1:
2: \documentclass[12pt,preprint]{aastex}
3:
4: %\usepackage{epsfig}
5:
6: %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
7: \shortauthors{Zhang and Flyer} \shorttitle{Dependence of Helicity
8: Bound on Boundary Condition}
9:
10: %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
11: \begin{document}
12:
13: \title{The dependence of the helicity bound of force-free magnetic fields on
14: boundary conditions}
15:
16: \author{Mei Zhang\altaffilmark{1,2} \& Natasha Flyer\altaffilmark{3}}
17:
18: \altaffiltext{1}{National Astronomical Observatory, Chinese Academy
19: of Sciences, 20A Datun Road, Chaoyang District, Beijing 100012,
20: China} \altaffiltext{2}{High Altitude Observatory, National Center
21: for Atmospheric Research, PO Box 3000, Boulder, CO 80307, USA}
22: \altaffiltext{3}{Institute for Mathematics Applied to the
23: Geosciences, National Center for Atmospheric Research, PO Box 3000,
24: Boulder, CO 80307, USA}
25:
26: %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
27: \begin{abstract}
28: This paper follows up on a previous study showing that in an open
29: atmosphere such as the solar corona the total magnetic helicity of a
30: force-free field must be bounded and the accumulation of magnetic
31: helicity in excess of its upper bound would initiate a
32: non-equilibrium situation resulting in an expulsion such as a
33: coronal mass ejection (CME). In the current paper, we investigate
34: the dependence of the helicity bound on the boundary condition for
35: several families of nonlinear force-free fields. Our calculation
36: shows that the magnitude of the helicity upper bound of force-free
37: fields is non-trivially dependent on the boundary condition. Fields
38: with a multipolar boundary condition can have a helicity upper bound
39: ten times smaller than those with a dipolar boundary condition when
40: helicity values are normalized by the square of their respective
41: surface poloidal fluxes. This suggests that a coronal magnetic field
42: may erupt into a CME when the applicable helicity bound falls below
43: the already accumulated helicity as the result of a slowly changing
44: boundary condition. Our calculation also shows that a monotonic
45: accumulation of magnetic helicity can lead to the formation of a
46: magnetic flux rope applicable to kink instability. This suggests
47: that CME initiations by exceeding helicity bound and by kink
48: instability can both be the consequences of helicity accumulation in
49: the corona. Our study gives insights into the observed associations
50: of CMEs with the magnetic features at their solar surface origins.
51: \end{abstract}
52:
53: \keywords{MHD --- Sun: magnetic fields --- Sun: corona
54: --- Sun: coronal mass ejections (CMEs) }
55:
56: %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
57:
58: \section{Introduction}
59:
60: Magnetic helicity is a physical quantity that measures the
61: topological complexity of a magnetic field, such as the degree of
62: linkage and/or twistedness in the field (Moffatt 1985, Berger \&
63: Field 1984). In a previous paper (Zhang et al. 2006, hereafter
64: referred to as ZFL) we proposed that in an open atmosphere such as
65: the solar corona there is an upper bound on the total magnetic
66: helicity that a force-free field can contain. The accumulation of
67: magnetic helicity in excess of this upper bound would initiate a
68: non-equilibrium situation, resulting in a coronal mass ejection
69: (CME) as a natural product of coronal evolution.
70:
71: Our approach (Zhang \& Low 2003, 2005, ZFL) shifts the traditional
72: focus on the mechanism for CME evolution from the storage of
73: magnetic energy to the accumulation of magnetic helicity, although
74: the two types of considerations are not necessarily exclusive with
75: each other. The advantage of using magnetic helicity as a more
76: fundamental physical quantity is that from observations we know that
77: the magnetic fields are emerging from the solar interior with a
78: preferred helicity sign in each solar hemisphere (Pevtsov et al.
79: 1995, Rust \& Kumar 1996, Bao \& Zhang 1998, Zhang 2006). As a
80: result, an accumulation of the total magnetic helicity in the corona
81: becomes unavoidable because the total magnetic helicity is
82: approximately conserved in the corona during coronal processes
83: including fast magnetic reconnection (Berger 1984).
84:
85: In this paper, we study how the magnitude of the upper bound of the
86: total magnetic helicity depends on the boundary condition. Section 2
87: presents the model with two new boundary conditions. Results and
88: analysis are presented in Section 3. In Section 4, a brief summary
89: of the paper is given.
90:
91:
92: %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%5
93:
94: \section{The Model}
95:
96: \subsection{The governing equation}
97:
98: Following Flyer et al. (2004) and ZFL, we use the families of
99: power-law axisymmetric force-free fields to understand the basic
100: physical properties of interest.
101:
102: With axisymmetry, the solenoidal magnetic field ${\bf B}$ in
103: $r > 1$ can be written in the form of
104: \begin{equation}
105: {\bf B} = {1 \over r \sin \theta} \left[
106: {1 \over r}{\partial A \over \partial \theta} ~,
107: ~ - {\partial A \over \partial r}~, ~Q (A) \right] ~,
108: \end{equation}
109: \noindent where the flux function $A$ defines the poloidal magnetic
110: field and the function $Q$ defines the toroidal (or azimuthal)
111: field.
112:
113: $Q$ is defined as a strict power-law in $A$ with the form
114: \begin{eqnarray}
115: Q^2 (A) = \frac{2 \gamma}{n+1} A^{n+1} ~,
116: \end{eqnarray}
117: \noindent where $n$ is an odd constant index required to be no less
118: than 5 in order for the field to possess finite magnetic energy in
119: $r > 1$ and $\gamma$ is a free parameter which we choose to be
120: positive without loss of generality. This form of $Q$ reduces the
121: force-free condition to the following governing equation for the flux
122: function $A$:
123: \begin{eqnarray}
124: {\partial^2 A \over \partial r^2} + {1 - \mu^2 \over r^2}
125: {\partial^2 A \over \partial \mu^2} + \gamma A^n = 0 ~~.\label{eqnA}
126: \end{eqnarray}
127:
128: This governing equation was solved numerically as a boundary value
129: problem within domain $r > 1$ in Flyer et al. (2004) and ZFL,
130: subject to the prescribed boundary flux distribution of
131: \begin{equation}
132: A|_{r = 1} = \sin^2 \theta ~~.
133: \label{dipolar}
134: \end{equation}
135: \noindent This boundary flux distribution and its associated normal
136: field distribution are plotted in the top panels of Figure 1. All
137: force-free fields discussed in Flyer et al. (2004) and ZFL share the
138: boundary flux distribution given by (\ref{dipolar}), which is the
139: same as that for a dipole potential field corresponding to the
140: solution of equation (\ref{eqnA}) with $\gamma=0$. The solutions of
141: (\ref{eqnA}) with boundary condition (\ref{dipolar}) shall be
142: referred to as dipolar force-free fields or dipolar fields for
143: short. We refer interested readers to Flyer et al. (2004) and ZFL
144: for various properties of dipolar force-free fields for the cases of
145: $n = 5, 7, 9$.
146:
147:
148: %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
149:
150: \subsection{Two new boundary conditions}
151:
152: In this paper, we solve the same governing equation (\ref{eqnA})
153: using the numerical methods described in Flyer et al. 2004, 2005
154: but subject to two new and distinctively different boundary
155: conditions. We use these numerical solutions to investigate how
156: properties we discussed in ZFL would change with the new boundary
157: conditions.
158:
159: The first new boundary condition has the flux distribution of
160: \begin{equation}
161: A|_{r = 1} = \sin^{12}\theta ~~.
162: \end{equation}
163: \noindent This flux distribution and its associated normal field
164: distribution are plotted in the middle panels of Figure 1. We see
165: that this new boundary condition has its flux concentrated nearer to
166: the equator than that of the dipolar field. This makes it more like
167: solar active regions with its normal field strength much higher at
168: equatorial regions than that near the poles. We shall refer this
169: family of power-law axisymmetric force-free fields as bipolar
170: force-free fields or bipolar fields hereafter.
171:
172: The second new family of solutions are those that we shall refer to
173: as multipolar force-free fields or multipolar fields. They are also
174: the solutions to (\ref{eqnA}) but subject to the following boundary
175: condition:
176: \begin{equation}
177: A|_{r = 1} = \sin^2\theta \times (1- 5 \times \cos^2\theta) ~~~.
178: \end{equation}
179: \noindent Its flux distribution and associated normal field
180: distribution are plotted in the bottom panels of Figure 1. We see
181: that fields with this boundary condition have both positive and
182: negative magnetic fields in one hemisphere, distinctively different
183: from those in dipolar force-free fields and bipolar force-free
184: fields where fields in one hemisphere have the same magnetic
185: polarity.
186:
187: Table 1 presents a comparison of the three boundary conditions. We
188: see that their common feature is that they all have $A|_{r=1}=0$ at
189: solar poles and $A|_{r=1}=1$ at the equator. The dipolar boundary
190: and bipolar boundary differ by the contrast of their respective
191: maximum $B_r|_{r=1}$ values to their $B_r|_{r=1}$ values at the
192: northern pole. The multipolar boundary condition differs from the
193: dipolar and bipolar boundary conditions by the existence of both
194: magnetic polarities in each hemisphere, indicated by its negative
195: minimum $A$ value.
196:
197: In Table 1 we also present the total surface poloidal flux ($F_p$)
198: for the three boundary conditions, where
199: \begin{equation}
200: F_p = \int_{r=1} B_r(>0) ds = 2\pi\int^{\frac{\pi}{2}}_0 |B_r|
201: \sin\theta d\theta ~~~.
202: \end{equation}
203: \noindent We see that dipolar and bipolar fields have the same total
204: surface poloidal flux ($2\pi$) and the multipolar fields have a
205: larger total surface poloidal flux ($5.2\pi$). In the latter
206: development of the paper, we will normalize calculated magnetic
207: helicity by $F_p^2$ in order to make fields with different boundary
208: conditions comparable.
209:
210:
211: %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
212:
213: \section{Results and Analysis}
214:
215: As in ZFL, for each new boundary condition, we numerically solve
216: (\ref{eqnA}) for three cases: $n=5$, $n=7$ and $n=9$. In each case, the
217: numerical method, that is, the Newton's iteration combined with a
218: pseudo-arc length continuation scheme, guarantees the completeness
219: of each solution branch generated by the $\gamma$ values.
220:
221: Also as in ZFL, for each obtained solution, we calculate three
222: physical quantities of the field. They are: total magnetic energy
223: \begin{equation}
224: E = \int_{r > 1} {B^2 \over 8 \pi} dV
225: = \frac{1}{4} \int_{r = 1}
226: \left( B_r^2 - B_{\theta}^2 - B_{\varphi}^2
227: \right) \sin \theta d\theta ~~,
228: \end{equation}
229: \noindent total azimuthal flux
230: \begin{equation}
231: F_{\varphi} = \int_{r > 1} |B_{\varphi}| r dr d\theta =
232: \sqrt{\frac{\gamma}{m+1}} \int_{r > 1} |A|^{m+1} {dr d\theta \over
233: \sin \theta} ~~,
234: \label{Fphi}
235: \end{equation}
236: \noindent and total relative magnetic helicity
237: \begin{equation}
238: H_R = 4 \pi \int_{r>1} A B_{\varphi} r dr d \theta = 4 \pi
239: \sqrt{\frac{\gamma}{m+1}} \int_{r > 1} A^{m+2} {dr d\theta \over
240: \sin \theta} ~.
241: \end{equation}
242: \noindent The derivation of these formula can be found in ZFL. The
243: only difference is that since we are also considering multipolar
244: fields the absolute value of $B_r$ (that is, $|B_r|$) is introduced
245: when calculating total azimuthal flux.
246:
247: In the geometric simplicity of these force-free fields, the
248: equilibrium in each case is due to the magnetic tension force of the
249: poloidal flux confining the magnetic pressure of the azimuthal flux.
250: The magnitude of the tension force is sensitively related to the
251: poloidal flux at the inner boundary ($r=1$) that serves as an anchor
252: agent. Moreover, the poloidal flux and its tension force become
253: weaker in the outward radial direction. Hence, if the azimuthal flux
254: becomes too large it can not be confined by such a mechanism as was
255: shown in ZFL for dipolar fields. The magnetic pressure is
256: independent of the sign of the field. Therefore, the total unsigned
257: azimuthal flux given by (\ref{Fphi}) is more relevant for the
258: consideration of flux confinement than the total signed azimuthal
259: flux.
260:
261: %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
262:
263: \subsection{Helicity upper bound of bipolar fields}
264:
265: Figure 2 presents the variations of the total magnetic energy ($E$),
266: total azimuthal flux ($F_{\varphi}$) and total magnetic helicity
267: ($H_R$) versus $\gamma$ along the solution curve for bipolar force-free
268: fields with $n=5$, $n=7$ and $n=9$. The figure is similar to the
269: Figure 2 of ZFL except that Figure 2 in ZFL is for dipolar
270: force-free fields. Each point along the solution curve, denoted by a
271: plus symbol in the figure, represents a solution to (\ref{eqnA}). By
272: solution curve we mean that all solutions along the curve are
273: obtained with the same boundary condition and the same constant
274: index $n$ but with a monotonically increasing magnitude of total
275: azimuthal flux.
276:
277: From these curves of solutions we see that there may be upper bounds
278: on the total magnetic energy, total azimuthal flux as well as total
279: magnetic helicity for bipolar fields, as that we have suggested in
280: ZFL for dipolar fields.
281:
282: From Figure 2 we can also see that the magnitude of the helicity
283: upper bound for bipolar fields is smaller than that for dipolar
284: fields. While the upper bound for dipolar fields is close to 14, the
285: upper bound for bipolar fields only approaches 9. To illustrate this
286: further, we plot in Figure 3 the variation of the total magnetic
287: helicity ($H_R$) versus azimuthal flux ($F_{\varphi}$) along the
288: solution curve for fields with the dipolar (top panels) and bipolar
289: (bottom panels) boundary conditions. Here we have normalized the
290: values of total magnetic helicity ($H_R$) of each field by the
291: square of their corresponding surface poloidal fluxes ($F_p^2$) as
292: those in Demoulin et al. (2002), van Driel-Gesztelyi et al. (2003)
293: and Demoulin (2007). We see that while the upper bound of
294: $H_R/F_p^2$ is about 0.35 for dipolar fields, for bipolar fields it
295: is significantly lower at 0.22.
296:
297: This result suggests a dependence of the helicity upper bound on the
298: boundary condition, which in our view gives an insight into observed
299: associations between CMEs and magnetic features at their solar
300: surface origins. Observationally we know that CMEs can be triggered
301: by flux emergence (e.g. Feynman \& Martin 1995, Subramanian \& Dere
302: 2001) and converging motions (e.g. Martin 1990). Different
303: theoretical models have also proposed that CME-type eruptions can be
304: triggered by various surface field variations (e.g. Chen \& Shibata
305: 2000, Amari et al. 2000, 2003a, 2003b).
306:
307: We suggest that the common physics underlying the different
308: mechanisms associated with such surface field variations is the
309: dependence of helicity upper bound on the boundary condition. When a
310: magnetic field has accumulated a certain amount of magnetic helicity
311: (but not yet enough for an eruption) then a change of the boundary
312: condition could lower the helicity upper bound, resulting in a
313: non-equilibrium situation and hence a CME eruption under the new
314: boundary condition. For example, if $H_R/F_p^2$ were 0.3 for a
315: dipolar boundary condition, then an evolutionary change to a bipolar
316: boundary condition would result in a CME eruption because the
317: applicable upper bound on the conserved total helicity has been
318: reduced as suggested by our numerical experiments.
319:
320: A note to address here is that although in this paper we have
321: emphasized the role of boundary condition variations this does not
322: mean that the role of magnetic helicity accumulation becomes less
323: important. A change of the boundary condition may bring in an
324: eruption only when the field has accumulated enough helicity for an
325: eruption under the new boundary condition. If not, the field does
326: not erupt even when the boundary flux distribution is changing. This
327: is consistent with the observation (Zhang et al. 2007) that although
328: flux emergences are indeed found to be associated with CME
329: eruptions, the same rate of flux emergence can also be found when
330: there is no CME or solar activities. This means that flux emergence
331: may be a trigger of a CME eruption, but flux emergence alone do not
332: guarantee an eruption.
333:
334: Another interesting result from our calculations is that these normalized
335: $H_R/F_p^2$ helicities, estimated from simple axisymmetric power-law
336: force-free fields, lie close to those estimated from observations.
337: Van Driel-Gesztelyi et al. (2003) and Demoulin (2007) pointed out
338: that the $H_R/F_p^2$ numbers, estimated from the extrapolated
339: magnetic fields based on observed photospheric magnetograms, are
340: between 0.02 to 0.2. Our helicity upper bound numbers of dipolar and
341: bipolar fields are just a little higher. Notice that the numbers
342: estimated from the observations may be somehow underestimated
343: because of the limited spatial resolution of the observed
344: magnetograms. So there seems to be consistency between the
345: theoretical $H_R/F_p^2$ helicities and those estimated from
346: observations.
347:
348: Figure 4 presents four field configurations selected from the $n=9$
349: solution curve, positions of which along the curve are illustrated
350: in Figure 2. We see that starting from the potential field (Panel A)
351: the curve first reaches the maximum-$\gamma$ field (Panel B) and
352: then the field with maximum total magnetic energy (Panel C) and
353: then the one with maximum total azimuthal flux (Panel D). A clear
354: bubble (representing a flux rope) is presented in the field of Panel
355: C but not in the field of Panel D where the latter actually
356: possesses more total azimuthal flux than the former. This tells us
357: that although the existence of a flux rope (or flux ropes) in the
358: low corona does represent a storage of a certain amount of magnetic
359: helicity (see more discussions in Zhang \& Low 2003) it is not
360: necessary that they are present in the field with a maximum helicity
361: storage.
362:
363:
364: %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%5
365:
366: \subsection{Helicity upper bound of multipolar fields}
367:
368: As in Figure 2, we present in Figure 5 the variations of the total
369: magnetic energy, total azimuthal flux and total magnetic helicity versus
370: $\gamma$ along the solution curve with $n=5$, $n=7$ and $n=9$ but
371: for the multipolar fields. We see that these curves also suggest the
372: existence of upper bounds on the total magnetic energy, total
373: azimuthal flux as well as total magnetic helicity as those for
374: dipolar fields and bipolar fields.
375:
376: The figure also shows that the magnitude of the helicity upper bound
377: for multipolar fields is smaller than that for dipolar fields. This
378: reduction is even more evident in Figure 6 where we plot $H_R/F_p^2$
379: versus $F_\varphi$. We see that the upper bound of $H_R/F_p^2$
380: for multipolar fields is below 0.04, almost ten times smaller than
381: that of dipolar fields.
382:
383: Such a severe reduction of $H_R/F_p^2$ upper bound not only further
384: confirms our previous result that the helicity upper bound is
385: dependent on the boundary condition, but also brings our theoretical
386: $H_R/F_p^2$ value even closer to those estimated from
387: observations (Regnier et al. 2005). Furthermore, the severe
388: reduction of the helicity upper bound in terms of $H_R/F_p^2$ values
389: may also explain why solar eruptions such as CMEs are more likely to
390: happen in complicated active regions where the multipolar field, by
391: the above property, will take less time to reach its helicity bound,
392: producing an eruption.
393:
394: As in Figure 4, Figure 7 presents four field configurations selected
395: from the $n=7$ solution curve of the multipolar fields. We see again
396: a clear bubble in the field of Panel C but not in Panel D. Therefore,
397: as with bipolar fields, multipolar fields with maximum helicity storage
398: need not contain a flux rope.
399:
400: %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%5
401:
402: \subsection{Kink instability}
403:
404: A rope of highly helical field is susceptible to an instability that
405: causes the rope to kink (Friedberg 1987). From elementary
406: calculations, this kink instability sets in if a critical twist is
407: exceeded ($T>T_c$). The exact value of $T_c$ depends on the detailed
408: field models, and could increase from the traditional
409: Kruskal-Shafranov limit $T_c=2\pi$ to $T_c=2.5\pi$ (Hood \& Priest
410: 1981) and $T_c=4.8\pi$ (Mikic et al. 1990).
411:
412: Since we have helical flux tubes (or flux ropes) present in our
413: solutions, it is interesting to investigate whether these flux ropes
414: have exceeded the kink instability. Figure 8 presents the variation
415: of the average twist ($T$) versus $\theta_0$ for two fields. One of the
416: fields is the bipolar $n=9$ maximum-energy field, the one presented
417: in Panel C of Figure 4. The other is the multipolar $n=7$
418: maximum-energy field, the one presented in Panel C of Figure 7. Here
419: $\theta_0$ is the angle from the equator. The average twist ($T$) is
420: obtained from $H_R^\prime/(F^\prime_p)^2$, where $H_R^\prime$ is the
421: total relative magnetic helicity in domain $\Omega^\prime$, enclosed
422: by the $r=1$ surface and the flux surface with $A =
423: A|_{r=1,\theta=\theta_0}$, and $F_p^\prime$ is the total surface
424: ($r=1$) polodial flux of the domain $\Omega^\prime$.
425:
426: We see that in both fields the average twist of the central part of
427: the field (that is, where the flux rope is located) has exceeded the
428: kink instability criteria, $T_c=2.5\pi$ of Hood \& Priest (1981) or
429: $T_c=4.8\pi$ of Mikic et al. (1990). This tells us that with the
430: accumulation of a certain amount of magnetic helicity, the flux rope
431: formed in the field can possess a twist number that is larger than
432: the kink instability criteria. If other necessary conditions are
433: favorable, for example, if the field has accumulated enough free
434: magnetic energy, an eruption may happen even before the field has
435: reached its helicity upper bound state. In that sense, reaching the
436: helicity upper bound state may not be a necessary condition for
437: eruption, but the helicity upper bound is a sufficient condition
438: upon which an eruption will become unavoidable. This also shows that
439: CME eruptions that are initiated by the kink instability (e.g. Torok
440: and Kliem 2005, Fan \& Gibson 2007) or by the existence of helicity
441: upper bound could both be viewed as the consequences of magnetic
442: helicity accumulation and they are not mutually exclusive.
443:
444: %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%5
445:
446: \section{Conclusion}
447:
448: In this paper, we continue our study on the hydromagnetic origin of
449: CMEs in terms of magnetic helicity accumulation. As in a previous
450: paper (ZFL), we numerically solve (\ref{eqnA}) to get families of
451: axisymmetric power-law force-free fields, but subject to two new
452: boundary conditions.
453:
454: By analyzing and comparing obtained solutions for three different
455: boundary conditions we conclude the following:
456:
457: \begin{enumerate}
458: \item The suggestion that there may be an upper bound on the total magnetic
459: helicity for force-free fields is also found for the two new
460: boundary conditions.
461:
462: \item The magnitude of the helicity upper bound of force-free fields
463: is non-trivially dependent on the boundary condition. In our
464: examples, the fields with a surface flux distribution more like a
465: simple active region (bipolar fields) have their helicity upper
466: bound smaller than that of fields with dipolar boundary condition.
467: For multipolar fields, the helicity upper bound ($H_R/F_p^2$) can be
468: ten times smaller than that of dipolar fields. These results provide
469: some insights into the observed association of CMEs with flux
470: emergence and surface field variation. These results also suggest a
471: physical reason why eruptions are more likely to happen in
472: complicated active regions.
473:
474: \item CME initiations by kink instability and by the existence of a
475: helicity upper bound can both be the result of magnetic helicity
476: accumulation in the corona. They do not exclude each other.
477:
478: \end{enumerate}
479: %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
480:
481: \acknowledgements
482:
483: We thank Pascal Demoulin, BC Low, Alexander Nindos and the anonymous
484: referee for helpful comments. This work was partly supported by the
485: One-Hundred-Talent Program of Chinese Academy of Sciences, Chinese
486: National Key Basic Research Science Foundation (G2006CB806300),
487: Chinese National Science Foundation Grant 40636031, and USA NSF
488: ATM-0548060 Programs. The National Center for Atmospheric Research
489: is sponsored by the National Science Foundation. Natasha Flyer would
490: like to acknowledge the support of NSF grant ATM-0620100.
491:
492:
493: %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
494:
495:
496: \begin{thebibliography}{}
497:
498: \bibitem[Amari et al. 2000]{amari00}
499: Amari, T., Luciani, J. F., Mikic, Z., et al. 2000, ApJ, 529, L49
500:
501: \bibitem[Amari et al. 2003a]{amari03a}
502: Amari, T., Luciani, J. F., Aly, J. J., et al. 2003a, ApJ, 585, 1073
503:
504: \bibitem[Amari et al. 2003b]{amari03b}
505: Amari, T., Luciani, J. F., Aly, J. J., et al. 2003b, ApJ, 595, 1231
506:
507: \bibitem[Bao and Zhang 1998]{bao98}
508: Bao, S., \& Zhang, H. Q. 1998, ApJ, 496, L43
509:
510: \bibitem[Berger 1984]{berger84a}
511: Berger, M. A. 1984, Geophys. Astrophys. Fluid Dyn., 30, 79
512:
513: \bibitem[Berger and Field 1984]{berger84b}
514: Berger, M. A., \& Field, G. B. 1984, J. Fluid Mech., 147, 133
515:
516: \bibitem[Chen and Shibata 2000]{chen00}
517: Chen, P. F., \& Shibata, K. 2000, ApJ, 545, 524
518:
519: \bibitem[Demoulin 2007]{demoulin07}
520: Demoulin, P. 2007, Adv. Space Res., 39, 1674
521:
522: \bibitem[Demoulin et al. 2002]{demoulin02}
523: Demoulin, P., Mandrini, C. H., van Driel-Gesztelyi, L., et al. 2002,
524: A\&A, 382, 650
525:
526: \bibitem[Fan and Gibson 2007] {fan07}
527: Fan, Y., \& Gibson, S. E. 2007, ApJ, 668, 1232
528:
529: \bibitem[Feynman and Martin 1995]{feynman95}
530: Feynman, J., \& Martin, S. F. 1995, J. Geophys. Res., 100, 3355
531:
532: \bibitem[Flyer et al. 2004]{flyer04}
533: Flyer, N., Fornberg, B., Thomas, S., \& Low, B. C. 2004, ApJ, 606,
534: 1210
535:
536: \bibitem[Flyer et al. 2005]{flyer05}
537: Flyer, N., Fornberg, B., Thomas, S., \& Low, B. C. 2005, ApJ, 631,
538: 1239
539:
540: \bibitem[Friedberg 1987]{friedberg87}
541: Friedberg, J. P. 1987, Ideal Magnetohydrodynamics, Plenum : New York
542:
543: \bibitem[Hood and Priest 1981]{hood81} Hood, A. W., \&
544: Priest, E. R. 1981, Geophys. Ap. Fluid Dyn., 17, 297
545:
546: \bibitem[Martin 1990]{martin90}
547: Martin, S. F. 1990, in Dynamics of Quiescent Prominences, ed. V.
548: Ruzdjak \& E. Tandberg-Hanssen (Berlin: Springer), 1
549:
550: \bibitem[Mikic et al. 1990]{mikic90}
551: Mikic, Z., Schnack, D. D., \& van Hoven, G. 1990, ApJ, 361, 690
552:
553: \bibitem[Moffatt 1985]{moffatt85}
554: Moffatt, H. K. 1985, J. Fluid Mech., 159, 359
555:
556: \bibitem[Pevtsov et al. 1995] {pevtsov95}
557: Pevtsov, A. A., Canfield, R. C., \& Metcalf, T. R. 1995, ApJ, 440,
558: L109
559:
560: \bibitem[Regnier et al. 2005] {regnier05}
561: Regnier, S., Amari, T., \& Canfield, R. C. 2005, A\&A, 442, 345
562:
563: \bibitem[Rust \& Kumar 1996] {rust96}
564: Rust, D. M., \& Kumar, A. 1996, ApJ, 464, L199
565:
566: \bibitem[Subramanian and Dere 2001] {subramanian01}
567: Subramanian, P., \& Dere, K. P. 2001, ApJ, 561, 372
568:
569: \bibitem[Torok and Kliem 2005] {torok05}
570: Torok, T., \& Kliem, B. 2005, ApJ, 630, L97
571:
572: \bibitem[van Driel-Gesztelyi et al. 2003]{Gesztelyi03}
573: van Driel-Gesztelyi, L., Demoulin, P., \& Mandrini, C. H. 2003, Adv.
574: Space Res., 32, 1855
575:
576: \bibitem[Zhang 2006] {zhang06}
577: Zhang, M. 2006, ApJ, 646, L85
578:
579: \bibitem[Zhang and Low 2003] {zhang03}
580: Zhang, M., \& Low, B. C. 2003, ApJ, 584, 479
581:
582: \bibitem[Zhang and Low 2005] {zhang05_01}
583: Zhang, M., \& Low, B. C. 2005, ARAA, 43, 103
584:
585: \bibitem[Zhang and Low 2005] {zhang05_02}
586: Zhang, M., Flyer, N., \& Low, B. C. 2006, ApJ, 644, 575 (ZFL)
587:
588: \bibitem[Zhang et al. 2007] {zhang07}
589: Zhang, Y., Zhang, M., \& Zhang, H. Q. 2007, Adv. Space Res., 39,
590: 1762
591:
592: \end{thebibliography}
593:
594: %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
595: \clearpage
596: \begin{table}
597: \caption{Comparison of the three boundary conditions}
598:
599: \begin{tabular}{|c|c|c|c|c|c|c|c|c|c|}
600: \hline
601: & \multicolumn{4}{|c|}{$A$ (at $r=1$)} & \multicolumn{4}{|c|}{$B_r$ (at $r=1$)} & $F_p$ \\
602: \hline
603: & north pole & equator & max. & min. & north pole & equator & max. & min. & \\
604: \hline
605: dipolar & 0 & 1 & 1 & 0 & 2 & 0 & 2 & -2 & $2\pi$ \\
606: \hline
607: bipolar & 0 & 1 & 1 & 0 & 0 & 0 & 2.25 & -2.25 & $2\pi$ \\
608: \hline
609: multipolar & 0 & 1 & 1 & -0.8 & -8 & 0 & 8 & -8 & $5.2\pi$ \\
610: \hline
611: \end{tabular}
612: \end{table}
613:
614: \clearpage
615: %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
616:
617: \begin{figure}
618: \centerline{\includegraphics[width=160mm]{f1.eps}} \caption{Flux
619: distribution (left panels) and normal field distribution (right
620: panels) of dipolar fields (top panels), bipolar fields (middle
621: panels) and multipolar fields (bottom panels) at $r=1$.}
622: \end{figure}
623:
624:
625: \begin{figure}
626: \centerline{\includegraphics[width=130mm]{f2.eps}} \caption{Variation of
627: the total magnetic energy ($E$), azimuthal flux ($F_{\varphi}$) and
628: total magnetic helicity ($H_R$) vs $\gamma$ along the solution curve
629: for $n=5$ (top panels), $n=7$ (middle panels) and $n=9$ (bottom
630: panels) fields with the bipolar boundary condition. Each point,
631: denoted by a plus symbol in the plot, represents a solution to
632: (\ref{eqnA}). Letters (A, B, C and D) in the bottom panels indicate
633: the positions along the solution curve of the four fields plotted in
634: Figure 4.}
635: \end{figure}
636:
637:
638: \begin{figure}
639: \centerline{\includegraphics[width=160mm]{f3.eps}}
640: \caption{Variation of the total magnetic helicity ($H_R$) vs
641: azimuthal flux ($F_{\varphi}$) along the solution curve for fields
642: with the dipolar (top panels) and bipolar (bottom panels) boundary
643: conditions. Here the total magnetic helicity ($H_R$) of each field
644: has been normalized by the square of their corresponding poloidal
645: flux ($F_p^2$).}
646: \end{figure}
647:
648:
649: \begin{figure}
650: \centerline{\includegraphics[width=180mm]{f4.eps}} \caption{Field
651: configurations of four fields selected from the $n=9$ bipolar
652: fields. These contours of flux function $A$ represent the lines of
653: force of the axisymmetric field projected on the $r-\theta$ plane.
654: Panel A: the potential field; Panel B: the field with the maximum
655: $\gamma$ value among all $n=9$ solutions; Panel C: the field
656: possessing the maximum total magnetic energy among all $n=9$
657: solutions. The thick line in this plot outlines the range within
658: which the average twist is greater $4\pi$ (see text in Section 3.3
659: for details); Panel D: the field possessing the maximum azimuthal
660: flux among all $n=9$ solutions.}
661: \end{figure}
662:
663:
664: \begin{figure}
665: \centerline{\includegraphics[width=150mm]{f5.eps}} \caption{Same as
666: Figure 2, but for fields with the multipolar boundary condition.}
667: \end{figure}
668:
669:
670: \begin{figure}
671: \centerline{\includegraphics[width=160mm]{f6.eps}} \caption{Variation of
672: the total magnetic helicity ($H_R$), normalized by the square of
673: poloidal flux ($F_p^2$), vs azimuthal flux ($F_{\varphi}$) for
674: fields with the multipolar boundary condition.}
675: \end{figure}
676:
677:
678: \begin{figure}
679: \centerline{\includegraphics[width=180mm]{f7.eps}} \caption{Same as
680: Figure 4, but for the $n=7$ multipolar case.}
681: \end{figure}
682:
683:
684: \begin{figure}
685: \centerline{\includegraphics[width=120mm]{f8.eps}}
686: \caption{Variation of average twist ($T$) vs $\theta_0$ (angle from
687: the equator) for the bipolar $n=9$ maximum-energy field (top panel)
688: and multipolar $n=7$ maximum-energy field (bottom panel). See text
689: for the calculation of $T$. The dashed lines indicate the critical
690: twist of kink instability where $T_c=2.5\pi$ (Hood \& Priest 1981)
691: or $T_c=4.8\pi$ (Mikic et al. 1990). See also Panel C in Figures 4
692: and 7 for field configurations.}
693: \end{figure}
694:
695:
696: %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
697:
698: \end{document}
699: