1: %\documentclass[12pt,preprint]{aastex}
2: \documentclass[preprint]{aastex}
3: %\documentclass[manuscript]{aastex}
4: %\usepackage{psfig}
5: %\usepackage{lineno}
6:
7: \def\references{\bibliographystyle{/net/star/Users/schryver/book/references/swanbib}\bibliography{/net/star/Users/schryver/book/references/ref_karel}}
8:
9:
10: \def\citet#1{\citeauthor{#1} (\citeyear{#1})}
11:
12: \def\citep#1{\citeauthor{#1}, \citeyear{#1}}
13:
14: \hyphenation{Schrij-ver}
15:
16: %\def\referee#1{{\sc #1}}
17: \def\referee#1{{#1}}
18:
19: %%%%%%%%%%%%%%%%% END OF PREAMBLE %%%%%%%%%%%%%%%%
20:
21: \begin{document}
22: %\linenumbers
23:
24: \title{Observations and modeling of
25: the early acceleration phase of erupting filaments involved in
26: coronal mass ejections}
27:
28: \author{Carolus J. Schrijver$^1$, Christopher Elmore$^1$,
29: Bernhard Kliem$^{2,3}$, Tibor T{\"o}r{\"o}k$^{4,5}$, and Alan M. Title$^1$}
30: \affil{$^1$ Lockheed Martin Advanced Technology Center,
31: 3251 Hanover Street, Palo Alto, CA 94304, U.S.A; \\
32: $^2$ Astrophysical Institute Potsdam, An der Sternwarte 16, 14482 Potsdam,
33: Germany;\\
34: $^3$ Kiepenheuer Institute for Solar Physics, Sch{\"o}neckstr. 6,
35: 79104 Freiburg, Germany;\\
36: $^4$ University College London,
37: Mullard Space Science Laboratory, Holmbury St. Mary, Dorking,
38: Surrey RH5 6NT, United Kingdom;\\
39: $^5$ LESIA, Observatoire de Paris, CNRS. Univerit{\'e} Paris Diderot,
40: 5 Place Jules Janssen, 92190 Meudon, France}
41: \email{schryver@lmsal.com; celmore@lmsal.com;
42: bkliem@aip.de; tt@mssl.ucl.ac.uk; title@lmsal.com}
43:
44: \date{\today}
45:
46: \begin{abstract}
47: We examine the early phases of two near-limb filament destabilizations
48: involved in coronal mass ejections on 16 June and 27 July 2005, using
49: high-resolution, high-cadence observations made with the Transition
50: Region and Coronal Explorer (TRACE), complemented by coronagraphic
51: observations by Mauna Loa and the SOlar and Heliospheric Observatory
52: (SOHO). The filaments' heights above the solar limb in their
53: rapid-acceleration phases are best characterized by a height
54: dependence $h(t)\propto t^m$ with $m$ near, or slightly above, 3 for
55: both events. Such profiles are incompatible with published results for
56: breakout, MHD-instability, and catastrophe models. We show numerical
57: simulations of the torus instability that approximate this height
58: evolution in case a substantial initial velocity perturbation is
59: applied to the developing instability. We argue that the sensitivity
60: of magnetic instabilities to initial and boundary conditions requires
61: higher fidelity modeling of all proposed mechanisms if observations of
62: rise profiles are to be used to differentiate between them. The
63: observations show no significant delays between the motions of the
64: filament and of overlying loops: the filaments seem to move as part of
65: the overall coronal field until several minutes after the onset of the
66: rapid-acceleration phase.
67: \end{abstract}
68:
69: \keywords{Sun: coronal mass ejections (CMEs) -- Sun: filaments}
70: %\maketitle
71:
72: \section{Introduction}
73: Observations of the early rise phase of filaments and their overlying fields
74: can in principle help constrain the mechanisms involved in
75: the destabilization of the magnetic configuration through comparison
76: with numerical simulations (e.g.,
77: \citep{fan2005}; \citep{torok+kliem2005}; \citep{williams+etal2005};
78: and references therein), because the detailed evolution depends
79: sensitively on the model details. For example, a power-law rise with
80: an exponent $m=2.5$ was obtained for a slender flux tube in the
81: two-dimensional version of the catastrophe model
82: (\citep{Priest&Forbes02}). An MHD instability triggered by an
83: infinitesimal perturbation implies an exponential rise, as was
84: verified, for example, for a three-dimensional flux rope subject to a
85: helical kink instability (\citep{Torok&al04},
86: \citep{torok+kliem2005}). The same holds for the torus (expansion)
87: instability (TI), which starts as a $\sinh(t)$ function
88: (\citep{Kliem&Torok06}) that is very similar to a
89: pure exponential early on. The CME rise in a breakout model
90: simulation was well described by a parabolic profile
91: (\citep{Lynch&al04}).
92:
93: The early rise phase of erupting filaments is best observed near the
94: solar limb using high-resolution data, both in space and in time. Such
95: data can be obtained by, for example, Big Bear Solar Observatory
96: H$\alpha$ observations (e.g., \citep{kahler+etal1988}), the Mauna Loa
97: K-coronameter (e.g., \citep{gilbert+etal2000}), the Nobeyama
98: Radioheliograph (e.g., \citep{gopalswamy+etal2003};
99: \citep{kundu+etal2004}), and the Transition Region
100: and Coronal Explorer, TRACE (e.g., \citep{vrsnak2001};
101: \citep{gallagher+etal2003}; \citep{goff+etal2005};
102: \citep{sterling+moore2004}; \citep{sterling+moore2005};
103: \citep{williams+etal2005}).
104: In those few cases where observers had the field of view for an
105: appropriate diagnostic to attempt to establish whether the high loops
106: or the filaments were accelerated first, the temporal resolution often
107: was not adequate (see, e.g., \citep{sterling+moore2004}, who use the
108: standard 12-min.\ cadence of SOHO/EIT).
109:
110: These studies show that filaments that are about to erupt often --~but
111: not always~-- exhibit a slow initial rise during which both the
112: filament and the overlying field expand with velocities in the range
113: of $1-15$\,km/s. Then follows a rapid-acceleration phase during which
114: velocities increase to a range of $100$\,km/s up to over
115: $1000$\,km/s. The rapid-acceleration phase finally transitions into a
116: phase with a nearly constant velocity or even a deceleration into the
117: heliosphere.
118:
119: The height evolution immediately following the onset of the rapid
120: acceleration phase is often approximated by either an exponential
121: curve (e.g., \citep{gallagher+etal2003}; \citep{goff+etal2005};
122: \citep{williams+etal2005} --~who also show systematic deviations from
123: that fit up to 2$\sigma$ in position~-- ) or by a
124: constant-acceleration curve (e.g., \citep{kundu+etal2004}; and
125: \citep{gilbert+etal2000} --~ who show one case in which
126: a third-order curve improves the fit to the earliest evolution, and
127: leave others for future analysis); \citet{kahler+etal1988} fit curves
128: for the acceleration $a=ct^b$ to the first $10-50$\,Mm for four
129: erupting filaments, but do not list the best-fit values.
130: \cite{alexander&al02} find a best fit for the height of the early
131: phase of a CME observed in X-rays by YOHKOH's SXT of the form
132: $h_0+v_0t+ct^{3.7\pm0.3}$. For 184 prominence events observed by the
133: Nobeyama Radioheliograph, \citet{gopalswamy+etal2003} show that higher
134: in the corona velocity profiles include decelerating, constant
135: velocity, and accelerating ones for heights from $\sim 50$\,Mm to
136: 700\,Mm above the solar surface.
137:
138: In many cases, the detailed study of the evolution of the early
139: phase is hampered by insufficient temporal
140: coverage or by gaps between the fields of view of two complementing
141: instruments that can be as large as a few hundred Mm. This results in
142: substantial uncertainties in the height evolution. \citet{vrsnak2001},
143: for example, concludes that ``[t]he main acceleration phase \ldots\ is
144: most often characterized by an exponential-like increase of the
145: velocity'', but notes that polynomial or power-law functions fit at
146: comparable confidence levels.
147:
148: In this study, we examine two events displaying the early
149: destabilization and acceleration of ring filaments leading to coronal
150: mass ejections. The high cadence down to 20\,s, and the high spatial
151: resolution of 1\,arcsec, for the early evolution result in relatively
152: small uncertainties in the height profiles. This enables a sensitive
153: test of the height evolution against exponential, parabolic, and
154: power-law fits. We find that a power-law with exponent near $3$, or
155: slightly higher, is
156: statistically preferred in both cases. As no published model matches
157: that profile, we experiment with a numerical model for the
158: torus instability, and find that this model can indeed approximate the
159: observations provided that a sufficiently large initial velocity
160: perturbation is applied (without which an exponential-like
161: profile would be found). This finding reminds us of the sensitivity of
162: developing instabilities to both initial and boundary conditions,
163: and shows that the models, particularly their parametric
164: dependencies, need to be worked out in greater detail in order to
165: use observations of the height-time observations
166: to differentiate successfully between competing models.
167:
168: \section{Observations}\label{sec:Observations}
169: Primary data for this study were collected by the Transition Region
170: and Coronal Explorer (TRACE; see \citep{traceinstrument}), and
171: ancillary data by the Mauna Loa Solar Observatory Mark IV
172: K-Coronameter (MLSO MK4) and the SOHO/LASCO C2 and C3 instruments
173: (\citet{brueckner+etal1995}). The events we studied occurred on 16
174: June 2005 19:10\,UT to 20:24\,UT (emanating from NOAA Active Region
175: 10775), and on 27 July 2005 from 03:00\,UT to 06:20\,UT (from
176: AR\,10792).
177:
178: \subsection{16 June 2005}
179: This eruption in AR\,10775 was associated with an M4.0 X-ray flare.
180: TRACE data are examined
181: from 19:10:42 through 20:08:37\,UT; MLSO MK4 data were available
182: from 20:06:59 to 20:23:25\,UT to characterize the later positions of the
183: filament. SOHO's LASCO did not observe at this time.
184: A characteristic TRACE image is shown in Fig.~\ref{fig:1a}, with
185: a sampling of outlines for filament ridge, loops, and position tracks.
186:
187: Initial data were taken at 19:10:42, followed by a few frames
188: beginning at 19:25:32\,UT. There is a gap in the TRACE data from
189: 19:29:34 to 19:47:35\,UT as the spacecraft traversed a zone of
190: enhanced radiation in its orbit. Starting at 19:47:35 each available
191: image was used for tracking, with a characteristic cadence of
192: approximately 40\,s, changing with exposure time and depending on data
193: gaps associated with orbital zones of enhanced background radiation.
194:
195: As no distinct features could be tracked in the filaments or in the
196: overlying loop structures, we use outlines of the top segments of the
197: filament and of some outstanding overlying loops as indicated in
198: Fig.~\ref{fig:1a}. We assign confidence intervals to these positions
199: by estimating the range of pixels that provides a reasonable
200: approximation of a feature.
201:
202: The rising filament loses a traceable form mid-way through the
203: acceleration. Once this occurs, short bright 'streaks' of plasma
204: parcels show up that are blurred by their motion during the exposures.
205: The positions of the midpoints of these streaks were used to extend
206: the position data for the filament rise. The length of the objects
207: was estimated by correcting for motion blur estimated from their
208: displacement from one exposure to the next, and then their average
209: positions were obtained, complemented by an uncertainty estimate.
210:
211: The MLSO MK4 data do not provide the same clarity of features to track
212: as do the TRACE data, and their observations are at a lower spatial
213: and temporal resolution. Thus, only an estimate of the filament
214: position was tracked, and was chosen as the point furthest from the
215: limb on the innermost feature on each of the images.
216:
217: The displacement of the approximate outlines was tracked by fitting
218: parabolas to sets of three adjacent points on each outline. For each
219: exposure, a vector was computed normal to the approximating parabola
220: from the central point at time $t_i$ to where it intersects a
221: subsequent parabolic fit for time $t_{i+1}$. That intersection point
222: is then used as the central position for the next step in the tracking
223: algorithm, thus moving from beginning to end in the image sequence.
224: The track of the filament ridge and of two overlying loops thus
225: measured are identified in Fig.~\ref{fig:1a}. The streaks observed in
226: the later phases were tracked as described above; their positions are
227: also shown in Fig.~\ref{fig:1a}.
228:
229: The filament evolves through three stages (Fig.~\ref{fig:1b}): 1) an
230: initial slow rise phase at a near-constant velocity, followed by 2) a
231: rapid-acceleration phase, and finally 3) a constant-velocity phase
232: high in the corona beginning at about 1$R_\odot$ above the surface.
233:
234: TRACE data for phase 1 up to 19:54:58\,UT show the features to exhibit
235: an approximately
236: constant velocity relative to the solar EUV limb. During this phase,
237: the filament moves 11,500\,km at an average of 4.4\,km/sec. We note
238: that the contribution of the solar rotation to this is negligible: for
239: a filament at geometric height $H(t)$ above the photosphere, the
240: apparent velocity $\dot{h}$ relative to the solar limb induced by the
241: perspective change as the Sun rotates is approximated by $\dot{h}
242: \approx \dot{H} + R_\odot \,\alpha\,\dot{\alpha},$ for a small angle
243: $\alpha$ between limb direction and current longitude. For $\alpha
244: \sim 5^\circ$, the apparent motion due to rotation only would be no
245: more than $0.2$\,km/s, much less than the observed velocity.
246:
247: The beginning of the acceleration phase was determined by a
248: combination of visual inspection of the raw images, inspection of
249: displacement charts, and minimization of $\chi^2$ values for the fits.
250: These three methods agreed in each case to within a tens of seconds.
251: The position data were fit with three different functional dependences
252: of time: a parabolic fit $a+bt+ct^2$, a power law allowing for an
253: initial rise velocity $a+bt+ct^m$, and an exponential $a+c\exp(dt)$.
254:
255: The rapid acceleration phase begins at 19:54:58, at which time we note
256: the initial appearance of a brightening feature across the lower end
257: of the central barb of the filament. This time is at the beginning of
258: a data gap from 19:54:58 to 19:57:36. The rapid acceleration phase
259: continues at least until the remnants of the filament leave the TRACE
260: field of view at 20:09:15.
261:
262: We find that the rise of the left-hand segment of the filament is best
263: fit by a power law. The power-law fit is superior to the exponential
264: fit in the range $2.7 \le m \le 3.9$.
265: Fits with $\chi_\nu^2\le 1.31$, i.e., up to the 99\%\ confidence level,
266: are found for $2.9
267: \lesssim m \lesssim 3.6$, with a best-fit value of $m=3.25$.
268: Setting $m=3$, we obtain $a=21.3\pm0.7$~Mm,
269: $b=1.7\pm2.8$~km\,s$^{-1}$, and $c=0.25\pm0.01$~m\,s$^{-3}$, with
270: $\chi_\nu^2=1.05$; if $b\equiv 0$ and $m\equiv 3$, then
271: $\chi_\nu^2=1.07$, only marginally worse than the best fit. The best
272: fit yields a constant jerk of $6c=1.47\pm0.03$~m\,s$^{-3}$. At the
273: edge of the MLSO field of view, the velocity approaches a terminal
274: value of $\sim 750$~km\,s$^{-1}$.
275:
276: The above near-cubic fit characterizes the data better than the
277: quadratic or exponential fits ($\chi_\nu^2$ of 4.7 and 2.2,
278: respectively), and agrees better with the MLSO data for position and
279: velocity needed farther from the limb.
280:
281: The initial phase of the destabilization behaves as if the loops and
282: filament are parts of a rapidly-expanding volume with no discernible
283: delays between the motions: the separations between the filament ridge
284: and two loops traced above it (lower dashed and upper solid curves in
285: Fig.~\ref{fig:1a}) appear to be essentially constant until the field
286: is disrupted in the mass ejection (see Fig.~\ref{fig:1c}): filament
287: and high loops destabilize and begin moving at the same time, and the
288: distance between them stays close to constant. For both the higher
289: and slightly lower loops discernible in the upper field, their
290: distance from the filament is almost unchanged until 20:01:41 for the
291: outer loop and 19:59:39 for the inner, lower loop. At this time, the
292: aggregate distance increases, as the loops begin to move laterally to
293: the primary motion of the expanding filament quickly. This indicates
294: overall that the high field is not evolving substantially to allow
295: the filament through, as might be expected in, e.g., the breakout process.
296:
297: \subsection{27 July 2005}
298: TRACE data for the eruption associated with the 2005/07/27 M3.7 flare
299: (Fig.~\ref{fig:2a}) were analyzed for 03:00:18 to
300: 04:43:38\,UT. LASCO C2/C3 data of the leading edge of the associated CME
301: were available from 04:56:37 to
302: 06:18:05\,UT to characterize the later phase.
303: This eruption also exhibits three
304: stages (Figs.~\ref{fig:2b} and~\ref{fig:2c}):
305: an initial constant velocity stage, a second rapid
306: acceleration phase, and a final coasting phase at near-constant
307: velocity.
308:
309: The initial slow rise lasts until 04:30:13\,UT. This rise
310: is already underway when TRACE data start at 03:00:18\,UT.
311: The early data establish that the
312: filament and the high field form one slowly expanding system.
313: The early rise velocity is calculated to be 13.4\,km/sec --~considerably
314: faster than for the event of 16 June~-- with the filament moving by 17,500\,km
315: prior to 04:30:13 when the rapid acceleration phase begins.
316:
317: The rapid acceleration phase lasts from 04:30:13 to at least
318: 04:43:38\,UT when the filament leaves the TRACE field of view. The position
319: data shown in Fig.~\ref{fig:2b}
320: through 04:38:53 is for the filament's top ridge,
321: and from 04:39:20 to 04:43:21 is
322: for bright streaks similar to those seen in the 16 June 2005
323: event.
324:
325: The data show this filament is also accelerating with a nearly
326: constant jerk. The function $h=a+bt+ct^m$ fits the data {very} well
327: for $2.9\lesssim m \lesssim 3.7$, with $\chi_\nu^2=0.63$ for $m\equiv
328: 3,\, b\equiv 0$. With $m=3$, the fit yields $a=45.0\pm0.6$~Mm,
329: $b=4.3\pm3.3$~km\,s$^{-1}$, and $c=0.31 \pm 0.01$\,m\,s$^{-3}$, which
330: corresponds to a constant jerk of $1.9\pm0.06$~m\,s$^{-3}$. The
331: essentially cubic fit is also the only one of our fits that reaches the
332: appropriate height and velocity to follow the leading edge
333: of the ejection as observed with LASCO C2/C3. The
334: exponential fit does not fit the acceleration phase as well
335: ($\chi^2_\nu=2.2$) and makes for a much poorer transition to the LASCO
336: data past 4:55\,UT. The velocity for the quadratic fit provides an
337: even poorer fit to the acceleration phase observed by TRACE
338: ($\chi^2_\nu=9.1$), and it appears far too slow to match the high
339: transition to constant velocity.
340:
341: Two arcs in the overlying field were tracked for this event (outlined
342: in Fig.~\ref{fig:2a}). The left of the upper field loops is tracked through
343: 04:41:21, and the right loop of the upper field through 04:42:34.
344: The separation from the rising filament, as
345: shown in Fig.~\ref{fig:2c}, shows that for both features there is
346: little to no difference in distance between the filament and loops for
347: a majority of the early rise and acceleration phases. At 04:33:34\,UT
348: the filament begins to approach the right of the upper field loops in
349: the primary direction of the acceleration, but keeps a constant
350: distance in the direction normal to the acceleration. The track on
351: the right loop, however, was not at the very top, so this is partially
352: an effect of the filament moving up as the loop moved off to the
353: side. Once the loop was to the side, the filament could pass by, and
354: the distance perpendicular to the direction of motion of the filament
355: remains unchanged. At the same time, the distance in the direction of
356: travel of the filament is decreasing between the filament and the left
357: high loop, but the distance normal to the primary acceleration is
358: increasing quickly. This increase is most likely due to the track on
359: the left loop being closer to directly above the filament, so it had
360: farther to move up and to the side. So, as the filament was moving up,
361: the left side of the high field significantly displaced in the
362: direction of the filament's acceleration, as well as normal to the
363: primary direction of acceleration.
364:
365: The third stage, seen in the LASCO C2/C3 data, indicates a constant
366: velocity of $\sim 1,250$\,km/sec. Although the leading edge always
367: propagates much faster than the filament in a CME core, the difference
368: to the velocity at the last of the \textsl{TRACE} data is substantial,
369: and implies that the acceleration continues at least part
370: of the way out to the first C2 data at 1.41$R_\odot$.
371:
372: \section{Comparison with models}
373: \label{sec:Comparison}
374: The two filament eruptions analyzed here are best fit by a power-law
375: height evolution with a power-law index $m$ near 3 or perhaps slightly
376: higher ($\chi_\nu^2$ values reach unity for values of $m$ of 3.3 and
377: 3.6, respectively; note that these values match the value of
378: $3.7\pm0.3$ found in the study by \citep{alexander&al02}). The nearly
379: constant rate of increase for the acceleration by 1.4--1.9~m\,s$^{-3}$
380: persists for about $10-15$~min in both events. These phases
381: were shown to be statistically inconsistent with either a constant
382: acceleration or an exponential growth.
383:
384: The jerk values, $d^3h/dt^3$, of 1.4 and 1.9\,m/s$^3$
385: for the two filament eruptions studied here are
386: very similar. Estimated values using $6\Delta h/ (\Delta t)^3$ based
387: on erupting filaments up to $\sim 200$\,Mm in the studies referenced
388: in \S~1 and here range from $\sim 0.3$\,m/s$^3$ (for an M6.5 event
389: described by \citep{hori+etal2005}, and a C4 event
390: observed by \citep{maricic+etal2004} and modeled by \citep{torok+kliem2005})
391: up to $\sim 50$\,m/s$^3$ (in an X2.5 event described by
392: \citep{williams+etal2005}). There is no clear correlation between
393: flare magnitude and jerk value for the small sample of events,
394: other than that the largest outlying flare shows the largest outlying
395: value of jerk (we note that there is also no clear dependence of
396: eventual CME speed and flare magnitude --~see \citet{zhang+golub2003}~--
397: although the class of fast CMEs has a 3 times higher maximum X-ray brightness
398: than the class of slow CMEs). It thus remains unknown what determines
399: the value of $d^3h/dt^3$, but the similarity of the values for the two
400: cases studied here may be fortuitous.
401:
402: Our observations of two erupting filaments do not match the results of
403: catastrophe, MHD instability, or breakout models published thus
404: far. The catastrophe model comes closest with a power-law rise with an
405: index of 2.5, which is near, but significantly below, the range
406: readily allowed by the observations. The simplifying assumptions of a
407: two-dimensional slender flux rope with unrestricted reconnection below
408: it may, of course, have modified the height evolution for the
409: model. Here we explore another effect, namely that of different
410: initial conditions, specifically for the torus instability (TI). The
411: TI results if the outward pointing hoop force of a current ring
412: decreases more slowly with increasing
413: ring radius than the opposing Lorentz force
414: due to an external magnetic field (Bateman, 1978): we investigate
415: whether the instability can describe the rapid-acceleration phase of
416: the two events and its transition to a nearly constant terminal
417: velocity.
418:
419: The geometry of the two events appears compatible with a torus
420: instability: the eruption on 16 June 2005 exhibits an expanding main
421: loop that approaches a toroidal shape within the range observed by
422: \textsl{TRACE}, and the eruption on 27 July 2005 is consistent with
423: such a shape seen side-on. Neither shows indications of helical
424: kinking. The $\sinh(t)$ profile obtained analytically for the TI by
425: \cite{Kliem&Torok06} relied on the simplifying assumption
426: that the external poloidal field varies with the major torus radius
427: $R$ as $B_\mathrm{ex}\propto R^{-n}$ with a constant decay index $n$,
428: and it is exact only as long as the displacement from the equilibrium
429: position remains (infinitesimally) small.
430:
431: Allowing for a height dependence of the decay index $n$
432: likely will
433: cause the height evolution in the model to differ even more from the
434: observations: because $n(h)$ is in reality an increasing function on
435: the Sun \cite[see, e.g., Fig.~2 in][]{vanTend&Kuperus78}, the
436: acceleration profile would likely increase more steeply than the
437: initially nearly exponential $\sinh(t)$ function.
438:
439: We have performed numerical MHD simulations of the TI to study the
440: evolution for finite displacements. For some parameter settings, the
441: exponential expansion was found to hold up to several initial radii of
442: the current ring, while for others a power-law-like expansion with
443: exponents scattering around $m\sim3$ could be found. The latter turns
444: out to be related to the influence of the initial velocity on the rise
445: profile in these MHD simulations. We focus on this aspect below.
446:
447: Our simulations are largely similar to those of kinking flux ropes in
448: \cite{torok+kliem2005}, and we refer to that study for model details.
449: The flux rope equilibrium by \cite{Titov&Demoulin99} (TD99) is used as
450: initial condition. The line current in that model, which introduces a
451: stabilizing external toroidal field is here set to zero, and the
452: stabilizing influence of line tying is kept small by choosing a torus
453: center only one tenth of the initial apex height below the bottom plane. In
454: order to preclude the helical kink instability, a sub-critical twist of
455: $\Phi\approx2.5\pi$ is chosen, which requires the flux rope to be
456: relatively thick (the minor radius is
457: 0.6 times the initial apex height, yielding an aspect
458: ratio of only 1.83). The approximation of a slender flux tube used in
459: TD99 becomes relatively inaccurate for these settings, so that the
460: simulations start with a short phase of relaxation toward a numerical
461: equilibrium, lasting about a dozen Alfv\'en times ($\tau_A$).
462:
463:
464: The TI is triggered by the motions set up in the relaxation phase,
465: which may reach one tenth of the Alfv\'en speed ($V_A$, measured at
466: the flux rope apex in the initial configuration),
467: depending on parameters. In a first set of
468: simulations, we set the decay index of the external poloidal field at
469: the initial apex height to be $n=1.20$, close to its critical value
470: analytically derived to be $1.23$ for the parameters given \cite[see
471: Eq.~(5) in][]{Kliem&Torok06}. The TI then develops very gradually, in
472: a period of $\sim100~\tau_A$, while the perturbations caused by the
473: initial relaxation decay in $\approx15\mbox{--}20~\tau_A$. This
474: simulation yields a clearly exponential rise profile (solid
475: lines in Fig.~\ref{fig:TIseries}).
476:
477: In four subsequent runs in this set, an upward, linearly rising
478: perturbation velocity is imposed at the flux rope apex of the same
479: initial configuration at the start of the runs with an increasing
480: duration (from $6~\tau_A$ up to $10~\tau_A$).
481: Figure~\ref{fig:TIseries} shows the resulting transition from exponential to
482: power-law-like rise profiles for these TI simulations.
483:
484: The fourth run (dashed lines) approaches a constant-jerk rise
485: profile best. This best-fit run has an initial velocity of $0.03~V_A$
486: at the onset of the TI-driven rise of the acceleration at
487: $t\approx15~\tau_A$ and $h=1.74$. It approximates constant jerk up to
488: $t\sim30~\tau_A$ (i.e., nearly until the peak acceleration is
489: reached) and $h\sim2.8$.
490:
491: Figure~\ref{fig:constant_jerk_TI} shows another case of close approach
492: to a cubic rise profile by a flux rope in a strongly torus-unstable
493: TD99 equilibrium \referee{ (The included scaling of the simulation
494: data to the rise profile of the 2005/06/16 eruption is discussed
495: below.)}. {Here the decay index of the external field at the initial apex
496: height is strongly supercritical, $n=2.85$, close to the asymptotic
497: value for a dipole field ($n\to3$) in the TD99 equilibrium
498: (Figure~\ref{fig:simseq} shows a rendering of this simulation). On the
499: other hand, with the depth of the torus center chosen to be 3/8 of the
500: initial apex height, the line tying has a stronger stabilizing
501: effect. Except for a somewhat larger aspect ratio of 2.3, the other
502: parameters are identical to those of the runs shown in
503: Fig.~\ref{fig:TIseries}.} The initial velocity at the onset of the
504: TI-driven rise of the acceleration is, again, approximately
505: $0.03~V_A$. This velocity results from the initial, more vigorous
506: relaxation towards a numerical equilibrium and from the early onset of
507: magnetic reconnection in the vertical {current sheet, which is formed
508: below the flux rope similar to the simulation shown in
509: \cite*{Kliem&al04}. By the end of the relaxation
510: ($t\approx10~\tau_A$), both upward and downward reconnection outflow
511: jets from the current sheet are formed and the upward jet reaches
512: $0.03~V_A$, reducing the marked decrease of the upward perturbation
513: velocities observed in the first run in
514: Fig.~\ref{fig:TIseries}. During the whole phase of nearly
515: constant-jerk rise, the TI-driven rise of the flux rope apex and the
516: upward reconnection outflow jet grow synchronously, reaching similar
517: velocities.
518:
519: Such close coupling between the ideal instability and reconnection can
520: obviously support a power-law rise of the unstable flux rope, but it
521: is neither a necessary nor a sufficient condition for its occurrence,
522: as the comparison with Run~4 in Fig.~\ref{fig:TIseries} and with the
523: CME simulation in \cite{torok+kliem2005} shows. Run~4 exhibits a
524: nearly power-law rise, but reconnection outflow jets from the vertical
525: current sheet develop here only after the acceleration of the flux
526: rope has passed its peak ($t>40~\tau_A$). The CME simulation in
527: \cite{torok+kliem2005} showed a similar coupling between the ideal MHD
528: instability (the helical kink in this case) and reconnection as the
529: run shown in Fig.~\ref{fig:constant_jerk_TI}, but with an initial
530: velocity of $\approx0.01~V_A$ the rise was clearly exponential.
531:
532: While all data in Fig.~\ref{fig:TIseries} and the solid line in
533: Fig.~\ref{fig:constant_jerk_TI} monitor the apex of the magnetic axis
534: of the flux rope, the dashed lines in Fig.~\ref{fig:constant_jerk_TI}
535: show the rise of a fluid element near the bottom of the flux rope,
536: \referee{which is a likely
537: location for the formation of filaments. Lying initially at $0.65\,h_0$,
538: it belongs to an outer flux surface of the rope.}
539: Although the flux rope in the simulation
540: expands during the rise, both the axis and the bottom part show an
541: approximately constant jerk, and no significant timing differences
542: between the acceleration profiles.
543:
544: \subsection{Scaling simulation to observation}
545: \label{ssec:Scaling}
546:
547: \referee{
548: Figure~\ref{fig:constant_jerk_TI} presents a scaling of the simulation
549: data to the rise profile of the 2005/06/16 eruption, determined in
550: three steps. First, the time of the velocity minimum near $10\,\tau_A$
551: in the simulation is associated with the onset time, $t_0$, of the
552: rapid-acceleration phase, 19:54:58~UT, as obtained in Sect.~2.
553: Second, the time $t_1$ of maximum simulated velocity is associated
554: with the time halfway between the final MLSO data points, which yields a
555: substantially better match between the acceleration profiles than
556: assuming that the acceleration ceased at or after the final MLSO data.
557: These two
558: choices yield $\tau_A=32.5$~sec. Third, the simulated and observed
559: heights are matched at $t_1$, resulting in a length unit for the
560: simulation of $h_0=44.4$~Mm, an Alfv\'en speed
561: $V_A=h_0/\tau_A=1370$~km/s, and a normalization value for the
562: acceleration of $a_0=V_A/\tau_A$. Figure~\ref{fig:constant_jerk_TI}
563: shows the observed heights on a linear scale, with derived velocity
564: and acceleration data (based on central differences, with a 7-point
565: boxcar averaging to smooth the heights and velocities, and a 5-point
566: boxcar averaging for the accelerations).}
567:
568: \referee{
569: Both the rise of the magnetic axis of the flux rope (solid line in
570: Figure~\ref{fig:constant_jerk_TI}) and the rise of a fluid element
571: originally below the magnetic axis (dashed line) are scaled to the
572: data. The lower fluid element yields the best match, and is shown in
573: Fig.~\ref{fig:constant_jerk_TI}.}
574:
575: \referee{
576: We note that a correction of the observed heights for perspective
577: foreshortening may improve the fit of model to observations. The
578: \textsl{TRACE} images suggest that the direction of ascent may have
579: been inclined from the vertical direction by $\sim45^\circ$ at the
580: onset of the accelerated rise, and it is plausible to assume that it
581: had become vertical by the time of the final MLSO data point. Such a
582: correction brings all height data points even closer to the dashed
583: line in Fig.~\ref{fig:constant_jerk_TI}. However, since such a
584: correction introduces a degree of uncertainty while the effects are
585: relatively minor, we do not attempt to apply such a correction. }
586:
587: \referee{
588: Not only is the overall match between the observations
589: and the scaled simulation quite satisfactory,
590: the scaling also yields plausible values for the Alfv\'en velocity and the
591: footpoint spacing of the model flux rope, $D_\mathrm{foot}=98$~Mm. The
592: latter agrees well with the observed value, which Fig.~1 suggests
593: to be $\approx94$~Mm (from $x\approx230$ to $x\approx490$), or slightly
594: larger owing to foreshortening. The TD99 equilibrium
595: by construction tends to yield a systematically large initial apex
596: height, so that the footpoint distance provides a far better check of the
597: length scale when the model is confronted with observations.}
598:
599: \referee{
600: The scaling also shows that the filament velocity at the onset of the
601: rapid rise (first data points after 19:55~UT) nearly reaches the value
602: of $\sim0.03\,V_A$ required in the simulations of
603: Figs.~\ref{fig:TIseries} and \ref{fig:constant_jerk_TI} for the
604: transition from an exponential to a nearly cubic height-time profile. The
605: observed velocities at $t_0$
606: even exceeded
607: the initial velocity of $0.017\,V_A$ of the run shown dotted in
608: Fig.~\ref{fig:TIseries}, which developed an intermediate rise profile
609: quite close to the observed profile. We infer from this
610: that the initial velocity is a parameter which helps
611: control the detailed properties of the rise profile.}
612:
613: \referee{
614: The observations of the eruption on 2005/07/27 do not constrain the
615: scaling of the simulation as well as the 2005/06/16 data. The LASCO data
616: at large distances refer to the leading edge of the CME, i.e., to a
617: different part of the ejection than the \textsl{TRACE} data, and the two
618: sets do not join to form as nearly a continuous $h(t)$ profile as the
619: 2005/06/16 data. Only the \textsl{TRACE} data can be used for the
620: scaling, leaving more ambiguity in the scaling for this event. The best
621: match between the simulation and the data is obtained when the final
622: \textsl{TRACE} height measurement is assumed to lie slightly past the
623: time of peak acceleration, by $2\mbox{--}5\,\tau_A$. Equating the
624: simulated and observed heights at this time gives a match of comparable
625: quality to the one in Fig.~\ref{fig:constant_jerk_TI} for both the
626: magnetic axis and the lower fluid element. We present the former in
627: Fig.~\ref{fig:scaling20050727}, which yields the scaled parameters
628: $\tau_A=26$~sec, $V_A=940$~km/s, and $D_\mathrm{foot}=55$~Mm. Scaling the
629: rise of the lower fluid element to the observations yields
630: $\tau_A=29$~sec, $V_A=1500$~km/s, and $D_\mathrm{foot}=99$~Mm instead. As
631: with the 2005/06/16 data, the observed velocity closely approaches the
632: scaled simulation velocity shortly after the estimated onset time of
633: the fast rise (within $\sim5\,\tau_A$).}
634:
635: \referee{
636: The scalings support the hypothesis that the torus instability of a
637: flux rope has been a possible driver of both eruptive filaments in
638: their rapid-acceleration phase. We note that the only parameter that
639: was adjusted particularly to fit the observations is the decay index
640: for the overlying field ($n=2.85$), since both eruptions evolved into
641: a moderately fast CME and the TI requires $n\gtrsim 2$ to produce a
642: fast ejection (\citep{Torok&Kliem2007}).}
643:
644: \subsection{Dynamics of overlying loops}
645: \label{ssec:OverlyingLoops}
646:
647: \referee{
648: Figure~\ref{fig:TI_4frame_rendering} shows that field lines that
649: initially pass over the legs of the flux rope, lean strongly sideways
650: during the rope's rapid acceleration phase, similar to the motion of
651: the observed overlying loops. Their lateral motions in Figs.~3 and\,6
652: commence with little or no delay to the beginning rapid acceleration
653: of the filament (except for a much weaker lateral motion of the left
654: overlying loop in the slow rise phase of the 2005/07/27 event), and
655: they combine with the vertical motions such that the total distance
656: between loop apex and filament apex varies only little in the first
657: $\approx5$ minutes of the rapid-acceleration phase (corresponding to
658: $\sim10\,\tau_A$), but increases rapidly thereafter.}
659:
660: \referee{
661: We emphasize that the observations of the two events do not permit us
662: to determine the delay between the start of the displacement of the
663: overlying loops relative to the filament's rapid acceleration to
664: better than an Alfv\'en travel time: the Alfv{\'e}n velocities of
665: order 1,000\,km/s and the instrument cadence mean that signals can
666: propagate between the overlying loops and the filament within 1 to 2
667: imaging intervals. Consequently, we can only conclude that the data
668: are compatible with a delay of at most one Alfv\'en travel time.}
669:
670: \referee{
671: Figure~\ref{fig:OverlyingLoops} plots the distances for a set of loops
672: in a format similar to Figs.~3 and~6. These loops were selected such
673: that their apex points have equally-spaced initial distances on a
674: straight line from the origin, inclined by $25^\circ$ from the
675: vertical. The second lowest of these loops is marked by an asterisk in
676: Fig.~\ref{fig:TI_4frame_rendering}. We find that the model's
677: horizontal and vertical distances combine to a slowly varying total
678: distance for about $10\,\tau_A$ after TI onset (at
679: $t\approx10\,\tau_A$), followed by a rapid increase of the total
680: distance, as in the observations. This behavior occurs in an angular
681: range between the vertical and the initial origin-apex line of,
682: roughly, 20--35$^\circ$. For larger inclinations of the overlying
683: loop the initial ratio of vertical and horizontal distance is smaller
684: than observed, and for smaller inclinations the horizontal motion
685: commences too late.}
686:
687: \referee{
688: Figure~11 also reveals two types of perturbations in this simulation. The
689: first is an initial phase of relaxation from the analytical TD99 field to
690: a nearby, numerically nearly potential-field state, which occurs in the
691: whole surrounding field of the flux rope and is of nearly uniform
692: duration of $2\mbox{--}3\,\tau_A$. The second is a wave-like
693: perturbation, launched by the (more vigorous) initial relaxation of the
694: current-carrying flux rope, of duration $\sim10\,\tau_A$, and propagating
695: outward trough the whole box at about the Alfv\'en speed. The motion of
696: the overlying loops is seen to commence with the passage of the second
697: perturbation, i.e., with a delay of only one Alfv\'en travel time, and to
698: continue smoothly after its passage (similar to the behavior of the flux
699: rope, whose instability develops out of the initial relaxation). A delay
700: this short is consistent with the observations.}
701:
702: \referee{
703: The feature of an initially only slowly varying total distance occurs in
704: a substantial height range, so that one cannot conclude that the observed
705: overlying loops give a good indication of the edge of the flux rope in
706: the two events considered. However, with increasing initial height of the
707: loops, the phase of rapid increase of the distance to the rope occurs
708: progressively delayed. The scalings place the observed transition between
709: the two phases at $t\sim20\,\tau_A$, in agreement with the lowest two or
710: three loops included in Fig.~\ref{fig:OverlyingLoops}, indicating that
711: the overlying loops were located in the range between the surface of the
712: flux rope and about three minor radii from its axis.}
713:
714: \section{Conclusions}
715: \label{sec:conclusions}
716: We study two well-observed filament eruptions, and find that their
717: rapid acceleration phases are well fit by a cubic height-time curve
718: that implies a nearly constant jerk for $10-15$\,minutes, followed by
719: a transition to a terminal velocity of $\sim 750$\,km/s and $\sim
720: 1250$\,km/s, respectively. Simulations of a torus instability (TI) can
721: reproduce such a behavior, provided that a substantial initial
722: velocity perturbation is introduced. Without that perturbation, an
723: exponential rise profile would be found.
724:
725: We note that the initial slow rise and the onset of the subsequent rapid
726: acceleration phase are shared between the filament and overlying loop
727: structures: neither \referee{leads the other to within the temporal
728: resolution. For characteristic
729: Alfv{\'e}n speeds over active regions of $\sim 1,000$\,km/s, the
730: propagation of a perturbation over the separation of $\sim 75,000$\,km
731: would require only $\sim 1.2$\,min., which corresponds to only one or
732: two exposures. Thus the observations allow for Alfv{\'e}nic propagation
733: of a signal between filament and overlying loops, but suggest no longer-term
734: differential evolution.}
735:
736: We observe no significant changes in the separation of erupting
737: filament and overlying loops within that interval (Figs.~\ref{fig:1c}
738: and~\ref{fig:2c}). After that, the distance increases in the
739: 2005/06/16 eruption, suggesting the overlying field moves to the side
740: for some time faster than the filament rises. For the 2007/07/27
741: eruption, the distance stays the same for one loop and decreases for
742: another for up to 10~min after the start of the rapid acceleration
743: phase, which reflects the significant sideways motion component of the
744: rising filament. The observed configuration of the filament and high
745: loops may be part of a larger overall destabilizing field
746: configuration. Our numerical modeling has assumed that, in the rapid
747: acceleration phase, the overlying field starts to move rapidly only
748: as a consequence of the flux rope's destabilization. This is
749: consistent with the data. However, we cannot exclude that the filament
750: and the overlying field were destabilized simultaneously by a process
751: different from the one considered here. More study is needed
752: to establish whether the common evolution of the filament and high
753: loops has a significant diagnostic value as to the cause of the
754: instability.
755:
756: Comparison with other model studies in the literature leads us to
757: conclude that the catastrophe model and the TI model are both
758: marginally consistent with the observations of the two erupting
759: filaments. The catastrophe model predicts a power-law
760: exponent near the lower edge of the range of acceptable fits, but we
761: have to allow for the possibility that changing that model's details
762: may change the acceleration profile.
763: \referee{In order to yield the observed nearly cubic power-law rise (with
764: $m$ slightly exceeding 3), our TI model requires an initial
765: perturbation velocity that is in agreement with the observed rise
766: velocity at the onset of the rapid-acceleration phase. If a nearly
767: exact cubic rise were to be matched, however, initial velocities
768: moderately exceeding the observed ones, by a factor $\approx1.5$, were
769: required. In any case, our modeling is consistent with the observed
770: velocities after the first few minutes of the eruption.}
771:
772: Having established that the model for the TI instability is very
773: sensitive to the initial conditions, we should of course also
774: acknowledge that it depends sensitively on the model details
775: itself. These include the details of the external field and of the
776: rates and locations of the reconnection that occurs behind the
777: erupting filament. That such reconnection occurs in reality is
778: suggested for both events by the occurrence of brightenings mainly at
779: the bottom side of the filaments at the onset of the
780: rapid-acceleration phase. These brightenings develop later into the
781: streaks used for position determination in
782: Sect.~\ref{sec:Observations}. The onset of reconnection even before
783: the rapid-acceleration phase of the filament eruption on 27 July 2005
784: is strongly suggested by precursor soft and hard X-ray emission during
785: about 04:00--04:30~UT, whose analysis revealed heating to 15~MK and
786: the acceleration of non-thermal electrons to energies $>10$~keV
787: (\citep{Chifor&al06}).
788:
789: The observed rise velocity early in the filament eruption may
790: be an underestimate of the true expansion
791: velocity of the hoop formed by the flux rope: the filament channel in
792: the pre-eruption phase of AR\,10775 is strongly curved, and one of the
793: two possible channels in AR\,10792 is too (ambiguity exists here
794: because the eruptions occurred very near the limb, so that the
795: configurations of the filament channels can only be observed some days
796: before and after the events, respectively). If the initial expansion
797: of the flux rope would have a strong component in the general
798: direction of the inclined plane of the curved filament channel rather
799: than be purely normal to the solar surface, projection effects could
800: cause us to underestimate the expansion velocity in particular early in the
801: evolution. In addition to that, we must realize that the TI model
802: assumes a flux rope that stands normal to the solar surface and that
803: erupts radially. Future more detailed modeling will have to show how
804: deviations from that affect the evolution of the eruption.
805:
806: The fact that the torus-instability model yields qualitatively
807: different rise profiles (exponential vs.\ power law) in different
808: parts of parameter space, cautions against expectations that precise
809: measurements of the rise profile of filament eruptions by themselves
810: permit a determination of the driving process: the non-linearities in
811: the eruption models clearly require high-fidelity modeling if such
812: observations are to be used to differentiate successfully between
813: competing models. Our initial modeling discussed here suggests
814: that the torus instability is a viable candidate mechanism for at
815: least some filament eruptions in coronal mass ejections.
816: \referee{Given the dependence of nonlinear models on the details
817: of boundary and initial conditions, it will be necessary to investigate
818: how other models for erupting filaments compare to the data, as well as
819: how the fidelity of our modeling of the torus instability can be improved
820: before we can reach definitive conclusions about the mechanism(s) responsible
821: for filament eruptions in general.}
822:
823: \acknowledgements
824: We thank Joan Burkepile for providing us with MLSO MK4 observations,
825: and Terry Forbes for helpful discussions. We are grateful to the
826: referee for constructive and helpful comments that led us to pursue
827: the model-observation parallels in this study in more detail. This
828: work was supported by NASA under the TRACE contract NAS5-38099 with
829: NASA Goddard Space Flight Center, by NSF grant ATM 0518218 to the
830: University of New Hampshire, by the European Commission through the
831: SOLAIRE Network (MTRN-CT-2006-035484), and by the Deutsche
832: Forschungsgemeinschaft.
833:
834: %\references
835: \begin{thebibliography}{}
836:
837: \bibitem[Alexander et al.(2002)]{alexander&al02}
838: Alexander, D., Metcalf, T.~R., \& Nitta, N.~V.\ 2002, \grl, 29, 41
839:
840: \bibitem[Bateman (1978)]{bateman1978}
841: Bateman, G.\ 1978, MHD Instabilities, MIT, Cambridge, MA.
842:
843: %\bibitem[\protect\citeauthoryear{{Bong} {\em et~al.\/}}{2006}]{bong+etal2006}
844: %{Bong}, S.-C., {Moon}, Y.-J., {Cho}, K.-S., {Kim}, Y.-H., {Park}, Y.~D., {\&}
845: % {Choe}, G.~S.: 2006,
846: %\newblock ApJL 636, 169
847:
848: \bibitem[\protect\citeauthoryear{{Brueckner} {\em
849: et~al.\/}}{1995}]{brueckner+etal1995}
850: {Brueckner}, G.~E., {Howard}, R.~A., {Koomen}, M.~J., {Korendyke}, C.~M.,
851: {Michels}, D.~J., {Moses}, J.~D., {Socker}, D.~G., {Dere}, K.~P., {Lamy},
852: P.~L., {Llebaria}, A., {Bout}, M.~V., {Schwenn}, R., {Simnett}, G.~M.,
853: {Bedford}, D.~K., {\&} {Eyles}, C.~J.: 1995,
854: \newblock SPh 162, 357
855:
856: \bibitem[Chifor et al.(2006)]{Chifor&al06}
857: Chifor, C., Mason, H.~E., Tripathi, D., Isobe, H., \& Asai, A.\
858: 2006, \aap, 458, 965
859:
860: \bibitem[\protect\citeauthoryear{{Fan}}{2005}]{fan2005}
861: {Fan}, Y.: 2005,
862: \newblock ApJ 630, 543
863:
864: \bibitem[\protect\citeauthoryear{{Gallagher} {\em
865: et~al.\/}}{2003}]{gallagher+etal2003}
866: {Gallagher}, P.~T., {Lawrence}, G.~R., {\&} {Dennis}, B.~R: 2003,
867: \newblock ApJL 588, 53
868:
869: \bibitem[\protect\citeauthoryear{{Gilbert} {\em
870: et~al.\/}}{2000}]{gilbert+etal2000}
871: {Gilbert}, H.~R., {Holzer}, T.~E., {Burkepile}, J.~T., {\&} {Hundhausen},
872: A.~J.: 2000,
873: \newblock ApJ 537, 503
874:
875: \bibitem[\protect\citeauthoryear{{Goff} {\em et~al.\/}}{2005}]{goff+etal2005}
876: {Goff}, C.~P., {van Driel-Gesztelyi}, L., {Harra}, L.~K., {Matthews}, S.~A.,
877: {\&} {Mandrini}, C.~H.: 2005,
878: \newblock A{\&}A 434, 761
879:
880: \bibitem[\protect\citeauthoryear{{Gopalswamy} {\em
881: et~al.\/}}{2003}]{gopalswamy+etal2003}
882: {Gopalswamy}, N., {Shimojo}, M., {Lu}, W., {Yashiro}, S., {Shibasaki}, K., {\&}
883: {Howard}, R.~A.: 2003,
884: \newblock ApJ 586, 562
885:
886: \bibitem[\protect\citeauthoryear{Handy {\em et~al.\/}}{1999}]{traceinstrument}
887: Handy, B.~N., Acton, L.~W., Kankelborg, C.~C., {\em et~al.\/}: 1999,
888: \newblock SPh 187, 229
889:
890: \bibitem[\protect\citeauthoryear{{Hori} {\em et~al.\/}}{2005}]{hori+etal2005}
891: {Hori}, K., {Ichimoto}, K., {Sakurai}, T., {Sano}, I., {\&} {Nishino}, Y.:
892: 2005,
893: \newblock ApJ 618, 1001
894:
895: \bibitem[\protect\citeauthoryear{{Kahler} {\em
896: et~al.\/}}{1988}]{kahler+etal1988}
897: {Kahler}, S.~W., {Moore}, R.~L., {Kane}, S.~R., {\&} {Zirin}, H.: 1988,
898: \newblock ApJ 328, 824
899:
900: \bibitem[Kliem and T\"{o}r\"{o}k(2006)]{Kliem&Torok06}
901: Kliem, B., \& T\"{o}r\"{o}k, T.\ 2006, Phys.\ Rev.\ Lett., 96,
902: 255002
903:
904: \bibitem[Kliem et al.(2004)]{Kliem&al04}
905: Kliem, B., Titov, V. S., \& T\"{o}r\"{o}k, T.\ 2004, \aap, 413, L23
906:
907: \bibitem[\protect\citeauthoryear{{Kundu} {\em et~al.\/}}{2004}]{kundu+etal2004}
908: {Kundu}, M.~R., {White}, S.~M., {Garaimov}, V.~I., {Manoharan}, P.~K.,
909: {Subramanian}, P., {Ananthakrishnan}, S., {\&} {Janardhan}, P.: 2004,
910: \newblock ApJ 607, 530
911:
912: \bibitem[Lynch et al.(2004)]{Lynch&al04}
913: Lynch, B.~J., Antiochos, S.~K., MacNeice, P.~J., Zurbuchen,
914: T.~H., \& Fisk, L.~A.\ 2004, \apj, 617, 589
915:
916:
917: \bibitem[Mari{\v c}i{\'c} et al.(2004)]{maricic+etal2004}
918: {{Mari{\v c}i{\'c}}, D., {Vr{\v s}nak}, B., {Stanger}, A.~L., \&
919: {Veronig}, A.}\ 2004, \solphys, 225, 337
920:
921: \bibitem[Priest and Forbes(2002)]{Priest&Forbes02}
922: Priest. E.~R., \& Forbes, T.~G.\ 2002, \aapr, 10, 313
923:
924: \bibitem[\protect\citeauthoryear{{Sterling} and
925: {Moore}}{2004}]{sterling+moore2004}
926: {Sterling}, A.~C. {\&} {Moore}, R.~L.: 2004,
927: \newblock ApJ 613, 1221
928:
929: \bibitem[\protect\citeauthoryear{{Sterling} and
930: {Moore}}{2005}]{sterling+moore2005}
931: {Sterling}, A.~C. {\&} {Moore}, R.~L.: 2005,
932: \newblock ApJ 630, 1148
933:
934: \bibitem[Titov and D\'emoulin(1999)]{Titov&Demoulin99}
935: Titov, V. S., \& D\'emoulin, P.\ 1999, \aap, 351, 707 (TD99)
936:
937: \bibitem[\protect\citeauthoryear{{T{\" o}r{\" o}k} and
938: {Kliem}}{2005}]{torok+kliem2005}
939: {T{\" o}r{\" o}k}, T. {\&} {Kliem}, B.: 2005,
940: \newblock ApJL 630, 97
941:
942: \bibitem[T\"or\"ok \& Kliem(2007)]{Torok&Kliem2007}
943: T\"or\"ok, T., \& Kliem, B. 2007,
944: \newblock Astron. Nachr. 328, 743
945:
946: \bibitem[T\"{o}r\"{o}k et al.(2004)]{Torok&al04}
947: T\"{o}r\"{o}k, T., Kliem, B., \& Titov, V. S.\ 2004, \aap, 413, L27
948:
949: \bibitem[van Tend and Kuperus(1978)]{vanTend&Kuperus78}
950: van Tend, W., \& Kuperus, M.\ 1978, \solphys, 59, 115
951:
952: \bibitem[\protect\citeauthoryear{{Vr{\v s}nak}}{2001}]{vrsnak2001}
953: {Vr{\v s}nak}, B.: 2001,
954: \newblock JGR 106, 25249
955:
956: \bibitem[\protect\citeauthoryear{{Williams} {\em
957: et~al.\/}}{2005}]{williams+etal2005}
958: {Williams}, D.~R., {T{\" o}r{\"o}k}, T., {D{\'e}moulin}, P., {van
959: Driel-Gesztelyi}, L., {\&} {Kliem}, B.: 2005,
960: \newblock ApJL 628, 163
961:
962: \bibitem[\protect\citeauthoryear{{Zhang} and {Golub}}{2003}]{zhang+golub2003}
963: {Zhang}, M. {\&} {Golub}, L.: 2003,
964: \newblock ApJ 595, 1251
965:
966: \end{thebibliography}
967:
968:
969: \vfill\eject
970: \section*{Figure captions:}
971:
972: Figure~\ref{fig:1a}: TRACE 171\,\AA\ image taken at 2005/06/16 19:25:32\,UT.
973: Sample outlines of the top edge of the rising filament over time
974: and of two overlying loop
975: structures are shown for the time interval from 19:25\,UT to 20:04\,UT.
976: The positions for which the heights are shown in Fig.~\ref{fig:1b} are marked.
977:
978: Figure~\ref{fig:1b}: Distances from the solar EUV limb for the
979: rising and erupting filament on 2005/06/16
980: (see Fig.~\ref{fig:1a} for the tracked
981: positions). The bottom panel shows the central phase with rapid
982: filament acceleration in detail. Both panels show several
983: fits to the data (see legend). The top panel also shows positions
984: derived from the MLSO coronagraphic data for the later phase when the eruption
985: turns into a proper mass ejection.
986:
987: Figure~\ref{fig:1c}: Distances between the tracked ridge of the
988: filament shown in Fig.~\ref{fig:1a} and the lower (top panel)
989: and upper (bottom panel) overlying loops. The total distance is
990: shown by the solid line; distances in the figure's $x$ and $y$ directions
991: are shown separately
992: by dashed and dashed-dotted lines, respectively.
993:
994: Figure~\ref{fig:2a}: TRACE 171\,\AA\ image taken at 2005/07/27 03:00:08\,UT.
995: This figure, similar to Fig.~\ref{fig:1a}, identifies segments of
996: overlying loops on the left and right side of the rising filament.
997:
998: Figure~\ref{fig:2b}: As Fig.~\ref{fig:1b} for the event observed
999: on 2005/07/27. Note that the exponential and quadratic fits
1000: are shown offset by +2\,min. in the top panel to reduce overlap,
1001: but shown properly placed in time in the lower panel.
1002:
1003: Figure~\ref{fig:2c}: Distances between the tracked ridge of the
1004: filament shown in Fig.~\ref{fig:2a} and the left (top panel)
1005: and right (bottom panel) overlying loops.
1006:
1007: Figure~\ref{fig:TIseries}:
1008: Transition from exponential to approximately power-law rise profile
1009: with increasing initial velocity for a torus-unstable flux rope
1010: equilibrium with an external field decay index of $n\ge 1.2$ (see text
1011: for other parameter values). Apex height $h(t)$, velocity $u(t)$,
1012: acceleration $a(t)$, and jerk $j(t)=da/dt$ are normalized using the
1013: initial apex height $h_0$, the Alfv\'en speed $V_A$, and the
1014: corresponding derived quantities. Time is normalized by
1015: $\tau_A=h_0/V_A$. Solid lines show the unperturbed run, i.e., the
1016: development of the instability from rest. For the further runs of the
1017: series a velocity perturbation at the apex is linearly ramped up until
1018: 6, 8, 9.25, and 10~$\tau_A$ (dashed-dotted, dotted, dashed,
1019: dashed-triple dotted, respectively).
1020:
1021: Figure~\ref{fig:constant_jerk_TI}:
1022: Nearly constant-jerk rise profile for an unperturbed torus-unstable
1023: flux rope equilibrium with steeper field decrease above the flux rope
1024: than in Fig.~\ref{fig:TIseries}; the field decay index in this case is
1025: $n\ge 2.85$, i.e., near the value for the far field in the dipolar case
1026: (see text for other parameter differences for aspect ratio and initial
1027: torus depth). Solid lines show the rise profile of the apex point of
1028: the magnetic axis as in Fig.~\ref{fig:TIseries}, dashed lines show the
1029: rise profile of a fluid element below the apex, initially at
1030: $h=0.65\,h_0$. \referee{ The simulation data for this lower fluid element are
1031: scaled to the rise profile of the 2005/06/16 filament eruption, and
1032: the resulting Alfv\'en time, Alfv\'en speed, and footpoint distance
1033: are given.}
1034:
1035: Figure~\ref{fig:simseq}: Side view of a torus instability simulation
1036: (see Fig.~\ref{fig:constant_jerk_TI}). The field lines of the torus
1037: are shown lying in a flux surface at half the minor torus
1038: radius. Sample field lines for the overlying field are also shown. The
1039: starting points in the bottom plane for the traced field lines are the
1040: same for all panels. The times (expressed in Alfv\'en crossing times,
1041: as in Figs.~\ref{fig:TIseries}--\ref{fig:OverlyingLoops}) are
1042: $=0,\,20,\,30$, and 40, respectively. \referee{The motion of the loop apex
1043: marked by an asterisk is shown in Fig.~\ref{fig:OverlyingLoops}.}
1044:
1045: Figure~\ref{fig:scaling20050727}:
1046: \referee{Scaling of the simulation data from Fig.~\ref{fig:constant_jerk_TI} to
1047: the rise profile of the 2005/07/27 filament eruption; here the rise of
1048: the magnetic axis' apex point (solid line) is scaled.}
1049:
1050: Figure~\ref{fig:OverlyingLoops}:
1051: \referee{Distances of the apex point of representative loops, initially
1052: overlying the flux rope at an angle of $25^\circ$ from the vertical,
1053: to the lower fluid element of the simulation shown in
1054: Figs.~\ref{fig:constant_jerk_TI} and \ref{fig:scaling20050727} (dashed
1055: line in these figures). The format is similar to Figs.~3 and~6. For
1056: clarity, horizontal and vertical distances are included only for the
1057: lowest and highest of the selected loops. The second
1058: lowest of these loops is marked by an asterisk in
1059: Fig.~\ref{fig:TI_4frame_rendering}.}
1060:
1061: \vfill\eject
1062:
1063: \begin{figure}[ht!]
1064: \epsscale{.65}
1065: %\plotone{20050616/all_tracks_n2.eps}
1066: \plotone{fig1.eps}
1067: \caption{\em }\label{fig:1a}
1068: %\epsscale{1.0}
1069: \end{figure}
1070:
1071: \begin{figure}[ht!]
1072: \epsscale{.65}
1073: %\plotone{Accell/20050616_full.eps}
1074: %\plotone{Accell/20050616_accell.eps}
1075: \plotone{fig2a.eps}
1076: \plotone{fig2b.eps}
1077: \caption{\em }\label{fig:1b}
1078: %\epsscale{1.0}
1079: \end{figure}
1080:
1081:
1082: \begin{figure}[ht!]
1083: \epsscale{.65}
1084: %\plotone{Accell/20050616_upperDiff.eps}
1085: %\plotone{Accell/20050616_innerDiff.eps}
1086: \plotone{fig3a.eps}
1087: \plotone{fig3b.eps}
1088: \caption{\em }\label{fig:1c}
1089: %\epsscale{1.0}
1090: \end{figure}
1091:
1092: \begin{figure}[ht!]
1093: \epsscale{.65}
1094: %\plotone{20050727/all_tracks_n2.eps}
1095: \plotone{fig4.eps}
1096: \caption{\em }\label{fig:2a}
1097: %\epsscale{1.0}
1098: \end{figure}
1099:
1100: \begin{figure}[ht!]
1101: \epsscale{.65}
1102: %\plotone{Accell/20050727_full_time_log.eps}
1103: %\plotone{Accell/20050727_accell.eps}
1104: \plotone{fig5a.eps}
1105: \plotone{fig5b.eps}
1106: \caption{\em }\label{fig:2b}
1107: %\epsscale{1.0}
1108: \end{figure}
1109:
1110:
1111: \begin{figure}[ht!]
1112: \epsscale{.65}
1113: %\plotone{Accell/20050727_leftdiff.eps}
1114: %\plotone{Accell/20050727_rightdiff.eps}
1115: \plotone{fig6a.eps}
1116: \plotone{fig6b.eps}
1117: \caption{\em }\label{fig:2c}
1118: %\epsscale{1.0}
1119: \end{figure}
1120:
1121:
1122: \begin{figure} %[t]
1123: \epsscale{.65}
1124: % \includegraphics[width=3.25in]{jerk_acc6_all_paper.ps}
1125: % \includegraphics[width=4.25in]{fig7.eps}
1126: \plotone{fig7.eps}
1127: \caption[]
1128: {}
1129: \label{fig:TIseries}
1130: \end{figure}
1131:
1132: \begin{figure} %[t]
1133: \epsscale{.65}
1134: % \includegraphics[width=3.25in]{jerk_accd300_nopert_paper.ps}
1135: % \includegraphics[width=4.25in]{fig8.eps}
1136: \plotone{fig8.eps}
1137: \caption[]
1138: {}
1139: \label{fig:constant_jerk_TI}
1140: \end{figure}
1141:
1142: \begin{figure}[ht!]
1143: \epsscale{.5}
1144: \plotone{simseqbw.eps}
1145: \caption{\em }\label{fig:simseq}\label{fig:TI_4frame_rendering}
1146: \end{figure}
1147:
1148: \begin{figure} %[t]
1149: \centering
1150: % \includegraphics[width=3.25in]{fit_accd300nf_20050727_h-fit-axis_paper.ps}
1151: \includegraphics[width=3.25in]{fig10.eps}
1152: \caption[]
1153: {}
1154: \label{fig:scaling20050727}
1155: \end{figure}
1156:
1157: \begin{figure} %[t]
1158: \centering
1159: \includegraphics[width=2.5in]{fig11.eps} %[width=3.25in]
1160: % {overloop2-9_accd300nf.ps}
1161: % % {overloop2-9_accd300nf_.ps}
1162: \caption[]
1163: {}
1164: \label{fig:OverlyingLoops}
1165: \end{figure}
1166:
1167: \end{document}
1168:
1169: