Pt Love0208

For centuries, navigators of the worlds oceanshave been familiar with an effect of Earths magnetic field: Itimparts a directional preference to the needle of a compass. Al-though in some settings magnetic orientation remains impor-tant, the modern science of geomagnetism has emerged fromits romantic nautical origins and developed into a subject ofgreat depth and diversity. The geomagnetic field is used to ex-plore the dynamics of Earths interior and its surroundingspace environment, and geomagnetic data are used for geo-physical mapping, mineral exploration, risk mitigation, andother practical applications. A global distribution of ground-based magnetic observatories supports those pursuits by pro-viding accurate records of the magnetic-field direction and in-tensity at fixed locations and over long periods of time.

Magnetic observatories were first established in the early19th century in response to the influence of Alexander vonHumboldt and Carl Friedrich Gauss. Since then, magneticmeasurement has advanced significantly, progressing fromsimple visual readings of magnetic survey instruments to in-clude automatic photographic measurement and modernelectronic acquisition. To satisfy the needs of the scientificcommunity, observatories are being upgraded to collect datathat meet ever more stringent standards, to achieve higheracquisition frequencies, and to disseminate data in real time.

To appreciate why data from magnetic observatories canbe used for so many purposes, one needs only to recall thatthe geomagnetic field is a continuum, connecting the differ-ent parts of Earth to each other and to nearby space. Beneathour feet and above our heads, electric currents generate mag-netic fields that contribute to the totality of the geomagneticfield measured at an observatory on Earths surface. Themany physical processes that operate in each geophysical do-main give rise to a complicated field that exhibits a wide va-riety of time-dependent behavior.1 In this article I review thestatus of the global community of magnetic observatories,show how Earth and space can be monitored for purposes ofscientific understanding and practical application, and high-light the role played by magnetic observatories in the historyof geomagnetism research.

Measurement and dataTo support a wide range of geophysical studies, magnetic ob-servatories such as that shown in figure 1 need to produceaccurate measurements of the geomagnetic field over a widerange of time scales. The longest time scale is defined by thelifetime of the observatory. Naturally, that depends on manypractical factors, including long-term funding and staffing.

Some observatories operate for only a few years, but others,such as the Sodankyl Geophysical Observatory in Finlandand the Apia Observatory in Samoa, have operated continu-ously for well over a century. The shortest time scale of rele-vance is the time between sequential measurements. Olderanalog photographic systems typically produce data with aone-hour cadence. Modern digital systems provide data at much higher acquisition rates. These days, one-minute-average data is a standard observatory product, but one-second-average data production is becoming more common.As an example, the Kakioka Magnetic Observatory in Japanhas produced one-second data continuously since 1983arecord of magnetic-field variation over time scales spanningalmost nine orders of magnitude.

To reliably produce a long-period geomagnetic time se-ries, an observatory must operate under carefully controlledconditions.2 Typically, the site of an observatory is largeenough to isolate the measurements from most sources of an-thropogenic magnetic interference, and many observatoriesare in relatively remote locations. Buildings on the site pro-vide stable operating conditions for the sensors, calibrationsystems, and associated instrument and data-acquisitionelectronics.

A modern observatory has a fluxgate magnetometer,which gives vectorial data conventionally expressed in termsof either the Cartesian components (X [north], Y [east], andZ [down]) or the horizontalpolar components (horizontalintensity H = [X2 + Y2]1/2, declination D = arctan[Y/X], andZ [down]). Note that declination is the direction in which acompass needle points. More formally, it is the angle of thedirection of the magnetic fields horizontal component. Awell-run observatory will produce fluxgate data that showlittle drift in accuracyusually less than 20 nanotesla annu-ally. For many real-time nonresearch applications, that stan-dard of accuracy is sufficient.

But more stringent ionospheric and magnetospheric re-search projects, as well as long-term mapping of the globalmagnetic field, require more accurate data. For that reason a modern magnetic observatory has a proton precessionmagnetometer that measures the total absolute field intensityF = (X2 + Y2 + Z2)1/2. An observatory also has a pier-mountedtheodolite, a familiar surveying instrument, but one having asmall fluxgate fixed to its telescope. About once a week, an ob-server visits the site and makes a series of measurements usingthe theodolite to obtain D and the inclination I = arctan(Z/H).Those absolute magnetic-direction data are then used to cali-brate the fluxgate data, so as to compensate for long-term drift

2008 American Institute of Physics, S-0031-9228-0802-010-3 February 2008 Physics Today 31

Magnetic monitoringof Earth and spaceJeffrey J. Love

With data provided by magnetic observatories, geophysicists can gaininsights into our planets interior and nearby space environment withouteven leaving the ground.

Jeffrey Love ([email protected]) is the US Geological Survey adviser for geomagnetic research and a member of the Intermagnet executive council.

in the fluxgate magnetometer. Production of definitive obser-vatory data involve processing. The resulting data have an ab-solute accuracy of better than 5 nT, which permits meaningfulanalysis of magnetic variation that can occur over time scalesranging from the acquisition cadence out to the observatoryslifetime.

Fluxgate magnetometers and proton precession magne-tometers are the two types most commonly used in magneticobservatories. The Quick Study by Uli Auster on page 76 de-scribes how they work and includes a photo of a theodolitewith fluxgate.

A global networkApproximately 170 magnetic observatories oper-ate worldwide. Most are supported by nationalgovernments, some by universities, and a few byprivate companies. During the International Geo-physical Year in 195758, many new observa-tories were established as part of a coordinatedeffort to enhance the global collection of geo-physical data, and many existing observatorieswere improved (see the article by Fae L. Korsmo,PHYSICS TODAY, July 2007, page 38). Today about120 observatories produce and routinely reportdigital data with an acquisition cadence ofone minute or better; figure 2 shows where theyare located around the globe. The remaining 50or so observatories use older, analog systems orreport their data only years after acquisition.Note that the geographic distribution of observa-tories is far from uniform, with a general sparsityin, for example, the Southern Hemisphere and inthe central Pacific. To promote observatory oper-ation according to consistent standards and to fa-cilitate the prompt dissemination of digital data,the international observatory network organiza-tion Intermagnet (http://www.intermagnet.org)was formed in 1987. As of January 2008, 42 coun-tries and 108 observatories participate in and fol-low the modern standards set by Intermagnet.

Ground-based fluxgate networks, some-times called variometer networks, and satellite-based magnetometers fill niches that are comple-mentary to that filled by the observatories. Mostfluxgate networks are maintained by universitiesand various national governmental programs;they typically operate for a few years for space-physics research. Because of their more special-

ized nature, the networks do not need the laborious stan-dards adopted by full-fledged magnetic observatories.Satellite magnetometers measure the part of Earths magneticfield that is in space; over the course of many orbits, they canprovide good global coverage, albeit from only a relativelysmall number of locations at any particular time.

In the future almost all users of observatory data will expect greater accuracy, and many will require real-time magnetic-observatory data streams and easier accessto data from all parts of the globe. Demand for higher-frequency data acquisition will increase, especially from

space physicists. To meet those needs, the inter-national programs that support magnetic obser-vatories will have to be even better integratedthan they are today. Older observatories willneed to be modernized and all parts of observa-tory operations made more automatic.

Secular variation and Earths coreHow do geophysicists interpret the abundance ofmagnetic-observatory data? To answer that ques-tion, let us take a tour of magnetic signals. The tourbegins deep inside Earth, in the iron core where themajority of the geomagnetic field originates. Fromthere, we trace the magnetic field up to the surface,where the observatories are located. Next, we con-tinue onward and upward through the ionosphereto the magnetosphere. The physics encompassed inthe tour is classical and includes electricity and mag-netism, fluid mechanics, and plasma dynamics.

Figure 2. Worldwide distribution, as of 2006, of magneticobservatories that report one-minute-average digital data.

Figure 1. The observation tower building of the magnetic obser-vatory in Alibag, India, south of Mumbai. Here, careful measure-ments of the magnetic field are made in order to calibrate datacollected by digital systems. The online version of this article in-cludes links to photographs of other observatories.

32 February 2008 Physics Today www.physicstoday.org

Earths core lies some 2900 km below the surface. In theouter part of the core, a combination of thermal and chemi-cal buoyancy sustains convective fluid motion and estab-lishes what is essentially a naturally occurring electrical gen-erator. As the electrically conducting core fluid flows throughthe geomagnetic field, motional induction generates electriccurrents. Those currents, in turn, generate their own mag-netic fields. If it is sufficiently complicatedlacking simplesymmetrythe magnetic field that partakes in the motionalinduction is the same field that is sustained by the inducedelectric currents. The process is efficient enough to overcomethe effects of ohmic dissipation, and so Earths core is a self-sustaining dynamo.3 The mathematics of the geodynamo issometimes described as being a bit like that of oceanographyand meteorology, but with the additional complication pre-sented by the magnetic field itself. Scientists still dont knowor understand many things about dynamo theory; not sur-prisingly, it is the subject of ongoing research (see the articleby Raymond Jeanloz and Barbara Romanowicz, PHYSICSTODAY, August 1997, page 22).

Part of the magnetic field generated in the core extendsoutward, passes through the weakly electrically conductingmantle, and reaches the surface. The field at Earths surfacetypically has an intensity of 30 00060 000 nT and is ap-proximately dipolar, with an axis tilted by about 10 with re-spect to Earths rotational axis. But the magnetic field alsohas important ingredients that are nondipolar. One way toappreciate that is to make a map of declination. The compassneedle aligns itself with the horizontal direction of the localmagnetic field. As figure 3a shows, a compass needle almostnever points due north. Indeed, because of the nondipolarfield, declination is a complicated function of latitude andlongitude. As a result of core convection, the magnetic fieldalso exhibits secular variation over time scales of decades to

millions of years. And so, at a given location, the directionthat a compass points changes over time. The two maps ofdeclination in figure 3a show the progression of geomag-netic secular variation over the past century.4 Indeed, be-cause the field changes in time, maps of it are updated everyfive years or so.

Figure 3b shows year-to-year differences in declinationmeasured during the 20th century from five different obser-vatories. The secular variation not only is different in differ-ent locations but also occasionally accelerates. An interestingfeature of the data is the apparent presence around 1970 of adiscontinuous change, or jerk, in the rate of secular variation.Jerks are clearly seen in the European and Australian data.On the other hand, a jerk isnt obvious in the Japanese data.And although the Alaskan data show a jerk, it is of the op-posite sign of that for Europe and Australia.

Clearly, a global description of the secular variation iscomplicated. Still, geophysicists have made progress in re-lating jerks and secular variation to decade-scale changes inEarths rotational rate that arise from exchanges of angularmomentum between the core and mantle. With certain as-sumptions, core angular momentum can be deduced fromgeomagnetic secular variation models. Then, assuming thatEarths total angular momentum is conserved, one can esti-mate the changes that should have occurred in the mantle an-gular momentum over the past century or so. Predicted vari-ations in the length of a day are close to those actuallyobserved, and that gives researchers some confidence thattheir theories are reasonable.

Crust, ocean, and mantleTime-dependent magnetic-field variations sustained by cur-rents in the ionosphere and magnetosphere induce electriccurrents in the crust, ocean, and mantle.5 Those currents, in

www.physicstoday.org February 2008 Physics Today 33

1900

2000

YEAR

1900 20001980196019401920

CH

AN

GE

IND

EC

LIN

AT

ION

(arc

min

ute

s/y

ear)

30

20

10

0

0

0

0 0

0

0

10

5+10

+10

20

30

10

10

10

10

ESKScotland

15 /yr+

VLJ/CLFFrance+ 10 /yr

WAT/GNAAustralia+ 5 /yr

SITAlaska

KAKJapan 25 /yr

a b

Figure 3. Magnetic declination changes over time. (a) Contour maps of declination D for the years 1900 and 2000 show significant differences over the century. Each contour line represents 5; red is declination to the east and blue is to the west. (b) Data from five observatories show the yearly rate of change in declination. Note, in particular, the abrupt changes, or jerks,in the rate of secular variation, around 1970. For clarity of presentation, the data have been separated by the ordinate valueslisted on the right.

turn, generate internal magnetic fields. The skin depth towhich the induced currents diffusively penetrate is a func-tion of Earths subsurface electrical conductivity and the fre-quency of the overhead magnetic-field variations. So, for ex-ample, magnetic variations with periods ranging from asecond to tens of minutes penetrate into the crust some20100 km. An observatory, of course, measures the totalmagnetic field, a superposition of the external, inducing fieldand the internal, induced field. Mathematical separation ofthe two requires a large number of simultaneous measure-ments, densely distributed across Earths surface; it is a tasknot ideally suited to the relatively sparse distribution of mag-netic observatories. For that reason, detailed regional studiesof Earths conductivity structure often involve the deploy-ment of temporary arrays of sensors to measure both themagnetic field at Earths surface and the induced electric fieldin the crust.

Qualitative insight into geomagnetic induction can beobtained through direct inspection of observatory data. Fig-ure 4 shows magnetograms from four European observato-ries that recorded a large magnetic storm. Each observatoryshows a similar variation in H. Most of that is the magneticsignature of a large-scale, overhead ionospheric and magne-tospheric current system sustained during the storm.

In contrast, the magnetogram traces in Z vary signifi-cantly from site to site, due to localized subsurface differ-ences in electrical conductivity in the vicinity of each obser-vatory. Note, for example, the Z traces for the two Spanishobservatories: San Pablo Toledo (SPT), in the center of Spain,and Ebro (EBR), on the coast. Much of the variation in Z re-flected in the Ebro magnetogram comes from electric currentsinduced in the Mediterranean Sea. Since ocean water is agood electrical conductor compared with dry rock and hy-drated sediments, the nearby coastal inlets and sea-depthvariations establish a local conductivity heterogeneity; the re-sult is the complicated electromagnetic response seen in thedata. On the other hand, differences in Z variation betweenthe Italian and Romanian observatories are related to localgeology. Both areas are tectonically complicated, but the for-mation of the Carpathian Mountains created a zone of activerock metamorphism and unusually high electrical conduc-tivity that is manifested in the complexity of the lowest trace.

Quiet time variation and the ionosphereAbove Earths surface the magnetic field threads its waythrough the ionosphere, the electrically conducting part ofthe upper atmosphere where solar radiation maintains par-tial ionization.6,7 The degree of ionization is a function of al-titude, latitude, time of day, season, and solar-cycle phase. Ataltitudes of 90300 km or so, winds driven by daynight tem-perature differences and tides driven by the gravity of theMoon and Sun sustain motional induction. During quiettimes, when the magnetic field is relatively undisturbed bysolar activity, the electric currents of the ionospheric dynamogive a distinct diurnal variation to observatory magne-tograms, as evidenced in figure 5a. A detailed Fourier analy-sis of longer time series reveals frequencies corresponding tocoupled modulations driven by the solar cycle, Earths orbitaround the Sun, the Moons orbit around Earth, and Earthsrotation. Through application of Ampres law, the corre-sponding ionospheric currents can be mapped;8 figure 5b re-veals that the quiet-time current system is dominated by twoday-side current gyres. Earths rotation under that currentsystem gives rise to the quiet-time daily variation of the mag-netic field.

Prominent in Figure 5a is the daily variation in the mag-

netogram from Huancayo, Peru. First observed in the 1920s,soon after the Carnegie Institution of Washington estab-lished the Huancayo observatory, the variation is the resultof the ionospheres anisotropic electrical conductivity. OnEarths day side, in a roughly 5-wide band near the mag-netic equator, the horizontal ambient magnetic field and avertical electrical field maintained by charge separationacross the thickness of the ionosphere combine to facilitatethe eastwest motion of charge carriers. That gives a con-centrated flow of daytime electric current toward the eastand the observed enhancement of diurnal magnetic varia-tion at observatories like Huancayo that are located veryclose to the magnetic equator.

Magnetic storms and the magnetosphereThe extent of the geomagnetic field in near-Earth space de-fines the magnetosphere (see references 7 and 9 and the arti-cle by Syun-Ichi Akasofu and Louis J. Lanzerotti, PHYSICSTODAY, December 1975, page 28). The shape of the magne-tosphere, depicted in figure 6a, is determined by a supersonicsolar wind of electrons and ionized hydrogen and heliumthat moves at speeds of 2502000 km/s. Inside a resultingshock wave, the magnetic field of the magnetosphere on theSun side is compressed, with a magnetopause at about 10Earth radii (10 R). On the opposite, night side, the magne-tosphere is drawn out into a long tail whose length can exceed 100 R.

One can also describe the magnetosphere in terms of itsconstituent electric currents. The magnetopause is then de-fined by a surrounding current that flows eastward near theequatorial plane. The magnetotail can be defined in terms of

34 February 2008 Physics Today www.physicstoday.org

SPTSpain-4.3

EBRSpain0.5

AQUItaly13.3

SUARomania26.2

H

Z

300 nT100 nT

TIME (hours)0 5 10 15 20 25 30 35

Figure 4. Magnetic-field variation reveals differences insubsurface electrical conductivity. Shown here are magne-tograms taken on 2021 November 2003 at European ob-servatories of specified longitudes. Horizontal-intensity (H,blue) variation is driven by a storm in the magnetosphere.Site-to-site differences in the downward (Z, red) componentare due to induced electric currents in the ocean and crust.

a westward equatorial current sheet. The magnetospheric in-terior within about 36 R contains a neutral plasma of 1- to200-keV hydrogen and oxygen ions and lower-energy elec-trons. Those particles undergo a complicated dance consist-ing of cyclotron motion around magnetic-field lines, bouncesbetween mirror points in the Northern and Southern Hemi-spheres where field lines converge, and a slow migrationacross field lines due to gradients in the magnetic field. Thenet result is that ions tend to drift westward and electronseastward, a contrary motion that gives rise to a westwardequatorial ring current.

Along with the solar wind, the interplanetary magneticfielditself an extension of the heliomagnetic fieldcontrolsthe behavior of the magnetosphere. Diffusion allows the in-terplanetary magnetic field and geomagnetic field to connect,opening the interior of the magnetosphere to interplanetaryspace. Figure 6a depicts magnetic connection on the magne-tospheres day side (for additional discussion of magneticconnection, see PHYSICS TODAY, October 2001, page 16); onthe night side, an opposite process occurs with disconnectionof interplanetary and geomagnetic field lines in the magne-totail. With an open magnetosphere, the dragging of fieldlines across Earths polar cap by the solar wind establishes asolar-wind dynamo. That induces convection-like motion ofplasma in the magnetosphere and polar ionosphere, and itcan energize the ring current.

Occasionally, abrupt ejections or high-speed streams ofplasma from the Sun push the magnetosphere into a highlydynamic, time-dependent state called a magnetic storm. Thatcolorful expression was coined by von Humboldt in 1808 todescribe occasional periods during which ground-basedmeasurements show large, rapid, and irregular variation ofthe geomagnetic field. A magnetic storm can last from sev-eral hours to several days. Some also exhibit shorter-durationsubstorms.10 The cause and effect of substorms is controver-

sial, but, generally speaking, substorms result from a tempo-rary buildup of energy in the magnetotail that is released ex-plosively through a sudden collapse of part of the tail currentand diversion of current along magnetic-field lines. As a re-sult, the magnetospheric electric circuit closes through theionosphere, a detour that can give rise to beautiful auroraldisplays at high latitudes.

The Halloween stormOne of the largest magnetic storms on record occurred justbefore Halloween 2003. Figure 6b shows horizontal-intensitymagnetograms of that storm, which was initiated by a coro-nal mass ejection associated with a large sunspot group. Ob-servatory magnetograms recorded a sudden impulsivechange during the distinctive initial phase (I) that resultedfrom solar-wind compression of the magnetopause and mag-netic connection. The following main phase (M) was distin-guished by a general decrease in H at low magnetic latitudes,the signature of an increasing equatorial ring current: Thewestward ring current generates a southward magnetic fieldin its interior, the Earth side. Since the generated field pointsopposite to Earths prevailing northward dipole field, it de-creases H. Indeed, a longitudinal average of the disturbancein H from low-latitude observatories is proportional to theaverage increase in strength of the ring current. The recoveryphase (R) of the storm corresponds to ring-current diminu-tion and a return of low-latitude H to prestorm levels. TheHalloween storm is somewhat unusual in that it exhibitedtwo main phases, each followed by a recovery period.

A detailed comparison of low- and high-latitude mag-netograms reveals substorm occurrences during the Hal-loween storm. For example, figure 6b displays data from twocomparable longitude pairs of observatories that show in-termittent periods of anticorrelation. During those timespartial collapse of the ring or tail current gives an increase

www.physicstoday.org February 2008 Physics Today 35

BFEDenmark+55.4

TEOMexico+28.8

HUAPeru1.8

HBKSouth Africa27.1

EYRNew Zealand47.1

281 282 283 284 285 286

LOCAL DAY NUMBER

a b

50 nT

Figure 5. Magnetograms recording quiet times. (a) Datafrom five observatories show the horizontal intensity H for812 October 2003. The magnetograms are arranged bygeomagnetic latitude, which is measured relative to the axisof Earths tilted dipole. Diurnal variation is prominent, espe-cially for the Huancayo (HUA) observatory located near themagnetic equator. Note the change in polarity of the varia-tion at high latitudes. (b) Observatory data can be used tomap the electric currents in the ionosphere that sustain quiet-time variation. Shown here is a map for noon, GreenwichMean Time. Contour lines indicate 10-kA increments; more

densely packed contour lines indicate higher local current density. Red contours indicate clockwise-circulating current; bluecontours, counterclockwise current.

in H at low latitudes, while closure of the current systemalong field lines and through the ionosphere gives a simul-taneous decrease in H at high latitudes. Thus magnetic-observatory data can be used to monitor the electric circuitof the coupled magnetosphericionospheric system,notwithstanding that every magnetic storm has its ownunique and often complex character.

Space climate and weatherThe Suns dynamo is oscillatory. As a result, the Suns mag-netic polarity reverses once every 11 years or so and sunspotnumber and solar irradiance wax and wane (see reference 7and the article by Judith Lean, PHYSICS TODAY, June 2005,page 32). Figure 7 shows that magnetic activity, as measuredby the monthly standard deviation in H, is modulated inphase with the solar cycle. The spikes in magnetic standarddeviation during, for example, the years 1921, 1941, and 1989correspond to large magnetic storms. The discovery that

magnetic storms are more likely to occur during periods ofsunspot maxima and less likely to occur with sunspot min-ima was one of the most important in the history of spacephysics. It was made in 1852 by the astronomer and Britishmajor general Edward Sabine, who carefully analyzed a longtime series of data collected by various magnetic observato-ries including one located, at the time, in Toronto. In hiswords, the discovery gave to geomagnetism a much higherposition in the scale of distinct natural forces than was pre-viously assigned to it.11

Understanding magnetic storms is important for riskmitigation. Storm-induced currents in the crust can be a nui-sance for the electric power industry, since they can find theirway into power lines and transformers through ground con-nections.12 The most prominent example of that particularhazard occurred in March 1989, when a large magnetic stormled to the collapse of the electrical power grid serving the en-tire Canadian province of Quebec. Magnetic storms interferewith magnetic crustal surveys undertaken for mapping andmineral exploration, and they interfere with in situ magneticorientation systems used for directional drilling. Duringmagnetic storms, long-distance radio communication can bedifficult, and the accuracy of global positioning systems canbe reduced. In space, satellite electronics can be damaged andsatellite orbital drag enhanced. Astronauts and high-altitudepilots might be subjected to increased radiation.

Since magnetic storms were first identified throughground-based magnetic measurement, it is perhaps not sur-prising that standard measures of magnetic-storm size aredefined using magnetic-observatory data. Real-time obser-vatory data are used for low-cost monitoring or nowcast-ing of space weather. And historical observatory data enablestatistical studies of how storms are distributed in time andhow big they can be. Because of the potential risk to the ac-tivities and infrastructure of our modern, technology-basedsociety, the US federal government supports the interagencyNational Space Weather Program. Similar programs alsoexist in Japan and Europe.

The wide-ranging utility of magnetic-observatory datatestifies to the importance of programs dedicated to accurateand long-term geophysical measurement. And the datathemselves are a lasting legacy of the many hard-working in-dividuals who have supported observatory operations for al-most 170 years.

I thank William S. Leith for encouraging me to write this article andArnaud Chulliat, Carol A. Finn, David J. Kerridge, Stefan Maus, RobertL. McPherron, Lawrence R. Newitt, Leif Svalgaard, Jeremy N. Thomas,and E. William Worthington for reviewing a draft manuscript.

36 February 2008 Physics Today www.physicstoday.org

Interplanetary magnetic field

Solarwind

Connectionpoint

Ringcurrent

Tailcurrent

a

bABKSweden18.8QSBLebanon35.6

TIKRussia129.0BMTChina113.3

CMOAlaska212.2HONHawaii202.0

PBQCanada282.2SJGPuerto Rico293.9

4000 nT400 nT

I M R M R

302 303 304 305

TIME (days)

Figure 6. Earths magnetosphere is shaped and controlled bythe Sun. (a) Schematic representation of the magnetosphere.Blue lines represent magnetic fields, and red lines representcurrents. Magnetic connection is shown on the day side of themagnetosphere, where a southward-pointing interplanetaryfield merges with a northward-pointing geomagnetic field. (b) Data from eight observatories show the disturbed horizon-tal intensity H during the Halloween magnetic storm of 2831October 2003. The magnetograms are arranged by longi-tude; red traces correspond to low-magnetic-latitude observa-tories and blue lines to observatories at high latitudes. Theinitial (I), main (M), and recovery (R) phases are discussed inthe text. The circled anticorrelations between high- and low-latitude magnetograms are indicative of substorms.

References1. S. Chapman, J. Bartels, Geomagnetism, Clarendon Press, Oxford,

UK (1962); S. Matsushita, W. H. Campbell, eds., Physics of Geomag-netic Phenomena, Academic Press, New York (1967); J. A. Jacobs,ed., Geomagnetism, Academic Press, San Diego, CA (198791).

2. J. Jankowski, C. Sucksdorff, Guide for Magnetic Measurements andObservatory Practice, International Association of Geomagnetismand Aeronomy, Warsaw, Poland (1996).

3. F. H. Busse, Annu. Rev. Earth Planet. Sci. 11, 241 (1983); P. H.Roberts, A. M. Soward, Annu. Rev. Fluid Mech. 24, 459 (1992).

4. V. Courtillot, J. L. Le Moul, Annu. Rev. Earth Planet. Sci. 16, 389(1988); J. Bloxham, D. Gubbins, A. Jackson, Philos. Trans. R. Soc.London Ser. A 329, 415 (1989); S. Maus et al., Phys. Earth Planet.Inter. 151, 320 (2005).

5. I. I. Rokityansky, Geoelectromagnetic Investigation of Earths Crustand Mantle, N. L. Chobotova, G. M. Pestrylakov, B. G. Shilman,trans., Springer, New York (1982); F. Simpson, K. Bahr, PracticalMagnetotellurics, Cambridge U. Press, New York (2005).

6. M. C. Kelley, The Earths Ionosphere: Plasma Physics and Electrody-namics, Academic Press, San Diego, CA (1989); A. D. Richmond,EOS Trans. Am. Geophys. Union 77, 101 (1996); R A. Heelis, J. Atmos. Sol.-Terr. Phys. 66, 825 (2004).

7. M. G. Kivelson, C. T. Russell, eds., Introduction to Space Physics,Cambridge U. Press, New York (1995); G. W. Prlss, Physics of theEarths Space Environment: An Introduction, Springer, New York(2004).

8. S. R. C. Malin, Philos. Trans. R. Soc. London Ser. A 274, 551 (1973);D. E. Winch, Philos. Trans. R. Soc. London Ser. A 303, 1 (1981).

9. D. P. Stern, N. F. Ness, Annu. Rev. Astron. Astrophys. 20, 139(1982); C. T. Russell, Annu. Rev. Earth Planet. Sci. 19, 169 (1991);J. G. Lyon, Science 288, 1987 (2000).

10. C. T. Russell, R. L. McPherron, Space Sci. Rev. 15, 205 (1973);Y. Kamide et al., J. Geophys. Res. 103, 17705 (1998).

11. E. Sabine, Proc. R. Soc. London 8, 395 (185657).12. J. G. Kappenman et al., IEEE Trans. Plasma Sci. 28, 2114 (2000);

D. H. Boteler, Nat. Hazards 23, 101 (2001); R. Pirjola, Adv. SpaceRes. 36, 2231 (2005).

See www.pt.ims.ca/16297-14

50

10

DE

VIA

TIO

N(n

T)

100

10

1

SU

NS

PO

TN

UM

BE

R

1900 1910 1920 1930 1940 1950 1960 1970 1980 1990 2000

YEAR

a

b

Figure 7. Magneticactivity is driven by theSun. (a) The monthlystandard deviation inthe horizontal intensityH as measured atGerman observatories.(b) Monthly averagesfor sunspot number.