Upward high-energy field-aligned electron beams above the polar edge of auroral oval: Observations from the SKA-3 instruments onboard the Auroral Probe (Interball-2)

Abstract

A new phenomenon was found at the polar edge
of the auroral oval in the postmidnight-morning sectors: field-aligned (FA)
high-energy upward electron beams in the energy range 20–40 keV at altitudes
about 3RE, accompanied by bidirectional electron FA beams of keV
energy. The beam intensity often reaches more than 0.5·103
electrons/s·sr·keV·cm2, and the beams are observed for a
relatively long time (~3·102–103s), when the satellite at
the apogee moves slowly in the ILAT-MLT frame. A qualitative scenario of the
acceleration mechanism is proposed, according to which the satellite is within a
region of bidirectional acceleration where a stochastic FA acceleration is
accomplished by waves with fluctuating FA electric field components in both
directions.Key words. Ionosphere (particle acceleration;
wave-particle interactions) · Magnetospheric physics (magnetosphere-ionosphere
interactions)

Full text

Upward high-energy ®eld-aligned electron beams
above the polar edge of auroral oval: observations from the SKA-3
instruments onboard the Auroral Probe (Interball-2)
V. A. Stepanov1, Y. I. Galperin1, A. K. Kuzmin1, F. K. Shuiskaya1, L. S. Gorn2, B. A. Ilyin2, M. V. Iovlev2,
A. A. Klimashov2, I. I. Cherkashin2, B. I. Khazanov2, and A. Y. Safronov2
1Space Research Institute of Russian Academy of Sciences, Profsoyuznaya str., 84/32, Moscow 117810, Russia
2Scienti®c Engineering Center (SNIIP), Raspletina str., 5, Moscow 123060, Russia
Received: 25 August 1997 / Revised: 30 April 1998 / Accepted: 12 May 1998
Abstract. A new phenomenon was found at the polar
edge of the auroral oval in the postmidnight±morning
sectors: ®eld-aligned (FA) high-energy upward electron
beams in the energy range 20±40 keV at altitudes about
3R
E, accompanied by bidirectional electron FA beams
of keV energy. The beam intensity often reaches more
than 0:5103electrons=ssr keV cm2, and the beams
are observed for a relatively long time 3102±103s,
when the satellite at the apogee moves slowly in the
ILAT-MLT frame. A qualitative scenario of the accel-
eration mechanism is proposed, according to which the
satellite is within a region of bidirectional acceleration
where a stochastic FA acceleration is accomplished by
waves with ¯uctuating FA electric ®eld components in
both directions.
Key words. Ionosphere (particle acceleration; wave-
particle interactions) áMagnetospheric physics
(magnetosphere-ionosphere interactions)
1 Introduction
The Interball Auroral Probe carries the SKA-3 exper-
iment whose goal was to investigate high energy particle
intensities at high latitudes (i.e. above the auroral oval
and inside the northern polar cap). In several cases at
the polar edge of the oval in the postmidnight ± morning
sectors in the energy range 20±40 keV the SKA-3
instruments detected ®eld-aligned high-energy upward
electron beams (see Galperin, 1997; Stepanov et al.,
1997). These beams were observed as ®eld-aligned (FA)
monodirectional (upward), while simultaneous measure-
ments made by the ION spectrometer (data courtesy of
J.-A. Sauvaud and R.A. Kovrazhkin, see Sauvaud et al.,
this issue) showed that the keV energy electron ¯uxes
were bidirectional. In some cases, upward accelerated
ions Oof 100±1000 eV energy (at least partly folded
ion conics) were detected simultaneously with the SKA-
3 upward electron beam detection. For one case, upward
beam of low-energy (about 10 eV) H
+
was detected by
the onboard thermal mass-spectrometer HYPE-
RBOLOID (data courtesy of N. Dubouloz). However,
interesting features of the low-energy ions observed
simultaneously with the upward high-energy electron
beams are beyond the scope of this study.
Several authors have reported observations of the
upward electron beams in auroral regions (e.g., see
Burch et al., 1983; and references therein). Hultqvist
et al., 1988 were the ®rst to observe relatively energetic
(up to 8 keV) bidirectional ®eld-aligned electron beams
at altitudes 2R
E, i.e., above the main region of
downward auroral electron acceleration. They have
presented measurements from VIKING of upward
moving electrons and ions of comparable energy at
altitudes about 13 500 km in the invariant latitude
(ILAT) ± magnetic local time (MLT) sector of 76and
9 h, respectively, where the ®eld lines were most
probably closed, thus implying the possibility of electron
bouncing. Observations were made during quite mag-
netic conditions after a period with several substorms.
The upward electron beams detected by VIKING had a
conical distribution at energies of several keV, while at
the lowest energies they were clearly beamed, and those
electron conics were much narrower than ion conics.
The lower energy electrons had much higher ¯uxes than
those at high energies. It was suggested that the
distributions of low-energy and high-energy electron
beams were produced by dierent processes. Authors
have proposed an interpretation in which the upward
®eld-aligned acceleration of both electrons and ions was
due to the ¯uctuating electric ®eld with its mean parallel
component dierent from zero.
The SKA-3 observations of energetic upward elec-
tron beams, presented in this work, in some aspects are
Correspondence to: Y. I. Galperrin
Ann. Geophysicae 16, 1046±1055 (1998) ÓEGS ± Springer-Verlag 1998

similar to those reported by Hultqvist et al. 1988, but at
the same time are quite distinct from them. We present a
case study of upward electron beams in the energy range
20±40 keV only, i.e., at energies which are substantially
higher than typical auroral electron energies. Four
orbits in the postmidnight sector were chosen for a
detailed analysis during which the invariant latitude of
the Auroral Probe exceeded 75.
It is emphasized in this work that the upward
acceleration of electrons to tens of keV observed by
the Auroral Probe at altitudes of about 3 REat, or close
to the polar boundary of the auroral oval, cannot be
signi®cantly in¯uenced by a parallel quasi-DC electric
®eld component and conjugate eects. These observa-
tions imply acceleration processes distributed in a wide
altitude interval, both below and above the satellite
apogee 4R
Egeocentric), caused by wave turbulence
which leads to a stochastic particle acceleration. This
means that during the upward electron beam registra-
tions, the satellite at its apogee was within such extended
regions of intense plasma turbulence and auroral FA
acceleration.
2 The SKA-3 instrumentation set
The SKA-3 set of particle detectors installed onboard
the Auroral Probe is aimed to measure low and medium
energy electrons and ions. It consists of two energy-
angle analyzers EA-2 and EA-3 for low energy particles,
and two identical oppositely directed time-of-¯ight
(TOF) analyzers EM-1-1 and EM-1-2 for high energies.
A short description of the instruments is given (for more
information see Shuiskaya et al., 1995, 1998).
All instruments are mounted on the back of the
satellite, opposite to Sunward direction, with ®elds of
view perpendicular to the satellite rotation axis (which is
Sunward directed). The satellite spinning thus scans the
range of the pitch angles depending on the angle
between the Sunward direction and magnetic ®eld. The
spin period of the satellite is 120 s. Temporal resolution
of observation modes varies from 0.1 to 0.8 s, spatial
resolution at apogee is about 0:510 km.
2.1 Analyzers of low-energy charged particles
(EA-2 and EA-3)
EA-2 and EA-3 are energy-angle spectrometers based on
electrostatic toroidal analyzers. Analyzers dier in
polarity of voltages applied to systems of electrostatic
de¯ection and microchannel plates (MCP) circuitry.
EA-2 measures energy-angle distributions of electrons,
EA-3 measure ion distributions, During a satellite
rotation each window (of 22:5width with a gap of
22:5in between, and 5:7height) scans all the phase
angle range. The instruments allow us to measure
simultaneously in eight angular sectors evenly distrib-
uted at the angle 2pperpendicular to the satellite
rotation axis, and sequentially in seven energy intervals
for each angular sector. The central energies of energy
intervals are as follows: 30, 85, 240, 680, 1900, 5300,
15000 eV, the energy resolution DE/E is 10%. Particles
are recorded by means of the MCPs with spaced anodes
for eight angular directions. In order to improve the
angular resolution of the detector, a mask is applied to
the MCPs corresponding to location of anodes, and a
special sectioned collimator is attached to the entrance
aperture. The calculated geometric factor is
610ÿ3cm2sr.
2.2 Analyzers of medium-energy charged particles
(EM-1-1 and EM-1-2)
The instruments EM-1-1 and EM-1-2 are identical time-
of-¯ight solid-state energy-mass spectrometers. Elec-
trons are de¯ected from the ion aperture by a magnet.
Ions are selected with masses, 1, 4 and 16 in seven energy
intervals within energy range from 20 to 500 keV.
Spectrometers are based on a combination of TOF
device with a thick silicon semiconductor detector of the
full energy at the end. Fields-of-view of these detectors
are directed oppositely and are perpendicular to the
satellite rotation axis. Each ®eld-of-view is a cone of full
angle 8. Table 1 presents the ranges of energy intervals
for each type of particles. For the 20±40 keV electrons
the geometric factor is 5:710ÿ2cm2sr, and
310ÿ3cm sr for higher energy electrons.
3 Experimental results
Some results of observations made by the time-of-¯ight
instruments EM-1-1 and EM-1-2 (hereafter TOF ana-
lyzers) are presented here. Only some cases of the
observed phenomenon are presented, when the upward
high-energy electron beams were easily identi®able and
of long duration. There are other cases, not included
here, when the beam identi®cation was shorter, or less
reliable. The EA analyzers data as well as data from
other onboard instruments are now being used for a
more detailed analysis of those cases, and will be
reported elsewhere.
During a satellite spin period the ®eld-of-view of each
TOF analyzer scans a complete circle in a plane
perpendicular to the satellite rotation axis (which is
directed towards the Sun). Depending on the angle
Table 1. Ranges of energy intervals for EM-1-1 and EM-1-2
Interval Energy ranges, keV
eM=1
(p
+
)
M=4
(a)
M=16
(O
+
)
1 20±40 30±50 35±55 55±90
2 40±70 50±85 55±90 90±130
3 70±110 85±115 90±130 130±185
4 110±150 115±170 130±185 185±245
5 150±220 170±225 185±240 245±315
6 220±290 225±300 240±315 315±395
7 290±390 300±395 315±415 395±510
V. A. Stepanov et al.: Upward high-energy ®eld-aligned electron beams above the polar edge of auroral oval 1047

between the satellite rotation axis and the magnetic ®eld
line, each analyzer scans a limited pitch-angle range
90u;u90, where u90is a rather rare case. If
a particle beam is more or less ®eld-aligned, monodi-
rectional and crosses the rotating ®eld-of-view, it will be
seen by the instrument once in a period. Then, as ®elds-
of-view of the two analyzers point in opposite directions,
that beam will be seen by both instruments once in a
period but with a half-period shift.
Below we have used for comparisons some prelimi-
nary data from the ION onboard particle spectrometer
(courtesy of J.-A. Sauvaud and R. Kovrazhkin, see
Sauvaud et al., this issue), and data from the onboard
thermal mass-spectrometer HYPERBOLOID (courtesy
of N. Dubouloz).
3.1 Overall data description
Case 1: pass 218 (see Fig. 1). The ®rst of the events
presented here was detected during the northern polar
cap crossing on the 20 October, 1996. At 12:00 UT the
satellite location was near the apogee (altitude about
19 200 km). The invariant latitude ILAT 78:0, mag-
netic local time MLT 6 h at 12:00 UT, then the
position changed to ILAT 74:8, MLT 7 h at 12:30
UT. The Kp value during the beam observation was 2,
Dst index was about ÿ30 nT. This pass took place
during a quieting after a long disturbed period. The data
collected by both TOF analyzers revealed the presence
of a beam of 20±40 keV electrons at large pitch angles
(up to 175, so the beam moved upward from the
Earth). The beam was rather narrow, the angular width
of its most intensive portion was about 20. Maximum
beam intensity reached more than 0:5103
electrons=ssr keV cm2at 175pitch angle, while at
90it was not greater than 4 electrons=ssr keV cm2
The closest ground stations from which magnetic
®eld data were available at that time were Fort Simpson
Latitude 61:8, Longitude 238:8;MLT 3h at
12:00 UT) and Leirvogur (Latitude 64:2;
Longitude 338:3,MLT10 h at 12:00 UT). The
magnetogram from Fort Simpson indicated a series of
a moderate substorm-like bursts from 03:00 to 10:30
UT, and the moment of our observations corresponds to
a quieting period or a recovery phase of a small
substorm. The magnetogram from Leirvogur also
indicated a disturbed period from 00:00 to 08:00 UT,
and then quieting. The solar wind and IMF data from
the WIND spacecraft showed southward turning of
the IMF Bz at the time corresponding to our observa-
tions (delayed by 22 min according to the solar wind
velocity). Previous three hours of data indicated
frequent southward turnings of the IMF Bz (WIND
position at 11:40 UT was XGSE 112:4R
E;
YGSE ÿ69:8R
E;ZGSE ÿ1:6R
E, solar wind velocity
Vx ÿ535 km/s.
The beam detection lasted for 15 min (i.e. for eight
satellite rotation periods), with distinct ®eld-aligned
peaks and threshold ¯uxes around 90pitch angle. Then
at 12:17 UT both TOF instruments detected a 3-min
interval with a burst of rather isotropic ¯ux with
intensity of 102103electrons=ssr keV cm2at 20±
40 keV. The isotropic burst gradually decayed and the
¯ux returned to the ®eld-aligned beam which was more
intense than before the burst, and had some ®ne
structure around 90pitch angle. Its registration lasted
until 12:30 UT.
Simultaneous measurements made by the ION in-
strument onboard the Auroral Probe (see Sauvaud et al.,
180
90
0
1000
1000
100
100
10
10
Pitch
for EM-1-1
EM-1-1
imp/s
EM-1-2
imp/s
ALT
L
MLT
Lo
12:00:09
19165
23.4
6.0
78.0
12:07:39
19055
20.9
6.3
77.3
12:15:09
18876
18.5
6.6
76.5
12:22:39
18625
16.4
6.8
75.7
12:30:09
18304
14.4
7.0
74.7
Fig 1. Case 1, October 20, 1996). Intensities
of upward beam of electrons of 20±40 keV
by time-of ¯ight spectrometers EM-1-1 and
EM-1-2 in counts per second. In the top
panel the pitch angle for EM-1-1 is shown
(the pitch angle for EM-1-2 is shifted by half
period in respect to that of EM-1-1, see Sect.
2 for details). The middle and lower panels
show the 20±45 keV electrons data from
EM-1-1 and EM-1-2, respectively. At the
bottom there are also shown (from top to
bottom): universal time; ALT satellite alti-
tude; L-shell; MLT, magnetic local time, and
L0, invariant latitude.
1048 V. A. Stepanov et al.: Upward high-energy ®eld-aligned electron beams above the polar edge of auroral oval

this issue) indicated presence of a combination of an
isotropic plasmasheet-type background intensity with
intermittent FA counterstreaming electron beams in the
energy range of 100 1000 eV, at the time of the
upward electron beam observations from the SKA-3.
Also a beam of upward accelerated ions Hand O
of 100 1000 eV energy was seen from 11:30 till 12:34
UT. At 12:17 UT, simultaneously with the isotropizat-
ion burst observed by SKA-3, the ION spectrometer
detected a 3-minute interval of isotropic ¯ux of electrons
in the energy range of 100 1000 eV. The velocity
dispersed burst of Hions was detected at 12:17 UT in
the energy range of 800 eV±10 keV, along with the
upward beam of Hin the range of 100 1000 eV. The
superthermal ion data available (thermal mass-spec-
trometer HYPERBOLOID) did not show any signi®-
cant features which could be correlated with the
simultaneously observed upward high-energy electron
beam.
Case 2: pass 247 (see Fig. 2). At 11:00 UT on 27
October, 1996 the satellite position was ILAT 79:1
and MLT 2h;altitude about 17 800 km. The satellite
was moving to the apogee (altitude about 19 000 km), at
11:30 UT it reached ILAT 79:3and MLT 4h:
During this time interval both TOF analyzers detected
an upward beam of 20±40 keV electrons with intensity
of about 10 electrons=ssr keV cm2at pitch angles of
13020since 11:00 UT. The ¯ux around 90was
below the threshold. At 11:10 UT both instruments
simultaneously detected a 3-min isotropic intensity burst
with an intensity up to 0:5103electrons=s
sr keV cm2. Then instruments continued to detect
the upward electron beam, now with an intensity
reaching 0:6102103electrons=ssr keV cm2, and
a ®ne structure at pitch angles of 90(intensity there was
about 10 electrons=ssr keV cm2.
Simultaneous ION spectrometer measurements
showed counterstreaming electron beams in the energy
range of 40 200 eV after 11:10:30 UT and upward FA
accelerated Oions of 100 eV after 11:11 UT. At
11:10:30 UT ILAT 79:5;MLT 2:8 h) a 3-min
isotropic intensity burst of electrons of 100 1000 eV
energy was seen, and at 11:11 UT a 3-min isotropic burst
of intensity of Hions of 400 5000 eV was detected.
Before 11:10:30 UT the ¯ux of ions in all energy ranges
observed by ION was almost below the threshold, while
the electron ¯ux was detectable only in the range below
500 eV. From 11:05:30 UT till 11:09:30 UT, counter-
streaming electron ¯uxes were seen in the range of
80 500 eV. Unfortunately, for this case there were no
superthermal ion observations.
Ground magnetometer data from Fort Simpson
MLT 2:1 h at 11:10 UT) indicated a local substorm
from 07:30 to 10:30 UT, so the upward electron beam
was detected during the recovery phase of that sub-
storm. WIND IMF and solar wind data indicated a
decrease of proton number density (from 25 cmÿ3to 12
cmÿ3) along with an increase of IMF jBjmagnitude and
northward turning of the IMF Bz approximately at the
moment corresponding to our observations (delayed
according to solar wind velocity by 35 min). From 08:00
to 12:00 UT there were two broad peaks of the solar
wind proton number density up to 30 cmÿ3separated by
a period of lower density from 10:35 to 11:15 UT
(average proton number density during the day was
about 10 cmÿ3). The WIND position at 10:40 UT was
XGSE 121:7R
E;YGSE ÿ57:2R
E;ZGSE 1:9R
E, solar
wind velocity Vx ÿ369 km/s:Value of Kp at the
moment of observations was 1+, Dst index was ÿ8nT.
Case 3: pass 250 (see Fig. 3). At 03:55 UT on 28
October, 1996 both TOF analyzers indicated a weak
(less than 10 electrons=ssr keV cm2beam of
180
90
0
1000
1000
100
100
10
10
Pitch
for EM-1-1
EM-1-1
imp/s
EM-1-2
imp/s
ALT
L
MLT
Lo
11:00:08
17812
28.3
2.0
79.1
11:07:38
18222
29.9
2.5
79.4
11:15:08
18560
30.5
3.0
79.5
11:22:38
18826
30.2
3.5
79.5
11:30:08
19021
29.0
3.9
79.3
Fig 2. (Case 2, October 27, 1996). For
details see Fig. 1
V. A. Stepanov et al.: Upward high-energy ®eld-aligned electron beams above the polar edge of auroral oval 1049

20±40 keV electrons a pitch angles about 13020,
with almost no counts above the threshold around 90
of pitch angle. At 04:03 UT the TOF instruments
detected a 3-min isotropic burst with an intensity of
0:5102electrons=ssr keV cm2. Then till 04:21 UT
an upward electron beam was seen with peaks of the
same intensity at pitch angles of 12525. Around 90
the observed ¯ux was ®ne-structured, while the intensity
did not exceed 4 electrons=ssr keV cm2:
At 04:21 UT both TOF analyzers detected an 11-min
isotropic intensity burst with intensities up to 0 5103
electrons=ssr keV cm2with some ®ne structure. Af-
ter that until 04:39 UT an upward electron beam was
seen with rather distinct peaks at pitch angles of
13515intensity 0:5103electrons=ssr keVcm2.
At 90the intensity was rather weak-number ¯ux not
greater than 10 electrons=ssr keV cm2. Other parti-
cle instruments of the Auroral Probe either were not
operating at this time or the data is still unavailable.
The satellite position at 03:55 UT was 15 650km
altitude, MLT 2:5 h and ILAT 78:5;by04:39 UT
the satellite moved to an altitude of 18 600 km,
MLT 5:5 h and ILAT 77:3.Kp value was 3-, Dst
was about )3 nT. The magnetogram from the nearest
ground station Leirvogur MLT 3:5hindicated a
moderate disturbance or possible recovery phase of a
small substorm. The WIND data indicated southward
turning of the IMF Bz at the moment roughly corre-
sponding to the observation time, shifted by the delay
calculated from the solar wind velocity (26 min). The
WIND position at 10:40 UT was XGSE 122:1R
E,
YGSE ÿ55:5R
E,ZGSE 2:3R
E, solar wind velocity
Vx ÿ500 km/s.
Case 4: pass 251 (see Fig. 4). The last case we consider
here took place at 10:12 UT on 28 October, 1996. After
10:12 UT for 40 min the TOF analyzers detected an
upward beam of 20±40 keV electrons with intensities up
to 0:5102electrons=ssr keV cm2, pitch angles about
13515. Then at 10:52 UT (ILAT
79:1;MLT 4:8hsome ¯ux intensi®cation began,
up to 0:5103electrons/s sr keV cm2, along with
isotropization and intensi®cation of ¯ux around pitch
angles of 90. From 11:16:50 to 11:23:45 UT
76:9<ILAT <76:2,5:8<MLT <6:0h, distinct
peaks at maximum pitch angles were seen and almost
no detectable ¯ux at 90of pitch angle, after that a 3-
min isotropic burst was seen.
Other particle instruments observations are available
only after 11:00 UT. The ION particle spectrometer
detected a rather isotropic ¯ux of 100 1000 eV elec-
trons until 11:13 UT. Then ION detected a series of
isotropic 100 1000 eV electron bursts: from 11:16 till
11:17, from 11:19 till 11:21, from 11:24 till 11:27 UT.
After that the electron ¯ux became isotropized. Several
velocity dispersed H
+
structures below 8 keV were also
detected by ION from 11:03 till 11:16 UT, and then
after 11:24 UT. An isotropic H
+
burst in the energy
range 100 3000 eV was detected by ION from 11:16
till 11:18 UT. In addition, an upward 100 eV O
+
beam
was detected from 11:00 till 11:53 UT. The measure-
ments of the thermal plasma analyzer HYPERBOLOID
showed an upward beam of 10 eV H
+
from 11:00 till
11:55 UT, slightly isotropized from 11:37 till 11:45 UT.
From the WIND data for the case an unstable IMF
Bz behavior is seen: there was a 10-min southward IMF
Bz turning near the moment corresponding to the
observation time shifted by the appropriate delay
(22 min). WIND position at 10:40 UT was
XGSE 122:1R
E,YGSE ÿ54:7R
E,ZGSE 2:5R
E, solar
wind velocity Vx ÿ585 km=s. The Auroral Probe
position was changing from 18 000 km altitude, MLT
2 h and ILAT 80to 18 200 km, MLT 6:2 h and
180
90
0
1000
1000
100
100
10
10
Pitch
for EM-1-1
EM-1-1
imp/s
EM-1-2
imp/s
ALT
L
MLT
Lo
03:55:08
15680
25.4
2.5
78.5
04:06:20
16691
26.6
3.4
78.8
04:17:32
17523
25.5
4.3
78.5
04:28:44
18183
23.3
4.9
78.0
04:39:56
18677
20.6
5.5
77.2
Fig 3. (Case 3, October 28, 1996). For
details see Fig. 1
1050 V. A. Stepanov et al.: Upward high-energy ®eld-aligned electron beams above the polar edge of auroral oval

ILAT 75. During the beam registration the satellite's
orbit crossed its apogee at 11:03 UT, MLT 5:4h,
ILAT 78:5. At the moment of observations Kp was
3-, Dst about ÿ22 nT. Ground magnetometer data from
College Latitude 64:8, Longitude 212:2;MLT
0hand Fort Simpson MLT 1:3hindicated that
the beam was detected during a quieting period between
two substorm-like bursts, or just before the breakup
phase of a local activation.
3.2 Observations summary
As can be seen from the data description, the four cases
detected were located from 2 to 7h MLT and from 75
to 80ILAT near the apogee of the Auroral Probe orbit
(i.e. at the polar edge of the auroral oval and/or inside
the northern polar cap), during quieting, or moderate
disturbance levels. ION particle spectrometer measure-
ments onboard the Auroral Probe (Sauvaud et al., 1997,
this issue) for Case 2 also show that the satellite
probably crossed the poleward auroral oval boundary
during, or before, the beam registration. Observations
revealed the existence of narrow ®eld-aligned upward
beams of energetic electrons in that region interrupted
by isotropic intensity burst of about 3 min duration,
after which the upward beam intensity increased. At the
same time, in the keV energy range, counterstreaming
FA electron beams of variable intensity (generally much
higher than that of the upward beam of 20±40 keV)
were seen simultaneously with nearly isotropic intensity
of plasmasheet-like electrons. Sometimes upward beams
of low-energy ions (probably, conics narrowed in pitch
angle due to magnetic moment conservation) were seen
more or less simultaneously with the upward high-
energy electron beam. From a limited set of the data
available it is not clear whether the presence of such
energetic upward electron beams is caused by some
distinct signature in the IMF and solar wind. Cases 1, 3
and 4probably correspond to the southward turnings
of the IMF Bz, but Case 2 seems to be detected during
northward turning of the IMF Bz. Global magneto-
spheric conditions were not very disturbed, as is shown
by Kp and Dst values, and loosely could be described as
quieting periods. Ground magnetograms from the
nearby stations correspond to intervals after or between
disturbances of moderate activity levels, or probably to
a recovery or late recovery phase of a small substorm-
like bursts. There is no clear evidence that the events
observed are a product of a preceding substorm
activity. However this is not inconsistent with the data
available.
Simultaneously with the energetic upward beam
observations by SKA-3, measurements performed by
the ION particle instrument show variable counter-
streaming electron beams of 100 1000 eV embedded in
the plasmasheet-like background. An isotropic but ®ne-
structured intensity bursts of 100 1000 eV electrons
and Hions of up to some keV energy were observed
after several minutes of the upward beam appearance at
least for Cases 1 and 2. In addition an upward acceler-
ated ion beam of keV energy was observed in Case 1,as
well as a set of velocity-dispersed H
+
structures below 8
keV and upward 100 eV O
+
beam for Case 4. An upward
beam of 10 eV H
+
was also detected by the thermal
mass-spectrometer HYPERBOLOID in Case 4.
Further studies of features accompanying the upward
high-energy electron beams require detailed analysis of
simultaneous observations performed by other particle
detectors onboard the Auroral Probe such as ION,
PROMICS-3 and DOK-2, as well as wave and electric
®eld instruments and onboard magnetometers. The
work of data collection and analysis during the upward
high-energy electron beam registrations and some mod-
eling eorts are under way and will be reported
elsewhere.
In the next section we propose a qualitative scenario
of the mechanism which could be responsible for the
180
90
0
1000
1000
100
100
10
10
Pitch
for EM-1-1
EM-1-1
imp/s
EM-1-2
imp/s
ALT
L
MLT
Lo
10:10:09
17826
33.6
2.1
80.0
10:20:52
18384
35.3
2.9
80.3
10:31:35
18797
34.5
3.6
80.2
10:42:17
19066
31.8
4.3
79.8
10:53:00
19193
28.1
4.8
79.1
11:03:43
19180
24.2
5.3
78.2
11:14:26
19027
20.5
5.8
77.2
11:25:09
18732
17.2
6.1
76.0
Fig 4. (Case 4, October 28, 1996). For details see Fig. 1
V. A. Stepanov et al.: Upward high-energy ®eld-aligned electron beams above the polar edge of auroral oval 1051

formation of energetic upward electron beams in that
ILAT-MLT region above the polar ionosphere.
4 An interpretation scenario
4.1 Features of the upward high-energy electron beams
In an attempt to construct a plausible theoretical scheme
for explanation of the observed FA upward beams of
energetic electrons, the following morphological prop-
erties of the phenomenon will be taken into account:
1. The high-energy beams 20 ÿ40 keVare observed
for a relatively long time  3102±103s, when the
satellite stays almost at the same place, or moves
slowly in the ILAT-MLT frame. Thus they may be
considered as quasi-stationary.
2. The high-energy FA upward electron beam is nearly
monodirectional, while the simultaneously present
electron FA beams of keV energy are observed as
bidirectional and highly variable. In some cases high-
energy electron ¯uxes at 90also appear, but are
much weaker than the FA upward beam.
3. The ILAT-MLT location of the beams at, or close to,
the polar border of the oval in the postmidnight-
morning MLT sectors is consistent with a region of
large-scale downward FA current (region I). But the
preliminary onboard magnetometer data (courtesy of
V. Styazhkin, IMAP-3 experiment) are inconclusive
to prove the existence of the downward current.
Recent results by FAST group (see, e.g., Carlson
et al., 1998, and references therein) show high wave
activity and intermittent small-scale potential drops
in the downward current region. The inference about
the downward FA current at the upward beam
locations is also consistent with the ``Matreshka''
scheme of nested FA current loops at the polar edge
of the postmidnight oval (Timofeev and Galperin,
1991).
4. Simultaneous but preliminary data on ELF plasma
waves (courtesy of M. Mogilevsky, IESP instrument)
indicate that while some wave activity is usually
present during an event, no straightforward correla-
tions have been found so far.
4.2 A qualitative scenario
In view of these characteristics, a scenario for the high-
energy acceleration phenomenon is proposed. In this
scenario, during registration of the upward energetic
electron beam the satellite is supposed to be within a
bidirectional electron acceleration region, or at least, the
FA acceleration of electrons occurs both above and
below the satellite. This is evidenced by the counter-
streaming low-energy electron beams embedded in the
plasmasheet-like particle environment. Thus the FA
low-energy electrons  0:11 keV, reaching the sat-
ellite from both directions, have been accelerated by
some wave-particle interactions while crossing a part of
the acceleration region lying respectively below or above
the satellite (``single transit'').
The upward beam of high-energy electrons is pre-
sumed here to originate from the acceleration of a part
of the FA accelerated plasmasheet electrons. Before the
formation of the upward beam, those electrons have
passed downward through the whole acceleration region
both above and then below the satellite (``double
transit''). As a result of the FA acceleration they formed
electron beams, mostly reaching auroral altitudes, with
excitation of respective auroral phenomena, probably in
form of diuse (or weakly structured) features, some-
times with additional activation. The portion of the
accelerated electron distribution lying outside the loss
cone, after being mirrored below the acceleration region
(but above the ionosphere), and now moving upward, is
able to reach the satellite again after passing the lower
part of the acceleration region (``triple transit''). Thus
particles that formed an upward beam of high-energy
electrons have experienced much more eective acceler-
ation in a ``triple transit'' than those registered at the
satellite as counterstreaming electron beams which
experienced only a ``single transit'', or auroral electrons
at the footprint of the satellite that have experienced a
``double transit'' through the region of stochastic wave-
particle acceleration within the downward FA current
region. This scenario is shown schematically in Fig. 5.
4.3 On possible acceleration mechanisms
Consider now a possible mechanism for the particle
acceleration process. For the scenario discussed it must
be bidirectional, i.e., in these wave-particle interactions,
Plasma sheet electrons
"single transit""single transit"
"double transit"
"triple transit"
Plasma turbulence
FA acceleration region
FA ionospheric electrons
(downward FA current)
Ionosphere
Fig 5. Schematics of the acceleration scenario
1052 V. A. Stepanov et al.: Upward high-energy ®eld-aligned electron beams above the polar edge of auroral oval

the acceleration is experienced by electrons moving either
upward or downward through the layer of wave-particle
interactions. For that, a stochastic type of FA acceler-
ation is suitable if accomplished by waves with FA
electric ®eld components in either direction. At the same
time, as no comparable acceleration eects are observed
for ions, and considering the available wave and other
particle data onboard, some restrictions probably may
be devised for the choice of the wave-particle interactions
involved. This work is in progress now.
According to our observations, the acceleration by
wave-particle interactions is present in a wide range of
altitudes along the particle trajectory. It occurs both
above and below the satellite and is not con®ned in the
altitude region of 0.5±2 RE, where most of the auroral
FA acceleration takes place. It is not clear at what
altitudes above the satellite the FA downward acceler-
ation of the plasmasheet electrons occurs. It must be
con®ned at most to 6R
E, because otherwise the
magnetic moment conservation tendency will isotropize
the electron distribution more than is observed. At the
same time, the observed weakness of the high-energy
electron intensities around 90pitch angle in compar-
ison to the FA component indicates that the perpendic-
ular acceleration process by plasma waves is not local.
Thus, we cannot expect a good correlation of the
upward beam intensity with some particular type of
waves registered by the Auroral Probe. A search for
such waves, both from measurements on the Auroral
Probe and from theoretical analysis, is in progress.
Turbulent acceleration processes in the aurora have
been discussed in many papers (see, e.g., Bryant et al.,
1991; Tetreault et al., 1995), however, no de®nite
comparisons with the data are known so far. Intense
IC and LH wave packets and their interaction with
particles have been studied using a kinetic approach by
many authors (see, e.g., Andre and Chang, 1992;
Kintner et al., 1992; Retterer et al., 1994; and recently
Dubouloz et al., 1995 for LHR wave packets). Suppos-
ing the initial plasmasheet electron temperature of 0.5
keV, model results of Dubouloz et al., 1995, show the
possibility for such electrons to be FA accelerated up to
several keV due to LH wave packets in a single particle
transit through the upper part of the acceleration region
(above the satellite), or through its lower part (below the
satellite), if LH waves of sucient intensity are present
there. These accelerated electrons will then constitute
the observed FA counterstreaming beams of about
1keV. However, the resonance condition for high-energy
electrons of the upward beam requires very large wave
phase velocities hardly achievable by the LH wave
packets. But much higher energies, up to 20 or more
keV, are observed to be attainable by a small part of the
initial plasmasheet electron population which is pre-
sumed here to mirror above the ionosphere after
traversing the whole accelerating layer, forming an
upward accelerated energetic electron beam. Thus, the
theory outlined already is unable to explain observations
fully by itself and additional mechanisms are needed.
The described scenario seems to be qualitatively
consistent with observations of the energetic upward
electron beams as described, while the nature of the
wave-particle interactions involved remains obscure.
According to the scenario, the observed energetic
upward beam is the extended tail of the FA accelerated
electrons of lower energies, presumably with a power
law spectral form. The accompanying feature expected
from the scenario is a wave turbulence in extended
regions somewhere above, and especially below the
satellite, but not necessarily at its altitude. In addition,
for the LHR wave packets e.g., a perpendicular ion
heating and thus ion conic formation is another expect-
ed consequence. We are now studying the available data
with respect to these inferences.
5 Discussion
5.1 Location in magnetosphere and time development
We have concentrated here on a few cases of long
duration events when, during many satellite rotations (2
min each), both oppositely directed TOF analyzers
indicated the presence of an upward high-energy
electron beam. At this time no reliable statistics of such
cases could be presented.
The four cases described occur in the postmidnight-
morning sector of MLT near, or at, the polar edge of the
auroral oval, or within the polar diuse auroral zone
(see, Feldstein and Galperin, 1985, 1996; Galperin and
Feldstein, 1991). These magnetic ¯ux tubes in stationary
conditions are expected to be traced to, or close to, the
open-closed magnetic ®eld boundary, or the distant
reconnection line. A suggestion about an extended
region in the distant tail with turbulent magnetic ®eld
as an alternative to the single distant reconnection line
was recently put forward by Galperin, (1995) and
Galperin and Feldstein 1996. However, the conditions
in which the four events studied here took place cannot
be considered as stationary states of the magnetosphere.
They all belong to some quieting period after, or
between, substorm-like disturbances or activations,
when the distant tail still could not be equilibrated,
and ®laments of plasma with dierent properties could
extend from the plasmasheet to tail lobes, or plasmoid(s)
in the tail could be still present, causing variable FA
currents in the regions of beam observations. As there is
no returning high-energy electron beam in these cases, a
quasi-regular closed magnetic ®eld along the beam
plasma ¯ux tube is excluded (in contrast to the late
morning case described by Hultqvist et al., 1988). Thus,
either an open magnetic ®eld, or a turbulent weak
magnetic ®eld in the distant tail where strong scattering
of the high-energy electrons occurs, seem to be the
alternatives for these plasma ¯ux tubes.
Now let us consider the time development of the
observed energetic electron beams and ambient plasma
populations. In the four cases described qualitatively a
very similar behavior was found: a weak upward beam
of 20±40 keV electrons after several minutes was
followed by an isotropic intensity burst that lasted for
about 3 min. It occurred simultaneously with the
V. A. Stepanov et al.: Upward high-energy ®eld-aligned electron beams above the polar edge of auroral oval 1053

isotropic burst of electrons and ions in the keV energy
range. After the burst the quasi-steady high-energy
electron beam reappeared and was more intense than
before the burst. We note that the 3 min interval of the
burst is a characteristic time for the transit of protons of
0.3±3 keV from the distant tail region at 50±160 RE.
Thus, the origin of the burst in some active processes in
the far tail seems plausible, and deserves further
analysis.
5.2 A possible relation to energetic particle bursts in the
distant tail
The Auroral Probe orbit is chosen so that the satellite at
the apogee moves only very slowly in the ILAT-MLT
frame for a relatively long time (20 min), thus the time
dependence of the events can be distinguished from the
data more easily. This was the case during the registra-
tion of the upward electron beams. As could be seen
from our case studies, there is each time a period of a
relatively weak ¯ux (as long as several satellite spin
periods of 120 s) before an intensity burst and
isotropization interval followed by subsequent upward
beam intensi®cation.
Sarris et al., 1996, have reported Geotail observations
of the high-energy electron and ion bursts in the deep
magnetotail at a distance of XGSM ÿ128 RE. The
source location of these high-energy particles is reported
to be at ÿ103 RE. Then it could be of the same type as
the isotropic intensity bursts observed by the SKA-3
after some time of quasi-steady upward beam existence.
However, no sign of the preceding weak high-energy
beam intensity was noted in Sarris et al.'s 1996, study.
We note that if the source region was in the closed ®eld
line region, or a reconnection region, such a location
could be mapped to the ILAT-MLT region of our
observations. Thus energetic particle bursts in the tail
during the SKA-3 observations cannot be excluded, and
at least, in some cases, such checks will be possible from
simultaneous Tail Probe data.
5.3 A speculation concerning possible local destabilization
of the cross-tail current by high-energy electrons of the
upward beam
It is known that bursts of energetic electrons occur in the
distant tail (see Lui and Krimigis, 1984; Sarris et al.,
1996). It has been shown by Sitnov et al., 1996, that the
loss cone electron population could aect signi®cantly
the linear stability conditions of the tearing mode in the
quasi-neutral magnetotail current sheet. So it might be
speculated from this limited data set that the initial
upward energetic electron beam (even while rather
weak) could aect the distant tail stability conditions,
as the ®eld lines on which the measurements were
performed map to more than ÿ70 REgeocentric ac-
cording to the Tsyganenko-87 model. Supposedly, the
induced local destabilization of the tearing mode in the
distant tail plasma and currents could cause the
observed isotropic burst of energetic particles which is
detected by the SKA-3 instruments. Presumably the
plasma turbulence level at altitudes below and above the
satellite, which is responsible for the acceleration of
high-energy electrons, increases after the particle burst,
and qualitatively it can cause the upward beam inten-
si®cation after the burst.
One possibility to signi®cantly alter locally the cross-
tail current in the distant tail is to inject particles there
with a Larmor radius comparable to the local magnetic
®eld line curvature in the neutral sheet. According to the
data from the ISEE-3 and GEOTAIL, the average
magnetic ®eld in the neutral sheet at 100 REis about
0.3 nT (Fair®eld, 1992, 1993; Nishida et al., 1995; see
also Lui and Krimigis, 1984) and the curvature radius
can be less than 2 RE. For a 25 keV electron the
respective Larmor radius amounts then to about 1:7R
E,
so that its strong scattering due to non-adiabatic motion
is quite feasible. It may be supposed that, due to
appearance of new charge carriers with a dierent type
of particle motion in the cross-tail current sheet, the
charge neutrality and cross-tail current characteristics
could be locally disturbed. It is speculated here that this
disturbance could lead to the plasma particle burst
observed in the middle of the upward beam observa-
tions.
Clearly this point is just a speculation based on the
medium-altitude measurements and requires further
analysis and comparisons with the data from other
satellites and ground-based observations.
6 Summary and conclusions
1. A quasi-stable upward ®eld-aligned electron beam
with super-auroral energies (20±40 keV) was discovered
at altitudes 3R
E, high above the post-midnight polar
border of auroral oval (or at, or near, the open-closed
magnetic ®eld boundary) during quieting conditions.
2. From four case studies presented, the time devel-
opment of the beam starts from a low monodirectional
upward ¯ux, but after 5±15 min an isotropic intensity
burst appears, decays in 3 min and the upward
monodirectional beam reappears with higher intensity
than before the burst.
3. The appearance of bidirectional electron beams at
altitudes 3R
Eduring the upward high-energy electron
beams demonstrates that electron FA acceleration
processes, presumably by the wave-particle interactions,
can encompass a very extended altitude range, up to the
Auroral Probe apogee and higher, at least for the cases
considered.
4. A qualitative scenario is proposed for FA accel-
eration of the tail of the plasmasheet electron distribu-
tion function both above and below the satellite altitude,
by electron velocity diusion due to unspeci®ed wave-
particle interactions. The scenario includes additional
acceleration of the mirrored, upward moving, electrons,
which gives them higher energy than those reaching
auroral altitudes.
1054 V. A. Stepanov et al.: Upward high-energy ®eld-aligned electron beams above the polar edge of auroral oval

5. A local destabilizating action is suggested from the
high-energy FA electron beam for the cross-tail current
in the distant tail plasma sheet as a possible cause of the
isotropic intensity burst observed after several minutes
of the upward beam persistence.
Acknowledgements. The Interball Project was accomplished with
in the frame of the contract with the Russian Space Agency, N025-
7535/94. The work of V. Stepanov, Y. Galperin, F. Shuiskaya and
A. Kuzmin was supported by grants from Russian Foundation for
Basic Research (RFBR) N 94-02-04299 and 97-02-16333; and by
grant INTAS N 94±1695. We thank Russian Space Agency,
Lavochkin Space Association and Babakin Space Center for their
continuous eorts which led to the success of the Interball project.
We wish to thank N. Dubouloz, J.-J. Berthelier, A. Volosevich and
L. Zelenyi for many valuable discussions, N. Dubouloz (HYPE-
RBOLOID thermal mass-spectrometer PI), J.-A. Sauvaud and R.
Kovrazhkin (ION particle instrument PIs), M. Mogilevsky and S.
Perraut (IESP instrument Pis), V. Styazhkin (IMAP-3 onboard
magnetometer PI) for their permission to use, respectively, the
preliminary HYPERBOLOID, ION, IESP and IMAP-3 data in
our analysis. Also we would like to thank R. Lepping and K.
Ogilvie from NASA/GSFC, WIND magnetic ®eld and solar wind
experiments PIs, for the key magnetic ®eld and solar wind
parameters available at http://rumba.gsfc.nasa.gov/cdaweb/; and
NGDC sta made available the ground magnetometers data from
College, Fort Simpson and Leirvogur stations at the WWW site
http://www.ngdc.noaa.gov:8080/production/html/GEOMAG/geo-
ÿsearchÿframes.html.
Topical Editor K.-H. Glassmeier thanks H. E. J. Koskinen and
J. Retterer for their help in evaluating this paper.
References
Andre, M., and T. S. Chang, Ion heating perpendicular to the
magnetic ®eld, in Physics of Space Plasmas, Number 12, Eds.
Chang, T. and J. R. Jasperse, Scienti®c Publishers, Inc.,
Cambridge, Mass, USA, 35±71, 1992.
Bryant, D. A., A. C. Cook, Z.-S. Wang, U. de Angelis, and C. H.
Perry, Turbulent acceleration of auroral electrons, J. Geophys.
Res.,96, 13 829±13 839, 1991.
Burch, J. L., P. H. Rei, and M. Sugiura, Upward electron beams
measured by DE-1: a primary source of dayside region-1
Birkeland currents, Geophys. Res. Lett., 10, 753±756, 1983.
Carlson, C. W., J. P. McFadden, R. E. Ergun, M. Temerin, W.
Peria, F. S. Mozer, D. M. Klumpar, E. G. Shelley, W. K.
Peterson, E. Moebius, R. Elphic, R. Strangeway, C. Cattell, and
R. Pfa, FAST observations in the downward auroral current
region: energetic upgoing electron beams, parallel potential
drops, and ion heating, Geophys. Res. Lett., 25, 2017±2020,
1998.
Dubouloz, N., R. A. Treumann, R. Pottelette, and K. A. Lynch,
Electron acceleration from localized lowerhybrid waves, Geo-
phys. Res. Lett., 22, 2969±2972, 1995.
Fair®eld, D. H., On the structure of the distant magnetotail: ISEE
3, J. Geophys. Res., 97, 1403±1410, 1992.
Fair®eld, D. H., Solar wind control of the distant magnetotail:
ISEE 3, J. Geophys. Res., 98, 21 265±21 276, 1993.
Feldstein, Y. I., and Y. I. Galperin, The auroral luminosity structure
in the high-latitude upper atmosphere: its dynamics and
relationship to the large-scale structure of the Earth's magne-
tosphere, Rev. Geophys. Space Phys., 23, 217±275, 1985.
Feldstein, Y. I., and Y. I. Galperin, Structure of the auroral
precipitations in the nightside sector of the magnetosphere,
Cosmic Res. 34, 209±227, 1996.
Galperin, Y. I., Magnetospheric tail structure: concepts, problems,
and storm-time development of the auroral oval, J. Atmos. Terr.
Phys. 57, 1397±1414, 1995.
Galperin, Y. I., Polar magnetosphere during a solar minimum ± a
look from INTERBALL-2 (Abstr). Annales Geophysicae, 1.15
(3) C 678, 1997.
Galperin Y. I., and Y. I. Feldstein, Auroral luminosity and its
relationship to Magnetospheric plasma domains. In Auroral
Physics, Eds. Meng, C.-I., M. J. Rycroft, and L. A. Frank,
Cambridge UP, Cambridge, 207±222, 1991.
Galperin, Y. I., and Y. I. Feldstein, Mapping of the precipitation
regions to the plasma sheet, J. Geomag. Geoelectr., 48, 857±875,
1996.
Hultqvist, B., On the acceleration of electrons and positive ions in
the same direction along magnetic ®eld lines by parallel electric
®elds, J. Geophys. Res., 93, 9777±9784, 1988.
Hultqvist, B., R. Lundin, K. Stasiewicz, L. P. Block, P.-A. Lindkvist,
G. Gustafsson, H. Koskinen, A. Bahnsen, T. A. Potemra, and L.
J. Zanetti, Simultaneous observation of upward moving ®eld-
aligned energetic electrons and ions on auroral zone ®eld lines,
J. Geophys. Res., 93, 9765±9776, 1988.
Kintner, P. M., J. L. Vago, S. W. Chesney, R. L. Amoldy, K. A.
Lynch, C. J. Pollock, and T. E. Moore, Localized lower hybrid
acceleration of ionospheric plasma, Phys. Rev. Lett., 68, 2448±
2451, 1992.
Lui, A. T. Y., and S. M. Krimigis, Association between energetic
particle bursts and Birkeland currents in the geomagnetic tail, J.
Geophys. Res., 89, 10741±10748, 1984.
Nishida, A., T. Mukai, Y. Saito, S. Kokobun, and K. Maezawa,
GEOTAIL observation of magnetospheric convection in the
distant tail at 200 REin quiet times, J. Geophys. Res., 100,
23663±23675, 1995.
Retterer, J. M., T. S. Chang, and J. R. Jasperse, Transversely
accelerated ions in the topside ionosphere, J. Geophys. Res., 99,
13189±13201, 1994.
Sarris, E. T., V. Angelopoulos, R. W. McEntire, D. J. Williams, S.
M. Krimigis, A. T. Y. Lui, E. C. Roelof, and S. Kokobun,
Detailed observations of a burst of energetic particles in the
deep magnetotail by GEOTAIL, J. Geomag. Geoelectr. 48, 649±
656, 1996.
Shuiskaya, F. K., A. K. Kuzmin, Y. I. Galperin, V. A. Stepanov, R.
A. Kovrazhkin, L. S. Gom, B. A. Ilyin, M. V. Iovlev, A. A.
Klimashov, I. I. Cherkashin, B. I. Khazanov, and A. Y. Safronov,
Measurements of charged particles by means of the experiment
SKA-3, in: Interball Mission and Payload, Eds. Galperin, Y., T.
Muliarchik, and J.-P. Thouvenin, RKA, IKI, CNES, Paris,
262±267, 1995.
Shuiskaya, F. K., Y. I. Galperin, R. A. Kovrazhkin, A. K. Kuzmin,
V. A. Stepanov, L. S. Gorn, B. I. Khazanov, Measurements of
energetic charged particles at high latitudes: the SKA-3 exper-
iment on the Auroral probe (INTERBALL-2) satellite, in:
Cosmic Res. 36, 81±91, 1998.
Sitnov, M. I. H. V. Malova, and A. S. Sharma, Role of the
temperature ratio in the linear stability of the quasineutral sheet
tearing mode, Geophys. Res. Lett., 25, 269±272, 1998.
Stepanov, V. A., F. K. Shuiskaya, A. K. Kuzmin, and A. A. Skalsky,
Energetic particle beams at about 3 REabove the high-latitude
polar cap (Abstr.), Annales Geophysicae, 15(3), C 687, 1997.
Tetreault, D. J., Theory of electric ®elds in the auroral acceleration
region, J. Geophys. Res., 96, 3549±3563, 1991.
Timofeev, E. E., and Y. I. Galperin, Convection and currents in
stable auroral arcs and inverted-V's. J. Geomag. Geoelectr., 43,
Suppl., 259±274, 1991.
V. A. Stepanov et al.: Upward high-energy ®eld-aligned electron beams above the polar edge of auroral oval 1055