1: %%
2: %aastex manuscript preparation.
3:
4: \documentclass[12pt,preprint]{aastex}
5:
6: %% manuscript produces a one-column, double-spaced document:
7:
8: %\documentclass[manuscript]{aastex}
9:
10: %% preprint2 produces a double-column, single-spaced document:
11:
12: %\documentclass[preprint2]{emulateapj}
13: %\documentclass[preprint2]{aastex}
14:
15: %% If you want to create your own macros, you can do so
16: %% using \newcommand. Your macros should appear before
17: %% the \begin{document} command.
18:
19: \newcommand{\vdag}{(v)^\dagger}
20: \newcommand{\myemail}{skywalker@galaxy.far.far.away}
21:
22: %% You can insert a short comment on the title page using the command below.
23:
24: \shorttitle{On the Weakening of Chromospheric Magnetic Field}
25: \shortauthors{Nagaraju, Sankarasubrmanian \& Rangarajan}
26:
27: %% This is the end of the preamble. Indicate the beginning of the
28: %% paper itself with \begin{document}.
29:
30: \begin{document}
31:
32: \title{ON THE WEAKENING OF CHROMOSPHERIC MAGNETIC FIELD IN ACTIVE REGIONS}
33:
34: \author{K. Nagaraju}
35: \affil{Indian Institute of Astrophysics, Koramangala, Bangalore-560034, India; nagaraj@iiap.res.in}
36:
37: \author{K. Sankarasubramanian}
38: \affil{ISRO Satellite Center, Airport road, Monera, Bangalore-560017, India;
39: sankark@isac.gov.in}
40:
41: \and
42:
43: \author{K. E. Rangarajan}
44: \affil{Indian Institute of Astrophysics, Koramangala, Bangalore-560034, India;
45: rangaraj@iiap.res.in}
46:
47: \begin{abstract}
48: Simultaneous measurement of line-of-sight (LOS) magnetic and
49: velocity fields at the photosphere and chromosphere are presented.
50: Fe I line at $\lambda6569$ and $H_{\alpha}$ at $\lambda6563$
51: are used respectively for
52: deriving the physical parameters at photospheric and chromospheric heights.
53: The LOS magnetic field obtained through the center-of-gravity
54: method show a linear relation between photospheric and chromospheric field
55: for field strengths less than $700$ G.
56: But in strong field regions, the LOS magnetic field values derived
57: from $H_{\alpha}$ are much weaker than what one gets from the linear relationship
58: and also from those expected from the extrapolation of the photospheric
59: magnetic field.
60: We discuss in detail the properties of magnetic field observed in $H_{\alpha}$
61: from the point of view of observed velocity gradients.
62: The bisector analysis of $H_{\alpha}$ Stokes $I$ profiles show larger velocity
63: gradients in those places where strong photospheric magnetic fields
64: are observed. These observations may support the view that the stronger fields
65: diverge faster with height compared to weaker fields.
66:
67: \end{abstract}
68:
69: \keywords{Sun: magnetic fields - Sun: photosphere - Sun: chromosphere - sunspots}
70:
71: \section{INTRODUCTION}
72:
73: It is inferred through various observations that the magnetic field is
74: playing a central role in the solar energetics that take place in the
75: higher layers of the atmosphere (see for eg. \citet{regnier06}).
76: But, it is somewhat difficult to
77: obtain a reliable
78: vector magnetic field measurements at the chromosphere and corona
79: \citep{hector05, judge07}.
80: %CORRECTION: FOLLOWING LINES ARE MOVED FROM BOTTOM
81: The measurements of the photospheric magnetic fields are
82: relatively well established.
83: However, simultaneous vector magnetic field observations at the
84: photospheric and chromospheric heights will give a better handle on the
85: understanding of the magnetic structuring of the solar atmosphere.
86: More direct and reliable measurements of the magnetic field at the chromosphere
87: will also serve as a better boundary condition for extrapolating the magnetic
88: fields to the coronal heights.
89:
90: Measurements of the vector magnetic field in $H_{\alpha}$ is particularly
91: important in understanding the connection between photospheric, chromospheric
92: and coronal magnetic field as its height of formation ranges almost
93: from the upper photosphere to upper chromosphere \citep{vernaza81}.
94: Also, this is one of the most widely used spectral lines for the
95: study of solar chromosphere (see a recent review by \citet{rutten07}).
96: %CORRECTION: BOTTOM LINES ARE SHIFTED UP..
97: %The measurements of the photospheric magnetic fields are
98: %relatively well established.
99: %Hence, the simultaneous vector magnetic field observations at the
100: %photospheric and chromospheric heights will give a better handle on the
101: %understanding of the magnetic structuring of the solar atmosphere.
102: %More direct and reliable measurements of the magnetic field at the chromosphere
103: %will also serve as a better boundary condition for extrapolating the
104: %magnetic fields to the coronal heights.
105:
106: Comparison of the active region magnetic fields measured in $H_{\alpha}$
107: and in the photosphere show a one-to-one correspondence in the weak field regions.
108: %CORRECTION: THE REFRENCE MOVED TO THE FOLLOWING SENTENCE
109: %\citep{bala04, hanaoka05}.
110: But in the strong field regions, like umbra, there is a
111: considerable deviation from the linearity \citep{bala04, hanaoka05}.
112: Infact the LOS magnetic field measured in $H_{\alpha}$ weakens much faster
113: for the corresponding strong field regions of the photosphere.
114: At this point it may not be appropriate to demarcate the field strength
115: above which this deviation happens because different observations show
116: different deviation points. \citet{gosain03} have also reported the systematic
117: weakening of the magnetic
118: field derived from Mg I $\lambda5173/5184$ lines in comparison with that of
119: Fe I $\lambda6301.5/6302.5$ lines. In their case, the magnetic field
120: measured by Mg I lines agree with the potential field extrapolation of the photospheric LOS magnetic field in
121: weak field regime. In strong field regions there is a systematic shift
122: towards lower values but still linear.
123: Simultaneous observations of He I at $\lambda10830$ and Si I lines at $\lambda10827.1$ by \citet{choudhary02} suggest that
124: the field diverges faster in the upper layers of the chromosphere.
125:
126: However there is no conclusive explanation for the observed
127: weaker chromospheric magnetic fields for the corresponding
128: strong photospheric fields.
129: Various possibilities have been discussed by \citet{bala04} \&
130: \citet{hanaoka05} about weaker chromospheric fields observed in umbra.
131: \citet{hanaoka05} have discussed the possibility of scattered light and/or peculiarity of the
132: atmosphere (radiative transfer effects) for the decrease in
133: the polarization signal which in turn cause the underestimation of the
134: magnetic field.
135: \citet{bala04} have suggested that the strongest fields measured at the photosphere
136: diverge spatially and more quickly than the weak fields when propagating upward.
137: In this paper, we address these issues from the point of view of the observed
138: velocity gradients.
139:
140: It is well known that, velocities and magnetic fields are coupled to each other
141: in the solar atmosphere (see for eg. \citet{rajaguru06}).
142: Any change in the magnetic field is expected to alter the plasma motion and hence
143: the observed velocity.
144: If the field strength decreases with height then the velocity is expected to
145: increase and vice versa, for the simple reason that the inhibition of the
146: plasma depends on the magnetic field strength \citep{spruit79}.
147: It is also well known that the magnetic field strength in the umbral region at the
148: photosphere is large.
149: If the magnetic field measured with $H_{\alpha}$ in the
150: umbral region is weaker, then it shows that the magnetic field gradient is
151: larger than expected which implies larger velocity gradients in the
152: umbral region compared to the penumbral region at chromospheric heights.
153: In order to observationally verify this we have analyzed the velocity and magnetic
154: fields estimated from the simultaneous spectropolarimetric
155: observations of an active region using $H_{\alpha}$ and Fe I spectral lines.
156:
157: The spectropolarimetric observations which were carried out at the Kodaikanal
158: solar telescope using a dual beam polarimeter are discussed
159: in section(\ref{sect:instr&obs}).
160: Sections (3) and (4) discuss the data reduction and analysis
161: procedures respectively.
162: A comparison of LOS magnetic field at the chromosphere and photosphere is presented
163: in section (\ref{sect:losmag}). Velocity gradients obtained through bisector technique
164: are discussed in section (\ref{sect:velgrad}).
165:
166: \section{INSTRUMENT AND OBSERVATIONS}
167: \label{sect:instr&obs}
168:
169: Spectropolarimetric observations were carried out using a newly added dual beam
170: polarimeter to the spectrograph at Kodaikanal solar telescope (\citet{nagaraj07} and
171: also see \citet{bappu67} for details about the spectrograph and telescope setup).
172: The wavelength region of the observations presented in this paper
173: includes $H_{\alpha}(\lambda6563)$ and Fe I ($\lambda6569$) lines.
174: Both $H_{\alpha}$ and Fe I lines are magnetically sensitive
175: with effective Land$\acute{e}$ 'g' factor of 1.048 \citep{casini94}
176: and 1.4 \citep{kobanov03} respectively.
177: The measured instrumental broadening and the linear dispersion close to this
178: wavelength region in second order of diffraction are $38\pm0.5 m\AA$ and
179: $10.15m\AA/pixel$ respectively.
180: The wavelength calibration was done using the telluric lines ($H_2O$) at
181: $\lambda6570.63$ and $\lambda6561.097$ \citep{moore66}.
182:
183: An eight stage modulation scheme was used for the measurement of the general
184: state of polarization \citep{nagaraj07}.
185: In this modulation scheme the measurement of the Stokes parameters are well
186: balanced over the duration of eight stages of intensity measurements.
187: The rotation of the retarders were done manually for the modulation of
188: the input light.
189:
190:
191: The spectropolarimetric data of an active region NOAA0875 presented in
192: this paper were obtained on 28th April, 2006.
193: The heliographic coordinates of the sunspot during observations were
194: $11^o$ south and $18^o$ west.
195: Scanning of the sunspot was done by moving the Sun's image
196: in an east-west direction in steps of $\approx 5''$.
197: For each slit position, the modulated intensities were recorded by
198: the CCD detector. The CCD is a $1K \times 1K $ Photometrics detector with
199: the pixel size of $24 \mu$.
200: Eight stages of modulation took about $90$ s with a typical exposure time
201: of $0.5$ s.
202:
203: \section{DATA REDUCTION}
204:
205: The spectral images were corrected for dark current and gain table
206: variation of the pixels (details about the flat fielding of spectropolarimetric data
207: can be found in \citet{schl02,beck05}).
208:
209: Model independent velocity and magnetic field gradients can be derived
210: from the bisector technique for the range of heights over which the
211: spectral line is formed, provided the full spectral profile is available without any
212: blend \citep{bala97, sankar02}.
213: The application of bisector technique to $H_{\alpha}$ is restricted
214: because of the blend in its red wing by the Zeeman sensitive
215: Co I ($\lambda6563.4$) line which forms at the photospheric heights.
216: In the following section we discuss the procedure to remove this blend
217: and its limitations.
218:
219: \subsection{Blend Removal Procedure}
220: \label{sect:blendrem}
221:
222: Individual profiles of $H_{\alpha}$ corresponding to the orthogonally
223: polarized beams in each stage of modulation were considered for removing the
224: blend. A function which is a linear combination of a Gaussian and a quadratic term was
225: fitted to the blend region of the observed $H_{\alpha}$ profile.
226: This non-linear least square fit (available in IDL) takes into account the curvature in the
227: intensity profile of $H_{\alpha}$ line along with the Co I line profile
228: approximated as a Gaussian. The Gaussian, constructed out of the fitted parameters,
229: was removed from the observed profile.
230: Then the eight intensity measurements were combined to obtain
231: Stokes $I,~Q,~U$ and $V$ spectral images through the demodulation
232: procedure explained in \citet{nagaraj07}.
233: A typical $H_{\alpha}$ intensity profile before and after the blend removal
234: is shown in Fig. \ref{fig:blendrem} with the bisectors marked
235: as diamond symbols.
236:
237: Even though it appears in the total intensity profile that
238: the effect of Co I line is completely removed, the blend residuals
239: still appear in Stokes $Q,~U$, and $V$ profiles.
240: The residuals are due to the Gaussian approximation used for the Co~I line profile.
241: However, it is demonstrated in the section \ref{sect:velgrad} that the blend
242: residuals do not have any effect on the velocity gradients calculated
243: from the $H_{\alpha}$ intensity profiles.
244:
245: The better way to remove the blend may be to synthesise the Stokes profiles of Co I using radiative transfer equations
246: along with the atmospheric parameters obtained through Fe I ($\lambda6569$) line.
247: However in this paper only Stokes $I$ profiles
248: are considered for the velocity gradients estimation and a restricted
249: spectral range of $H_{\alpha}$ about line center for the
250: LOS magnetic field estimation. Hence, the simple blend removal method outlined in this paper
251: is found to be sufficient.
252:
253: \subsection{Correction for Polarimeter Response and Telescope Induced
254: Cross-talks}
255:
256: The polarimetric data was corrected for polarimeter response
257: (for details see \citet{nagaraj07}).
258: The instrumental polarization introduced by the telescope was corrected
259: by using the telescope model developed by \citet{bala85} and \citet{sankar00}.
260: The refractive index values for Aluminium coating
261: used in the model are obtained from the catalog \citep{walter78}.
262: Since these values may
263: be different from the actual values, there still remain residual cross-talks
264: among Stokes parameters. These residual cross-talks are removed by the
265: statistical method given in \citet{schl02}.
266:
267: \section{DATA ANALYSIS}
268: \subsection{LOS Magnetic Field}
269: \label{sect:losb}
270:
271: The LOS magnetic fields at the photospheric and chromospheric
272: heights are derived using the Fe I line and
273: $H_{\alpha}$ respectively. For the derivation of
274: the LOS magnetic fields, center-of-gravity (COG) method was used
275: \citep{rees79, cauzzi93, han03, bala04}.
276:
277: For the COG method, the LOS field strength is given by
278:
279: \begin{equation}
280: B_{LOS} = \frac{(\lambda_+ - \lambda_-)/2}
281: {4.667\times 10^{-13} \lambda_0^2g_L},
282: \end{equation}
283:
284: where $\lambda_0$ is the central wavelength of the line in $\AA$, $g_L$ is
285: the Land$\acute{e}$ `g' factor of the line, and $\lambda_{\pm}$ are the
286: COG wavelengths of the positive and negative circularly polarized
287: components respectively. The COG components are calculated as
288:
289: \begin{equation}
290: \lambda_{\pm} = \frac{\int [I_{cont} - (I \pm V)] \lambda d\lambda}
291: {\int I_{cont} - (I \pm V)d\lambda}
292: \end{equation}
293:
294: where $I_{cont}$ is the local continuum intensity. The integration is over
295: the spectral range of a given spectral line. Since Fe I line is free of blend,
296: the full spectral range is available for the COG method.
297: For the $H_{\alpha}$ line, the spectral range is restricted due to the blending
298: of the Zeeman sensitive Co I line. Even though the blend is removed in Stokes $I$
299: profile using the Gaussian fit technique explained in the section (\ref{sect:blendrem}),
300: there still remains a residual in polarization profiles. To avoid introducing
301: any artifacts in the LOS magnetic field values derived using COG method, a restricted
302: spectral range about the $H_{\alpha}$ line core is used. This spectral region was
303: selected by looking at the strongest Stokes $V$ signal of Co I line. Hence the derived
304: LOS magnetic field using $H_{\alpha}$ would correspond mostly to the upper
305: chromosphere \citep{rutten07}.
306:
307: The expected maximum underestimation of the photopsheric LOS magnetic field
308: is about $12\%$ for the kind of field strengths presented in this
309: paper.
310: Because of the large intrinsic Doppler width of $H_{\alpha}$,
311: the underestimation of the
312: LOS field is not expected even for the strongest magnetic field
313: observed at the chromosphere \citep{han03}.
314: %CORRECTION: MODIFIED THE BELOW SENTENCE
315: However we checked the reliability of the COG method in deriving LOS field
316: from $H_{\alpha}$ by applying it to synthetic Stokes $I$ and $V$ profiles
317: of $H_{\alpha}$ obtained using the radiative transfer code of \citet{han98}.
318: From these studies it was found that the COG method overestimates the LOS field
319: by about $1.5\%$ in the field strength range 0 to 2000 G.
320: %CORRECTION: BELOW SENTENCE IS MODIFIED
321: However, the main error involved in the determination of the LOS magnetic field,
322: for this observation, is due to
323: the error in estimating the shifts between the COG wavelength
324: positions $\lambda_+$ and $\lambda_-$ which is expected to be
325: less than 70 G.
326:
327: \subsection{Photospheric Vector Magnetic Field}
328: \label{sect:vmag:photo}
329: The strength and the orientation of the chromospheric magnetic field are
330: very much dependent on its vector nature at the photosphere
331: (\citet{wiegelmann06} and references there in).
332: Inference of the vector magnetic field at the chromosphere from the
333: $H_{\alpha}$ observations is difficult because,
334: Stokes $Q$ and $U$ are hardly discernible.
335: Only at few locations of penumbra they are above the noise level
336: ($2 \times 10^{-3} I_{cont}$ in our observations).
337: Hence for the correct interpretation of the LOS magnetic field obtained
338: from $H_{\alpha}$ it is important to have the complete information
339: about the vector magnetic field at the photosphere.
340:
341:
342: The photospheric vector magnetic fields are obtained by inverting the observed Stokes profiles of Fe I line.
343: Milne-Eddington Line Analysis using a Numerical Inversion
344: Engine (MELANIE)
345: \footnote{http://www.hao.ucar.edu/public/research/cic/index.html}
346: was used to perform the inversion.
347: %CORRECTION: PLEASE ADD REFERENCE TO MELANIE!
348: MELANIE performs
349: non-linear least-square fitting of the observed Stokes profiles under
350: Local-Thermodynamic-Equilibrium (LTE) condition by assuming Milne-Eddington atmosphere.
351: Inversion code returns magnetic field strength, inclination angle with
352: respect to LOS, azimuth, line strength, damping parameter,
353: LOS velocity, source function and its gradient with optical depth,
354: macroturbulence and fraction of stray light/fill factor of the non-magnetic
355: component.
356:
357: The fit error in estimating the magnetic field ranges from
358: $50$ G for the Stokes profiles which are symmetric and well above
359: the noise level to $250$ G for the Stokes profiles which are highly
360: asymmetric and close to noise level. The maximum fit error in the estimation of
361: the field orientation is about $5^o$ and its azimuth is $6^o$.
362: Other physical parameters returned by the MELANIE are not used in the
363: current work and hence are not discussed here.
364:
365: \subsection{Velocity Gradients}
366: The measurement of the magnetic field in $H_{\alpha}$ has been
367: difficult to interpret \citep{bala04, hanaoka05}.
368: However it is possible to address some of these difficulties from the
369: study of plasma motions as magnetic field and plasma motion
370: influence one another in the solar atmosphere \citep{gary01}.
371: Hence the study of velocity and its gradient may help in better
372: interpretation of the observed magnetic field.
373:
374: The velocities at the photosphere and chromosphere are obtained through
375: COG method \citep{han03}.
376: The COG wavelength $\lambda_{COG}$ of a line profile $I$ is defined as the
377: centroid of its residual intensity profile:
378: \begin{equation}
379: \lambda_{COG} = \frac{\int \lambda(I_{cont}-I)d\lambda}
380: {\int (I_{cont}-I)d\lambda}.
381: \end{equation}
382: The LOS velocity with respect to the average quiet Sun
383: reference ($\lambda_{ref}$) is defined as
384: \begin{equation}
385: v_{LOS} = \frac{c(\lambda_{COG}-\lambda_{ref})}{\lambda_{ref}},
386: \end{equation}
387: where $c$ is the speed of the light.
388:
389: Bisector technique has been applied to Stokes $I$ profiles
390: to derive the velocity gradients both at
391: the photosphere and chromosphere.
392: The Stokes $I$ profiles of Fe I and $H_{\alpha}$ are separately considered
393: for bisector analysis.
394: Bisectors are obtained at 9 equal intensity levels between line core and the wing for
395: Fe I and 14 equal intensity levels between line core and wing for $H_{\alpha}$ respectively.
396: Out of 9 bisectors of Fe I line only 7 are considered, namely, the bisectors between second
397: and eighth counting from the line core. Similarly for $H_{\alpha}$ the bisectors between
398: second and thirteenth are considered totalling 12 bisectors. The bisectors very close to line
399: core and wing are not considered because of the lower signal-to-noise ratio and to avoid the
400: influence of the continuum respectively. For the Fe I line, the wavelength position of the
401: second bisector which corresponds to higher atmospheric layer was subtracted from the seventh
402: bisector which corresponds to lower atmospheric layer. Similarly for $H_{\alpha}$, the wavelength
403: position of the second bisector was subtracted from the thirteenth bisector.
404: The wavelength differences
405: ($\Delta\lambda$s) thus obtained are converted into velocities - which would then represent the
406: velocity difference between the lower and higher atmospheric layers - using the
407: following relation,
408: \begin{equation}
409: \label{eq:veldifbisec}
410: \Delta V_{bs} = \frac{c \Delta\lambda}{\lambda_0}.
411: \end{equation}
412: Where $\lambda_0$ is the rest wavelength of the spectral line under
413: consideration. The velocity difference defined
414: in Eq. (\ref{eq:veldifbisec}) represents the velocity gradient over the line
415: formation height.
416:
417: The errors in estimating the velocity differences are mainly due to the
418: errors involved in finding the wavelength shifts.
419: The maximum error in estimating
420: the velocity gradients is about $0.09$ km s$^{-1}$.
421:
422: \subsection{Stokes $V$ Amplitude Asymmetry}
423: Amplitude and area asymmetries of Stokes $V$ profiles are caused
424: by the gradients in the velocity and magnetic fields
425: (see for eg. \citet{sanchez92, sankar02}). Hence their analysis will give
426: us some handle on the understanding of the field gradients.
427:
428: If $a_r$ and $a_b$ represent the amplitudes of red and blue wings of Stokes $V$
429: respectively then the amplitude asymmetry is defined as,
430:
431: \begin{equation}
432: \delta a = \frac{|a_b|-|a_r|}{|a_b|+|a_r|}.
433: \end{equation}
434:
435: The area asymmetry is not considered in this paper due to the difficulty in estimating
436: it for the $H_{\alpha}$ in the presence of Co I line blend.
437:
438: \section{RESULTS AND DISCUSSIONS}
439:
440: \subsection{Comparison of Photospheric and Chromospheric LOS Magnetic Fields}
441: \label{sect:losmag}
442:
443: The scatter plot of LOS magnetic field derived from $H_{\alpha}$ and Fe I is
444: shown in Fig. \ref{fig:28fld}. The plot shows that the chromospheric
445: magnetic field is weaker in general compared to its photospheric counterpart.
446: The important point to note from this figure is that the chromospheric
447: fields are much weaker in the locations where the strong photospheric fields
448: are observed.
449: Similar kind of observations were reported earlier by \citet{bala04} using
450: simultaneous observations in $H_{\alpha}$ and Fe I line at $\lambda6301.5$ and
451: \citet{hanaoka05} who compared the LOS field measured in $H_{\alpha}$ with the
452: magnetograms of SOHO/MDI.
453: These observations may imply that the stronger fields weaken much faster
454: when they propagate upward in the solar atmosphere.
455:
456: To illustrate the quicker weakening of the stronger fields, plots of the
457: photospheric and the chromospheric LOS field strengths along two different
458: radial slices of the sunspot are shown on the right side of the Fig. \ref{fig3}.
459: The radial slices considered for this purpose are marked as 1 and 2 on
460: the SOHO/MDI intensitygram showing the sunspot analysed in this paper
461: \footnote{Because of the coarser step size used for scanning the sunspot
462: ($\approx$ 5 arc sec), the raster images do not give a good representation of
463: the observed region and also there was no imaging facility available during
464: these observations and therefore intensitygram from SOHO/MDI is used.
465: The radial cuts marked on the image represent the approximate slit
466: positions of the observations.}.
467: The top panel on the right side of the Fig. \ref{fig3} shows the plots
468: of photospheric and chomospheric LOS magnetic field strengths along the
469: radial cut (marked as 1 on the sunspot image) passing close to the central
470: umbra. Note that, along this radial cut the photospheric field strength
471: increases systematically from the penumbral region to the umbral region.
472: While the chromospheric field strength increases upto about 800 G, more or
473: less linear with the photospheric field, and then starts decreasing towards the
474: umbra. The decrease in the chromospheric field is much larger close to central
475: umbra. The photospheric field along the radial cut 2 as shown in the bottom
476: panel on the right side of the Fig. \ref{fig3} shows the behavior very
477: similar to that was seen along the radial cut 1. That means the photospheric
478: field strength is larger in the umbral region and decreases towards
479: the edges of the sunspot. In the case of chromospheric field strength,
480: the values do not decrease towards the umbra as was seen for the radial cut 1
481: but, they are considerably smaller compared to its photospheric counter part.
482: The main difference between these two radial cuts is that the field strength
483: in the umbral photosphere for radial cut 1 is larger compared to that of
484: radial cut 2.
485: This may be an indication of the faster divergence of the stronger fields.
486: However, there are other possibilities which can cause this observed weaker fields
487: in the umbral chromosphere and they are : weakening of the polarization signal due to
488: scattered light within the instrument and from nearby quiet Sun; peculiarity of
489: the atmosphere such as discussed by \citet{hanaoka05};
490: the methods used to derive LOS magnetic field (inversion of $H_{\alpha}$
491: Stokes profiles is yet to be established);
492: inclination of the magnetic field;
493: and most importantly the height of formation which
494: is highly ambiguous \citep{hector04}.
495: We will discuss these issues in detail after looking at the results from the
496: analysis of velocity gradients both at the photosphere and chromosphere.
497:
498: \subsection{Velocity Gradients and the Nature of the Magnetic Fields}
499: \label{sect:velgrad}
500:
501: The velocities calculated using COG method show a
502: typical behavior of Evershed flows both at the photosphere and chromosphere.
503: That means the limb side penumbra shows red shift with
504: respect to quiet Sun where as center side penumbra shows blue shift
505: at the photospheric heights. The situation is exactly opposite
506: at chromospheric heights which are consistent with the well known
507: inverse Evershed effect.
508: COG velocities in umbral regions both at photosphere and chromosphere
509: are smaller compared to penumbral regions.
510:
511: The plots of bisector velocity differences (Eq. \ref{eq:veldifbisec})
512: as a function of photospheric magnetic field strength are shown
513: in Fig. \ref{fig:velgradfe} for the Fe I line.
514: Top panel in this figure is for all the points over
515: the total field of view (FOV), where as the bottom left and right panel is for umbral and
516: penumbral regions of the observed spot, respectively.
517: Note from these figures that the large number of points correspond to
518: umbra have smaller velocity gradients than the penumbra.
519: Closer investigation of bisector velocity differences of Fe I line
520: show the flow pattern consistent with the well known Evershed flow.
521: That means larger portion of the limb side penumbra shows net downflow
522: (both core and wing side bisectors show redshifts)
523: and disk center side penumbra shows net upflow (both
524: core and wing bisectors show blue shifts with respect to the reference).
525: For the definition of the net up- and down-flow see \citet{bala97}.
526:
527: We also found that the bisector velocity gradients and COG velocities
528: observed in Fe I show a good correlation in agreement with the earlier
529: observations. That means, larger the COG velocity larger the velocity
530: gradient at the photosphere.
531:
532: In Fig. \ref{fig:vgradhalpha_full} the plots of bisector velocity
533: gradients for $H_{\alpha}$ v/s the photospheric magnetic fields are shown.
534: Top panel in this figure is for the total FOV and left bottom panel for
535: umbral region and right bottom panel for penumbral region.
536: Notice the increase in bisector velocity gradients measured in $H_{\alpha}$
537: with increase in the photospheric magnetic field strength.
538: Plots of velocity gradients along two radial slices of sunspot (marked as
539: 1 and 2 in Fig. \ref{fig3}) also show that the larger velocity gradients at the
540: chromosphere are located at which strong photospheric fields are observed
541: (Fig. \ref{fig6}). Top panels in Fig. \ref{fig6} show the plots of
542: photospheric field strengths along the radial cut 1 and the radial cut 2.
543: The corresponding plots of velocity gradients are shown in the bottom panels.
544: Except at few locations (mostly in the limb side penumbra such as shown in the
545: bottom right panel of Fig. \ref{fig6}) which show less variation in the values
546: of velocity gradients inspite of variation in the photospheric field
547: strengths (which is evident also in Fig. \ref{fig:vgradhalpha_full}), most of
548: the places the velocity gradients are larger where the photospheric magnetic
549: field strengths are larger.
550:
551: Closer examination of the bisector wavelength positions
552: of $H_{\alpha}$ Stokes $I$ profiles in different
553: regions of the sunspot with respect to the reference wavelength
554: reveals the following results (see figure
555: \ref{fig:comprbisec}).
556: %CORRECTION IS MADE IN THE LAST SENTENCE OF ABOVE PARAGRAPH
557:
558: \begin{itemize}
559: \item{} In the limb side penumbra both line core and wing side bisectors
560: show blue shift with respect to quiet Sun. The shifts in the line core
561: bisectors are large compared to the line wing bisectors
562: indicating the net upflow.
563: \item{} In the umbral region also both line core side and wing side bisectors
564: show blue shifts with respect to the reference wavelength. The shifts in the
565: line core side bisectors for the umbra are almost comparable to that of the
566: limb side penumbra. However, the shifts in the line wing side bisectors for the umbra
567: are much smaller with respect to that of the limb side penumbra.
568: This would indicate again the net upflows but with larger velocity gradients.
569: \item{} In the center side penumbra the line wing bisectors show redshift in most of
570: the places with respect to the quiet Sun. At few locations they show a small blue
571: shift or no shift. The line core side bisectors show slight blue shift with respect
572: to quiet Sun reference wavelength position.
573: \end{itemize}
574:
575: It is found through these analyses that the velocity gradients are larger
576: in the umbra at the chromospheric heights compared to the penumbra
577: (see Fig. \ref{fig:vgradhalpha_full}).
578: Which is exactly in contrast with the flow pattern observed at the photosphere.
579: The analysis of bisector velocity differences also indicate accelerated
580: upflows in the umbral region.
581:
582: There is a little concern due to the residuals present after the Co I line
583: blend is removed. To make sure that there are no artifacts introduced due to
584: the residuals, the bisector velocity differences calculated by considering
585: the spectral region of the line which is not affected by the blend are shown
586: in Fig. \ref{fig:vgradhalpha_part} as a function of photospheric field
587: strength. This figure also shows the trend that the velocity gradients
588: increase with increase in photospheric magnetic field strength confirming
589: the observations made with the blend removed full intensity profiles. As
590: expected, the gradients are smaller due to the smaller wavelength regions
591: considered in this case.
592:
593: To summarise, wavelength shifts analysis of Stokes $I$ bisectors show
594: that the velocity gradients are larger in the umbral region than in
595: the penumbral region at the chromospheric heights.
596: Most importantly, accelerated upflows are observed in the umbral region.
597: In other words, LOS velocity increases upward more rapidly in
598: the umbral region than in the penumbral region at the chromosphere.
599:
600: Let us now address some of the possibilities, mentioned in the
601: beginning of this section, which can cause
602: the observed weaker chromospheric field in the umbral region from
603: the point of view of the velocity gradients.
604:
605: We found that the stray light
606: within the instrument is less than $2\%$ by comparing the quiet Sun
607: spectrum with the atlas \citep{wallace00}. This amount of stray light
608: is too small to account for the observed weaker Stokes $V$ signals
609: in the umbral region. Also, it can not account for the observed
610: velocity gradients in $H_{\alpha}$.
611:
612: The next question raised was about the reliability of the methods used to
613: estimate the LOS magnetic field. Despite the availability of more accurate
614: estimation of photospheric magnetic field through inversion, we
615: have used the COG method to maintain the uniformity while comparing
616: the LOS field at the photosphere and chromosphere (Fig. \ref{fig:28fld}).
617: This is because the inversion of $H_{\alpha}$ profiles to derive magnetic field
618: is yet to be established.
619: As discussed in section \ref{sect:losb} the underestimation of the LOS
620: field obtained from Fe I is expected to be more while the values
621: %CORRECTION: THE BELOW LINE IS MODIFIED
622: obtained from $H_{\alpha}$ will be less as confirmed through numerical simulations.
623: %should be more close to the actual values.
624: Also, the observed umbral fields are larger at the photosphere where as
625: they are smaller at the chromosphere.
626: Hence the departure of LOS magnetic fields from the actual values
627: due to the method used to estimate them can not explain the
628: %CORRECTION: THE BELOW LINE IS MODIFIED
629: observed weakening of the magnetic fields.
630:
631: One of the possibilities which can cause the reduction in polarization
632: (in effect the weaker magnetic field)
633: proposed by \citet{hanaoka05} is the peculiarity of the atmosphere.
634: However, the kind of peculiarity discussed by him can not explain the
635: velocity gradients and the Stokes profiles of $H_{\alpha}$ discussed in this
636: paper.
637:
638: It is a general wisdom gained from the extrapolation technique that the
639: magnetic topology at the chromosphere is very much dependent on the field
640: configuration at the photosphere. Extrapolation techniques like potential field
641: approximation suggest that larger the field inclination at the photosphere
642: chromospheic fields should also show larger inclination.
643: From the inversion of Fe I line it was
644: found that the fields are oriented more close to LOS in the umbral region than
645: in the penumbral region. Hence we expect larger field orientation in
646: penumbral region at the chromospheric heights also.
647: As mentioned in section \ref{sect:vmag:photo}
648: $H_{\alpha}$ shows Stokes $Q$ and $U$ signals which are above the noise level
649: in penumbral region but not in umbral region indicating that the orientation
650: of the fields are larger in the penumbral region compared to the umbral region
651: even at the chromosphere.
652: Hence we believe that the weaker LOS field observed in
653: the umbral region may not be due to the larger orientation angle with respect
654: to the LOS. This scenario is also verified using the velocity gradients observed
655: at the chromosphere, as larger inclination means smaller LOS velocity in
656: contradiction to the observed velocities (Fig. \ref{fig:vgradhalpha_full}).
657:
658: Another major difficulty in interpreting the $H_{\alpha}$ observations
659: is the ambiguity in its height of formation. Studies based on the
660: response functions by \citet{hector04} show that the $H_{\alpha}$
661: is sensitive mostly to chromospheric magnetic field in the umbral model.
662: While in the quiet Sun model it shows sensitivity to both photospheric and
663: chromospheric magnetic fields. Since the magnetic fields at the photosphere
664: is large, the magnetic field measured in quiet Sun regions will mostly
665: be photopsheric.
666: If the $H_{\alpha}$ were to show the sensitivity to magnetic field
667: in the penumbral model similar to that of quiet Sun model then one would expect
668: larger field values in the penumbra which will be an average of photospheric and
669: chromospheric fields while the umbral fields are exclusively chromospheric.
670: To confirm this, more forward modeling is needed which includes
671: the comparison of $H_{\alpha}$ line formation at different regions
672: with different field configurations.
673: More over, response of $H_{\alpha}$ line to various physical
674: parameters needs to be studied and consistently explain
675: the observations such as velocity and velocity gradients presented
676: in this paper.
677:
678: Other possibility which can cause the observed weaker chromospheric fields
679: in the umbral region is the faster divergence of stronger fields as suggested
680: by \citet{bala04}. This scenario may consistently explain both the observed
681: magnetic properties (Fig. \ref{fig:28fld} ) as well as the velocity properties
682: (Figs. \ref{fig:vgradhalpha_full}/\ref{fig:vgradhalpha_part}).
683: Because of the decrease in the field strength, the plasma becomes less
684: inhibited by the magnetic fields and hence more free to move. That means if the
685: stronger fields diverge faster compared to weaker fields then the the velocities
686: should increase faster in the stronger field regions which are consistent with
687: the observations.
688:
689:
690: \subsection{Stokes $V$ Amplitude Asymmetry}
691:
692: From the close examination of Stokes $V$ profiles of $H_{\alpha}$ we found that the Co I
693: line has not intruded up to the extent that the amplitudes are affected.
694: Hence the results presented from $H_{\alpha}$ based on amplitude
695: asymmetry are reliable.
696:
697: From the Fig. \ref{fig:ampasym_fe} it is clear that the amplitude asymmetry
698: of Fe I line tends toward zero
699: with increase in photospheric magnetic field strength. This means, at the
700: photospheric heights the gradients are smaller in umbra which corresponds
701: to the region of strong fields.
702: In contrast amplitude asymmetry observed in $H_{\alpha}$ tend to increase
703: with photospheric field strength as shown in the Fig. \ref{fig:ampasym_hi}.
704: This implies that the field gradients are larger in the umbra at the
705: chromosphere compared to penumbra.
706: Hence the analysis of amplitude asymmetry confirm the results obtained from the
707: bisector analsysis in section \ref{sect:velgrad}.
708:
709: \section{CONCLUSIONS}
710: \label{sect:concl}
711:
712: In this paper we have used the multiwavelength spectropolarimetric tool to
713: understand the stratification of the magnetic and velocity fields in the solar
714: atmosphere. Out of the two lines considered for spectropolarimetry,
715: one forms at photospheric height (Fe I $\lambda6569$) and the other spans
716: almost from the upper photosphere to upper chromosphere ($H_\alpha$). Hence,
717: these lines were useful in studying the connection between physical parameters
718: at the photospheric and chromospheric heights. The main physical parameters
719: studied in this paper are the magnetic and velocity fields in an active region.
720:
721: As discussed in section \ref{sect:losmag}, the LOS magnetic field measured in
722: chromosphere is in general weak compared to its photospheric counterpart. The
723: weakening of the chromospheric field is much faster for the corresponding strong
724: photospheric field. The magnetic field strengths observed in the umbral
725: chromosphere are much weaker than those expected from the extrapolation of
726: the photospheric magnetic field. For instance, the field strength inferred
727: through $H_{\alpha}$ observations is about 400 G where as the field strength
728: obtained through the extrapolation of the observed photosperic field under
729: potential field approximation to an height of 2000 km is about 1000 G
730: (assuming that the height of formation of $H_{\alpha}$ is about 2000 km).
731: Various possibilities have been discussed which can cause the weaker fields
732: observed in the umbral chromosphere.
733: The most probable ones are the fast divergence of the stronger fields when
734: they propagate upward in the atmosphere
735: \citep{bala04} and ambiguity in the height resolution of
736: $H_{\alpha}$ magnetic sensitivity which may be
737: photospheric and/or chromospheric depending on the region of observation
738: \citep{hector04}. If former is the reason, then it can explain the observed
739: properties of both velocity and the magnetic fields presented in this paper.
740: This is because, as shown in Fig. \ref{fig3}, observed chromospheric fields
741: are systematically weaker at the locations where strong photospheric fields
742: are observed. If the weaker field strengths observed at the umbral chromosphere
743: are truly of solar origin then this implies that the umbral fields
744: decrease more rapidly with height compared to penumbral fields.
745: Rapid decrease in field strengths cause rapid increase in velocity with
746: height as there is a small upflow in the umbral photosphere.
747: Observations also show that the velocity increases more rapidly in the
748: umbral region compared to penumbral region (section \ref{sect:velgrad}).
749: Earlier observations by \citet{gosain03} have also indicated the weakening
750: of the magnetic field which is larger for stronger fields.
751: In their observations quicker weakening of the stronger fields is not apparent,
752: probably, because of the lines (Mg I b1 and b2 at $\lambda5173$ and
753: $\lambda5184$) used to infer
754: chromospheric magnetic field that originate at the lower chromosphere.
755: While the observations based on $H_{\alpha}$ presented in this paper
756: as well as earlier by \citet{bala04,hanaoka05} show clearly that the
757: strong fields weaken quickly with height because in these works, to infer
758: chromospheric magnetic field only the spectral region close to its line
759: core is considered which samples mostly the higher layers of the
760: chromosphere \citep{rutten07}.
761: Hence there is a possibility that the weaker LOS field strengths observed in
762: the umbral chromosphere are caused due to the faster divergence of the stronger
763: fields. However, we would like to caution that the reliability of the COG
764: method in estimating the LOS magnetic field strengths discussed in this paper
765: (section \ref{sect:losb}) is for simple solar atmospheric model.
766: But, in reality the solar atmosphere may be complicated.
767: More studies on the magnetic and velocity response functions for $H_{\alpha}$
768: in different regions will help in better interpretation of the observations.
769: Simultaneous multiline spectropolarimetry, which includes $H_{\alpha}$ and at
770: least one more line which is formed at the chromosphere (preferably Infrared
771: $Ca$ Triplet line at $\lambda$8542) and a photospheric line, is needed to get
772: further insight into the physical processes that take place in the chromosphere.
773:
774: \acknowledgments
775: Hector Socas-Navarro and Han Uitenbroek are thankfully acknowledged for
776: providing us MELANIE and RH codes. We thank K. S. Balasubramaniam for
777: his critical comments on our earlier version of this paper. We also thank
778: S. P. Rajaguru for useful discussions regarding the connection between
779: velocity and magnetic field gradients and B. Ravindra for providing us the
780: extrapolated magnetic field strengths under potential field approximation.
781: Thanks to the help of Devendran and Hariharan during the observations.
782: The sunspot image was obtained from SOHO/MDI data acrhive. SOHO is a mission
783: of international cooperation between ESA and NASA.
784:
785: \begin{thebibliography}{100}
786:
787: \bibitem[Balasubramaniam et al.(2004)]{bala04}
788: Balasubramaniam, K. S., Christopoulou, E. B., \& Uitenbroek, H. 2004, \apj, 606, 1233
789:
790: \bibitem[Balasubramaniam et al.(1997)]{bala97}
791: Balasubramaniam, K. S., Keil, S. L., \& Tomczyk, S. 1997, \apj, 482, 1065
792:
793: \bibitem[Balasubramaniam et al.(1985)]{bala85}
794: Balasubramaniam, K. S., Venkatakrishnan, P., \& Bhattacharya, J. C.
795: 1985, \solphys, 99, 333
796:
797: \bibitem[Bappu(1967)]{bappu67}
798: Bappu, M. K. V. 1967, \solphys, {\bf 1}, 151
799:
800: \bibitem[Beck et al.(2005))]{beck05}
801: Beck, C., Schmidt, W., Kentischer, T., \& Elmore, D. F.,
802: 2005, \aap, 437 1159
803:
804: \bibitem[Casini \& Landi Degl'Innocenti(1994))]{casini94}
805: Casini, R., \& Landi Degl'Innocenti, E. 1994, \aap, 291, 668
806:
807: \bibitem[Cauzzi et al.(1993)]{cauzzi93}
808: Cauzzi, G., Smaldone, L. A., Balasubramaniam, K. S., \& Keil, S. L. 1993,
809: \solphys, 146, 207
810:
811: \bibitem[Choudhary et al.(2002)]{choudhary02}
812: Choudhary, D. P., Suemastu, Y., \& Ichimoto, K. 2002, \solphys, 209, 349
813:
814: \bibitem[Gary(2001)]{gary01}
815: Gary, G. A. 2001, \solphys, 203, 71
816:
817: \bibitem[Gosain \& Choudhary(2003)]{gosain03}
818: Gosain, S., \& Choudhary, D. P. 2003, \solphys, 209, 349
819:
820: \bibitem[Hanaoka(2005)]{hanaoka05}
821: Hanaoka, Y., 2005, Publ. Astron. Soc. Japan, 57, 235
822:
823: \bibitem[Judge(2007)]{judge07}
824: Judge, P. G. 2007, \apj, 662, 677
825:
826: \bibitem[Kobanov et al.(2003)]{kobanov03}
827: Kobanov, N. I., Makarchik, D. V., \& Sklyar, A. A. 2003, \solphys, 217, 53
828:
829: \bibitem[Moore et al.(1966)]{moore66}
830: Moore, C. E.,Minneart, M. G. J., Houtgast, J., \& Rowland, H. A. 1966, The
831: Solar Spectrum 2935 \AA to 8770 \AA. (Washington:NIST)
832:
833: \bibitem[Nagaraju et al.(2007))]{nagaraj07}
834: Nagaraju, K., Ramesh, K. B., Sankarasubramanian, K., \& Rangarajan, K. E.
835: 2007, Bull. Astro. Soc. India, 35, 307
836:
837: \bibitem[Rajaguru et al.(2006)]{rajaguru06}
838: Rajaguru, S. P., Wachter, R., \& Hasan, S. S. 2006, ILWS, Goa
839:
840: \bibitem[Rees \& Semel(1979)]{rees79}
841: Rees, D. E., \& Semel, M. D. 1979, \aap, 74, 1
842:
843: \bibitem[Regnier \& Canfield(2006)]{regnier06}
844: Regnier, S., \& Canfield, R. C. 2006, \aap, 451, 319
845:
846: \bibitem[Rutten(2007)]{rutten07}
847: Rutten, R. J., 2007, in ASP Conf.Ser.368, The Physics of Chromospheric Plasmas, ed. P. Heinzel, I. \& R. J. Rutten, 27
848:
849: \bibitem[Sanchez Almeida \& Lites(1992)]{sanchez92}
850: Sanchez Almeida, J., \& Lites, B. W. 1992, \apj, 398, 359
851:
852: \bibitem[Sankarasubramanian \& Rimmele(2002)]{sankar02}
853: Sankarasubramanian, K., \& Rimmele, T. R. 2002, \apj, 576, 1048
854:
855: \bibitem[Sankarasubramanian(2000)]{sankar00}
856: Sankarasubramanian, K. 2000, Ph.D. thesis, Bangalore Univ.
857:
858: \bibitem[Schlichenmaier \& Collados(2002)]{schl02}
859: Schlichenmaier, R., \& Collados, M. 2002, \aap, 381, 668
860:
861: \bibitem[Socas-Navarro(2005)]{hector05}
862: Socas-Navarro, H. 2005, \apj, 631, L67
863:
864: \bibitem[Socas-Navarro \& Uitenbroek(2004)]{hector04}
865: Socas-Navarro, H., \& Uitenbroek, H. 2004, \apj, 603, L129
866:
867: \bibitem[Spruit \& Zweibel(1979)]{spruit79}
868: Spruit, H. C., \& Zwibel, E. G. 1979, \solphys, 62, 15
869:
870: \bibitem[Uitenbroek(1998)]{han98}
871: Uitenbroek,H. 1998, \apj, 498, 427
872:
873: \bibitem[Uitenbroek(2003)]{han03}
874: Uitenbroek,H. 2003, \apj, 592, 1225
875:
876: \bibitem[Vernazza et al.(1981)]{vernaza81}
877: Vernazza, J. E., Avrett, E. H., \& Loeser, R. 1981, \apjs, 45, 635
878:
879: \bibitem[Wallace et al.(2000)]{wallace00}
880: Wallace, L., Hinkle, K., \& Livingston, W. 2000,
881: An atlas of sunspot umbral spectra in the visible,
882: from 15,000 to 25,500 $cm^{-1}$ (3920 to 6664 $\AA$)
883:
884: \bibitem[Walter (1978)]{walter78}
885: Walter, D. G. 1978, Hand Book of Optics, McGraw-Hill Book Company, New York
886:
887: \bibitem[Wiegelmann et al.(2006)]{wiegelmann06}
888: Wiegelmann, T., Inhester, B., Kleim, B., Valori, G., \& Neukirch, T. 2006, \aap, 453, 737
889:
890: \end{thebibliography}
891:
892: \clearpage
893:
894: \begin{figure}[h]
895: %\includegraphics{demo_blendrem.eps}
896: \plotone{f1.eps}
897: \caption{Typical $H_{\alpha}$ Stokes $I$ profile with the bisectors
898: marked in diamond symbols. Dotted line in the top panel shows
899: the spectral region of the profile after removing the blend.}
900: \label{fig:blendrem}
901: \end{figure}
902: \clearpage
903:
904: \begin{figure}
905: %\includegraphics{blos_fe_halpha_bary_28apr06.eps}
906: \plotone{f2.eps}
907: \caption{Scatter plot of photospheric and chromospheric LOS magnetic field
908: derived using COG method applied to Fe I($\lambda6569$) and
909: $H_{\alpha}$ respectively. Solid line is a linear fit to the field
910: strengths correspond to the penumbral region.}
911: \label{fig:28fld}
912: \end{figure}
913: \clearpage
914:
915: \begin{figure}
916: %\includegraphics{blos_fe_halpha_bary_28apr06.eps}
917: %\plotone{f3.eps}
918: \plottwo{f3a.eps}{f3b.eps}
919: \caption{On the right side of this figure shown are the plots of LOS
920: magnetic field strengths at the photosphere (dashed lines) and
921: the chromosphere (dotted lines) along two radial slices of the sunspot.
922: For reference, these two radial cuts are marked on the SOHO/MDI intensitygram.
923: The arrow on the sunspot image indicates the disk center direction.}
924: \label{fig3}
925: \end{figure}
926: \clearpage
927:
928: \begin{figure}
929: %\includegraphics{plot_vgrad_bisec_fld_fe.eps}
930: \plotone{f4.eps}
931: \caption{Plots of velocity gradients at the photosphere
932: v/s photospheric magnetic field strength.}
933: \label{fig:velgradfe}
934: \end{figure}
935: \clearpage
936:
937: \begin{figure}
938: %\includegraphics{plot_vgrad_bisec_fld_halpha_full.eps}
939: \plotone{f5.eps}
940: \caption{Plots of velocity gradients at chromosphere derived using $H_{\alpha}$ with the bisectors considered for the full line profile
941: v/s the total field strength (photospheric). Solid lines are the linear
942: fit to the data points.}
943: \label{fig:vgradhalpha_full}
944: \end{figure}
945:
946: \begin{figure}
947: \plotone{f6.eps}
948: \caption{The top panels in this figure show the plots of magnetic field
949: strength at the photosphere along two radial slices of the sunspot.
950: The radial slices are same as those shown in Fig. \ref{fig3}.
951: The corresponding plots of velocity gradients
952: at the chromosphere are shown in the bottom panels.}
953: \label{fig6}
954: \end{figure}
955: \clearpage
956:
957: \begin{figure}
958: %\includegraphics{bisec_regions_i.eps}
959: \includegraphics{f7.eps}
960: \caption{Representative bisectors of $H_{\alpha}$ line profiles
961: correspond to different regions of the sunspot are plotted on an
962: average quiet Sun profile.
963: The solid line represents the reference wavelength (which is a COG wavelength
964: position of the nearby average quiet Sun profile). Dotted line represents
965: the bisectors location typical of umbral profiles, dashed line-limb
966: side penumbra, dash dotted line-center side penumbra.}
967: \label{fig:comprbisec}
968: \end{figure}
969: \clearpage
970:
971: \begin{figure}
972: %\includegraphics{plot_vgrad_bisec_fld_halpha_part.eps}
973: \plotone{f8.eps}
974: \caption{ Same as Fig. \ref{fig:vgradhalpha_full} but
975: with the limited spectral range about the line core considered for
976: calculating velocity gradients to avoid the influence of blending due to Co I
977: line.
978: The axes scales are kept same as Fig. \ref{fig:vgradhalpha_full} to
979: indicate that the velocity gradients are smaller compared to the case when the
980: full Stokes $I$ profile is considered.}
981: \label{fig:vgradhalpha_part}
982: \end{figure}
983: \clearpage
984:
985: \begin{figure}
986: %\includegraphics{plot_ampl_asym_fld_fe_linfit.eps}
987: \includegraphics{f9.eps}
988: \caption{Plots of amplitude asymmetry of Stokes V profiles of Fe I line v/s the
989: total field strength (photospheric). The solid lines are the linear fits
990: to the data points.}
991: \label{fig:ampasym_fe}
992: \end{figure}
993: \clearpage
994:
995: \begin{figure}
996: %\includegraphics{plot_ampl_asym_fld_halpha_linfit.eps}
997: \includegraphics{f10.eps}
998: \caption{Plots of amplitude asymmetry of Stokes V profiles of $H_{\alpha}$ line
999: v/s the total field strength (photospheric). The solid lines are the linear fits
1000: to the data points.}
1001: \label{fig:ampasym_hi}
1002: \end{figure}
1003:
1004:
1005: \end{document}
1006:
1007: %%
1008: %% End of file `sample.tex'.
1009: