Response of the magnetosheath-cusp region to a coronal mass ejection

Transcription

1 Click Here for Full Article JOURNAL OF GEOPHYSICAL RESEARCH, VOL. 112,, doi: /2006ja012122, 2007 Response of the magnetosheath-cusp region to a coronal mass ejection N. Balan, 1 H. Alleyne, 1 S. Walker, 1 H. Reme, 2 E. Lucek, 3 N. Cornilleau-Wehrlin, 4 A. N. Fazakerley, 5 S.-R. Zhang, 6 and A. P. van Eyken 7 Received 13 October 2006; revised 18 July 2007; accepted 7 September 2007; published 19 December [1] Cluster made an unusual magnetosheath-exterior cusp crossing during the first 2.5 hours of a coronal mass ejection (CME) that flowed past Earth for about 7 hours on 24 October During the first 2.5 hours ( UT) the solar wind dynamic pressure remained high and stable though the CME had a discontinuity after 40 min (1605 UT), when the azimuthal flow turned dawnward up to 100 km s 1 and IMF B y and B z changed from highly negative to positive up to 25 nt. The responses of the magnetosheath-cusp region during the unusual event are presented using Cluster and ground-based (EISCAT VHF radar; 69.6 N, 19.2 E) observations. The unusual Cluster crossing (compared to the usual midaltitude cusp crossing at this time of the year) occurred owing to a large compression of the magnetosphere. Cluster, which was in the southern magnetospheric lobe, suddenly found itself in the magnetosheath at the arrival of the CME at 1524:45 UT. Cluster then crossed through the compressed magnetosheath (highly compressed after the discontinuity in the CME) for about 1.5 hours ( UT), magnetopause with strong signatures of lobe reconnection (1655 UT), stagnant but compressed exterior cusp for about an hour ( UT), and then entered the dayside magnetosphere. The observations also show strong signatures of magnetosphere-ionosphere coupling through a late afternoon (17 MLT) cusp during the first 40 min ( UT) of the event when IMF B z remained negative. Strong magnetic waves are also generated in the magnetosheath-cusp region. Citation: Balan, N., H. Alleyne, S. Walker, H. Reme, E. Lucek, N. Cornilleau-Wehrlin, A. N. Fazakerley, S.-R. Zhang, and A. P. van Eyken (2007), Response of the magnetosheath-cusp region to a coronal mass ejection, J. Geophys. Res., 112,, doi: /2006ja Introduction [2] The magnetospheric cusps, originally proposed as magnetic null points [Chapman and Ferraro, 1930], represent topological boundaries separating the dayside field lines from those extending into the nightside tail lobes. The topological boundaries (or cusps) are observed as longitudinally extended funnel shaped regions in the highlatitude dayside magnetosphere, one in each hemisphere, through which solar wind mass and energy can directly enter the magnetosphere and find their way down to the ionosphere without crossing the geomagnetic field lines 1 Automatic Control and Systems Engineering, University of Sheffield, Sheffield, UK. 2 Centre d Etude Spatiale des Rayonnements, Toulouse, France. 3 Blackett Laboratory, Imperial College, London, UK. 4 Centre d Etudes des Environnements Terrestre et Planétaires du Centre National de la Recherche Scientifique, Vélizy, France. 5 Mullard Space Science Laboratory, University College London, Surrey, UK. 6 Haystack Observatory, Massachusetts Institute of Technology, Westford, Massachusetts, USA. 7 EISCAT Scientific Association, Kiruna, Sweden. Copyright 2007 by the American Geophysical Union /07/2006JA012122$09.00 [e.g., Haerendel et al., 1978]. The cusps are also known as favorable regions to observe the signatures of magnetic reconnection [e.g., Dungey, 1961; Lockwood and Smith, 1994] and associated flux transfer events (FTEs) [Russell and Elphic, 1979]. Using energetic electrons as field line topology tracers, Zong et al. [2005] showed that magnetospheric cusps are regions where energetic particles are trapped. The region just equatorward of the cusp opening through which magnetosheath plasma enters the closed field lines through turbulence and diffusion is known as the entry layer (EL) and the region poleward of the cusp opening through which part of the cusp plasma escapes to the nightside tail lobe is known as the plasma mantle (PM) [Haerendel et al., 1978]. [3] The cusps, having a latitude width of about 2.5 degrees, are centered at around 80 invariant latitude at magnetic noon at a geocentric distance of 8 R e. However, they are found to extend down to about 70 invariant latitudes, 07 to 17 hours MLTs (magnetic local times), and 6 to 9 R e geocentric distances [Stubbs et al., 2004]. Between the cusp and magnetosheath there is an intermediate region called exterior cusp. The exterior cusp was predicted by the gas dynamic model of Spreiter and Summers [1967]. Lavraud et al. [2002] identified it as a pocket of hot and stagnant 1of10

2 plasma of magnetosheath origin. Its external boundary with the magnetosheath has sharp bulk velocity gradient and it appears stagnant under northward IMF conditions while more convective (antisunward) during southward IMF conditions [Cargill et al., 2004; Lavraud et al., 2004a, 2004b, 2005]. The exterior cusp is generally associated with depressed magnetic field and lies predominantly within MLTs and extends to both earlier and later MLTs [e.g., Russell, 2000; Zhang et al., 2005]. [4] The cusps are also known to move with the changes in solar wind dynamic pressure and IMF (interplanetary magnetic field) [Newell et al., 1989; Woch and Lundin, 1992]. The magnetopause and cusp move back and forth during periods of high and low solar wind pressure, and dawnward/duskward during periods of dawnward/duskward solar wind velocity component [Lundin et al., 2001]. Since the exterior cusps are regions where plasma pressure and magnetic field pressure balance [Cargill et al., 2004], a small change in the solar wind dynamic pressure can lead to a large effect in the cusp [Balan et al., 2006]. The occurrence of double cusp during periods of large IMF B y and small IMF B z [Wing et al., 2001], and cusp regions that look like multiple cusps have also been reported [Newell et al., 1989; Zong et al., 2004]; multiple cusp regions are due to magnetospheric (or cusp) oscillations (with respect to observing spacecraft) driven by changes in solar wind pressure, azimuthal flow and IMF B y [e.g., Zong et al., 2004]. [5] Extreme solar events such as Coronal Mass Ejections (CMEs) are known to drive extraordinary responses of the terrestrial environment. A number of extreme CMEs occurred in October November Gopalswamy et al. [2005a, 2005b] have presented a review of the 61 papers describing these CME events and their geoeffectiveness in the Journal of Geophysical Research and of another 13 papers in Geophysical Research Letters. The papers describe the solar sources, propagation and energetic particle distribution of the CMEs. The geoeffectiveness of the CMEs including their impacts on the magnetosphere, and effects on the ionosphere, thermosphere and atmosphere are also described in a number of papers [e.g., Terasawa et al., 2005; Mannucci et al., 2005; Rosenqvist et al., 2005; Borovsky and Steinberg, 2006; Lavraud et al., 2006]. An extremely large solar eruption on 28 October caused an intense geomagnetic storm, and a second solar eruption on 29 October resulted in a reintensification of the geomagnetic storm about a day later. Within the second storm, some of the strongest substorms in the history of magnetic recordings occurred in northern Scandinavia, and geomagnetically induced currents in ground conductors caused power failures for 20 to 50 min in southern Sweden [Rosenqvist et al., 2005]. [6] Though an extreme solar wind (following a CME) flowed past Earth on 24 October 2003, its geoeffectiveness has not been studied. The present paper reports the response of the magnetosheath-cusp region to the extreme solar wind using Cluster and ground-based observations. Cluster made an unusual magnetosheath-exterior cusp crossing after the arrival of the CME. Such crossings are rare, and are important to understand the dynamics of the magnetosphere and magnetosphere-ionosphere coupling during extreme solar wind conditions. The instruments and data are briefly described in section 2. Section 3 describes the CME. The observations are presented and discussed in section 4, with a summary in section Instruments and Data [7] The Cluster mission [Escoubet et al., 2001] provides a unique opportunity to investigate the three-dimensional (3-D) structure, dynamics and populating mechanisms of the cusp and surrounding regions [e.g., Owen et al., 2001; Lavraud et al., 2002; Cargill et al., 2004]. In the present study, Cluster data from a fluxgate magnetometer (FGM) [Balogh et al., 2001], Composition Ion Spectrometer (CIS) [Reme et al., 2001], Plasma Electron and Current Experiment (PEACE) [Johnston et al., 1997], and Spatiotemporal Analysis of Field Fluctuation (STAFF) [Cornilleau-Wehrlin et al., 2003] instruments are used. The data presented are in the GSE coordinate system and have a time resolution of four seconds. Ground-based data are obtained from the EISCAT VHF radar at Tromso (69.67 N, E) [van Eyken et al., 1984]. The Cluster data correspond to the southern magnetosphere while ground-based data correspond to the northern ionosphere. The solar wind and IMF conditions are assessed using ACE (Advanced Composition Explorer) satellite, which was at the L1 point at a distance of 234 R e away from Earth [Skoug et al., 2004]. The study also uses the magnetosphere model of Tsyganenko and Stern [1996], magnetopause model of Shue et al. [1998], and orbit visualization tool of the OVT Team (see [8] The Cluster period of interest is 1524: :00 UT, which begins with the arrival of the CME at Cluster at 1524:45.25 UT (from high time resolution data). The CME produced a moderate geomagnetic storm whose onset and early main phase cover the period of the study, with Dst around 40 nt and Kp = 7. At the beginning of the period (1524:45 UT), Cluster was at approximately 7.0 R e away from Earth, and its location during the period of interest (1524: :00 UT) varied from X = 2.11 R e,y=2.23r e and Z = 7.83 R e to X = 3.46 R e,y= 1.26 R e and Z = 4.31 R e. The spacecraft configuration was an elongated cigar-like tetrahedron (Figure 1) and spacing varied from about 170 to 1000 km during the period of interest. Figure 1 shows the Cluster orbit during UT projected on the X-Z plane that contains part of a model magnetosphere [Tsyganenko and Stern, 1996] (see also at 1524:45 UT, obtained with appropriate input parameters (Dst = 40 nt and Kp = 7 ; solar wind pressure = 30 npa and velocity = 600 km s 1 ;IMFB x = 5 nt,b y = 2 nt and B z = 18 nt). As shown by Figure 1, Cluster (1524 UT) appears outside the magnetopause by about 0.7 R e just after the impact of the CME (1524:45 UT), and Cluster was inside the magnetopause (southern magnetospheric lobe) just before the impact (figure not shown). The relative positions of Cluster just before and after the impact of the CME suggest a sudden magnetopause compression at the impact. Such a compression by about 4 R e (figure not shown) is also predicted by the magnetopause model of Shue et al. [1998]. 3. Solar Wind and IMF [9] As shown by Figure 2, the ACE spacecraft encountered the CME cloud at 1448:53 UT [Skoug et al., 2004]. 2of10

3 4. Observations and Discussion 4.1. Cluster Observations Identification of Regions [11] Figure 3 shows the ion and electron fluxes measured by Cluster. As shown, Cluster was in the southern magnetospheric lobe before the arrival of the CME cloud. Starting at the arrival of the cloud at 1524:45 UT, there is a sudden appearance of magnetosheath-like fluxes, which indicates that the CME cloud has compressed and deformed the magnetosphere in such a way that Cluster suddenly found itself in the magnetosheath. After observing a compressed but nearly stable magnetosheath for about 40 min, Cluster encountered a sharp change in the fluxes at around 1605 UT, which appears like a boundary. The detailed examination of the data (below) shows that Cluster was still in the magnetosheath and the change corresponds to the discontinuity in the CME. Cluster crossed the magnetosheathcusp boundary at around 1655 UT, entered the exterior cusp at around 1700 UT, and came out of the cusp and Figure 1. Cluster orbit during UT on 24 October 2003 projected on the X-Z plane containing part of a model (T96) magnetosphere at 1524:45.25 UT, obtained with appropriate input conditions (see section 2). Cluster configuration is also shown. The solar wind speed suddenly increased from about 430 to 600 km s 1 (Figure 2b) and proton density increased from about 5 to 25 cm 3 (Figure 2c), which caused the pressure to raise from about 2 to 30 npa (Figure 2a). The proton temperature increased from about 0.1 to 0.5 MK (Figure 2d). The high-speed, high-density, high-pressure, hot solar wind flowed past Earth for about seven hours. A comparison of the ACE (Figure 2) and Cluster data (section 4.1) shows that the solar wind took about 35.9 min to travel from ACE (1448:53 UT) to reach Cluster at 1524:45 UT. The ACE period between the vertical lines (1448: :08 UT, Figure 2) corresponds to the Cluster period of interest (1524: :00 UT). [10] During the first 40 min of this period ( UT), IMF B x (Figure 2e) was slightly antisunward, IMF B y (Figure 2f) was dawnward up to 20 nt, and IMF B z (Figure 2g) was southward up to 20 nt. At around 1527 UT, the CME showed a large change (or discontinuity); IMF B y and B z changed from negative to positive (up to 25 nt) and IMF B x became further antisunward; solar wind azimuthal flow also became strongly dawnward up to 100 km s 1. These changes that happened when the solar wind speed and pressure remained high and nearly stable had strong effects on the magnetosheath-cusp region as discussed below. Magnetic reconnection could take place in the dayside magnetopause before the arrival of the discontinuity owing to negative IMF B z [Lockwood and Smith, 1994; Onsager et al., 2001], the effect of which, as shown below, was detected by EISCAT radar. Figure 2. (a g) Solar wind pressure, speed, density, and temperature, and IMF components B x,b y and B z (in GSM) during UT on 24 October 2003 measured by ACE at L1 point at 234 Re away from Earth. An extreme solar wind, following a coronal mass ejection, arrived at 1448:53 UT. The Advanced Composition Explorer (ACE) time between the vertical lines (1448: :08 UT) corresponds to the Cluster period (1524: :00 UT) of interest. 3of10

4 Figure 3. Ion and electron fluxes measured by Composition Ion Spectrometer (CIS) and Plasma Electron and Current Experiment (PEACE) instruments on Cluster SC4 during UT on 24 October Cluster, which was in the southern magnetospheric lobe (M lobe) before the arrival of the coronal mass ejection (CME), suddenly found itself in the magnetosheath (MSH) at the impact of the CME (1524:45 UT) and crossed through the MSH ( UT) and exterior cusp ( ), and entered the dayside magnetosphere (D-MSP). The change in the MSH at around 1605 UT is due to the discontinuity in the CME (see section 4.1.1). entered the dayside magnetosphere (D-MSP) at around 1802 UT. [12] The different regions crossed by Cluster are also identified in Figure 4 that compares the total magnetic fields and their Y components measured by Cluster (Figure 4a) and ACE (Figure 4b; shifted by 35.9 min); the clock angles of the two magnetic fields are also compared (Figure 4c). The clock angle is the angle between the projection of the magnetic field on the YZ plane and the Z axis of the GSE coordinate (or arctan(b y /B z ). The clock angles of the IMF and dayside magnetosheath field correlate well if there is only a bow shock in between [Spreiter and Stahara, 1980]. As shown by Figure 4, the two magnetic fields and their clock angles undergo similar variations during UT, confirming that this region including the effects of the discontinuity in the CME is magnetosheath. The discontinuity in IMF also appears as a dip in the total field at 1605 UT (Figure 4b). By interacting with the bow shock, this discontinuity could have been amplified, and was observed by Cluster in the magnetosheath as a deep and long transition (Figure 4a) owing to the slow magnetosheath flow. After about 1655 UT Cluster enters the exterior cusp region where the magnetic fields and clock angles differ. However, there are short durations in the cusp where the two clock angles agree. This is also in line with the finding of Zhang et al. [2005] that the clock angle could be used as a criterion for only 73% of the cusp events they studied Effects on Magnetosheath-Cusp Region [13] The effects of the CME on the magnetosheath-cusp region are shown in Figure 5 which compares the Cluster FGM and CIS data (Figures 5c 5f) with the solar wind velocity components (Vy and Vz; Figure 5a) and IMF components (Figure 5b); solar wind and IMF are shifted by 35.9 min. The solar wind velocity Vx and dynamic pressure are not shown as they remain almost stable during the period of interest (Figure 2). The CIS data shown are from the Hot Ion Analyser (HIA) detector, which measures the 3-D distribution function of the ions (H +,He +,He ++ and O + ) in the energy range from 5 ev/e to 32 kev/e without mass selection but with a good angular resolution [Reme et al., 2001]. [14] The sudden shift of the magnetosheath to the Cluster position at the impact of the CME cloud is seen in the Cluster data starting from 1524:45 UT. The magnetic field (Figure 5c) becomes similar to IMF (Figure 5b); ion velocity (Figure 5d) turns strongly antisunward and southward (poleward) up to 200 km s 1 ; ion density (Figure 5e) increases nearly 400 times from about 0.5 to 200 cm 3 ; and ion temperatures (Figure 5f) are nearly isothermal and increases by 10 times from about 0.5 to 5 MK. The ion velocity, density and temperature data (Figures 5d 5f) during UT confirm that Cluster was crossing through the compressed magnetosheath during this period; the ACE and Cluster magnetic fields (Figures 5b and 5c) are also similar during this period. A comparison of the ACE and Cluster data (Figure 5) also confirm that the sharp change in the Cluster data at 1605 UT that appeared like a boundary is caused by the discontinuity in the CME, with the solar wind azimuthal flow (Figure 5a) becoming highly dawnward up to 100 km s 1 and IMF B y changing from 4of10

5 suggests the exit of Cluster from the cusp. Cluster then entered the dayside magnetosphere. The magnetic field and plasma data from all four spacecraft (not shown) are found to agree well, which suggest that the plasma characteristics are consistent within the Cluster spacing (up to 1000 km). [16] Figure 6 shows the magnetic wave intensities parallel to and perpendicular to the magnetic field (B o ) measured by STAFF. As shown, strong magnetic waves are generated in all frequency ranges at the arrival of the extreme solar wind (1525:45 UT). The effect is similar to but much stronger than that usually observed at the magnetopause boundary under normal solar wind conditions [e.g., Cargill et al., 2005]. In the present case, strong waves are generated as the magnetopause is suddenly compressed. The waves also remain powerful throughout the period of interest because the field is compressed and undergoes fluctuations though the wave power is larger in the magnetosheath ( UT) than in the exterior cusp ( UT). The wave power, in general, also maximizes at the magnetic field reversals (1605 and 1620 UT), magnetosheath-cusp boundary (1655 UT) and cusp magnetosphere boundary (1802 UT) as noticed in case studies and statistical studies [e.g., Cargill et al., 2005] Radar Observations [17] The EISCAT VHF radar data are shown in Figure 7. The panels from top show the line of sight ion velocity, ion Figure 4. Comparison of the total magnetic field (dotted curves) and its Y component (solid curves) measured by (a) fluxgate magnetometer on Cluster SC3 with those measured by (b) Advanced Composition Explorer (ACE). (c) Comparison of the clock angles of the two magnetic fields; ACE data is shifted by 35.9 min. The comparison identifies the different regions crossed by Cluster during UT and illustrates the effects on the magnetosheath of the discontinuity in the coronal mass ejection (1605 UT Cluster time). highly negative to highly positive up to 25 nt (Figure 5b). However, the ion densities in the magnetosheath are higher after the discontinuity than before though the solar wind dynamic pressure remained almost stable (Figure 2); that is discussed below. [15] Cluster passed through the magnetosheath-cusp boundary (or magnetopause) at around 1655 UT with a strong velocity reversal. The velocity reversal agrees with the possible Earthward injection of ions owing to the lobe reconnection that could happen owing to northward IMF B z [Lockwood and Smith, 1994; Kessel et al., 1996]. At around 1700 UT Cluster entered the exterior cusp in which the plasma becomes almost stagnant; the ion densities in the cusp are also lower and ion temperatures are higher than those in the magnetosheath as during normal solar wind [Lavraud et al., 2004a, 2004b; Zhang et al., 2005]. However, the ion densities in the magnetosheath-cusp region during the CME passage are much higher than those during normal solar wind, which is due to compression. Though the CME lasted for about seven hours (Figure 2), the ion densities reduced after about 1802 UT, which Figure 5. Comparison of (a) solar wind velocity components Vy and Vz and (b) IMF components, shifted by 35.9 min, with (c g) Hot Ion Analyser/Composition Ion Spectrometer (HIA/CIS) data from Cluster SC3. The comparison identifies the different regions crossed by Cluster and illustrates the response of the magnetosheathcusp region to the coronal mass ejection, including the effect of the discontinuity at 1605 UT Cluster time. 5of10

6 Figure 6. Intensity of the magnetic waves (top) parallel to and (bottom) perpendicular to the magnetic field (B 0 ) measured by Spatiotemporal Analysis of Field Fluctuation (STAFF) on Cluster SC3 during UT. The wave power in all frequencies suddenly increased when Cluster, which was in the southern magnetospheric lobe, suddenly found itself in the magnetosheath (MSH) at the impact of the coronal mass ejection at 1524:45 UT. The wave power remains high in the MSH ( UT), slightly decreases in the exterior cusp ( UT), and generally maximizes at the magnetic field reversals and magnetospheric boundaries (see section 4.1.2). temperature, electron density and electron temperature at altitudes from 100 to 500 km. The radar was operated with an elevation of 30 so that the km altitude range covers latitudes of N, and the usual ionospheric peak (300 km) falls near 74 N. The vertical lines (1524:45 and 1802:00 UT) correspond to the beginning and end of the effects of the CME observed by Cluster. [18] As shown by Figure 7, the effects of the CME on the ionosphere follow the start of the effects on the magnetosphere (left vertical line), but with a small time delay for the separation between the observation points (Cluster and EISCAT). The ionospheric effects begin with a large downward turning of the line of sight ion velocity (Figure 7a) in the entire field of view, and strong ion heating (Figure 7b) centered around the ionospheric peak; the velocity exceeds 600 m s 1 and temperature rises above 3000 K. The ionospheric density (Figure 7c) increases in the lower ionosphere down to E region altitudes for a short period centered at 1545 UT, indicating the precipitation of highenergy particles, possibly through the cusp as discussed below. The ionospheric effects (Figure 7) remain strong while IMF B z remains negative (Figure 5). When IMF B z turned positive (at the CME discontinuity), the ion flow and ion heating become weak (Figures 5a and 5b) though electron density and electron temperature (Figures 5c and 5d) continue to show the after effects of the earlier strong ion heating. The high ion temperature combined with the production of ionization going on during the sunlit hours seems to expand the ionosphere which also becomes denser (Figure 5c) after the ion heating. By about 1640 UT, the ionospheric peak rises to the high-altitude, high-latitude range of the radar (Figure 5c), after which there is a short period of strong upward turning of the ion velocity at around 1655 UT. The electron temperature (Figure 7d), which was less than the ion temperature during the ion heating period, rises above the ion temperature through ionelectron energy transfer, and nearly follows the usual daytime anticorrelation with electron density (Figure 7c) [Balan et al., 1997]. [19] The radar observations indicate the possibility of the solar wind plasma finding its way down to the ionosphere through the afternoon cusp that descends to low latitudes during the early stages of the CME event (1524: UT) when IMF B z remained negative. The red density spot observed at the bottom of the ionosphere (Figure 7c; around 1545 UT) seems to be the ionospheric footprint of the magnetospheric cusp as shown by the red square (66.1 N magnetic latitude, 1702 MLT) in Figure 8, which shows the expected footprint (blue stars) of the openclosed field line boundary (or cusp) obtained using the T96 magnetospheric model [Tsyganenko and Stern, 1996] for the conditions of the red density spot in Figure 7 (1545 UT, solar wind pressure = 30 npa, velocity = 600 km s 1 ;IMFB z = 20 nt, B x = 3 nt,b y = 18 nt; Dst = 40 nt and Kp = 7 ). Figure 8 also illustrates that the cusp footprint (green stars), which was at around 78 magnetic latitude and symmetric with respect to noon before the arrival of the extreme solar wind (1520 UT, solar wind pressure = 3 npa, velocity = 440 km s 1 ;IMFB z = 5 nt,b x = 3 nt,b y = 2 nt; Dst = 16 nt and Kp = 1 + ), has descended to low latitudes and been deformed in latitude and longitude after the arrival of the wind (blue stars). [20] The occurrence of the cusp at low latitudes (66 N magnetic latitude) in the late afternoon hours (17 MLT) is a rare effect of the extreme solar wind. Earlier observations during normal solar wind conditions showed the cusp occurring at low latitudes in the afternoon hours though not as late as 17 MLT. Opgenoorth et al. [2001] observed a transient excursion of the cusp-like features from dawn to dusk by over five hours in MLT under undisturbed solar wind conditions. In a statistical survey, Stubbs et al. [2004] found the cusp-like regions extending from near the pole to below 70 invariant latitude and centered around magnetic noon, with small number of occurrence after 15 MLT. Zhang et al. [2005] also found the cusp-like region extending to earlier and later MLTs from their preferential time of MLT Discussion [21] As shown by the observations, the response of the magnetosheath-cusp region to the CME is controlled both by the strong but stable solar wind dynamic pressure (25 npa) (Figure 2) and the discontinuity in the CME. 6of10

7 Figure 7. EISCAT VHF radar data averaged at 5-min time intervals and 10-km-altitude intervals (at altitudes <170 km) during UT. The panels, from top to bottom, show the line of sight ion velocity, ion temperature, electron density, and electron temperature at altitudes from 100 to 500 km. The radar elevation was 30, so that the 100- to 500-km altitude range corresponds to latitudes of N. The vertical lines correspond to the time of start and end of the responses observed by Cluster. Strong magnetosphere-ionosphere coupling is observed during the first 40 min of the coronal mass ejection ( UT), when IMF B z remained negative. The discontinuity (1605 UT) was such that the solar wind azimuthal flow turned strongly dawnward up to 100 km s 1 and IMF B y and B z changed from highly negative to highly positive up to 25 nt (Figure 5). The discontinuity also appeared as a drop in the total IMF at ACE, which was observed by Cluster as a deep and long transition in the magnetosheath field (Figure 4) owing to possible amplification at the bow shock and slow magnetosheath flow. [22] Cluster made the uncommon magnetosheath-exterior cusp crossing owing to the large compression of the magnetosphere by the CME. At the impact of the CME, the magnetosphere is compressed and deformed in such a way that Cluster, which was in the southern magnetospheric lobe, suddenly found itself in the magnetosheath at 1525 UT (Figures 1 and 3). While crossing through the compressed magnetosheath for about 40 min, Cluster encountered a sharp change (1605 UT) that appeared like a boundary where the total magnetic field dropped, magnetic field components B y and B z reversed, and ion velocities and densities became low (Figure 5). Detailed investigation of the data showed that the change was not due to Cluster crossing any boundary but was due to the discontinuity in the CME. After the discontinuity, the ion velocities, densities and temperatures, and the clock angles of the magnetic field became similar to those before the discontinuity, confirming that Cluster was still in the magnetosheath. [23] After the discontinuity, the ion densities increased (Figure 5) despite the solar wind dynamic pressure (density and speed) remaining stable and IMF B z northward. There are a number of possible reasons for the ion density increase. One possibility is the location of the spacecraft relative to the orientation of the shock normal with respect to the upstream field direction. Figure 9 shows the model shock position (curves) calculated following Dmitriev et al. [2003] at times before (1530 UT, blue) and after (1630 UT, red) the discontinuity together with the shock normal directions directly upstream of the spacecraft (stars); the green lines show the upstream field directions. The black curve is the shock position before the arrival of the CME for comparison. As shown in Figure 9, the shock becomes more quasi-perpendicular after the discontinuity, Q BN changes from 50 before the discontinuity to 70 afterward. This change does not seem to contribute to the high ion density in the magnetosheath after the discontinuity because the density compression ratio downstream of the shock decreases as the shock normal angle increases Kennel et al. [1985]. Further checks on the compression ratio of the shock as function of Q BN cannot be performed as the observations on either side of the shock are not available. [24] Another possibility for the high ion density after the discontinuity (Figure 5) could be due to the ion velocity turning more toward the magnetopause after the discontinuity than before (Figure 10) possibly as a result of the solar wind velocity component Vy turning strongly dawnward [Zong et al., 2004] and IMF B y turning strongly duskward [Newell et al., 1989]. When the ion velocity turns more 7of10

8 magnetic reconnection could take place on the dayside magnetopause. As the reconnected field lines convect toward nightside the associated plasma enters the cusps (on both hemispheres) [e.g., Lockwood and Smith, 1994; Zong et al., 2005] and reaches the ionosphere along the cusp field lines as discussed also by Heikkila and Winningham [1971]. After IMF B z turned positive (1605 UT), the ionospheric effects over the radar became weak as expected from weak ion injection owing to lobe reconnection and poleward movement of the Cusp. By the time Cluster entered the exterior cusp (1700 UT), the ionospheric cusp might have also disappeared owing to evening hours (1815 MLT). Figure 8. Footprint of the open-closed field line boundary (inset) in GSM coordinates at 1520 UT (green stars) and 1545 UT (blue stars) calculated for the solar wind, IMF, and geomagnetic conditions at these times (see section 4.2) using the magnetospheric model of Tsyganenko and Stern [1996]. The circles are north magnetic latitudes at 4 intervals, and the lines are longitudes at 8 intervals. The red square corresponds to the magnetic location of the red density spot in Figure 7c. toward the magnetopause, the solar wind becomes more effective in compressing the plasma against the magnetopause and hence ion density increases. The ion density increase is also evident at the impact of the CME (1525 UT), when the velocity vector was most strongly directed toward the magnetopause (Figure 10) while the density attained its maximum value (Figure 5). The total magnetic field also decreased after discontinuity (Figure 4a; dotted curve), which can contribute to the high ion density after the discontinuity because the low magnetic field restricts the divergence of plasma. The ion temperature does not seem to have any significant effect on the ion density changes since it remains more or less stable before and after the discontinuity (Figure 5f). [25] At the magnetopause cusp boundary (1655 UT), Cluster observed a large ion velocity reversal (Figures 5 and 10), which agrees with the effect of possible lobe reconnection taking place owing to positive IMF B z [e.g., Lockwood and Smith, 1994; Onsager et al., 2001]. In the cusp region ( UT), the ions remain almost stagnant (Figures 5 and 10) as during normal solar wind with northward IMF B z as reported by Lavraud et al. [2004a, 2004b] and Zhang et al. [2005]. [26] The EISCAT radar detected strong magnetosphereionosphere coupling during the CME event. During the first 40 min ( UT) of the event, when IMF B z remained negative (Cluster was in the magnetosheath), the radar detected large downward line of sight ion velocity, strong ion heating and high ion density at the base of the ionosphere (Figure 7). That indicated a late afternoon (17 MLT) cusp (Figure 8) descending to low latitudes (66 N magnetic latitude) during this period ( UT, MLT) of negative IMF B z when 5. Summary [27] The response of the magnetosheath-cusp region to the CME that flowed past Earth for about seven hours on 24 October 2003 has been presented using Cluster (FGM, CIS, PEACE, and STAFF) and ground-based (EISCAT VHF radar; 69.6 N, 19.2 E) observations. Cluster made an unusual magnetosheath-exterior cusp crossing during the first 2.5 hours ( UT) of the event when the solar wind dynamic pressure remained high and stable (25 npa) though the CME had a discontinuity after 40 min (1605 UT, Cluster time) when the azimuthal flow turned dawnward up to 100 km s 1 and IMF B y and B z changed from highly negative to positive up to 25 nt. The observations show the cause of the unusual crossing of Cluster (compared to the usual mid altitude cusp crossing at this time of the year), the effects in the magnetosheath of the discontinuity Figure 9. Model shock position (curves) calculated at times before (1530 UT; blue) and after (1630 UT; red) the discontinuity in the coronal mass ejection (CME) together with the shock normal directions directly upstream of the spacecraft (stars); the green lines show the upstream field directions. The black curve is the shock position before the arrival of the CME for comparison. 8of10

9 the magnetopause (1655 UT) suggesting lobe reconnection owing to positive IMF B z. The exterior cusp ( UT) was stagnant and hot as normal though the density was high owing to solar wind compression. Strong magnetic waves are generated at the impact of the CME, and the wave power, in general, maximized at field reversals and magnetosphere boundaries. [29] Acknowledgments. We thank P. I. N. Ness and P. I. D. McComas for the solar wind and IMF data from Advanced Composition Explorer (ACE). We also acknowledge the use of the magnetosphere model of N. A. Tsyganenko and D. P. Stern (1996), magnetopause model of J. H. Shue et al. (1998), and orbit visualization technique of the OVT Team (2002). The work at Sheffield is supported by PPARC (UK) grant PPA/G/S/1999/ [30] Zuyin Pu thanks the reviewers for their assistance in evaluating this paper. Figure 10. The total ion velocity from Hot Ion Analyser/ Composition Ion Spectrometer (HIA/CIS) on Cluster SC3 during the magnetosheath ( UT) and exterior cusp ( UT) crossings of Cluster plotted at 5-min intervals on Cluster orbit (1419 UT). The turning of the velocity toward the magnetopause during UT caused high ion densities in the magnetosheath (Figure 5). in the CME, signatures of magnetic reconnection due both negative and positive IMF B z, magnetosphere-ionosphere coupling through a late afternoon cusp, and generation of strong magnetic waves. [28] The CME had compressed and deformed the magnetosphere in such a way that Cluster, which was in the southern magnetospheric lobe at about 7.0 R e away from Earth, suddenly found itself in the magnetosheath at the impact of the CME (1524:45 UT). During the first 40 min ( UT), when Cluster was crossing the compressed magnetosheath, the EISCAT radar observed large downward turning of the line of sight ion velocity ( 600 m s 1 ), high ion temperature (3000 K) and high ion density down to E region altitudes. That suggests signatures of magnetic reconnection in the dayside magnetopause owing to negative IMF B z, and strong magnetosphere-ionosphere coupling through a late afternoon (17 MLT) cusp that descended to low latitudes. Though the solar wind pressure remained stable, the magnetosheath showed a change at the discontinuity (1605 UT) in the CME, after which the ion density in the sheath increased; this is found to be due the turning of the total ion velocity more toward the magnetopause so that the solar wind pressure became more effective in compressing the plasma against the magnetopause. The orientation of the bow shock normal with respect to the upstream magnetic field does not seem to have any significant effect on the ion density increase in the sheath following the discontinuity in the CME. The ion velocity, which was antisunward and poleward in the magnetosheath, showed a large reversal at References Balan, N., Y. Otsuka, and S. Fukao (1997), New aspects in the ionosphere observed by the MU radar, Geophys. Res. Lett., 24, Balan, N., et al. (2006), Cluster observations of a structured magnetospheric cusp, Ann. Geophys, 24, Balogh, A., et al. (2001), The Cluster magnetic field investigation: overview of in-flight performance and initial results, Ann. Geophys., 19, Borovsky, J. E., and J. T. Steinberg (2006), The calm before the storm in CIR/magnetosphere interactions: Occurrence statistics, solar wind statistics, and magnetospheric preconditioning, J. Geophys. Res., 111, A07S10, doi: /2005ja Cargill, P. J., M. W. Dunlop, B. Lavraud, R. C. Elphic, D. L. Holland, K. Nykyri, A. Balogh, I. Dandouras, and H. Reme (2004), CLUSTER encounter with the high altitude cusp: Boundary structure and magnetic field depletions, Ann. Geophys., 22, Cargill, P. J., et al. (2005), Cluster at the magnetospheric cusps, Space Sci. Rev., 118, 321. Chapman, S., and V. C. A. Ferraro (1930), A new theory of magnetic storms, Nature, 126, 129. Cornilleau-Wehrlin, N., et al. (2003), First results obtained by the Cluster STAFF experiment, Ann. Geophys., 21, 437. Dmitriev, A. V., J. K. Chao, and D. J. Wu (2003), Comparative study of bow shock models using Wind and Geotail observations, J. Geophys. Res., 108(A12), 1464, doi: /2003ja Dungey, J. W. (1961), Interplanetary magnetic field and the auroral zones, Phys. Rev. Lett., 6, 47. Escoubet, C. P., M. Fehringer, and M. Goldstein (2001), The Cluster mission, Ann. Geophys., 19, Gopalswamy, N., L. Barbieri, G. Lu, S. P. Plunkett, and R. M. Skoug (2005a), Introduction to the special section: Violent Sun-Earth connection events of October November 2003, Geophys. Res. Lett., 32, L03S01, doi: /2005gl Gopalswamy, N., S. Yashiro, Y. Liu, G. Michalek, A. Vourlidas, M. L. Kaiser, and R. A. Howard (2005b), Coronal mass ejections and other extreme characteristics of the 2003 October November solar eruptions, J. Geophys. Res., 110, A09S15, doi: /2004ja Haerendel, G., G. Paschmann, N. Sckopke, H. Rosenbauer, and P. C. Hedgecock (1978), The frontside boundary layer of the magnetosphere and the problem of reconnection, J. Geophys. Res., 83, Heikkila, W. J., and J. D. Winningham (1971), Penetration of magnetosheath plasma to low altitude through the dayside magnetospheric cusp, J. Geophys. Res., 76, 883. Johnston, A. D., et al. (1997), PEACE: A plasma electron and current experiment, Space Science Rev., 79, 351. Kennel, C. F., J. P. Edmiston, and T. Hada (1985), A quarter century of collisionless shock research, in Collisionless Shocks in the Heliosphere: Reviews of Current Research, Geophys. Monogr. Ser., vol. 35, edited by R. G. Stone and B. T. Tsurutani, p. 1, AGU, Washington, D. C. Kessel, R. L., S.-H. Chen, J. L. Green, S. F. Fung, S. A. Boardsen, L. C. Tan, T. E. Eastman, J. D. Craven, and L. A. Frank (1996), Evidence of high-latitude reconnection during northward IMF: Hawkeye observations, Geophys. Res. Lett., 23, 583. Lavraud, B., et al. (2002), Cluster observations of the exterior cusp and its surrounding boundaries under northward IMF, Geophys. Res. Lett., 29(20), 1995, doi: /2002gl Lavraud, B., A. Fedorov, E. Budnik, A. Grigoriev, P. J. Cargill, M. W. Dunlop, H. Reme, I. Dandouras, and A. Balogh (2004a), Cluster survey of the high-altitude cusp properties: A three-year statistical study, Ann. Geophys., 22, of10

10 Lavraud, B., et al. (2004b), The exterior cusp and its boundary with the magnetosheath: Cluster multi-event analysis, Ann. Geophys., 22(8), Lavraud, B., A. Fedorov, E. Budnik, M. F. Thomsen, A. Grigoriev, P. J. Cargill, M. W. Dunlop, H. Rme, I. Dandouras, and A. Balogh (2005), High-altitude cusp flow dependence on IMF orientation: A 3-year Cluster statistical study, J. Geophys. Res., 110, A02209, doi: / 2004JA Lavraud, B., M. F. Thomsen, J. E. Borovsky, M. H. Denton, and T. I. Pulkkinen (2006), Magnetosphere preconditioning under northward IMF: Evidence from the study of coronal mass ejection and corotating interaction region geoeffectiveness, J. Geophys. Res., 111, A09208, doi: /2005ja Lockwood, M., and M. F. Smith (1994), Low- and middle-altitude cusp particle signatures for general magnetopause reconnection rate variations: 1. Theory, J. Geophys. Res., 99, Lundin, R., B. Aparicio, and M. Yamauchi (2001), On the solar wind flow control of the polar cusp, J. Geophys. Res., 106, 13,023. Mannucci, A. J., B. T. Tsurutani, B. A. Iijima, A. Komjathy, A. Saito, W. D. Gonzalez, F. L. Guarnieri, J. U. Kozyra, and R. Skoug (2005), Dayside global ionospheric response to the major interplanetary events of October 29 30, 2003 Halloween Storms, Geophys. Res. Lett., L12S02, doi: /2004gl Newell, P. T., C. I. Meng, D. G. Sibeck, and R. Lepping (1989), Some low altitude dependencies on the interplanetary magnetic field, J. Geophys. Res., 94, Opgenoorth, H. J., et al. (2001), Co-ordinated ground-based, low altitude satellite and Cluster observations on global and local scales during a transient post noon sector excursion of the magnetospheric cusp, Ann. Geophys., 19, Onsager, T. G., J. D. Scudder, M. Lockwood, and C. T. Russell (2001), Reconnection at the high latitude magnetopause during northward interplanetary magnetic field conditions, J. Geophys. Res., 106, 25,467. Owen, C. J., et al. (2001), Cluster PEACE observations of electrons during magnetospheric flux transfer events, Ann. Geophys., 19, Reme, H., et al. (2001), First multispacecraft ion measurements in and near the Earth s magnetosphere with the identical Cluster ion spectrometry (CIS) experiment, Ann. Geophys., 19, Rosenqvist, L., H. Opgenoorth, S. Buchert, I. McCrea, O. Amm, and C. Lathuillere (2005), Extreme solar-terrestrial events of October 2003: High-latitude and Cluster observations of the large geomagnetic disturbances on 30 October, J. Geophys. Res., 110, A09S23, doi: / 2004JA Russell, C. T. (2000), POLAR eyes the cusp, Eur. Space Agency Spec. Publ., 449, 47. Russell, C. T., and C. Elphic (1979), ISEE observations of flux transfer events at the dayside magnetopause, Geophys. Res. Lett., 6, 33. Shue, J.-H., et al. (1998), Magnetopause location under extreme solar wind conditions, J. Geophys. Res., 103, 17,700. Skoug, R. M., J. T. Gosling, J. T. Steinberg, D. J. McComas, C. W. Smith, N. F. Ness, Q. Hu, and L. F. Burlaga (2004), Extremely high speed solar wind: October 2003, J. Geophys. Res., 109, A09102, doi: / 2004JA Spreiter, J. R., and S. S. Stahara (1980), A new predictive model for determining solar wind-terrestrial planet interactions, J. Geophys. Res., 85, Spreiter, J. R., and A. L. Summers (1967), On conditions near the neutral points on the magnetosphere boundary, Planet. Space Sci., 15, 787. Stubbs, T. J., P. J. Cargill, M. Lockwood, M. Grande, B. J. Kellett, and C. H. Perry (2004), Extended cusp-like regions and their dependence on the Polar orbit, seasonal variations, and interplanetary conditions, J. Geophys. Res., 109, A09210, doi: /2003ja Terasawa, T., et al. (2005), Determination of shock parameters for the very fast interplanetary shock on 29 October 2003, J. Geophys. Res., 110, A09S12, doi: /2004ja Tsyganenko, N. A., and D. P. Stern (1996), Modeling the global magnetic field of the large-scale Birkeland current systems, J. Geophys. Res., 101, 27,187. van Eyken, A. P., H. Rishbeth, D. M. Willis, and S. W. H. Cowley (1984), Initial EISCAT observations of plasma convection at invariant latitudes 7077, J. Atmos. Terr. Phys., 46, 635. Wing, S., P. T. Newell, and J. M. Ruohoniemi (2001), Double cusp: Model predictions and observational verification, J. Geophys. Res., 106, 25,571. Woch, J., and R. Lundin (1992), Magnetosheath plasma precipitation in the polar cusp and its control by the interplanetary magnetic field, J. Geophys. Res., 97, Zhang, H., T. A. Fritz, Q.-G. Zong, and P. W. Daly (2005), Stagnant exterior cusp region as viewed by energetic electrons and ions: A statistical study using Cluster Research with Adaptive Particle Imaging Detectors (RAPID) data, J. Geophys. Res., 110, A05211, doi: / 2004JA Zong, Q.-G., T. A. Fritz, H. Zhang, A. Korth, P. W. Daly, M. W. Dunlop, K.-H. Glassmeier, H. Reme, and A. Balogh (2004), Triple cusps observed by Cluster Temporal or spatial effect?, Geophys. Res. Lett., 31, L09810, doi: /2003gl Zong, Q.-G., T. A. Fritz, A. Korth, P. W. Daly, M. Dunlop, A. Balogh, J. F. Fennell, J. D. Sullivan, R. W. H. Friedel, and H. Reme (2005), Energetic electrons as a field line topology tracer in the high latitude boundary/cusp region: Cluster Rapid observations, Surv. Geophys., 26(13), 215, doi: /s z. H. Alleyne, N. Balan, and S. Walker, Automatic Control and Systems Engineering, University of Sheffield, Mappin Street, Sheffield S1 3JD, UK. (b.nanan@sheffield.ac.uk) N. Cornilleau-Wehrlin, Centre d Etudes des Environnements Terrestre et Planétaires du Centre National de la Recherche Scientifique, Avenue de l Europe, F Vélizy-Villacoublay, France. A. N. Fazakerley, Mullard Space Science Laboratory, University College London, Holmbury St. Mary, Dorking, Surrey RH5 6NT, UK. E. Lucek, Blackett Laboratory, Imperial College, London SW7 2BZ, UK. H. Reme, Centre d Etude Spatiale des Rayonnements, 9 Avenue du Colonel Roche, Boite Postale 44346, F Toulouse Cedex 4, France. A. P. van Eyken, EISCAT Scientific Association, P. O. Box 164, SE-981 Kiruna, Sweden. S.-R. Zhang, Haystack Observatory, Massachusetts Institute of Technology, Off Route 40, Westford, MA , USA. 10 of 10