1,2009 28-05-2009 13:45 Pagina 15

16

A. Chambodut, D. Di Mauro, J.J. Schott, P. Bordais, L. Agnoletto and P. di Felice

stated by Intermagnet. From this point of view,the lack of balance with the Northern hemi-sphere at similar latitudes is striking. In addi-tion, each of DRV, SBA and MAW observato-ries is located on extremely magnetized base-ments, causing severe observatory biases whichare a drawback to be taken into account in glob-al or regional models (e.g. Mandea andLanglais, 2002). They are also submitted tocoast effects which influence the power spec-trum at diurnal frequencies and beyond.

DomeC (hereafter quoted DMC) is operatedjointly by Ecole et Observatoire des Sciencesde la Terre (EOST), Strasbourg, France, and byIstituto Nazionale di Geofisica e Vulcanologia(INGV), Rome, Italy, with the logistic and part-ly financial support of the French Institution In-stitut Polaire Français Paul Emile Victor(IPEV) and the Italian Institution Ente per leNuove Tecnologie, l’Energia e l’Ambiente(ENEA). The aim is to operate DMC in accor-dance to the accuracy requirements stated byIntermagnet and hence, to provide data con-straining helpfully global and regional modelsof the main field and its secular variation (De

Santis et al., 2002; Torta et al., 2002; Gaya-Piqué et al., 2006). Regarding the external field,DMC is situated inside the polar cap, close tothe invariant pole and the geomagnetic Southpole. The auroral zones and polar caps are areaswhere the spatial variation of phenomena likemagnetic storms, substorms, and pulsations issharp. Hence, their study requires the availabil-ity of a dense network of stations. Once again,this is the case in the Northern hemisphere. Inthe Southern hemisphere, VOS, Casey (CSY),DMC, DRV, SBA and Terra Nova Bay (TNB)would build up a fairly dense network in theparticular region mentioned above. The paperpublished by Lepidi et al., 2003, based upondata from TNB and DMC, highlights the inter-est for DMC from external point of view.

Some other specifications of the observato-ry may be found in Schott et al. (2005).

2. Equipment and protocol of measure-ments

The description of the observatory equip-ment was already given in Schott et al. (2005).It is provided again hereafter with some up-dates, for the sake of completeness.

The observatory is equipped with standardinstruments for continuous three-componentfield variation recording and absolute measure-ments. The field variations are recorded with asuspended triaxial fluxgate magnetometer(FGE type) especially manufactured by theDanish Meteorological Institute for low tem-peratures, with suitable damping silicon oil andsilicon connection cables. It is put on a pillarpenetrating about one meter into the ice. Al-though the horizontal component is weak(around 10400nT), the variometer is orientedwith respect to the local magnetic meridian.The technical specifications are: dynamic range±4000nT; resolution 0.2nT; temperature coeffi-cient of sensor lower than 0.2nT/°C; tempera-ture coefficient of electronics lower than0.1nT/°C; band pass DC to 1Hz; sampling rate1Hz. An Overhauser proton magnetometerSM90R records the field intensity at a samplingrate of 0.2Hz. Both instruments are installed ina cave, two meters deep, situated beneath the

Fig. 1. Location of magnetic observatories inAntarctica. Square: Intermagnet observatories; dia-mond: observatories with various status; star:DomeC observatory.

Vol52,1,2009 28-05-2009 13:45 Pagina 16

17

Three years continuous record of the Earth’s magnetic field at Concordia station (DomeC, Antarctica)

shelter containing the electronics and acquisi-tion system. The acquisition system was builtup by INGV. It comprises mainly a 24 bit ADconverter and a PC driving both the AD con-verter and the SM90R. Accurate timing is pro-vided by GPS signal. A second PC connected tothe acquisition displays a control board and cur-rent records. Electronics and acquisition are lo-cated inside a thermally controlled box. Dataare transmitted to the main building via a Wi-Ficonnection. One minute records and absolutemeasurements are sent manually by the observ-er, weekly both to EOST and INGV. Thus, afairly timely remote control of the observatoryis made possible.

The absolute measurements are performedon a pillar made with a polyethylene tube,standing temperatures down to less than –40°C.Its dimensions are: length 3 m, diameter 40 cm,thickness 2.27 cm. It penetrates 1.75 m into theice. It is located inside a shelter which is contin-uously heated at about 10°C. The absolutemeasurements consist of D, I measurementswith a D/ I flux non-magnetic theodolite and Fmeasurements performed manually with a sec-ond Overhauser proton magnetometer SM90R.The theodolite is a Zeiss 010A type with read-ing in grades, 0.5 second of arc resolution. Thesensor and driving electronics are Bartingtondevices with 0.1nT resolution. An azimuthmark was fixed onto the variometer shelter,about 25 m distant. Its true North bearing wasdetermined from sun observations (Newitt etal., 1996). Due to magnetic activity, declinationand inclination are measured using a close-to-zero method (Jankowski and Sucksdorff, 1996)and all readings are reduced to a common time,nearly the mean time of the whole sequencetime span.

To summarize, three component field varia-tions are recorded at 1Hz sampling rate. Inten-sity is recorded at 0.2Hz sampling rate. Bothkinds of data are filtered according to Intermag-net recommendations and resampled at oneminute sampling rate.

The results obtained during three prelimi-nary summer campaigns achieved in December1999-January 2000, December 2001 and De-cember 2003-January 2004 were published inSchott et al. (2005). Therefore, this paper fo-

cuses on the more or less continuous recordsbeginning in February 2005 when the base(named Concordia) opened permanently. Al-though, the observatory is at the moment oper-ated by non dedicated and non qualified ob-servers, we believe that the data quality is goodenough to be of interest for the scientific com-munity.

3. Base lines

3.1. Adopted base lines

Let us recall that one aim with base linemeasurements is to compute absolute values ofthe field components. Thus, a smooth continuousseries , where the subscript bl stands forbase line, has to be estimated from the set of spotvalues by some fitting method.

may be represented either by its Carte-sian components in the geographic referenceframe, or, regarding the horizontal components,by its traditional polar form H(ti), D(ti). In ourcase, according to Section 2, this latter form is theimmediate outcome of the measurement proto-col. Current choices for the fitting are either para-metric (for instance a set of polynomials orthog-onal over the set of sampled times) or non para-metric (for example cubic smoothing splines).

Unfortunately, during the 2006-2007 sum-mer campaign, a very inappropriate decisionwas taken to deploy an electric power line closeto the shelters of the observatory, despite theagreement upon a clean area around the obser-vatory. As a result, no absolute measurementscould be performed during most of the year2007 due to the disturbance from this powerline. Luckily, it was removed during the 2007-2008 summer campaign and proper measure-ments could be resumed since the end of De-cember 2007. The very last measurements wereincorporated into the data set in order to betterconstrain the fitting (as can be noted in fig. 2,the range for adopted base line computationmay be strictly included in the range of avail-able data used for the computation of splinefunctions or orthogonal polynomials. In gener-al, this avoids boundary effects. In addition, inthe present study, base lines are needed only till

tBbl i] g

tB ,bl i i K KN1=] g

tBbl ] g

Vol52,1,2009 28-05-2009 13:45 Pagina 17

18


the end of December 2007). Due to the hugegap in the series, the fitting is not straightfor-ward. This is an important issue because the fi-nal values depend of course on this fitting andhence the time variation of the field. We havetested both a spline and a low degree (up to 3)orthogonal polynomials fitting. Figures 2 and 3show the results for the spline fitting. Figure 2shows the spot values as well as the fitting forH, D, Z and F which is the total field differencebetween the absolute pillar and the variometercave. As expected, the small values of Fbl re-flect the absence of crustal contamination. Thenon zero mean value is due to the variousequipments stored in the variometer house.Figure 3 displays some statistical aspects of the

goodness of fit, with statistical parameters de-rived using the method described by Silverman(1985). One well-known criterion (e.g. Seber,1977) is provided by the time series of theresiduals. Figure 3 shows that they are fairlyrandomly distributed, with a distribution con-sistent with the underlying Gaussian assump-tion. However, the oscillations displayed in fig.2 for the year 2007 are evidently questionablealthough they can be seen also on better sam-pled parts of the curves. They may account, atleast regarding the fluxgate vector magnetome-ter, for an annual variation of the magnetometer(sensors and electronics) and /or acquisitiontemperature, although the temperature variationis not well documented (see below, Section 4).

Fig. 2. 2005-2007 base line fitting using cubic splines for the horizontal H component, declination D, verticalcomponent Z, total field difference F between the proton magnetometer measurements on the absolute pillar andvariometer measurements on the absolute pillar and variometer cave respectively. Crosses: spot values; solidlines: spline adjustment; dashed lines: limits of the 95% confidence interval.

Vol52,1,2009 28-05-2009 13:45 Pagina 18

19


As to the total field difference, Flb, derived fromproton magnetometers, thermal effects cannotnot be precluded either (Sapunov et al., 2001)all the more as so, the sensor is working at un-usually low temperatures. In addition, regard-ing the whole set of base lines, we also have tokeep in mind the annual disturbances broughtabout during the summer campaigns (for in-stance, these disturbances may account for thescattering observed on the very last measure-ments). A more conservative fitting is providedby the low degree orthogonal polynomials ascan be seen in fig. 4. Despite the time distribu-tion of residuals displayed in fig. 5 contains un-derfitting features, we will adopt this base lineadjustment for the computation of the absolute

values. The discussion given in the next Sectionis based upon this model as well.

3.2. Base line drift

Figures 2 and 3 show a conspicuous drift ofthe base lines over the whole time span. Wehave tentatively explained this drift by a pro-gressive rotation of the magnetometer, assum-ing that the sensors are mutually orthogonal andthe base lines constant in the sensor referenceframe (in this frame, they are merely the biasand offset fields). Writing the base linevector at time t in the geographic referenceframe, we have estimated the rotation matrix

B tbl

G ] g

Fig. 3. Left part: studentized (dimensionless) residuals versus time for each component of the base lines shownon fig. 2 var stands for the variance of the statistical model, p for the smoothing factor. Right part: distributionof studentized residuals compared to a Normal Gaussian distribution.

Vol52,1,2009 28-05-2009 13:45 Pagina 19

20


which maps onto .

and are the adjusted base lines at timest and t0, described in the previous Section. t0 issome time origin, taken as the time of the firstvalue of the series in the range 2005-2007. The next approximation we made was toassume that the orientation of the sensors wasknown at time t0. This is of course not true, al-though a common practice, with magnetome-ters aligned in the local geomagnetic referenceframe, is to assimilate the direction of the H-sensor to the base line declination, and to as-sume that the Z-sensor is actually vertical.However, this assumption will only influencethe actual drift of the sensors, not the estimationof the rotation matrix, which anyway requires afurther constraint. Indeed, the problem of com-puting the matrix knowing only a pair ofR t

1] g

B tbl

G ] g

B tbl

G0] g

B tbl

G ] gB tbl

G ] gB tbl

G0] gR t

1] g vectors and has an infinite num-

ber of solutions. Among this set, we have se-lected the rotation having the smallest angle.

The results are shown in figs. 6 and 7. Figu-re 6 shows the paths followed by each of thethree sensor orientations, with a time step ofone month. In every case, somewhat unexpect-edly, the movement is mainly in vertical planes,with an increasing dip for sensors H (negative)and D (positive), whereas the Z-sensor movesaway from the vertical line. Figure 7 shows thatthe rotation vector is essentially horizontal, butthat its axis itself rotates counter-clockwise.The rotation angle varies linearly with time forroughly the years 2005 and 2006, but after-wards exhibits a bending which indicates thatthe drift seems to decrease. The causes of thedrift are either instrumental of environmental.Let us recall that the magnetometer is suspend-

B tbl

G ] gB tbl

G0] g

Fig. 4. Same as fig. 2, apart from the fitting model wich is a sequence of orthogonal polynomials up to degree3. See text for further comments.

Vol52,1,2009 28-05-2009 13:53 Pagina 20

21


ed. According to the technical specifications,the suspension device should compensate tiltsup to 0.5°, that is in a range significantly largerthan the estimated dip of the sensors. However,the proper working of the suspension assumesthat the damping oil is not too viscous (or evenfrozen up). The choice of the damping oil as-sumed temperatures not lower than around -40°Celsius. It turns out that the winter tempera-tures are lower (see Section 4). Thus we cannotpreclude a hardening of the oil. Another sourceof drift might be some unexpected strain of thebrace plates which support the magnetometer.However, any drift due to either of these effectsshould have stopped long ago, because themagnetometer was installed in 1999. On theother hand, if we suspect some movement of

the pillar, the magnetometer has to be rigidlyfixed to its housing, so that this assumption,likewise, implies that the suspension is frozenin somehow. We should also keep in mind theassumption which the estimation of the rotationrelies upon, namely the absence of drift due tothe bias fields, and which is not supported byany direct evidence. However, on one hand, wehave never noticed such a large drift with FGEmagnetometers, though in less tough condi-tions, and, on the other hand, the very small het-erogeneity between absolute and variometershelters, as assessed from Flb range, makes thedrift of D base line inconsistent with a meredrift in H bias field. Anyway, further inquiriesare necessary in order to suggest a plausiblereason for the drift.

Fig. 5. Same as fig. 3 apart from the misfit parameter which is the usual square misfit in linear regression models.

Vol52,1,2009 28-05-2009 13:53 Pagina 21

22


4. Temperature records

Temperatures have been recorded since Jan-uary 2006 with small data loggers, independent-ly from the acquisition device itself. One is putin the cave containing the magnetometer, aboutone meter above the bottom, the second one islocated in the box housing the electronics andacquisition system. Unfortunately, only piece-wise records are available, but nevertheless theyare meaningful enough to illustrate the difficul-ty to fairly control the temperature variations.

The upper curve in fig. 8 shows hourly meantemperature variations in the vicinity of the elec-tronics and acquisition device. The unexpectedhigh values reached in the first part of the year2006 are due to the box being closed by an insu-lating door. Even with the heating supplyswitched off, the heat dissipated by the variouspieces of equipment was efficient enough tomake the temperature rise to surprising levels.By mid-June, the door was left open and, as a re-sult, the temperature dropped to around 10°C.The two large drops noticeable around February10 and September 25 are due to power supplyinterruptions. Most of the remaining jumps pre-sumably reflect the temperature gradient insidethe variometer house, as the temperature datalogger was not exactly put on the same place af-ter each data download. The small oscillationsoccurring mainly in the first and last part of thecurve are daily variations not fully understood.

The lower curve in fig. 8 displays a short se-quence of hourly mean temperature recorded inthe cave. It shows that the temperature variationis unexpectedly large, all the more as the data log-ger fails to record temperatures lower than -44°C.Thus, unfortunately, we have so far no idea of thefull range of variations. We refer to the previoussection for the consequences this large range ofexcursion and decrease to unexpected low tem-peratures might have upon the base lines.

5. Magnetic field daily means

The purpose of this paper is to focus on thelong-term change of the magnetic field. Dailyvariations are a good compromise for showingit along with some features of the external field.

Fig. 6. Drift of the vector magnetometer sensorsover the time span 2005-2007. H is nearly horizontal,close to the magnetic North direction; D is nearlyhorizontal, close to the magnetic East direction; Z isclose to vertical. The time step is one month; the an-gles are expressed in degrees and minutes. Full cir-cles: positive inclinations; open circles: negative in-clinations.

Fig. 7. Left: drift of the rotation vector accountingfor the assumed rotation of the magnetometer sen-sors over the time span 2005-2007. Right: rotationangle versus time. See Section 3.2 for further expla-nations on the meaning of the rotation.

Vol52,1,2009 28-05-2009 13:45 Pagina 22

23


Fig. 9. Daily mean values of the magnetometer outputs. Xv, Yv, Zv, stand for the components measured in thesensor reference frame (close to the local magnetic reference frame). Fv is the total field recorded by the protonmagnetometer. dF is the difference Frecorded minus Fcomputed where Fcomputed is the intensity of the field resultingfrom the combination of the magnetometer outputs and the base lines. The jump of dF in the range March 1 –mid-September 2007 is due to a move of the sensor of the scalar magnetometer.

Fig. 8. Top: Hourly mean temperature record in the box housing the instrument electronics and acquisition de-vice. There are mainly two parts, before and after the door of the box was left open (mid June 2006 onwards).See Section 4 for details regarding the additional jumps and peaks. Bottom: Hourly mean temperature record inthe cave, close to the magnetometer. The temperature drops to values lower than -44°C, which is the lower lim-it of the data logger recording range.

Vol52,1,2009 28-05-2009 13:45 Pagina 23

24


Regarding the daily variations, we refer to thework published by Cafarella et al. (2007) whichdescribes some peculiarities of this variation inconnection with IMF conditions. Figures 9 and10 highlight clearly the seasonal variation dueto large variation of the ionospheric conductiv-ity controlled by the solar radiation.

As to the secular variation, the comparisonbetween figs. 9 (field variation without baseline correction) and 10 (with base lines added)illustrates the difference between an automaticobservatory and an observatory providing ab-solute field components. The long-term correc-tion afforded by the base line control is particu-larly striking in the current situation where astrong base line drift prevails over the naturalsecular variation. Besides, figures 9 and 10 dis-play numerous data gaps in 2007 due to acqui-sition failures. The acquisition device is notprotected against power supply breakdowns

which occurred frequently during these threeyears of permanent working of the base. The in-corporation of an uninterruptible power supplyin April did not improve the working much.The repetition of power supply failures proba-bly progressively damaged the acquisition de-vice. It could not be replaced before the end of2007.

Last, we may notice a jump amounting 8 nTin dF due to an accidental move of the scalarmagnetometer which was put again on its pre-vious place in December.

In order to evaluate the accuracy improve-ment due to the base line correction, we com-pared the observed secular variation with thesecular variation predicted by the IGRF10model. Figure 11 shows that the agreement be-tween observed and predicted values is fairlygood for the vertical component and total field,but not for the horizontal components. Howev-

Fig. 10. Daily mean values of the absolute field. X, Y, Z are the components in the geographical referenceframe. F and dF are equivalent to Fv and dF, respectively, on fig. 9.

Vol52,1,2009 28-05-2009 13:45 Pagina 24

25


er, the discrepancy is constant, and hence, thepredicted and calculated secular variations arein good agreement. Bearing in mind the behav-iour of the base lines and, in addition, the un-certainty on the azimuth mark orientation withrespect to the true North, we might suspect adeclination error. But an elementary calcula-tion shows that a constant angular error cannotexplain the constant shifts, all the more as Xand Y do not vary in the same way. Althoughfurther controls are desirable, especially withregional models like ARM (i.e. Gaya-Piqué etal., 2006) we are confident in the relevance ofthe corrected, absolute field values yielded bythe combination of adopted base lines andmagnetometer outputs.

6. Future improvements and conclusions

Despite the numerous, partly unforeseendifficulties, the progress towards a high stan-dard observatory is encouraging. Three maindifficulties remain to be overcome: a) propertemperature measurements and control; b) pow-er supply stabilization; c) explanation and rem-edy to the base line drift. A solution should bebrought to the two first points thanks to the de-ployment of a new acquisition which will incor-porate temperature recording elements and willbe powered by batteries which should standmain supply cuts and damp its fluctuations. Thethird point is more difficult to tackle. As the ab-solute measurements could luckily be resumed,

Fig. 11. Comparison of the daily mean values of the absolute field with the field predicted by IGRF10 mod-el. Note the fairly good agreement for components Z and F and the discrepancy for X and Y components,amounting around 80 and -270 nT respectively (observed minus predicted values). In any case, there is a goodagreement in secular variation.

Vol52,1,2009 28-05-2009 13:45 Pagina 25

26


we will be able to follow the base line driftmore closely. The analysis of the piecewise pre-vious data should provide an additional clue.On the other hand, technical suggestions are theset up of tiltmeters and the installation of anoth-er magnetometer of a different type, which ispossible without difficulty on a second, at themoment unused pillar.

The data collected within these three firstyears of permanent opening show that it makessense to deploy an observatory on the ice capdespite the potential disturbances which mightbe caused by the ice drift and despite the unusu-al environmental conditions. It is too early toevaluate the impact of these data on the knowl-edge of the internal and external Earth’s field,although some preliminary studies have alreadyshown their importance (Lepidi et al., 2003;Cafarella et al., 2007). This paper focuses onthe long-term field variation, whose validationis of prime importance for the internal fieldknowledge, but we have to mention the collec-tion of the one second data, not yet worked out.Along with the high frequency data recordedwith induction coils in a project conducted inparallel by l’Aquila University, these datashould be very helpful in the realm of solar-ter-restrial physics. Last, due to the exceptional en-vironmental conditions, the observatory de-ployed at DomeC offers very attractive facili-ties for testing magnetometers intended forplanetary exploration.

But in order to strengthen the position of themagnetic observatory project, to improve the rou-tinely working and data quality, and finally tomake the observatory recognized as a high stan-dard one, unfortunate decisions like the one men-tioned in Section 3.1 should be avoided in the fu-ture and a dedicated and educated observershould be devoted to the observatory operation.

Acknowledgements

The authors express their thanks to J.M.Torta and S. Marsal who critically reviewed themanuscript and made helpful suggestions forimproving it. Three of us (P. Bordais, L. Agno-letto, P. Di Felice) operated the observatory

during their one year term at Concordia. The fi-nancial and logistic support of IPEV and ENEAare vital for the carrying out of the project. Wewould like to thank them for this support. Thisis a contribution of C.N.R.S. UMR 7516.

REFERENCES

CAFARELLA, L., D. DI MAURO, S. LEPIDI, A. MELONI, M.PIETROLUNGO, L. SANTARELLI and J.J. SCHOTT (2007):Daily variation at Concordia station (Antarctica) andits dependence on IMF conditions, Annales Geophysi-cae, 25, 2045-2051.

DE SANTIS, A., J.M. TORTA and L.R. GAYA-PIQUE (2002): Thefirst Antarctic geomagnetic reference model (ARM), Geo-phys. Res. Lett., 29 (8), doi: 10.1029/2002GL0144675.

GAYA-PIQUE, L.R., D. RAVAT, A. DE SANTIS and J.M. TORTA

(2006): New model alternatives for improving the repre-sentation of the core magnetic field of Antarctica, Antarc-tic Science, 18 (1), doi: 10.1017/SO954102006000095.

JANKOWSKI, J. and C. SUCKSDORFF (1996): Guide for mag-netic measurements and observatory practice, (Inter-national Association of Geomagnetism and Aerono-my), pp. 235.

LEPIDI, S., L. CAFARELLA, P. FRANCIA, A. MELONI, P. PALAN-GIO and J.J. SCHOTT (2003): Low frequency geomag-netic field variations at Dome C (Antarctica), AnnalesGeophysicae, 21, 923-932.

MANDEA, M. and B. LANGLAIS (2002): Observatorycrustal magnetic biases during MAGSAT and Orstedsatellite missions, Geophys. Res. Lett., 29 (15), doi:10.1029/2001GL013693.

NEWITT, L.R., C.E. BARTON and J. BITTERLY (1996): Guidefor magnetic repeat station surveys, (International As-sociation of Geomagnetism and Aeronomy, WorkingGroup V-8), pp. 111.

SAPUNOV, V., A. DENISOV, O. DENISOVA and D. SAVELIEV

(2001): Proton and Overhauser magnetometers tech-nology, Contributions to Geophysics & Geodesy, Vol.31, pp. 119-124.

SCHOTT, J.J. and J.L. RASSON (2007): Observatories inAntarctica, in Encyclopedia of Geomagnetism and Pa-leomagnetism, edited by D. GUBBINS and E. HERRERO-BERVERA, pp. 1054.

SCHOTT, J.J., D. DI MAURO, A. PERES, L. CAFARELLA, L.MAGNO, A. ZIRIZOTTI and A. MELONI (2005): Towardsthe opening of a magnetic observatory at DomeC(Antarctica), Proceedings XIth Workshop on Geomag-netic Observatories, Instruments, Data Acquisition andProcessing, (Kakioka, November 17-24, 2004).

SEBER, G.A.F. (1977): Linear Regression Analysis, (JohnWiley and Sons (Eds), pp. 465.

SILVERMAN, B.W. (1985): Some aspects of the splinesmoothing approach to non-parametric regressioncurve fitting, J.R. Statist. Soc. B, 47 (1), 1-52.

TORTA, J.M., A. DE SANTIS, M. CHIAPPINI and R.R.B. VON

FRESE (2002): A model of the secular change of the ge-omagnetic field for Antarctica, Tectonophysics, 347,179-187.

Vol52,1,2009 28-05-2009 13:45 Pagina 26

three years continuous record of the earth’s magnetic

Documents