1: \documentclass[12pt,preprint]{aastex}
2: %
3: \newcommand{\pref}{\protect\ref}
4: \newcommand{\solrad}{\ifmmode{R}_{\rm S}\else${R}_{\rm S}$\fi}
5: \newcommand{\solmas}{\ifmmode{M}_{\rm S}\else${M}_{\rm S}$\fi}
6: \newcommand{\tintu}{$\,$ergs$\,$cm$^{-2}\,$s$^{-1}\,$sr$^{-1}$}
7: \newcommand{\intu}{$\,$ergs$\,$cm$^{-2}\,$s$^{-1}\,$sr$^{-1}\,$\AA$^{-1}$}
8: \newcommand{\ctn}{\ifmmode\kappa\else$\kappa$\fi}
9: \newcommand{\ergu}{$\,$ergs$\,$s$^{-1}$}
10: \newcommand{\flxu}{$\,$ergs$\,$cm$^{-2}\,$s$^{-1}$}
11: \newcommand{\velu}{$\,$km$\,$s$^{-1}$}
12: \newcommand{\dynu}{$\,$dyn$\,$cm$^{-2}$}
13: \newcommand{\term}[2]{\mbox{$\,^{#1}{\rm #2}$}}
14: \def\term#1 #2/{\mbox{$\,^{#1}{\rm #2}$}}
15: \newcommand{\muv}{$2s\,2p^2~^2\!D$ $-$ $2s^22p~^2\!P^o$}
16: \newcommand{\meuv}{$2s\,2p^2~^2\!S$ $-$ $2s^22p~^2\!P^o$}
17: \def\aspcs{{ASP Conf.\ Ser.}}
18:
19: \newcommand{\pder}[2]{{{\partial {#1} \over {\partial {#2}}}}}
20:
21: \renewcommand{\vec}[1]{{\bf #1}}
22: \newcommand{\cross}{\times}
23: \newcommand{\divg}{{\rm div\,}}
24: \newcommand{\grad}{{\bf grad\,}}
25: \newcommand{\curl}{{\bf curl\,}}
26: \newcommand{\jcb}{\ifmmode\vec{j}\cross\vec{B}\else$\vec{j}\cross\vec{B}$ \fi}
27: \newcommand{\coords}[2]{$x,y$=#1\arcsec,#2\arcsec}
28: \newcommand{\coordm}[2]{$x,y$=#1,#2\,Mm}
29:
30: \newcommand\tabone{
31: \protect\begin{deluxetable}{llll}
32: \renewcommand{\baselinestretch}{1.0}
33: \tablecaption{Sensitivity of MDI, KPVT and Hinode-SP
34: longitudinal magnetograms \label{tab:sens}}
35: \tablehead{Instrument/mode & noise per pixel & pixel size & noise in
36: flux \\
37: & Mx~cm$^{-2}$ & arc seconds & units of 10$^{15}$ Mx }
38: \startdata
39: MDI/full disk & 17 & $1\farcs 984\times 1\farcs984$ & 350\\
40: KPVT/synoptic & 2.8 & $1\farcs 148\times 1\farcs148$& 19\\
41: Hinode SP/normal map & 3& $0\farcs 164\times 0\farcs164$& 0.42\\
42: (Kitt Peak 40 channel magnetograph & 0.4 &$\dag$ & $\approx13$ \\
43: \citealp{Livingston+Harvey1971}) & & &\\
44: \enddata
45: %
46: \tablecomments{1\arcsec{} on the Sun corresponds to 725 km
47: \protect\citep{
48: Allen1973}. $^\dag$Seeing limited, here we use an effective pixel size
49: of $2.5\times2.5\arcsec$ corresponding to half of the quoted resolution of
50: $5\arcsec$. }
51: \end{deluxetable}
52:
53: }
54:
55:
56: \newcommand\figone{
57: %\clearpage
58: \begin{figure}[!ht]
59: \epsscale{0.9}
60: \plotone{f1.ps}
61: \caption{\label{fig:dowdy86} A sketch of the likely magnetic structure in
62: a quiet region of the Sun's magnetic network, according to
63: \protect\citet{Dowdy+Rabin+Moore1986}. The ``coronal funnels''
64: are similar to the structures modeled by \protect\citet{Gabriel1976},
65: which can account for emission above about $2\times10^5$ K. The
66: ``network loops'', arising from mixed polarity magnetic fields
67: {\em within} network boundaries, explain the cool transition region
68: emission, in Dowdy's picture.}
69: \end{figure}
70: }
71: %
72:
73: \newcommand\figtwo{
74: %\clearpage
75: \begin{figure}[!ht]
76: \epsscale{1.15}
77: \plotone{f2.ps}
78: \caption{\label{fig:context} Context images showing the VAULT FOV
79: (boxed region) for the observation at 18:12 UT on 14 June
80: 2002. Shown are
81: images of longitudinal field from MDI (between $\pm 200$ Mx~cm$^{-2}$), and of
82: He II 304 emission and 171 \AA{} Fe IX/X emission from EIT, both
83: instruments on the SOHO spacecraft. These images were
84: differentially rotated to 18:12:00 UT,
85: the epoch of these VAULT-2 observations.
86: The active region in the SE
87: corner contains a filament along a magnetic neutral line, another filament
88: is seen in the He II image, oriented roughly N-S on the western edge
89: of the VAULT FOV. Also shown are (lower right) VAULT-2 data.
90: }
91: \end{figure}
92: }
93: %
94:
95:
96:
97: \newcommand\figthree{
98: %\clearpage
99: \begin{figure}[!ht]
100: \epsscale{.98}
101: \plotone{f3.ps}
102: \caption{\label{fig:vault} The ``quiet'' part of the VAULT image
103: for the observation at 18:12 UT on 14 June
104: 2002. $+100$ Mx~cm$^{-2}$) contours of magnetic flux density from the MDI longitudinal
105: magnetogram are shown. The co-alignment of MDI
106: and VAULT-2 is reliable only to within a few arcseconds. The
107: lower panel shows an edge-enhanced image highlighting the morphology
108: of L$\alpha$ threads.
109: }
110: \end{figure}
111: }
112: %
113:
114:
115: \newcommand\figfour{
116: %\clearpage
117: \begin{figure}[!ht]
118: \epsscale{.98}
119: \plotone{f3.ps}
120: \caption{\label{fig:tricolor} A false color image of
121: VAULT-2 data
122: (yellow) together with TRACE 171\AA{} data of the lower corona
123: (blue). The TRACE data were
124: obtained 30 minutes before the VAULT-2 image. Four TRACE 171\AA{} images
125: were added together.
126: }
127: \end{figure}
128: }
129: %
130:
131: \newcommand\figfive{
132: %\clearpage
133: \begin{figure}[!ht]
134: \epsscale{.6}
135: \includegraphics[angle=90,scale=0.45]{f4a.ps}
136: \includegraphics[angle=90,scale=0.45]{f4b.ps}
137: \caption{\label{fig:kpvtvault} Magnetic field contours and
138: potential field lines superposed on the VAULT-2 image. The upper panel
139: shows loops with lengths $<$10 Mm, the lower shows
140: longer loops. Contours are at
141: $\pm 2\sigma$
142: ($\pm 5.6$ Mx cm$^{-2}$), negative
143: contours are dashed lines, positive solid. Not all field lines from
144: each pixel are plotted even if their signal exceeds the noise (see
145: text), to avoid confusion. Field lines reaching heights
146: $\le 5$ Mm are plotted as a black line, others are
147: shown as a black on top of a white line.
148: The figure origins are the same but
149: arbitrary, the center of the figure is Sun center.
150: }
151: \end{figure}
152: }
153: %
154:
155:
156:
157: \newcommand\figfivec{
158: %\clearpage
159: \begin{figure}[!ht]
160: \epsscale{1}
161: \includegraphics[width=0.55\textwidth,clip,angle=90]{f5a.ps}
162: \includegraphics[width=0.55\textwidth,clip,angle=90]{f5b.ps}
163: \caption{\label{fig:kpvtvaultsm}
164: Potential field lines
165: superposed on a portion of the VAULT-2
166: image, centered at
167: \coords{-500}{+290} in Figure \pref{fig:context}. The area covers
168: perhaps 10 supergranular areas. The upper panel shows loops of 10Mm
169: length and less, the lower panel longer loops. Contours of $\pm
170: 2\sigma$
171: ($\pm 5.6$
172: Mx cm$^{-2}$) are shown, where $\sigma$ is the rms uncertainty in
173: line of sight field strength. Field lines were plotted with no limit
174: set on the signal-to-noise ratios of the magnetogram.
175: There is little correspondence between
176: short loops and the bright L$\alpha$ emission, instead the brightest
177: emission originates from the bases of loops longer than 10Mm.
178: }
179: \end{figure}
180: }
181: %
182:
183:
184: \newcommand\figsix{
185: %\clearpage
186: \begin{figure}[!ht]
187: \epsscale{1}
188: \includegraphics[width=0.6\textwidth,clip,angle=90]{f6a.ps}
189: \includegraphics[width=0.6\textwidth,clip,angle=90]{f6b.ps}
190: %\plottwo{f6a.ps}{f6b.ps}
191: \caption{\label{fig:spqr}
192: Line of sight field strength images,
193: obtained with the SP on Hinode on 17 November 2007 beginning at 11:31 UT,
194: are shown with
195: potential field lines superposed.
196: The magnetograms are shown on a linear scale between
197: -50 and +50 Mx~cm$^{-2}$ to show weak flux regions.
198: The upper panel
199: shows loops with total length below 10 Mm, the lower shows those with
200: longer lengths. The coordinates have an
201: arbitrary origin.
202: }
203: \end{figure}
204: }
205: %
206:
207: \newcommand\figseven{
208: %\clearpage
209: \begin{figure}[!ht]
210: \epsscale{.98}
211: \includegraphics[width=0.6\textwidth,clip,angle=90]{f7a.ps}
212: \includegraphics[width=0.6\textwidth,clip,angle=90]{f7b.ps}
213: \caption{\label{fig:spar} A similar plot to
214: figure \pref{fig:spqr}, but for the active region observed with the SP
215: on Hinode, on 29
216: December 2006 beginning at 01:51 UT. (Field line cusps near $y=47$ are artifact of
217: the Fourier method used.)
218: }
219: \end{figure}
220: }
221: %
222:
223: \newcommand\figeight{
224: %\clearpage
225: \begin{figure}[!ht]
226: \epsscale{.98}
227: \includegraphics[width=0.6\textwidth,clip,angle=90]{f8a.ps}
228: \includegraphics[width=0.6\textwidth,clip,angle=90]{f8b.ps}
229: %\plotone{f8.ps}
230: \caption{\label{fig:spqrc} Close-up magnetograms of flux
231: concentrations observed by the SP on Hinode
232: from the central portions of the quiet region dataset from
233: November 27 2007 (upper panel),
234: and the more active area from
235: and 29 December 2006 (lower panel).
236: The regions show cover about the area of one supergranule.
237: Short magnetic loops abound in both regions, as the strong network
238: concentrations return to cell interiors of opposite polarity or to
239: nearby concentrations.
240: }
241: \end{figure}
242: }
243: %
244:
245: \newcommand\fignine{
246: %\clearpage
247: \begin{figure}[!ht]
248: \epsscale{.8}
249: \plotone{f9.ps}
250: \caption{\label{fig:azimuths} Magnetic field azimuths from inversions
251: of the measured
252: Hinode SP data (top), those computed from the potential field approximation
253: using just the line-of-sight component of the magnetic field
254: (center), and the difference (bottom).
255: The plot shows the lower half of the dataset obtained on 29
256: December 2006 which contains significant Stokes $Q,U$
257: signals. White regions in the bottom panel are most likely the
258: result of the 180$^\circ$ ambiguity in inverted fields. The qualitative agreement suggests that
259: such differences, if typical, do not affect our conclusions concerning
260: the locations of loops of various lengths.
261: }
262: \end{figure}
263: }
264:
265:
266:
267: \slugcomment{}
268: %\RequirePackage{rrbib}
269: %\input{rrmacros}
270: %\input{bbmacros}
271:
272: \begin{document}
273:
274: \title{\large On the magnetic structure of the solar transition region}
275:
276: \author{Philip Judge and Rebecca Centeno}
277: \affil{High Altitude Observatory,
278: National Center for Atmospheric Research\altaffilmark{1},
279: P.O. Box 3000, Boulder CO~80307-3000, USA\\ \vbox{}}
280:
281:
282: \begin{abstract}
283:
284: We examine the hypothesis that ``cool loops'' dominate emission
285: from solar transition region plasma below temperatures
286: of $2\times10^5$K. We compare published VAULT
287: images of H L$\alpha$, a lower transition region line,
288: with near-contemporaneous magnetograms from Kitt Peak, obtained during
289: the second flight (VAULT-2) on 14 June 2002.
290: The measured surface fields and potential
291: extrapolations suggest that there are
292: too few short loops, and that L$\alpha$ emission is associated with
293: the base regions of longer, coronal loops.
294: VAULT-2 data of
295: network boundaries have an
296: asymmetry on scales larger than supergranules, also
297: indicating an association with long loops.
298: We complement the Kitt Peak data with very sensitive vector
299: polarimetric data from the Spectro-Polarimeter on board Hinode,
300: to determine the influence of very small magnetic
301: concentrations on our analysis. From these data
302: two
303: classes of behavior are found: within the cores of strong
304: magnetic flux concentrations ($> 5\times10^{18}$ Mx) associated with active network
305: and plage, small-scale mixed fields are absent and any short loops can
306: connect just the peripheries of the flux to cell interiors. Core fields
307: return to the surface via longer, most likely coronal, loops. In weaker
308: concentrations, short loops can connect between concentrations and
309: produce mixed fields within
310: network boundaries as suggested by Dowdy
311: and colleagues. The VAULT-2 data which we examined are associated with strong
312: concentrations. We conclude that the cool loop model applies only to
313: a small fraction of the VAULT-2 emission, but we cannot discount a
314: significant role for cool loops in quieter regions. We suggest a
315: physical picture for how network L$\alpha$ emission may occur through
316: the cross-field diffusion of neutral atoms from chromospheric into coronal plasma.
317:
318: \end{abstract}
319:
320: \keywords{Sun: atmosphere - Sun: chromosphere - Sun: transition region
321: - Sun: corona - Sun: magnetic fields}
322:
323: \section{Introduction}
324: \label{sec:introduction}
325:
326:
327: In spite of a century or so of research, the solar atmosphere
328: continues to challenge our understanding. As well as the well-known problems
329: of identifying the causes of coronal and chromospheric heating, other
330: phenomena such as spicules, flares, explosive events, and atmospheric
331: dynamics in general, remain only
332: partly understood.
333:
334: A particularly stubborn puzzle relates to the structure of the solar
335: transition region (henceforth ``TR'') - plasma between the chromosphere and corona. The
336: chromosphere and TR are particularly bright over
337: magnetic field concentrations. Outside of sunspots, the concentrations
338: form the supergranular network boundaries (henceforth, ``NB''s). Cell
339: interiors (``CI''s) are less bright. The supergranular network
340: pattern eventually disappears at coronal temperatures $T_e \ga 10^6$ K
341: \citep{Tousey1971,Reeves1976}. The upper TR (where the
342: electron temperature $T_e$ lies between say $2\times 10^5$ K and
343: coronal temperatures of $10^6$ K) is adequately described by magnetic
344: field-aligned thermal conduction down from the corona
345: \citep{Gabriel1976,Jordan1980b}, but that the lower TR
346: ($10^4$ K $< T_e < 2\times 10^5$ K) is not so easily understood
347: \citep[see, for example, the reviews
348: by][]{Mariska1992,Anderson-Huang1998}. Models dominated by heat
349: conduction along magnetic field lines fail to produce enough bright
350: emission from the lower transition by orders of magnitude. In other
351: words, the differential emission measure in the lower transition
352: region is far higher than predicted by such models.
353: Two points argue,
354: however, in favor of conduction as a source of energy for the lower
355: TR: First, the conductive flux is of the right order of
356: magnitude to explain the radiation losses there. Second, the relative
357: intensities of coronal and TR lines vary surprisingly
358: little
359: over the solar surface (i.e., the differential emission
360: measure has a universal shape), suggesting an energetic link.
361:
362: Attempts to fix the problem have either largely
363: failed, or raised questions of an equally troubling nature. For
364: example, models based on heat flow parallel to magnetic field lines
365: were explored which dropped earlier restrictions on magnetic field
366: geometry and a static picture. But such calculations fail to account
367: for the brightness of the lower transition
368: region (\citealp{Gabriel1976}, \citealp{Pneuman+Kopp1978}, \citealp{Athay1981},
369: \citealp{Woods1986}). A variation on such models was
370: developed by
371: \citet{Cally1990}, introducing enhanced
372: heat fluxes using transport by turbulent eddies. Cally found
373: that the mixing length (a free parameter) should scale as
374: $T_e^{-1.5}$ to account for observations. But this is somewhat
375: unsatisfactory because there is no physical reason why the
376: turbulent transport should behave like this.
377: \citet{Ashbourn+Woods2001} presented a more promising
378: model in which the ion-acoustic
379: instability sets in owing to the high electron drift speed associated
380: with heat conduction, in the
381: middle TR. The instability makes the
382: perpendicular ion thermal conductivity large and dominant below 10$^5$
383: K, thereby providing a lower temperature gradient and higher emission
384: measure than fluid approximations give.
385: The downward heat flux was also
386: explored by \nocite{Fontenla+Avrett+Loeser1990,Fontenla+Avrett+Loeser1991,
387: Fontenla+Avrett+Loeser1993,Fontenla+Avrett+Loeser2002}
388: Fontenla {\em et al.} (1990,1991,1993,2002; henceforth ``FAL'') in
389: multi-fluid
390: models where diffusion and bulk flows transport cool material to
391: hot regions, thereby extracting and radiating much of the available
392: energy in the corona.
393:
394: The essential property of all such models is that they produce a
395: geometrically thin lower TR, where emission is confined to a layer just a few tens of
396: km deep. As discussed in a series of papers by Feldman and
397: colleagues \citep[e.g.,][]{Feldman1983}, UV, EUV and X-ray images
398: present serious challenges for such thin models. Such observations
399: have prompted others to set aside conduction, instead focusing on
400: local sources of heat within the TR, such as Joule
401: dissipation (e.g., \citealp{Rabin+Moore1984},
402: \citealp{Roumeliotis1991}) or parameterized forms of heat
403: dissipation \citep[e.g.][]{Antiochos+Noci1986,
404: Patsourakos+Gouttebroze+Vourlidas2007}. Such models are similar to
405: chromospheric models \citep[e.g.][]{Vernazza+Avrett+Loeser1981},
406: in the sense that local heating is balanced by radiation
407: losses\footnote{However, unlike the chromosphere, at transition
408: region temperatures the plasma is almost fully ionized and there is no
409: internal heat sink (latent heat of ionization) which acts, in the
410: chromosphere, as a thermostat. Locally heated and cooled models of
411: the TR are therefore susceptible to thermal instability
412: \citep{Cally+Robb1991}.}. In such models, conduction plays no role,
413: and instead the
414: heights of the structures or ``cool loops''
415: are physically limited by the pressure scale height, the
416: temperature scale heights being larger
417: \citep{Antiochos+Noci1986}. Even in loops near $10^5$K, the height
418: cannot exceed 5Mm for equilibrium to exist, for heating functions
419: which are not of a special form \citep{Cally+Robb1991}. Limb
420: observations of active regions reveal cool material much higher, but
421: such structures are almost always dynamic
422: \citep[e.g.][]{diGiorno+other2003}.
423:
424: \citet{Dowdy+Rabin+Moore1986} pointed out that one would expect
425: magnetic field of mixed polarity {\em within} the supergranular NBs,
426: a feature absent from the model of \citep{Gabriel1976}.
427: Dowdy and colleagues suggested TR emission
428: below $\sim 10^5$ K is dominated by radiation from locally heated loops
429: which connect such opposite polarity photospheric fields, on
430: scales of $\la 10$Mm. Their picture is illustrated by
431: figure~\pref{fig:dowdy86}, taken from their 1986 article.
432:
433: \citet{Sanchezalmeida+others2007} searched for
434: magnetic signatures of footpoints of cool loops, using data from
435: the SUMER \citep{Wilhelm+others1995a}
436: and MDI \citep{Scherrer+others1995}
437: instruments on the SOHO spacecraft, and G band bright point data
438: from the Dutch Open Telescope \citep{Hammerschlag+Bettonvil1998}.
439: Their analysis was inconclusive,
440: because, as we will find, neither SUMER nor MDI (full-disk) data
441: have sufficient spatial resolution ($1\farcs5$ and $4\arcsec$
442: respectively), or magnetic sensitivity.
443:
444: The present paper studies the proposal of Dowdy et al. by examining
445: very high angular resolution ($\approx 0\farcs3$) images of the lower
446: transition from the VAULT instrument
447: \citep{Korendyke+others2001}, and comparing these with
448: magnetic fields and extrapolations from nearly-simultaneous
449: magnetogram data. Since the magnetic data available do not have the
450: sensitivity of recent data, we also examine spectroheliograms of the
451: Sun from the spectropolarimeter
452: (SP, \citealp{Lites+Elmore+Streander2001}) on the Hinode spacecraft
453: \citep{Kosugi+others2007}. We will argue that several observed
454: properties of the VAULT L$\alpha$ emission are not compatible with the cool loop
455: hypothesis.
456:
457:
458: \section{UV data of the cool transition region and corona}
459:
460: We have re-analyzed data described by
461: \citet{Patsourakos+Gouttebroze+Vourlidas2007}, augmented by
462: UV and EUV data from SOHO and TRACE \citep{Handy+others1999}. The VAULT data
463: were obtained during the second instrument flight
464: (VAULT-2) on 14 June 2002 near 18:12 UT. The field of view is
465: $375\arcsec \times 257\arcsec$ centered near coordinates
466: ($-600\arcsec,+260\arcsec$) relative to solar disk center.
467: The instrument captured both active
468: plage areas south and east of the center of the FOV, and
469: the quieter areas to the NW.
470: The pixel size is $0\farcs125$ and resolution $\approx 0\farcs3$.
471: Figure~\pref{fig:context} shows contextual images of the photosphere
472: (MDI instrument) and corona (EIT, \citealp{Delaboudiniere+others1995})
473: from the SOHO spacecraft, together with a VAULT-2
474: image\footnote{The image analyzed is {\tt img04.png} from {\tt
475: http://wwwsolar.nrl.navy.mil/rockets/vault/archives.html}.}.
476: The MDI, EIT 171\AA{} and EIT 304\AA{} images were
477: acquired at 17:59:30, 18:04:37, and 17:45:15 UT respectively.
478: The VAULT-2 image was coaligned
479: with the the EIT He~II 304\AA{} image, crudely by eye. It was found to be
480: centered near $(-612\arcsec,260\arcsec)$, and rotated clockwise by
481: $9^\circ$. Furthermore the pixel sizes were found to be $0\farcs125$
482: and $0\farcs115$ in $x$ and $y$ respectively. (The different pixel
483: scales found are not in disagreement with measurements by the VAULT
484: team,
485: C. Korendyke, private communication 2008).
486: This co-alignment is
487: reliable only to within $\approx \pm 2-3$\arcsec{}, but is sufficiently
488: accurate for our purposes.
489:
490: Focusing on the ``quiet'' NW sector of the images, the brightest
491: L$\alpha$ emission occurs over fairly strong, unipolar magnetic flux
492: concentrations (see section \pref{sec:magvault}). A feature of the
493: data receiving much attention
494: \citep[see][in particular their fig.~2]{Patsourakos+Gouttebroze+Vourlidas2007} is the fine
495: thread-like structure. The threads are most obvious seen against the
496: darker CIs in the neighborhood of the bright concentrations. Similar
497: fine structure has been seen for many years in the related, but
498: notoriously difficult to interpret, H$\alpha$ line
499: \cite[e..g.][]{Kiepenheuer1953,Martin+others1994}. The chromospheric
500: network in L$\alpha$ contains many such threads, combining to form a
501: collective network emission of order 10$\arcsec$ width \citep[][their
502: fig.~2]{Patsourakos+Gouttebroze+Vourlidas2007}. These authors
503: showed that the fine structure is missing from data with lower
504: angular resolution than $\approx1\arcsec$. In their picture, much of
505: the emission from each concentration originates from the threads,
506: which are assumed to be small loops.
507:
508: \citet{Patsourakos+Gouttebroze+Vourlidas2007} also noted that ``The
509: threads are, at a first approximation, radially distributed or
510: slightly bent around the cell centers, which suggests that they could
511: correspond to closed structures, i.e. small loops with temperatures in
512: the temperature formation range of Ly$\alpha$.''
513: However, such patterns are not actually
514: common in the VAULT-2 data analyzed by them. Inspection of Fig.~2 of
515: \citet{Patsourakos+Gouttebroze+Vourlidas2007}, an area of
516: $100\arcsec\times60\arcsec$ containing say 4 or so supergranular
517: cells, shows that the threads are all oriented between $\approx
518: -40^\circ$ and $10^\circ$ of the direction of the $y$ axis (see also
519: figure~\pref{fig:tricolor}). Thus {\em the thread orientations are
520: organized on supergranular scales (i.e. much larger than granules),
521: and larger still near active regions}.
522: The
523: bright L$\alpha$ patches appear like small ``comets'' whose tails are
524: comprised of the threads, the ``heads'' of the comets pointing
525: generally towards the active region in the SE of the image (bottom
526: left of the figure). Such thread orientations are at odds with
527: magnetic loop footpoints which might be expected from more random
528: granular and
529: supergranular forcing of surface magnetic fields. The ``comets'' are
530: reminiscent of patterns seen in H$\alpha$ data, particularly in the
531: neighborhood of filaments, for many years
532: \citep{Martin+others1994,Low1996}.
533:
534: Figure~\pref{fig:tricolor} shows the VAULT-2 data (yellow) together with
535: TRACE 171\AA{} data (four summed images)
536: of the lower corona (blue), obtained close to
537: 17:45 UT. TRACE data were aligned to VAULT-2 data using the EIT co-alignment.
538: While 171\AA{} coronal emission is present almost
539: everywhere, it is brightest only in the vicinity of the plage and
540: filament. White regions in the image shows areas where both features
541: are bright, including some ``moss'' emission (slightly granulated
542: structure) in the neighborhood of the filament, but also including
543: some loops slightly further from the filament region. Use of a
544: different color table shows coronal moss emission throughout the plage,
545: with weak 171 emission even in the ``quiet'' region (upper right
546: corner). 171 \AA{} moss is associated with conductive heating from
547: overlying hotter loops \citep{Fletcher+dePontieu1999, Berger+others1999}.
548: The coronal emission will be related to magnetic field
549: extrapolations below.
550:
551: VAULT images
552: from an earlier flight (VAULT-1) were analyzed by
553: \citet{Korendyke+others2001,Vourlidas+others2001}. These images
554: contain qualitatively similar features to those seen in the data
555: studied here, including moss and threads organized
556: into comet-like patterns in the region within $\approx$ 100\arcsec{} of the
557: filament. The VAULT-1 DATA contain more quiet areas which contain
558: network concentrations hosting significant, but less ordered,
559: thread emission \citep[fig.~5 of][]{Korendyke+others2001}.
560:
561: To summarize,
562: the essential features of the VAULT-2 data under scrutiny are:
563:
564: \begin{itemize}
565: \item{} Much of the emission is organized into thin threads,
566: originating in the network cell boundaries, most
567: visible against the dark cell interiors,
568: \item{} near large-scale coronal structures and organized
569: photospheric magnetic fields (active regions), the threads have
570: characteristic orientations on scales in excess of supergranules,
571: \item{} within plage or large active network boundaries, the
572: emission appears to have a small scale granular structure associated
573: with conductively heated coronal emission called
574: ``moss'' \citep{Vourlidas+others2001}.
575: \end{itemize}
576:
577: \section{Magnetic fields measured at the epoch of the VAULT-2 observations}
578: \label{sec:magvault}
579:
580: MDI full-disk data have a 1$\sigma$ noise level of 17 Mx~cm$^{-2}$ per
581: $1\farcs 984\times 1\farcs984$ pixel, determined from
582: the spatial power spectrum assuming the noise is ``white''.
583: Fortunately, higher quality longitudinal magnetograms from the Kitt
584: Peak Vacuum Telescope synoptic program are available, here we analyze
585: a scan taken between 16:00 and 16:55 UT on June 14 2002. These data
586: have pixels of $1\farcs 148\times 1\farcs148$, each with a noise of
587: $\approx 2.8$ Mx~cm$^{-2}$. Table~\pref{tab:sens} lists the relative
588: sensitivities of these and other magnetograms. The KPVT
589: instrument can detect the ubiquitous ``salt and pepper''
590: weak longitudinal field, with fluxes of $(65\rightarrow100)\times10^{15}$ Mx,
591: discovered many years ago by \citet{Livingston+Harvey1971}. Such
592: small fluxes are invisible to MDI.
593:
594: We aligned the KPVT data with VAULT-2 by eye, using the unsigned
595: magnetic flux, we estimate the accuracy of the alignment is no worse
596: than 3\arcsec, probably better. This apparently large uncertainty
597: arises because although there
598: is a strong correlation between chromospheric and TR emission and
599: absolute magnetic flux, the correlation is only on scales in excess of a few seconds of
600: arc \citep{Skumanich+Smythe+Frazier1975,Sanchezalmeida+others2007}. Variations with time and
601: small scale spatial variations limit the
602: co-alignment accuracy.
603: We rotated both datasets to disk center assuming that L$\alpha$ and
604: the KPNO data arise from the same spherical surface. This is
605: manifestly incorrect, given that L$\alpha$ is formed at least 2Mm
606: above the photosphere and has contributions from spicules extending to
607: 10Mm or so. Thus after the rotation we re-did the co-alignment. Lastly,
608: we assumed that the observed fields are statistically radial, and
609: re-computed the line-of-sight field on the rotated surface. (Our
610: comparisons of field line morphology with L$\alpha$ images
611: are insensitive to the exact choice of radial vs. vertical
612: field, and our calculations are only potential anyway).
613:
614: We computed the potential magnetic field from this surface,
615: assumed to be flat, using the Fourier method\footnote{In all figures
616: except \pref{fig:spqr} and \pref{fig:spar}, a larger field of
617: view was used to compute the potential fields than is plotted, to try
618: to avoid periodic artifacts.}. These
619: calculations crudely indicate the morphology of magnetic fields
620: overlying the surface fields, assuming that sources (i.e. currents) of
621: magnetic field are negligible except those sub-surface currents
622: responsible for the surface fields. Figure \pref{fig:kpvtvault} shows
623: VAULT data, contours of magnetic flux density, and field lines
624: superposed. The field lines were plotted only if both footpoints
625: exceeded twice the noise level of 2.8 Mx~cm$^{-2}$,
626: but the
627: extrapolations themselves include all pixels.
628: In this way we reject
629: more than 95\% of field lines which arise solely from noise.
630: (Figure~\pref{fig:kpvtvaultsm} below includes all field lines
631: regardless of signal-to-noise ratios, to illustrate that rejection of
632: noisy data is not a critical issue).
633: The plotted field lines were
634: evenly sampled to avoid overcrowding here and in later figures. The
635: region observed by VAULT is dominated by flux of positive polarity.
636: Both short and long loops are aligned locally in generally the same
637: direction. The large positive flux region connects to negative
638: polarity fields in the SW corner and outside of the plotted field of
639: view. The overwhelming amount of positive flux guides essentially all
640: field lines away from it, in the potential fields shown.
641:
642: Let us consider the extrapolated field lines as possible
643: plasma loops, and first examine
644: those
645: of length $\le 10$Mm. (Lengths and heights of potential field
646: loops are on average related, but we also
647: discriminate between low and high lying loops in the following
648: figures). Such short loops are
649: candidates for the ``cool loop'' model of L$\alpha$ emission. If the
650: emission were dominated by such structures, we would expect to find a
651: correlation between bright thread emission and the position and
652: density of these field lines. There are indeed places where this
653: is the case, near \coordm{70}{150} and \coordm{140}{110} in the figure,
654: for example, but in most cases the bright emission and short loop
655: densities are poorly correlated. Figure \pref{fig:kpvtvaultsm} shows
656: a close up of flux concentrations\footnote{We refine our definition of
657: ``concentration'' here to refer to an aggregation of flux of one polarity which defines
658: part of a NB. Typically these have an area of
659: $\ga 1$Mm$^2$ and a flux $\ga 10^{18}$ Mx, about three times the
660: detection limit of MDI full disk data (table~\pref{tab:sens}).
661: They
662: are closely correlated with ``the network'' (clumps) of L$\alpha$ emission.}
663: centered near \coords{-500}{+290} in
664: figure \pref{fig:context}, showing many more short loops and their
665: properties compared to the L$\alpha$ threads.
666: In this
667: case we set no limits on the footpoint signal-to-noise ratios in
668: selecting field lines, to see if excluding noisy pixels
669: might introduce an important bias. It does not, it simply reduces the
670: number of field lines with at least one footpoint in the CI regions.
671: Thus, in figure~\pref{fig:kpvtvaultsm}, one finds areas where cool
672: loops are indeed plausible (e.g. between \coords{15,60} and
673: \coords{55,60}), but there are other areas where short loops are found
674: without L$\alpha$ threads (most other locations in the figure).
675: In any case, these short loops tend
676: to connect only the {\em peripheries} of flux concentrations to
677: CI regions, but L$\alpha$ is bright also directly over
678: the flux concentrations themselves.
679:
680: In fact, L$\alpha$ is brightest
681: where loops much {\em longer} than 10Mm originate in the photosphere
682: (lower panel of the figure). This
683: is because there simply is not room for short loops to reach the central,
684: brightest and unipolar regions of the flux concentrations from neighboring
685: CI fields. These longer and taller loops, if
686: they contain much plasma, are likely to have most of their volume
687: filled with hot (i.e. coronal) plasma in order that they have a
688: reasonable lifetime (i.e. close to
689: hydrostatic equilibrium, \citealp[e.g.][]{Rosner+Tucker+Vaiana1978}).
690: {\em Thus, L$\alpha$ emission from the NBs appears to be
691: dominated by processes near the footpoints of {\em long}
692: loops}. The longer loops not only originate in the bright cores of the
693: concentrations, but they also show ``comet''-like structure in broad
694: agreement with the L$\alpha$ thread morphology. These field line shapes
695: arise because surface field lines initially directed towards the SE are forced to avoid
696: the
697: strong same polarity flux concentration there, and so turn towards the
698: NW, attracted by opposite polarities outside of the FOV.
699:
700: These results present difficulties for the ``cool loop'' model,
701: which attempts to explain {\em most} of the network emission.
702: However, care is needed before arriving at stronger conclusions.
703: At the sensitivity of the KPVT
704: data, there are simply too few small loops to account for all threaded
705: L$\alpha$ structures. It may be that, as suggested by
706: figure~\pref{fig:dowdy86}, there are unresolved mixed polarities within
707: the cores of the flux concentrations, as well as in CI regions.
708: To address this problem, we turn to more sensitive
709: measurements of surface magnetic fields, at higher angular resolution,
710: to see if such fields
711: exist on the Sun at a level undetected by these data, and discuss the
712: role of fields unresolved even by Hinode later.
713:
714: \section{Vector polarimetry from Hinode}
715: \label{sec:hinode}
716:
717: \subsection{Longitudinal fields}
718:
719: We examined data for quiet and active regions obtained with the
720: SP on Hinode.
721: Such data are
722: unique, stable, seeing-free measurements of the full Stokes vector,
723: with pixels of only 0.164\arcsec. Image stability is critical for
724: accurate polarimetry since seeing-induced errors can be large
725: \citep{Lites1987}, even for ground-based images captured with adaptive
726: optics \citep{Judge+others2004}. The noise in longitudinal
727: magnetograms from a ``normal map'' (4.8 second acquisition time)
728: acquired with the SP corresponds to just $3$ Mx~cm$^{-2}$ in each
729: pixel \citep{Lites+others2008}. These data are sensitive to very
730: small magnetic flux concentrations, the noise levels are $4\times
731: 10^{14}$ Mx (Table ~\pref{tab:sens}).
732:
733: Active region data from December 29 2006 and January 19 2007 were
734: analyzed with quiet data obtained on November 27
735: 2007. Here we address the quiet (figure~\pref{fig:spqr}) and December
736: 2006
737: (figure~\pref{fig:spar})
738: datasets. Since
739: the SP is a slit instrument, the regions took 135 and 43 minutes to
740: scan respectively.
741: The
742: latter small active region contains a small group of pores
743: and a single pore of opposite positive polarity.
744: These pores have associated with them
745: magnetically disturbed granulation. Figure~\pref{fig:spqrc} shows
746: close-up views of these magnetic
747: concentrations,
748: again plotting groups of
749: representative field lines whose lengths are less than and greater
750: than 10 Mm, and which have flux densities in each pixel greater than
751: $6$ Mx~cm$^{-2}$ (i.e. $2\sigma$ above the noise).
752: When
753: observed at high magnetic sensitivities and spatial resolution,
754: the quiet Sun and active
755: regions have network-like flux concentrations which are surrounded by
756: mixed polarity fields. The SP data show significant signal almost
757: everywhere (there is no obvious ``white noise'' component in the
758: spatial power spectra). The SP thus
759: has not yet reached a limiting small scale of the photospheric
760: magnetic structure.
761:
762: By binning these Hinode
763: data to
764: the pixel size and
765: sensitivity of the KPVT, we find that, as expected, the KPVT
766: misses magnetic flux. But the missing flux amounts to just 25\% of
767: the detectable flux within the
768: supergranular CIs. Thus, although a significant number of short loops connecting NBs to CIs
769: are certainly missing from figures~\pref{fig:kpvtvault} and
770: \pref{fig:kpvtvaultsm}, the missing flux would not be
771: sufficient to account for all observed
772: L$\alpha$ threads as cool loops.
773:
774: NB fields are organized into concentrations of various sizes, so care
775: must be taken in discussing field line lengths and their relation to
776: underlying flux and associated UV emission. It is convenient to
777: discuss two groups- those unipolar concentrations in excess of $5\times
778: 10^{18}$ Mx, and those below.
779:
780: The VAULT-2 and active region SP data fall into the large flux group, the
781: quiet Sun SP data into the small flux group. In the large flux
782: group, the SP data contain no examples of small scale ($\la$
783: $10^{18}$ Mx) mixed polarities within the cores of the network
784: concentrations themselves, at the detection limit of $\sim 0.4 \times
785: 10^{15}$ Mx, as is suggested by figure~\pref{fig:dowdy86}. If the cool
786: loop picture is generally applicable, there is a point where we should
787: observe them. But the SP data show that, at a resolution of
788: 240 km, the typical large concentration is unipolar. Any missing
789: mixed polarity fields on scales below 240 km would
790: yield extremely short loops (lengths $\ell \la 240\pi/2 \approx 365$ km)
791: which would not even be visible in features like L$\alpha$
792: formed higher above the photosphere (see section \pref{subsec:unresolved}).
793: The
794: SP data therefore confirm
795: that short loops from larger concentrations associated with active
796: network (an example might be that at $x=15$ in the lower panel of
797: figure~\pref{fig:spqrc}) generally only extend from the {\em
798: peripheries} of these flux concentrations. The core regions of the
799: concentrations are almost always connected via longer loops between
800: different concentrations and not between the concentration and CI
801: field (note the absence of short loops connected to centers of large
802: concentrations of flux in the lower panel of figure \pref{fig:spqrc}).
803: The brightest UV and EUV emission from the TR sits
804: (statistically) directly over the concentrations, as exemplified by
805: the VAULT-2 data above. The cool loop picture therefore fails to account
806: for the bulk of L$\alpha$ emission over such NBs.
807:
808: The story is different in the small flux group. The SP data contain
809: examples of what are probably mixed polarities {\em within} NBs, as
810: proposed by \citet{Dowdy+Rabin+Moore1986}. For example, if the
811: negative polarity ribbon of condensations along $x=5$ are considered
812: part of the larger positive polarity condensations seen along $x=7.5$
813: in the upper panel of figure \pref{fig:spqrc}, then this situation
814: qualifies as a mixed polarity network structure as suggested by Dowdy
815: et al.. The short loops clustered between and around these
816: concentrations are likely candidates for cool loops, and may explain
817: the rosette like structures seen in the TR in the quiet Sun (see the
818: compilation of images from the SUMER instrument on SOHO
819: \citealp{Feldman+others2003}).
820:
821: %These regions lie near
822: %\coords{80}{43} and at the boundaries of a cell whose center is at
823: %\coords{60}{12}. In these regions, the field lines appear to be bent
824: %away from the regions of strong concentrations, though these are some
825: %10-15Mm distant.
826:
827: \subsection{Vector fields}
828:
829: The SP data consist of the full Stokes vectors of the photospheric Fe {\sc i}
830: 6301.5 and 6302.5 \AA\ lines.
831: We can therefore assess how the potential
832: fields compare to the transverse field properties derived from the Stokes $Q,U$
833: (linear polarization) measurements. This is meaningful only for the
834: areas of active region data with sufficient signal-to-noise
835: ratios.
836: Figure~\pref{fig:azimuths} shows the field azimuths both measured and
837: computed using the MERLIN\footnote{
838: {\tt http://www.hao.ucar.edu/projects/csac/nextgen.php\#merlin}. See
839: \citet{Lites+others2007}.} inversion scheme
840: for the 29 December 2006 dataset. (Field azimuths are the
841: angles measured counter-clockwise from the direction pointing solar west, of the field vector projected on to
842: the local solar surface. The azimuths are subject to the well known
843: 180$^\circ$ ambiguity). There is broad agreement.
844: Significant departures from potential fields exist
845: in both
846: active region datasets, some arising simply from the limits
847: of using Fourier transforms which assume
848: that the domain is periodic in $x$ and $y$. (The ``cusps'' seen near
849: $y=47$ in
850: figure \pref{fig:spar} are such artifacts, and the azimuths close to the
851: boundaries in Figure~\pref{fig:azimuths} also reflect this problem). Others are of solar
852: origin, caused by electrical currents above the surface $z=0$.
853: Examples are the many patches of gray near the center of the lower panel of
854: in Figure~\pref{fig:azimuths}.
855: These differences are of prime interest for the physics of the atmosphere, but
856: here we ask simply how our conclusions concerning the
857: validity of the ``cool loop'' model might be changed.
858:
859: The lack of short loops in the cores of larger network flux
860: concentrations appears to be a robust result- opposite polarities
861: simply do not exist there for any length of time. The directions and
862: connections of field lines will depart from the potential
863: calculations. We do not speculate on such effects, simply noting the
864: broad agreement in the measured and computed azimuths. But it is
865: difficult to see how the ``comet'' structures could remain aligned as
866: they are seen in the VAULT-2 data by {\em small}-scale current
867: systems. Indeed, it is well known that filaments must carry
868: significant electrical currents to provide support against gravity via
869: the Lorentz force \citep[e.g.][]{Low1996}. ``Comet''-like structures
870: seen in upper chromospheric H$\alpha$ are aligned on scales larger
871: than supergranulation along filament channels
872: \citep{Martin+others1994}. The large scale currents associated with
873: filament channels may therefore also be responsible for the observed L$\alpha$
874: ``comet'' alignment. As in the potential field case, it is only if
875: the ``comets'' are associated with large scale ($>$ supergranules)
876: coronal structure that this observation makes physical sense.
877:
878:
879: \section{Discussion and conclusions}
880:
881: On the basis of cool loop models of the VAULT-2 L$\alpha$ emission
882: analyzed here,
883: \citet{Patsourakos+Gouttebroze+Vourlidas2007} concluded that `` The
884: reasonable agreement between the models and the observations indicates
885: that an explanation of the observed fine structure in terms of cool
886: loops is plausible.'' Their motivation for interpretation of the
887: threads in terms of such models is that the thread-like structure,
888: assumed to be aligned with magnetic fields associated with the
889: chromospheric network, is incompatible with heating via field-aligned
890: heat conduction down from overlying coronal plasma. The cool loop
891: model has emerged as a viable explanation for the anomalous brightness
892: of TR emission in features formed below $2\times10^5$
893: K, in spite of some significant physical problems, notably the
894: tendency for instability of such classes of model
895: \citep{Cally+Robb1991}. Nevertheless, the work by
896: \citet{Patsourakos+Gouttebroze+Vourlidas2007} lends support to this
897: picture.
898:
899: In contrast, by studying VAULT-2 data in terms of the magnetic structure of the chromospheric
900: network, we find:
901:
902: \begin{itemize}
903: \item{} The location and orientation of some of the L$\alpha$ threads
904: are only rarely compatible with the idea of cool, short, loops
905: originating from the NBs and extending into
906: neighboring CI regions, but
907: \item{} such short loops usually connect the CI to the
908: {\em peripheries} of network flux concentrations with unipolar
909: fluxes in excess of, say $5\times 10^{18}$ Mx.
910: \item{} The bulk of the L$\alpha$ NB emission, arising from the cores of
911: such concentrations, seems to be associated with far longer magnetic
912: loops
913: which connect to other concentrations. In this way these
914: concentrations seem analogous to plages.
915: \item{} In the ``quiet'' region of the VAULT-2 data, the
916: longer loops diverge non-radially from their NB concentrations in a
917: manner reminiscent of the ``comet'' L$\alpha$ patterns.
918: \item{}Hinode SP observations at $0\farcs33$ resolution reveal
919: that short loops can exist in what appear to be NBs,
920: provided the concentrations making up the magnetic concentrations
921: have small enough areas (fluxes $\la 5\times10^{18}$ Mx).
922: Thus, cool loops emitting L$\alpha$ may indeed
923: be present
924: in quiet regions, where they may cluster around small concentrations
925: of flux, and have the appearance of clumped rosettes of emission of
926: $\la 10$Mm diameter (see the upper panel of Figure~\pref{fig:spqrc}).
927: \item{} Hinode data in the neighborhood of active
928: regions, while showing mixed polarities, tend to reveal short loops
929: only connecting peripheries of the larger flux concentrations
930: found there to the CI regions (see the lower panel of
931: Figure~\pref{fig:spqrc}). This is also the case found for
932: the particular VAULT-2 data analyzed
933: here. The bulk of bright network
934: L$\alpha$ emission is difficult to accounted for by such structures.
935: \item{} Non-potential fields are clearly present in the Hinode data,
936: but we have argued that such fields, on the small scales associated
937: with the cool loop model, do not affect our overall conclusions.
938: \end{itemize}
939:
940: Our work suggests that the cool loop picture cannot be universally
941: valid. Below we propose a different qualitative picture of the emission from
942: the cores of these flux concentrations.
943:
944: \subsection{The ``comet'' patterns seen in VAULT data}
945:
946: The aligned ``comet-tail'' patterns of the L$\alpha$
947: threads initially presented us with a puzzle. Why would such large
948: scale order be characteristic of a small scale process involving the
949: formation of physically far smaller loops, whose footpoints might be
950: controlled by random convective processes?
951:
952: The explanation may be that the region termed ``quiet'' by Patsourakos and
953: colleagues is not really ``quiet''. There is close by a large
954: net flux which imposes a large scale order on potential fields. Two other observations
955: indicate organization on a large scale: the threads largely point away
956: from the coronal Fe IX/X emitting loops associated with the active
957: region in the SE corner of the field of view; the ``quiet'' region in
958: fact lies between two filament channels- the SE channel and another
959: lying along the western edge of the VAULT FOV (see
960: fig.~\pref{fig:context}).
961: Free magnetic energy associated with the magnetostatic balance of
962: filaments, in the form of atmospheric current systems, is surely present
963: which might modify the overlying magnetic field from the potential
964: state we have calculated. Such currents would not be incompatible
965: with the large scale order implied by the comets.
966: It is only if one considers the structures to be formed from
967: convection-driven short loops that large scale organization would be
968: surprising.
969: Our analysis instead points to an association of long loops with most of
970: the L$\alpha$ emission, including the threads modeled previously as
971: cool loops. Instead of cool loops, the emission seems to
972: arise, perhaps as chromospheric
973: material is launched as spicules and heated along much longer field lines.
974: The difficulty in this picture is to explain why L$\alpha$ emission is
975: bright and extended over several Mm lengths, given the obvious failure of
976: field-aligned heat conduction to achieve this.
977:
978: \subsection{The role of unresolved magnetic fields}
979: \label{subsec:unresolved}
980:
981: We noted above that the SP data show signal on all scales down to the
982: Lunqvist limit (2 pixels $=0\farcs328$). The smallest scales of solar
983: photospheric magnetic fields are as yet unknown. It is therefore
984: likely that tiny loop structures are missing from our analysis.
985: However, the existence of smaller scale structures does not weaken our
986: conclusions, for several reasons.
987:
988: Firstly, to the extent that the magnetic fields are potential,
989: structures of horizontal scale $\ell$ in the photospheric normal magnetic field
990: will extend only to heights $\approx \ell$ higher into the
991: atmosphere. This result is a general property of solutions of
992: Laplace's equation (see, e.g. \citealp{Gary1989}). Now L$\alpha$
993: radiation cannot emerge from heights less than the height where the
994: continuum optical depth is unity, which occurs near 0.8Mm because of
995: opacity and atmospheric stratification (FAL). In FAL's models, the
996: bulk of the emission arises
997: at least 2Mm above the photosphere.
998: Thus, for any small-scale bipoles observable at the
999: solar surface, {\em only those with footpoints separated by 0.8Mm or
1000: so can
1001: contribute to observable L$\alpha$ emission}. Those
1002: separated by smaller scale lengths likely return to the
1003: photosphere before reaching 1 Mm heights. The thermal signatures of
1004: such loops would be primarily visible in lines and continua less opaque than
1005: L$\alpha$, influencing L$\alpha$ itself only marginally.
1006:
1007: Other reasons arise from the nature of the observed L$\alpha$ threads.
1008: They extend over 5-10Mm lengths, so that if formed within tubes of
1009: magnetic flux, footpoint separations are at least this large. Threads
1010: of lengths between 2 and 5 Mm would extend high enough to emit
1011: L$\alpha$ but are not seen in the data. The thread orientations are
1012: collectively organized over scales of several supergranules, and the
1013: threads, if both footpoints are anchored in the photosphere,
1014: require opposite polarity flux surrounding the flux
1015: concentrations associated with network boundaries. None of these
1016: observations are consistent with random ``salt and pepper''
1017: distributions of flux associated with supergranules on any scale below
1018: say 5 Mm in length.
1019:
1020:
1021: For all these reasons we believe that unobservably
1022: small scale magnetic fields are irrelevant to the problem of
1023: understanding the essential properties of the observed L$\alpha$
1024: line, both in plage and other regions.
1025:
1026:
1027: \subsection{Speculation on the origin of the bulk of NB emission}
1028:
1029: Field-aligned diffusive and flowing models of the type computed by FAL
1030: may indeed account for the moss L$\alpha$ emission, but as noted in
1031: the introduction, they are incompatible with L$\alpha$ thread
1032: emission. In the absence of cool loops as a viable proposal of
1033: L$\alpha$ emission except in quiet regions, we can speculate on what
1034: might be the cause of the L$\alpha$ properties seen by VAULT.
1035:
1036: Figure \pref{fig:tricolor} reveals that the corona overlying the
1037: observed region is bright on the SE side, dim elsewhere.
1038: The L$\alpha$ emission from network patches
1039: seems to care little of the intensity of overlying coronal emission.
1040: If energy for L$\alpha$ radiation arises
1041: directly from coronal plasma, surely there must be some
1042: correlation between L$\alpha$ and coronal brightness?
1043:
1044: Independent evidence suggests that the observed L$\alpha$
1045: structure is correlated with observations of H$\alpha$, whose
1046: morphology is complicated, but whose properties on fine scales are
1047: known to relate to conditions in the overlying corona
1048: \citep{Berger+others1999}.
1049: \citet{Berger+others1999} found that, on scales of arcseconds and less,
1050: coronal ``moss''
1051: emission at 171 \AA{} is dark where H$\alpha$ wings are strong,
1052: suggesting that the EUV corona and cooler H$\alpha$ plasma are
1053: separated by a thermal interface which lies parallel to magnetic field
1054: lines
1055: (it is not therefore the ``classical'' thin TR, but
1056: more of a sheath.) We can speculate that this
1057: interface might by a place where energy from the coronal plasma can be
1058: transferred to hydrogen atoms, by diffusion of neutral atoms into the
1059: hot coronal regions \citep{Pietarila+Judge2004}, via cross-field
1060: conduction which occurs because of proton dynamics \citep{Athay1990},
1061: or because of some as yet undetermined (Rayleigh-Taylor like?)
1062: instability \citep{Gabriel1976}.
1063:
1064: Given this interface, consider the neutral atom diffusion scenario
1065: (Judge 2008, submitted).
1066: Neutral atoms, at an interface with hot corona, experience no Lorentz
1067: force until ionized. The probability that a hydrogen atom is ionized
1068: by collisions with hot coronal particles (electrons) is related to the
1069: probability for excitation to the $n=2$ level, which is almost immediately
1070: ($10^{-8}$ s)
1071: followed by the emission of a L$\alpha$ photon. Roughly one L$\alpha$
1072: photon is emitted before it is ionized, independent of the density of
1073: the coronal plasma, provided it is ionized somewhere in the coronal
1074: plasma. It turns out that the L$\alpha$ intensities expected from
1075: this process are proportional to the thermal energy density of the
1076: corona and the neutral diffusion speed. Thus, there is a different
1077: dependence of coronal and L$\alpha$ emission on thermal parameters, so
1078: it is possible in principle to explain why similar L$\alpha$
1079: intensities may arise from regions with different coronal
1080: intensities.
1081: The essential difference between this and the cool loop
1082: picture is that coronal thermal energy is drained by diffusion of
1083: neutrals across magnetic field lines to generate much of the emission
1084: seen in L$\alpha$ and
1085: other typical ``lower TR transitions''. Some chromospheric process
1086: is assumed to launch spicules to get the process started. Such a
1087: model can explain the puzzle of the ``comet'' asymmetry noted in
1088: the present paper, in that cool plasma threads along long coronal loops
1089: already have the large-scale organization required by observations of the L$\alpha$
1090: threads. If it proves feasible, this process bypasses thermal
1091: stability problems presented by cool-loop models, and it is appealing
1092: in that the downward directed conductive flux density of $\approx
1093: 10^6$ \flxu{}, unaccounted for in cool loop models, is radiated by
1094: cool, strong TR lines. Further work on this
1095: scenario is in progress (Judge, 2008, submitted).
1096:
1097: We thank
1098: Scott McIntosh and the referee for helpful comments on this paper.
1099: Hinode is a Japanese mission developed and launched by ISAS/JAXA,
1100: collaborating with NAOJ as a domestic partner, NASA and STFC (UK) as
1101: international partners. Scientific operation of the Hinode mission is
1102: conducted by the Hinode science team organized at ISAS/JAXA. This team
1103: mainly consists of scientists from institutes in the partner
1104: countries. Support for the post-launch operation is provided by JAXA
1105: and NAOJ (Japan), STFC (U.K.), NASA, ESA, and NSC (Norway).
1106:
1107: \def\aspcs{{ASP Conf.\ Ser.}}
1108:
1109:
1110: \begin{thebibliography}{}
1111:
1112: \bibitem[\protect\astroncite{Allen}{1973}]{Allen1973}
1113: Allen, C.~W.: 1973,
1114: \newblock {\em Astrophysical Quantities\/},
1115: \newblock Athlone Press, Univ.\ London
1116:
1117: \bibitem[\protect\astroncite{Anderson\-{-}Huang}{1998}]{Anderson-Huang1998}
1118: Anderson\-{-}Huang, L.~S.: 1998,
1119: \newblock {\em Space Sci.\ Rev.\/} {\bf 85}, 203
1120:
1121: \bibitem[\protect\astroncite{{Antiochos} and {Noci}}{1986}]{Antiochos+Noci1986}
1122: {Antiochos}, S.~K. and {Noci}, G.: 1986,
1123: \newblock {\em Astrophys.\ J.\/} {\bf 301}, 440
1124:
1125: \bibitem[\protect\astroncite{Ashbourn and Woods}{2001}]{Ashbourn+Woods2001}
1126: Ashbourn, J. M.~A. and Woods, L.~C.: 2001,
1127: \newblock {\em Proc. R. Soc. A.\/} {\bf 457(2012)}, 1873
1128:
1129: \bibitem[\protect\astroncite{Athay}{1990}]{Athay1990}
1130: Athay, R.: 1990,
1131: \newblock {\em Astrophys.\ J.\/} {\bf 362}, 364
1132:
1133: \bibitem[\protect\astroncite{Athay}{1981}]{Athay1981}
1134: Athay, R.~G.: 1981,
1135: \newblock {\em Astrophys.\ J.\/} {\bf 249}, 340
1136:
1137: \bibitem[\protect\astroncite{{Berger} {\em et~al.}}{1999}]{Berger+others1999}
1138: {Berger}, T.~E., {De Pontieu}, B., {Schrijver}, C.~J., and {Title}, A.~M.:
1139: 1999,
1140: \newblock {\em Astrophys.\ J.\ Lett.\/} {\bf 519}, L97
1141:
1142: \bibitem[\protect\astroncite{Cally}{1990}]{Cally1990}
1143: Cally, P.~S.: 1990,
1144: \newblock {\em Astrophys.\ J.\/} {\bf 355}, 693
1145:
1146: \bibitem[\protect\astroncite{Cally and Robb}{1991}]{Cally+Robb1991}
1147: Cally, P.~S. and Robb, T.~D.: 1991,
1148: \newblock {\em Astrophys.\ J.\/} {\bf 372}, 329
1149:
1150: \bibitem[\protect\astroncite{{Delaboudiniere} {\em
1151: et~al.}}{1995}]{Delaboudiniere+others1995}
1152: {Delaboudiniere}, J.-P., and 27 co-authors: 1995,
1153: \newblock {\em Solar Phys.\/} {\bf 162}, 291
1154:
1155: \bibitem[\protect\astroncite{{Di Giorgio} {\em
1156: et~al.}}{2003}]{diGiorno+other2003}
1157: {Di Giorgio}, S., {Reale}, F., and {Peres}, G.: 2003,
1158: \newblock {\em Astron.\ Astrophys.\/} {\bf 406}, 323
1159:
1160: \bibitem[\protect\astroncite{{Dowdy} {\em
1161: et~al.}}{1986}]{Dowdy+Rabin+Moore1986}
1162: {Dowdy}, J.~F., J., {Rabin}, D., and {Moore}, R.~L.: 1986,
1163: \newblock {\em Solar Phys.\/} {\bf 105}, 35
1164:
1165: \bibitem[\protect\astroncite{Feldman}{1983}]{Feldman1983}
1166: Feldman, U.: 1983,
1167: \newblock {\em Astrophys.\ J.\/} {\bf 275}, 367
1168:
1169: \bibitem[\protect\astroncite{{Feldman} {\em et~al.}}{2003}]{Feldman+others2003}
1170: {Feldman}, U., {Dammasch}, I.~E., {Wilhelm}, K., {Lemaire}, P., {Hassler},
1171: D.~M., and {Battrick}, B.: 2003,
1172: \newblock {\em {Images of the solar upper atmosphere from SUMER on SOHO}\/},
1173: \newblock ESA SP-1274.~Noordwijk, Netherlands: ESA Publications Division
1174:
1175: \bibitem[\protect\astroncite{{Fletcher} and {de
1176: Pontieu}}{1999}]{Fletcher+dePontieu1999}
1177: {Fletcher}, L. and {de Pontieu}, B.: 1999,
1178: \newblock {\em Astrophys.\ J.\ Lett.\/} {\bf 520}, L135
1179:
1180: \bibitem[\protect\astroncite{Fontenla {\em
1181: et~al.}}{1990}]{Fontenla+Avrett+Loeser1990}
1182: Fontenla, J.~M., Avrett, E.~H., and Loeser, R.: 1990,
1183: \newblock {\em Astrophys.\ J.\/} {\bf 355}, 700
1184:
1185: \bibitem[\protect\astroncite{Fontenla {\em
1186: et~al.}}{1991}]{Fontenla+Avrett+Loeser1991}
1187: Fontenla, J.~M., Avrett, E.~H., and Loeser, R.: 1991,
1188: \newblock {\em Astrophys.\ J.\/} {\bf 377}, 712
1189:
1190: \bibitem[\protect\astroncite{Fontenla {\em
1191: et~al.}}{1993}]{Fontenla+Avrett+Loeser1993}
1192: Fontenla, J.~M., Avrett, E.~H., and Loeser, R.: 1993,
1193: \newblock {\em Astrophys.\ J.\/} {\bf 406}, 319
1194:
1195: \bibitem[\protect\astroncite{{Fontenla} {\em
1196: et~al.}}{2002}]{Fontenla+Avrett+Loeser2002}
1197: {Fontenla}, J.~M., {Avrett}, E.~H., and {Loeser}, R.: 2002,
1198: \newblock {\em Astrophys.\ J.\/} {\bf 572}, 636
1199:
1200: \bibitem[\protect\astroncite{Gabriel}{1976}]{Gabriel1976}
1201: Gabriel, A.: 1976,
1202: \newblock {\em Phil Trans. Royal Soc. Lond.\/} {\bf 281}, 339
1203:
1204: \bibitem[\protect\astroncite{{Gary}}{1989}]{Gary1989}
1205: {Gary}, G.~A.: 1989,
1206: \newblock {\em Astrophys.\ J.\ Suppl.\ Ser.\/} {\bf 69}, 323
1207:
1208: \bibitem[\protect\astroncite{Hammerschlag and
1209: Bettonvil}{1998}]{Hammerschlag+Bettonvil1998}
1210: Hammerschlag, R.~H. and Bettonvil, F. C.~M.: 1998,
1211: \newblock {\em New Astronomy Reviews\/} {\bf 42}, 485
1212:
1213: \bibitem[\protect\astroncite{{Handy} {\em et~al.}}{1999}]{Handy+others1999}
1214: {Handy}, B.~N., and 47 co-authors: 1999,
1215: \newblock {\em Solar Phys.\/} {\bf 187}, 229
1216:
1217: \bibitem[\protect\astroncite{Jordan}{1980}]{Jordan1980b}
1218: Jordan, C.: 1980,
1219: \newblock {\em Astron.\ Astrophys.\/} {\bf 86}, 355
1220:
1221: \bibitem[\protect\astroncite{Judge {\em et~al.}}{2004}]{Judge+others2004}
1222: Judge, P.~G., Elmore, D.~F., Lites, B.~W., Keller, C.~U., and Rimmele, T.:
1223: 2004,
1224: \newblock {\em Applied Optics: optical technology and medical optics\/}
1225: submitted
1226:
1227: \bibitem[\protect\astroncite{Kiepenheuer}{1953}]{Kiepenheuer1953}
1228: Kiepenheuer, K.~O.: 1953,
1229: \newblock in G.~P. Kuiper (Ed.), {\em The Sun\/}, Chicago University Press,
1230: Chicago, p.~322
1231:
1232: \bibitem[\protect\astroncite{{Korendyke} {\em
1233: et~al.}}{2001}]{Korendyke+others2001}
1234: {Korendyke}, C.~M., {Vourlidas}, A., {Cook}, J.~W., {Dere}, K.~P., {Howard},
1235: R.~A., {Morrill}, J.~S., {Moses}, J.~D., {Moulton}, N.~E., and {Socker},
1236: D.~G.: 2001,
1237: \newblock {\em Solar Phys.\/} {\bf 200}, 63
1238:
1239: \bibitem[\protect\astroncite{{Kosugi} {\em et~al.}}{2007}]{Kosugi+others2007}
1240: {Kosugi}, T., and 24 co-authors: 2007,
1241: \newblock {\em Solar Phys.\/} {\bf 243}, 3
1242:
1243: \bibitem[\protect\astroncite{{Lites} {\em et~al.}}{2007}]{Lites+others2007}
1244: {Lites}, B., {Casini}, R., {Garcia}, J., and {Socas-Navarro}, H.: 2007,
1245: \newblock {\em Memorie della Societa Astronomica Italiana\/} {\bf 78}, 148
1246:
1247: \bibitem[\protect\astroncite{Lites}{1987}]{Lites1987}
1248: Lites, B.~W.: 1987,
1249: \newblock {\em Applied Optics\/} {\bf 26}, 3838
1250:
1251: \bibitem[\protect\astroncite{{Lites} {\em
1252: et~al.}}{2001}]{Lites+Elmore+Streander2001}
1253: {Lites}, B.~W., {Elmore}, D.~F., and {Streander}, K.~V.: 2001,
1254: \newblock in M. {Sigwarth} (Ed.), {\em Advanced Solar Polarimetry -- Theory,
1255: Observation, and Instrumentation\/}, Vol. 236 of {\em Astronomical Society of
1256: the Pacific Conference Series\/}, ~33
1257:
1258: \bibitem[\protect\astroncite{Lites {\em et~al.}}{2008}]{Lites+others2008}
1259: Lites, B.~W., Kubo, M., Socas-Navarro, H., Berger, T., Frank, Z., Shine, R.,
1260: Tarbell, T., Title, A., Ichmoto, K., Katsukawa, Y., Tsuneta, S., Sumematsu,
1261: Y., Shimizu, T., and Nagata, S.: 2008,
1262: \newblock {\em Astrophys.\ J.\/} {\bf 672}, 1237
1263:
1264: \bibitem[\protect\astroncite{Livingston and
1265: Harvey}{1971}]{Livingston+Harvey1971}
1266: Livingston, W. and Harvey, J.: 1971,
1267: \newblock in R. Howard (Ed.), {\em Solar Magnetic Fields\/}, IAU Symposium 43,
1268: Reidel, Dordrecht, p.~51
1269:
1270: \bibitem[\protect\astroncite{{Low}}{1996}]{Low1996}
1271: {Low}, B.~C.: 1996,
1272: \newblock {\em Solar Phys.\/} {\bf 167}, 217
1273:
1274: \bibitem[\protect\astroncite{Mariska}{1992}]{Mariska1992}
1275: Mariska, J.~T.: 1992,
1276: \newblock {\em The Solar Transition Region\/},
1277: \newblock Cambridge Univ.\ Press, Cambridge UK
1278:
1279: \bibitem[\protect\astroncite{{Martin} {\em et~al.}}{1994}]{Martin+others1994}
1280: {Martin}, S.~F., {Bilimoria}, R., and {Tracadas}, P.~W.: 1994,
1281: \newblock in R.~J. {Rutten} and C.~J. {Schrijver} (Eds.), {\em Solar Surface
1282: Magnetism\/}, 303
1283:
1284: \bibitem[\protect\astroncite{{Patsourakos} {\em
1285: et~al.}}{2007}]{Patsourakos+Gouttebroze+Vourlidas2007}
1286: {Patsourakos}, S., {Gouttebroze}, P., and {Vourlidas}, A.: 2007,
1287: \newblock {\em Astrophys.\ J.\/} {\bf 664}, 1214
1288:
1289: \bibitem[\protect\astroncite{Pietarila and Judge}{2004}]{Pietarila+Judge2004}
1290: Pietarila, A. and Judge, P.~G.: 2004,
1291: \newblock {\em Astrophys.\ J.\/} {\bf 606}, 1239
1292:
1293: \bibitem[\protect\astroncite{Pneuman and Kopp}{1978}]{Pneuman+Kopp1978}
1294: Pneuman, G. and Kopp, R.: 1978,
1295: \newblock {\em Sol. Phys.\/} {\bf 57}, 49
1296:
1297: \bibitem[\protect\astroncite{Rabin and Moore}{1984}]{Rabin+Moore1984}
1298: Rabin, D. and Moore, R.: 1984,
1299: \newblock {\em Astrophys.\ J.\/} {\bf 285}, 359
1300:
1301: \bibitem[\protect\astroncite{Reeves}{1976}]{Reeves1976}
1302: Reeves, E.~M.: 1976,
1303: \newblock {\em Solar Phys.\/} {\bf 46}, 53
1304:
1305: \bibitem[\protect\astroncite{Rosner {\em
1306: et~al.}}{1978}]{Rosner+Tucker+Vaiana1978}
1307: Rosner, R., Tucker, W.~H., and Vaiana, G.~S.: 1978,
1308: \newblock {\em Astrophys.\ J.\/} {\bf 220}, 643
1309:
1310: \bibitem[\protect\astroncite{{Roumeliotis}}{1991}]{Roumeliotis1991}
1311: {Roumeliotis}, G.: 1991,
1312: \newblock {\em \apj\/} {\bf 379}, 392
1313:
1314: \bibitem[\protect\astroncite{{S{\'a}nchez Almeida} {\em
1315: et~al.}}{2007}]{Sanchezalmeida+others2007}
1316: {S{\'a}nchez Almeida}, J., {Teriaca}, L., {S{\"u}tterlin}, P., {Spadaro}, D.,
1317: {Sch{\"u}hle}, U., and {Rutten}, R.~J.: 2007,
1318: \newblock {\em Astron.\ Astrophys.\/} {\bf 475}, 1101
1319:
1320: \bibitem[\protect\astroncite{{Scherrer} {\em
1321: et~al.}}{1995}]{Scherrer+others1995}
1322: {Scherrer}, P.~H., {Bogart}, R.~S., {Bush}, R.~I., {Hoeksema}, J.~T.,
1323: {Kosovichev}, A.~G., {Schou}, J., {Rosenberg}, W., {Springer}, L., {Tarbell},
1324: T.~D., {Title}, A., {Wolfson}, C.~J., {Zayer}, I., and {MDI Engineering
1325: Team}: 1995,
1326: \newblock {\em Solar Phys.\/} {\bf 162}, 129
1327:
1328: \bibitem[\protect\astroncite{Skumanich {\em
1329: et~al.}}{1975}]{Skumanich+Smythe+Frazier1975}
1330: Skumanich, A., Smythe, C., and Frazier, E.~N.: 1975,
1331: \newblock {\em Astrophys.\ J.\/} {\bf 200}, 747
1332:
1333: \bibitem[\protect\astroncite{{Tousey}}{1971}]{Tousey1971}
1334: {Tousey}, R.: 1971,
1335: \newblock {\em Royal Society of London Philosophical Transactions Series A\/}
1336: {\bf 270}, 59
1337:
1338: \bibitem[\protect\astroncite{Vernazza {\em
1339: et~al.}}{1981}]{Vernazza+Avrett+Loeser1981}
1340: Vernazza, J., Avrett, E., and Loeser, R.: 1981,
1341: \newblock {\em Astrophys.\ J.\ Suppl.\ Ser.\/} {\bf 45}, 635
1342:
1343: \bibitem[\protect\astroncite{{Vourlidas} {\em
1344: et~al.}}{2001}]{Vourlidas+others2001}
1345: {Vourlidas}, A., {Klimchuk}, J.~A., {Korendyke}, C.~M., {Tarbell}, T.~D., and
1346: {Handy}, B.~N.: 2001,
1347: \newblock {\em Astrophys.\ J.\/} {\bf 563}, 374
1348:
1349: \bibitem[\protect\astroncite{{Wilhelm} {\em
1350: et~al.}}{1995}]{Wilhelm+others1995a}
1351: {Wilhelm}, K., and 17 co-authors: 1995,
1352: \newblock {\em proc. SPIE\/} {\bf 2517}, 2
1353:
1354: \bibitem[\protect\astroncite{Woods}{1986}]{Woods1986}
1355: Woods, D.~T.: 1986,
1356: \newblock {\em NCAR/CT\/} 105
1357:
1358: \end{thebibliography}
1359:
1360: \clearpage
1361: \tabone
1362: \clearpage
1363: \figone
1364: \figtwo
1365: %\figthree
1366: \figfour
1367: \figfive
1368: \figfivec
1369: \figsix
1370: \figseven
1371: \figeight
1372: \fignine
1373: \end{document}
1374:
1375:
1376:
1377: