Eos, Vol. 91, No.

51, 21 December 2010

Volume 91
EOS, TranSacTiOnS, amErican GEOphySical UniOn

number 51

21 DeCember 2010 pages 501512

measurements and global geomagnetic models (Figure1a). It followed more than 150 years of slow drift at less than 15kilome ters per year, starting with the first location of the NMP by Ross. This sudden accelera tion contrasts with the behavior of the south magnetic pole, which has a drift speed that has never exceeded 15kilometers per year since the beginning of the twentieth century [Olsen and Mandea, 2007]. Magnetic recordings from polar cap observatories, in Resolute Bay (in Nuna vut, Canada) and Qaanaaq (formerly the city of Thule, Greenland), combined with global models, revealed that the first time derivative of the field of internal origin (i.e., the secular variation) also experienced an unusually large increase over the 1990s, by more than 50nanoteslas per year in the northern component. A detailed analysis of the relationship between geomagnetic secu lar variation and NMP drift speed has shown that the increase in the secular variation was indeed responsible for about 75% of the increase of the drift speed, the remain der being caused by changes in the local gradient of the main field [Chulliat etal., 2010]. That is, the NMP accelerated primar ily because the rate of change of the mag netic field originating in the core suddenly changed at the Earths surface in the north polar region. It is important to note that the NMP posi tion does not seem to have any special physical meaning with respect to the geo dynamo. This is unlike geomagnetic poles, which are defined as the intersections of the Earths surface with the axis of the dipole part of the core field (Earths northern geo magnetic pole is located north of Greenland, and the south geomagnetic pole is antipo dal). Geomagnetic poles are distinct from the magnetic poles because the core field is not exactly dipolar. Also, for the same rea son, the NMP position on the Earths surface cannot be continued radially downward to the core surface and is not even a unique point at the core surface.

What Caused Recent Acceleration of the North Magnetic Pole Drift?

PAGES 501502
The north magnetic pole (NMP) is the point at the Earths surface where the geo magnetic field is directed vertically down ward. It drifts in time as a result of core con vection, which sustains the Earths main magnetic field through the geodynamo pro cess. During the 1990s the NMP drift speed suddenly increased from 15kilometers per year at the start of the decade to 55kilo meters per year by the decades end. This acceleration was all the more surprising given that the NMP drift speed had remained less than 15kilometers per year over the pre vious 150 years of observation. Why did NMP drift accelerate in the 1990s? Answering this question may require revis ing a longheld assumption about processes in the core at the origin of fluctuations in the intensity and direction of the Earths mag netic field on decadal to secular time scales, and hints at the existence of a hidden plume rising within the core under the Arctic. Ross were different, hence suggesting some drift, but still very close to each other (less than 50 kilometers apart). Similar navigational motivations led Natu ral Resources Canada (NRCan) to undertake regular measurement campaigns from 1948 to 1994 to follow the slow drift of the NMP. Two additional NMP locations have recently been determined, in 2001 [Newitt etal., 2002] and 2007 [Newitt etal., 2009]. These were derived from data gathered during two campaigns that resulted from the successful international collaboration between Asso ciation PolyArctique (a French fundraising group), NRCan, the Institut de Physique du Globe de Paris (IPGP), and Frances Bureau de Recherches Gologiques et Minires (BRGM). As the NMP drifts away from the nearest airfield in Eureka (a small research base in Nunavut, Canada) toward Siberia, measuring the field in the vicinity of the NMP has become increasingly difficult due to the limited flying range of aircraft able to land on sea ice. Fortunately, the recent drift over the Arc tic Ocean has happened at a time when low Earth orbiting satellites carrying high precision magnetometers, such as rsted and the Challenging Minisatellite Payload (CHAMP), have made it possible to calcu late global geomagnetic field models with unprecedented spatial and temporal resolu tion [e.g., Olsen etal., 2009]. As the highly successful rsted and CHAMP missions are reaching an end (CHAMP ceased on 19Sep tember), geomagnetic measurements from space are expected to be continued thanks to the European Space Agencys Swarm mission, to be launched in 2012 [FriisChristensen etal., 2006].

From Field Surveys to Satellite Monitoring

On 1 June 1831, James Clark Ross made the first determination of the location of the north magnetic pole. He found it on Boothia Peninsula, in the Canadian Arctic: Amidst mutual congratulations, we fixed the British flag on the spot, and took possession of the north magnetic pole and its adjoining terri tory, in the name of Great Britain and King William the Fourth [Ross, 1835]. This suc cess was hailed as the greatest achievement of the 5year polar expedition led by John Ross (Jamess uncle) in search of the North west Passage. Indeed, knowing the poles position was important for navigation, which at that time relied heavily on the use of com passes and accurate declination charts. Seventy three years passed before Roald Amundsen made a second determination of the location of the NMP, in 1904, during the first successful navigation of the Northwest Passage. The locations by Amundsen and By A. ChulliAt, G. hulot, l. R. Newitt, ANd J.-J. oRGevAl

Sudden Acceleration in the 1990s

Observations from satellites, magnetic observatories, and field surveys show that the NMP drift speed suddenly increased in the 1990s [Newitt and Barton, 1996; Newitt etal., 2002], from 15kilometers per year in 1990 to about 60kilometers per year in 2002, after which it slowly decreased [Olsen and Mandea, 2007; Newitt etal., 2009]. This phe nomenon was observed in both field survey

Core Plume Hypothesis

Why should scientists and society pay attention to the acceleration of NMP drift? The answer lies in what this acceleration may reveal about the Earths core, a region that can be studied only through indirect means. Studies show that the large change in secular variation observed in the north

Eos, Vol. 91, No. 51, 21 December 2010

polar region in the 1990s is mainly caused by a similar change of the secular variation at the core surface in a relatively small area (about 1000 kilometers in diameter) located under the New Siberian Islands (Figure1b). This can be seen by how the observed main field and secular variation continues down ward from the Earths surface to the core mantle boundary (CMB), assuming that the effect of the mantle electrical conductivity is negligible. In fact, analysis of the mathemati cal function that relates the magnetic field at the CMB to that at the Earths surface [Chulliat etal., 2010] shows that the NMPwhich is to some degree independent of dynamics vertically under it at the CMBwas at the right place (2000 kilometers from New Sibe rian Islands) at the right time to bear the full force of the large secular variation change under the New Siberian Islands in the 1990s. Interpretation of the observed secular variation at the CMB is usually made by solv ing an inverse problem yielding core surface flows, assuming that magnetic diffusion is negligible on decadal to secular time scales, an assumption referred to as the frozenflux hypothesis [Roberts and Scott, 1965]. How ever, inferring core flows that would gener ate the NMP acceleration proved to be prob lematic due to the geometry of the secular variation under the New Siberian Islands, which does not seem to be compatible with the frozenflux hypothesis. Global core field models based upon satellite data indicate that some magnetic flux was expelled from the core in this area during the 1990s (Fig ures1c and1d) through a necessarily diffu sive process as shown by Backus [1968]. Further insight into core processes at the origin of the NMP acceleration is pro vided by recent three dimensional numeri cal simulations of the geodynamo. Accord ing to some recent models [Aubert etal., 2008], plumes of less dense fluid form at the inner core boundary and subsequently rise within the cylinder tangent to the inner core whose central axis is the Earths rota tion axis. Such plumes undergo a strong helical motion due to the Earths rapid rota tion, a phenomenon also observed in lab oratory experiments with water [Aurnou etal., 2003]. In the core, helical plumes advect and twist the magnetic field lines, forming what scientists call polar mag netic upwellings. Upon reaching the CMB, a polar magnetic upwelling leads to the expulsion of magnetic field lines into the mantle and the formation of a pair of mag netic flux concentrations, detectable as patches of an intense radial field at the core surface, each having a flux opposite to the other. Numerical simulation scaling laws suggest that such magnetic flux expulsion could happen in only a few decades. The similarity between the CMB radial field pattern from simulations of the dynamo and that observed under the New Siberian Islands is striking. Whether the observed magnetic flux expulsion at the origin of the NMP acceleration could result from such a polar magnetic upwelling

Fig. 1. (a)Average north magnetic pole (NMP) drift speeds from the observed NMP positions (in blue), represented by stairs delimited by the observation dates, with error bars.These are shown along with NMP drift speeds from the gufm1model [Jackson etal., 2000], the Comprehensive Model CM4 [Sabaka etal., 2004], and the CHAOS-2 model [Olsen etal., 2009]. (bd)Polar views (from colatitude 0 to 35, parallel lines every 10) at the core surface of the total radial secular variation change (in nanoteslas per year) between 1989 and 2002 (Figure1b) and the radial magnetic field (in microteslas) in 1989 (Figure1c) and 2002 (Figure1d); all are calculated from the CM4 model.The NMP (at the Earths surface) is shown as a red dot.The location of the maximum secular variation change is shown as a red triangle.The intersection of the cylinder tangent to the inner core and the core surface is shown as a black solid circle.

remains a hypothesis [Chulliat etal., 2010], the validity of which will have to be assessed by more detailed modeling and numerical simulations. Interestingly, the existence of another magnetic plume was hypothesized under the large patch of field lines directed opposite to its surroundings located under the NMP [Olson and Aurnou, 1999], where some magnetic flux also seems to be expelled from the core, but this plume would not contribute to the NMP drift (according to the mathematical func tion relating the CMB field to the surface field); for geometrical reasons, only the plume under the New Siberian Islands has an effect on the NMP drift.

Magnetic Flux Expulsions at the Core Surface

The interpretation of the NMP accelera tion in terms of magnetic flux expulsion

would probably not have been easily accepted only a decade ago. At that time it was widely believed that the frozenflux hypothesis was valid on decadal to secu lar time scales and that the contribution of magnetic diffusion to magnetic field varia tions was only significant on longer time scales. Yet both numerical simulations [Amit and Christensen, 2008] and recent observational studies relying on satellite data [Chulliat and Olsen, 2010] have shown that this hypothesis might be significantly violated in some areas of the core surface, casting doubts about core flows inferred from main field and secular variation obser vations in these regions. One such area is located under the southern Atlantic Ocean, where at least two patches of magnetic flux directed opposite to their surroundings are at the origin of a large area of anomalously low field intensity referred to as the South

Eos, Vol. 91, No. 51, 21 December 2010

Atlantic anomaly. The regular deepening of this anomaly, i.e., the decrease in mini mum field intensity and increase in the anomaly size, seems to be related to a reg ular increase of the magnetic flux being expelled through these two patches at the core surface [Bloxham etal., 1989]. Precise knowledge of the long term behavior of this anomaly is lacking because prior to 1980 models based on high quality satellite data did not exist. Magnetic flux expulsions at the core sur face, whether under the South Atlantic anomaly or within the cylinder tangent to the inner core, can be seen as Earths ana logs of sunspots. Future magnetic satel lite missions as well as more sophisticated numerical simulations should lead to a bet ter characterization of this phenomenon and a better understanding of its relation ship with deeper geodynamo processes.
cal dynamos, Geophys. J.Int., 172, 945956, doi:10.1111/j.1365 246X.2007.03693.x. Aurnou, J., S. Andreadis, L. Zhu, and P.Olson (2003), Experiments on convection in Earths core tangent cylinder, Earth Planet. Sci. Lett., 212(12), 119134, doi:10.1016/S0012821X(03)002371. Backus, G. E. (1968), Kinematics of geomagnetic secular variation in a perfectly conducting core, Philos. Trans. R. Soc. London A, 263(1141), 239 266, doi:10.1098/rsta.1968.0014. Bloxham, J. B., D. Gubbins, and A.Jackson (1989), Geomagnetic secular variation, Philos. Trans. R.Soc. London A, 329(1606), 416502, doi:10.1098/rsta.1989.0087. Chulliat, A., and N. Olsen (2010), Observation of magnetic diffusion in the Earths outer core from Magsat, rsted, and CHAMP data, J.Geophys. Res., 115, B05105, doi:10.1029/2009JB006994. Chulliat, A., G. Hulot, and L. R. Newitt (2010), Magnetic flux expulsion from the core as a pos sible cause of the unusually large acceleration of the north magnetic pole during the 1990s, J.Geophys. Res., 115, B07101, doi:10.1029/ 2009JB007143. Friis Christensen, E., H. Lhr, and G.Hulot (2006), Swarm: A constellation to study the Earths mag netic field, Earth Planets Space, 58, 351358. Jackson, A., A. R. T. Jonkers, and M.R. Walker (2000), Four centuries of geomagnetic secu lar variation from historical records, Philos. Trans. R.Soc. LondonA, 358(1768), 957990, doi:10.1098/rsta.2000.0569. Newitt, L. R., and C. E. Barton (1996), The position of the north magnetic dip pole in 1994, J.Geomagn. Geoelectr., 48(2), 221232. Newitt, L. R., M. Mandea, L. A. McKee, and J.J. Orgeval (2002), Recent acceleration of the north magnetic pole linked to magnetic jerks, Eos Trans. AGU, 83(35), 381, doi:10.1029/2002EO000276. Newitt, L. R., A. Chulliat, and J.J.Orgeval (2009), Location of the north magnetic pole in April 2007, Earth Planets Space, 61(6), 703710. Olsen, N., and M.Mandea (2007), Will the mag netic north pole move to Siberia?, Eos Trans. AGU, 88(29), 293, doi:10.1029/2007EO290001. Olsen, N., M. Mandea, T. J. Sabaka, and L.Tffner Clausen (2009), CHAOS2A geomagnetic field model derived from one decade of continuous satellite data, Geophys. J.Int., 179(3), 14771487, doi:10.1111/j.1365 246X .2009.04386.x. Olson, P., and J. Aurnou (1999), A polar vor tex in the Earths core, Nature, 402, 170173, doi:10.1038/46017. Roberts, P. H., and S. Scott (1965), On analysis of the secular variation: I.A hydromagnetic constraintTheory, J.Geomagn. Geoelectr., 17(2), 137151. Ross, J. (1835), Narrative of a Second Voyage in Search of a North-West Passage, and of a Residence in the Arctic Regions, During the Years 1829, 1830, 1831, 1832, 1833: Including the Reports of Commander, Now Captain, James Clark Ross, R.N., F.R.S., F.L.S., etc., and the Discovery of the Northern Magnetic Pole, A.W. Webster, London. Sabaka, T. J., N. Olsen, and M.Purucker (2004), Extending comprehensive models of the Earths magnetic field with rsted and CHAMP data, Geophys. J.Int., 159, 521547, doi:10.1111/j.1365 246X .2004.02421.x.

Acknowledgments
A portion of the funds collected by Asso ciation PolyArctique goes to support medi cal research on rheumatoid polyarthritis.

References
Amit, H., and U. R. Christensen (2008), Accounting for magnetic diffusion in core flow inversions from geomagnetic secular variation, Geophys. J.Int., 175, 913924, doi:10.1111/j.1365 246X.2008.03948.x. Aubert, J., J. Aurnou, and J.Wicht (2008), The magnetic structure of convection driven numeri

Author Information
Arnaud Chulliat and Gauthier Hulot, Institut de Physique du Globe de Paris, Paris, France; Email: chulliat@ipgp.fr; Lawrence R.Newitt, Boreal Lan guage and Science Services, Ottawa, Ontario, Cana da; and JeanJacques Orgeval, Projet PolyArctique, Olivet, France

nEWS
Eos Interviews John Holdren, President Obamas Science Advisor
PAGE 503
With the Obama administration about to face a Republicanled House of Repre sentatives in January, presidential science advisor John Holdren sat down with Eos for an exclusive and wideranging inter view following a policy speech he deliv ered on 13December at the AGU Fall Meet ing in San Francisco. During the interview, Holdren, who also is director of the White House Office of Science and Technology Policy (OSTP), focused on the challenging congressional and budgetary environment, the administrations priorities related to the Earth sciences, and the responsibility of sci entists in helping to communicate the soci etal benefits of science, educate the public, and improve science education and literacy. Holdren said the Obama administrations top priorities related to the Earth sciences include improving observations of the Earth, making progress in dealing with climate change, and rebalancing NASAs focus. Continuing to invest in and build up the nations capacities for Earth observations of all kinds, including from the air, from space, and on and under the oceans, is a top priority, he said. We dont yet have the observation networks and capacities that we ought to have to keep track of whats happening on and to the Earth, he said. That priority on maintaining and expand ing the data sets, the observations, the monitoring, is absolutely key. If you dont do that, you can never make up for it, in the sense that we will never know what the Earth was doing in places and times when we werent monitoring it, aside from through such studies as paleoclimatology, he said. We will be ashamed of ourselves if we allow the kinds of data sets from space that we have had from LandSat and we are having from the current generation of polarorbiting satellites, if we allow those sequences, those series, to lapse. Well never forgive ourselves, and our successors

John Holdren in Earth science will never forgive us, he said. Regarding climate change, Holdren said the failure in getting comprehensive cli mate and energy legislation passed during the administrations first 2 years in office is discouraging but that the administration is not giving up on making significant prog ress in the area. I dont think by any means we have missed the moment. What needs to be rec ognized is this is not a challenge that is going away. Its only going to become more obvious, he said, indicating that 2010 likely will be the hottest or secondhottest year in the instrumental record. Eventually, every body is going to catch on.