The Astrophysical Journal, 566:L117–L120, 2002 February 20
䉷 2002. The American Astronomical Society. All rights reserved. Printed in U.S.A.
ARE HOMOLOGOUS FLARE–CORONAL MASS EJECTION EVENTS TRIGGERED
BY MOVING MAGNETIC FEATURES?
Jun Zhang1,2 and Jingxiu Wang1
Received 2001 November 2; accepted 2002 January 16; published 2002 January 28
ABSTRACT
Coronal mass ejections (CMEs) often present destabilization and eruption of a global (or large-scale) magnetic
structure in the solar atmosphere. Furthermore, the Earth-directed CMEs are the primary driver of the disastrous
space weather. Here we report on five homologous CMEs. They initiated in the early phase of five homologous
X-class flares seen in NOAA Active Region 9236 on 2000 November 24–26. The flares appeared between the
main sunspot with positive magnetic field and the moat of the active region. We examined the magnetic evolution
in the source active region for the first three of the five flare-CME events and found that the main magnetic
changes are magnetic flux emergence in the form of moving magnetic features (MMFs) in the vicinity of the
main positive magnetic field. There are three peaks in the flux (number) distribution of the emerging MMFs. If
each peak corresponds to an X-class flare and associated halo CME, there is a time lag of 10 hr between flux
emergence and flaring. The flux appearing in the form of MMFs is 1.1 # 10 22 Mx, with a net flux of ⫺2.1 #
10 21 Mx. Around the main spot region, about 9.1 # 10 21 Mx of flux disappeared in the 2 day interval. This is
indicative that the repeated flare-CME activities are triggered by the continuous emergence of MMFs.
Subject headings: Sun: coronal mass ejections (CMEs) — Sun: flares — Sun: magnetic fields —
Sun: UV radiation
homologous CMEs, and they defined the homologous CMEs
as follows: (1) each member of homologous CMEs must associate with a member of a homologous flare, (2) extended
EUV (X-ray) dimming must be similar among the homologous
CME and flare, and (3) the coronagraph appearances of the
homologous CME and flare must also resemble each other.
Moving magnetic features (MMFs) are small magnetic elements that are carried away from the sunspot to the periphery
by plasma flows (Vrabec 1971; Harvey & Harvey 1973; Muller
& Mena 1987; Brickhouse & LaBonta 1988; Lee 1992). There
are two kinds of MMFs: unipolar and mixed polarity (Harvey
& Harvey 1973; Ryutova et al. 1998; Yurchyshyn, Wang, &
Goode 2001). The model by Harvey & Harvey (1973) suggests
that magnetic flux is removed from the sunspot at the photospheric level. This would produce pairs of MMFs in which
magnetic elements of polarity opposite to that of the sunspot
tend to be formed farther out. An alternative possibility was
suggested by Wilson (1973, 1986). In this case, the magnetic
flux tube is detached from the main bundle of tubes well below
the surface. The detached tubes float turbulently to the surface
developing twists and kinks, which then are seen as MMFs.
Spruit, Title, & van Ballegooijen (1987) assumed that a large
loop rises from the convection zone and breaks into many small
loops as it crosses the surface. Unfortunately, the true nature
of MMFs is still an open question.
Fast and furious action was observed in NOAA Active Region
9236 on 2000 November 24–26, when the region not only produced numerous bright jets visible in the EUV, but also five
homologous X-class flares. At least five halo CMEs (Nitta &
Hudson 2002) were recorded by the Solar and Heliospheric
Observatory (SOHO) Large Angle and Spectrometric Coronagraph (LASCO) Experiment. Here we report five homologous
CMEs that are identified to be associated with the five homologous X-class flares, we show the magnetic evolution in the
source region of the homologous flare-CME events.
The high-resolution UV observations were obtained from the
Transition Region and Coronal Explorer (TRACE) satellite
(Handy et al. 1999). TRACE offers an unprecedented opportunity
1. INTRODUCTION
Coronal mass ejections (CMEs) are known as the most spectacular form of solar magnetic activity whose analog has not
been clearly identified in other astrophysical objects (Wang et
al. 2002). Great efforts and important progress in CME studies
have been made since the first detection of CMEs in the early
1970s (Rust & Hildner 1976; Wanger 1984; Kahler 1992; Gosling 1996). However, it is still poorly understood how CMEs
initiate. The reason for this lies in the fact that the CMEs are
observed with a coronagraph above the occulting disk as projections on the sky plane. Based on these observations, it is
difficult to correlate the observed CMEs with configuration and
evolution of the underlying magnetic fields, which are believed
to power the CMEs. The puzzles concerning the initiation of
CMEs raise serious difficulties in any attempt to predict the
Earth-directed mass ejections. As argued by Feynman & Martin, only with the disk identity of CME initiation can an accurate
correlation of CMEs with the surface magnetism and activity
be obtained (Feynman & Martin 1995). Real progress in understanding the physics of CME initiation cannot be made before
we have enough knowledge about CME-associated magnetic
changes on the solar surface. Recently, efforts to associate
CMEs with surface sources have started to be made (Dere et
al. 1997; Delannée & Aulanier 1999; Innes et al. 1999, 2001;
Maia et al. 1999; Delannée, Delaboudinière, & Lamy 2000;
Zhang et al. 2001; Zhang & Wang 2001; Zhang, Wang, & Nitta
2001).
The definition of homologous flaring states that members of
a series must have the same main footpoints as defined by Ha
or EUV kernels. They must also share the same general shape
in the main phase as defined by Ha ribbons or EUV images
(Woodgate et al. 1984). Wang et al. (2002) first reported on
1
National Astronomical Observatories, Chinese Academy of Sciences, 20
Datun Road, 100012 Beijing, China; [email protected], [email protected].
ac.cn.
2
Max-Planck-Institut für Aeronomie, Katlenburg-Lindau D-37191, Germany; [email protected].
L117
L118
HOMOLOGOUS FLARE-CME EVENTS
Vol. 566
five associated flares were homologous. Figure 1 demonstrates
the generally similar shape and the identical location of the flare
ribbons relative to the active region structures. The first column
(from left to right) shows three flares observed by TRACE at the
1600 Å wavelength, on 2000 November 24. The second column
shows five flares observed by SOHO/EIT, from 2000 November
24 to 26; three windows outline the fields of view of the TRACE
1600 Å images shown in the first column. Second, the five CMEs
exhibited similar extended EUV dimming, which is shown in
the third column. Arrows indicate dimming regions. Third, the
five CMEs resemble each other in appearance in LASCO images
(see Fig. 1, fourth column). They all manifested themselves as
halo CMEs as NOAA AR 9236 was not far from the central
meridian. Each later CME appeared as a magnification of the
former. The behavior is exactly the same as in the homologous
CMEs reported by Wang et al. (2002). As the active region
rotated to the limb, the weight centers of the CMEs shifted
slightly from the north to the equator.
A traditional back extrapolation of the CME fronts is used
to roughly infer the CME onset time and the projected ejection
velocity for each CME. Table 1 shows the main properties of
the five homologous flare-CME events. Although several factors have limited the accuracy of the extrapolation, such as the
rather poor temporal resolution of EIT and LASCO data, the
uncertainty of the CME initiated height, and the angle between
the ejection and the sky plane, the data display that each CME
initiated around the time of the early phase of the corresponding
flare.
Fig. 1.—Appearance of the five homologous flare-CME events. From left
to right, first column: Three flares observed by TRACE 1600 Å wavelength,
on 2000 November 24. Second column: Five flares observed by SOHO/EIT,
from 2000 November 24 to 26; three windows outline the field of view of
TRACE 1600 Å images shown in first column. Third column: Running difference of SOHO/EIT 195 Å images; arrows indicate dimming regions. Fourth
column: Appearance of the five CMEs in LASCO C2 coronagraph; white circle
in the center of each image shows the location of the solar limb. The field of
view of TRACE 1600 Å images is about 128 # 128, and the field of view
of EIT images 1008 # 1008.
to learn the fine structure and dynamics of the corona and transition region. The 1600 Å images used in this study have spatial
resolution of 1⬙. 0, temporal resolution of 20–60 s, and fields of
view 8⬘. 5 # 8⬘. 5. The full-disk EUV images, coronagraphs, and
high-resolution line-of-sight magnetic field observations were
taken by the EUV Imaging Telescope (EIT), LASCO, and the
Michelson Doppler Imager (MDI) on board SOHO. Detailed
descriptions of these instruments are provided by Delaboudinière
et al. (1995), Brueckner et al. (1995), and Scherrer et al. (1995).
2. THE PROPERTIES OF HOMOLOGOUS FLARE-CME EVENTS
The homology of the five CMEs is concluded from their general resemblance in the three key aspects: the associated surface
activity, the solar disk signature, and the coronagraph appearance.
First, each CME was associated with an X-class flare, and the
3. MOVING MAGNETIC FEATURES IN THE SOURCE REGION
OF THE FIVE FLARE-CME EVENTS
The active region NOAA AR 9236, a large, fast-growing
beta region consisting of a large leader spot and smaller trailing
spots, emerged on the eastern limb of the Sun on 2000 November 19. Full-disk MDI magnetograms show that at the early
phase of this active region evolution, the leader spot with positive polarity flux was simple and isolated, and there is almost
no opposite polarity flux near its boundary. However, MMFs
emerged successively around the leader spot just after the spot
first appeared at the solar surface. On 2000 November 23, the
active region was located near the solar center, and some negative flux developed around the leader positive spot due to the
successive emergence of MMFs. At this time, violent activity
started. In Figure 2 we compare a TRACE 1600 Å image with
an MDI magnetogram on 2000 November 24. The UV brightening (shown by ellipses in the UV image) always appears at
sites where opposite polarity magnetic patches (indicated by
ellipses in magnetogram) cancel. TRACE and SOHO/EIT observations show that these bright points subsequently extended
and developed into an X2.1 flare at 04:55 UT, the first X-class
flare of the five. The other four flares also developed from
bright points. These bright points were located near the positions of those of the first flare.
TABLE 1
Properties of Five Homologous Flare-CME Events
Flare Number
1
2
3
4
5
.............
.............
.............
.............
.............
Date
Nov
Nov
Nov
Nov
Nov
24
24
24
25
26
Class
Start Time
(UT)
Peak Time
(UT)
Stop Time
(UT)
Time
(UT)
Speed
(km s⫺1)
X2.1
X2.3
X2.0
X1.9
X4.0
04:55
14:51
21:48
18:33
16:34
05:02
15:13
22:00
18:44
16:48
05:08
15:21
22:24
18:55
16:56
04:54
15:04
21:45
18:45
16:33
1050
1220
990
660
980
No. 2, 2002
ZHANG & WANG
L119
Fig. 2.—Comparison of a TRACE 1600 Å image (left) and an MDI magnetogram (right). White patches in the magnetogram present positive polarity
fields, and black patches, negative fields. The field of view is about 145 #
145. Symbols and arrows are described in the text.
To show the magnetic evolution on some key sites, we present
in Figure 3 the time series of MDI magnetograms in the region
shown by a window in Figure 2. There are four relatively large
magnetic patches (shown by arrows and circles in Figs. 2 and
3, respectively, and marked by numbers 1–4) and several MMFs.
It is identified from the history of magnetic evolution that the
four patches (1–4) are the collections of the elements of MMFs.
Patch 1 first appeared at 13:16 UT on 2000 November 23. From
18:28 to 23:38 UT, many small magnetic elements merge with
the former part of patch 1, thus making the flux increasingly
large. Patches 2–4 undergo a similar evolution process as patch
1. The interaction among these magnetic fields produces several
bright points (e.g., b1, b2, and b3 in Fig. 2), the early appearances
of the first flare.
We use here a 2 day (from 2000 November 23 to 24) sequence
of high-resolution MDI magnetograms (0⬙. 625 per CCD pixel
and 1 minute per line-of-sight magnetogram) to study the magnetic evolution of NOAA AR 9236, the source region of five
flare-CME events. The obvious magnetic changes in the course
of the events are magnetic flux emergence in the form of MMFs.
From the sequence of magnetograms studied, we can see that
most elements of MMFs appear as clusters of mixed polarity. Some elements are born as small ephemeral regions. We
are able to follow 452 MMFs from birth to death and measure
the flux of each element. About 1.1 # 10 22 (9.1 # 10 21) Mx
of flux emerged (disappeared) in the 2 days of 2000 November
23 and 24. The flux emergence (disappearance) rate is 2.3 #
10 20 (1.9 # 10 20) Mx hr⫺1. Among all the emergence flux, the
negative flux exceeds the positive by ⫺2.1 # 10 21 Mx.
Figure 4 presents the hour-averaged emerging flux and the
number of MMFs. There are three peaks in both the flux (P1,
P2, and P3) and number (N1, N2, and N3) distributions in
temporal evolution. The dotted curve shows the mean relative
intensity of TRACE 1600 Å emission, inside the region shown
in the first column of Figure 1. The three peaks of intensity
represent three flares: F1, F2, and F3. If the peaks of flux
(number) distribution correspond to the flares, e.g., P1 (N1)
corresponds to F1, P2 (N2) to F2, and P3 (N3) to F3, we can
say that the MMFs appear about 10 hr prior to the corresponding flare. Support for this idea is obtained from Figures 2 and
3, which show that the bright points b1, b2, and b3, which
represent the early stage of the first flare, are triggered by
MMFs. In most cases, the MMFs emerge about 5 hr before
the corresponding bright points. The bright points first appear
about 6–7 hr before the flare. During the time when the later
two flares happened, NOAA AR 9236 was out of the field of
view of the high-resolution MDI magnetograph. However, full-
Fig. 3.—Time series of line-of-sight magnetograms from SOHO/MDI, in
the area shown by the window in Fig. 2. The field of view is about 42 #
42. As described in the text, the positions of magnetic features 1–4 are identified at appropriate times to illustrate their appearance, growth, and movement.
disk MDI observations show that in the active region, the magnetic evolution pattern during the later two flare onsets was
similar to that of the earlier three flares.
4. RESULTS AND DISCUSSION
The solar events on 2000 November 24–26 exhibit simultaneous and homologous flares and CMEs. These events provide a good opportunity to study the surface magnetic activity
that results in the successive magnetic instability. In this Letter
we examine the magnetic evolutions in the course of the solar
events, with an emphasis on the evolutions that characterize
this homologous flare-CME active region. We suggest that the
homologous flare-CME events are triggered by the successive
emergence of MMFs.
A flare will erupt when new magnetic flux emerges within
or adjacent to the unipolar magnetic fields in an orientation
favorable for reconnection (Feynman & Martin 1995). Flux
cancellation in the vicinity of a neutral line is suggested to be
Fig. 4.—Mean relative intensity (dotted curve) of TRACE 1600 Å emission
in the region presented in the first column of Fig. 1, the hour-averaged emerging
flux (upper light curve) and number (lower heavy curve) of moving magnetic
features, from the beginning of 2000 November 23 to the beginning of November 25. Details are described in the text.
L120
HOMOLOGOUS FLARE-CME EVENTS
a necessary condition for flare onset. The association of flares
and cancelling magnetic fields was first noted by Martin, Livi,
& Wang (1985) and was later discussed by Livi et al. (1989)
and Wang & Shi (1993). Feynman & Martin (1995) and Wang
& Sheeley (1999) identified the association of emerging flux
and CMEs. However, very few systematic efforts to identify
the distribution and evolution of surface magnetic flux that
leads to CME initiation have been made so far.
Our observations show that on November 24, when three
X-class flares took place, flux cancellation was observed at
many sites. Moreover, the TRACE 1600 Å brightenings always
appear at the sites of cancelling magnetic features. This indicates that some energy release takes place at the sites of flux
cancellation. Thus, magnetic reconnection must have taken
place in cancelling magnetic features. Wang & Shi (1993) presented a two-step reconnection scenario for the flare process.
The first-step reconnection takes place in the photosphere, or
lower atmosphere, which manifests itself as flux cancellation
observed in the photosphere. It is slow but continuous. This
slow reconnection might convert the magnetic energy into heat
and kinetic energy; but more importantly, it transports the magnetic energy and complexity into the larger scale magnetic
structure higher in the corona. The second-step reconnection,
which is explosive in nature and directly responsible for transient solar activities, can take place only when some critical
status is achieved in the corona. The key idea of these authors
is the overwhelming importance of magnetic reconnection in
the lower atmosphere in the energy process of explosive solar
activities.
Vol. 566
We suggest that at the time the flare-CME events happened,
NOAA AR 9236 consisted of at least three kinds of magnetic
structures: (1) active region loops (ARLs), connecting the
leader spot and smaller trailing spots; (2) a moat, developed
from the collections of former MMFs; and (3) MMFs. These
three kinds of structures are the key interplaying entities in the
homologous CMEs. The MMFs are the driver, while the moat
and ARLs are the responders that are energized by “helicity
charging” in a slow magnetic reconnection with the driver. With
the increase of self-helicity, the ARLs grow and erupt outward
when the helicity exceeds a certain magnitude. This instability
characterizes the early phase of the CMEs. The magnetic skeleton remains, and the basic processes repeat as long as the
MMFs keep emerging. In each CME, only a fraction of ARLs
are involved, but more and more fractions are energized and
erupt in successive CMEs.
We are grateful to the referee, D. E. Innes, whose suggestions
and comments led to improvement of the manuscript. J. Zhang
thanks Professor S. Solanki for valuable discussions. The authors are indebted to the TRACE, SOHO/EIT, LASCO, and
MDI teams for providing the wonderful data. This work is
supported by the National Natural Science Foundation of China
under grant G19973009, the National Key Basic Research Science Foundation grant G2000078404, and the cooperation
agreement between the Chinese Academy of Sciences and the
Max-Planck Society.
REFERENCES
Brickhouse, N. S., & LaBonta, B. J. 1988, Sol. Phys., 115, 43
Brueckner, G. E., et al. 1995, Sol. Phys., 162, 357
Delaboudinière, J.-P., et al. 1995, Sol. Phys., 162, 291
Delannée, C., & Aulanier, G. 1999, Sol. Phys., 190, 107
Delannée, C., Delaboudinière, J. P., & Lamy, P. 2000, A&A, 355, 725
Dere, K. P., et al. 1997, Sol. Phys., 175, 601
Feynman, J., & Martin, S. F. 1995, J. Geophys. Res., 100, 3355
Gosling, J. T. 1996, ARA&A, 34, 35
Handy, B. N., et al. 1999, Sol. Phys., 187, 229
Harvey, K., & Harvey, J. 1973, Sol. Phys., 28, 61
Innes, D. E., et al. 1999, Sol. Phys., 186, 337
———. 2001, ApJ, 549, L249
Kahler, S. 1992, ARA&A, 30, 113
Lee, J. W. 1992, Sol. Phys., 139, 267
Livi, S. H. B., Martin, S. F., Wang, H., & Ai, G. 1989, Sol. Phys., 121, 197
Maia, D., et al. 1999, J. Geophys. Res., 104(A6), 12,507
Martin, S. F., Livi, S. H. B., & Wang, J. 1985, Australian J. Phys., 38, 929
Muller, R., & Mena, B. 1987, Sol. Phys., 112, 295
Nitta, N. V., & Hudson, H. S. 2002, Geophys. Res. Lett., 28, 3801
Rust, D. M., & Hildner, E. 1976, Sol. Phys., 48, 381
Ryutova, M., Shine, R., Title, A., & Sakai, J. I. 1998, ApJ, 492, 402
Scherrer, P. H., et al. 1995, Sol. Phys., 162, 129
Spruit, H. C., Title, A. M., & van Ballegooijen, A. A. 1987, Sol. Phys., 110,
115
Vrabec, D. 1971, in IAU Symp. 43, Solar Magnetic Fields, ed. R. Howard
(Dordrecht: Reidel), 329
Wang, J., & Shi, Z. 1993, Sol. Phys., 143, 119
Wang, J., et al. 2002, ApJ, submitted
Wang, Y.-M., & Sheeley, N. R. 1999, ApJ, 510, L157
Wanger, W. J. 1984, ARA&A, 22, 267
Wilson, P. R. 1973, Sol. Phys., 32, 435
———. 1986, Sol. Phys., 106, 1
Woodgate, B. E., et al. 1984, Adv. Space Res., 4, 11
Yurchyshyn, V. B., Wang, H., & Goode, P. R. 2001, ApJ, 550, 470
Zhang, J., & Wang, J. 2001, ApJ, 554, 474
Zhang, J., Wang, J., Deng, Y., & Wu D. 2001, ApJ, 548, L99
Zhang, J., Wang, J., & Nitta, N. V. 2001, Chinese Astron. Astrophys., 1, 85