JOURNAL OF GEOPHYSICAL RESEARCH, VOL. 107, NO. 0, XXXX, doi:10.1029/2001JA007553, 2002
Relationship between multiple substorm onsets and the IMF:
A case study
C.-C. Cheng1 and C. T. Russell
Institute of Geophysics and Planetary Physics, University of California, Los Angeles, California, USA
M. Connors
Centre for Science, Athabasca University, Athabasca, Canada
P. J. Chi
Institute of Geophysics and Planetary Physics, University of California, Los Angeles, California, USA
Received 29 August 2001; revised 7 March 2002; accepted 8 April 2002; published XX Month 2002.
[1] The relationship between a series of Pi2 pulsations and their associated substorm
onsets with the IMF is examined on 5 May 1999. The magnetograms from the
CANOPUS array show the occurrence of three substorm onsets, accompanied by Pi2
bursts. Analysis of the auroral electrojet location and current intensity as deduced from the
Churchill line in the CANOPUS array shows that the following onsets move poleward.
ACE and Wind magnetic field data show that the first onset occurs during a period of
weak southward interplanetary magnetic field (IMF) and that the second onset occurs
about an hour later when the IMF has become slightly northward. The third onset occurs
during strongly northward IMF. Magnetic disturbances at GOES 8 and GOES 10 confirm
that each onset is accompanied by disturbances in the field like those expected from the
substorm current wedge. Meanwhile, from the IMP 8 observations in the magnetotail, the
decreases of the total and X components of the magnetic field imply that significant
removal of magnetic flux from the tail has occurred after the third onset. Thus a single
cycle of dayside reconnection transport of open flux and reconnection of that open flux
may contain within it multiple onsets of tail reconnection and dipolarization of the night
INDEX TERMS:
magnetosphere, some possibly not involving the open flux in the lobes.
2788 Magnetospheric Physics: Storms and substorms; 2752 Magnetospheric Physics: MHD waves and
instabilities; 2740 Magnetospheric Physics: Magnetospheric configuration and dynamics; KEYWORDS:
substorm onset, Pi2 pulsations, IMF control
Citation: Cheng, C.-C., C. T. Russell, M. Connors, and P. J. Chi, Relationship between multiple substorm onsets and the IMF: A case
study, J. Geophys. Res., 107(0), XXXX, doi:10.1029/2001JA007553, 2002.
1. Introduction
[2] Pi2 pulsations, in the period range, 40 to 150 seconds
s, are impulsive and damped oscillations of the geomagnetic
field. This type of pulsations is predominantly a nighttime
phenomenon and associated with substorm onset [see
review by Yumoto, 1986; Olson, 1999]. Earlier studies
reported that Pi2 generally occurs in the expansive phase
of a magnetic bay and tends to occur successively to form a
group of pulsations [cf. Saito, 1969]. Rostoker [1968] found
that a substorm often has two Pi2 pulsations and two
individual ‘‘bays’’ in the horizontal component of the
magnetic field that he denoted as the trigger bay and the
main bay. Subsequent study by Kisabeth and Rostoker
1
Also at Department of Physics, National Huwei Institute of
Technology, Hu-Wei, Taiwan.
Copyright 2002 by the American Geophysical Union.
0148-0227/02/2001JA007553$09.00
SMP
[1971] showed some evidence of multiple onsets accompanied by Pi2 pulsations in a magnetospheric substorm. In
addition, using the magnetic contour maps from midlatitude
observations, Clauer and McPherron [1974] found that two
onsets occurred in a single magnetospheric substorm. More
recently observations by Mishin et al. [2000, 2001] also
reported the occurrence of two distinct onsets in a magnetospheric substorm on the basis of ground data. They speculate on the causes of the two onsets but this speculation
needs to be verified with data obtained closer to the onset
region.
[3] Recently, Russell [2000] extended the near-Earth
neutral point model of substorms to emphasize the role of
the distant neutral point. This model attempts to explain
both the observations of Mishin et al. [2000, 2001] and how
substorm onset could be triggered by the northward turning
of the interplanetary magnetic field (IMF). In the model, the
interplay between near Earth and distant neutral points in
the magnetotail creates two onsets, one when reconnection
X-1
SMP
X-2
CHENG ET AL.: CONSECUTIVE Pi2 BURSTS AND SUBSTORM ONSETS
[2002] analyzed consecutive bursts of low-latitude Pi2
pulsations and related them to IMF observations from the
ACE and Wind satellites. They reported that the first Pi2
burst at the SMALL array followed a southward turning of
the IMF and the second one occurred shortly after the
northward turning of the IMF. They also pointed out that
the delay time of two consecutive bursts of low-latitude Pi2
is correlated with the estimated flux pileup by the incident
southward IMF into the magnetosphere. But how consecutive bursts of Pi2 pulsations are related to the multiple
substorm onsets was not comprehensively understood with
systematic observations from the solar wind, into the inner
nightside magnetosphere, and down to the ground.
[5] On 5 May 1999 the ACE satellite observed the
southward turning of the IMF at 0636 UT. The IMF Bz
first dropped to close to 3 nT, remained at that level for
about 15 min, and then slowly increased for more than one
hour until the IMF became northward at 0840 UT. This IMF
structure was also observed by the Wind satellite, which had
just exited the dayside magnetosphere. The Wind observation verified that the IMF structure as seen by ACE
remained fairly constant as it was convected toward the
Earth’s magnetosphere. Meanwhile, IMP 8 began to cross
the plasma sheet, GOES 10 moved toward the midnight
sector, and GOES 8 moved into the postmidnight sector.
Figure 1a shows at the time of the consecutive Pi2 bursts,
the locations of these satellites except for ACE that is 224 RE
upstream. Figure 1b is similar to Figure 1a except that it
shows the x – z plane. In Figure 1, the cross denotes the
satellite location at the onset time of the first Pi2 burst and
the triangle for the second burst. These satellites provided us
an opportunity to acquire a global picture of the development of multiple substorm onsets relevant to the southward
and northward turnings of the IMF. The relative weakness
of the southward field resulted in a slow evolution of the tail
and the resulting pair of disturbances were well separated.
In this study, we use magnetic records obtained from the
IGPP/LANL array and CANOPUS in comparison with the
magnetic field data at multiple satellites in an attempt to
understand these consecutive ‘‘substorm onsets.’’
2. Ground Observations
Figure 1. (a) Satellite locations on the x-y plane in the
GSM coordinates relevant to consecutive Pi2 bursts seen at
the IGPP/LANL array on 5 May 1999. The cross denotes
the satellite location at the onset time of the first Pi2 burst
and the triangle for the second burst. (b) Same as Figure 1a,
except for the x-z plane.
at the near-Earth neutral point first begins on closed field
lines within the plasma sheet and one when reconnection
reaches the open flux of the tail lobes. Thus, during substorm onsets, there can be two or more Pi2 bursts both on
the ground and in space.
[4] More recently with data from the Sino Magnetic
Array at Low Latitudes (SMALL) in 1999, Cheng et al.
2.1. Consecutive Bursts of Pi2 Pulsations at the IGPP/
LANL Array
[6] To study the propagation of solar wind disturbances in
the magnetosphere, the IGPP/LANL magnetometer array
has been jointly set up by Institute of Geophysics and
Planetary Physics (IGPP) at University of California, Los
Angeles (UCLA), University of California at Berkeley, Los
Alamos National Laboratory (LANL), and the US Air Force
Academy since 1998. Each station is installed with a
fluxgate magnetometer of high temporal resolution and
equipped with GPS receiver for accurate timing. More
details about the setup and instruments of the IGPP/LANL
array are given by Le et al. [1998] and also accessible on the
Web site (http://www-ssc.igpp.ucla.edu/uclamag). The locations of five available stations at the IGPP/LANL array are
listed in Table 1.
[7] Figure 2 shows that there were three consecutive
bursts of Pi2 pulsations in the H components from ATH
to TEO on 5 May 1999. In Figure 2, the first Pi2 burst
CHENG ET AL.: CONSECUTIVE Pi2 BURSTS AND SUBSTORM ONSETS
SMP
X-3
Table 1. Locations of Five Available Stations at the IGPP/LANL Array
Station Name
Athabasca
Edmonton
Boulder
USAFA
Teoloyucan
Abbreviation
Geographic
Latitude
Geographic
Longitude
Corrected Geomagnetic
Latitude
Corrected
Geomagnetic
Longitude
L
ATH
EDM
BLD
AFA
TEO
54.72
53.52
40.13
39.01
19.74
246.72
246.47
254.76
255.12
260.81
62.31
61.07
49.13
48.05
29.04
305.56
305.69
319.58
320.22
329.28
4.63
4.27
2.34
2.24
1.31
occurred at 0836 UT, the second one at 0938 UT and the
third at 1031 UT. Note that henceforth in this study #1
denotes the first Pi2 burst, #2 the second one and #3 the
third at the IGPP/LANL array. The solid vertical line
denotes the onset time for Pi2 bursts at the IGPP/LANL
array in the following figures. The waveforms at ATH look
like those at EDM. The waveforms at both BLD and AFA
are also similar to each other. The reason is that the two
Figure 2. Two consecutive Pi2 bursts in the H components at IGPP/LANL on 5 May 1999. The first
Pi2 event occurs at 0836 UT and the second one at 0938 UT. #1 denotes the first Pi2 burst and #2 for the
second burst. The solid vertical line denotes the onset time of Pi2 bursts.
SMP
X-4
CHENG ET AL.: CONSECUTIVE Pi2 BURSTS AND SUBSTORM ONSETS
stations of each pair are close to the same L value.
Compared to other stations, TEO seems to be more affected
by noise. Hence, the H component at TEO was low-pass
filtered with running average in every 16 s data points.
From close inspection of Figure 2, one may find that the
wave period for the first Pi2 burst is longer than the second
one. Moreover, the Pi2 wave period at the lower latitude is
shorter than at the higher latitude. This phenomenon is
similar to those studied by Li et al. [1998] and Cheng et al.
[2002] as well. However, both ATH and EDM stations are
near auroral latitudes, while most stations in the IGPP/
LANL array are usually inside the plasmasphere. Hence, the
Pi2 period at the both ATH and EDM may be determined by
bouncing Alfvén waves between the auroral ionosphere and
the neutral sheet during substorm onset [Baumjohann and
Glassmeier, 1984; Bauer et al., 1995]. There is also the
possibility that plasmaspheric cavity (virtual) resonances
[e.g., Cheng et al., 2000, and references therein] may play a
dominant role in determining the Pi2 wave period at lower
latitudes.
2.2. Magnetograms at the CANOPUS Array
[8] In this study, we also use the magnetograms at the
CANOPUS array to verify the occurrence of consecutive
Pi2 bursts relevant to multiple substorm onsets. On 5 May
1999 there are three larger magnetic disturbances in the
magnetograms at the CANOPUS array. Table 2 shows the
locations of five stations at the CANOPUS array. Figure 3
shows that the first larger disturbance in the X components
occurs at 0835 UT, the second at 0937 UT and the third at
1031 UT. Figure 4 corresponds to Figure 3 but shows the
Z components. While the auroral electrojet generally
causes the X component at a station under the electrojet
to decrease sharply over a wide latitude range, the sign of
the Z component indicates whether the current flows to the
north or the south of the stations. In Figure 3, the first two
sharp negative bays appear at the FSIM station but reach a
deeper minimum at the RABB station. This implies that
the two substorm expansions occur closer to RABB than
FSIM. We can use the Z component to give a more precise
location of the onset. For a westward electrojet, an
increase in current causes the Z component to become
more positive for a station north of it. In contrast, the Z
component becomes more negative for stations situated to
the south of the westward electrojet. Thus from Figure 4,
the first two magnetic differences in the Z components at
RABB and MCMU indicate that the auroral electrojet
during the first substorm is equatorward of RABB, but
during the second and third substorms it is poleward of
RABB.
2.3. Electrojet Inversion With Churchill Line Data at
CANOPUS
[9] By using an electrojet model based on adjustment of
parameters to obtain an optimal match to magnetic data
from available stations [Connors, 1998], we can be more
quantitative about both the location and strength of the
electrojet. Due to rather large latitude gaps between the
stations at the longitudes of RABB, FSIM, and MCMU,
the Churchill line data at CANOPUS were used instead for
electrojet inversion. In the Churchill meridian the declination of the magnetic field is near zero so modeling can be
done in XYZ coordinates. In Figure 5, the upper part
shows the latitudes of the electrojet borders in centered
dipole (CD) coordinates, which are roughly standard
magnetic coordinates. The westward current normal to
the meridian plane is shown at the bottom. This simple
model has uniform current density between the latitudes
given. Initially the currents are small. Nevertheless, the
auroral oval is rather latitudinally extended and similar
variations are seen over RABB, FSIM, and MCMU. The
first few points have too small a current to get solutions
but the consistent results after about 0815 UT seem to
indicate an extended oval. At the first onset the current
rises significantly but not rapidly: abrupt changes may
show at a ground station as one filament switches on but
the time constant shown for the overall current rise. Of
interest with regard to this study is that the latitude of the
electrojet at onset is as low as 60 degrees magnetic, but
after the onset, currents likely flow mainly at 64 – 70
degrees (consistent with what one sees in magnetograms).
The Pi2 may cause some of the jitter near onset. Before
the second onset is a period when the electrojet inversion
switches to a ‘‘narrow’’ form; this is likely to some extent
real although it partly reflects structure within the electrojet. The second onset is not as clear but is again accompanied by a poleward retreat of the equatorward border.
The third onset at 1031 UT broadens the auroral electrojet
and carries its northern border to higher latitudes. As
evident in Figure 5, the southern border of the electrojet
was about 4 degrees further north at the time of the second
onset than at the time of the first. After the second onset
the southern border again moved northward but returned to
the more southerly position for the third onset. One must
also bear in mind that the onset regions are likely well
west of this meridian.
3. Satellite Observations of the IMF
[10] In this section the magnetic field data at the ACE and
Wind satellites are used to examine how multiple substorm
onsets are related to the southward and northward turnings
of the IMF. Figures 6a – 6c show the solar wind data with
the southward IMF obtained by the ACE satellite from 0600
UT to 1000 UT on 5 May 1999. During the time of interest,
ACE was located about 224 Re (Earth radii) in front of the
Earth. It is apparent in Figure 6c that there is a steady
southward component of the IMF lasting for more than one
hour beginning at 0636 UT. Afterward, there is a clear
northward turning of the IMF occurring at 0743 UT in
Figure 6c. To check whether this IMF structure reaches the
Earth, we compare to the data from the Wind satellite.
Figures 6d –6f show the IMF data at Wind on 5 May 1999
from 0600 UT to 1000 UT. In Figure 6f, the southward IMF
begins at 0732 UT and lasts for more than one hour. There is
a clear northward turning of the IMF at 0842 UT in Figure
6f. The Wind satellite is located just outside of the Earth’s
dayside magnetosphere (see Figure 1 in this study). At 400
km/s the solar wind needs about 57 min to flow from ACE
to the Earth. Comparison of Figures 6c and 6f shows that
the steady southward component and northward turning of
the IMF are similar at both ACE and Wind. This verifies the
same IMF structures to persist as they propagate from ACE
to the Earth.
CHENG ET AL.: CONSECUTIVE Pi2 BURSTS AND SUBSTORM ONSETS
SMP
Table 2. Locations of Five Stations at the CANOPUS Array
Station Name
Abbreviation
Geographic
Latitude
Geographic
Longitude
Corrected Geomagnetic
Latitude
Corrected Geomagnetic
Longitude
L
Contwoyto Lake
Fort Smith
Fort Simpson
Rabbit Lake
Fort McMurray
CONT
FSMI
FSIM
RABB
MCMU
65.75
60.02
61.76
58.22
56.66
248.75
248.05
238.77
256.32
248.79
73.28
67.74
67.54
67.43
64.62
302.36
304.96
292.25
317.39
307.48
12.08
6.97
6.85
6.79
5.44
Figure 3. The X component of the magnetic field data at CANOPUS from 0800 UT to 1100 UT on 5
May 1999. The solid vertical line denotes the onset time of Pi2 bursts at IGPP/LANL. Same as Figure 2,
#1 and #2 denote ground Pi2 bursts, respectively.
X-5
SMP
X-6
CHENG ET AL.: CONSECUTIVE Pi2 BURSTS AND SUBSTORM ONSETS
Figure 4. Same as Figure 3, except for the Z component.
[11] In Figures 6a – 6c the solid vertical line denotes the
equivalent time of the Pi2 bursts referenced to the ACE and
Wind data. It is evident in Figures 6c and 6f that the first Pi2
burst occurred some time after the IMF turned southward
but the second Pi2 burst occurred shortly after the IMF
reached a horizontal and then northward orientation. The
third burst occurred after a period of quite strongly northward IMF. As discussed by Russell [2000], the near-Earth
neutral point (NENP) is expected to form after a certain
span of the southward IMF but the reconnection point does
not reach the open field lines of the lobe where rapid
reconnection can occur until shortly after the IMF turns
northward. The first Pi2 probably signals the onset of
reconnection on closed field lines. The third Pi2 seems to
be consistent with open flux reconnection and the northern
border of the electrojet moved further into the polar cap.
However, if so, what does Pi2 burst number two signal? We
now examine data in the night magnetosphere and tail to
explore this question. One of the advantages of the interval
chosen for this study is that the southward IMF is weak
even at its strongest so that time scales for significant
changes are long and the events are well separated.
4. Satellite Observations in the Nightside
Magnetosphere
[12] In this section the development of successive substorm onsets in the nightside magnetosphere is examined
CHENG ET AL.: CONSECUTIVE Pi2 BURSTS AND SUBSTORM ONSETS
SMP
X-7
Figure 5. The time series of the electrojet border and current intensity inverted with the Churchill line
data at CANOPUS on 5 May 1999. The dashed vertical lines denote the onset time for Pi2 bursts at IGPP/
LANL.
with the magnetic field data from IMP 8, GOES 8 and
GOES 10 satellites. On 5 May 1999, IMP 8 orbited across
the magnetotail at about x = 24 RE. Figure 7 suggests that
IMP 8 stays near the plasma sheet from 0600 UT to
1100 UT. In Figures 7a and 7b, there is a trend for the total
component Bt and the X component Bx of the magnetic field
to increase from 0640 UT and reach a maximum at
0913 UT. This trend could be caused either by the satellite
moving out of the plasma sheet as it thins or due to an
increased flaring of the tail as magnetic flux is added to it by
dayside merging. This result is similar to those reported by
Lyons et al. [2001]. In Figure 7, a broad slow decline of the
magnetic field begins after about 0945 UT and hastens after
the third Pi2 at 1031 UT. According to Russell and
McPherron [1973], such a decrease of the Bt component
of the magnetic field in the tail indicates that magnetic flux
has been removed from it. In the NENP model of substorms
this occurs by reconnection. In summary, the first Pi2 seems
unrelated to the transport of open flux to the tail or the
removal of open flux from the tail; the second Pi2 occurred
when net transport to the tail ceased; and the third Pi2
occurred when transport of flux out of the tail lobes began.
SMP
X-8
CHENG ET AL.: CONSECUTIVE Pi2 BURSTS AND SUBSTORM ONSETS
Figure 6. (a) The Bx component of the IMF data at ACE from 0600 UT to 1000 UT on 5 May 1999.
ACE was moving from (224.0, 12.487, 22.42) Re to (224.0, 12.485, 22.34) Re in GSM coordinates.
As in Figure 2, #1 and #2 denote ground Pi2 bursts, respectively. The solid vertical line denotes the
equivalent onset time of ground Pi2 bursts at ACE correcting for solar wind transit time in the X direction.
(b) As in Figure 6a, except for the By component. (c) As in Figure 6a, except for the Bz component. (d)
As in Figure 6a, except for Wind moving from (4.1, 9.8, 31.1) Re to (1.3, 10.7, 29.4) Re in GSM
coordinates. (e) Same as Figure 6d, except for the By component. (f ) Same as Figure 6d, except for the Bz
component.
[13] From 0800 UT and 1100 UT on 5 May 1999, both
GOES 8 and GOES 10 stayed in the nightside magnetosphere. In the time corresponding to from 0300 LT to 0600
LT, GOES 8 was moving from ( 4.5, 4.5, 2.0) Re to (0.0,
6.5, 1.25) Re in the GSM coordinate system. In the time
corresponding to 2248 LT to 0154 LT, GOES 10 was moving
from ( 6.15, 2.0, 1.58) Re to ( 5.5, 3.0, 2.2) Re in the
GSM coordinate system. These synchronous orbit satellites
provide us an opportunity to examine the dynamic process of
multiple substorm onsets in the inner magnetosphere. Note
that the solid vertical lines in Figure 8 denote the onset time
of consecutive Pi2 bursts at the IGPP/LANL array. The
CHENG ET AL.: CONSECUTIVE Pi2 BURSTS AND SUBSTORM ONSETS
SMP
X-9
Figure 7. (a) The total component Bt of IMP 8 from 0600 UT to 1100 UT on 5 May 1999. IMP 8 was
moving from ( 25.5, 8.8, 8.4) Re to ( 22.8, 10.7, 11.2) Re in GSM coordinates. As in Figure 2, #1
and #2 denote ground Pi2 bursts, respectively. The solid vertical line denotes the onset time for Pi2 bursts
at IGPP/LANL. (b) Same as Figure 7a, except for the Bx component. (c) Same as Figure 7a, except for
the By component. (d) Same as Figure 7a, except for the Bz component.
magnetic field at GOES 8 and GOES 10 satellites is defined
as: Hp, perpendicular to the satellite orbital plane (or parallel
to the Earth spin axis in the case of a zero degree inclination
orbit); He, perpendicular to Hp and directed earthwards; and
Hn, perpendicular to Hp and directed eastward. During the
time of interest, GOES 8 was moving from the postmidnight
sector into the dawn sector (see Figure 1 in this study).
Figures 8a and 8c show that at the onset of the first Pi2 burst
the He component begins to decrease and the Hp component
begins to increase. This signals a return of magnetic flux into
the nightside magnetosphere. Figure 8b shows that the
increase in fluctuations in the Hn direction confirms the
onset of activity signaled by the Pi2 on the ground, albeit in
space the disturbance lasts much longer. The increasing Hn
component and quasiperiodic oscillations have been studied
by Saka et al. [1996]. With simulation of forced field line
oscillations by the sudden increase in the plasma pressure in
the equatorial midnight sector region, they suggested that the
source could be particle injections from the magnetotail at
substorm expansion onset.
[14] During the time of interest, GOES 10 moved across
the midnight sector (see Figure 1 in this study). In Figures
8e – 8f there are three clear magnetic disturbances, especially
in the He and Hn components, the first disturbance begin-
ning at 0836 UT and continuing to 0900 UT and the second
one commencing at 0937 UT and the third at about 1031
UT. The magnitudes of the first and third disturbances are
larger than the second one. In Figures 8d and 8e, the Hn
component at onset has positive perturbations and the He
component has negative perturbations. This is consistent
with the location of GOES 10 with respect to the substorm
current wedge of which perspective view was illustrated in
Figure 7 of Clauer and McPherron [1974], and is also
consistent with the results of Sakurai and McPherron
[1983]. Moreover, Figures 8d– 8f shows that the magnetic
disturbances at GOES 10 start about 1– 2 min before ground
Pi2 onsets.
[15] The signatures are also consistent with the ground
signature at the CANOPUS array described above, and
show that all three Pi2 pulsations have substorm like
effects on the night magnetosphere even though the IMF
source is a single southward and northward turning and
the tail shows a single filling and emptying cycle.
5. Discussion and Summary
[16] As mentioned in section 1, from earlier studies to
recent observations, ground based studies have revealed
SMP
X - 10
CHENG ET AL.: CONSECUTIVE Pi2 BURSTS AND SUBSTORM ONSETS
Figure 8. (a) The He component of GOES 8 from 0800 UT to 1100 UT on 5 May 1999. From 0300 LT
to 0600 LT, GOES 8 was moving from ( 4.5, 4.5, 2.0) Re to (0.0, 6.5, 1.25) Re in the GSM
coordinates. Same as Figure 2, #1 and #2 denote ground Pi2 bursts, respectively. The solid vertical line
denotes the onset time for Pi2 bursts at IGPP/LANL. (b) Same as 8a, except for the Hn component. (c)
Same as 8a, except for the Hp component. (d) Same as 8a, except for GOES 10. In the time
corresponding to from 2248 LT to 0154 LT, GOES 10 was moving from ( 6.15, 2.0, 1.58) Re to ( 5.5,
3.0, 2.2) Re in the GSM coordinates. (e) Same as Figure 8d, except for the Hn component. (f ) Same as
Figure 8d, except for the Hp component.
the occurrence of multiple onsets in a single magnetospheric substorm sequence. While some might prefer to
call these multiple, closely spaced substorms, the important and nonsemantic point is that the multiple onsets
were associated with a single north to south to north again
sequence of IMF changes. Until now, except the attempt
of Cheng et al. [2002] to compare ground pulsations to
IMF observations, there have been few studies to investigate the global development of successive substorm
onsets related to consecutive Pi2 bursts with both space
CHENG ET AL.: CONSECUTIVE Pi2 BURSTS AND SUBSTORM ONSETS
and ground observations. With the availability of satellite
observations in the nightside magnetosphere, the 5 May
1999 event is the first observational evidence to clarify
the relationship of consecutive Pi2 bursts with successive
substorm onsets relative to the southward and northward
turnings of the IMF.
[17] During substorm onsets, the X component at a
station under the auroral electrojet may decrease sharply
over a wide latitude range. But the sign of the Z
component indicates whether the electrojet flows to the
north or the south of the stations. As a result, one may
speculate from Figures 3 and 4 in this study that the
reconnection site for the first substorm onset is located
more earthward than for the second one. For justification
of above speculation, the electrojet border and current
intensity were inverted from the Churchill line data within
the CANOPUS array with an electrojet model. Figure 5 of
this study shows that the latitude of electrojet at onset is
as low as 60 degrees magnetic but the onset currents
move quickly to 64 –70 degrees. This is consistent with
the signature of magnetic disturbances caused by the
auroral electrojet in Figures 3 and 4 of this study. The
second onset is not as clear but is again accompanied by a
poleward retreat of the equatorward border. Moreover, the
southern border of the electrojet was about 4 degrees
further north at the time of the second onset. This
significant change during the third pulsation was for the
poleward border to move further into the polar cap region.
Thus each onset caused the currents to move further
poleward.
[18] Recently, Russell [2000] discussed how in a twoneutral-point model substorm onsets could be triggered by
the northward turning of the IMF. In the model, the distant
neutral point supplies the magnetized plasma on closed
field lines surrounding the NENP that reconnect on closed
field lines after the IMF has turned southward. We use the
acronym NENP rather than the more common NENL for
near-Earth neutral line to emphasize the importance of the
localized nature of the reconnection region. When the IMF
turns northward, the reconnection at the distant neutral
point ceases. But the reconnection at the NENP continues
and soon reaches the low-density field lines where the rate
becomes rapid driving a full expansion. As a result, there
are two onsets in the model and the first one occurs at a
lower latitude as the reconnection at the NENP initiates on
closed field lines, while the second ensues when reconnection reaches low density open field lines at the edge of
the plasma sheet and rapid increase in the rate of reconnection occurs into the open flux of the tail lobes. This
interpretation is consistent with the many substorms
studied with ground-based data [Mishin et al., 2000,
2001]. However, in this study we find three onsets, not
two, in a single substorm sequence, marked by a single
cycle of southward IMF interval followed by a northward
IMF interval and marked by a single cycle of tail flux
buildup and decay. Nevertheless, the night magnetosphere
showed three onsets marked by Pi2s and by near-synchronous orbit dipolarization in the magnetic field. This
observation indicates that tail reconnection is very nonsteady and can stop and restart in what otherwise is a
single substorm. Thus, the model of Russell [2000] is
oversimplified.
SMP
X - 11
[19] In summary, the relationship between multiple
bursts of Pi2 pulsations and substorm onsets is investigated with both ground and satellite magnetic data on 5
May 1999. The magnetograms at the CANOPUS array
verify the occurrence of three substorm onsets, accompanied
by successive Pi2 bursts at the IGPP/LANL array. Interpreted
with the line current model for the auroral electrojet, the
latitudinal variation in the H and Z components shows a
stepwise polarward progression of activity. A comparison of
ground observations with the ACE and Wind magnetic field
data shows that the first Pi2 burst occurs well after the
southward turning of the IMF, the second one occurs shortly
after the IMF becomes slightly northward, and the third after
a strong northward turning. Magnetic disturbances at GOES
8 and GOES 10 show similar signatures resulting from the
substorm current wedge at all three onsets. These satellites
results are consistent with ground observations from both
CANOPUS and IGPP/LANL array. The behavior of these
multiple Pi2 bursts and the nightside signatures at synchronous orbit shows that a single cycle of dayside reconnection
and flux transport to the tail can generate multiple onsets of
activity.
[20] Acknowledgments. This work was performed while the first
author (C.-C. Cheng) was on study leave at IGPP/UCLA supported by the
scholarship to study abroad from the Ministry of Education, R. O. C. on
Taiwan. Operation of ATH and EDM magnetometers was supported by the
Academic Research Fund of Athabasca University. M. Connors acknowledges support from the Canadian NSERC. The magnetometer data at the
AFA and TEO stations were obtained with assistance from F. K. Chun and
J. A. L. Cruz-Abeyro. The CANOPUS data were provided by Canadian
Space Agency. We thank H. Singer at NOAA for providing the GOES 8 and
GOES 10 magnetic field data on the CDAWeb.
References
Bauer, T. M., W. Baumjohann, and R. A. Treumann, Neutral sheet oscillations at substorm onset, J. Geophys. Res., 100, 23,737 – 23,742, 1995.
Baumjohann, W., and K. H. Glassmeier, The transient response mechanism
and Pi 2 pulsations at substorm onset: Review and outlook, Planet. Space
Sci., 32, 1361 – 1370, 1984.
Cheng, C.-C., J. K. Chao, and K. Yumoto, Spectral power of low-latitude Pi
2 pulsations at 210° magnetic meridian stations and plasmaspheric cavity
resonances, Earth Planet. Space, 52, 615 – 627, 2000.
Cheng, C.-C., C. T. Russell, Y. F. Gao, and P. J. Chi, On consecutive bursts
of low-latitude Pi 2 pulsations, J. Atmos. Sol. Terr. Phys., in press, 2002.
Clauer, C. R., and R. L. McPherron, Mapping the local time-universal time
development of magnetospheric substorms using mid-latitude magnetic
observations, J. Geophys. Res., 79, 2811 – 2820, 1974.
Connors, M., Auroral current systems studied using automated forward
modelling, Ph.D. thesis, 427 pp., Univ. of Alberta, Edmonton, Canada,
1998.
Kisabeth, J. L., and G. Rostoker, Development of the polar electrojet during
polar magnetic substorms, J. Geophys. Res., 76, 6815 – 6828, 1971.
Le, G., C. T. Russell, P. J. Chi, V. Angelopoulos, G. Reeves, F. K. Chun,
and W. Goedecke, Initial results from the IGPP/LANL ground-based
magnetometer array, Eos. Trans. AGU, 79(17), S302, Spring Meet.
Suppl., 1998.
Li, Y., B. J. Fraser, F. W. Menk, D. J. Webster, and K. Yumoto, Properties
and sources of low and very low latitude Pi 2 pulsations, J. Geophys.
Res., 103, 2343 – 2358, 1998.
Lyons, L. R., R. L. McPherron, E. Zesta, G. D. Reeves, J. B. Sigwarth, and
L. A. Frank, Timing of substorm signatures during the November 24,
1996 geospace environment modeling event, J. Geophys. Res., 106,
349 – 359, 2001.
Mishin, V. M., C. T. Russell, T. I. Saifudinova, and A. D. Bazarzhapov,
Study of weak substorms obeserved during December 8, 1990, geospace
environment modeling campaign: Timing of different types of substorm
onsets, J. Geophys. Res., 105, 23,263 – 23,276, 2000.
Mishin, V. M., T. Saifudinova, A. Bazarzhapov, C. T. Russell, W. Baumjohann, B. Nakamura, and M. Kubyshkina, Two distinct substorm onsets,
J. Geophys. Res., 106, 13,105 – 13,118, 2001.
SMP
X - 12
CHENG ET AL.: CONSECUTIVE Pi2 BURSTS AND SUBSTORM ONSETS
Olson, J. V., Pi 2 pulsations and substorm onsets: A review, J. Geophys.
Res., 104, 17,499 – 17,520, 1999.
Rostoker, G., Microstructure of geomagnetic bays, J. Geophys. Res., 73,
4217 – 4229, 1968.
Russell, C. T., How northward turnings of the IMF can lead to substorm
expansion onsets, Geophys. Res. Lett., 27, 3257 – 3259, 2000.
Russell, C. T., and R. L. McPherron, The magnetotail and substorms, Space
Sci. Rev., 15, 205 – 266, 1973.
Saito, T., Geomagnetic pulsations, Space Sci. Rev., 10, 319 – 412, 1969.
Saka, O., H. Akaki, O. Watanabe, and D. N. Baker, Ground-satellite correlation of low-latitude Pi 2 pulsations: A quasi-periodic field line oscillation in the magnetosphere, J. Geophys. Res., 101, 15,433 – 15,440, 1996.
Sakurai, T., and R. L. McPherron, Satellite observation of Pi 2 activity at
synchronous orbit, J. Geophys. Res., 88, 7015 – 7027, 1983.
Yumoto, K., Generation and propagation mechanisms of low-latitude magnetic pulsations: A review, J. Geophys. Res., 91, 79 – 105, 1986.
C.-C. Cheng, P. J. Chi, and C. T. Russell, Institute of Geophysics and
Planetary Physics, University of California, Los Angeles, CA 90095-1567.
(ctrussell@igpp.ucla.edu)
M. Connors, Centre for Science, Athabasca University, 1 University
Drive, Athabasca, Alberta, Canada T9S 3A3.