1: \documentclass[12pt,preprint]{aastex}
2:
3: %\usepackage[dvips]{graphicx}
4: %\usepackage{verbatim}
5: %\usepackage{multirow}
6: %\usepackage{lscape}
7: %\usepackage[square]{natbib}
8: %\usepackage{subfigure}
9:
10: \def\gsim{\;\lower4pt\hbox{${\buildrel\displaystyle >\over\sim}$}\;}
11: \def\lsim{\;\lower4pt\hbox{${\buildrel\displaystyle <\over\sim}$}\;}
12: \def\grls{\;\lower4pt\hbox{${\buildrel\displaystyle >\over <}$}\;}
13:
14: \newcommand{\unsure}[1]{{\color[rgb]{1,0.3,0}\it #1}}
15: \newcommand{\modpart}[1]{{\bf #1}}
16: %\setlength{\parskip}{5pt plus 0.2pt minus 0.1pt} %\setlength{\parindent}{16pt}
17:
18: \shortauthors{WANG \& ZHANG} \shorttitle{Non-CME Associated Flares}
19:
20:
21: \begin{document}
22:
23: %\newcommand{\runningtitle}{Geoeffectiveness of Shock overtaking MC}
24: %\newcommand{\runningauthor}{Wang et al.}
25:
26: \title{A Comparative Study Between Eruptive X-class Flares Associated
27: with Coronal Mass Ejections and Confined X-class Flares}
28:
29: \author{Yuming Wang\altaffilmark{1, 2}, and Jie Zhang\altaffilmark{1}}
30: \altaffiltext{1}{Department of Computational and Data Sciences,
31: College of Science, George Mason University, 4400 University Dr.,
32: MSN 6A2, Fairfax, VA 22030, USA, Email: ywangf@gmu.edu,
33: jzhang7@gmu.edu}
34:
35: \altaffiltext{2}{School of Earth \& Space Sci., Univ. of Sci. \& Tech. of
36: China, Hefei, Anhui 230026, China, Email: ymwang@ustc.edu.cn}
37:
38: \begin{abstract}
39: Following the traditional naming of ``eruptive flare" and
40: ``confined flares" but not implying a causal relationship between
41: flare and coronal mass ejection (CME), we refer to the two kinds
42: of large energetic phenomena occurring in the solar atmosphere as
43: eruptive event and confined event, respectively: the former type
44: refers to flares with associated CMEs, while the later type refers
45: to flares without associated CMEs. We find that about 90\% of
46: X-class flares, the highest class in flare intensity size, are
47: eruptive, but the rest 10\% confined. To probe the question why
48: the largest energy release in the solar corona could be either
49: eruptive or confined, we have made a comparative study by
50: carefully investigating 4 X-class events in each of the two types
51: with a focus on the differences in their magnetic properties. Both
52: sets of events are selected to have very similar intensity (X1.0
53: to X3.6) and duration (rise time less than 13 minutes and decaying
54: time less than 9 minutes) in soft X-ray observations, in order to
55: reduce the bias of flare size on CME occurrence. We find no
56: difference in the total magnetic flux of the photospheric source
57: regions for the two sets of events. However, we find that the
58: occurrence of eruption (or confinement) is sensitive to the
59: displacement of the location of the energy release, which is
60: defined as the distance between the flare site and the
61: flux-weighted magnetic center of the source active region. The
62: displacement is 6 to 17 Mm for confined events, but is as large as
63: 22 to 37 Mm for eruptive events, compared to the typical size of
64: about 70 Mm for active regions studied. In other words, confined
65: events occur closer to the magnetic center while the eruptive
66: events tend to occur closer to the edge of active regions.
67: Further, we have used potential-field source-surface model (PFSS)
68: to infer the 3-D coronal magnetic field above source active
69: regions. For each event, we calculate the coronal flux ratio of
70: low corona ($<$ 1.1 $R_\odot$) to high corona ($\geq$ 1.1
71: $R_\odot$). We find that the confined events have a lower coronal
72: flux ratio ($< 5.7$), while the eruptive events have a higher flux
73: ratio($> 7.1$). These results imply that a stronger overlying
74: arcade field may prevent energy release in the low corona from
75: being eruptive, resulting in flares without CMEs. A CME is more
76: likely to occur if the overlying arcade field is weaker.
77: \end{abstract}
78: \keywords{Sun: coronal mass ejections --- Sun: flares --- Magnetic
79: Fields}
80:
81: \section{Introduction}
82: Coronal mass ejections (CMEs) and flares are known to be the two
83: most energetic phenomena that occur in the atmosphere of the Sun,
84: and have profound effects on the physical environment in the
85: geo-space environment and human technological systems. In this
86: paper, we intend to understand the physical origins of CMEs and
87: flares by comparatively studying two different kinds of energetic
88: phenomena, both of which have almost identical flares, but one is
89: associated with CMEs and the other not. In the past when lacking
90: direct CME observations, these two kinds of phenomena have been
91: called as eruptive flares and confined flares,
92: respectively~\citep[e.g.,][]{Svestka_Cliver_1992}. An eruptive
93: flare (also called a dynamic flare) is usually manifested as
94: having two-ribbons and post-flare loops in H$\alpha$ imaging
95: observations and long duration (e.g., tens of minutes or hours) in
96: soft X-ray, while a confined flare occurs in a compact region and
97: last for only a short period (e.g., minutes). Following this
98: convention but trying not to imply a causal relation between flare
99: and CME, hereafter we refer the flare event associated with an
100: observed CME as an ``eruptive event", and the flare event not
101: associated with a CME as a ``confined event".
102:
103: It has been suggested based on observations that CMEs and flares
104: are the two different phenomena of the same energy release process
105: in the corona~\citep[e.g.,][]{Harrison_1995, Zhang_etal_2001a,
106: Harrison_2003}. They do not drive one another but are closely
107: related. In particular,~\citet{Zhang_etal_2001a}
108: and~\citet{Zhang_etal_2004} showed that the fast acceleration of
109: CMEs in the inner corona coincides very well in time with the rise
110: phase (or energy release phase) of the corresponding soft X-ray
111: flares, strongly implying that both phenomena are driven by the
112: same process at the same time, possibly by magnetic field
113: reconnections. However, this implication raises another important
114: question of under what circumstances the energy release process in
115: the corona leads to an eruption, and under what circumstances it
116: remains to be confined during the process. An answer to this
117: question shall shed the light on the origin of flares as well as
118: CMEs.
119:
120: The occurrence rate of eruptive events depends on the intensity
121: and duration of flares. A statistical study performed
122: by~\citet{Kahler_etal_1989} showed that the flares with longer
123: duration tend to be eruptive, while more impulsive flares tend to
124: be confined. By using the CME data from Solar Maximum Mission
125: (SMM) and the flare data from GOES satellites during 1986 --
126: 1987,~\citet{Harrison_1995} found that the association ratio of
127: flares with CMEs increases from about 7\% to 100\% as the flare
128: class increases from B-class to X-class, and from about 6\% to
129: 50\% as the duration of flares increases from about 1 to 6
130: hours.~\citet{Andrews_2003} examined 229 M and X-class X-ray
131: flares during 1996 -- 1999, and found that the CME-association
132: rate, or eruptive rate is 55\% for M-class flares and 100\% for
133: X-class flares. With a much lager sample of 1301 X-ray
134: flares,~\citet{Yashiro_etal_2005} obtained a similar result that
135: the eruptive rates of C, M, and X-class flares are $16-25$\%,
136: $42-55$\%, and $90-92$\%, respectively.~\citet{Yashiro_etal_2005}
137: work showed that a flare is not necessarily associated with a CME
138: even if it is as intense as an X-class flare. Such kind of
139: confined but extremely energetic events were also reported
140: by~\citet{Feynman_Hundhausen_1994} and~\citet{Green_etal_2002}.
141:
142: The studies mentioned above indicate the probability of CME
143: occurrence for a given flare. On the other hand, there is also a
144: probability of flare occurrence for a given CME. There are CMEs
145: that may not be necessarily associated with any noticeable X-ray
146: flares.~\citet{Zhang_etal_2004} reported an extremely gradually
147: accelerated slow CME without flare association, implying that the
148: non-flare eruptive event tends to be slowly driven. By Combining
149: the corona data from SMM and 6-hr soft X-ray data from GOES
150: satellite,~\citet{St_Cyr_Webb_1991} reported that about 48\%
151: frontside CMEs were associated with X-ray events near the solar
152: minimum of solar cycle 21. Based on SOHO
153: observations,~\citet{Wang_etal_2002a} studied 132 frontside halo
154: CMEs, and found that the association rate of CMEs with X-ray
155: flares greater than C-class increased from 55\% at solar minimum
156: to 80\% near solar maximum. With 197 halo CMEs identified during
157: 1997 -- 2001,~\citet{Zhou_etal_2003} concluded that 88\% CMEs were
158: associated with EUV brightenings.
159:
160: Some attempts have been made to explain the confinement or the
161: eruptiveness of solar energetic events in the context of the
162: configuration of coronal magnetic field.~\citet{Green_etal_2001}
163: analyzed the 2000 September 30 confined event utilizing
164: multi-wavelength data, and suggested that the event involve
165: magnetic reconnection of two closed loops to form two newly closed
166: loops without the opening of the involved magnetic
167: structure.~\citet{Nindos_Andrews_2004} made a statistical study of
168: the role of magnetic helicity in eruption rate. They found that
169: the coronal helicity of active regions producing confined events
170: tends to be smaller than the coronal helicity of those producing
171: eruptive events.
172:
173: In this paper, we address the eruptive-confinement issue of solar
174: energetic events with the approach of a comparative study. We
175: focus on the most energetic confined events that produce X-class
176: soft X-ray flares but without CMEs. While the majority of X-class
177: flares are eruptive, a small percentage (about 10\%) of them are
178: confined. The magnetic properties of these confined events shall
179: be more outstanding than those less-energetic confined events. For
180: making an effective comparison, we select eruptive events with
181: X-ray properties, in terms of intensity and duration, very similar
182: to those selected confined events. The differences on magnetic
183: configuration between these two sets of events shall most likely
184: reveal the true causes of eruption or confinement. How to select
185: events and the basic properties of these events are described in
186: section 2. Detailed comparative analyses of the two sets of events
187: are given in section~\ref{sec_photosphere} and~\ref{sec_corona},
188: which are on the properties of the photospheric magnetic field
189: distribution and the extrapolated coronal magnetic field
190: distribution, respectively. In section~\ref{sec_summary}, we
191: summarize the paper.
192:
193:
194: \section{Selection of Events and Observations}\label{sec_selection}
195:
196: \subsection{Confined Events: X-class Flares without CMEs}
197: From 1996 to 2004, there are 104 X-class soft X-ray flares
198: reported by the NOAA (National Oceanic and Atmosphere
199: Administration) Space Environment Center (SEC). Flares are
200: observed by Geosynchronous Earth Observing Satellites (GOES),
201: which record in high temporal resolution (every 3 seconds) the
202: disk-integrated soft X-ray flux in two pass-bands: 1.0 -- 8.0~\AA\
203: and 0.5 -- 4.0~\AA, respectively. The flare catalog provides the
204: peak intensity, beginning time, peak time, and ending time of
205: flares. Based on peak intensity, flares are classified into five
206: categories: A, B, C, M and X in the order of increasing strength.
207: An X-class flare, in the strongest category, is defined by the
208: peak flux in the 1.0 -- 8.0~\AA\ band exceeding $10^{-4}$
209: Wm$^{-2}$.
210:
211: To find out whether a flare is associated with a CME or not, we
212: make use of both the CME observations by Large Angle and
213: Spectrometric Coronagraph (LASCO,~\citealp{Brueckner_etal_1995})
214: and coronal disk observations by Extreme Ultraviolet Imaging
215: Telescope (EIT,~\citealp{Delaboudiniere_etal_1995}); both
216: instruments are on-board the Solar and Heliospheric Observatory
217: (SOHO) spacecraft. The search process started with the CME
218: catalog\footnote{the NRL-GSFC-CUA CME catalog at
219: http://cdaw.gsfc.nasa.gov/CME\_list/}~\citep{Yashiro_etal_2004}
220: for an initial quick look. A flare becomes a candidate of confined
221: type if there is no any CME whose extrapolated onset time is
222: within the 60-minute time window centered at the fare onset time.
223: The onset time of a CME is calculated by linearly extrapolating
224: the height-time measurement in the outer corona back to the
225: surface of the Sun, which shall give the first-order approximation
226: of the true onset time of CME. Further, we visually examine the
227: sequence of the LASCO and EIT images around the flare time to
228: verify that indeed there is no CME associated with the flare
229: studied. One common property of these confined events is the lack
230: of EUV dimming in EIT images, even though they show strong compact
231: brightenings in EIT images. Following the compact brightening,
232: there is no corresponding CME feature appearing in subsequent
233: LASCO images. This scenario is in sharp contrast to that of an
234: eruptive event, in which an EIT dimming accompanies the
235: brightening, and within a few frames, a distinct CME feature
236: appears in appropriate position angle in LASCO images. After
237: applying this process on all X-class flares, we find 11 events are
238: confined; they are listed in Table~\ref{tb_flares_list}. We notice
239: that event 7 to 11 occurred within three days between July 15 -
240: 17, 2004, and they all originated from the same solar active region
241: (NOAA AR10649).
242:
243: Among the 11 confined X-class flares from 1996 to 2004, the first
244: four events have been reported earlier
245: by~\citet{Yashiro_etal_2005}.The third one has also been reported
246: and studied by~\citet{Green_etal_2002}. During 1996 -- 2004, there
247: are in total 104 X-class flares. Thus the percentage of confined
248: X-class flares is about $10\%$. As shown in
249: Table~\ref{tb_flares_list}, all these confined flares are
250: impulsive. Their rise time does not exceed 13 minutes except for
251: event 8 (23 minutes). The decay time does not
252: exceed 10 minutes for all the 11 events. The rise and decay times
253: are derived from the begin, maximum and end times of
254: flares, which are defined and compiled by
255: NOAA/SEC~\footnote{http://www.sec.noaa.gov/ftpdir/indices/events/README}.
256: The peak intensity of all
257: these events was less than X2.0 level except for event 10 (X3.6).
258: Any event stronger than X3.6 is found always associated with a
259: CME.
260:
261: Out of the 11 confined events, we are able to select 4 of them
262: suitable for further in-depth analysis. These events are numbered
263: as 4, 5, 6 and 11, respectively. They are suitable because (1) the
264: flare is isolated, which means that there is no other flare
265: immediately preceding and following that flare, and (2) there is
266: no other coronal dimming or CME eruption in the vicinity of the
267: flare region within a certain period. Events 1 and 3 are not
268: selected, because they mixed up with a flare-CME pair from the
269: same source regions. In the presence of an eruptive flare
270: immediately preceding or following a confined event, we are not
271: certain how well the confined event is related with the earlier or
272: later eruptive one. In order to make our analysis as ``clean" as
273: possible, such events are discarded. Event 2 is also excluded
274: because its source region is right behind the western limb and
275: hence no timely magnetogram data is available. Events 7 through 11
276: were all from the same active region. By over-plotting EIT images
277: showing flare locations on the MDI magnetogram images, we find
278: that these flares essentially occurred at the same location within
279: the active region. Thus we choose only the last event representing
280: all the five events. The four confined events selected for further
281: analysis are labelled by $C_1$ through $C_4$ in the second column
282: of Table~\ref{tb_flares_list}.
283:
284:
285: \subsection{Eruptive Events: X-class Flares with CMEs}
286: For making the comparative study with the 4 confined events
287: mentioned above, a set of four eruptive X-class flares are
288: selected. These eruptive flares are chosen to have similar
289: properties in X-rays as those confined events:(1) their rise time
290: and decay time are less than 13 minutes, (2) their intensities are
291: between X1.0 to X2.0. Further, their locations are within
292: $60^\circ$ from solar central meridian in longitude in order to
293: reduce the projection effect in magnetograms. These flares, which
294: are relatively impulsive, are indeed associated with CMEs, as
295: shown in LASCO images. The four events are also listed in
296: Table~\ref{tb_flares_list} labelled as $E_1$ through $E_4$,
297: respectively.
298:
299: An overview of the two sets of events is given in
300: Figure~\ref{fg_flare_overview}. The upper panels show the four
301: confined events and the lower panels show the four eruptive
302: events. For each event we show the GOES soft X-ray flux profile
303: (all in a 2-hour interval), running-difference EIT image and
304: running-difference LASCO image in the top, middle and low panels,
305: respectively. Apparently, the temporal profiles of GOES soft X-ray
306: fluxes exhibit no noticeable difference between the two sets of
307: events, due to the constraint in our selection of events.
308: Moreover, as seen in the EIT images, the two sets of events are
309: all associated with compact coronal brightening indicating the
310: occurrence of flares. However, for eruptive flares, the
311: accompanying CMEs are clearly seen in those LASCO images. In
312: contrast, there is no apparent brightening CME feature seen in
313: LASCO images for those confined events (only one image is shown
314: here to represent the observed sequence of images, which all
315: indicate a non-disturbed corona). For eruptive events, the speeds
316: and angular widths of CMEs are also listed in the
317: Table~\ref{tb_flares_list}.
318:
319:
320:
321:
322: \section{Magnetic Properties in the Photosphere}\label{sec_photosphere}
323: \subsection{Flare Location and Active Region Morphology}
324:
325:
326: To explore what physical factors lead the two similar sets of
327: flares, all strong and impulsive, to have difference in CME
328: production, we first study the magnetic properties of their
329: surface source regions. SOHO/MDI provides the observations of
330: photospheric magnetic field (the component along the line of
331: sight) every 96 minutes. The spatial resolution of MDI
332: magnetograms is about 4 arcsec with a plate scale of 2 arcsec per
333: pixel, at which detailed magnetic features across the source
334: active regions are reasonably resolved. To reduce the projection
335: effect of line-of-sight magnetic field, we have only chosen the
336: events within $60^\circ$ from solar central meridian.
337:
338: For each event, we determine the location of the flare seen in EIT
339: relative to the magnetic features seen in MDI. We first align the
340: MDI image with the corresponding EIT image. The difference in the
341: timing between MDI and EIT images have been taken into account.
342: Figure~\ref{fg_pmdieit_example} illustrates the alignment for the
343: 2001 June 23 event. The soft X-ray flare began at 04:02 UT and
344: peaked at 04:09 UT. In EIT 195~\AA\ image showing the flare was
345: taken at 04:11 UT. The nearest full disk MDI image prior to the
346: flare was taken at 03:11 UT. The MDI magnetogram is rotated to fit
347: the EIT time, and then superimposed in contours onto the EIT
348: image. In the right panel of Figure~\ref{fg_pmdieit_example}, we
349: display the aligned images; only the region of interest is shown.
350: Using this method, we are able to determine the location of
351: flares, which is just above the neutral lines seen in the
352: magnetogram.
353:
354: In Figure~\ref{fg_segmdi_noncme} and~\ref{fg_segmdi_cme}, we show
355: the magnetogram images for the four confined events and the four
356: eruptive events, respectively. The flare sites, or bright patches
357: seen in EIT, are marked by red asterisks in the images. The
358: magnetogram images have been re-mapped onto the Carrington
359: coordinate, which reduces the spherical projection effect of the
360: image area. The $x$-axis is Carrington longitude in units of
361: degree, and the $y$-axis is the sine of latitude. The area of the
362: images shown in Figure~\ref{fg_segmdi_noncme}
363: and~\ref{fg_segmdi_cme} are all $30^\circ\times30^\circ$ squares,
364: which usually cover the entire active regions producing the
365: CMEs/flares of interest. To highlight the magnetic features, the
366: displayed images have been segmented into three different levels:
367: strong positive magnetic field ($\geq 50$ Gauss) indicated by
368: white color, strong negative magnetic field ($\leq -50$ Gauss)
369: indicated by black color, and weak field (from $-50$ to 50 Gauss)
370: indicated by gray color. Note that the noise level of a MDI
371: magnetogram image is typically at about 10 Gauss. As shown in the
372: figures, an active region is naturally segmented into many
373: individual pieces. Those pieces with magnetic flux larger than
374: $10^{13}$ wb are labelled by a letter with a number in the
375: neighbouring bracket indicating the magnetic flux in units of
376: $10^{13}$ wb.
377:
378:
379: \subsection{Results}
380: We find that there is no apparent difference in term of total
381: magnetic flux of the source region between the confined events and
382: eruptive events. The total magnetic flux, combining both positive
383: and negative flux, are listed in Table~\ref{tb_photosphere}. The
384: total flux for confined events varies from about 5 to
385: 36$\times10^{13}$ wb, while for eruptive events it varies from
386: about 11 to 24$\times10^{13}$ wb.
387:
388: However, there is a noticeable pattern that the confined flares
389: all originated in a location relatively closer to the center of
390: the host active region. Figure~\ref{fg_segmdi_noncme}a shows the
391: confined event of 2001 June 23. There are three relatively large
392: magnetic pieces labelled by `A', `B' and `C'. The flare site is
393: surrounded by the pieces `A' and `B'.
394: Figure~\ref{fg_segmdi_noncme}b shows the 2003 August 9 confined
395: event. The flare location is associated with three small negative
396: patches (marked by the red asterisks) embedded in a large positive
397: piece `A'. Figure~\ref{fg_segmdi_noncme}c shows the 2004 February
398: 26 confined event. The flare occurred just above the neutral line
399: between thye large positive piece `A' and the large negative piece
400: `B'. Figure~\ref{fg_segmdi_noncme}d shows the 2004 July 17
401: confined event. The flare was located in a complex active region
402: with a large number of sunspots. It occurred right at the boundary
403: between pieces `C' and `F'. From the view of the entire active
404: region, pieces `C' and `F' were further enclosed by two much
405: larger and stronger pieces `A' and `D' whose fluxes were about 10
406: times larger.
407:
408: For those eruptive flares, on the other hand, the flare sites were
409: all relatively further from the center of the magnetic flux
410: distribution. In other words, they were closer to the edge of
411: hosting active regions. Figure~\ref{fg_segmdi_cme}a shows the 1998
412: May 2 event. The strongest pair of magnetic pieces are `A' and
413: `D', but the eruptive flare occurred at the neutral line between
414: pieces `D' and `C', which was the smallest amongst the 4 labelled
415: pieces in the active region. Figure~\ref{fg_segmdi_cme}b shows the
416: 2000 March 2 event. Similarly, the strongest pair of pieces were
417: `A' and `C', but the flare was from the neutral line between
418: pieces `C' and `B', which was the smallest labelled piece.
419: Figure~\ref{fg_segmdi_cme}c shows the 2000 November 24 event. The
420: flare occurred at the outer edge of the strongest piece `A', which
421: was neighbored by a very small region with negative flux.
422: Figure~\ref{fg_segmdi_cme}d shows the 2004 October 30 event. The
423: flare site was also close to the edge of the entire active region.
424:
425:
426:
427:
428:
429: To quantify this observation of different displacements of flare
430: locations, we here introduce a flare displacement parameter, which
431: is defined by the surface distance between the flare site and the
432: weighted center of the magnetic flux distribution of the host
433: active region, or center of magnetic flux (COM) for short. The COM
434: might be the place that has the most overlying magnetic flux. The
435: COM is calculated based on the re-mapped $30^\circ\times30^\circ$
436: MDI images (without segmentation). It is a point, across which any
437: line can split the magnetogram into two flux-balanced halves, and
438: can be formulated as $x_c=\frac{\sum_iF_i*x_i}{\sum_iF_i}$ and
439: $y_c=\frac{\sum_iF_i*y_i}{\sum_iF_i}$. The COM of these events
440: have been marked by the diamonds in Figure~\ref{fg_segmdi_noncme}
441: and~\ref{fg_segmdi_cme}. With known COM, it is easy to derive the
442: displacement parameter, which is listed in
443: Table~\ref{tb_photosphere}. Consistent with the earlier
444: discussions, it is found that the displacement parameters for the
445: four confined events are all smaller than 17 Mm, while for the
446: four eruptive events, they are all larger than 22 Mm.
447:
448: We now consider possible errors in calculating the displacement
449: parameter. The error main arises from the uncertainty in the
450: recorded weak magnetic field around the active regions. However,
451: in the selected region of study that is $30^\circ\times30^\circ$
452: across, the highlighted white and black pieces contain about 99\%
453: of the total magnetic flux in the region. Therefore the
454: uncertainty of the flux is expected to be at the order of 1\%.
455: Considering the formula of the coordinates of COM given in the
456: last paragraph and assuming a typical scale of 100 Mm, the error
457: of the calculated distance is about 1 Mm. Further, considering the
458: spatial resolution of MDI of $\sim1$ Mm (varing from $\sim0.7$ Mm
459: at central meridian to $\sim1.4$ Mm at longitude of
460: $\pm60^\circ$), the overall uncertainty should be about $\pm2$ Mm.
461: With these consideration, the displacement parameters and their
462: uncertainties are plotted in Figure~\ref{fg_ratio_distance}. The
463: confined events are indicated by diamonds, and the eruptive events
464: are indicated by asterisks. The vertical dashed line, which
465: corresponds to a displacement of about 19.5 Mm, effectively
466: separates the two sets of events.
467:
468:
469:
470:
471:
472: \section{Magnetic Properties in the Corona}\label{sec_corona}
473: \subsection{Method}
474: Having studied the magnetic field distributions in the
475: photosphere, we further investigate into magnetic field
476: distributions in the 3-D corona. The magnetic field configuration
477: in the corona shall ultimately determine the eruption/confinement
478: since the energy releases occurs in the corona. There is so far no
479: direct observations of coronal magnetic fields. We have to utilize
480: certain models to calculate the coronal magnetic field based on
481: observed photospheric boundary. In this paper, we apply the
482: commonly used potential-field source-surface (PFSS)
483: model~\citep[e.g.,][]{Schatten_etal_1969, Hoeksema_etal_1982}. The
484: PFSS model is thought to be a useful first-order approximation to
485: the global magnetic field of the solar corona. Nevertheless, we
486: realize that the current-carrying core fields, which are low-lying
487: and near the magnetic neutral line, are far from the potential
488: field approximation; these core fields are likely to be the
489: driving source of any energy release in the corona. Therefore, the
490: usage of PFSS model in this study is limited to calculate the
491: total flux of the overlying large scale coronal field, which is
492: believed to be closer to a potential approximation. These
493: overlying fields are thought to constrain the low-lying field from
494: eruption.
495:
496: A modified MDI magnetic field synoptic chart is used as input to
497: our PFSS model. The high-resolution MDI magnetic field synoptic
498: chart\footnote{http://soi.stanford.edu/magnetic/index6.html} is
499: created by interpolating data to disk-center resolution, resulting
500: in a $3600\times1080$ pixel map. The X and Y axis are linear in
501: Carrington longitude ($0.1^\circ$ intervals) and sine latitude,
502: respectively. This high resolution is useful in creating detailed
503: coronal magnetic field above the source regions of interest. Since
504: an MDI synoptic chart is created from the magnetogram images over
505: a $\sim27$-day solar rotation, the synoptic chart does not exactly
506: represent the photospheric magnetic field in the region of
507: interest at the flare/CME time. The details of the source region
508: may be different because of the evolution of photospheric magnetic
509: field. To mitigate this problem, we use the MDI daily magnetogram
510: to update the original MDI synoptic chart. The process is to
511: re-map the snapshot magnetogram image prior to the flare
512: occurrence to the Carrington grid, and then slice out the region
513: of interest, $30^\circ$ in longitude and $60^\circ$ in latitude.
514: This sliced region, to replace the corresponding portion in the
515: original synoptic chart.
516:
517: Since a PFSS model makes use of the spheric harmonic series
518: expansion, we realize that a high-resolution data requires a high
519: order expansion in order to have a consistent result. We calculate
520: the spheric harmonic coefficients to as high as 225 order for the
521: input $3600\times1080$ boundary image. We find that, at this
522: order, we can get the best match between the calculated
523: photospheric magnetic field and the input synoptic chart. The mean
524: value of the difference between them is less than 0.5 Gauss, and
525: the standard deviation is generally $\lsim 15$ Gauss for solar
526: minimum and $\lsim 25$ Gauss for solar maximum, which are
527: comparable to the noise level of MDI magnetograms. That means, we
528: can effectively reproduce the observed photospheric magnetic field
529: with the 225-order PFSS model. With spheric harmonic coefficients
530: known, it is relatively straightforward to calculate the magnetic
531: field in the 3-D volume of the corona.
532:
533:
534:
535: \subsection{Results}
536: Figure~\ref{fg_mag_n_20040717} and~\ref{fg_mag_c_20041030} show
537: the extrapolated coronal magnetic field lines for one confined
538: event (2004 July 17) and one eruptive event (2004 October 30),
539: respectively. In each figure, the left panel shows field lines
540: viewed from top, and the right panel shows field lines viewed from
541: the side by rotating the left panel view of $60$ degrees into the
542: paper. The green-yellow colors denote the closed field lines, with
543: green indicating the loop part of outward magnetic field (positive
544: magnetic polarity at the footpoints) and yellow the inward
545: (negative magnetic polarity at the footpoints), and the blue color
546: indicates the open field lines. The two examples show that the
547: location of the confined flare, which is near the center of the
548: active region, is covered by a large tuft of overlying magnetic
549: loop arcades, while the location of the eruptive flare, which is
550: near the edge of the active region, has relatively few directly
551: overlying loop arcade. In particular, for the eruptive event, the
552: nearby positive and negative magnetic field lines seem to connect
553: divergently with other regions, instead of forming a loop arcade
554: of its own.
555:
556: To quantify the strength of the overlying field, we calculate the
557: total magnetic flux cross the plane with the $x$ direction
558: extending along the neutral line and the $y$ direction vertically
559: along the radial direction. The thick blue lines on the
560: photospheric surface in Figure~\ref{fg_segmdi_noncme}
561: and~\ref{fg_segmdi_cme} indicate the neutral lines used in the
562: calculation. The length of the neutral lines is determined as it
563: encompasses the major part of the eruption region. The overlying
564: magnetic field flux then is normalized to the length of the
565: neutral line. Thus obtained normalized overlying magnetic flux is
566: a better quantity to be used for comparison between different
567: events, because this parameter is not sensitive to the exact
568: length of the neutral lines selected, which may vary
569: significantly from event to event.
570:
571:
572: Such calculated magnetic fluxes for the 8 events are listed in
573: Table~\ref{tb_flux}. In calculating the flux, we do not consider
574: that the crossing direction of the field lines over the neutral
575: line is from one side to the other or opposite. The relative
576: uncertainty of the calculated magnetic flux can be estimated as
577: $\sigma_B/B_0$ where $\sigma_B$ is the uncertainty of calculated
578: magnetic field strength in the corona and $B_0$ is the magnetic
579: field strength in the active regions at photosphere. Considering
580: $\sigma_B$ is about 15 to 25 Gauss, the standard deviation
581: mentioned before, and $B_0$ is usually hundreds of Gauss, we infer
582: that the uncertainty of overlying magnetic flux is about 10\%.
583: The estimate should be true in case that the coronal magnetic field
584: is correctly obtained by our extrapolation method. If the
585: extrapolated field largely deviates from the realistic status, the
586: uncertainty may probably be slightly different.
587:
588:
589: The total overlying flux, $F_{total}$, in the height range from
590: 1.0 to 1.5 $R_\odot$, is given in the third column. It seems that
591: there is no systematic difference between the confined events and
592: eruptive events. The value for the confined events varies from
593: 0.40 to 1.27 $\times 10^{10}$ wb/Mm, and that for the eruptive
594: events from 0.73 to 1.34 $\times10^{10}$ wb/Mm.
595:
596: We further calculate the flux in two different height ranges, the
597: lower flux from 1.0 to 1.1 $R_\odot$ and the higher flux from 1.1
598: to 1.5 $R_\odot$. A common accepted scenario is that the lower
599: flux shall correspond to the inner sheared core field (or
600: fully-fledged flux rope if filament present) that tends to move
601: out, while the outer flux is the large scale overlying arcade that
602: tends to constrain the inner flux from eruption. Note that the
603: chosen of 1.1 $R_\odot$, which corresponds to a height of about 70
604: Mm above the surface, is rather arbitrary. However, slightly
605: changing this number will not affect any overall results that will
606: be reached. The magnetic flux in the low corona, $F_{low}$, and in
607: the high corona, $F_{high}$, are listed in the fourth and fifth
608: columns of Table~\ref{tb_flux}, respectively.
609:
610: There is a trend that the low-corona flux for the eruptive events
611: is generally larger than that for the confined events. Three out
612: of the four eruptive events have their low-corona overlying flux
613: more than 1.0 $\times10^{10}$ wb/Mm, while three out of the four
614: confined events have the flux less than 1.0 $\times10^{10}$ wb/Mm.
615: On the other hand, the high-corona flux for the eruptive events
616: seems smaller than that for the confined events. Three confined
617: events have their high-corona flux more than 0.15 $\times10^{10}$
618: wb/Mm, while all four eruptive events have the flux less than 0.15
619: $\times10^{10}$ wb/Mm.
620:
621: We further calculate the flux ratio parameter, which is defined as
622: $R=F_{low}/F_{high}$. This quantity is independent of the normalization.
623: It may serve as an index of how weak the
624: constraint on the inner eruptive field is. Interestingly, the flux
625: ratios for the two sets of events fall into two distinct groups.
626: For the confined events, $R$ varies from 1.59 to 5.68, while for
627: the eruptive events, the value of $R$ is larger, from 7.11 to
628: 10.17. The value of 6.5 may be used as a boundary separating the
629: two sets of events. This value probably implies a threshold for
630: confinement or eruptiveness. This is to say, if the flux ratio is
631: less than 6.5, a flare is likely to be confined, otherwise
632: eruptive. The higher the ratio, the higher the possibility of a
633: coronal energy release being eruptive.
634:
635:
636:
637:
638:
639: \section{Summary and Discussions}\label{sec_summary}
640: In summary, among the 104 X-class flares occurred during 1996 --
641: 2004, we found a total of 11 ($\sim10\%$) are confined flares
642: without associated CMEs, and all the others ($\sim90\%$) are
643: eruptive flares associated with CMEs. Four suitable confined
644: flares are selected to make a comparative study with four eruptive
645: flares, which are similar in X-ray intensity and duration as those
646: confined events. We have carefully studied the magnetic properties
647: of these events both in the photosphere and in the corona. The
648: following results are obtained:
649:
650: (1) In the photosphere, we can not find a difference of the total
651: magnetic fluxes in the surface source regions between the two sets
652: of events. However, there is an apparent difference in the
653: displacement parameter, which is defined as the surface distance
654: between the flare site and the center of magnetic flux
655: distribution. For the confined events, the displacement is from 6
656: to 17 Mm, while for those eruptive events it is from 22 to 37 Mm.
657: This result implies that the energy release occurring in the
658: center of an active region is more difficult to have a complete
659: open eruption, resulting in a flare without CME. On the other
660: hand, the energy release occurring away from the magnetic center
661: has a higher probability to have an eruption, resulting in both
662: flares and CMEs. Whether an eruption could occur or not may be
663: strongly constrained by the overlying large scale coronal magnetic
664: field. The overlying coronal magnetic field shall be strongest and
665: also longest along the vertical direction over the center of an
666: active region. On the other hand, the overlying constraining field
667: shall be weaker if the source is away from the center. This
668: scenario is further supported by our study of coronal magnetic
669: field.
670:
671: (2) Calculation of coronal magnetic field shows that the flux
672: ratio of the magnetic flux in the low corona to that in the high
673: corona is systematically larger for the eruptive events than that
674: for the confined events. The magnetic flux ratio for the confined
675: events varies from 1.6 to 5.7, while the ratio for the eruptive
676: events from 7.1 to 10.2. However, there is no evident difference
677: between the two sets of events in the total magnetic flux straddling
678: over neutral lines, and there is only a weak trend indicating a
679: systematic difference in the low- and high-corona magnetic fluxes.
680: This low-to-high corona magnetic flux ratio serves as a proxy of
681: the strength of the inner core magnetic field, which may play an
682: erupting role, relative to the strength of the overlying large
683: scale coronal magnetic field, which may play a constraining role
684: to prevent eruption. The lower this ratio, the more difficult the
685: energy release in the low corona can be eruptive.
686:
687: There is variety of theoretical models on the initiation mechanism
688: of CMEs and the energy release of flares
689: ~\citep[e.g.,][]{Sturrock_1989, Chen_1989,
690: vanBallegooijen_Martens_1989, Forbes_Isenberg_1991,
691: Moore_Roumeliotis_1992, Low_Smith_1993, Mikic_Linker_1994,
692: Antiochos_etal_1999, Lin_Forbes_2000}. These models differ in
693: pre-eruption magnetic configurations, trigger processes, or where
694: magnetic reconnection occurs. Nevertheless, in almost all these
695: models, the magnetic configuration involves two magnetic regimes,
696: one is the core field in the inner corona close to the neutral
697: line, the other is the large scale overlying field or background
698: field. The core field is treated as highly sheared or as a
699: fully-fledged flux rope; in either case, the core field stores
700: free energy for release. On the other hand, the overlying field is
701: regarded as potential and considered to be main constraining force
702: to prevent the underlying core field from eruption or
703: escaping.~\citet{Torok_Kliem_2005} and~\citet{Kliem_Torok_2006}
704: recently pointed out that the
705: decrease of the overlying field with height is a main factor in
706: deciding whether the kink-instability (in their twist flux rope
707: model) leads to a confined event or a CME. On the other
708: hand,~\citet{Mandrini_etal_2005} reported the smallest CME event ever
709: observed by 2005, in which the CME originated from the smallest source
710: region, a tiny dipole, and developed into the smallest magnetic cloud.
711: They suggested that the ejections of tiny flux ropes are possible.
712: Therefore, it is reasonable to argue that, whether an energy release in the corona
713: is eruptive or confined, is sensitive to the balance between the
714: inner core field and the outer overlying field. Our observational
715: results seem to be consistent with this scenario.
716:
717: This study is only a preliminary step to investigate the
718: confinement and/or eruptiveness of solar flares, or coronal energy
719: releases in general. However, it demonstrates that the
720: distribution of magnetic field both in the photosphere and in the
721: corona may effectively provide the clue of the possible nature of
722: an energetic event: whether a flare, a CME or both. To further
723: evaluate the effectiveness of this methodology, a more robust
724: study involving more events is needed.
725:
726:
727: \acknowledgments{We acknowledge the use of the solar data from the
728: LASCO, EIT and MDI instruments on board SOHO spacecraft. The
729: SOHO/LASCO data used here are produced by a consortium of the
730: Naval Research Laboratory (USA), Max-Planck-Institut fuer
731: Aeronomie (Germany), Laboratoire d'Astronomie (France), and the
732: University of Birmingham (UK). SOHO is a project of international
733: cooperation between ESA and NASA. We also acknowledge the use of
734: CME catalog generated and maintained at the CDAW Data Center by
735: NASA and The Catholic University of America in cooperation with
736: the Naval Research Laboratory, and the solar event reports
737: compiled by the Space Environment Center of NOAA. We thank the
738: useful discussion with X. P. Zhao at Stanford University, who
739: provides the procedure of PFSS model. This work is supported by
740: NSF SHINE grant ATM-0454612 and NASA grant NNG05GG19G. Y. Wang is
741: also supported by the grants from NSF of China (40525014) and
742: MSTC (2006CB806304), and J. Zhang is also supported by NASA
743: grants NNG04GN36G.}
744:
745:
746: %\bibliographystyle{kluwer}
747: %\bibliography{../../ahareference}
748:
749: \begin{thebibliography}{xx}
750:
751: \harvarditem{Andrews}{2003}{Andrews_2003}
752: Andrews, M.~D.: 2003, A search for {CMEs} associated with big flares, {\em Sol.
753: Phys.} {\bf 218},~261--279.
754:
755: \harvarditem[Antiochos et~al.]{Antiochos, DeVore and
756: Klimchuk}{1999}{Antiochos_etal_1999}
757: Antiochos, S.~K., DeVore, C.~R. and Klimchuk, J.~A.: 1999, A model for solar
758: coronal mass ejections, {\em Astrophys. J.} {\bf 510},~485--493.
759:
760: \harvarditem[Brueckner et~al.]{Brueckner, Howard, Koomen, Korendyke, Michels,
761: Moses, Socker, Dere, Lamy, Llebaria, Bout, Schwenn, Simnett, Bedford and
762: Eyles}{1995}{Brueckner_etal_1995}
763: Brueckner, G.~E., Howard, R.~A., Koomen, M.~J., Korendyke, C.~M., Michels,
764: D.~J., Moses, J.~D., Socker, D.~G., Dere, K.~P., Lamy, P.~L., Llebaria, A.,
765: Bout, M.~V., Schwenn, R., Simnett, G.~M., Bedford, D.~K. and Eyles, C.~J.:
766: 1995, The large angle spectroscopic coronagraph ({LASCO}), {\em Sol. Phys.}
767: {\bf 162},~357--402.
768:
769: \harvarditem{Chen}{1989}{Chen_1989}
770: Chen, J.: 1989, Effects of toroidal forces in current loops embedded in a
771: background plasma, {\em Astrophys. J.} {\bf 338},~453--470.
772:
773: \harvarditem[Delaboudiniere et~al.]{Delaboudiniere, Artzner, Brunaud and
774: et~al.}{1995}{Delaboudiniere_etal_1995}
775: Delaboudiniere, J.-P., Artzner, G.~E., Brunaud, J. and et~al.: 1995, {EIT}:
776: Extreme-ultraviolet imaging telescope for the {SOHO} mission, {\em Sol.
777: Phys.} {\bf 162},~291--312.
778:
779: \harvarditem{Feynman and Hundhausen}{1994}{Feynman_Hundhausen_1994}
780: Feynman, J. and Hundhausen, A.~J.: 1994, Coronal mass ejections and major solar
781: flares: The great active center of march 1989, {\em J. Geophys. Res.} {\bf
782: 99(A5)},~8451--8464.
783:
784: \harvarditem{Forbes and Isenberg}{1991}{Forbes_Isenberg_1991}
785: Forbes, T.~G. and Isenberg, P.~A.: 1991, A catastrophe mechanism for coronal
786: mass ejections, {\em Astrophys. J.} {\bf 373},~294--307.
787:
788: \harvarditem[Green et~al.]{Green, Harra, Matthews and
789: Culhane}{2001}{Green_etal_2001}
790: Green, L.~M., Harra, L.~K., Matthews, S.~A. and Culhane, J.~L.: 2001, Coronal
791: mass ejections and their association to active region flaring, {\em Sol.
792: Phys.} {\bf 200},~189--202.
793:
794: \harvarditem[Green et~al.]{Green, Matthews, {van Driel-Gesztelyi}, Harra and
795: Culhane}{2002}{Green_etal_2002}
796: Green, L.~M., Matthews, S.~A., {van Driel-Gesztelyi}, L., Harra, L.~K. and
797: Culhane, J.~L.: 2002, Multi-wavelength observations of an x-class flare
798: without a coronal mass ejection, {\em Sol. Phys.} {\bf 205},~325--339.
799:
800: \harvarditem{Harrison}{1995}{Harrison_1995}
801: Harrison, R.~A.: 1995, The nature of solar flares associated with coronal mass
802: ejection, {\em Astron. \& Astrophys.} {\bf 304},~585--594.
803:
804: \harvarditem{Harrison}{2003}{Harrison_2003}
805: Harrison, R.~A.: 2003, Soho observations relating to the association between
806: flares and coronal mass ejections, {\em Adv. Space Res.} {\bf
807: 32(12)},~2425--2437.
808:
809: \harvarditem[Hoeksema et~al.]{Hoeksema, Wilcox and
810: Scherrer}{1982}{Hoeksema_etal_1982}
811: Hoeksema, J.~T., Wilcox, J.~M. and Scherrer, P.~H.: 1982, Structure of the
812: heliospheric current sheet in the early portion of sunspot cycle 21, {\em J.
813: Geophys. Res.} {\bf 87},~10331--10338.
814:
815: \harvarditem[Kahler et~al.]{Kahler, {Sheeley, Jr.} and
816: Liggett}{1989}{Kahler_etal_1989}
817: Kahler, S.~W., {Sheeley, Jr.}, N.~R. and Liggett, M.: 1989, Coronal mass
818: ejections and associated {X-ray} flare durations, {\em Astrophys. J.} {\bf
819: 344},~1026--1033.
820:
821: \harvarditem{Kliem and Torok}{2006}{Kliem_Torok_2006}
822: Kliem, B., and Torok, T.: 2006, Torus Instability, {\em Phys. Rev. Lett.} {\bf
823: 96},~255002.
824:
825: \harvarditem{Lin and Forbes}{2000}{Lin_Forbes_2000}
826: Lin, J. and Forbes, T.~G.: 2000, Effects of reconnection on the coronal mass
827: ejection process, {\em J. Geophys. Res.} {\bf 105(A2)},~2375--2392.
828:
829: \harvarditem{Low and Smith}{1993}{Low_Smith_1993}
830: Low, B.~C. and Smith, D.~F.: 1993, The free energies of partially open coronal
831: magnetic fields, {\em Astrophys. J.} {\bf 410},~412--425.
832:
833: \harvarditem[Mandrini et~al.]{Mandrini, Pohjolainen, Dasso, Green, Demoulin, van Driel-Gesztelyi, Copperwheat and Foley}{2005}{Mandrini_etal_2005}
834: Mandrini, C.~H., Pohjolainen, S., Dasso, S., Green, L.~M., Demoulin, P., van Driel-Gesztelyi, L., Copperwheat, C. and Foley, C.: 2005, Interplanetary flux rope ejected from an X-ray bright point: The smallest magnetic cloud source-region ever observed, {\em Astron. \& Astrophys.} {\bf 434},~725--740.
835:
836: \harvarditem{Mikic and Linker}{1994}{Mikic_Linker_1994}
837: Mikic, Z. and Linker, J.~A.: 1994, Disruption of coronal magnetic field
838: arcades, {\em Astrophys. J.} {\bf 430},~898--912.
839:
840: \harvarditem{Moore and Roumeliotis}{1992}{Moore_Roumeliotis_1992}
841: Moore, R.~L. and Roumeliotis, G.: 1992, Triggering of eruptive flares -
842: destabilization of the preflare magnetic field configuration, {\em in}
843: Z.~Svestka, B.~V. Jackson and M.~Machado (eds), {\em Eruptive Solar Flares},
844: Proceedings of IAU Colloquium No. 113, Springer-Verlag, New York,
845: pp.~107--121.
846:
847: \harvarditem{Nindos and Andrews}{2004}{Nindos_Andrews_2004}
848: Nindos, A. and Andrews, M.~D.: 2004, The association of big flares and coronal
849: mass ejections: What is the role of magnetic helicity?, {\em Astrophys. J.}
850: {\bf 616},~L175--L178.
851:
852: \harvarditem[Schatten et~al.]{Schatten, Wilcox and
853: Ness}{1969}{Schatten_etal_1969}
854: Schatten, K.~H., Wilcox, J.~M. and Ness, N.~F.: 1969, A model of interplanetary
855: and coronal magnetic fields, {\em Sol. Phys.} {\bf 6},~442--455.
856:
857: \harvarditem{{St. Cyr} and Webb}{1991}{St_Cyr_Webb_1991}
858: {St. Cyr}, O.~C. and Webb, D.~F.: 1991, Activity associated with coronal mass
859: ejections at solar minimum - {SMM} observations from 1984-1986, {\em Sol.
860: Phys.} {\bf 136},~379--394.
861:
862: \harvarditem{Sturrock}{1989}{Sturrock_1989}
863: Sturrock, P.~A.: 1989, The role of eruption in solar flares, {\em Sol. Phys.}
864: {\bf 121},~387--397.
865:
866: \harvarditem{Svestka and Cliver}{1992}{Svestka_Cliver_1992}
867: Svestka, Z. and Cliver, E.~W.: 1992, History and basic characteristics of
868: eruptive flares, {\em in} Z.~Svestka, B.~V. Jackson and M.~E. Machado (eds),
869: {\em Eruptive Solar Flares}, Proceedings of IAU Colloquium No. 113,
870: Springer-Verlag, New York, pp.~1--11.
871:
872: \harvarditem{Torok and Kliem}{2005}{Torok_Kliem_2005}
873: Torok, T. and Kliem, B.: 2005, Confined and ejective eruptions of kink-unstable
874: flux ropes, {\em Astrophys. J.} {\bf 630},~L97--L100.
875:
876: \harvarditem{{van Ballegooijen} and
877: Martens}{1989}{vanBallegooijen_Martens_1989}
878: {van Ballegooijen}, A.~A. and Martens, P. C.~H.: 1989, Formation and eruption
879: of solar prominences, {\em Astrophys. J.} {\bf 343},~971--984.
880:
881: \harvarditem[Wang et~al.]{Wang, Ye, Wang, Zhou and Wang}{2002}{Wang_etal_2002a}
882: Wang, Y.~M., Ye, P.~Z., Wang, S., Zhou, G.~P. and Wang, J.~X.: 2002, A
883: statistical study on the geoeffectiveness of earth-directed coronal mass
884: ejections from {March} 1997 to {December} 2000, {\em J. Geophys. Res.} {\bf
885: 107(A11)},~1340, doi:10.1029/2002JA009244.
886:
887: \harvarditem[Yashiro et~al.]{Yashiro, Gopalswamy, Akiyama, Michalek and
888: Howard}{2005}{Yashiro_etal_2005}
889: Yashiro, S., Gopalswamy, N., Akiyama, S., Michalek, G. and Howard, R.~A.: 2005,
890: Visibility of coronal mass ejections as a function of flare location and
891: intensity, {\em J. Geophys. Res.} {\bf 110(A12)},~A12S05.
892:
893: \harvarditem[Yashiro et~al.]{Yashiro, Gopalswamy, Michalek, {St. Cyr},
894: Plunkett, Rich and Howard}{2004}{Yashiro_etal_2004}
895: Yashiro, S., Gopalswamy, N., Michalek, G., {St. Cyr}, O.~C., Plunkett, S.~P.,
896: Rich, N.~B. and Howard, R.~A.: 2004, A catalog of white light coronal mass
897: ejections observed by the soho spacecraft, {\em J. Geophys. Res.} {\bf
898: 109(A7)},~A07105.
899:
900: \harvarditem[Zhang et~al.]{Zhang, Dere, Howard, Kundu and
901: White}{2001}{Zhang_etal_2001a}
902: Zhang, J., Dere, K.~P., Howard, R.~A., Kundu, M.~R. and White, S.~M.: 2001, On
903: the temporal relationship between coronal mass ejections and flares, {\em
904: Astrophys. J.} {\bf 559},~452--462.
905:
906: \harvarditem[Zhang et~al.]{Zhang, Dere, Howard and
907: Vourlidas}{2004}{Zhang_etal_2004}
908: Zhang, J., Dere, K.~P., Howard, R.~A. and Vourlidas, A.: 2004, A study of the
909: kinematic evolution of coronal mass ejections, {\em Astrophys. J.} {\bf
910: 604},~420--432.
911:
912: \harvarditem[Zhou et~al.]{Zhou, Wang and Cao}{2003}{Zhou_etal_2003}
913: Zhou, G., Wang, J. and Cao, Z.: 2003, Correlation between halo coronal mass
914: ejections and solar surface activity, {\em Astron. \& Astrophys.} {\bf
915: 397},~1057.
916:
917: \end{thebibliography}
918:
919: \clearpage
920:
921: \begin{table}[t]%\vskip -100pt
922: \linespread{1.5} \caption{List of Confined X-class Flares from
923: 1996 to 2004 and Selected Eruptive Flares} \label{tb_flares_list}
924: \tabcolsep 2pt
925: \footnotesize %\centering
926: \begin{tabular}{lcccccccccp{100pt}}
927: \hline
928: \# &Label &Date &Begin &$T_R^a$ &$T_D^b$ &Class &Location &NOAA &CME$^c$ &Comment \\
929: & & &UT &min &min & & &AR &V(km/s)/Width & \\
930: \hline
931: \multicolumn{11}{c}{Confined Flares}\\
932: 1 & &2000/06/06 &13:30 &9.0 &7.0 &X1.1 &N20E18 &9026 &- &Contained by a preceding and a following M-class flares (Y)\\
933: 2 & &2000/09/30 &23:13 &8.0 &7.0 &X1.2 &N07W91 &9169 &- &Limb event (G, Y)\\
934: 3 & &2001/04/02 &10:04 &10.0 &6.0 &X1.4 &N17W60 &9393 &- &Contained by a preceding eruptive flare (Y)\\
935: 4 &$C_1$ &2001/06/23 &04:02 &6.0 &3.0 &X1.2 &N10E23 &9511 &- &(Y) \\
936: 5 &$C_2$ &2003/06/09 &21:31 &8.0 &4.0 &X1.7 &N12W33 &10374 &- & \\
937: 6 &$C_3$ &2004/02/26 &01:50 &13.0 &7.0 &X1.1 &N14W14 &10564 &- & \\
938: 7 & &2004/07/15 &18:15 &9.0 &4.0 &X1.6 &S11E45 &10649 &- & \\
939: 8 & &2004/07/16 &01:43 &23.0 &6.0 &X1.3 &S11E41 &10649 &- & \\
940: 9 & &2004/07/16 &10:32 &9.0 &5.0 &X1.1 &S10E36 &10649 &- & \\
941: 10 & &2004/07/16 &13:49 &6.0 &6.0 &X3.6 &S10E35 &10649 &- & \\
942: 11&$C_4$ &2004/07/17 &07:51 &6.0 &2.0 &X1.0 &S11E24 &10649 &- &Event 7--11 all from the same AR \\
943: \hline
944: \multicolumn{11}{c}{Eruptive Flares}\\
945: 1 &$E_1$ &1998/05/02 &13:31 &11.0 &9.0 &X1.1 &S15W15 &8210 &936/halo &\\
946: 2 &$E_2$ &2000/03/02 &08:20 &8.0 &3.0 &X1.1 &S18W54 &8882 &776/62$^\circ$ &\\
947: 3 &$E_3$ &2000/11/24 &04:55 &7.0 &6.0 &X2.0 &N19W05 &9236 &1289/halo &\\
948: 4 &$E_4$ &2004/10/30 &11:38 &8.0 &4.0 &X1.2 &N13W25 &10691 &427/halo &\\
949: \hline
950: \end{tabular}\\
951: $^a$ Rise time of flares.\\
952: $^b$ Decay time of flares.\\
953: $^c$ Apparent speed and angular width of CMEs. Adopted from the online GSFC-NRL-CUA CME catalog.\\
954: G and Y in comment column mean that the corresponding events have
955: been reported by~\citet{Green_etal_2002}
956: and~\citet{Yashiro_etal_2005}, respectively.
957: \end{table}
958:
959: \clearpage
960:
961: \begin{table}[t]%\vskip -100pt
962: \linespread{1.5} \caption{Magnetic Properties of the Source Active
963: Regions of the Confined and Eruptive Flares}
964: \label{tb_photosphere} \tabcolsep 2pt
965: \footnotesize %\centering
966: \begin{tabular}{lccc}
967: \hline
968: Event &Date &Flux$^a$ &Distance$^b$ \\
969: & &$10^{13}$ wb &Mm \\
970: \hline
971: \multicolumn{4}{c}{Confined Flares} \\
972: $C_1$ &2001/06/23 &5 &6 \\
973: $C_2$ &2003/06/09 &36 &17 \\
974: $C_3$ &2004/02/26 &23 &8 \\
975: $C_4$&2004/07/17 &34 &10 \\
976: \multicolumn{4}{c}{Eruptive Flares} \\
977: $E_1$ &1998/05/02 &17 &22 \\
978: $E_2$ &2000/03/02 &24 &33 \\
979: $E_3$ &2000/11/24 &18 &37 \\
980: $E_4$ &2004/10/30 &11 &29 \\
981: \hline
982: \end{tabular}\\
983: $^a$ Total magnetic flux in active regions measured in MDI magnetogram.\\
984: $^b$ Surface distance between the flare site and the COM of the
985: associated active region.
986: \end{table}
987:
988: \clearpage
989:
990: \begin{table}[t]%\vskip -100pt
991: \linespread{1.5} \caption{Magnetic flux per unit length overlying the neutral lines}
992: \label{tb_flux} \tabcolsep 2pt
993: \footnotesize %\centering
994: \begin{tabular}{lccccc}
995: \hline
996: Event &Date &$F_{total}$ &$F_{low}$ &$F_{high}$ &Ratio$=\frac{F_{low}}{F_{high}}$ \\
997: & &$10^{10}$ wb/Mm &$10^{10}$ wb/Mm &$10^{10}$ wb/Mm & \\
998: \hline
999: \multicolumn{6}{c}{Confined Flares} \\
1000: $C_1$ &2001/06/23 &0.40 &0.34 &0.06 &5.67 \\
1001: $C_2$ &2003/06/09 &0.83 &0.61 &0.22 &2.77 \\
1002: $C_3$ &2004/02/26 &1.27 &1.08 &0.19 &5.68 \\
1003: $C_4$ &2004/07/17 &1.19 &0.73 &0.46 &1.59 \\
1004: \multicolumn{6}{c}{Eruptive Flares} \\
1005: $E_1$ &1998/05/02 &1.34 &1.22 &0.12 &10.17 \\
1006: $E_2$ &2000/03/02 &1.17 &1.06 &0.11 & 9.64 \\
1007: $E_3$ &2000/11/24 &1.14 &1.03 &0.11 & 9.36 \\
1008: $E_4$ &2004/10/30 &0.73 &0.64 &0.09 & 7.11 \\
1009: \hline
1010: \end{tabular}
1011: \end{table}
1012:
1013: \clearpage
1014:
1015: \begin{figure}[tb]
1016: \centering
1017: \includegraphics[width=0.8\hsize]{f1.eps}
1018: \caption{An overview of the four confined ($C_1 - C_4$, upper panels) and four eruptive
1019: ($E_1 - E_2$, lower panels) flares. For each events, we display
1020: its GOES X-ray flux profile (spanning 2 hours), running-difference
1021: images of EIT 195\AA\ and LASCO/C2 in the three sub-panels from
1022: top to bottom.}
1023: \label{fg_flare_overview}
1024: \end{figure}
1025:
1026: \clearpage
1027:
1028: \begin{figure}[tb]
1029: \centering
1030: \includegraphics[width=\hsize]{f2.ps}
1031: \caption{An example showing flare and its source region. The left
1032: image is a full disk MDI magnetogram taken before the flare onset.
1033: The right image shows the EIT image in green-white false colors;
1034: the white patch at the center denotes the flare location. The
1035: superimposed contours show the magnetogram, with yellow the
1036: positive and blue the negative field.}
1037: \label{fg_pmdieit_example}
1038: \end{figure}
1039:
1040: \clearpage
1041:
1042: \begin{figure}[tb]
1043: \centering
1044: \includegraphics[width=\hsize]{f3.eps}
1045: \caption{Segmented MDI magnetograms for the four confined events, in which
1046: strong positive ($\geq 50$ Gauss) and negative ($\leq -50$ Gauss)
1047: magnetic fields are highlighted as white and black colors,
1048: respectively. Red asterisk symbols indicate the flare sites, the
1049: red diamond symbols indicate the COM of the active regions, and
1050: the blue lines denote the neutral lines over which the flares
1051: occurred. See text for more details.}
1052: \label{fg_segmdi_noncme}
1053: \end{figure}
1054:
1055: \clearpage
1056:
1057: \begin{figure}[tb]
1058: \centering
1059: \includegraphics[width=\hsize]{f4.eps}
1060: \caption{Segmented MDI magnetograms for the four eruptive events. See the caption in Figure~\ref{fg_segmdi_noncme}.}
1061: \label{fg_segmdi_cme}
1062: \end{figure}
1063:
1064: \clearpage
1065:
1066: \begin{figure}[tb]
1067: \centering
1068: \includegraphics[width=\hsize]{f5.eps}
1069: \caption{Calculated coronal magnetic field of one confined event.
1070: The closed field lines are denoted by the green \& yellow colors,
1071: corresponding to the outward and inward direction respectively,
1072: and the open field lines are denoted by the blue color. The left
1073: image is of a top view while the right image is of a side view.}
1074: \label{fg_mag_n_20040717}
1075: \end{figure}
1076:
1077: \clearpage
1078:
1079: \begin{figure}[tb]
1080: \centering
1081: \includegraphics[width=\hsize]{f6.eps}
1082: \caption{An example of eruptive events showing the extrapolated magnetic
1083: field above the active region.}
1084: \label{fg_mag_c_20041030}
1085: \end{figure}
1086:
1087: \clearpage
1088:
1089: \begin{figure}[tb]
1090: \centering
1091: \includegraphics[width=\hsize]{f7.ps}
1092: \caption{The scattering plot showing magnetic properties of
1093: both confined events (diamond symbols) and the eruptive events
1094: (asterisk symbols). The $x$-axis denotes the distance between the
1095: flare site and the center of magnetic flux distribution (COM) of
1096: the active region, and the $y$-axis denotes the ratio of magnetic
1097: flux in the low to high corona above the neutral line.}
1098: \label{fg_ratio_distance}
1099: \end{figure}
1100:
1101:
1102:
1103: \end{document}
1104: