• No results found

Statistics of Plasma Sheet Convection

N/A
N/A
Protected

Academic year: 2022

Share "Statistics of Plasma Sheet Convection"

Copied!
35
0
0

Laster.... (Se fulltekst nå)

Fulltekst

(1)

Statistics of Plasma Sheet Convection

1

L. Juusola,1N. Østgaard, and1E. Tanskanen1,2

L. Juusola, Finnish Meteorological Institute, P.O. Box 503, FIN-00101 Helsinki, Finland.

([email protected])

N. Østgaard, Department of Physics and Technology, University of Bergen, Postboks 7803, NO-5020 Bergen, Norway.

([email protected])

E. Tanskanen, Finnish Meteorological Institute, P.O. Box 503, FIN-00101, Helsinki, Finland.

([email protected])

1Department of Physics and Technology, University of Bergen, Bergen, Norway.

2Finnish Meteorological Institute, Helsinki, Finland.

(2)

Abstract. We study statistically plasma sheet convection using ion and

2

magnetic field data obtained by Cluster 1 and 3 (years 2001–2007), Geotail

3

(1995–2006), and THEMIS 1–5 (2007–2009). The condition ionβ > 0.5 is

4

used to find the plasma sheet regime. Plasma sheet convection is observed

5

to be dominated by slow speed (<100 km/s) flows that circulate around Earth

6

on both sides towards the dayside. Higher speeds flows are concentrate around

7

the aberrated midnight meridian. With increasing speed, the sunward com-

8

ponent of the flow velocity becomes more pronounced, such that flows with

9

V >500 km/s are directed almost purely sunward, not circulating around Earth

10

like the slower flows. Interplanetary magnetic field (IMF) y- and z-components

11

are observed to penetrate the plasma sheet, causing a rather uniform change

12

of a few nT in the same magnetic field component. Moreover, during IMFz<0

13

conditions, a channel of increased Bz is created in the nightside around the

14

aberrated midnight axis. It is suggested that the channel is caused by dipo-

15

larization and magnetic flux pileup related to fast flows. Compared to IMFz<0

16

conditions, the nightside region of highest mean flow speed is shifted towards

17

dusk during IMFy<0 conditions, and towards dawn during IMFy>0 condi-

18

tions. For the V >100 km/s flows, no correlation is found between the plasma

19

sheet flow speed and the solar wind electric field magnitude, but between the

20

flow speed and IMF clock angle there is a clear correlation, with increasing

21

speed as IMF turns southward.

22

(3)

1. Introduction

According to the open magnetosphere model byDungey [1961], the closed geomagnetic

23

field lines reconnect with the interplanetary magnetic field (IMF) on the dayside mag-

24

netopause. The resulting open field lines are dragged tailwards across the northern and

25

southern polar caps by the solar wind. On the nightside they reconnect again, and the re-

26

sulting closed, but far stretched field lines migrate back toward the dayside magnetopause.

27

This process is known as the magnetospheric convection. Occasionally the dayside and

28

nightside reconnection rates are balanced, resulting in periods of steady magnetospheric

29

convection [e.g., Sergeev et al., 1996], but more often more magnetic flux is either opened

30

or closed [e.g.,Milan et al., 2007]. The substorm, for instance, consists of such periods of

31

loading and unloading.

32

According to the two-fluid approximation, two components contribute to the ion velocity in the plasma sheet. These are the electric drift and the diamagnetic drift:

Vi =qiE× B

qiB2 −∇Pi

Ni × B

qiB2, (1)

whereqi,Pi,Ni, andViare the ion charge, thermal pressure, number density, and velocity,

33

and E and B the electric and magnetic field. In the plasma sheet, ions and electrons

34

drift diamagnetically in opposite direction, creating the cross-tail current. The relative

35

strengths of the two terms in Eq. (1) are determined by two factors. First, the electric

36

field and pressure gradients vary significantly in the near-Earth region. Second, electric

37

drift does not depend on particle energy, but diamagnetic drift does. Thus, the motion

38

of low-energy ions is dominated by electric drift, and the motion of high-energy ions by

39

diamagnetic drift.

40

(4)

Several statistical studies based on ion moments computed from satellite measurements

41

have addressed the characteristics of plasma sheet convection. In most cases, the focus

42

has either been on the slow [Angelopoulos et al., 1993; Zhu, 1993; Wang et al., 2006] or

43

fast [Baumjohann et al., 1990;Angelopoulos et al., 1992, 1994; Shiokawa et al., 1997;Raj

44

et al., 2002] plasma sheet flows, and the studied region limited to the nightside plasma

45

sheet.

46

Using data from the AMPTE/IRM satellite (four months from 1986) in the region

47

−19 RE < xGSM < −9 RE (Earth radii) and |yGSM| < 15 RE, Baumjohann et al. [1989]

48

inferred that the average ion flow speed in the central plasma sheet is low, below 100

49

km/s, due to the predominance of low-speed flows. Magnetotail lobe or plasma sheet

50

boundary layer (PSBL) samples were excluded from the data set by assuming that in the

51

plasma sheet spacecraft charging effects should be absent. High-speed flows were found

52

to occur in the plasma sheet, but in bursts mostly less than one minute in duration, with

53

intermittent intervals of near stagnant plasma. The bursts were more frequent during

54

geomagnetically disturbed times, but occurred also during low Auroral Electrojet (AE)

55

index conditions.

56

Baumjohann et al. [1990] used AMPTE/IRM data (eight months from spring 1985 and

57

spring 1986) to study the occurrence rates and typical characteristics of high-speed (V >

58

400 km/s) ion flows in the plasma sheet and PSBL. The largest occurrence rates of high-

59

speed flows were found near the midnight meridian at the largest radial distances accessible

60

to the spacecraft (≈19 RE), and their occurrence strongly peaked in the sunward direction.

61

The high-speed flows were found to be bursty, with the majority of the flows lasting less

62

than 10 sec.

63

(5)

Angelopoulos et al. [1992] studied the high-speed flows reported by Baumjohann et al.

64

[1989] using AMPTE/IRM data (two months from 1985). They found that the high-speed

65

flows organize themselves in 10-min time scale flow enhancements, which they termed

66

bursty bulk flow (BBF) events. The flow peaks were usually associated with magnetic

67

field dipolarization and ion temperature increases.

68

Angelopoulos et al. [1993] constructed the average flow pattern in the quiet, that is,

69

non-BBF, plasma sheet (β > 0.5) using AMPTE/IRM (6 months from 1985) and ISEE

70

2 data (12 months from 1978 and 1979). BBFs were defined as plasma sheet segments of

71

continuous ion flow of V >100 km/s, during which V exceeded 400 km/s at least once.

72

The ISEE 2 orbit extended the analysis beyond the apogee of IRM to 22 RE. The flow

73

was found to be predominantly duskward at local midnight, while closer to the flanks of

74

the examined region it was mostly earthward.

75

Zhu [1993] constructed the average convection pattern in the plasma sheet (β > 0.2)

76

using ISEE 1 and 2 data (14 months in 1977–1979) in the region−20 RE < xGSM <−10 RE

77

and|yGSM|<15 RE. Only flows withqVx2+Vy2 <250 km/s were included. They showed

78

that the convective flows tend to follow contours of constant unit magnetic flux volume

79

as they move around the Earth, which helps to avoid the pressure balance inconsistency

80

found in two dimensional magnetotail models [e.g., Erickson and Wolf, 1980].

81

Angelopoulos et al.[1994] studied the statistical properties of BBFs in the plasma sheet

82

using AMPTE/IRM and ISEE 2 data. The plasma sheet was defined as the regime where

83

β > 0.5, and BBFs as segments of V > 100 km/s in the plasma sheet, during which

84

V exceeded 400 km/s for at least at one sample. Samples of V > 400 km/s that were

85

less than 10 minutes apart were considered to belong to the same BBF event, even if the

86

(6)

speed dropped below 100 km/s between these samples. They inferred that BBFs were

87

responsible for 60–100% of the earthward transport of mass, energy and magnetic flux,

88

even though they occurred only 10–15% of the observation time the spacecraft spent in

89

the plasma sheet. Earthward ofxGSM=−19 RE, the occurence rate of BBFs was observed

90

to decrease. Braking of the flow speed was suggested as a possible explanation.

91

Shiokawa et al.[1997] studied possible braking mechanisms of high-speed (>400 km/s)

92

ion flows in the near-Earth plasma sheet using the same data set as Baumjohann et al.

93

[1989]. The high-speed flows were found almost always to be directed earthward between

94

−19 RE< xGSM<−9 RE, indicating that their source was beyondxGSM<−19 RE. The

95

occurrence rate of the high-speed flows was observed to substantially decrease towards

96

Earth, but flows with speeds >600 km/s were still observed at xGSM = −9 RE. It was

97

suggested that the high-speed flows are stopped at a clear boundary between the regions

98

of dipolar and tail-like field in the plasma sheet, which corresponds to the inner edge of

99

the plasma sheet. The average jump of the magnetic field at this boundary was estimated

100

to be 6.7 nT.

101

The fast flows observed in the PSBL typically consist of unidirectional or counter stream-

102

ing ion beams strongly aligned with the magnetic field [e.g.,Raj et al., 2002, and references

103

therein]. These field-aligned beams have sharp cutoffs at low energies, and occur within

104

relatively steady magnetic field and plasma conditions. For counter streaming beams the

105

velocity moments do not represent the bulk motion of the plasma. Quasi-steady recon-

106

nection tailward of the observation point has been suggested as the source of these beams

107

[Onsager et al., 1991]. High speed flows observed in the plasma sheet, on the other hand,

108

typically consist of a single bulk flow population directed mostly sunward, independent of

109

(7)

the local magnetic field direction. These bulk flows do not have low-energy cutoffs, and

110

are often associated with magnetic field dipolarization and plasma temperature enhance-

111

ments.

112

Generally, it is assumed that field-aligned beams and bulk flows occur in the PSBL and

113

plasma sheet, respectively, and these two regions can be distinguished from each other

114

based on, for instance, plasma β. Raj et al. [2002] surveyed all high-speed (>250 km/s)

115

flows detected by Wind during 17 perigee passes across the near-Earth (−25 RE < xGSE <

116

0) plasma sheet between 1995 and 1997. Instead of plasma moments or the regions in

117

which the flows were detected, they classified the high-speed flow events based on their

118

ion distribution characteristics. They discovered that bulk flows are perpendicular to

119

the magnetic field when detected at the neutral sheet (Bx ≈ 0) but have a large field-

120

aligned component at higher magnetic latitudes. Field-aligned bulk flow and field-aligned

121

beams were similar in terms of their velocity moments and could occur at the same

122

magnetic latitudes, but were easily distinguishable based on their ion distributions. No

123

single moment-based parameter or threshold could cleanly separate beam from bulk flow

124

distributions, because a range of values of these parameters existed where both types of

125

fast flows were observed. Moreover, they observed little or no temperature enhancements

126

in cases where the spacecraft resided near the neutral sheet before the arrival of bulk

127

flows, suggesting that the temperature enhancements seen in other bulk flow events might

128

in part be a spatial effect instead of true heating of the plasma. The occurrence of high-

129

speed bulk flow events had a dawn-dusk asymmetry, with higher occurrence rates in the

130

premidnight sector of the plasma sheet, while beam events occurred with equal probability

131

on both sides of midnight.

132

(8)

Wang et al. [2006] used Geotail data (November 1994 to April 1998) in the region

133

|yGSM| < 22.5 RE and −30 RE < xGSM < 0 to study the distributions of plasma sheet

134

(β >1) ions and magnetic fields under northward and southward IMF conditions. They

135

utilized the condition VGSM,x > −100 km/s to exclude magnetosheath crossings, and

136

V < 200 km/s to exclude large BBFs. The perpendicular flow around |yGSM| = 10 RE

137

was observed to be stronger in the premidnight than in the postmidnight region, and this

138

asymmetry was shown to result from the westward diamagnetic drift. The flow pattern

139

was not significantly different during northward and southward IMF conditions, but the

140

overall flow speed was higher during southward IMF.

141

The several years of Geotail, Cluster, and THEMIS data now available provide an excel-

142

lent opportunity to revisit the question of the characteristics of plasma sheet convection.

143

The abudance of data enables us to reveal more detailed as well as new infomation on

144

this important process, by which disturbances are communicated from the tail to the

145

inner magnetosphere. In this study, we analyze statistically plasma sheet bulk flows at

146

radial distances<30 RE from Earth. In Sect. 3.1, we construct the average plasma sheet

147

flow pattern, and discuss the relative abundance of sunward and tailward flows. Such

148

convection maps could also be useful, when validating magnetospheric simulation codes.

149

Sect. 3.2 deals with dawn-dusk symmetries and asymmetries observed in the plasma sheet

150

parameters, and in Sect. 3.3 the effect of the IMF clock angle is examined. As noted

151

by Angelopoulos et al. [1992], the often used fast flow speed threshold of 400 km/s does

152

not represent any physically significant quantity, and also flows with speeds below the

153

threshold are an integral part of the BBF structure. Thus, in Sect. 3.4, we have divided

154

all samples according to speed into 10 bins (0–100 km/s, 100–200 km/s, ..., 900–1000

155

(9)

km/s), and examined the properties of each bin separately. The convection electric field

156

related to the average flow pattern is derived in Sect. 3.5. The final Sect. 4 contains dis-

157

cussion and conclusions. We begin by introducing the used instruments and data analysis

158

techniques (Sect. 2).

159

2. Data

In this study we have used magnetospheric ion and magnetic field data obtained by the

160

Cluster 1 and 3 satellites from 2001 to 2007, by the Geotail satellite from 1995 to 2006,

161

and by the THEMIS 1–5 satellites from September 2007 to the end of 2009.

162

From the Cluster 1 and 3 spacecraft we have used 4-s (spin average) resolution data

163

from the Fluxgate magnetometer [FGM,Balogh et al., 2001] and ion moments from Cluster

164

Ion Spectrometry [CIS,R`eme et al., 2001] Hot Ion Analyser (HIA) instrument. CIS/HIA

165

measures energies from 5 eV/q to 32 keV/q. From the Geotail spacecraft we have utilized

166

3-s (spin average) resolution data from the Magnetic Field experiment [MGF, Kokubun

167

et al., 1994], and 12-s resolution ion moments from the Low Energy Particle experiment

168

[LEP, Mukai et al., 1994]. LEP/EA measures energies from several eV/q to 43 keV/q.

169

From the THEMIS 1–5 spacecraft we have used 3-s (spin average) resolution data from

170

the Fluxgate magnetometer [FGM,Auster et al., 2008], and ion moments from the Elec-

171

trostatic Analyzer [ESA, McFadden et al., 2008a, b]. ESA measures energies from 1.6

172

eV/q to 25 keV/q. Solar wind data at 1-min resolution propagated to Earth’s bow shock

173

nose were extracted from NASA/GSFC’s OMNI data set through the OMNIWeb interface

174

(http://omniweb.gsfc.nasa.gov/).

175

Figure 1 displays the observation times of ion β > 0.5 (this choice is discussed below)

176

on a 2.5 RE×2.5 RE grid (a): for Cluster 1 and 3 from 2001 to 2007, (b): for Geotail

177

(10)

from 1995 to 2006, and (c): for THEMIS 1–5 from 2007 to 2009, presented as a functionx

178

and y. Here and elsewhere in this study we have used Cartesian (x,y,z) Geocentric Solar

179

Magnetospheric (GSM) coordinates. The gray circle represents the geostationary orbit at

180

the 6.6 RE radial distance, and the gray curve the magnetopause according toShue et al.

181

[1997], with Pdyn = 1 nPa and IMF Bz = 0. Both the magnetopause and the solid grid

182

have been rotated 4.8 clockwise to take into account the aberration due to Earth’s orbital

183

motion. The thick gray line shows the aberrated noon-midnight axis. Approximately a

184

circle of 30 RE radius was covered by the observations.

185

Our aim is to study statistically plasma sheet bulk flows, which means that first we had

186

to separate those data from other samples. With such a large amount of data involved,

187

it was not feasible to examine every particle distribution. Therefore, we compromised

188

between the amount of data and accurate selection of plasma sheet bulk flows by using

189

moment-based criteria. We chose the criterion ionβ >0.5 for the plasma sheet. According

190

to Raj et al. [2002], this would include 95% of high-speed (V > 250 km/s) bulk flow

191

samples, but at the same time, 55% of undesired high-speed field-aligned beam samples

192

would also satisfy the criterion. Although increasing the limit to β > 2 would eliminate

193

more than 95% of beam samples, only 45% of bulk samples would be included, decreasing

194

the amount of data significantly. As is obvious from Fig. 1, towards plasma sheet flanks,

195

particularly on the dayside, the criterion also included magnetosheath and solar wind

196

samples.

197

Because of the large amount of data, we began by dividing our samples into “flow events”, represented by averages of the ion and magnetic field measurements during the event. Also other parameters describing the ambient conditions, such as solar wind data,

(11)

were included. Because plasma sheet convection is dominated by slow speed flows, a flow event was defined as a continuous block of samples during which β > 0.5 and the flow speed remained within one of the bins: 0–100 km/s, 100–200 km/s, ..., 900–1000 km/s.

Typical durations of the events varied from seconds to minutes. In order to emphasize the bulk flows in the center of the plasma sheet, the ion moment and magnetic field averages were weighted by β:

x=

Pn i=1βixi

Pn i=1βi

. (2)

The solar wind averages tagged to each event included not just the event period but

198

also the preceding 30 min in order to better reflect the conditions in the plasma sheet.

199

These event-averages were then gridded according to their xy-locations, and averages of

200

the event-averages, weighted by the durations of the events, were computed for each grid

201

point. The results are displayed in plots such as those in Figure 2.

202

3. Results

3.1. Sunward and Tailward Flows

The color scale in panel (a) of Fig. 2 shows the time Cluster 1 and 3, Geotail, and

203

THEMIS 1–5 had altogether spent observing β >0.5. Due to the orbits of the satellites,

204

some parts within the 30 RE radius circle were better covered than others, but almost

205

everywhere there was at least one day of observations. The vectors in panel (a) of Fig.

206

2 display the mean velocity. The pattern revealed three distinct regions: Upstream the

207

undisturbed solar wind flow had an average angle of 4.8 with the −x axis due to Earth’s

208

orbital motion, in the magnetosheath the solar wind flow was deflected around the magne-

209

(12)

topause, and inside the magnetosphere the average flow was near stagnant at the displayed

210

scale.

211

Panel (b) of Fig. 2 displays the percentage of time spent observing Vx ≥ 0 relative to

212

all β > 0.5 samples. On the dawn and dusk flanks of the plasma sheet the flows had

213

predominantly a sunward component, but towards the dayside the percentage decreased

214

due to the increasing amount of antisunward magnetosheath flow samples. Also in the

215

nightside plasma sheet, mainly in the region −30 RE < x < −18 RE and −6 RE < y <

216

12 RE, a significant percentage of the flows were antisunward.

217

At the equatorial plane, antisunward flows observed on open field lines behind an X-line

218

would generally be expected to be associated with a southward magnetic field component,

219

while flows on closed field lines would have a northward component. Panels (c) and (d)

220

of Fig. 2 display the percentage of time spent observingBz <0 relative to all β >0.5 and

221

Vx ≥0 and Vx <0 samples. As expected, most sunward flows in the tail were associated

222

with a northward magnetic field component, but so were the majority of antisunward

223

flows. The majority of both sunward and antisunward flows had very low speeds, as is

224

demonstrated by the small mean velocities displayed in panels (e) and (f). Time series

225

of the velocity moments of such low speed flows typically fluctuate around zero, thus

226

containing both positive and negative values. Although the geomagnetic field generally

227

has a positive z component at the equatorial plane, north and south of this plane the sign

228

depends on the distance from Earth, such that close to Earth Bz < 0 and farther away

229

>0. Thus, there are large numbers of samples inside the geostationary orbit in panel (c)

230

of Fig. 2, originating from high magnetic latitudes. Instead of midnight, the majority of

231

the flows with a southward magnetic field component were observed in the premidnight

232

(13)

region. Fast flows with Vx > 0 and Bz < 0 or Vx < 0 and Bz > 0 could be observed,

233

for instance, during passages of magnetic flux ropes or plasmoids. On closed field lines,

234

antisunward flows also take place at the flanks of sunward fast flow channels.

235

Panels (e) and (f) of Fig. 2 show the mean speed (color) and flow direction (vectors)

236

for β > 0.5 and Vx ≥ 0 or Vx < 0. The lengths of the velocity vectors have been

237

normalized to unity, and the xy-component is displayed. For the Vx ≥0 flows, the mean

238

speed was largest inside the approximate nightside region −30 RE < x < −18 RE and

239

−6 RE < y <12 RE, and diminished rather sharply towards Earth and the dawn and dusk

240

sides. In this region the flows were mainly aligned with the x-axis, with a small positive

241

y-component, which strengthened towards Earth. Outside this region, the flows diverged

242

towards the dawn and dusk flanks. The convection circulated azimuthally around Earth,

243

and converged towards noon on the dayside. Outside the magnetopause there were some

244

flows with a sunward component diverging away from the subsolar point. These could

245

be reflected magnetosheath flows. For the Vx <0 flows, only the nightside plasma sheet

246

could be seen, as the dayside and flanks were dominated by magnetosheath flows. The

247

mean flow pattern resembles that of the sunward flows with negative y-components in the

248

postmidnight region and positive elsewhere.

249

Because of the large amount of data, it was not feasible to examine every orbit for

250

magnetopause crossings. As there is no other clean way to separate antisunward mag-

251

netosheath flows from antisunward plasma sheet flows based on moment data only, from

252

now on only samples with a sunward velocity component have been included.

253

3.2. Dawn-Dusk Symmetries and Asymmetries

(14)

The convection pattern in panel (e) of Fig. 2 shows some dawn-dusk asymmetries:

254

In GSM coordinates, the nightside region of higher mean flow speed was not located

255

symmetrically around the x-axis, but shifted to the premidnight side. However, the region

256

appears to have been fairly symmetrical with respect to the aberrated midnight axis.

257

Furthermore, the diamagnetic drift causes the positive ions to drift westward around

258

Earth, and this effect became stronger closer to Earth due to the magnetic field gradient.

259

Figure 3 displays the mean ion number density, temperature, magnetic field strength,

260

and plasma β. The scales are logarithmic. The ion number density distribution did not

261

present any clear dawn-dusk asymmetries, but in the near-Earth region the temperature

262

appeared to be higher on the duskside than on the dawnside. In the plasmasphere inside

263

the geosynchronous orbit and in the magnetosheath, the plasma was colder and denser

264

than in the plasma sheet.

265

The magnetic field strength, on the other hand, was clearly weaker in the premidnight

266

region tailward of aboutx≈ −20 RE than elsewhere. The magnetic field strength dawn-

267

dusk asymmetry also caused the mean β, which is the ration between ion thermal and

268

magnetic pressures, to be asymmetric with larger values in the premidnight region than

269

in the postmidnight region. To make sure that the effect was not artificial and caused by

270

orbital bias, panels (e) and (f) of Fig. 2 show β separately for the northern (Br <0) and

271

southern (Br > 0) hemispheres. As the effect is still present in both panels, it seems to

272

be real.

273

3.3. Effect of IMF Clock Angle

Panel (a) of Figure 4 displays Byduring IMFBy<0 conditions (clock angle|θ+ 90|<

274

45) with a background of IMFBz>0 (|θ|<45) conditions subtracted. Similarly, panels

275

(15)

(c) and (e) display By during IMF By > 0 (|θ−90| < 45) conditions, and Bz during

276

IMF Bz<0 (|θ|>135) conditions.

277

All three panels indicate that an ambient IMFByorBzcauses a rather uniform change

278

of a few nT in the same magnetic field component in the plasma sheet. However, during

279

southward IMF conditions, a channel of enhanced positive Bz was also created around

280

the aberrated midnight axis. In x-direction, the channel extended from x ≈ −30 RE

281

to approximately to the geostationary orbit. In y-direction, the width of the channel

282

was roughly 20 RE, and corresponded to the region where higher mean flow speeds were

283

observed in panel (e) of Fig. 2. The channel could have been caused by dipolarization and

284

magnetic flux pileup related to the sunward fast flows.

285

Panel (b) of Fig. 4 displaysV during IMFBy<0 conditions with a background of IMF

286

Bz >0 conditions subtracted. Similarly, panels (d) and (f) display V during IMF By >0

287

and IMF Bz <0 conditions. In general, the mean flow speed on the nightside was in all

288

cases higher than during northward IMF conditions, but highest during southward IMF

289

conditions. Compared to IMF Bz < 0 conditions, the nightside region of highest flow

290

speed was shifted towards dusk during IMFBy<0 conditions, and towards dawn during

291

IMF By >0 conditions.

292

3.4. Flow Speed

The color scales in Figure 5 display the percentage of time spent observing flows with

293

β > 0.5, Vx ≥ 0, and (a): 0 ≤ V < 100 km/s, (b): 100 km/s ≤ V < 200 km/s,

294

(c): 200 km/s ≤ V < 300 km/s, (d): 300 km/s ≤ V < 400 km/s, (e): 400 km/s ≤

295

V < 500 km/s, (f): 500 km/s ≤ V < 600 km/s, (g): 600 km/s ≤ V < 700 km/s, (h):

296

700 km/s ≤ V < 800 km/s, (i): 800 km/s ≤ V < 900 km/s, and (j): 900 km/s ≤

297

(16)

V < 1000 km/s relative to flows with β > 0.5, Vx ≥ 0, and 0 ≤ V < 1000 km/s. Note

298

that the scales vary between the panels. The vast majority of the flows in the plasma

299

sheet had speeds <100 km/s. Only around the aberrated midnight axis, in the region

300

−6 RE < y < 12 RE and −30 RE < x < −18 RE, about 50% of the flows had higher

301

speeds. With increasing flow speed, the flows became more concentrated around midnight

302

and farther downtail. The duration of each speed bin relative to all β >0.5, Vx≥0, and

303

0≤V < 1000 km/s observations are also shown in panel (a) of Figure 6. The centers of

304

the bins have been used as the x-axis, and the y-axis scale is logarithmic. 95% of the time

305

the flows had speeds <100 km/s.

306

The vectors in Fig. 5 show the mean flow direction for each speed bin. The slow speed flows were mainly azimuthal with respect to Earth, but with increasing speed the flow direction turned sunward. This is illustrated more quantitatively in panel (b) of Fig. 6, which shows an mean of the angle α of all flow events in each 100 km/s speed bin. α for one flow event was defined as

cos(α) = Vxy·Bxy

VxyBxy . (3)

As the angles would otherwise be α and 180 −α for the oppositely oriented magnetic

307

field lines of the southern and northern hemispheres for the same velocity vector, all field

308

lines of the southern hemisphere (Br > 0) were turned anti-parallel: Bxy → −Bxy. The

309

lengths of the error bars correspond to 20 × the standard deviation of the mean. The

310

very small values indicate that the means are significant.

311

Panel (c) of Fig. 6 displays the mean solar wind electric field| −Vsw×IMF|of all flow

312

events in each speed bin. The electric field was clearly weakest for the slow speed (<100

313

(17)

around 2 mV/m. Panel (d) of Fig. 6 shows the mean of the absolute value of the IMF clock

315

angle (|θ|) of all flow events in each speed bin. Again, the V <100 km/s bin was clearly

316

different from the others, with |θ| <90. For the other bins, |θ|>90, and increased for

317

higher speeds.

318

3.5. Convection Electric Field

Figure 7 shows the meanE=−V×Bmagnitude (color) and direction (vectors) for the

319

β >0.5 and Vx ≥ 0 samples. The lengths of the vectors have been normalized to unity,

320

and the xy-component is displayed. As the sunward flows on the nightside tended to occur

321

around the aberrated midnight axis, also the relatively pure y-directed field was restricted

322

to that region. Towards dusk and dawn, the field direction turned radially away from and

323

towards Earth, respectively, although the effect was more strong on the duskside. Near

324

the outer reaches of the magnetosphere, the field became normal to the magnetopause.

325

Mapped to the ionospheric Hall currents, which can be observed by ground-based mag-

326

netometers, the azimuthal flow in the dusk and dawn flanks would correspond to the

327

eastward and westward electrojets, respectively. The Harang discontinuity would be ex-

328

pected to correspond to the region where the sunward flow of the midnight region turned

329

azimuthal, or electric field direction turned from y to radial. In this region, the mean flow

330

speed was also observed to brake down significantly.

331

4. Discussion and Conclusions

In this study, magnetospheric ion and magnetic field data obtained by Cluster 1 and 3

332

between 2001 and 2007, by Geotail between 1995 and 2006, and by THEMIS 1–5 between

333

2007 and 2009 have been used to construct statistical maps of the plasma sheet convection.

334

(18)

Radial distances from Earth up to approximately 30 RE were covered. The condition ion

335

β > 0.5 was used to find the plasma sheet regime. However, on the dayside and near

336

the flanks of the magnetosphere, this condition also included magnetosheath and solar

337

wind flow samples. As we did not want to restrict our study only to the nightside plasma

338

sheet, and checking every orbit for magnetopause crossings was not feasible because of

339

the large amount of data, we concentrated mainly on the flows with a sunward velocity

340

component, which were the dominant population in the plasma sheet. As the produced

341

convection maps are statistical, they do not represent the actual convection pattern at

342

any given instant.

343

On the nightside, the mean configuration was observed to comprise mainly sunward

344

convection around midnight that diverged towards the dawn and dusk flanks. The con-

345

vection circulated azimuthally around Earth, and converged towards noon on the dayside.

346

The mean speed was largest inside a roughly 20 RE wide channel around the aberrated

347

midnight axis, and diminished rather sharply towards Earth and the dawns and dusk

348

sides. This is consistent with the results of Baumjohann et al.[1990], according to whom

349

the largest occurrence rates of high-speed flows are found near the midnight meridian. Al-

350

though flows with speeds as high as 900–1000 km/s could still be observed atx≈ −10 RE,

351

the occurrence frequency of high-speed flows started to decrease noticeably earthward of

352

about x ≈ −20 RE. This is in agreement with the observations of Baumjohann et al.

353

[1990] and Shiokawa et al. [1997] between−19 RE< x <−9 RE.

354

The ion number density distribution did not present any obvious dawn-dusk asym-

355

metries, and the temperature asymmetries were restricted near Earth (r < 10 RE), in

356

agreement with Wang et al. [2006]. The magnetic field strength, on the other hand, was

357

(19)

noticeably weaker in the premidnight region than in the postmidnight region. This caused

358

the mean ion β also to be asymmetric, with larger values in the premidnight region than

359

in the postmidnight region.

360

Ambient IMFByand Bz were observed to penetrate the plasma sheet, causing a rather

361

uniform change of a few nT in the same magnetic field component. Moreover, during

362

southward IMF conditions, a channel of enhanced positiveBz was created in the nightside

363

around the aberrated midnight axis. In y-direction, the width of the channel corresponded

364

to the region where fast flows were most frequently observed, but in x-direction it extended

365

closer to Earth, approximately to the geostationary orbit. We suggest that the channel

366

was caused by dipolarization and magnetic flux pileup related to fast flows. Consistent

367

with the results of Zhu [1993] and Wang et al. [2006], flow speeds on the nightside were

368

observed to be higher during southward IMF conditions than during northward IMF

369

conditions. We also found that compared to IMFBz<0 conditions, the nightside region

370

of highest flow speed was shifted towards dusk during IMFBy <0 conditions, and towards

371

dawn during IMF By >0 conditions.

372

During 95% of all β >0.5 and Vx ≥0 observation time, plasma sheet flows had speeds

373

<100 km/s. Only in the region −30 RE < x < −18 RE and −6 RE < y < 12 RE

374

about 50% of the flows had higher speeds. With increasing flow speed, the flows became

375

more and more concentrated around midnight and farther downtail. The slow speed

376

flows were mainly azimuthal with respect to Earth, but with increasing speed, the flow

377

direction turned increasingly sunward. This is consistent with the results ofBaumjohann

378

et al. [1990], who found that the occurrence of high-speed (V > 400 km/s) flows peaked

379

strongly in the sunward direction.

380

(20)

During times when slow flows (<100 km/s) were observed in the plasma sheet, the

381

solar wind electric field was on average weaker and IMF more northward than during

382

times when faster flows (>100 km/s) were observed. For the faster flows, no correlation is

383

found between the observed flow speeds and the mean solar wind electric field magnitude.

384

However, there is a clear correlation between the IMF clock angle and flow speed, with

385

increasing speed when IMF turns southward. According to our knowledge, this has not

386

been reported earlier.

387

Acknowledgments. We acknowledge the principal investigators E. Lucek (Clus-

388

ter/FGM), I. Dandouras (Cluster/CIS), T. Nagai (Geotail/MGF), and Y. Saito (Geo-

389

tail/LEP). We acknowledge NASA contract NAS5-02099 and V. Angelopoulos for use of

390

data from the THEMIS Mission. Specifically: C. W. Carlson and J. P. McFadden for

391

use of ESA data and K. H. Glassmeier, U. Auster and W. Baumjohann for the use of

392

FGM data provided under the lead of the Technical University of Braunschweig and with

393

financial support through the German Ministry for Economy and Technology and the

394

German Center for Aviation and Space (DLR) under contract 50 OC 0302. Cluster data

395

were acquired through Cluster Active Archive, and Geotail (Editor-A) and THEMIS data

396

through CDAWeb. We acknowledge NASA/GSFC’s Space Physics Data Facility’s OM-

397

NIWeb service, and OMNI data. This study was supported by the Norwegian Research

398

Council, through the Norwegian Cluster project 197639/V30.

399

References

Angelopoulos, V., W. Baumjohann, C. Kennel, F. Coroniti, M. Kivelson, R. Pellat,

400

R. Walker, H. L¨uhr, and G. Paschmann (1992), Bursty bulk flows in the inner cen-

401

(21)

tral plasma sheet, J. Geophys. Res., 97(A4), 4027–4039.

402

Angelopoulos, V., C. F. Kennel, F. V. Coroniti, R. Pellat, H. E. Spence, M. G. Kivelson,

403

R. J. Walker, W. Baumjohann, W. C. Feldman, J. T. Gosling, and C. T. Russell (1993),

404

Characteristics of ion flow in the quiet state of the inner plasma sheet, Geophys. Res.

405

Lett., 20(16), 1711–1714.

406

Angelopoulos, V., C. Kennel, F. Coroniti, R. Pellat, M. Kivelson, R. Walker, C. Rus-

407

sell, W. Baumjohann, W. Feldman, and J. Gosling (1994), Statistical characteristics of

408

bursty bulk flow events, J. Geophys. Res.,99(A11), 21,257–21,280.

409

Auster, H. U., K. H. Glassmeier, W. Magnes, O. Aydogar, W. Baumjohann, D. Constan-

410

tinescu, D. Fischer, K. H. Fornacon, E. Georgescu, P. Harvey, O. Hillenmaier, R. Kroth,

411

M. Ludlam, Y. Narita, R. Nakamura, K. Okrafka, F. Plaschke, I. Richter, H. Schwarzl,

412

B. Stoll, A. Valavanoglou, and M. Wiedemann (2008), The themis fluxgate magnetome-

413

ter, Space Science Reviews, 141(1–4), 235–264.

414

Balogh, A., C. M. Carr, M. H. Ac˜una, M. W. Dunlop, T. J. Beek, P. Brown, K.-H.

415

Forna¸con, E. Georgescu, K.-H. Glassmeier, J. Harris, G. Musmann, T. Oddy, and

416

K. Schwingenschuh (2001), The cluster magnetic field investigation: overview of in-

417

flight performance and initial results, Ann. Geophys., 19(10), 1207–1217.

418

Baumjohann, W., G. Paschmann, and C. Cattell (1989), Average plasma properties in

419

the central plasma sheet, J. Geophys. Res., 94(A6), 6597–6606.

420

Baumjohann, W., G. Paschmann, and H. L¨uhr (1990), Characteristics of high-speed ion

421

flows in the plasma sheet, J. Geophys. Res., 95(A4), 3801–3809.

422

Dungey, J. W. (1961), Interplanetary magnetic field and the auroral zones, Phys. Rev.

423

Lett., 6, 47–48.

424

(22)

Erickson, G. M., and R. A. Wolf (1980), Is steady convection possible in the earth’s

425

magnetotail?, Geophys. Res. Lett., 7(11), 897–900.

426

Kokubun, S., T. Yamamoto, M. H. Ac˜una, K. Hayashi, K. Shiokawa, and H. Kawano

427

(1994), The geotail magnetic field experiment, Journal of geomagnetism and geoelec-

428

tricity,46(1), 7–21.

429

McFadden, J. P., C. W. Carlson, D. Larson, M. Ludlam, R. Abiad, B. Elliott, P. Turin,

430

M. Marckwordt, and V. Angelopoulos (2008a), The themis esa plasma instrument and

431

in-flight calibration, Space Science Reviews, 141(1–4), 277–302.

432

McFadden, J. P., C. W. Carlson, D. Larson, J. Bonnell, F. Mozer, V. Angelopoulos, K.-H.

433

Glassmeier, and U. Auster (2008b), Themis esa first science results and performance

434

issues, Space Science Reviews,141(1–4), 477–508.

435

Milan, S. E., G. Provan, and B. Hubert (2007), Magnetic flux transport in the dungey

436

cycle: A survey of dayside and nightside reconnection rates, J. Geophys. Res., 112,

437

A01209, doi:10.1029/2006JA011642.

438

Mukai, T., S. Machida, Y. Saito, M. Hirahara, T. Terasawa, N. Kaya, T. Obara, M. Ejiri,

439

and A. Nishida (1994), The low energy particle (lep) experiment onboard the geotail

440

satellite : Geotail instruments and initial results. ii, Journal of geomagnetism and geo-

441

electricity, 46(8), 669–692.

442

Onsager, T., M. Thomsen, R. Elphic, and J. Gosling (1991), Model of electron and ion

443

distributions in the plasma sheet boundary layer, J. Geophys. Res., 96(A12), 20,999–

444

21,011.

445

Raj, A., T. Phan, R. P. Lin, and V. Angelopoulos (2002), Wind survey of high-speed bulk

446

flows and field-aligned beams in the near-earth plasma sheet, J. Geophys. Res., 107,

447

(23)

1419–, doi:10.1029/2001JA007547.

448

R`eme, H., C. Aoustin, J. M. Bosqued, I. Dandouras, B. Lavraud, J. A. Sauvaud, A. Barthe,

449

J. Bouyssou, T. Camus, O. Coeur-Joly, A. Cros, J. Cuvilo, F. Ducay, Y. Garbarowitz,

450

J. L. Medale, E. Penou, H. Perrier, D. Romefort, J. Rouzaud, C. Vallat, D. Alcayd´e,

451

C. Jacquey, C. Mazelle, C. D’Uston, E. M¨obius, L. M. Kistler, K. Crocker, M. Granoff,

452

C. Mouikis, M. Popecki, M. Vosbury, B. Klecker, D. Hovestadt, H. Kucharek, E. Kuen-

453

neth, G. Paschmann, M. Scholer, N. Sckopke, E. Seidenschwang, C. W. Carlson, D. W.

454

Curtis, C. Ingraham, R. P. Lin, J. P. McFadden, G. K. Parks, T. Phan, V. Formisano,

455

E. Amata, M. B. Bavassano-Cattaneo, P. Baldetti, R. Bruno, G. Chionchio, A. di Lel-

456

lis, M. F. Marcucci, G. Pallocchia, A. Korth, P. W. Daly, B. Graeve, H. Rosenbauer,

457

V. Vasyliunas, M. McCarthy, M. Wilber, L. Eliasson, R. Lundin, S. Olsen, E. G. Shelley,

458

S. Fuselier, A. G. Ghielmetti, W. Lennartsson, C. P. Escoubet, H. Balsiger, R. Friedel,

459

J.-B. Cao, R. A. Kovrazhkin, I. Papamastorakis, R. Pellat, J. Scudder, and B. Sonnerup

460

(2001), First multispacecraft ion measurements in and near the earth’s magnetosphere

461

with the identical cluster ion spectrometry (cis) experiment, Ann. Geophys., 19(10),

462

1303–1354.

463

Sergeev, V. A., R. J. Pellinen, and T. I. Pulkkinen (1996), Steady magnetospheric con-

464

vection: A review of recent results, Space Science Reviews, 75, 551–604.

465

Shiokawa, K., W. Baumjohann, and G. Haerendel (1997), Braking of high-speed flows in

466

the near-earth tail, Geophys. Res. Lett., 24(10), 1179–1182.

467

Shue, J.-H., J. K. Chao, H. C. Fu, C. T. Russell, P. Song, K. K. Khurana, and H. J. Singer

468

(1997), A new functional form to study the solar wind control of the magnetopause size

469

and shape, J. Geophys. Res., 102, 9497–9511.

470

(24)

Wang, C., L. R. Lyons, J. M. Weygand, T. Nagai, and R. W. McEntire (2006), Equatorial

471

distributions of the plasma sheet ions, their electric and magnetic drifts, and magnetic

472

fields under different interplanetary magnetic field bz conditions,J. Geophys. Res.,111,

473

A04215, doi:10.1029/2005JA011545.

474

Zhu, X. (1993), Magnetospheric convection pattern and its implications,J. Geophys. Res.,

475

98(A12), 21,291–21,296.

476

(25)

−30 −20 −10 0 10 20 30

−30

−20

−10 0 10 20 30

xGSM [R E]

yGSM [RE] Duration [days]

(a)

0 1 2 3 4 5 6 7 8 9 10

−30 −20 −10 0 10 20 30

−30

−20

−10 0 10 20 30

xGSM [R E]

yGSM [RE] Duration [days]

(b)

0.2 0.4 0.6 0.8 1 1.2 1.4 1.6 1.8 2

−30 −20 −10 0 10 20 30

−30

−20

−10 0 10 20 30

xGSM [RE]

yGSM [RE] Duration [days]

(c)

0 0.5 1 1.5 2 2.5 3

Figure 1. Observation times of β > 0.5 for (a): Cluster 1 and 3 from 2001 to 2007, (b):

Geotail from 1995 to 2006, and (c): THEMIS 1–5 from September 2007 to the end of 2009, presented as a function of GSM x and y. The gray circle represents the geostationary orbit at the 6.6 REradial distance, and the gray curve the magnetopause according toShue et al. [1997], withPdyn = 1 nPa and IMFBz= 0. Both the magnetopause and the solid grid have been rotated 4.8 clockwise to take into account the aberration due to Earth’s orbital motion. The thick gray line shows the aberrated noon-midnight axis.

(26)

−30 −20 −10 0 10 20 30

−30

−20

−10 0 10 20 30

xGSM [R E] yGSM [RE]

V

GSM,xy = 500 km/s

Duration [days]

(a)

0 1 2 3 4 5 6 7 8 9 10

−30 −20 −10 0 10 20 30

−30

−20

−10 0 10 20 30

xGSM [R E]

yGSM [RE] Duration [%]

(b)

0 10 20 30 40 50 60 70 80 90 100

−30 −20 −10 0 10 20 30

−30

−20

−10 0 10 20 30

xGSM [RE]

yGSM [RE] Duration [%]

(c)

0 5 10 15 20 25 30 35 40 45 50

−30 −20 −10 0 10 20 30

−30

−20

−10 0 10 20 30

xGSM [RE]

yGSM [RE] Duration [%]

(d)

0 5 10 15 20 25 30 35 40 45 50

−30 −20 −10 0 10 20 30

−30

−20

−10 0 10 20 30

xGSM [RE] yGSM [RE]

V

GSM,xy/V = 1

V [km/s]

(e)

0 10 20 30 40 50 60 70 80 90 100

−30 −20 −10 0 10 20 30

−30

−20

−10 0 10 20 30

xGSM [RE] yGSM [RE]

V

GSM,xy/V = 1

V [km/s]

(f)

0 10 20 30 40 50 60 70 80 90 100

(27)

Figure 2. (a): Time spent observingβ >0.5 (color). The vectors display the xy-component of the mean velocity. (b): Percentage of time spent observingVx≥0 relative to allβ >0.5 samples.

(c): Percentage of time spent observing Bz <0 relative to all β >0.5 and Vx ≥0 samples. (d):

Percentage of time spent observingBz<0 relative to allβ >0.5 andVx<0 samples. (e): Mean speed (color) and flow direction (vectors) for β > 0.5 and Vx ≥ 0 samples. The lengths of the velocity vectors have been normalized to unity, and the xy-component is displayed. (f ): Mean speed (color) and flow direction (vectors) for β >0.5 and Vx <0 samples.

(28)

−30 −20 −10 0 10 20 30

−30

−20

−10 0 10 20 30

xGSM [RE] yGSM [RE]

10x N [cm−3]

(a)

−2

−1.5

−1

−0.5 0 0.5 1 1.5 2

−30 −20 −10 0 10 20 30

−30

−20

−10 0 10 20 30

xGSM [RE] yGSM [RE]

10x

T [keV]

(b)

−1

−0.8

−0.6

−0.4

−0.2 0 0.2 0.4 0.6 0.8 1

−30 −20 −10 0 10 20 30

−30

−20

−10 0 10 20 30

xGSM [R E] yGSM [RE]

10x

B [nT]

(c)

−2

−1.5

−1

−0.5 0 0.5 1 1.5 2

−30 −20 −10 0 10 20 30

−30

−20

−10 0 10 20 30

xGSM [R E] yGSM [RE]

10x

β

(d)

−1

−0.5 0 0.5 1 1.5 2 2.5 3

−30 −20 −10 0 10 20 30

−30

−20

−10 0 10 20 30

xGSM [R E] yGSM [RE]

10x

β

(e)

−1

−0.5 0 0.5 1 1.5 2 2.5 3

−30 −20 −10 0 10 20 30

−30

−20

−10 0 10 20 30

xGSM [R E] yGSM [RE]

10x

β

(f)

−1

−0.5 0 0.5 1 1.5 2 2.5 3

(29)

Figure 3. (a): Mean ion density. (b): Mean ion temperature. (c): Mean magnetic field

strength. (d): Mean plasma β. (e): Mean β in the northern hemisphere (Br <0)(f ): Mean β in the southern hemisphere (Br >0). Only data with Vx ≥0 are included. The color scales are logarithmic.

Referanser

RELATERTE DOKUMENTER

From Figure 12 it is clear that the slope of the unit torque-unit speed at runaway ( T = 0 ) is less steep than measured, and the unit flow-unit speed slope at runaway (marked

Participants with low plasma choline concentrations had poorer cognitive performance in sensorimotor speed, percep- tual speed and executive function than subjects with high plasma

A superposed epoch analysis of the plasma sheet flows is presented in Figure 5, with the same format as Figure 1, though extended up to 4 hr after an IMF B y reversal. Since the

For earthward fast flows there is a tailward current in the inner central plasma sheet and an earthward current in the outer central plasma sheet on average.. For tail- ward fast

[ 11 ] Whether an ion escaping the polar cap ionosphere at a certain latitude is directly lost downtail into the solar wind or fed to the plasma sheet (recirculated) is thus

Below, we employ these radar measurements of the convection flow, in conjunction with our knowledge of the motion of the OCB, to estimate the dayside reconnection voltage during

However, because the metric subject to prediction is defined by the event ”this is the last registered financial statement, and bankruptcy is recorded within 3 years” a firm which

Tail reconnection allows a reconfiguration of the mag- netosphere, via convecting flux and plasma (e.g. The enhancement in earthward magne- tospheric convection subsequently