1:
2: \documentclass[12pt,preprint]{aastex}
3:
4: %% manuscript produces a one-column, double-spaced document:
5:
6: %%\documentclass[manuscript]{aastex}
7:
8: %% preprint2 produces a double-column, single-spaced document:
9:
10: %% \documentclass[preprint2]{aastex}
11:
12: %% Sometimes a paper's abstract is too long to fit on the
13: %% title page in preprint2 mode. When that is the case,
14: %% use the longabstract style option.
15:
16: %% \documentclass[preprint2,longabstract]{aastex}
17:
18: %% If you want to create your own macros, you can do so
19: %% using \newcommand. Your macros should appear before
20: %% the \begin{document} command.
21: %%
22: %% If you are submitting to a journal that translates manuscripts
23: %% into SGML, you need to follow certain guidelines when preparing
24: %% your macros. See the AASTeX v5.x Author Guide
25: %% for information.
26:
27: \def\bibfiles{grana_mag}
28: \def\aareferences{\bibliographystyle{aabib}
29: \bibliography{aajour,\bibfiles}}
30:
31:
32: \def\rmit#1{{\it #1}} %% italics (RR mode, Kluwer)
33: \def\etal{\rmit{et al.}}
34: \def\etc{\rmit{etc.}}
35: \def\ie{\rmit{i.e.}}
36: \def\eg{\rmit{e.g.}}
37: \def\degree{\hbox{$^\circ$}}
38: \def\arcsec{\hbox{$^{\prime\prime}$}}
39:
40: \def\kms{\hbox{km$\;$s$^{-1}$}}
41: \def\ms{\hbox{m$\;$s$^{-1}$}}
42:
43:
44:
45:
46: \shorttitle{MHS sunspot model} \shortauthors{Khomenko and
47: Collados}
48:
49: %% This is the end of the preamble. Indicate the beginning of the
50: %% paper itself with \begin{document}.
51:
52: \begin{document}
53:
54: %% LaTeX will automatically break titles if they run longer than
55: %% one line. However, you may use \\ to force a line break if
56: %% you desire.
57:
58: \title{MHS sunspot model from deep sub-photospheric to chromospheric layers}
59:
60: \author{E. Khomenko\altaffilmark{1,2}, M. Collados\altaffilmark{1}}
61: \email{khomenko@iac.es, mcv@iac.es}
62:
63:
64: \altaffiltext{1}{Instituto de Astrof\'{\i}sica de Canarias, 38205
65: La Laguna, Tenerife, Spain} \altaffiltext{2}{Main Astronomical
66: Observatory, NAS, 03680, Kyiv, Ukraine}
67:
68: \begin{abstract}
69: In order to understand the influence of magnetic fields on the
70: propagation properties of waves, as derived from different local
71: helioseismology techniques, forward modeling of waves is required.
72: Such calculations need a model in magnetohydrostatic equilibrium
73: as initial atmosphere to propagate oscillations through it. We
74: provide a method to construct such a model in equilibrium for a
75: wide range of parameters to be used for simulations of artificial
76: helioseismologic data. The method combine the advantages of
77: self-similar solutions and current-distributed models. A set of
78: models is developed by numerical integration of magnetohydrostatic
79: equations from the sub-photospheric to chromospheric layers.
80: \end{abstract}
81:
82:
83: \keywords{MHD; Sun: magnetic fields; Sun: sunspots}
84:
85:
86: %________________________________________________________________
87: \section{Introduction}
88:
89: In the recent years, local helioseismology has provided new
90: insights into the sub - photospheric structure of quiet and active
91: regions of the Sun \citep{Duvall+etal1993, Kosovichev1999,
92: Kosovichev2002, Kosovichev+etal2000, Zhao+Kosovichev2003,
93: Braun+Lindsey2000}.
94: %
95: However, the influence of magnetic fields on data interpretation
96: has not been fully explored. Theoretical efforts have been made by
97: \citet{Crouch+Cally2003}, \citet{Cally2005, Cally2006},
98: \citet{Schunker+Cally2006}, \citet{Cally+Goossens2007},
99: \citet{Schunker+etal2008} in order to include mode conversion and
100: to model ray path of waves in magnetized structures by means of
101: analytical theory.
102: %
103: These studies confirm the potential importance on helioseismic
104: measurements of the so called surface effects caused by the
105: presence of a magnetic field. A more complete understanding of the
106: problem will, probably, be better reached via direct forward
107: modeling of helioseismological data, since magnetic fields of
108: arbitrary configuration may be used.
109: %
110: There are several recent works reported in this direction as, \eg,
111: \citet{Gizon+etal2006, Khomenko+Collados2006,
112: Parchevsky+Kosovichev2007, Shelyag+etal2007, Hanasoge2008,
113: Cameron+etal2008}. In all the works cited above
114: \citep[except][]{Shelyag+etal2007}, the authors apply a similar
115: strategy. In particular, they assume the existence of an
116: equilibrium atmosphere containing a magneto-static structure whose
117: properties may resemble to a larger or lesser extent those of a
118: sunspot or a magnetic flux tube. A small-amplitude perturbation is
119: then applied to the system in order to study wave propagation,
120: wave mode transformations, amplitude and phase behaviour of waves
121: in a complex magnetic field topology, \etc
122: %
123:
124: The behaviour of waves observed in active regions is very
125: sensitive to their magnetic field configuration. Photosphere and
126: low chromosphere are regions where a small change of parameters
127: such as the size of the magnetic structure, or its temperature,
128: density or magnetic field strength and inclination, may produce
129: significant changes in the resulted wave field. Simulations can be
130: of invaluable help to explore, within a full parameter space,
131: different magneto-static structures in order to understand the
132: effects produced by the magnetic field on the measurable variables
133: used in local helioseismology.
134: %
135:
136: The latter task requires a robust procedure to construct
137: magneto-static structures of desired properties. In this paper, we
138: propose a strategy with that aim and apply it to obtain thick
139: structures, as prototypes for solar spots and pores.
140: %
141: As a minimum requirement, the model should fulfill the following
142: properties: (i) in the photosphere the model should, on average,
143: reproduce the properties of a typical sunspot; (ii) at the border,
144: the model should smoothly merge into a quiet-Sun non-magnetic
145: model atmosphere; (iii) there should be a possibility to choose
146: the profile of thermodynamic parameters at the sunspot axis;
147: Wilson depression should be taken into account; (iv) magnetic
148: field strength, inclination and the radius of the structure should
149: be adjustable; (v) the model should be easily extensible into an
150: arbitrary depth below the photosphere.
151:
152: There is a vast amount of works on magneto-static models reported
153: in the literature. Leaving apart small-scale flux tube models,
154: those for thick structures can be divided into those possessing a
155: current sheet \citep[\eg][]{Pizzo1990}, with a sharp magnetic-non
156: magnetic interface, and those with distributed currents
157: \citep[\eg][]{Pizzo1986}, showing a smooth transition. Without
158: discussing advantages and disadvantages of the both, we will
159: proceed here with current-distributed models.
160:
161: Present current-distributed models apply two different
162: philosophies. In the first category of models, the magnetic
163: structure is prescribed and the distribution of thermodynamic
164: variables is looked for to be in agreement with this structure.
165: This is the class of self-similar models, proposed by
166: \citet{Schluter+Temesvary1958} and then extended by, \eg,
167: \citet{Low1975, Low1980}. In the second category, the pressure
168: distribution is prescribed as the boundary condition at the axis
169: of the magnetic structure and in the far-away non-magnetic
170: atmosphere. Both, pressure and magnetic field, are iteratively
171: changed in the remaining points to reach an equilibrium situation
172: \citep{Pizzo1986}.
173:
174: From the point of view of the requirements set above, both classes
175: of models have advantages and disadvantages. The approach by
176: \citet{Pizzo1986} is more fruitful in the photosphere, since the
177: pressure distributions of the field-free and magnetized
178: atmospheres can be taken from observations and are relatively well
179: known. However, for deep sub-photospheric layers, the models that
180: can be taken as boundary conditions are scarce. More precisely,
181: the quiet-Sun non-magnetic pressure stratification can be taken
182: from helioseismological data, for example, from the standard solar
183: model of \citet{Christensen-Dalsgaard+etal1996}. As for the
184: sunspot axis, no precise data are available \citep[see,
185: however][]{Zhao+Kosovichev+Duvall2001, Kosovichev2002,
186: Couvidat+Birch+Kosovichev2006}. The model of \citet{Pizzo1986}
187: turns out to be very sensitive to the pressure deficit inside
188: sunspots and the method, in general, has poor convergence if the
189: simulation box is too deep and is very sensitive to the guess of
190: the pressure distribution at the sunspot axis. This makes the
191: method unsuitable for the purpose of our work.
192:
193: On the other side, the procedure proposed by \citet{Low1980} works
194: better in deep layers, where the gas pressure dominates over the
195: magnetic pressure. In the photosphere, where the plasma becomes
196: magnetically-dominated, negative pressures are frequently obtained
197: from the method of \citet{Low1980}. It is complicated to guess the
198: parameters of the magnetic field configuration in order to avoid
199: this problem. At the same time, if one wishes to extend the models
200: into the photosphere and higher layers, the magnetic field
201: strength is limited to rather low flux tube-like values, not
202: appropriate for sunspots \citep[see][]{Hanasoge2008,
203: Cameron+etal2008}.
204:
205: In this paper, we take advantage of both Pizzo-like and Low-like
206: approaches and propose a method to calculate the magneto-static
207: equilibrium of a thick sunspot-like structure with the properties
208: defined above. Below we describe the equations that allow to
209: successfully merge results from both methods and show examples of
210: MHS solutions for a wide range of parameters. The conclusions are
211: given in the last section.
212:
213:
214: %__________________________________________________________________
215: \section{Method}
216:
217: We solve the equilibrium force balance equation together with
218: divergence-free condition for the magnetic field:
219: %
220: \begin{eqnarray}
221: \label{eq:initial}
222: - & \vec{\nabla}P & + \rho\vec{g} +
223: \frac{1}{4\pi}(\vec{\nabla}\times\vec{B})\times\vec{B} = 0 \,, \\
224: \nonumber
225: %
226: & \vec{\nabla}\vec{B} & = 0 \,.
227: \end{eqnarray}
228: %
229: Following \citet{Pizzo1986}, the equations are solved in
230: cylindrical coordinates $(r,\phi,z)$ and axial symmetry is assumed
231: (\ie, all variables are independent of $\phi$).
232: %
233: Under these conditions the magnetic field vector can be
234: conveniently written in terms of the field line constant $u$:
235: %
236: \begin{equation}
237: \label{eq:bvec}
238: \vec{B}=\left( - \frac{1}{r}\frac{\partial{u}}{\partial{z}},
239: \frac{G(u)}{r}, \frac{1}{r}\frac{\partial{u}}{\partial{r}} \right)
240: \,,
241: \end{equation}
242: where $G(u)$ is a function related to the twist component of the
243: field.
244: %
245: The variable $u$ is used both in \citet{Pizzo1986} and
246: \citet{Low1980}.
247: %
248: The difference is that, in \citet{Low1980}, the analytical
249: expression for $u$ is postulated, while, in \citet{Pizzo1986}, the
250: shape of the field lines $u$ is looked for by solving iteratively
251: the equation of the force balance for given boundary conditions in
252: agreement with some pressure distribution.
253: %
254: Except for constants, the functional form of $u$ used by
255: \citet{Pizzo1986} as initial condition at the lower boundary of
256: the computational domain is exactly the same as the one postulated
257: by \citet{Low1980}.
258: %
259: Thus, both models can be joined in a natural way, assuming that
260: the deep layers of the model sunspot can be approximated by the
261: self-similar solution and the upper layers by the solution of
262: Pizzo.
263: %
264:
265: %dimensions of u is Re*Re*B_0
266:
267: In \citet{Low1980}, following the spirit of self-similar
268: solutions, the field line constant $u$ is expressed as a function
269: of one variable, $\varphi$:
270: %
271: \begin{equation}
272: u(r,z) = u(\varphi); \,\,\, \varphi = r^2 \cdot F(z); \,\,\,
273: F(z) = (z^2 + a^2)^{-1} \,,
274: \end{equation}
275: %
276: where $a$ is a constant parameter.
277: %
278: The field is untwisted and the azimuthal component, $B_{\phi}$, is
279: zero. This is equivalent to setting $G(u)=0$ in
280: Eq.~\ref{eq:bvec}. Using the above expression, Eq.~\ref{eq:bvec}
281: can be rewritten as a function of $\varphi$:
282: %
283: \begin{equation}
284: \label{eq:low} \vec{B}=\left( - r \frac{d F(z)}{d z} \frac{d u}{d
285: \varphi} ,\, 0,\, 2F(z)\frac{d u}{d \varphi} \right) \,.
286: \end{equation}
287: %
288: Following Low, the function $d u/d \varphi$ has to satisfy certain
289: normalizations in order to fulfill the force balance equation.
290: This leads to the following expression:
291: %
292: \begin{equation}
293: \label{eq:u} \frac{d u}{d \varphi} =B_0^L h^2 \cdot \exp(-\eta
294: \varphi) \,.
295: \end{equation}
296: %
297: Here, $B_0^L$ is a parameter that controls the magnetic field
298: strength and $h$ is a suitable length scale (note that, in the
299: original paper of Low, dimensionless variables are used, while
300: here we choose to use physical dimensions for all the variables).
301:
302:
303: Introducing Eq.~\ref{eq:u} into Eq.~\ref{eq:low}, the horizontal
304: and vertical components of the magnetic field vector in the Low's
305: model are written as:
306: %
307: \begin{equation}
308: \label{eq:brlow} B_r (r,z)=2 B_0^L \frac{(z-z_{\rm d})rh^2
309: }{((z-z_{\rm d})^2 + a^2)^2} \exp \left( \frac{-\eta r^2}
310: {(z-z_{\rm d})^2 + a^2} \right)
311: \end{equation}
312: %
313: \begin{equation}
314: \label{eq:bzlow} B_z (r,z) =2 B_0^L \frac{h^2}{(z-z_{\rm d})^2 +
315: a^2} \exp \left( \frac{-\eta r^2} {(z-z_{\rm d})^2+ a^2} \right)
316: \end{equation}
317: %
318: The parameter $z_{\rm d}$ is a reference height, where the
319: magnetic field is purely vertical. The latter expression is
320: directly comparable with the one used in Pizzo as a boundary
321: condition at the bottom boundary of the domain:
322: %
323: \begin{equation}
324: \label{eq:bzpizzo} B_z(r,z_0) = B_0^P \exp
325: \left(-\frac{r^2}{r_e^2} \right)
326: \end{equation}
327: %
328: Comparing these two expressions, we see that, if both models are
329: to be joined at some arbitrary height, $z=z_0$, the parameters of
330: the models should be related as:
331: %
332: \begin{equation}
333: \label{eq:ltp1} B_0^P = B_0^L \frac{2h^2}{(z_0-z_{\rm d})^2 + a^2}
334: \end{equation}
335: %
336: \begin{equation}
337: \label{eq:ltp2} r_e^2=((z_0-z_{\rm d})^2 + a^2)/\eta
338: \end{equation}
339:
340: Keeping this in mind, the model can be constructed following the
341: steps described below.
342:
343: \subsection{Step 1: Generation of a self-similar solution in deep layers}
344:
345: In deep sub-photospheric layers, we calculate a self-similar
346: solution for $\vec{B}$ after Eqs.~\ref{eq:brlow} and
347: \ref{eq:bzlow}. The pressure and density distributions with height
348: and radius are found from analytical expressions \citep[equations
349: 50 and 51 in][]{Low1980}. As a boundary condition at the right
350: boundary (field-free atmosphere) we take the pressure and density
351: from the model S of \citet{Christensen-Dalsgaard+etal1996}. The
352: choice of the field-free pressure and density is rather arbitrary,
353: however.
354: %
355: Because of azimuthal symmetry, the left boundary of the domain
356: corresponds to $r=0$, i.e., the axis of the magnetic structure.
357: The lower boundary is taken exactly at height $z_{\rm d}$, where
358: $B_r(r,z_{\rm d}) = 0$ at all distances $r$. In all the cases
359: presented here, $z_{\rm d} = -10$ Mm and the origin for the z-axis
360: is taken at the base of the photosphere.
361: %
362: Given pressure and density, the temperature distribution in the
363: model sunspot can be calculated using the equation of state either
364: in tabular form or the one for an ideal gas.
365: %
366: The parameters $\eta$, $a$ and $B_0^L$ can be chosen freely. An
367: additional free parameter is the height, $z_0$, that limits the
368: upper boundary of the self-similar model. Depending on the height
369: of the upper boundary, $B_0^L$ can be larger or smaller in order
370: to prevent from getting negative gas pressures.
371: %
372:
373:
374: The basic topology of the solution is given in Fig.~\ref{fig:low},
375: for the parameters indicated in the figure caption. These
376: parameters are chosen on purpose in order to demonstrate that the
377: method is able to deal with large field strengths.
378: %
379: The sunspot radius at the bottom boundary is roughly defined by $a
380: \eta^{-1/2}$ \citep{Low1980}. The inclination of the field at the
381: top boundary changes with the distance from the axis, from 0 to
382: about 70 degrees at the right-most point of the domain. The
383: magnetic field is concentrated inside the first 15 Mm from the
384: axis, being weak in the rest of the domain. The field strength
385: drops at the axis from 12 kG at $z=-10$ Mm to 4 kG at $z=-$1 Mm
386: depth.
387: %{\bf
388: The gas pressure is always above the magnetic
389: pressure.%}
390: %
391: %The plasma parameter $\beta$ is always above unity.
392: %
393: The Wilson depression is rather weak and the model sunspot is
394: almost thermally plane-parallel in deep layers, as follows from
395: the distribution of the acoustic speed, $c_S$. We do not have
396: criteria to judge how realistic is this description. Data of the
397: sub-photospheric distribution of the sound speed in sunspots are
398: scarce and uncertain \citep[see, however, time-distance analysis
399: results by][]{Zhao+Kosovichev+Duvall2001, Kosovichev2002,
400: Couvidat+Birch+Kosovichev2006}.
401:
402: %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
403: \begin{figure}
404: \centering
405: \includegraphics[width=9cm]{fig1.eps}
406: \caption{Topology of the Low solution with $a= 2h$, $h=3$ Mm,
407: $\eta=1.3$, $B_0=25000$ G and $z_0=-1$ Mm. Top: magnetic field
408: strength; middle: acoustic speed; bottom: log of the Alfv\'en
409: speed. White lines are magnetic field lines. Black lines with
410: labels are the contours of the ratio of the sound speed and the
411: Alfv\'en speed squared, $c_S^2/v_A^2$.} \label{fig:low}
412: \end{figure}
413: %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
414:
415:
416: \subsection{Step 2: Generation of potential solution in the overlaying atmosphere}
417:
418: Given the values of $\eta$, $a$, $B_0^L$, $z_{\rm d}$, and $z_0$
419: we calculate the initial parameters of the Pizzo model from
420: Eqs.~\ref{eq:ltp1} and \ref{eq:ltp2}. This gives us $B_0^P=4$ kG
421: and $r_e=9.4$ Mm.
422: %
423: We follow the same steps as in the original paper of
424: \citet{Pizzo1986} and start from computing the potential solution:
425: %
426: \begin{equation}
427: \label{eq:potencial} \frac{\partial^2 u}{\partial r^2} -
428: \frac{1}{r}\frac{\partial u}{\partial r} + \frac{\partial^2
429: u}{\partial z^2} =0
430: \end{equation}
431: %
432: The bottom boundary of the domain coincides with the top boundary
433: from the previous step and is located at $z=-1$ Mm, below the
434: photosphere.
435: %
436: The field line constant $u$ at the bottom boundary is approximated
437: by:
438: \begin{equation}
439: u=r_e^2 B_0^P \left(1-\exp \left(-\frac{r^2}{r_e^2} \right)
440: \right)/2
441: \end{equation}
442: %
443: At the left (sunspot axis) and top boundaries $u=0$ (vertical
444: field) and $u$ approaches a constant value at the right boundary
445: (horizontal field).
446: %
447: With this set of boundary conditions, the boundary value problem
448: posed by Eq.~\ref{eq:potencial} is solved by standard methods.
449:
450: %Describe what equations are solved. (laplas u=0). Define boundary
451: %conditions. (u is horizontal at the right boundary, supposing the
452: %discontinuity)
453:
454: \subsection{Step 3: Generation of magneto-static solution in the overlaying atmosphere}
455:
456:
457: The potential solution obtained in Step 2 is used as initial guess
458: in the integration of the complete force balance equation along
459: the magnetic field lines (equation 4 in the paper by Pizzo):
460: %
461: \begin{equation}
462: \label{eq:pizzo} \frac{\partial^2 u}{\partial r^2} -
463: \frac{1}{r}\frac{\partial u}{\partial r} + \frac{\partial^2
464: u}{\partial z^2} = -4 \pi r^2 \frac{\partial P(u,z)}{\partial u}
465: \end{equation}
466:
467: In order to start iterations we need an approximation for the
468: distribution of pressure along the magnetic field lines $P(u,z)$.
469: %
470: Following \citet{Pizzo1986} and \citet{Low1975} we take it of the
471: form:
472: %
473: \begin{equation}
474: P(u,z)=P_0(u)\exp\left(-\int^z_0 \frac{dz^{'}}{h(u,z^{'})} \right)
475: \,,
476: \end{equation}
477: %
478: where $P_0(u)$ is the gas pressure along the bottom boundary. The
479: function $h(u,z)$ is a scale height.
480: %
481: For a complete description of the problem, the representative
482: pressure distributions along the axis and in the field-free quiet
483: atmosphere need to be specified.
484: %
485: As field-free atmosphere, we use of the model S of
486: \citet{Christensen-Dalsgaard+etal1996}, smoothly joint to the
487: VAL-C model of the solar chromosphere
488: \citep{Vernazza+Avrett+Loeser1981}.
489: %
490: At the axis, we use the \citet{Avrett1981} model in the upper
491: layers. In deep layers, we take a model by
492: \citet{Kosovichev+etal2000}, obtained from helioseismic inversions
493: of the sound speed beneath sunspots. This model already takes into
494: account the Wilson depression, which is about 450 km. The complete
495: model on the axis can be shifted up or down, though, if smaller or
496: larger values of the Wilson depression are required.
497: %
498: The model has a cool region just below the surface and a hot one
499: below, down to about -10 Mm. For our purposes we take this model
500: starting from -1 Mm depth and, thus, the hot layer is not taken
501: into account.
502: %
503: Once these models are specified, we calculate a smooth transition
504: between them for the gas pressure $P(u,z)$ and scale height
505: $h(u,z)$ distributions, as given by \citet{Pizzo1986} in his
506: equations (13), (17) and (18). Then, the Eq.~\ref{eq:pizzo} is
507: iterated until a convergence criterion is reached.
508:
509:
510: %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
511: \begin{figure}
512: \centering
513: \includegraphics[width=9cm]{fig2.eps}
514: \caption{Topology of the Pizzo solution with $B_0^P=4$ kG and
515: $r_e=9.4$ Mm. Top: magnetic field strength; middle: acoustic
516: speed; bottom: log of the Alfv\'en speed. White lines are magnetic
517: field lines. White lines with labels are the contours of
518: $c_S^2/v_A^2$. Note that, for better visualization, the vertical
519: axis has been expanded.} \label{fig:pizzo}
520: \end{figure}
521: %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
522:
523: The model sunspot for the parameters given above is shown in
524: Fig.~\ref{fig:pizzo}. The magnetic field on the axis drops from 4
525: kG at $z=-1$ Mm to about 2 kG at $=2$ Mm, which is a rather large
526: value at this height. Due to such a large field strength, the
527: Alfv\'en speed exceeds $10^4$ \kms\ in the upper layers. The image
528: of the sound speed shows the presence of the Wilson depression
529: around $z=0$, \ie\ the temperature at the sunspot axis is smaller
530: than in the outside atmosphere at a given height. Note that at
531: higher layers, the effect is the opposite and the temperature
532: inside the sunspot is larger. {This effect is due to initial
533: distribution in the model atmospheres taken as boundary
534: conditions. The field lines are more inclined comparing to the Low
535: solution in Fig.~\ref{fig:low}.
536:
537:
538:
539: \subsection{Step 4: Concatenating the solutions}
540:
541: %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
542: \begin{figure}
543: \centering
544: \includegraphics[width=9cm]{fig3.eps}
545: \caption{Topology of the magnetic field lines before iterations
546: (dashed) and after a new equilibrium is reached (solid). }
547: \label{fig:cat}
548: \end{figure}
549: %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
550: %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
551: \begin{figure}
552: \centering
553: \includegraphics[width=9cm]{fig4.eps}
554: \caption{Topology of the complete solution with $B_0^L=25000$,
555: $a=2h$, $h=3$ Mm and $\eta=1.3$ ($B_0^P=4$ kG and $r_e=9.4$ Mm).
556: Top: magnetic field strength; middle: acoustic speed; bottom: log
557: of the Alfv\'en speed. White lines are magnetic field lines. White
558: lines with labels are the contours of $c_S^2/v_A^2$. }
559: \label{fig:complete}
560: \end{figure}
561: %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
562: %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
563: \begin{figure}
564: \centering
565: \includegraphics[width=9cm]{fig5.eps}
566: \caption{Distribution with radial distance (left panels) and with
567: depth (right panels) of the magnetic field strength, pressure,
568: ratio $c_S^2/v_A^2$ and the magnetic field inclination for the
569: sunspot with $B_0^L=25000$, $a=2h$, $h=3$ Mm and $\eta=1.3$
570: (corresponding to $B_0^P=4$ kG and $r_e=9.4$ Mm at $z= -1$ Mm).
571: The radial pressure distributions are normalized to their values
572: at the right boundary. } \label{fig:radius}
573: \end{figure}
574: %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
575:
576: Both solutions obtained in Step 1 and Step 3 are in MHS
577: equilibrium. In order to construct the complete model from deep to
578: high layers one has to put one model on top of the other.
579: %
580: However, despite $B_z$ at the bottom boundary of the Pizzo model
581: is calculated in agreement with $B_z$ at the top boundary of the
582: Low model, there is a discontinuity in the horizontal component of
583: the magnetic field.
584: %
585: This discontinuity can be appreciated in Fig.~\ref{fig:cat}, where
586: the field lines with a dashed style correspond to the two models
587: concatenated as they are.
588: %
589: The reason of this discontinuity is two-fold. On the one hand, the
590: physics of the solution changes abruptly from one model to
591: another, thus changing the gradients of the magnetic field, gas
592: pressure, \etc \ On the other hand, the boundary condition for the
593: field line constant $u$ is not the same in the both models. In the
594: case of Low model, there is no need to put a condition on $u$,
595: neither there is a possibility. The inclination of the magnetic
596: field lines at the right boundary is a consequence of the
597: parameters of the model and should not be necessarily horizontal.
598: %
599: Contrarily, in the case of the Pizzo model, we impose horizontal
600: magnetic field at the right boundary.
601: %
602: The dependence of $B_z$ on $r$ given by Eq.~\ref{eq:bzpizzo}
603: defines the vertical magnetic field strength but does not put
604: constraints on the horizontal field component.
605:
606: Thus, in order to obtain a smooth solution everywhere in the
607: domain, we repeat the Step 3 calculations for the complete model
608: sunspot. We take the pressure distributions at the axis and in the
609: field-free outside atmosphere from the joint model at all
610: heights. The boundary conditions for $u$ are the same as in the
611: Pizzo model. The distribution of $u$ at the bottom boundary is
612: taken from the Low model. Then we repeat the solution of
613: Eq.~\ref{eq:pizzo}. The resulted topology of the magnetic field
614: lines is plotted in Fig.~\ref{fig:cat} in solid line style. The
615: field lines in the final solution are more horizontal in low
616: layers, while they are more vertical in the upper layers.
617:
618:
619:
620: Fig.~\ref{fig:complete} gives the distribution of some parameters
621: in the complete model sunspot, at all layers. We can see that the
622: last iteration has re-distributed all the parameters compared to
623: the individual Low and Pizzo parts of the solution. In particular,
624: the magnetic field gradient at the axis is now more steep and the
625: field strength at high layers becomes lower. The field is in
626: general more inclined, being horizontal at the right hand side
627: domain boundary, consistent with our imposed boundary condition
628: there. The gas pressure is modified accordingly to maintain the
629: new force balance.
630:
631:
632:
633: Fig.~\ref{fig:radius} gives a more detailed view on the model spot
634: solution. It shows the distribution with radius (left panels) and
635: with depth (right panels) of some parameters of the sunspot
636: atmosphere. The field strength decreases rapidly with height at an
637: average rate of about 1 G/km at the axis. The magnitude of the
638: gradient decreases with height and with distance to the axis.
639: These gradients are in agreement with photospheric
640: spectropolarimetric observations \citep{Solanki2003}. The
641: magnitude of the pressure deficit inside the model sunspot
642: decreases with depth almost disappearing at about -2 Mm depth, in
643: accordance with our assumption of self-similarity of the MHS
644: solution at larger depths.
645:
646:
647: As can be seen from the radial pressure distribution, there is a
648: pressure excess observed at larger heights in the chromosphere at
649: some distance from the axis. This pressure excess would produce a
650: bright ring in the emergent intensity from the model sunspot and
651: it is present in the original model by \citet{Pizzo1986}. As shown
652: in the latter work, the bright ring can be removed by an improved
653: initial guess of the $P(u,z)$ distribution. It is unimportant for
654: the purpose of the present work since we only need an approximate
655: agreement between the average properties of the MHS solution and
656: the observed properties of sunspots. The magnetic field lines of
657: the model sunspot are inclined less than 30 degrees within the
658: first 10 Mm from the axis, which can be considered as the umbra.
659: Due to the boundary condition, the inclination changes gradually
660: becoming 90 degrees at the edge of the model, where the magnetic
661: field is already very weak.
662: %
663: The ratio between the sound speed and the Alfv\'en speed squared
664: (which gives the measure of the gas to magnetic pressure) changes
665: orders of magnitude from $10^6$ at $z=-10$ Mm to $10^{-6}$ at
666: $z=$2 Mm. Note that despite this, there is no problem with the
667: convergence of the solution.
668: %
669: %The plasma $\beta$ changes orders of
670: %magnitude from $10^6$ at $z=-10$ Mm to $10^{-6}$ at $z=$2 Mm.
671:
672:
673: In the next section, we give more examples of MHS solutions
674: comparing models obtained with various sets of parameters. In the
675: examples below we discuss the models calculated in a complete
676: domain from $z=-10$ to $z=2$ Mm.
677:
678: %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
679: \begin{figure}
680: \centering
681: \includegraphics[width=9cm]{fig6a.eps}
682: \includegraphics[width=9cm]{fig6b.eps}
683: \caption{Top panels: height dependence of the magnetic field and
684: pressure at the axis for the models with $a=2h$, $h=3$ Mm,
685: $\eta=1.3$ and $B_0^L$= 10000 (red line), 16000 (green line),
686: 25000 (blue line) and 40000 (magenta line) G. Bottom panel:
687: topology of the magnetic field lines for the same solutions (same
688: color coding). Contours of the magnetic field strength of $B=1000$
689: G are shown by dotted lines for each case. Horizontal solid lines
690: mark the levels of $c_S=v_A$. } \label{fig:height}
691: \end{figure}
692: %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
693:
694:
695: %__________________________________________________________________
696: \section{Examples}
697:
698:
699: {\bf Dependence on magnetic field strength.} Fig.~\ref{fig:height}
700: shows the magnetic field topology of models with different values
701: of the magnetic field strength (parameter $B_0^L$), all the other
702: parameters being exactly the same as previously.
703: %
704: As can be seen by comparing the different curves on the figures,
705: the resulting gas pressure stratifications only differ in the
706: highest layers by the amount of the pressure deficit.
707: %
708: The magnetic field topology is indistinguishable in all the cases.
709: %
710: This property originates from two effects. On the one side, the
711: self-similar solution in the bottom part of the domain scales with
712: magnetic field strength, i.e., the field line topology does not
713: depend on $B_0^L$.
714: %
715: On the other side, the Pizzo solution in the upper part is close
716: to potential, imposed by the solution at the bottom part. The
717: potential solution also scales with the magnetic field strength
718: and is independent of the thermodynamic properties.
719: %
720: Both effects lead to the independence of $B_0^L$ of the magnetic
721: topology of the final solution in the complete domain.
722: %
723: This is a useful property from the point of view of
724: helioseismology simulations. Using a set of models with different
725: magnetic field strength, but otherwise the same, the effects of
726: the magnetic field strength on waves can be checked independently
727: of the effects of the magnetic field inclination.
728:
729:
730: It should be noted that the above property originates only from
731: the particular choice of the parameters $a$ and $\eta$. This
732: choice produces $r_e$ large enough, so that the final solution in
733: the upper part of the domain approaches to potential and becomes
734: almost independent on the pressure distribution $P(u,z)$
735: \citep{Pizzo1986}.
736:
737: The models presented in this Section are available electronically
738: in FITS format from the Astrophysical Journal web site.
739:
740:
741:
742:
743: %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
744: \begin{figure}
745: \centering
746: \includegraphics[width=9cm]{fig7a.eps}
747: \includegraphics[width=9cm]{fig7b.eps}
748: \caption{Top panels: height dependence of the magnetic field and
749: pressure at the axis for the models with $a=2h$, $h=3$ Mm,
750: $B_0^L$= 25000 and $\eta=$ 1.3 (red line), 2.5 (green line) and
751: 3.5 (blue line). Bottom panel: topology of the magnetic field
752: lines for the same solutions (same color coding). Contours of the
753: magnetic field strength of $B=1000$ G are shown by dotted lines
754: for each case. Horizontal solid lines mark the levels of
755: $c_S=v_A$.} \label{fig:height_eta}
756: \end{figure}
757: %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
758:
759:
760: {\bf Dependence on $a$ and $\eta$.} Fig.~\ref{fig:height_eta}
761: shows the magnetic field topology of the models calculated with
762: different values of the parameter $\eta$ (Eqs.~\ref{eq:brlow} and
763: \ref{eq:bzlow}), the rest of the parameters being the same.
764: %
765: Note that, according to the above equations, the inclination of
766: the magnetic field is independent of $\eta$ in the Low model at
767: the bottom part of the domain. However, the initial radius of the
768: structure in the Pizzo part of the solution $r_e$
769: (Eq.~\ref{eq:ltp2}) depends on $\eta$, thus changing the
770: inclination of the magnetic field lines in the upper part of the
771: atmosphere. The final iteration performed in Step 4 takes that
772: into account, making the solution in the complete domain dependent
773: on $\eta$.
774:
775:
776: The change of $\eta$ produces two effects. By increasing $\eta$,
777: we decrease the magnetic field strength by a smaller amount than
778: by varying $B_0^L$, as in the previous example. At the same time
779: increasing $\eta$ produces an increase of the inclination of the
780: magnetic field lines in the solution in the complete domain. The
781: difference in the inclination is more pronounced in the deep
782: layers of the model.
783: %
784: The magnetic field topology of the solutions is different. The
785: gradient of the magnetic field at the axis is slightly larger for
786: larger values of $\eta$ in the sub-photospheric part of the model.
787: Again, this difference is produced after the final iteration in
788: Step 4 since the $B_z(r=0)$ given by Eq.~\ref{eq:bzlow} of the Low
789: part of the solution is independent of $\eta$. The pressure
790: distribution at the axis and the amount of the pressure deficit
791: are not very different between the given models. The models
792: presented in this Section are available electronically in FITS
793: format from the Astrophysical Journal web site.
794:
795:
796: Varying the parameter $a$ produces similar effects. The difference
797: is that, by varying $a$, we change mostly the curvature of the
798: magnetic field lines and the radius of the structure, not
799: affecting much the magnetic field strength.
800:
801:
802: %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
803: \begin{figure}
804: \centering
805: \includegraphics[width=9cm]{fig8a.eps}
806: \includegraphics[width=9cm]{fig8b.eps}
807: \caption{Top panels: height dependence of the magnetic field and
808: pressure at the axis for the models with $a=1.5h$, $h=3$ Mm,
809: $B_0^L$= 25000, $\eta=$ 3.5 and $z_0$=-1 Mm (red line), -2 Mm
810: (green line) and -3 Mm (blue line). Bottom panel: topology of the
811: magnetic field lines for the same models (same color coding).
812: Contours of the magnetic field strength of $B=1000$ G are shown by
813: dotted lines for each case. Horizontal solid lines mark the levels
814: of $c_S=v_A$. } \label{fig:height_zdeep}
815: \end{figure}
816: %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
817:
818:
819: {\sc \bf Dependence on $z_0$.} Another parameter introduced in our
820: modeling is the height where both solutions merge, $z_0$. Fig.
821: \ref{fig:height_zdeep} shows the topology of the magnetic field
822: lines and the pressure and density distributions with height at
823: the axis, for three models with different values of $z_0$. In this
824: example we take different values of $a$ and $\eta$ compared to
825: above examples (see figure caption) in order to produce a
826: structure with a smaller radius. This way, we show that the
827: procedure is robust and can produce magnetic structures with very
828: different properties.
829:
830: %
831: The magnetic field strength at the axis is almost independent of
832: the choice of $z_0$. The amount of the pressure deficit at the
833: near-surface layers increases with decreasing $z_0$ from $-1$ to
834: $-3$ Mm, extending to larger depths.
835: %
836: Note, however, that we can not shift the level of $z_0$ much
837: deeper than $-3$ Mm due to a poor convergence of the solution.
838: %
839: Despite the magnetic field strength is nearly the same, the
840: position of the $\beta=1$ level is different in all the solutions
841: due to the different amount of the pressure deficit.
842: %
843:
844: %
845: The inclination of the magnetic field lines is similar in the
846: central part of the model sunspots in the three solutions. At the
847: periphery, especially at larger depths, the field lines get more
848: inclined with decreasing $z_0$. Thus, the field is more
849: concentrated to the central part and the effective radius of the
850: structure is smaller.
851: %
852: The models presented in this Section are available electronically
853: in FITS format from the Astrophysical Journal web site.
854:
855:
856: %______________________________________________________________
857:
858: \section{Conclusions}
859:
860: In this study, we propose a method to construct a magnetostatic
861: structure with properties and size of a typical sunspot, from the
862: deep interior to the solar surface.
863: %
864: Previously published methods to construct such a model fail due to
865: a poor knowledge of the thermodynamic and magnetic parameters of
866: sunspots in sub-photospheric layers.
867: %
868: We make use of self-similar models in deep layers and show that
869: such models can naturally merge with models where the pressure
870: distribution is prescribed on the axis, as well as the field-free
871: atmosphere, allowing for a more realistic description of the
872: atmospheric layers of sunspots.
873: %
874: The procedure shows a rather good convergence and stability. By
875: changing the parameters of the solution, a set of models can be
876: produced with desired properties.
877: %
878: We suggest that these models may be used, among others, in
879: artificial helioseismology data simulations. Given a set of
880: models, a parametric study can be done, investigating the
881: influence of the topology and strength of the magnetic field of
882: sunspots on the parameters inferred by the local helioseismology
883: measurements in solar active regions.
884:
885:
886: \begin{acknowledgements}
887: This research has been funded by the Spanish
888: Ministerio de Educaci{\'o}n y Ciencia through projects
889: AYA2007-63881 and AYA2007-66502.
890: \end{acknowledgements}
891:
892: %\aareferences
893: \input{mslit.bbl}
894:
895: \end{document}
896: