The Cluster Mission Editorial

Home , Cluster (spacecraft), Polar (satellite)

INTERNATIONAL SPACE SCIENCE INSTITUTE No 9, June 2002

Published by the Association Pro ISSI

The Fourfold Way Through the Magnetosphere: The Cluster Mission Editorial

Imagine a lonely spacecraft trav- currently Director of the Inter- Impressum elling through space. Its mission national Space Science Institute in is to probe the value of, say, the Bern and Senior Scientists at the strength of the local magnetic Max Planck Institute for extra- field. Its instruments continuously terrestrial Physics in Garching SPATIUM monitor the field strength and de- (Germany) is one of the fathers of Published by the liver the data to the on-board Cluster. He is also the Principle Association Pro ISSI telecommunication system, which Investigator of the Electron Drift twice a year forwards them to the ground con- Instrument designed to determine trol station on Earth. There, scien- the strength of the local electric tists try to interpret the data. field around the spacecraft. INTERNATIONAL SPACE SCIENCE But now a basic dilemma arises: In November 2001, very shortly INSTITUTE due to the continuous motion and after the commissioning of the Association Pro ISSI temporal evolution of the ambi- Cluster spacecraft G. Paschmann Hallerstrasse 6, CH-3012 Bern ent medium and its boundaries, presented the first results to our Phone ++41 31 631 48 96 the recorded data are a complicat- members. In the mean time, this Fax ++41 31 631 48 97 ed mix. The observer cannot dis- mission has delivered a wealth of tinguish whether the magnetic additional data, which partially President field experienced a change in time could be considered in the present Prof. Hermann Debrunner, or it varied spatially or both hap- revised version of his lecture. We University of Bern pened simultaneously. are very indebted to G. Pasch- Publisher mann for his kind permission to Dr. Hansjörg Schlaepfer, That was the problem faced by publish the present text on the legenda schläpfer wort & bild, the fathers of the European Space fascinating world of magnetic Winkel Agency’s Cluster mission in 1982 fields. Layout intended to exploring the Earth’s Marcel Künzi, marketing · kom- magnetosphere. Their idea was to munikation, CH-8483 Kollbrunn solve the problem by flying a Printing quartet of closely spaced identical Hansjörg Schlaepfer Druckerei Peter + Co dpc spacecraft, since time and space Bern, June 2002 CH-8037 Zurich need a set of three co-ordinates to be defined completely. The instruments aboard the four spacecraft may now take measurements at identical times but at different places in space allowing the scientists to work out the rele- vant temporal and spatial charac- teristics of the magnetic field and Front Cover: The aurora is the the other quantities that charac- only visible proof of the Earth’s terize the ambient medium, such magnetosphere. This wonderful as the electric field, and plasma picture shows the glow from out- density, velocity and temperature. bursts of ionic particles from the Sun in the ionosphere. (Credit & The author of the present issue Copyright: Dennis Mammana, of Spatium, Dr. Götz Paschmann, Skyscapes)

SPATIUM 9 2 The Fourfold Way Through the Magnetosphere: The Cluster Mission *)

Götz Paschmann,International Space Science Institute,Bern and Max Planck Institut für extraterrestrische Physik,Garching

Introduction ral evolution. Cluster is the first forward. One should not forget attempt to do this in space, admit- though that because of the com- tedly with only the minimum plex structure of the magnetos- number that can be arranged in phere and the strong interrela- three dimensions, namely four. tionship between its parts, there is The Earth’s magnetosphere has There have previous attempts the danger that spacecraft flying in been investigated by direct in-situ with two spacecraft (ISEE and close formation will miss the measurements since the begin- AMPTE), but with two stations large-scale context. Fortunately, ning of the space age and has one can only infer variations along observations from the ground and turned out to be the site of many the line between the spacecraft, from other spacecraft are available fascinating phenomena and pro- and three stations always form a that serve to fill in this global cesses that must be studied in-situ plane. Thus Cluster is a giant step context. because they are not accessible by remote sensing techniques. Exam- ples are shock formation, magnetic reconnection, particle accelera- Interplanetary Magnetic Field tion, wave-particle interaction and Tail Current turbulence. While interesting in their own right, as part of the ex- ploration of the Earth’s environ- ment, they can also guide our Plasma Mantle understanding of similar processes on the Sun or the distant Magnetotail Northern Lobe universe, places where one can never make in-situ measure- Polar Cusp ments. Figure 1 shows a view of the Plasma Sheet Plasma- Neutral Sheet magnetosphere that emphasizes sphere Current

its main features, except for the t

bow shock that has been left out n e r for clarity. u r Ring C While much has been learned Magnetopause about the magnetosphere through Solar Magneto- measurements from single satel- pause Field-Aligned Current Current lites, a limitation has been the Wind fundamental inability of a single observer to unambiguously distinguish spatial from temporal changes. The magnetosphere is constantly changing its shape and Figure 1 size, and many of the processes Cutaway view of the Earth’s magnetosphere that identifies the key regions en- countered by Cluster: the solar wind approaching from the left; the magnetopause governing it are known to vary on (yellow surface) that is the outer boundary of the magnetosphere; the polar cusp, a short spatial and temporal scales. funnel-shaped indentation in the magnetopause; the magnetotail, with the hot A solution to this dilemma is a plasma sheet at the centre and the lobes on either side. Thin black lines with arrows fleet of mobile stations, very are the magnetic field lines, dashed lines show plasma flow. The black arrows mark electric current flow. The bow shock that stands in the solar wind ahead of the much like what one does to study magnetopause, and the region in between, the magnetosheath, have been omitted weather patterns and their tempo- for clarity.

th *) Pro ISSI lecture, Bern, 8 November 2001 SPATIUM 9 3 The Cluster Spatium Nr. 2. Cluster was not so and successfully launched on two fortunate. Ready for launch in Soyuz-Fregate launchers from Mission 1995, the four spacecraft were de- Baikonur on 16 July and 12 August stroyed when the Ariane 5 rocket 2000, respectively. After extensive exploded during its maiden flight commissioning of the payload, the on 4 June 1996. At first it seemed science phase officially began on that the only remaining option 1 February 2001. was to put together a single space- History craft from the leftover parts, ap- This being a Pro ISSI lecture, I propriately named Phoenix after should point out that several peo- Cluster’s history goes back to the mythical bird that rose from ple associated with ISSI were November 1982, when a group of the ashes. But recognizing the deeply involved in Cluster. Two European scientists proposed the unique importance of the mis- of its directors, Bengt Hultqvist mission. In February 1986 ESA sion, the ESA Science Programme and myself, and a long-time mem- chose Cluster and the Solar and Committee (SPC) on 3 April ber of its Science Committee, Heliospheric Observatory (SOHO) 1997 agreed that three new Gerhard Haerendel, were among as the first “Cornerstone” in its Cluster spacecraft should be built the seven who proposed the mis- Horizon 2000 Science Pro- alongside Phoenix, and thus sion in 1982. Roger Bonnet, then gramme. SOHO was launched in Cluster was reborn. Through a Director of Scientific Programmes December 1995 and has sent back tremendous effort by all involved, at ESA, was the key force behind a stream of exciting data about the fully instrumented satellites the recovery initiative, and Hans the Sun, as vividly described in were rebuilt in only three years Balsiger was chairman of ESA’s

ck magnetosheath sho w bo magnetopause

Bow shock Polar cusp cusp Auroral zone neutral sheet tra region Plasmas- phere lobes

Magnetopause Solar wind

Figure 2 Cluster orbit in February, as it cuts through the key regions of the magnetosphere.

SPATIUM 9 4 SPC, when the decision to repeat superimposed on a cut of the mag- 10 5 Nominal phase Extension phase Cluster was finally made in April netosphere, and illustrates that 1997. Last but not least, Hans Cluster moves outbound over the 10 4 Peter Schneiter and Contraves northern polar cap, crosses the mag- 10 3 Space had a leading role in the re- netopause and bow shock into the max vival efforts by the industry that solar wind, before recrossing those min 10 2 had built Cluster. It is therefore boundaries in reverse order and cusp only natural that the interpretation moving over the southern polar tail 10 1 of the Cluster measurements will cap back towards perigee. Figure 2 distance [km]Inter-spacecraft Aug-00 01 02 03 04 05 date be one of ISSI’s future workshop applies to February of each year. Figure 4 activities. As the orbit is inertially fixed, it Spacecraft separation distances as rotates around the Earth once a planned for the entire mission. year, as the latter revolves around Orbits and Separations the Sun. As a result, the apogee regular tetrahedron when cross- of the orbit will be located in the ing the major boundaries. The After their successful launch from geomagnetic tail half a year later, separation distances between the Baikonur, the Cluster spacecraft as shown in Figure 3. In the course spacecraft are being changed be- were placed in nearly identical, of the mission Cluster will thus tween 100 km and tens of thou- highly eccentric polar orbits, with encounter the regions of interest sands of km during the mission. an apogee of 19.6 Earth, a perigee several times. The orbits are tuned This allows studying the phe- of 4 Earth radii, and a period of so that the four spacecraft are nomena at sufficiently different 57 hours. Figure 2 shows the orbit located at the vertices of a nearly scales. Figure 4 shows the sequence

ck magnetosheath sho w bo magnetopause

Mid-altitude cusp Magneto- Polar cusp tail neutral sheet trapping region

lobes

Solar wind

Figure 3 Same as Figure 2, but for the month of August.

SPATIUM 9 5 2 5

2 1

4 1 2 3 2 1 3 1 2 3

1 STAFF (N. Cornilleau-Wehrlin, F) 1 PEACE (A. Fazakerley, UK) 1 FGM (A. Balogh, UK) Magentic and electric fluctuations Electrons (E < 30 [keV]) Magnetic field 2 EFW (M. Andre, S) 2 CIS (H. Reme, F) 2 EDI (G. Paschmann, D) Electric fields and waves Ion composition (E < 40 [keV]) Plasma drift velocity 3 DWP (H. Alleyne, UK) 3 RAPID (P. Daly, D) (0.1 < E < 10 [mV/m], Wave processor High energy electrons and ions 5 < B < 1000 [nT])

4 WHISPER (P. Decreau, F) (2 < Eions < 1500 [keV/nuc], 3 ASPOC (W. Riedler, A) Electron density and plasma waves 20 < Ee < 400 [keV]) Spacecraft potential control 5 WBD (D. Gurnett, USA) Electric field wave-forms

Figure 5 The science payload carried by each of the 4 Cluster spacecraft. The group listed on the left comprises the Wave Experiment Consortium, the three in the middle deal with the particle component, and the top two on the right address measurements of the magnetic field and the plasma drift velocity, respectively, while the last controls the spacecraft potential. of separation distances planned States, Canada, China, the Czech part of a “kinetic sculpture” called until the end of the mission. This Republic, Hungary, India, Israel, Moving Plates by their creators, separation strategy is possible be- Japan and Russia. Viennese artists Eva Wohlgemuth cause each Cluster spacecraft car- and Andreas Baumann. Other ried 600 kg of fuel at launch, of Figure 5 also gives an impression of such plates are mounted on sta- which about half was expended the large size of the spacecraft. Its tionary as well as moving objects in order to reach the operational diameter is almost 3 m and its all over the world. orbit. mass at launch 1200 kg, of which half was in form of fuel needed to reach the final orbit and to carry Excursion: The Electron Drift Science Payload out the separation strategy. The Instrument mechanical structure of the space- The four Cluster spacecraft carry craft was developed and built by The Cluster payload contains one identical sets of 11scientific instru- Contraves Space, while Dornier instrument, the Electron Drift In- ments, designed to measure the was ESA’s Prime Contractor for strument (EDI), which is unique ambient electromagnetic fields and Cluster, which included integra- because it actively emits beams of particle populations over a wide tion and test of the entire system, electrons to sense the electric range of frequencies and energies, including instruments. fields in the ambient plasma. This respectively (see Figure 5). Each in- instrument was developed by a strument was built by a team Almost unnoticed by even the consortium consisting of the under the leadership of a Principal teams involved, one of the space- University of New Hampshire, Investigator (PI). There are more craft is carrying an extra piece of the University of California at than 200 Co-Investigators from hardware, namely an engraved San Diego, Lockheed Palo Alto ESA member states, the United 10-by-10 cm Titanium plate that is Research Laboratory, and the

SPATIUM 9 6 distorted into cycloid orbits, such tions. By employing two beams that the electrons are displaced and two detectors, the two unique from their origin after one gyra- directions can be monitored con- tion. The displacement (the “drift tinuously and the displacement step”) is directly proportional to obtained by a triangulation proce- the electric field and inversely dure. Figure 6 illustrates the prin- GDU2 proportional to the square of the ciple, and also shows that the two magnetic field, and ranges from a beams travel different distances, GDU1 E few centimetres to hundreds of and thus have different travel meters. Under these conditions times. When the magnetic field the beam will return to the space- becomes small, the electric field is B v craft only when emitted in either obtained from these time-of-flight of two precisely defined direc- differences. Figure 6 EDI’s principle of operation. For any combination of magnetic field B and Cluster-EDI (HR) Triangulation 20-Oct-2000 16:10:02.500 0.250” drift velocity V (induced by an electric 10 field E), only a single electron trajectory exists that connects each electron gun in [m] with the detector located on the oppo- gyro site side of the spacecraft. The two y trajectories have different path lengths and thus different times-of-flight. 5

Max-Planck-Institut für extraterrestrische Physik in Garching. d = Vd T g Because I happen to be the PI for ~ E / B2 the EDI instrument, I will dwell d on it in some detail. 0

When injected perpendicular to a magnetic field, electrons perform a circular orbit (with a radius proportional to their velocity) that –5 returns them to their origin after a complete gyration. Gyro radius and gyro time both are inversely proportional to the strength of q = 3 the magnetic field. For electrons with 1 keV energy and a mag- –10 netic field of 100 nT, the gyro –10 –5 0510 x in [m] radius is 1000 m and the gyro gyro Figure 7 time 0.35 ms. Sample EDI triangulation plot for an interval of 0.5 s duration. Red and green lines are the firing lines of the two electron guns for those times when the beams were In the presence of an electric field, properly aimed to return to the detectors. The gun positions, projected into the plane the electrons execute a drift mo- perpendicular to the magnetic field, move on an ellipse because the spacecraft spin axis and the magnetic field direction are not aligned. The 0.5 s interval corresponds to tion in addition to their gyration. 1 /8 of a full spin. The measured “drift step” is the vector from the beam intersection As a result, the circular orbits are to the centre of the figure, and is proportional to E/B2.

SPATIUM 9 7 While exceedingly simple in prin- electrons. As a by-product, one Following ciple, it is the implementation that also obtains a very precise meas- has made the Electron Drift In- ure of the magnetic field this way Cluster from the strument a formidable challenge. (from the average of the two The first difficulty lies in the need times-of-flight, which is equal to Solar Wind to the to keep the beams precisely per- the gyro time). Further difficulties Ionosphere pendicular to the ambient mag- arise because the beams diverge. netic field, which can have ar- Thus fewer electrons return to bitrary direction and can vary the detectors the larger the dis- rapidly. This is overcome by a tance they have to travel, i.e., We now will follow Cluster unique electron gun design that the smaller the magnetic field through key regions of the mag- can emit a narrow beam in any di- strength is. This is partially over- netosphere and demonstrate its rection within more than a hemi- come by continuously adjusting capabilities with illustrative exam- sphere, and by having a direct the beam current (between 0.1 ples, starting from the solar wind on-board link to the Cluster mag- and 300 nA) and the detector and ending with the aurora in netometer instruments (FGM and sensitivity. the polar ionosphere. STAFF) that allows reconstruc- tion of the instantaneous magnetic The onboard software runs a 4-ms field direction. The next difficulty servo loop to execute the beam Solar Wind is to find the direction that returns acquisition and tracking tasks and each beam to its associated detector. to navigate through the various The solar wind is a plasma stream This is achieved by sweeping the control parameters. More than 50 that approaches the Earth with beam in the plane perpendicular patches of that software were speeds between 300 and 800 to the magnetic field and waiting needed before it worked satisfac- km/s. It carries with it the mag- until the detectors record a signal. torily under the variable condi- netic field of solar origin, the in- While this is going on, the detec- tions experienced along the Clus- terplanetary magnetic field (IMF). tor look direction is steered in ter orbit. Figure 7 shows one of Solar wind and IMF are both synchronism with the beam firing the first successful measurements, highly variable, which is one of directions. Once signal has been using the triangulation technique. the reasons the magnetosphere is acquired, the beam is swept back The vector, d, from the beam in- so dynamic. and forth in order to continuously tersection to the centre of the track the target direction. spacecraft is the “drift step”, which As an example, let us look at a is proportional to the ratio E/B2. sudden change in IMF direction Because there usually are plenty Using the magnetic field, B, meas- caused by the passage of the he- of natural electrons at the same ured by the FGM instrument, liospheric current sheet (HCS) energy as the beams, care must be the electric field, E, is readily ob- over the Cluster spacecraft. The taken to distinguish the beam tained. Results obtained with EDI HCS is a warped surface dividing electrons from the natural elec- are discussed in the next section. the heliosphere into two hemi- trons. This is achieved by coding spheres of oppositely directed the beams with a pseudo-noise magnetic fields. Cluster encoun- code, much the same way as in a tered the HCS on 13 February GPS system. By correlating the 2001 at 10:48 universal time received electrons with the code (UT), and the results are present- of the emitted electrons, one not ed in Figure 8. The figure shows only discriminates against the nat- the magnetic field measurements ural electrons, but one also meas- and their analysis by the FGM ures the times-of-flight of the team that highlights the unique

SPATIUM 9 8

B 360 Cluster 1 Φ 300 Cluster 2 Azimuthal Field Angle Cluster 3 240 Cluster 4

180

120

60 0

B 0.010 Linear Estimate of curl B components 0.005 curl [nT/km] 0.000

components (curl B) x –0.005 (curl B) y –0.010 (curl B) z B 0.015 Linear Estimate of div B

div 0.010

[nT/km] 0.005 0.000 10: 30 10: 35 10: 40 10:45 10:50 10:55 11:00 time (hours:minutes) Figure 8 An encounter of the heliospheric current sheet (HCS). The top panel shows the magnetic field direction that showed a large rotation near 10:47 UT as the HCS crossed the Cluster formation. The middle panel shows the curl of the magnetic field, estimat- ed from the four spacecraft magnetic field data: the large values during the passage of the HCS are caused by the current within the structure. The bottom panel shows an estimate of div B: this remains small throughout, providing a consistency check for the estimate of curl B (from Eastman et al., 2001). capability of Cluster to derive Bow Shock vantage point. Figure 9 shows a spatial gradients in three dimen- crossing of the bow shock when sions. Assuming that the field The first sign of any interaction be- the IMF had a direction that made varies linearly between the space- tween the solar wind and Earth’s it nearly aligned with the shock craft positions, the density of elec- magnetic field is a shock wave in surface, a geometry referred to tric current flowing in the volume space on the Sun-facing side of the as (quasi-)perpendicular because between the spacecraft can be de- Earth. This bow shock is rather the field is then perpendicular to rived from the magnetic field like the sonic boom caused when the normal to its surface. In such a measurements at the four points, an aircraft flies at supersonic situation the shock has a very sim- using Ampère’s law. The figure speed. The role of the bow shock ple step-like structure and the demonstrates the success of the is to slow down the solar wind to profiles recorded by the 4 space- technique: precisely when the subsonic speed such that it can craft are essentially just time-shift- HCS passes over the spacecraft, be deflected around the obstacle ed copies of each other. From the there is a spike in the computed formed by the Earth’s magnetic time shift and the known distance current density. Note that as a shield, the magnetosphere. between the spacecraft one ob- consistency check, one also com- tains the speed at which the shock putes the divergence of B, which Simply inspecting the raw magne- passed over the spacecraft, about according to Maxwell’s equations tometer data already reveals the 5 km/s in this case. Together with should be zero. superiority of a four-spacecraft the measured duration of the shock

SPATIUM 9 9 FGM Cluster 25 December 2000 transition, this speed yields the 50 thickness of the shock, a quantity Cluster 1 that is an important for under- Cluster 2 standing shock formation in colli- 40 Cluster 3 sionless plasmas. Cluster 4

30 When, by contrast, the IMF is more nearly aligned with the shock normal, it becomes a (quasi-)parallel 20 shock. Parallel shocks are characterized by oscillating transitions, and are much harder to analyse. 10 They are, however, exceedingly important, because it is such paral- 0 lel shocks (created by supernovae 50 explosions) that are thought to be Cluster 2 responsible for acceleration of galactic cosmic rays. As a conse- 40 quence of the parallel geometry, particles are free to traverse the 30 shock in both directions. This result is an extended “foreshock” region that is characterized by parti- 20 cle beams, waves, turbulence and first-order Fermi acceleration. Deep in this foreshock one finds

Magnetic Field Strength [nT] 10 Short Large Amplitude Magnetic Pulsations (SLAMP’s), which are 0 believed to play the dominant role 50 in the thermalisation of the in- Cluster 1 cident plasma (via reflection and Cluster 3 subsequent mixing analogous to 40 Cluster 4 that at quasi-perpendicular shocks). Figure 10 shows two SLAMPs ob- 30 served by Cluster on 2 February 2001, when the spacecraft were separated by about 600 km. De- 20 spite both theoretical and previous experimental estimates of SLAMP sizes larger than 3000 10 km, the Cluster measurements re- veal considerable spatial variabili- 0 ty on scales less than 600 km. It 02:12 02:15 02:18 02:21 02:24 will be interesting to see how the Time [hours:minutes] situation changes when Cluster Figure 9 Magnetic field measured by the FGM instruments on the four spacecraft for a cross- separations have been reduced to ing of a quasi-perpendicular shock. 100 km (see Figure 4).

SPATIUM 9 10 Magnetopause 11. The plasma density is deduced ties measured at a given time on in this case from the plasma fre- the four spacecraft, the density Because charged particles are deflect- quency measured by the WHIS- gradients inside the current layer ed by magnetic fields, the Earth’s PER instrument. Using the values can be determined. Similarly, the magnetic field is an obstacle that of the density measured simulta- current density can be determined the solar wind cannot simply pen- neously at the four spacecraft, the from the magnetic field differ- etrate. Instead it flows around it, density gradient can be calculated, ences, using the method illustrat- confining the terrestrial magnetic as has been done for a magne- ed in Figure 8. The gradient vectors field to a cavity, the magnetosphere topause crossing on 2 March 2002 are shown at the bottom of Figure (see Figure1). Its boundary, the mag- shown in Figure 11. At the top, the 11, projected onto one of three or- netopause, is located where the figure shows the measured plasma thogonal planes (the other two plasma pressure exerted by the densities on the four spacecraft. At projections not shown for clarity). solar wind is in equilibrium with the beginning of the interval, the Knowledge of the instantaneous the magnetic pressure from the spacecraft were in the magnetos- density gradients and current den- Earth’s magnetic field. Across the phere, characterized by a tenuous sities are essential for understand- magnetopause, the magnetic field plasma, then they crossed the ing the formation of the magne- usually undergoes a sharp change magnetopause around 03: 31 UT topause, and for distinguishing in magnitude and direction. The and went into the magnetosheath different plasma transport mecha- magnetopause is thus a current with a density about 20 cm–3 for nisms, but neither information was layer that separates the solar wind less than a minute before re-en- available before Cluster. and its embedded interplanetary tering the magnetosphere at about magnetic field from the Earth’s 03:32 UT. Note that at this time, magnetic field and plasma. the Cluster spacecraft were only Magnetopause Thickness about 100 km apart. As a result, and Motion they were all located in the thin Density Gradients magnetopause simultaneously, as As a result of dynamical variations demonstrated that the times of the in the solar wind and interplane- The four spacecraft measure- four density traces in Figure 11 tary magnetic field, the magneto- ments allow the measurement of clearly overlap. The same is true pause is in constant motion, which spatial gradients of scalar quanti- for the magnetic field traces from makes determination of its struc- ties, such as the gradient of the the FGM instrument (not shown). ture virtually impossible from sin- plasma density, as shown in Figure From the differences in the densi- gle-spacecraft data.

30 2nd February 2001

20 Magnetic Field [nT] 10

0 20:58:30 20:59:00 Figure 10 Shock Large Amplitude Pulsations observed by the 4 Cluster spacecraft in the vicinity of the quasi-parallel bow shock. Note the significant differences between the various spacecraft despite their modest (~600 km) separation (from Lucek et al., 2001).

SPATIUM 9 11 ] Figure 12

3 (top panel) shows time series of the magnetic field measured by the FGM instruments on 20 the four spacecraft for a magnetopause crossing on 4 July 2001. From the four crossing times and the known spacing between the 15 Density [particles/cm spacecraft one can derive the magnetopause velocity, 30 km/s in this case. The bottom panel shows 10 an expanded view of the magnetopause crossing by spacecraft C3. From Figure 12 one infers a thickness of the magnetopause current 5 layer of approximately 200 km, corresponding to about three ion Rumba Salsa Larmor radii, which is physically Samba reasonable, although theory makes Tango no definite prediction. 3:31:00Gradient density vectors3:31:30 [part/cm3/km] along 3:32:00 trajec Time [hh:mm:ss] While determination of the cur- onto 3 GSE planes, and magnitude of gradient dens rent layer thickness has impor-

) nd tance in its own right, the most E 2 March, 2002, 03:31:00 to 03:32:00 striking aspect of the data in Figure

Z (R 8.41 70 12 is the sharp rise in plasma density (from 0.1 to 24 cm3) across 60 the current layer (shown in the bottom panel), because it means that at this time essentiality no 50 solar wind plasma was able to 8.4 enter the magnetosphere, neither 40 at this location nor further up- stream. This is an unexpected result, because it is generally 30 thought that solar wind plasma can enter the magnetosphere by 8.39 Z 20 any number of processes.

The four-spacecraft timing pro- 10 vides the magnetopause velocity X 7E 02 only at four discrete instances in time, which is often not enough to 7.18 7.17 7.16 X (R ) x 1E.03 E infer the boundary thickness and Figure 11 its evolution. But it turns out that Plasma densities measured on the four Cluster spacecraft by the WHISPER instru- the measurements of the plasma ments for a magnetopause crossing on 2 March, 2002 (top), when the spacecraft separation distances were only 100 km. At the bottom, the figure shows the gradient vectors drift velocity with the Electron inferred from the instantaneous differences in the densities (from Decreau et al., 2002). Drift Instrument (EDI), referred

SPATIUM 9 12 60 50 40 Cluster 1 30 Cluster 2 FGM B [nt] 20 Cluster 3 10 Cluster 4 0 80

[km/s] 60

I t MP V 40 I

20 EDI 0 40 20 [km/s]

n 0 –20 EDI V –40 23:55:00 23:56:00 23:57:00 23:58:00 23:59:00 4th July, 2001

100.0 40 [nt]

x Bx 10.0 20 ] 3 200 km – 0 1.0 FGM B –20 [cm n 0.1 n –40 CIS 23:58:20 23:58:40 23:59:00 23:59:20 Figure 12 Magnetopause crossings by the four Cluster spacecraft on 4 July 2001, identified by the sudden drop in magnetic field strength measured by the FGM instruments (top panel). The second and third panels show, respectively, the components of the plasma convection velocity tangential (Vt) and normal (Vn) to the magnetopause, as measured by EDI on C3. The velocity Vn is interpreted as the magnetopause velocity. When integrated over time, it gives the instantaneous distance of spacecraft C3 from the magnetopause (not shown). The bottom panel shows, for an expanded time scale, the x-component of the magnetic field (from FGM) and the plasma density (from CIS), measured on C3 across the magnetopause. The horizontal bar shows a distance scale (200 km) obtained from the measured magnetopause velocity. to earlier, and by the ion spec- magnetosphere. Integrating this pletely separates the Earth’s mag- trometer CIS, can provide a con- velocity over time provides a con- netic field and the solar wind. The tinuous estimate of this motion, tinuous estimate of the magne- example in Figure 12 demonstrates because, to first order, plasma and topause distance (not shown. This that this situation can actually magnetopause move together. Fig- information permits reconstruc- occur. More commonly, however, ure 12 shows, as the third panel, the tion of the complete spatial pro- that separation is far from perfect, projection, Vn , of the drift velocity file, for example of the plasma and the investigations of the pro- measured by these two instru- density. cesses that allow transfer across ments on C3 onto the magneto- the magnetopause are at the heart pause normal direction. Vn varies, of the Cluster mission.The one pro- but is negative on average, corre- Magnetic Reconnection cess known to play a key role for sponding to a net in-ward motion transfer of mass, momentum, and of the magnetopause that eventu- To first order, the magnetopause is energy across the magnetopause, ally causes the spacecraft to exit the an impenetrable surface that com- is magnetic reconnection. This pro-

SPATIUM 9 13 ] 40 cess is the result of the breakdown 3 of the “frozen-in” picture. Under – S/C1 multiple MP 30 [cm ideal conditions, the interplane- i N tary magnetic field is frozen into 20 the solar wind plasma, and therefore slips along the magnetopause 10 M’sphere with the solar wind flow. But M’sheath when this concept breaks down 0 locally, by processes not yet identi- 600 S/C1 over fied, the IMF can diffuse relative jets 2 hours [km/s] to the plasma and connect with the I 400 of jets v terrestrial magnetic field across the I at MP magnetopause. This is sketched in 200 Figure 13. Once reconnected, the magnetic field lines are dragged CIS over the poles and into the mag- 0 600 netotail because they are embed- S/C3 ded in the solar wind flow. [km/s] I 400 V2 v I

200 B1 t 3 B2 CIS t t2 2 Magnetos- X-line 0 t phere t11 t1 600 V 1 S/C4 Solar

V1 [km/s] Wind I 400

Diffusion v Region I

200 CIS V 2 0 Figure 13 08:00 10:00 12:00 Schematic to illustrate reconnection hh:mm th across the magnetopause between mag- 26 January, 2001 netic field lines from the Sun and from

Earth. A pair of field lines that at time t1 are still entirely in the solar wind and Cluster magnetosphere, respectively, connect through the magnetopause at time t2 and become two sharply kinked field jet lines at time t3 that now have one end on the Earth, the other on the sun. The reconnection-line reconnection site is a line that is termed X-line or reconnection line. The solar wind plasma approaches at speed V1 and becomes accelerated to V2 as a result of the tension in the sharply bent field Figure 14 lines. It is those high-speed plasma jets Plasma density (on C1) and velocities (on C1, C3, C4) measured by CIS-HIA for that are the most directly observable the magnetopause encounter on 26 January 2001. The high velocities indicate the signature of magnetic reconnection at encounter with plasma jets emanating from a reconnection-line located somewhere the magnetopause. below the spacecraft (courtesy Phan et al.).

SPATIUM 9 14 When reconnection occurs, solar the cusp. Close to perigee, when wind plasma can enter the magne- these measurements are made, the tosphere by fluid flow along re- Cluster spacecraft have much connected field lines. Upon entry larger separation distances than it is accelerated to large speeds by further out along their orbit. This the magnetic tension forces that explains the large time shift be- exist across the magnetopause be- Exterior tween the encounters, especially cause the reconnected field lines Magnetosheath Cusp that by spacecraft C3. The differ- are sharply bent, as shown in ence in appearance is probably Figure 13. We were the first to caused by a reconfiguration of the observe the resulting plasma jets cusp before the encounter by C3. with an instrument on ISEE in Entry Layer 1979. This discovery formed the “smoking gun” evidence for re- Polar Cap connection at the magnetopause. Moving tailward from the polar One of the open questions is cusp, one enters the polar cap re- whether reconnection can happen gion that is permeated by magnet- in a quasi-stationary fashion, or ic field lines that extend deep into whether it is necessarily intermit- the geomagnetic tail where the tent. As the plasma jets are local- Figure 15 form the tail lobes (see Figure 1). ized in thin layers, they can be ob- Expectations for the distant polar The polar cap is often void of served only when this layer passes cusp region, as envisaged in the 1982 any plasma, but CIS/CODIF Cluster proposal. over the spacecraft. With a single data show the frequent presence spacecraft one could never be sure of cold oxygen ions on polar that reconnection had not actually wind. This is due to their funnel cap magnetic field lines. Figure 17 stopped in between successive shape and their small magnetic shows an energy-time spectrogram crossings. With Cluster one now field. This unique feature was one of O+ ions for an hour on 4 March has closely spaced crossings that of the drivers for the Cluster idea, 2001. The feature to note is the allow to fill the gaps. Figure 14 as illustrated by Figure 15, taken narrow band of intense fluxes that shows data from the CIS instru- from the 1982 proposal. As illus- proceeds to lower and lower ener- ment that demonstrate that the trated by Figure 2, the Cluster obit gies while the spacecraft moves plasma jets are observed when- crosses the distant cusp region from the cusp over the polar cap ever any of the spacecraft crosses when its apogee is one sunward towards the magnetotail. For the magnetopause. For this inter- side. In fact it was this property three selected times, t1–t3, the val of almost 1-hour duration, re- that drove the selection of the upper diagram shows the velocity connection must therefore have Cluster orbit. By contrast, the distribution of the oxygen ions, happened almost continuously, al- orbit crosses cusp magnetic field confirming that they have an ex- though not necessarily at a fixed lines at mid-altitudes when the tremely narrow spread in velocity, rate. apogee is on the night side (see with the mean velocity becoming Figure 3). smaller with time.

Polar Cusps Figure 16 shows what is observed at The schematic diagram at the bot- such a mid-altitude cusp crossing tom illustrates the interpretation. The polar cusps are weak spots in and illustrates the successive en- Ions injected with a broad energy the defence of the magnetosphere counter with the particles enter- spectrum from a narrow source at against the on-rush by the solar ing the magnetosphere through low altitudes on the dayside, all

SPATIUM 9 15 will move up the magnetic field that the convection velocities are of a northern and southern lobe lines, but end up at different loca- essentially identical, with zero containing very tenuous plasma tions, depending on the ratio of time shifts. This means that the and permeated by magnetic fields the speed at which they move observed variations in the convec- of opposite polarity that emanate along the magnetic field and the tion velocity, in particular the from the northern and southern speed with which those field lines short bursts of sunward convec- polar cap. At the centre there is a move over the polar cap, as shown tion (positive Vx) are temporal in region of hot dense plasma, the in the bottom panel. The black nature, probably caused by “dipo- plasma sheet (cf. Figure 1). The line is derived from the O+ distri- larizations" of the tail magnetic plasma sheet in turn straddles the butions, while the red symbols are field that occur as a result of a so-called neutral sheet, where the from the measurements of the magnetospheric substorm (see magnetic field reverses direction, drift of artificially injected elec- below). and which therefore is an ideal trons by EDI. Noting the entirely site for magnetic reconnection. different nature of the measure- Through magnetic reconnection ments, the agreement is simply Magnetotail in the tail, magnetic field lines that remarkable. were opened by reconnection at As a consequence of the solar wind the subsolar magnetopause (see Data shown in Figure 17 are taken flow past the Earth, magnetic field Figure 1), and were carried into the from one spacecraft only. But if lines are being stretched in the magnetotail by the action of the one compares the measurements anti-sunward direction, creating a solar wind flow, are being closed on the other spacecraft, one finds long tail. This magnetotail consists again, convect sunwards, and start

10 C4 Tango

cusp 0

0.1 C Ion Energy [KeV] 2 C C S 4 1 a 0.01 R ls u a T m a b n C1 Rumba a g 10 o cusp 0

0.1 Ion Energy [KeV]

a 0.01 b m a S C3 Samba 3 10 C Mid-altitude polar cusp cusp cusp 0

0.1 Ion Energy [KeV]

0.01 15 16 17 18 Equator Universal Time on 30th August 2001 Pole Figure 16 Crossing of the mid-altitude cusp (around 5 Re altitude) by the 4 spacecraft on 30 August 2001. The energy-time spectrograms from CIS on C1, C3 and C4 on the night show the encounter with the ions entering the magnetosphere through the cusp (from Bosqued et al., 2001).

SPATIUM 9 16 the process all over again once they + have reached the subsolar magne- O Beam 60 topause. The combination of magnetic reconnection at the dayside t t t V [km/sec] 3 2 1 magnetopause and in the magne- 30 totail sets up a global circulation of the plasma and magnetic field 0 in the magnetosphere. –30 The plasma sheet is the site of vio- –60 –30 0 30 60 lent events referred to as magne- VII [km/s] tospheric substorms, which form t1 t2 t3 one of the most intriguing phenomena in the magnetosphere. 10 3 O+ Substorms are manifestations E [eV] of the sudden release of energy stored in the geomagnetic tail. For 10 2 reasons not yet understood, the energy is not released at the same 100 V rate as it is stored, but quite sud- 80 II1 O+ V denly, with enormous conse- [km/s] 60 II2 II

V VII3 quences for the entire magnetos- 40 phere, including an intensification 20 and expansion of the aurora. Figure 0 18 shows how, at the onset of such 20 V 1 a substorm, the magnetic field 10 V 2 V 3

[km/s] 0 CIS measured by the Cluster spacecraft x –10 suddenly changes from a very V –20 stretched configuration, where the EDI –30 magnetic fled is essentially along 22:00 22:30 23:00 UT the tail axis, indicated by a small elevation angle, , to an orienta- tion characteristic of a magnetic Earth dipole, i.e., having a large . This + O Source O+ Motion dipolarisation is not simultane- ous, but propagates as a front pass- Cluster Orbit ing over the four spacecraft in t3 t succession. This allows for the first Magnetic Field Line V 2 t1 time to precisely determine the V3 VII z speed and direction of such a front. V2

V1 x Figure 17 Plasmasphere Oxygen ion observations and field line convection velocities for a polar cap pass on 4 March 2001. The top panel shows the distributions of O+ ions in velocity-space, The upper atmosphere and iono- measured by CIS-CODIF at three different times; below are the energy time spectrograms of these ions, their inferred parallel velocity, and the convection velocity, sphere are important sources of + Vx, inferred from the O measurements (solid black line) and from EDI (red sym- plasma for the magnetosphere. In bols). The interpretation of these observations is illustrated by the sketch at the bottom.

SPATIUM 9 17 the region of closed magnetic field Bx lines in the inner magnetosphere, 15 outflow of ionospheric plasma pro- duces the so-called plasmasphere, 10 SC 1 with a usually sharp outer bound- SC 2 nT 5 ary, the plasmapause. SC3 SC 4 0 Figure 19 shows a traversal of the plasmasphere, first inbound and B then outbound, as illustrated by 80 deg the insert in the lower left. In ad- 60 dition to the central plasmasphere, the figure shows some density struc- 40 tures further outward. The shape 20 of these structures looks strikingly similar in the four profiles. Replot- 0 V ting the density measurements 600 against the McIlwain L-parameter confirms that these structures are 400 fixed in space and do not vary much 200 km/s over time-scales of about one hour. (The L-parameter is a measure of 0 the equatorial distance of magnetic –200 field lines in the inner magnetos- 11:32:24 11:33 : 36 11:34 : 48 11:36: 00 11:37: 12 UT phere, originally introduced to char- Figure 18 acterize the radiation belts) Magnetic field observations from the four Cluster spacecraft during the onset of a magnetospheric sub storm, when the stretched magnetic configuration, indicated by

large values of the magnetic field along the tail axis, Bx , and small values of the field elevation angle , suddenly relaxes to a more dipolar configuration, characterized by Aurora small values of Bx and a large angle .

The only visible consequence of Black Aurora as “black aurora”. Figure 20 schemat- the interaction between the solar ically illustrates the juxtaposition wind and Earth’s magnetic field The visible aurora is associated of visible and black aurora. is the aurora. It occurs in an oval with converging electric field around both magnetic poles and is structures producing an upward Figure 21 shows, schematically, the generated by energetic particles directed electric field that acceler- observations by the EFW and that precipitate into the upper at- ates electrons towards Earth, gen- PEACE instruments from a cross- mosphere and cause light emis- erating the auroral light emission ing of a U-shaped potential struc- sion as well as ionisation. The cusp and upward field-aligned cur- ture around 04:30 UT on 14 Jan- and boundary layers on the day- rents. The opposite happens in the uary 2001. At the time of the first side and the plasma sheet on the auroral return current region: overpass, the electric field meas- night side are the sources of this there the electric field structures urements by EFW showed a weak precipitation. On the other hand, are divergent and accelerate elec- bipolar signal (signifying a sign ions escaping from the auroral oval trons upwards into space, causing reversal of the field, as required contribute to the plasma in the a void of electrons and a lack of by Figure 20), which then became magnetotail and in the magneto- auroral light emissions. This is stronger during the second and pause boundary layers. why this region is often referred to third overpasses, but which had

SPATIUM 9 18 010605 WHISPER and EFW pot 20h10–23h40 3 ] 10 3 –

Ne [cm Structure 1, L – 7 Structure 2, L – 7 South. hemisphere North. hemisphere

10 2

10 1

20:10:00 20:45:00 21:20:00 21:55:00 22:30:00 23:05:00 23:40:0 Universal Time 010605 3 ] 10 3 –

Ne [cm Structure 1 Structure 1 30 outbound inbound 10 2 z 22:00

1 22:00 10

x y 30

21:00 10 0

30 Rumba Salsa Samba Tango 10 –1 46810 12 14 Figure 19 L Plasma density deduced from the WHISPER and EFW instruments during the crossing of the plasmasphere on 5 June 2001 (top panel). The spacecraft orbit is shown on the lower left. When plotted against the McIlwain L-parameter (lower right), the density structures prove to be spatially fixed (from Decreau et al., 2001).

SPATIUM 8 19 essentially disappeared when the Cluster fourth spacecraft flew over the same region. Quite consistent with + – this, the PEACE instrument ob- EI EI EI EI served upward directed electron – beams with increasing energy, negative + positive potential – before the electrons disappeared structure potential structure as well. The results show that this (focused on – in this study) diverging electric field structure accelerated electrons e + extends to altitudes greater than downward electric 20,000 km and that it grows in B field – accelerated size, intensifies and decays over upward e electrons electric time scales of a few hundred sec- field N W E onds. S downward current upward carried by upward current ionospheric electrons Ion Outflow and Energy Input arctic circle

Electric fields directed such that Figure 20 they accelerate (negatively charged) U-shaped electric potential configuration over an aurora (left) and its counterpart electrons downward into the ion- over a dark region referred to as black aurora (right). In the structure on the left, osphere to cause auroral displays electrons are accelerated downward into the atmosphere to cause the characteristic auroral emissions, while the structure on the right accelerates electrons upward, will at the same time accelerate causing a lack of aurora emissions, hence its name. ions from the upper atmosphere/

0.7 keV 1.2 keV 2 keV ionosphere in the opposite direc-

e– e– e– tion because of their positive electric charge. Figure 22 shows a upward accelerated sequence of auroral images, to- electrons gether with Cluster ion measure- +++ ments taken during a traversal of the plasma sheet, recognized as Rumba Samba Salsa Tango the intense flux (shown in red) of energetic ions in the spectrogram +++ in the top panel. Inspection of the EI EI EI EI EI EI no structure fourth panel on the right shows observed that oxygen ions are flowing up +++ the magnetic field lines, as indicated by the predominance of large fluxes at pitch-angles of 180 de- grees, and are detected on Cluster EII EII EII whenever Cluster is magnetically t = 0 t = + 100 s t = + 200 s t = + 300 s B connected with the auroral re- Figure 21 gion. In return, the plasma sheet is Schematic representation of the evolution of the electric potential pattern that a source of electromagnetic ener- Cluster observed at around 04:30 UT on 14 January 2001 when the four spacecraft gy for the aurora. At 04:31:30 UT, sequentially flew over a black aurora, recording initially growing electric fields and electron energies, before electric field and electrons disappeared altogether (after when Cluster crosses the edge of Marklund et al., 2001). the plasma sheet, recognized as

SPATIUM 9 20 20010114 0343:14 UT LBHL C1 HIA / RPA H+, He+, O+ 2001–01–14

12 104 8.0 3 60 10 7.0 102 70 6.0 energy [eV] 101 80 H+ E: 666-852 eV 18 6 180

a 135 6.0

90 ] 1 5.0 – sr

45 1 – s 2

04:2004:2 0 4.0 – He+ E: 666-852 eV 180 0 [eflux/cm

10 Cluster a 135 6.0 log 90 0349:22 5.0 12 45

0 4.0 0.0 60 O+ E: 666-852 eV 70 180 a 80 135 6.0 90 18 6 5.0 45

0 4.0 0.0

2 0 –2 0 –4 dB dB dB GSE [nT] Cluster –6 x y z 0432:36 12 40 –1 20 dEx dEy dEz GSE [mV ] 60 0 70 –20 –40 80 04:20:0004:20:00

18 6 0.4 –2 –1 0.3 SII S [erg cm s ] 0.2 0.1 Cluster 0.0 04:20:00 04:25:00 04:30:00 04:35:00 04:40:00 04:45:00 0 ] 1

0438:44 – s 2

12 – Figure 22 60 On the left, a sequence of four auroral images from the UVI instrument on 2 NASA’s polar spacecraft, with the mapped Cluster positions shown as blue crosses. 70 10 On the right, Cluster particle and field measurements for the same time interval. The 80 top panel shows an energy-time spectrogram of ions with energies between 5 eV 18 6 and 27 keV from the CIS-HIA instrument. The next three panels show the pitch-

photons [cm angle distribution of hydrogen, helium and singly charged oxygen, respectively. The magnetic and electric field fluctuations measured by EFW and FGM are shown in the third and second panels from the bottom. The bottom panel, finally, shows the 1 10 energy flux density, S, computed from these electric field and magnetic field fluctua- Cluster tions. The key features are the appearance of upflowing oxygen ions at pitch-angles near 180º, and the large earthward directed energy flux right at the edge of the hot 0 plasma sheet, which serves to power the aurora (from Wilber et al., 2002).

SPATIUM 9 21 the drop in flux of energetic ions in the spectrogram in the top panel, a high flow of earthward directed electromagnetic energy, SC1 S, is observed. This energy flux pro- SC2 vides a way to power the aurora.

Auroral Kilometric Radiation 12 = 2 – 1 Since the mid 1960s it is known that the Earth is a powerful radio source in the frequency range from 50 to 500 kHz. Because electromagnetic radiation in this 2 1 frequency range has wavelengths in the kilometer range, and because of the close association with the aurora, this radiation is now called auroral kilometric radiation, abbreviated AKR. One of the primary objectives of the WBD instrument on Cluster is to use the unique four-spacecraft configuration to conduct very-long- baseline interferometry (VLBI) investigations of AKR, because they have the potential of locating the AKR source better than can be achieved with any other technique. Figure 23 illustrates the differential delay location determination method, simplified to only two spacecraft.

Figure 23 Illustration of the scheme for locating the source of the Auroral Kilomet- ric Radiation (AKR) by triangulation, based on differential delays between the radiation recorded by pairs of spacecraft. With only two spacecraft, the source is constrained to locations in a ring at some distance from the Earth’s centre (courtesy WBD team).

SPATIUM 9 22 Future

Originally intended for a two- year lifetime, the Cluster mission has now been extended until late 2005. Although at the time of this writing, the Cluster teams are still busy with final calibrations and validation of the analysis methods, it is clear from even the cursory glance at the data provided in this lecture, that at the end we will have a highly improved understanding of the various physical processes and phenomena operat- ing in collisionless plasmas. How- ever, it is also clear that in view of the many scale sizes that are involved, four spacecraft are simply the minimum. Ideas to fly swarms of dozens of spacecraft are present- ly being debated. Such a fleet would allow coverage of multiple spatial scales simultaneously. Cou- pled with global measurements, this would improve our understanding of the complex linkage between the various processes and regions, an important goal in its own night, but also a key in understanding space weather.

Acknowledgment

The author is indebted to his col- leagues in the Cluster community for their help in putting this lecture together.

SPATIUM 9 23 The author

of G. Haerendel and got his PhD At the time of the failure of the from the Technische Universität first Cluster attempt, G. Pasch- München in 1971. mann had already started his in- volvement with ISSI, initially as At MPE G. Paschmann worked the leader of a team that wrote a on instrument development and book on Analysis Methods for on the analysis of data from ESA’s Multi-Spacecraft Data that has be- Heos2 spacecraft on the polar re- come the reference for much of gions of Earth’s magnetosphere. the Cluster data analysis. In July He then joined forces with 1999 he became a Director at ISSI S. J. Bame and his group at the in Bern, where he now spends Los Alamos Scientific Laboratory 50% of the time, and the other to build plasma detectors for 50% still at MPE in Garching. the NASA/ESA ISEE mission launched in 1977. The discovery of high-speed plasma jets at the magnetopause as the “smoking Götz Paschmann was born in gun" evidence of magnetic recon- 1939 in Schwelm/Germany. He nection is one of the highlights studied physics, first at the Uni- of this work. Later, G. Paschmann versität Tübingen, then at the was involved in further magne- Technische Universität München, tosphere research missions like where he got his Diploma in the ill-fated Firewheel project 1965. Fascinated by the emerging and the highly successful AMPTE field of space science, he asked mission. Prof. Lüst, then Director of the newly founded Max-Planck-Insti- In 1982, G. Paschmann was among tut für extraterrestrische Physik the group of European scientists (MPE) in Garching for a PhD the- that proposed to ESA a mission sis opportunity. The theme was called Cluster, intended to ex- related to Earth’s Van Allen belts plore the magnetosphere with a and depended on data to be ob- fleet of four spacecraft flying in tained from a sounding rocket close formation. The article in flight into these radiation belts. this issue of Spatium describes When this failed, he was sent to Cluster’s long history, including the Lockheed Palo Alto Research its failure in 1996 and eventual Laboratory, where he worked in recovery, ending in the successful R. G Johnson’s group on auroral launch in the summer of 2000. particle precipitation. Back at G. Paschmann is the Principal In- MPE, he then wrote his thesis on vestigator for the Electron Drift this subject, under the supervision Instrument on Cluster .