1: \documentclass[12pt,preprint]{aastex}

2: 

3: \begin{document}

4: 

5: \title{Twisted flux tube emergence from the convection zone to the corona}

6: 

7: \author{Juan Mart\'inez-Sykora}

8: \and

9: \author{Viggo Hansteen} 

10: \and 

11: \author{Mats Carlsson }

12: \affil{Institute of Theoretical Astrophysics, University of Oslo, P.O. Box 1029 Blindern, N-0315 Oslo, Norway}

13: 

14: \newcommand{\myemail}{juanms@astro.uio.no}

15: \newcommand{\viscous}{\underline{\underline{\tau}}}

16: \newcommand{\resistive}{\underline{\underline{\eta}}}

17: 

18: \begin{abstract}

19: 

20: Three dimensional numerical simulations of the emergence of a horizontal magnetic flux tube 

21: with different twist levels are carried out in a computational domain spanning the upper 

22: layers of the convection zone to the lower corona. We use the {\em Oslo Staggered Code} (OSC) 

23: to solve the full MHD equations with non-grey and non-LTE radiative transfer and thermal 

24: conduction along the magnetic field lines. The emergence of the

25: magnetic flux tube input at the bottom boundary into a weakly magnetized atmosphere is presented. 

26: The photospheric and chromospheric response is described with magnetograms, synthetic continuum 

27: images at different wavelengths, as well as synthetic Ca~{\sc ii} H-line images, and velocity field 

28: distributions. 

29: 

30: The emergence of a magnetic flux tube into such an atmosphere

31: results in varied atmospheric responses. 

32: In the photosphere the granular size increases when the flux tube approaches from 

33: below, as has been reported previously in the literature. In the convective overshoot region some

34: 200~km above the photosphere adiabatic expansion produces cooling, darker regions with 

35: the structure of granulation cells. We also find collapsed granulation in the boundaries 

36: of the rising flux tube. Once the flux tube has crossed the photosphere, bright points related 

37: with concentrated magnetic field, vorticity, high vertical velocities and heating by compressed 

38: material are found at heights up to 500~km above the photosphere. At greater heights in the 

39: magnetized chromosphere, we find that the rising flux tube produces a large, cool, dim, magnetized

40: bubble that tends to expel the usual chromospheric oscillations. In addition the rising flux tube

41: dramatically increases the chromospheric scale height, pushing the transition region and corona

42: aside such that the chromosphere extends up to 6~Mm above the photosphere in the center of the

43: rising flux tube. The emergence of magnetic flux tubes through the photosphere and subsequent

44: rise through the chromosphere up to the bottom of the lower corona is a relatively slow process,

45: taking of order 1~hour to complete. In later papers we intend to discuss the merging of the 

46: newly emerged field with the pre-existing coronal field and its diagnostic consequences.

47: 

48: \end{abstract}

49: 

50: \keywords{Magnetohydrodynamics MHD ---Methods: numerical --- Radiative transfer --- Sun: atmosphere --- Sun: magnetic field}

51: 

52: \section{Introduction}

53: 

54: One of the stated goals of the newly launched Hinode satellite is to follow

55: the evolution of magnetic flux from the moment of emergence through the photosphere and

56: into the chromosphere and corona. Such observations come at a timely moment as they

57: should prove a fertile testing ground for the 3D numerical models of flux emergence that

58: are now becoming available due to the technological development of massively 

59: parallel computers and algorithms to utilize these.

60: 

61: A great variety of simplified simulations have been developed in order

62: to study the different phases of flux emergence. Originally, early simulations 

63: were performed in 1D due to computational limitations. 

64: Studies of rising flux tubes were carried out to gain an understanding of 

65: the emergence latitudes of active regions, their tilt angles as well as asymmetry between 

66: leading and following polarities and magnetic forces such as tension 

67: \citep{moreno1986,fan1993,Caligari1995}.  The early 80's saw the appearance

68: of 2D simulations where \citet{Shussler1979} and

69: \citet{longcope1996} found that an emerging magnetic flux tube without twist 

70: is fragmented. \citet{moreno1996} investigated the effect

71: of the surrounding flows on the flux tube's rise to the surface and

72: demonstrated that a minimum twist in the tube was required in order to

73: suppress the conversion of the tube into vortex pairs \citep{lites1995}.

74: \citet{emonet2001} studied the Von K{\'a}rm{\'a}n vortex street 

75: generated in the boundary layer of the tube \citep{Dorch+Nordlund1998}.

76: In the 90's 2D simulations were extended to 3D. \citet{fan1998}

77: investigated flux tubes moving through an adiabatic stratification 

78: found

79: that the degree of twist must be large enough to avoid fragmentation yet small 

80: enough to avoid the kink instability.  There are also other 3D emergence 

81: models that while not including radiative transfer, nor any detailed 

82: treatment of convection cells below the photosphere, have added 

83: schematic descriptions of the chromosphere and corona

84: \citep{matsumoto1998,fan2001,magara2004,Magara:2006lr,manchester2004,

85: archontis2004} in order to gain a handle on the effects to be expected.

86: 

87: On the other hand, in the last decade more realistic 3D MHD models of active

88: region flux emergence have been developed up to the photosphere. 

89: \citet{cheung2007} consider flux emergence through a convective layer 

90: including detailed photospheric radiative transfer, but excluding the layers above

91: the photosphere. Other simulations involve solar surface magnetoconvection

92: \citep{Stein:2002fk,Stein:2006qy,Stein:2003uq,Vogler:2005fj}, but do not consider

93: flux emergence from below.

94: 

95: The description of the outer layers of the solar atmosphere are also growing

96: in sophistication. \citet{Gudiksen:2005lr} recently performed realistic 

97: simulations of the solar corona driven by a stochastic approximation of 

98: granulation. \citet{Hansteen2004,Hansteen+Carlsson+Gudiksen2007} and \citet{abbett2007} 

99: studied 3D MHD simulations with a 

100: convective layer below the photosphere and included upper atmosphere (chromosphere,

101: transition region and lower-corona) with the radiative losses in the

102: quiet Sun and conduction along the magnetic field treated in detail.

103: 

104: In this paper our goal is to study horizontal flux tube emergence from the 

105: upper-convection zone to the lower corona in a realistic 3D MHD model. The model is 

106: designed to have a corona driven by magnetoconvection. The radiative

107: losses from the photosphere and lower chromosphere are computed in a 

108: non-grey manner and scattering in the chromosphere is handled to a certain 

109: approximation \citep{Nordlund1982,Skartlien2000}.  For the

110: upper-chromosphere and corona we assume effectively thin radiative losses.

111: A flux tube is injected into the bottom boundary of an atmosphere that contains 

112: a preexisting but weak magnetic field. We follow the evolution of the flux tube as

113: it rises through the various atmospheric layers: How much flux enters

114: the differing regions? When do we see observational signatures of flux emergence?

115: Where and how does the reconnection process happen? How does this affect chromospheric and 

116: coronal dynamics and structure? are questions we will discuss.

117: 

118: In section~\ref{sec:equations} we summarize how the MHD equations with

119: radiative transfer, thermal conduction, viscosity and resistivity 

120: are solved by the {\em Oslo Staggered Code}.  

121: Section~\ref{sec:condition} shows the initial and boundary

122: conditions of our numerical box for the different simulations. We

123: describe in more detail how the magnetic flux tube is input in the bottom

124: boundary. The results of our simulations are described in section~\ref{sec:results}

125: centered on the different events observed in the different layers of the

126: simulation box. We start with the photosphere, then treat the reverse granulation layer,

127: the end of the overshooting and beginning of the chromosphere and 

128: finally the upper chromosphere/transition region and lower corona. 

129: 

130: \section{Equations and Numerical Method}\label{sec:equations}

131: 

132: In order to model the rise of magnetic flux tubes through the upper convection layer and their 

133: emergence through the photosphere and into the chromosphere and corona we solve the equations 

134: of MHD using the {\it Oslo Stagger Code} (OSC):

135: 

136: \begin{equation}

137: \frac{\partial \rho}{\partial t} + \nabla ({\rho \bf u})=0 \label{cont}

138: \end{equation}

139: \begin{equation}

140: \frac{\partial {\bf u}}{\partial t}  + ({\bf u} \nabla) {\bf u} + \frac{1}{\rho} \nabla (P+\viscous) =

141: 	+ \frac{{\bf J} \times {\bf B}}{\rho}+{\bf g},\label{moment} 

142: \end{equation}

143: \begin{equation}

144: \frac{\partial (e)}{\partial t}  + \nabla (e{\bf u})  + {\bf u} \nabla P=

145: 	 \nabla {\bf F}_r+\nabla {\bf F}_c+Q_{Joule}+Q_{visc},\label{ener} 

146: \end{equation}

147: \begin{equation}

148: \frac{\partial{\bf B}}{\partial t}=\nabla \times ({\bf u} \times {\bf B})-\nabla\times (\resistive {\bf J}), \label{induction}

149: \end{equation}

150: 

151: \noindent where $\rho$ represents the mass density, ${\bf u}$ the fluid velocity, $P$ the

152: gas pressure, ${\bf J}$ the current density, ${\bf B}$ the magnetic field, ${\bf g}$ 

153: gravitational acceleration, and $e$ the internal energy. The viscous stress tensor is 

154: written $\viscous$ and the resistivity $\resistive$. ${\bf F}_r$ represents the radiative flux, 

155: ${\bf F}_c$ represents the conductive flux, and $Q_{Joule}$ and $Q_{visc}$ are the joule 

156: heating and viscous heating respectively. 

157: 

158: These equations are solved using an extended version of the numerical code described

159: in \citet{Dorch+Nordlund1998,Mackay+Galsgaard2001} and in more detail

160: by Nordlund \& Galsgaard at http://www.astro.ku.dk/$\sim$kg and 

161: \citet{Hansteen+Carlsson+Gudiksen2007}.

162: Extended in this case means the inclusion of

163: thermal conduction along the magnetic field, non-grey, non-LTE radiative transfer and 

164: characteristic boundary conditions on the upper and lower boundaries. The latter 

165: allow the passage of waves out of the computational box while at the same time allowing 

166: the specification of an input magnetic field and entropy at the lower boundary.

167: 

168: In short, the code functions as follows: The variables are represented

169: on staggered meshes, such that the density and the internal energy

170: are volume centered, the magnetic field components and the momentum densities

171: ${\bf p}=\rho{\bf u}$ are face centered, while the electric field ${\bf E}$

172: and the current density are edge centered. A sixth order accurate 

173: method involving the three nearest neighbor points on each side is

174: used for determining the partial spatial derivatives. In the cases where

175: variables are needed at positions other than their defined positions 

176: a fifth order interpolation scheme is used. The equations are stepped 

177: forward in time using the explicit 3rd order predictor-corrector 

178: procedure defined by \citet{Hyman1979}, modified for variable time steps. 

179: In order to suppress numerical noise, high-order artificial diffusion is added both

180: in the forms of a viscosity and in the form of a magnetic diffusivity.

181: 

182: The radiative flux divergence from the photosphere and lower chromosphere is obtained

183: by angle and wavelength integration of the transport equation. Assuming isotropic opacities 

184: $\chi_\lambda$ and emissivities we find

185: \begin{equation}

186: \nabla{\bf F}_r=4\pi\int_\lambda\epsilon_\lambda\chi_\lambda(B_\lambda-J_\lambda)d\lambda,

187: \end{equation}

188: 

189: \noindent where $\epsilon_\lambda$ is the photon destruction probability. 

190: The method of solving the transport equation assumes opacities are in LTE and coherent

191: scattering. \citet{Nordlund1982} devised a technique of solution based on group mean opacities, 

192: in which the spectrum is divided into four bins representing strong, medium and weak lines 

193: in addition to the continuum. The transfer equation is formulated for these bins and a group 

194: mean source function is calculated for each bin. These

195: source functions contain an approximate scattering term and an exact contribution from thermal

196: emissivity. The resulting 3D scattering problems are solved by iteration based on a one-ray

197: approximation in the angle integral for the mean intensity as developed by \citet{Skartlien2000}.

198: 

199: For the upper chromosphere and corona we assume optically thin radiative losses such that:

200: 

201: \begin{equation}

202: \nabla{\bf F}_r\propto \rho^2f(T)e^{-\tau}.

203: \end{equation}

204: 

205: \noindent where $f(T)$ is the optically thin radiative loss function based on the coronal

206: approximation and atomic data collected in the HAO spectral diagnostics package 

207: \citep{HAO_Diaper1994}; $f(T)$ is based on the elements hydrogen, helium, carbon, oxygen, neon 

208: and iron. The $e^{-\tau}$ term prevents this term from having an effect in the

209: deep photosphere, where we assume an ``optical depth'', $\tau$, proportional to the gas pressure.

210: 

211: In the mid and upper chromosphere we include non-LTE radiative losses from hydrogen continua, hydrogen

212: lines and lines from singly ionized calcium. These losses are calculated from

213: 

214: \begin{equation}

215: \nabla{\bf F}_r\propto N_e\rho f_i(T)\epsilon_i(m_c)

216: \end{equation}

217: 

218: \noindent where $N_e$ is the electron density, $f_i(T)$ is a pre-calculated function based on collisional

219: excitation and radiative de-excitation, separate for hydrogen and calcium, and $\epsilon_i(m_c)$ is

220: the non-LTE escape probability as a function of column mass ($m_c$). This escape probability 

221: function is calculated from 1D dynamical chromospheric models in which the radiative losses

222: are computed in detail (Carlsson \& Stein 1992, 1995, 1997, 2002).

223: \nocite{Carlsson+Stein1992}

224: \nocite{Carlsson+Stein1995}

225: \nocite{Carlsson+Stein1997}

226: \nocite{Carlsson+Stein2002}

227: In addition to these optically thin losses in the

228: upper atmosphere we have added an {\it ad hoc} heating term to prevent the 

229: atmosphere from cooling much below $2000$~K in the upper chromosphere.

230: 

231: The Courant condition for a diffusive operator such as that describing thermal conduction

232: scales with the square of the grid size $\Delta z^2$ instead of with $\Delta z$ as applies

233: to the magneto-hydrodynamic operator. 

234: This severely limits the time step $\Delta t$ the code can be stably run at. Our 

235: solution to this problem is to proceed by operator splitting such that the operator advancing 

236: the variables in time is $L=L_{hydro}+L_{conduction}$. Thus the conductive part of the energy 

237: equation is solved by discretizing.

238: 

239: \begin{equation}

240: {\partial e\over\partial t}=\nabla{\bf F}_c={-\nabla\kappa_\parallel\nabla_\parallel T}

241: \end{equation} 

242: 

243: \noindent using the Crank-Nicholson method and solving the resulting implicit problem using

244: a multi-grid solver. 

245: The formulation used in the code described here is based on the method used by Malagoli, 

246: Dubey, Cattaneo as shown at 

247: {\tt http://astro.uchicago.edu/Computing/On\_ Line/cfd95/camelse.html}, but extended to 

248: 3D and to confining conduction to follow the magnetic field lines.

249: 

250: The units employed in this works are in SI. Sometimes we refer the distance in Mm ($10^6$m) and the time in hs ($10^2$~s).

251: 

252: \section{Initial and boundary conditions}\label{sec:condition}

253: 

254: The five models described here are run on a grid of $256\times 128\times 160$

255: points spanning $8\times 4\times 16$~Mm$^3$ and $16\times 8\times 16$~Mm$^3$.  

256: At these resolutions the models have been run for roughly one hour solar time.

257: 

258: The average temperature at the bottom boundary is maintained by setting 

259: the entropy of the fluid entering the computational domain. 

260: The bottom boundary is otherwise open, allowing fluid to enter and leave as required. 

261: The upper boundary is set so that the temperature gradient is zero; no

262: conductive heat flux enters or leaves the computational domain through the top 

263: boundary. 

264: 

265: Of course, without coronal heating the corona would cool on a time scale of roughly 

266: an hour (depending on the magnetic field topology), when no heat flux enters the model

267: through the upper boundary. We have therefore seeded the 

268: initial model with a magnetic field in which sufficient stresses can be built up

269: to maintain coronal temperatures in the upper part of the computational domain, as 

270: previously shown to be feasible by \citet{Gudiksen+Nordlund2004}. 

271: The initial field was obtained by semi-randomly spreading some $20-30$ positive and negative 

272: patches of vertical field at the bottom boundary, then calculating the potential field

273: that arises from this distribution in the remainder of the domain. Stresses sufficient 

274: to maintain a minimal corona are built up by photospheric motions after roughly $20$~minutes solar 

275: time.

276: 

277: At the top boundary the hydrodynamic variables (aside from the temperature)

278: and the magnetic field are set by characteristic extrapolations. This method hinders 

279: most of the reflections that are due to the presence of the upper boundary. 

280: No joule heating is added in the top five computational zones.

281: 

282: \subsection{Initial model}

283: \label{sec:initial-model}

284: 

285: The initial atmosphere, into which magnetic field is injected, includes the 

286: upper convection zone, the photosphere, chromosphere, transition region, 

287: and corona. In the left panel of figure~\ref{fig:initsetup} we show the lower 12~Mm (of 16~Mm) 

288: minimum, maximum, and horizontally averaged temperatures and densities as a function of

289: height $z$. 

290: At the base of the model, $1.4$~Mm below the photosphere, the average temperature 

291: is some $16\,200$~K and average density $6.7\times10^{-3}$~kg m$^{-3}$. With increasing

292: height the temperature falls close to adiabatically up to the photosphere, which is located 

293: at $z\approx0$~Mm, where the temperature gradient becomes much steeper and the average 

294: temperature falls to $6\,000$~K. The average density falls roughly $1.5$ orders of magnitude

295: to $2.5\times 10^{-4}$~kg m$^{-3}$ in the same height range. 

296: Quantities vary much more in the chromosphere, which extends from the top of the photosphere 

297: up to a height of some 4~Mm in some locations. The chromospheric temperature varies from 

298: a low of $2000$~K up to almost $10\,000$~K in the strongest shocks and in regions of strong 

299: magnetic field.

300: The density also has a high horizontal contrast in the chromosphere while at the same time 

301: falling by $7-8$ orders of magnitude vertically. The location of the transition region varies 

302: from $1.5$ to $4$~Mm above the 

303: photosphere. In this region the temperature rises rapidly to coronal values of roughly 1~MK, 

304: while coronal densities are of order $10^{-12}$~kg m$^{-3}$ in this model. 

305: 

306: In the right panel of 

307: figure~\ref{fig:initsetup} we plot the minimum, maximum, and horizontally averaged gas and

308: magnetic pressures as a function of height. 

309: The average gas pressure is $8.7\times 10^5$~Pa at the bottom boundary while the average 

310: magnetic pressure at the same depth is much smaller at $1.6\times 10^3$~Pa. The ratio between

311: these is still large in the photosphere, but note that the maximum magnetic pressure is

312: as large as the maximum gas pressure here; photospheric motions can concentrate the magnetic

313: field to pressures that equal or exceed photospheric pressure in intergranular lanes. 

314: The average plasma $\beta$ (ratio of gas pressure to magnetic pressure) remains greater 

315: than one up to some $600$~km above the 

316: photosphere, but the $\beta=1$ plane is quite corrugated and can extend up to $z=1000$~km. 

317: This means that the upper chromosphere, transition region and lower corona all are 

318: low $\beta$ regions. In the initial model, there is little net magnetic 

319: flux and the field is small at great heights, we therefore find some high $\beta$ regions 

320: at heights greater than $z=8$~Mm or so. 

321: 

322: The numerical box and the axes used are shown in figure~\ref{fig:initbox}. The color scale 

323: represents the temperature. The red lines are the magnetic field lines in the initial background

324: field. Note that field lines extending into the corona are mainly rooted in regions that cover 

325: all $y$ values and around $x=7$~Mm and around $x=14$~Mm. In addition we show planes of temperature

326: at four heights: in the photosphere $z=10$~km, in the reverse granulation $z=234$~km, at

327: the layer $z=458$~km, and in the chromosphere at $z=900$~km. We will show details

328: of how the initial atmosphere reacts to flux emergence at these heights later in this paper.

329: The green lines show the magnetic field lines of the flux tube coming in from the bottom boundary.

330: 

331: \subsection{Injection of magnetic flux}

332: \label{seq:injection}

333: 

334: We introduce a magnetic sheet or magnetic flux tubes into the lower boundary of the model 

335: described in section~\ref{sec:initial-model}. The OSC uses a numerical method that, in principle,

336: does not change $\nabla \cdot{\bf B}$ in time. The solenoidal condition must be enforced on 

337: the boundary to ensure that no magnetic monopoles are introduced. This is implemented by applying 

338: the boundary condition to the {\it electric} field, the staggered mesh will then enforce 

339: $\nabla\cdot{\bf B}=0$ to the numerical accuracy of the operators at the boundary.

340: 

341: The magnetic field variation at the boundary is defined by:

342: 

343: \begin{equation}

344: \frac{\partial{\bf B}}{\partial t}=\nabla \times {\bf E},

345: \end{equation}

346: \noindent where, for example for the $x$ component of the electric field we have

347: \begin{equation}

348: E^n_x = E_x + \frac{\Delta (B_y)}{\tau}\Delta z

349: \end{equation}

350: \noindent where $E^n_x$ is the new {\it x} component of the electric field at the boundary,  

351: $\tau$ is the time step size, and $\Delta z$ is the vertical cell size, 

352: $\Delta (B_y) = B^n_y - B_y$ is the difference between the value of the magnetic field we would

353: like to impose at the boundary, $B^n_y$, and the current boundary field $B_y$. 

354: 

355: We have run models in which a constant horizontal field in the $y$ direction has been injected, 

356: and several models where a flux tube is introduced. The flux tube is horizontally rectilinear 

357: with twisted magnetic fields lines. The expression for the magnetic field 

358: has a structure given by

359: 

360: \begin{eqnarray}

361: {\bf B}_{long}&=& B_o\, \exp \left(-\frac{r^2}{R^2}\right)\, {\bf e}_{z}\label{eq:blong}\\

362: {\bf B}_{trans}&=&B_{long}\,r\,q\,{\bf e}_{\phi} ,\label{eq:btrans}

363: \end{eqnarray}

364: 

365: \noindent where  $r=\sqrt{(x-x_o)^2+(z-z_o)^2}$ is the radial distance to the center of the 

366: tube that has radius $R$. ${\bf B}_{long}$, ${\bf B}_{trans}$ are the longitudinal and 

367: transversal magnetic fields in cylindrical coordinates respectively. The 

368: parameter $q$ is used by \citet{linton1996} and \citet{fan1998} to define the twist of the 

369: magnetic field. Following \citet{mark2006}, we define a new twist parameter, $\lambda$, as

370: 

371: \begin{eqnarray}

372: \lambda\equiv q\,R. \label{lambda}

373: \end{eqnarray}

374: 

375: As the flux tube enters the computational box, the height of the center of the tube 

376: ($z_o$) changes in time. The speed of flux tube, 

377: $({dz_o/dt})$, is set to the average of the velocity of plasma inflow at the boundary 

378: in the region where the magnetic flux tube is located each time step.

379: 

380: The field defined by equations~\ref{eq:blong} and \ref{eq:btrans} is easily seen to be 

381: a horizontal axisymmetric magnetic flux tube in which the longitudinal field has

382: a gaussian profile in the radial direction.

383: 

384: \section{Results}\label{sec:results}

385: 

386: We have carried out five simulations with different twist and magnetic field strength in order 

387: to study the effects of flux emergence in the photosphere and in the chromosphere. A summary of the 

388: runs completed is shown in table \ref{tab:runs}. In describing

389: the reaction of the atmosphere to the introduction of new magnetic flux we will specifically

390: concentrate on four layers placed at heights of $z=10, 235, 450$, and $900$~km represented

391: in the figure \ref{fig:initbox}. 

392: 

393: We find that all runs show a similar series of events after the emerging magnetic flux

394: pierces the photosphere. These events, that occur at roughly the same time in all models, 

395: are summarized in table \ref{tab:events}, which shows the processes observed in the 

396: simulations ordered in time. These processes are described in the following sections according 

397: to where they happen.

398: 

399: In the following sections all the figures shown correspond to

400: simulation A4 unless otherwise noted, both because it has run for a long solar 

401: time ($T=3200$~s) and as it has the most pronounced processes as a result of flux tube 

402: emergence. 

403: 

404: \subsection{Photosphere}\label{sec:photosphere}

405: 

406: The photosphere is located at $z\approx0$~Mm and is, up to the time when 

407: the emerging flux penetrates, as described in section~\ref{sec:initial-model},

408: with a cellular temperature structure similar to that shown in the upper left 

409: panel of figure~\ref{fig:intphot}. As the flux tube approaches the photosphere from 

410: below the cells immediately above the flux tube expand as seen in the upper 

411: right panel of figure~\ref{fig:intphot}. Eventually, horizontal field pierces the 

412: photospheric surface in the granular cell centers as seen in the second panel of the 

413: right column of figure~\ref{fig:fieldphot}. The horizontal field is rapidly moved to the

414: intergranular lanes as a result of the granular flow (figure~\ref{fig:fieldphot} third row). 

415: The granular cells where the field penetrates remain larger than in the undisturbed 

416: state as the flux tube passes through the photosphere. 

417: At some point, the expansion lowers the temperature of the

418: expanding cells through adiabatic cooling. At later stages a small amount of

419: the magnetic flux that pierces the photosphere is returned to the convection zone.

420: The photospheric evolution of emerging flux as described above has already been 

421: extensively studied previously and reported in the literature e.g. \citet{cheung2007} and references 

422: cited therein.

423: 

424: In all the simulations we find that 

425: approximately 2100~s passes from the time the magnetic flux tube crosses the 

426: photosphere until the time the photosphere again appears to regain its ``normal'' state, 

427: {\it i.e.} until the photospheric again is similar to the initial state before 

428: the flux tube passes through (see figure~\ref{fig:intphot}). Let us study the figures

429: in more detail.

430: 

431: In figure~\ref{fig:3dpho} we show two 3D images of the A4 simulation 

432: from two different viewpoints; from above and from below the photosphere as seen at 

433: time $t=1700$~s. The figure shows how the temperature structure of the granular cells 

434: has been changed by the presence of the flux tube, and in addition how the structure 

435: of the flux tube itself has been modified. The tube has suffered an expansion and 

436: splitting, gaining an irregular structure due to the movements of the convective cells. 

437:  

438: The magnetic field evolution in the photosphere is shown in the figure

439: \ref{fig:fieldphot}. The maximum magnetic field-strength in the photosphere

440: varies between 850~G and 1060~G. However, during the first hectosecond (hs)

441: after the flux tube enters the photosphere the maximum rises to between 

442: 970~G and 1030~G. Later, the magnetic field is concentrated by the granular flow

443: and the maximum rises to some 1100~G . The increased maximum magnetic 

444: field-strength is observed 

445: from 1700~s until the simulation ends, in some instants rising to 1700~G 

446: during the last few hectoseconds of the simulation. The magnetic field is almost 

447: horizontal and centered in the cells when the flux tube starts to cross 

448: the photosphere. The magnetic field is confined by the fluid and 

449: displaced to cell edges at which time its orientation becomes vertical. 

450: After most of the original tube has crossed the photosphere,  some remaining 

451: magnetic flux that has remained below the photosphere appears. This remaining

452: flux is small compared to the main tube's emergence and appears far away on either side of 

453: the central tube axis. The emergence process for this remaining flux is similar 

454: to the main tube's emergence, in that almost horizontal field rises in the cell 

455: center to be subsequently transported to the intergranular lanes where it

456: becomes vertical as also reported earlier by 

457: \citet{cheung2007} (see the last row of figure~\ref{fig:fieldphot}).

458: 

459: The intensity in the opacity bin with smallest opacity, designed to mimic the continuum, 

460: is shown in figure \ref{fig:intphot} along with contours of the magnetic field strength 

461: at height $z=10$~km at 6 different instants in time; at 600, 1100, 1210, 1660, 

462: 2200, and 3220~s.

463: The evolution of the intensity is very similar to that of the temperature:

464: Granular cells, located where the flux tube emerges, become larger and darker 

465: as they expand. These large cells cool as

466: their size increases due both to radiative looses and to adiabatic expansion. 

467: After the tube has crossed the photosphere the granulation pattern returns to

468: normal, roughly at times greater than 2100~s.

469: 

470: The temperature range before the tube crosses the photosphere is 

471: [6.0,10.2]~kK. While the upper value of the photospheric temperature remains 

472: the same, the lower limit varies: when the tube is just below the photosphere 

473: it rises to 6.2~kK, thereafter decreasing to 5.6~kK as large granulations cells 

474: cool by expansion (at time 1300~s). In the latter stages of the simulation

475: the lower temperature stabilizes at 5.8~kK.

476: 

477: The range in vertical velocity is roughly $[-6,5.5]$~km s$^{-1}$ where positive means

478: upwards flow. It is well known that downflows in the intergranules are

479: faster than the upward granular flow. 

480: However, it appears that 

481: the maximum and minimum vertical velocity 

482: increase after time 1100~s, {\it i.e.}

483: during and after the tube crosses the photosphere. 

484: The maximum upward velocity increases to 8~km s$^{-1}$ when the 

485: tube is in the photosphere, {\it i.e.} from 1100~s to 1600~s. 

486: Later the maximum returns to normal values, but still shows large variations in time. 

487: The maximum horizontal velocity oscillates between [6,7]~km s$^{-1}$ except at 

488: time 1350~s when there is a big jump to $10$~km s$^{-1}$, that lasts until

489: $t\approx 2000$~s. We do not see any remarkable patterns in the extrema of vertical

490: velocity before, during or after the tube crosses the photosphere.

491: 

492: It is worth remarking that we see several examples of collapsed cells 

493: \citep[see][]{Skartlien:2000lr} visible near the boundary of the flux tube as it 

494: crosses the photosphere (see the top-right at time 1100~s and bottom left at time 1210~s panels

495: at $x\approx 3$~Mm and at $x\approx 5$~Mm, the blue line of the figure 

496: \ref{fig:intphot}). 

497: Squeezing of the granule occurs as the adjacent plasma

498: expands during the rise of the flux tube and by the tendency for downflows 

499: surrounding a small granule to merge in the boundary layers. 

500: There are also bright points visible after the tube crosses the

501: photosphere, as seen in the lower left panel of the figure~\ref{fig:intphot}

502: at position $(x,y)\approx (6.5,2.7)$~Mm at time 2200~s or the bottom-right

503: panel at the position $7,3.8$~Mm. At that position the temperature is low

504: at photospheric height, but the magnetic field is strong and the decreased gas density 

505: (and thus photospheric opacity) 

506: allows us to see deeper into the atmosphere where the temperature is higher

507: \citep[see also][]{Shelyag:2004lr,Carlsson+Stein+Nordlund+Scharmer2004}.

508: 

509: In figure~\ref{fig:fluxret} we show the evolution of the mean of

510: magnetic flux of simulations A1, A2, A3 and A4 by area.  

511: We define the mean of the magnetic flux per unit of area between 2 heights by:

512: 

513: \begin{equation}

514: <\phi>=\int^L_0\left(\int^{L_h}_0 \int^u_l {\bf B} d{\bf S} \right)dn/(L |{\bf S}|) \label{eq:phi1}

515: \end{equation}

516: 

517: \noindent where $\phi$ is the mean of magnetic flux integrated in the

518: area in the vertical plane between $z=l$ and $z=u$  in the vertical component of ${\bf S}$,{\it i.e.}

519: $L_z=u-l$ and

520: along the whole horizontal component of ${\bf S}$, {\it i.e.} from $0$ to $L_h$. The mean is 

521: carried out

522: along the perpendicular component of ${\bf S}$, {\it i.e.}

523: it is defined as the integral over the line perpendicular to the 

524: surface ${\bf S}$ divided by the length in that direction of the box domain.

525: We are interested in measuring the flux in the $y$ direction, then the expression 

526: \ref{eq:phi1} simplifies to:

527: 

528: \begin{equation}

529: <\phi>=\int^{L{_y}}_0\left(\int_0^{L_{x}}\int^u_l  B_y dzdx\right) dy/(L_y L_x L_z)\label{eq:phi2} 

530: \end{equation}

531: 

532: \noindent where $L_x$  is the length of the box in the $x$ direction, $L_y$ is the length 

533: in the $y$ direction and the integration in $z$ goes from $z=l$ to $z=u$.

534: We consider two different integrated regions in the figure \ref{fig:fluxret}, one is below the 

535: photosphere ($z<0.01$~Mm), {\it i.e.} from $l=-1.5$~Mm to $u=0.01$~Mm, in red color, 

536: and the second from $l=0.01$~Mm 

537: to $u=245$~km, in black. 

538: The figure shows that the greater the twist, the greater the amount of magnetic flux that crosses 

539: the photosphere. On the other hand, the time needed for the tube

540: to cross the photosphere is similar in all simulations.

541: (It should be kept in mind that there is also magnetic flux being lost

542: through the bottom boundary of the simulation box in addition to that rising 

543: beyond the photosphere into the upper atmosphere.) 

544: The simulation A2 (with weaker injected field $B_o$) has a smaller total flux and 

545: the slope ($d\langle\phi\rangle/dt$) is much smaller than in the other simulations. 

546: When the tube has finished crossing the photosphere ($t>2000$~s)

547: the flux retained above the photosphere is in quasi-equilibrium and 

548: oscillates with a period of roughly 5 minutes.

549: This oscillation is due in part to the magnetic field that is returned 

550: to the photosphere and in part because the direction of the tube 

551: changes in time.

552: The oscillation in flux strength coincides with oscillations in maximum downward 

553: velocity variation with time, but there is a small delay of around 30~s between the 

554: flux and the velocity variation. We find the same phenomenon in upper layers (see section

555: \ref{sec:revgran} and \ref{upcrom}). 

556: 

557: The average amount of magnetic flux in the convection zone is greater than in the photosphere, 

558: which means not all of the flux tube crosses into the photosphere. 

559: The only portion of the magnetic tube that reaches the photosphere is the part that has a 

560: steeply decreasing field strength with height. \citet{cheung2007} have pointed out that the 

561: higher the level of twist in the initial flux tube, the larger the fraction of magnetic flux 

562: that can pass into the regions above the photosphere. This can be explained as a consequence

563: of the magnetic tension of the transverse fields tending to keep the tube coherent as shown

564: by \citet{moreno1996}. Following the linear stability analysis of \citet{Acheson:1979lr} we find

565: that the photosphere has a superadiabatic excess profile, defined 

566: by $\delta_T=\nabla-\nabla_{ad}$, that is shaped like a gaussian with 

567: height. In this case the magnetic buoyancy instability sees a barrier in the 

568: photosphere and the emerging tube velocity decreases as it comes closer to 

569: the photosphere. Only the portion of the flux tube where the gradient of the magnetic field

570: strength is larger than the superadiabatic excess in the photosphere can reach the layers

571: above the photosphere. The relation between the buoyancy instability and the superadiabatic

572: excess has been studied by several authors \citep{Acheson:1979lr,magara2001,archontis2004}.

573:  

574: The field left behind after the tube has passed is more vertical than it was before the 

575: tube passes through a given particular height. This behavior is more evident near the end 

576: of the simulation where the magnetic field is more vertical in the 

577: chromosphere. The ratio between 

578: $B_z$ and $|B|$ oscillates in time in much the same manner as does the maximum downward 

579: velocity in the photosphere between $z=0$~km and $z=458$~km.

580: 

581: \subsection{Reverse granulation}\label{sec:revgran}

582: 

583: Intensities originating a few hundred kilometers above the

584: granulation show a pattern that is reverse that of granulation,

585: with low intensity above granules and higher intensity above

586: intergranular lanes \citep{Evans:1972fk,

587: Suemoto:1987qy,Suemoto:1990uq,Rutten:2003fj}. 

588: The reason for the inverse contrast is the cooling from 

589: a divergent velocity field above granules and heating from

590: a convergent velocity field above the intergranular lanes.

591: The final intensity pattern from this layer may also

592: be influenced by internal gravity waves but does not

593: appear to be related to the magnetic field

594: \citep{Rutten:2004kx,Leenaarts:2005yq}.

595: However, as we discuss below, when the magnetic flux crosses this region, 

596: the contrast and the structure of the reverse granulation is modified. 

597: 

598: Figure~\ref{fig:3drev} shows 3D images in two

599: different views at $t=1900$~s; as seen from above and from below the reverse granulation layer. 

600: The figure shows how the reverse

601: granulation has been changed by the presence of the rising magnetic flux tube and 

602: is colder in the interior of the reverse granulation cells in regions where the tube is 

603: located than outside. Also at this height we can see how the structure of the tube

604: has been disrupted as it has passed through the atmosphere. 

605: In fact, the structure of the tube is more split and fragmented than that seen in 

606: figure~\ref{fig:3dpho}.

607: 

608: We define a horizontal layer at $z= 234$~km. At this height the granular 

609: pattern is still clearly visible, but it is reversed compared to the photosphere 

610: (figure~\ref{fig:revgran}). We see expansion of the cells beginning at time 920~s. 

611: The flux tube starts to go

612: through the reverse-granulation layer some 80~s after it crosses the photosphere,

613: {\it i.e.} at time 1380~s.

614: As in the photosphere, the reversed granules increase in size as the flux

615: tube approaches. The magnetic field lines of the tube are seen in figure~\ref{fig:3drev} 

616: to go from one side of the expanded cells to the other.

617: 

618: Figure~\ref{fig:fieldlwch} shows the evolution of the vertical magnetic field $B_z$ in the

619: left column and the horizontal magnetic field in the right column, at three different

620: instants in time. The maximum magnetic field-strength, before the flux tube reaches the reverse

621: granulation, is of order 500~G. When the tube starts to cross the

622: height $234$~km the maximum rises to 600~G. 

623: At later times, but while parts of the tube are still to be found at this height,

624: the remaining magnetic field is confined to downflow regions and the maximum field strength 

625: rises to 650~G. After the tube has passed, some remaining confined flux rises to a  maximum

626: of up to 800~G. During the last 500~s of the simulation the maximum field strength oscillates

627: around 700~G. Again, as in the photosphere, the magnetic field is

628: initially horizontal and appears in the center of the granular cells 

629: (top panels). The magnetic field is then confined and moves with the fluid 

630: to the edges of the cells where it becomes vertical (see the second and third 

631: rows of the figure \ref{fig:fieldlwch}). 

632: At the cell edges 

633: a great number of bipoles form by $t=2100~s$.

634: However, at later times the bipoles are canceled

635: by magnetic diffusion; {\it i.e.} in the bottom panel on the left side of 

636: figure~\ref{fig:fieldlwch} we find generally that negative (downward) magnetic field 

637: is concentrated while we find positive (upward) flux on the right.

638: 

639: The top row of figure~\ref{fig:revgran} shows the continuum intensity at

640: 170~nm as calculated from the simulation at times 1620~s (left) and

641: 3220~s (right). The continuum opacity at this wavelength is dominated

642: by Fe bound-free opacity and the monochromatic optical depth unity is

643: around $250$~km height in the simulation. Note that real observations

644: around this wavelength (such as the TRACE 170~nm bandpass) are dominated

645: by line opacity, with the intensity thus coming from higher layers. The

646: intensity images shown here are, however, typical of a diagnostic formed

647: in the reverse granulation layer (like the inner wings of the resonance

648: lines from ionized calcium).

649: The figure also shows the vertical velocity in the second row with the same grey-scale

650: values, {\it i.e.} from 2.6~km s$^{-1}$ (white) to -4.1~km s$^{-1}$ (black), the divergence of 

651: the velocity in the third row and the vertical vorticity in the fourth row, all

652: these variables at height $234$~km. The synthetic 170~nm intensity shows reverse granulation

653: and it is clear that at $t=1620$~s the granular cells have expanded in regions co-spatial

654: with the flux tube. The same pattern is also found in plots of the temperature and the 

655: vertical velocity. Larger cells grow dimmer and colder due to the expansion as can be verified 

656: by considering the velocity divergence. While cool, dim regions are due expansion 

657: the hotter bright network points are seen to be related to contraction of the fluid at this height. 

658: At the end of simulation we find several bright points. This is seen clearest 

659: at time $t=3220$~s in the regions where we find concentrated vertical 

660: magnetic field (see  second column of the figure \ref{fig:revgran} at the position 

661: [$6.7,3.7$]~Mm). These bright points in intensity 

662: are related to large vertical velocities, but also to high temperatures compared 

663: to the surroundings, a large vertical vorticity and to 

664: compression of material (black color in the divergence of the velocity plots).  

665: The heating process for these bright points is clearly related to the high velocity 

666: convergence and high vorticity. 

667: We will find similar bright points at greater heights in the atmosphere, though 

668: with larger horizontal extent.

669: 

670: The temperature range is found to be [4.9,6.2]~kK before the tube reaches the $234$~km

671: layer,  while during and after the flux tube passes the temperature range is [4.6,6.1]~kK;

672: the lower limit decreases in time after 2400~s. The reason for this temperature drop is the

673: expansion as discussed above.

674: 

675: There are no

676: large changes in the upward maximum velocity with time before, during

677: or after the tube crosses the layer $234$~km height; it is found to be around 2.5~km s$^{-1}$. 

678: The maximum downward velocity does not change until after 

679: $t=1400$~s when it has a small increase for 200~s.

680: The maximum downward velocity then returns to normal before

681: increasing from 3~km s$^{-1}$ to 5~km s$^{-1}$ starting at $t=2200$~s.

682: These large downflow velocities are found to correspond to the small regions 

683: with a strongly confined magnetic field in the intergranular lanes. 

684: Examples of such a regions are found in the second row, right column, of figure \ref{fig:revgran} at position 

685: $[x,y]=[4.1,0.2]$~Mm and also at $[7.1,3.6]$~Mm. 

686: Towards the end of the simulation, at time 3100~s, the maximum downward velocity 

687: has returned to 3~km s$^{-1}$. 

688: The maximum horizontal velocity oscillates around 5.5~km s$^{-1}$.

689: However, it has a peak at 1480~s with a maximum velocity of 8~km s$^{-1}$ 

690: as the tube is crossing the layer $234$~km height. 

691: 

692: The expansion of the cells is observed at time $t=900$~s in the photosphere ($z=10$~km) and

693: roughly 25~s later in the reverse granulation layer $234$~km height. The separation between

694: these two layers is $224$~km. The propagation speed of the tube itself is much slower,

695: as table \ref{tab:events} shows, it spends 250~s in moving from the 

696: layer at $10$~km to the layer at $234$~km height. Another possibility for explaining 

697: the expansion of the cells at the layer $234$~km height is by the propagation of 

698: Alfv{\'e}n waves, but again, the time such waves need to propagate from 

699: the layer $10$~km to the layer $234$~km height is too long, because 

700: the mean Alfv{\'e}n speed is very small in that region ($<100$~m s$^{-1}$). 

701: On the other hand, the average sound speed between $10$~km to $234$~km height is 8.5~km s$^{-1}$, which 

702: gives a travel time close to the 25~s that fits with the time to move outward the 

703: horizontal expansion of the cells.

704: In other words we see the expansion of the granular cells some 400~s before the magnetic tube 

705: actually emerges through the photosphere, when the tube is situated 182~km below the photosphere. 

706: The speed of the sound below the photosphere is on average 10~km s$^{-1}$ which means the expansion

707: wave is produced roughly 420~s before the tube reaches the photosphere, {\it i.e} when the 

708: tube is roughly at -246~km. The average tube velocity from -246~km to 10~km is 0.5~km s$^{-1}$.

709: However, the tube decelerates when it approaches the photosphere as explained 

710: in the section \ref{sec:photosphere}.

711: The cells average horizontal expansion velocity is roughly 2.5~km s$^{-1}$ before the tube reaches

712: the photosphere.

713: 

714: The figure~\ref{fig:fluxretlow} shows the average magnetic flux in the same manner

715: as in figure~\ref{fig:fluxret}, but here the regions considered are from the 

716: layer $z=10$~km to the layer $z=234$~km (black color) and from the layer $z=234$~km 

717: to the layer $z=458$~km (red color). 

718: Each of the four displayed simulations conserve magnetic flux in the sense that 

719: when the magnetic flux tube crosses these regions no flux is lost or retained in the 

720: region, as opposed to what happens in the convection zone. The biggest 

721: difference between the two regions is seen to be in the slope of the flux

722: ($d\langle\phi\rangle/dt$) as the tube moves through the region. The slope of  

723: the flux is much smaller than that found in figure~\ref{fig:fluxret} when 

724: the tube crosses those layers, {\it i.e.} more time is spent in crossing 

725: each region. The oscillation of the mean magnetic flux occurs more or 

726: less at the same time, when the tube crosses the corresponding region,

727: but the amplitude of the variation is greater in higher regions.

728: 

729: \subsection{End of the photosphere-beginning of the Chromosphere}

730: \label{upcrom}

731: 

732: The top of the convection zone, photosphere, and lower-chromosphere 

733: are the regions where acoustic waves are excited by convective motions. 

734: While the waves propagate upwards, they steepen into shocks, dissipate, 

735: and deposit their mechanical energy as heat in the chromosphere. 

736: In studying the quiet non-magnetic chromosphere

737: \citet{Wedemeyer:2004} corroborates in 3D models the finding by 

738: \citet{Carlsson+Stein1994} that the chromospheric temperature rise

739: derived from semi-empirical modeling of the time average intensity does not 

740: necessarily imply an increase in the average gas temperature but rather can 

741: be explained by the presence of substantial spatial and temporal temperature 

742: inhomogeneities, such as those caused by non-linear waves. 

743: We are interested in finding how the rise of a flux tube

744: affects chromospheric dynamics and energetics, {\it i.e.} if there are other 

745: energy sinks or sources that become active with the appearance of new magnetic flux,

746: or if the new flux modifies the propagation of acoustic waves into

747: the middle and high chromosphere.

748: 

749: In our simulations the magnetic flux tube enters the height $z=458$~km 

750: some 300~s after it has crossed the photosphere. The layer at $458$~km height suffers 

751: various perturbations in response to the appearance of the flux tube. 

752: Of the most important effects is the expansion 

753: and the cooling related to that expansion. Each of these processes are 

754: explained below.

755: 

756: Figure~\ref{fig:3dcro} shows two 3D images of the A4 simulation at

757: $t=1980$~s 

758: from above and from below a temperature slice made at height $z=458$~km

759: above the photosphere. The temperature structure at that height ($z=458$~km)

760:  has been changed by the presence of the tube and is much cooler in regions

761: where the tube is located. However, on the borders of the rising tube we find regions with 

762: higher temperatures. As in the previous 3D images, featuring lower heights,

763: we see how the structure of the magnetic flux tube has been perturbed as a result of 

764: rising through the atmosphere; comparison with these figures shows the structure of 

765: the tube is at this late time even more splintered and fragmented than at earlier times 

766: (see figures \ref{fig:3dpho} and \ref{fig:3drev}).

767: 

768: The vertical magnetic field is shown in the left column of figure~\ref{fig:fieldcr} 

769: and the horizontal magnetic field in the right column,

770: both for 2 different instants at height $z=458$~km. The maximum magnetic

771: field strength before the magnetic flux tube reaches this height is 170~G. 

772: During the flux emergence process the maximum field strength increases to 270~G, and

773: at later times, as the magnetic flux is compressed and confined to smaller regions, 

774: the field strength in some points increases to 350~G. 

775: At this height there is a cellular structure at any given time, but

776: the cells do not directly correspond to the photospheric granular 

777: cells, but rather to patterns in the velocity field that change from instant to 

778: instant. The vertical magnetic field is mostly confined to downflow regions and dominates

779: the field. However, as the flux tube crosses the layer at $458$~km height the horizontal 

780: magnetic field is strongest and fills the upflow cells. 

781: There are some regions or ``points'' with strong vertical magnetic field. For instance

782: the bottom left panel of the figure~\ref{fig:fieldcr} at position $[7,4]$~Mm.  

783: These

784: points have increasing maximum magnetic field-strength with time

785: after the flux tube has passed through the layer and 

786: they move around with the fluid.

787: 

788: Figure~\ref{fig:tgcr} shows the calculated continuum intensity as described 

789: in the figure at

790: 160~nm and 130~nm for two snapshots of the simulation. These 

791: images show what could

792: be observed in the TRACE 160~nm filter and in a SUMER raster 

793: at 130~nm. At both wavelengths the continuum opacity is dominated

794: by bound-free opacity from silicon: at 160~nm from the first

795: excited level and at 130~nm from the ground state.

796: The monochromatic optical depth unity is

797: on average around $480$~km and $790$~km height in the simulation at 160~nm and 130~nm respectively 

798: (in the VALC model atmosphere the corresponding heights are $456$~km and $735$~km, respectively).

799: The intensity at $t=1620$~s shows expansion

800: of the plasma with subsequent dimming, somewhat more clearly at 

801: 160~nm. At 3220~s both the continua show a bright point

802: at position $[7,3.8]$~Mm. It is also possible that the tube expansion

803: could be affected by the finite horizontal size of the computational domain;

804: figure~\ref{fig:fieldcr} shows that the horizontal magnetic field almost reaches 

805: the boundary layers of the box. On the other hand the (limited number of) models we have run

806: with a larger box, $16\times 8\times 16$~Mm$^3$ do not show a fundamentally different 

807: behavior, so we believe the effects of limited simulation size are minor.

808: 

809: The vertical velocity is shown in figure~\ref{fig:divucr} in the first

810: row of panels using the same grey-scale intensity, from 1.5~km s$^{-1}$ (white) 

811: to -5.88~km s$^{-1}$ (black); the divergence of the velocity in the second row

812: and the vertical vorticity in the last row of panels, all at time 

813: 1620~s in the first column and at time 3220~s in the second column. 

814: We find similar structure to those found in the synthetic observations 

815: (figure~\ref{fig:tgcr}) before the tube crosses the layer at $458$~km height. 

816: The structure of the cells is strongly modified as the tube goes

817: through that layer. We find that the cell expansion is related

818: to the divergence of the velocity, also at this height, but that the 

819: correspondence between temperature/continuum emission and velocity 

820: divergence is not as clear as in the lower layers. There are 

821: large vertical velocities (first row of the figure~\ref{fig:divucr}), both up or down 

822: (see the left-panels of the figure~\ref{fig:divucr}  at position [$2.1,0.8$] 

823: and right panels  at position [$7,3.7$]) which

824: correspond to the bright points in the synthetic observations and also fit 

825: with the large concentration of magnetic flux and vertical vorticity. 

826: 

827: The maximum upward velocity oscillates around

828: 1.5~km s$^{-1}$ with no important changes in time except at $t=2420$~s when the

829: peak rises to 4~km s$^{-1}$. This peak is due to the rapid movement of the tube as

830: it rises. The maximum downward velocity oscillates around 2~km s$^{-1}$. Later,

831: the maximum downward velocity increases slowly with time from $t=1800$~s, up to

832: 5~km s$^{-1}$ towards the end of the simulation at $t=3220$~s. The downflowing plasma

833: is located in the same region as the bright points with high magnetic field strength and 

834: vorticity are found. The maximum horizontal

835: velocity varies around 3.5~km s$^{-1}$ in the period from 0~s to 1400~s.

836: After that time it 

837: increases with time and rises to 7~km s$^{-1}$ at time 1940~s due to the

838: expansion process. Afterwards the maximal upward velocity returns to 

839: roughly 5.5~km s$^{-1}$, though with big variations.

840: 

841: The temperature range before the tube crosses the layer $z=458$~km is 

842: [4,6]~kK from 0~s to 1300~s. As the tube approaches the 

843: lower limit decreases to 3~kK due to expansion. When the tube arrives,

844: the temperature range changes to  [3.2,6.2]~kK at time 1900~s.  The lower limit

845: increases to 3.9kK after 2500~s. 

846: 

847: After the magnetic flux tube has crossed the layer at $458$~km height the structure of the

848: plasma no longer appears granular. The magnetic field, shown in the figure~\ref{fig:fieldcr},

849: is almost horizontal and only in the boundary layer of the tube can one see any 

850: vertical component. 

851: With greater initial twist the tube 

852: remains more horizontal in the chromosphere compared to the cases with less initial 

853: twist, {\it i.e.} the ratio of the horizontal magnetic field

854: component to the vertical magnetic field component, in regions where the tube is located, 

855: increases with increasing twist. 

856: 

857: Figure~\ref{fig:divucr} also shows the expansion of the fluid as the tube 

858: crosses the layer at $458$~km height. The center of these cells are colder than 

859: the edges as the tube expands. 

860: However, as the tube crosses the layer at $458$~km height, 

861: the small cells seen in the initial model disappear and cold larger zones form near 

862: the center of the tube. This is presumably due to expansion 

863: (second row of the figure~\ref{fig:divucr}), but the velocity divergence

864: does not fit as well with the temperature or intensity maps as found in the lower 

865: layers of the height $458$~km. This suggests that at this height also other processes 

866: could be active in cooling the chromosphere. We will 

867: go into more detail on these alternate processes in the next section,

868: as they become steadily more important with increasing height into the chromosphere.

869: 

870: At the boundary layer of the magnetic flux tube the temperature

871: rises. This heating could be due to a wave, compression of the plasma, or reconnection

872: where the magnetic field is almost vertical.

873: Once 

874: the tube has disappeared from the layer at $458$~km, at time $t\approx 2300$~s, the temperature 

875: displays an irregular structure with some high temperature points. These points are coincident

876: with a high vertical vorticity and with strong vertical magnetic field 

877: (e.g. at position [7,3.8] Mm at time 3220~s, see figures~\ref{fig:fieldcr}-\ref{fig:divucr}). 

878: Note that the temperature structure has not returned to its 

879: pre emergence state. This could be due to the fact that the magnetic structure of the atmosphere

880: has changed: before the tube appeared the energy density of the

881: magnetic field was low and $\beta > 1$, but around $t=2200$~s we find several regions 

882: where $\beta \approx 0.5$ and the field dominates the chromospheric structure.

883: 

884: Bright points are seen to occur in the synthetic observations (figure~\ref{fig:tgcr}). 

885: In the layer at height $458$~km the bright points have greater contrast than found in the regions

886: below and also much greater temperature than in the surrounding plasma (The bright point in figure

887: \ref{fig:tgcr} right-top panel at the position $[7,4]$~Mm has a temperature at this height of

888: 6500~K compared with the surrounding temperature of 5100~K. At the same position at height

889: $234$~km (see figure \ref{fig:revgran}) the temperature is 5700~K compared with 5200~k in the surroundings)

890: As explained earlier for the lower layers, bright points are related to strong magnetic field, 

891: to vertical velocity, and vertical vorticity, and to a larger compression 

892: 

893: than in the surroundings. The heating seems to come primarily from the 

894: convergence of the velocity field as we find the most important heating term is the advection term; 

895: it is an order of magnitude greater than the joule and viscous heating terms.

896: 

897: 

898: Figure~\ref{fig:fluxretup} shows the total flux for the regions from 

899: $z=234$~km to $z=458$~km in red, and from  

900: $z=458$~km to $z=905$~km, in black. 

901: The  biggest difference between the figures \ref{fig:fluxret}, 

902: \ref{fig:fluxretlow} and \ref{fig:fluxretup} is the slope of the flux with time 

903: ($d\langle\phi\rangle/dt$). The total flux entering the chromosphere is much smaller than 

904: that reported in figures~\ref{fig:fluxret} and \ref{fig:fluxretlow} because the tube spends

905: more time in crossing the upper regions. This is presumably due to the horizontal 

906: expansion of the magnetic flux tube. The amplitude of the slope of the curves in the figure 

907: \ref{fig:fluxretup} increase with magnetic field strength. After the initial

908: large rise, when the tube crosses the corresponding section, the curve shows an oscillation (see

909: figure \ref{fig:fluxretup}) that is quite similar to that seen in the layers 

910: below (see figures \ref{fig:fluxret} and \ref{fig:fluxretlow}). However, by comparing 

911: each curve it is possible to see that there is a small delay between the different lines 

912: in the figures \ref{fig:fluxret}, \ref{fig:fluxretlow} and \ref{fig:fluxretup}. 

913: This retardation of the oscillation between the different layers is what we would 

914: expect from a wave that propagates at the sound speed between each layer. 

915: The variation of the mean magnetic flux strength increases with increasing height, 

916: twist, and/or the magnetic field strength.

917: The oscillations seem related to the oscillations in the downward maximum

918: velocity with time and also to oscillations in the ratio between the vertical and horizontal 

919: magnetic field, as mentioned in the previous sections.

920: 

921: Finally, we find that the slope of the flux with time, $d\langle\phi\rangle/dt$, at the 

922: time the tube crosses into the chromosphere, 

923: is larger with greater values of the twist parameter and with larger magnetic 

924: field strength. This effect is clearer in the chromosphere than it is in the photosphere.   

925: Previous 2D flux emergence simulations by \citet{Shibata:1989lr} and \citet{magara2001} have 

926: also identified magnetic buoyancy 

927: instabilities as important in the rise of magnetic field to the chromosphere and coronal layers.

928: 

929: \subsection{Chromosphere, Transition Region and Corona}

930: 

931: Let us finally discuss the effects of the magnetic flux tube as it reaches the 

932: upper chromosphere, transition region and lower corona. The variation of the

933: magnetic field in the chromosphere, at $z=906$~km, is shown in 

934: figure~\ref{fig:fieldupcr} with the vertical field in the left column and the

935: horizontal field in the right column, both at three instants; times 600~s, 2800~s and

936: 3220~s. The first signs of the magnetic flux tube reaching the chromosphere 

937: are evident at time $t=2100$~s. However, the tube does not appear uniformly 

938: in the region from $x=2$~Mm to $x=6$~Mm as it does in the levels below. Rather

939: patches of magnetic flux rise into different places randomly along the tube 

940: region at different times. The maximum absolute magnetic field, before the tube reaches 

941: the layer is 43~G. This value increases after the tube comes into the chromosphere

942: to 80~G. During the flux emergence into this level and after the tube has passed 

943: the average of the maximum field is of order of 65~G. 

944: At this height the magnetic field does not have any singular isolated points with strong 

945: amplitude. The field becomes generally more horizontal once the tube crosses this 

946: layer and has a smoother structure than in the layers below.

947: 

948: In figure \ref{fig:tgupcr} we show the intensity of the C~{\sc i}  continuum 

949: at  109.9~nm (as could be observed with SUMER). We have

950: calculated intensities at $t=1620$~s before the emerging magnetic flux

951: has entered the chromosphere  and at $t=3220$~s, some 15 minutes after the

952: emerging flux first appears. The mean monochromatic optical depth unity is

953: around $1100$~km height in the simulation of the C~{\sc i} continuum at 109.9~nm 

954: (in the VALC atmosphere model it is at $z\approx975$~km).

955: We also show synthetic Hinode observations of the 

956: Ca~{\sc ii} H-line at times 600, 1620, 2110 and 3220~s. Note that while the carbon continuum

957: is formed mainly in the chromosphere

958: the Ca~{\sc ii} as observed with the 0.22~nm wide Hinode filter has contributions from 

959: the entire atmosphere, stretching from the upper photosphere to the chromosphere

960: \citep{Carlsson+etal2007}.

961: The structure of the intensity formed at this height shows large changes before and 

962: after the magnetic flux tube reaches the chromosphere. 

963: The temperatures range in the chromosphere at height $906$~km before the tube reaches this level

964: is [2.2,7.2]~kK, while after the tube has reached this height

965: the range is [2.0,7.0]~kK.

966: However, note especially the large

967: dim (and cold) zones that become evident after the tube crosses into the chromosphere

968: at time 3000~s. These dim, cold, regions are much larger than in the lower layers (see

969: section \ref{sec:revgran} and \ref{upcrom}). 

970: 

971: It is worth remarking that we do not find the concentrated bright points that 

972: appear in the lower lying layers. This is presumably because the magnetic field 

973: is mainly horizontal and we do not observe the confinement of the magnetic field that

974: occurs in the regions below due to plasma motions. On the other hand,

975: the chromosphere and corona suffer large changes in size, structure

976: and temperature after the tube enters this zone. 

977: The cooling in the layers between $234$~km and $458$~km height is due to radiative losses and

978: superadiabatic expansion, but in the chromosphere the big cold 

979: regions do not fit as well with the local plasma expansion as in the lower layers. 

980: 

981: The vertical velocity in the chromosphere at height $906$~km is shown in figure~\ref{fig:uzupcr}

982: at four different instants; 600, 2110, 2800 and 3220~s. The maximum downward velocity 

983: is around 10~km s$^{-1}$ and the maximum upward velocity is around 7~km s$^{-1}$ throughout the whole 

984: simulation.

985: It is difficult to see any effects due the emergence of flux in the maximum 

986: horizontal velocity as these have large changes with time. The average maximum horizontal 

987: velocity is 13~km s$^{-1}$ and varies between 7~km s$^{-1}$ and 20~km s$^{-1}$. On the other hand the structure 

988: of the velocities change radically as is evident by comparing the pre- and post- emergent 

989: panels of figure~\ref{fig:uzupcr}. 

990: Shock waves dominate the upper chromosphere before

991: the tube crosses the chromosphere ($t=600$~s). However, during and after the tube

992: crosses the chromosphere the velocity is much smoother in the regions where the field

993: strength increases due to the rise of the flux tube into to upper chromosphere.

994: 

995: That the chromosphere and overlying transition region and lower corona suffer 

996: large structural changes as a result of flux emergence is driven home by considering a 

997: vertical slice of the atmosphere as shown in the left panel of figure~\ref{fig:tgxz}.

998: The transition region has been pushed by the emerging flux from a height that varies 

999: with chromospheric dynamics from around 2~Mm to almost 6~Mm. This panel also shows that the 

1000: flux tube has grown in horizontal extent and almost fills the computational box. We therefore

1001: also present the equivalent vertical temperature slice from simulation B1 which has a 

1002: computational domain that is twice as large in the horizontal directions. It is evident

1003: also in this model that the rising flux sheet has pushed the transition region and lower corona

1004: aside and has lifted chromospheric plasma to heights of roughly 6~Mm. These cool, 

1005: Lorentz force supported bubbles are long lived, and the simulation that has run the longest, 

1006: B1 to 4500~s, maintains an extended chromosphere for almost 2000~s and is still doing so 

1007: at the end of the simulation. The gas in the extended chromospheric bubble is quite cool, 

1008: down towards a uniform 2000~K, and shows no sign of reheating at heights greater than 1~Mm.

1009: The big bubbles can also be seen in the synthetic Hinode Ca~{\sc ii} H limb images

1010: as shown in figure~\ref{fig:caxz}. These synthetic limb

1011: observations, described in the figure,

1012: show that the chromosphere has expanded and cooled significantly as

1013: the flux tube rises through the chromosphere.

1014: Towards the end of the B1 simulation the emerging magnetic flux shows signs of having

1015: begun fragmentation as it reconnects with the previously present field.

1016: This looks to have interesting implications for transition region and coronal observables, but

1017: the process is not complete in any of our simulations and we will refrain from discussing 

1018: this until the next paper in this series. 

1019: 

1020: Further insight into this phenomena can be found by considering vertical slices of the

1021: B1 simulation at earlier times, as the flux tube bubble rises through the chromosphere,

1022: such as the slices of logarithmic temperature, vertical velocity and logarithmic pressure

1023: we show in figure~\ref{fig:tg_bubble}. We find that a small, low $\beta$ region forms just above the

1024: photosphere, centrally placed over the (partly stalled) rising flux tube. As plasma

1025: $\beta$ falls below one, the bubble becomes buoyant and rises at roughly the sound speed,

1026: expanding rapidly both horizontally and vertically as it does so. The expansion causes 

1027: the plasma to cool and the pressure in the bubble becomes much smaller than in the surrounding 

1028: chromosphere. At some point in the chromosphere the bubble reaches a quasi-equilibrium

1029: with its surroundings and the continued expansion is halted, or at least slows considerably. 

1030: 

1031: The appearance of the extended chromospheric bubble and its associated magnetic field has

1032: large repercussions to chromospheric dynamics and energetics. This is abundantly clear

1033: as we consider figure~\ref{fig:tgxt} where we have plotted the variations of the temperature,

1034: velocity, and logarithmic pressure at height 1.1~Mm as a function of time in the B1 simulation.

1035: The region where the bubble is found has very low temperatures and pressures, but also shows 

1036: that the otherwise all pervasive chromospheric oscillations are excluded from the bubble. 

1037: Acoustic waves seem to have difficulty entering this low $\beta$ region - and even when the

1038: bubble oscillates as the rest of the chromosphere at late times the temperature and pressure 

1039: plots show that these oscillation are, at best, only weakly compressive. This expulsion

1040: process appears in every simulation we have run and is important from roughly 800~km above 

1041: the photosphere and extends up to the transition region. 

1042: 

1043: \section{Conclusions}

1044: 

1045: The emergence of magnetic flux tubes has been studied in order to

1046: understand not only the dynamics of flux emergence evolution, but also its effects

1047: on the solar atmosphere. One of the major questions we are considering is how flux 

1048: tube emergence changes the energetics and the dynamics of the photosphere,

1049: chromosphere and corona. What contribution to atmospheric energetics 

1050: does the emerging flux represent? Another set of important questions relates to 

1051: the observational signatures of flux emergence throughout the atmosphere.

1052: This paper does not pretend to fully answer these questions, but is rather a 

1053: first attempt to consider a framework for some of the problems and processes 

1054: that arise as one follows the rise of a magnetic field up through the atmosphere.

1055: 

1056: Specifically, the aim of this work is to simulate a 3D box which

1057: contains a realistic convection zone and the atmospheric elements above;

1058: a photosphere and chromosphere with attendant non-grey radiative losses, 

1059: a transition region and lower corona, including realistic optically 

1060: thin cooling and Spitzer conductivity. The magnetic field injected at the

1061: lower boundary is either a horizontal flux tube with varying degrees of twist 

1062: and magnetic field strengths, or a non-twisted flux sheet. 

1063: 

1064: Considering first the region spanning the convection zone to photosphere, we find our

1065: results to agree with those reported by \citet{cheung2007} as well as the 

1066: observations of flux emergence of small-scale magnetic loops in the quiet-Sun 

1067: internetwork studied by \citet{Centeno:2007lr}.

1068: 

1069: We also find that the time evolution of the magnetic flux tube as it rises into the atmospheric

1070: layers above the photosphere is similar in all five simulations we have studied. The flux tube or 

1071: flux sheet crosses the different layers of the atmosphere at roughly the same time in each 

1072: simulation as shown in table \ref{tab:events}.

1073: 

1074: The results shown in this paper focus on four different

1075: layers; the photosphere at height $10$~km, the region that produces 

1076: reverse granulation at height $234$~km and is represented 

1077: by the bulk of the emission in the Hinode/SOT Ca~{\sc ii} H-line, 

1078: at height $458$~km which is typical for the TRACE 160~nm, and into the chromosphere proper

1079: at a height $906$~km above the photosphere where short wavelength continuum emission 

1080: at 130~nm and 109~nm are observable with SOHO/SUMER. Simulation A4 is the longest running 

1081: simulation that contains a twisted flux tube. In this simulation the flux 

1082: tube rises from the convection zone to the chromosphere at height around 1.1~Mm in 3000~s.

1083: The amount of magnetic flux that reaches each of these regions 

1084: is strongly related to the twist parameter and also to the 

1085: original magnetic field strength injected at the bottom boundary. 

1086: We find that with greater twist and/or greater field strength, 

1087: more flux crosses the photosphere into the chromosphere. The

1088: photosphere is the region where most of the flux is retained. 

1089: The flux that eventually enters the lower chromosphere rises through the various heights we 

1090: have recorded at less than 10~km s$^{-1}$ and does not halt until reaching the transition region/lower 

1091: corona. 

1092: 

1093: As previously reported in the literature, we find that an 

1094: emerging flux tube produces an expansion of the granulation cells in the 

1095: photosphere. This expansion begins some 400~s before the flux tube itself reaches 

1096: the photosphere.  The simulations show that also the chromospheric plasma expands 

1097: as the emerging flux tube rises. As in the photosphere, we find a delay between 

1098: cell expansion and the arrival of magnetic flux in the lower layers of the chromosphere.

1099: As a result of the expansion the tube plasma is colder than its surroundings,

1100: and emission from the central parts of the flux tube dim successively 

1101: at all heights of the atmosphere according to the schedule outlined in table~\ref{tab:events}.

1102: Such successive dimming has been observed by the SOT and EIS instruments aboard the

1103: Hinode spacecraft \citep{Hansteen+etalSOTEIS2007} for lines up to and including Fe~{\sc xii} 19.5~nm

1104: which is formed at a temperature of some 1.2~MK, but further observations are certainly 

1105: required to search for confirmation or discrepancies with the models described here.

1106: 

1107: In addition to dimming of the central portion of the flux tube, 

1108: the expansion produces other important effects: For instance,

1109: in the photosphere, the boundary layer of the rising flux tube is a region that 

1110: is conducive in producing granular collapse. 

1111: 

1112: We also find that chromospheric structure

1113: is substantially altered by the rise and expansion of the flux tube. The chromosphere 

1114: is pushed upward by the magnetic field and cool chromospheric material is found at 

1115: heights up to 6~Mm above the photosphere in contrast to the $2-3$~Mm extent of the 

1116: pre-emergence chromosphere. In expanding, the transition region and corona above 

1117: the flux tube are pushed aside by the chromospheric plasma. However, the pre-existing

1118: coronal magnetic field and transition region and coronal emission are remarkably 

1119: little changed at the times reached by our simulations. Only in simulation B1 do 

1120: we see some initial signs of reconnection between the pre-existing coronal field

1121: and the rising bubble of emerging flux. This reconfiguration, with attendant changes

1122: in transition region and coronal emission, occurs some $30-40$~minutes after the flux

1123: tube initially pierces the photosphere. Since the flux tube also expands dramatically 

1124: in the horizontal direction, reaching widths almost equal to the simulation box

1125: for simulations A1--A4, discussion of transition region and coronal

1126: effects will be postponed to a later paper when simulations contained in a larger 

1127: box are completed. 

1128: 

1129: While the cooling process in the photosphere and at the lower chromospheric heights

1130: ($234, 458$~km) is dominated by adiabatic cooling, we find that at higher layers where

1131: $\beta < 1$ other processes also come into play. The acoustic waves

1132: that are generated as a result of convective overshoot and/or granular dynamics and that 

1133: permeate the non-magnetic chromosphere are expelled from the rising cooling flux tube. 

1134: Thus, the temperature variations that are due to compressive shock waves propagating up 

1135: through the chromosphere are absent in the plasma rising with the flux tube which maintains

1136: a much lower temperature than its surroundings for the duration of our simulations. 

1137: Towards the end of the simulations the magnetic bubble does begin to oscillate with 

1138: the same vertical velocity as the rest of the atmosphere, but the modes penetrating

1139: the bubble are still non-compressive at the time of the completion of the simulation

1140: and the temperature remains fairly constant.

1141: 

1142: Another interesting effect is the increase of magnetic field strength

1143: in small regions or ``points'' in the atmosphere near the edges of the rising

1144: flux tube. For example, photospheric

1145: bright points are formed in the intergranular lanes edging the flux tube 

1146: after the flux tube has passed through the photosphere. 

1147: These are regions where the magnetic field is concentrated and the field reaches

1148: strengths greater than $10^3$~G. Such bright points have been extensively 

1149: observed in the G-band and other spectral windows previously and 

1150: theoretical and numerical work have explained their appearance: 

1151: The high intensity of bright points is explained by the high magnetic

1152: pressure causing a low gas density. The resulting low opacity means that 

1153: intensity is formed in regions below, where the gas temperature is higher.

1154: However, we also find bright points in the overlying regions formed as 

1155: a result of flux emergence. In the case of the higher lying regions the

1156: bright points are regions of concentrated high magnetic field strength, 

1157: high up- or down-flowing velocity, high velocity convergence and high vorticity.

1158: They are bright in the various spectral bands describing these heights

1159: as a result of high gas temperature due to the compression of the fluid

1160: in these regions.  

1161: 

1162: There are several interesting and important effects that we have not

1163: reported extensively in this paper, amongst these are the effects of joule 

1164: heating and transition region and coronal diagnostics of flux emergence. This

1165: is mainly due to the finding that the flux emergence process is a quite slow one:

1166: it takes on the order of one hour solar time from the time the magnetic field pierces

1167: the photosphere until it reaches and begins to recombine with the pre-existing 

1168: coronal field. We are currently undertaking simulations that have run long enough 

1169: and that have a large enough computational box to study these phenomena.

1170: 

1171: \section{Acknowledgements}

1172: 

1173: This research was supported by the Research Council of Norway through

1174: grant 170935/V30 and through grants of computing time from the Programme for

1175: Supercomputing and by a Marie Curie Early Stage

1176: Research Training Fellowship of the European Community's Sixth Framework

1177: Programme under contract number MESTCT-2005-020395.

1178: Financial support by the European Commission through the SOLAIRE Network (MTRN-CT-2006-035484) is gratefully acknowledged.

1179: To analyze the data we have used IDL and Vapor  (www.vapor.ucar.edu).

1180: 

1181: \clearpage

1182: 

1183: \begin{thebibliography}{}

1184: 

1185: \bibitem[\protect\citeauthoryear{{Abbett}}{{Abbett}}{2007}]{abbett2007}

1186: {Abbett} W.~P., 2007, \apj, 665, 1469

1187: 

1188: \bibitem[\protect\citeauthoryear{{Acheson}}{{Acheson}}{1979}]{Acheson:1979lr}

1189: {Acheson} D.~J., 1979, \solphys, 62, 23

1190: 

1191: \bibitem[\protect\citeauthoryear{{Archontis}, {Moreno-Insertis}, \&

1192:   {Galsgaard}}{{Archontis} et al.}{2004}]{archontis2004}

1193: {Archontis} A., {Moreno-Insertis} F.,  {Galsgaard} K., 2004, A\&A, 426, 1047

1194: 

1195: \bibitem[\protect\citeauthoryear{{Caligari}, {Moreno-Insertis}, \&

1196:   {Schussler}}{{Caligari} et al.}{1995}]{Caligari1995}

1197: {Caligari} P., {Moreno-Insertis} F.,  {Schussler} M., 1995, \apj, 441, 886

1198: 

1199: \bibitem[\protect\citeauthoryear{{Carlsson} et al.}{{Carlsson} et

1200:   al.}{2007}]{Carlsson+etal2007}

1201: {Carlsson} M., {Hansteen} V.~H., {De Pontieu} B., et al., 2007, ArXiv e-prints,

1202:   709

1203: 

1204: \bibitem[\protect\citeauthoryear{Carlsson \& Stein}{Carlsson \&

1205:   Stein}{1992}]{Carlsson+Stein1992}

1206: Carlsson M.,  Stein R.~F., 1992, \apj, 397, L59

1207: 

1208: \bibitem[\protect\citeauthoryear{{Carlsson} \& {Stein}}{{Carlsson} \&

1209:   {Stein}}{1994}]{Carlsson+Stein1994}

1210: {Carlsson} M.,  {Stein} R.~F., 1994, in {Carlsson} M. (ed.),

1211:   Chromospheric Dynamics, p.~47

1212: 

1213: \bibitem[\protect\citeauthoryear{Carlsson \& Stein}{Carlsson \&

1214:   Stein}{1995}]{Carlsson+Stein1995}

1215: Carlsson M.,  Stein R.~F., 1995, \apjl, 440, L29

1216: 

1217: \bibitem[\protect\citeauthoryear{Carlsson \& Stein}{Carlsson \&

1218:   Stein}{1997}]{Carlsson+Stein1997}

1219: Carlsson M.,  Stein R.~F., 1997, \apj, 481, 500

1220: 

1221: \bibitem[\protect\citeauthoryear{Carlsson \& Stein}{Carlsson \&

1222:   Stein}{2002}]{Carlsson+Stein2002}

1223: Carlsson M.,  Stein R.~F., 2002, \apj, 572, 626

1224: 

1225: \bibitem[\protect\citeauthoryear{Carlsson et al.}{Carlsson et

1226:   al.}{2004}]{Carlsson+Stein+Nordlund+Scharmer2004}

1227: Carlsson M., Stein R.~F., Nordlund {\AA}.,  Scharmer G.~B., 2004, \apjl, 610,

1228:   L137

1229: 

1230: \bibitem[\protect\citeauthoryear{{Centeno} et al.}{{Centeno} et

1231:   al.}{2007}]{Centeno:2007lr}

1232: {Centeno} R., {Socas-Navarro} H., {Lites} B., et al., 2007, \apjl, 666, L137

1233: 

1234: \bibitem[\protect\citeauthoryear{{Cheung} \& {Moreno-Insertis}}{{Cheung} \&

1235:   {Moreno-Insertis}}{2006}]{mark2006}

1236: {Cheung} M.~C.~M.,  {Moreno-Insertis} F., 2006, A\&A, 451, 303

1237: 

1238: \bibitem[\protect\citeauthoryear{{Cheung}, {Sch{\"u}ssler}, \&

1239:   {Moreno-Insertis}}{{Cheung} et al.}{2007}]{cheung2007}

1240: {Cheung} M.~C.~M., {Sch{\"u}ssler} M.,  {Moreno-Insertis} F., 2007, \aap, 467,

1241:   703

1242: 

1243: \bibitem[\protect\citeauthoryear{{Dorch} \& {Nordlund}}{{Dorch} \&

1244:   {Nordlund}}{1998}]{Dorch+Nordlund1998}

1245: {Dorch} S.~B.~F.,  {Nordlund} A., 1998, A\&A, 338, 329

1246: 

1247: \bibitem[\protect\citeauthoryear{{Emonet}, {Moreno-Insertis}, \&

1248:   {Rast}}{{Emonet} et al.}{2001}]{emonet2001}

1249: {Emonet} T., {Moreno-Insertis} F.,  {Rast} M.~P., 2001, ApJ, 549, 1212

1250: 

1251: \bibitem[\protect\citeauthoryear{{Evans} \& {Catalano}}{{Evans} \&

1252:   {Catalano}}{1972}]{Evans:1972fk}

1253: {Evans} J.~W.,  {Catalano} C.~P., 1972, \solphys, 27, 299

1254: 

1255: \bibitem[\protect\citeauthoryear{{Fan}}{{Fan}}{2001}]{fan2001}

1256: {Fan} Y., 2001, \apjl, 554, L111

1257: 

1258: \bibitem[\protect\citeauthoryear{{Fan}, {Fischer}, \& {DeLuca}}{{Fan} et

1259:   al.}{1993}]{fan1993}

1260: {Fan} Y., {Fischer} G.~H.,  {DeLuca} E.~E., 1993, Sol Phys, 405, 390

1261: 

1262: \bibitem[\protect\citeauthoryear{{Fan} et al.}{{Fan} et al.}{1998}]{fan1998}

1263: {Fan} Y., {Zweibel} E.~G., {Linton} M.~G.,  {Fischer} G.~H., 1998, ApJ, 505,

1264:   L59

1265: 

1266: \bibitem[\protect\citeauthoryear{{Gudiksen} \& {Nordlund}}{{Gudiksen} \&

1267:   {Nordlund}}{2004}]{Gudiksen+Nordlund2004}

1268: {Gudiksen} B.~V.,  {Nordlund} {\AA}., 2004, {An Ab Initio Approach to the Solar

1269:   Coronal Heating Problem}, in IAU Symposium, Vol. 219, {Dupree} A.~K.,  {Benz}

1270:   A.~O. (eds.), Stars as Suns : Activity, Evolution and Planets, p. 488

1271: 

1272: \bibitem[\protect\citeauthoryear{{Gudiksen} \& {Nordlund}}{{Gudiksen} \&

1273:   {Nordlund}}{2005}]{Gudiksen:2005lr}

1274: {Gudiksen} B.~V.,  {Nordlund} {\AA}., 2005, \apj, 618, 1020, 

1275: 

1276: \bibitem[\protect\citeauthoryear{{Hansteen}}{{Hansteen}}{2004}]{Hansteen2004}

1277: {Hansteen} V.~H., 2004, {Initial simulations spanning the upper convection zone

1278:   to the corona}, in IAU Symposium, Vol. 223, {Stepanov} A.~V.,

1279:   {Benevolenskaya} E.~E.,  {Kosovichev} A.~G. (eds.), Multi-Wavelength

1280:   Investigations of Solar Activity, p. 385

1281: 

1282: \bibitem[\protect\citeauthoryear{{Hansteen}, {Carlsson}, \&

1283:   {Gudiksen}}{{Hansteen} et al.}{2007}]{Hansteen+Carlsson+Gudiksen2007}

1284: {Hansteen} V.~H., {Carlsson} M.,  {Gudiksen} B., 2007, {3D Numerical Models of

1285:   the Chromosphere, Transition Region, and Corona}, in Astronomical Society of

1286:   the Pacific Conference Series, Vol. 368, {Heinzel} P., {Dorotovi{\v c}} I.,

1287:   {Rutten} R.~J. (eds.), The Physics of Chromospheric Plasmas, p. 107

1288: 

1289: \bibitem[\protect\citeauthoryear{{Hansteen} et al.}{{Hansteen} et

1290:   al.}{2007}]{Hansteen+etalSOTEIS2007}

1291: {Hansteen} V.~H., {De Pontieu} B., {Carlsson} M., et al., 2007, ArXiv e-prints,

1292:   711

1293: 

1294: \bibitem[\protect\citeauthoryear{{Hyman}, {Vichnevtsky}, \&

1295:   {Stepleman}}{{Hyman} et al.}{1979}]{Hyman1979}

1296: {Hyman} J., {Vichnevtsky} R.,  {Stepleman} R., 1979, Adv. in Comp. Meth,

1297:   PDE's-III, 313

1298: 

1299: \bibitem[\protect\citeauthoryear{{Judge} \& {Meisner}}{{Judge} \&

1300:   {Meisner}}{1994}]{HAO_Diaper1994}

1301: {Judge} P.~G.,  {Meisner} R.~W., 1994, {The `HAO spectral diagnostics package'

1302:   (HAOS-Diaper)}, in ESA Special Publication, Vol. 373, {Hunt} J.~J. (ed.),

1303:   Solar Dynamic Phenomena and Solar Wind Consequences, the Third SOHO Workshop,

1304:   p.~67

1305: 

1306: \bibitem[\protect\citeauthoryear{{Leenaarts} \&

1307:   {Wedemeyer-B{\"o}hm}}{{Leenaarts} \&

1308:   {Wedemeyer-B{\"o}hm}}{2005}]{Leenaarts:2005yq}

1309: {Leenaarts} J.,  {Wedemeyer-B{\"o}hm} S., 2005, \aap, 431, 687

1310: 

1311: \bibitem[\protect\citeauthoryear{{Linton}, {Longcope}, \& {Fisher}}{{Linton} et

1312:   al.}{1996}]{linton1996}

1313: {Linton} M.~G., {Longcope} D.~W.,  {Fisher} G.~H., 1996, ApJ, 469, 954

1314: 

1315: \bibitem[\protect\citeauthoryear{{Lites} et al.}{{Lites} et

1316:   al.}{1995}]{lites1995}

1317: {Lites} B.~W., V.~M., {Seagraves} P.,  {Skunanick} A., 1995, ApJ, 446, 877

1318: 

1319: \bibitem[\protect\citeauthoryear{{Longcope}, {Fischer}, \& {Arendt}}{{Longcope}

1320:   et al.}{1996}]{longcope1996}

1321: {Longcope} D.~W., {Fischer} G.~H.,  {Arendt} S., 1996, ApJ, 464, 999

1322: 

1323: \bibitem[\protect\citeauthoryear{{Mackay} \& {Galsgaard}}{{Mackay} \&

1324:   {Galsgaard}}{2001}]{Mackay+Galsgaard2001}

1325: {Mackay} D.~H.,  {Galsgaard} K., 2001, \solphys, 198, 289

1326: 

1327: \bibitem[\protect\citeauthoryear{{Magara}}{{Magara}}{2004}]{magara2004}

1328: {Magara} T., 2004, \apj, 605, 480

1329: 

1330: \bibitem[\protect\citeauthoryear{{Magara}}{{Magara}}{2006}]{Magara:2006lr}

1331: {Magara} T., 2006, \apj, 653, 1499

1332: 

1333: \bibitem[\protect\citeauthoryear{{Magara} \& {Longcope}}{{Magara} \&

1334:   {Longcope}}{2001}]{magara2001}

1335: {Magara} T.,  {Longcope} D.~W., 2001, \apjl, 559, L55

1336: 

1337: \bibitem[\protect\citeauthoryear{{Manchester} et al.}{{Manchester} et

1338:   al.}{2004}]{manchester2004}

1339: {Manchester} W., IV, {Gombosi} T., {DeZeeuw} D.,  {Fan} Y., 2004, \apj, 610,

1340:   588

1341: 

1342: \bibitem[\protect\citeauthoryear{{Matsumoto} et al.}{{Matsumoto} et

1343:   al.}{1998}]{matsumoto1998}

1344: {Matsumoto} R., {Tajima} T., {Chou} W., {Okubo} A.,  {Shibata} K., 1998, \apjl,

1345:   493, L43

1346: 

1347: \bibitem[\protect\citeauthoryear{{Moreno-Insertis}}{{Moreno-Insertis}}{1986}]{%

1348: moreno1986}

1349: {Moreno-Insertis} F., 1986, A\&A, 166, 291

1350: 

1351: \bibitem[\protect\citeauthoryear{{Moreno-Insertis} \&

1352:   {Emonet}}{{Moreno-Insertis} \& {Emonet}}{1996}]{moreno1996}

1353: {Moreno-Insertis} F.,  {Emonet} T., 1996, ApJ, 472, L53

1354: 

1355: \bibitem[\protect\citeauthoryear{{Nordlund}}{{Nordlund}}{1982}]{Nordlund1982}

1356: {Nordlund} {\AA}., 1982, Aap, 107, 1

1357: 

1358: \bibitem[\protect\citeauthoryear{{Rutten}, {de Wijn}, \&

1359:   {S{\"u}tterlin}}{{Rutten} et al.}{2004}]{Rutten:2004kx}

1360: {Rutten} R.~J., {de Wijn} A.~G.,  {S{\"u}tterlin} P., 2004, \aap, 416, 333

1361: 

1362: \bibitem[\protect\citeauthoryear{{Rutten} \& {Krijger}}{{Rutten} \&

1363:   {Krijger}}{2003}]{Rutten:2003fj}

1364: {Rutten} R.~J.,  {Krijger} J.~M., 2003, \aap, 407, 735

1365: 

1366: \bibitem[\protect\citeauthoryear{{Sch\"ussler}}{{Sch\"ussler}}{1979}]{Shussler1979}

1367: {Sch\"ussler} M., 1979, A\&A, 71, 79

1368: 

1369: \bibitem[\protect\citeauthoryear{{Shelyag} et al.}{{Shelyag} et

1370:   al.}{2004}]{Shelyag:2004lr}

1371: {Shelyag} S., {Sch{\"u}ssler} M., {Solanki} S.~K., {Berdyugina} S.~V.,

1372:   {V{\"o}gler} A., 2004, \aap, 427, 335

1373: 

1374: \bibitem[\protect\citeauthoryear{{Shibata} et al.}{{Shibata} et

1375:   al.}{1989}]{Shibata:1989lr}

1376: {Shibata} K., {Tajima} T., {Steinolfson} R.~S.,  {Matsumoto} R., 1989, \apj,

1377:   345, 584

1378: 

1379: \bibitem[\protect\citeauthoryear{{Skartlien}}{{Skartlien}}{2000}]{Skartlien200%

1380: 0}

1381: {Skartlien} R., 2000, \apj, 536, 465

1382: 

1383: \bibitem[\protect\citeauthoryear{{Skartlien}, {Stein}, \&

1384:   {Nordlund}}{{Skartlien} et al.}{2000}]{Skartlien:2000lr}

1385: {Skartlien} R., {Stein} R.~F.,  {Nordlund} {\AA}., 2000, \apj, 541, 468

1386: 

1387: \bibitem[\protect\citeauthoryear{{Stein}, {Bercik}, \& {Nordlund}}{{Stein} et

1388:   al.}{2003}]{Stein:2003uq}

1389: {Stein} R.~F., {Bercik} D.,  {Nordlund} {\AA}., 2003, {Solar Surface

1390:   Magneto-Convection}, in Astronomical Society of the Pacific Conference

1391:   Series, Vol. 286, {Pevtsov} A.~A.,  {Uitenbroek} H. (eds.), Current

1392:   Theoretical Models and Future High Resolution Solar Observations: Preparing

1393:   for ATST, p. 121

1394: 

1395: \bibitem[\protect\citeauthoryear{{Stein} \& {Nordlund}}{{Stein} \&

1396:   {Nordlund}}{2002}]{Stein:2002fk}

1397: {Stein} R.~F.,  {Nordlund} {\AA}., 2002, {Solar Surface Magneto-Convection and

1398:   Dynamo Action}, in ESA Special Publication, Vol. 505, {Sawaya-Lacoste} H.

1399:   (ed.), SOLMAG 2002. Proceedings of the Magnetic Coupling of the Solar

1400:   Atmosphere Euroconference, p.~83

1401: 

1402: \bibitem[\protect\citeauthoryear{{Stein} \& {Nordlund}}{{Stein} \&

1403:   {Nordlund}}{2006}]{Stein:2006qy}

1404: {Stein} R.~F.,  {Nordlund} {\AA}., 2006, \apj, 642, 1246

1405: 

1406: \bibitem[\protect\citeauthoryear{{Suemoto}, {Hiei}, \& {Nakagomi}}{{Suemoto} et

1407:   al.}{1987}]{Suemoto:1987qy}

1408: {Suemoto} Z., {Hiei} E.,  {Nakagomi} Y., 1987, \solphys, 112, 59

1409: 

1410: \bibitem[\protect\citeauthoryear{{Suemoto}, {Hiei}, \& {Nakagomi}}{{Suemoto} et

1411:   al.}{1990}]{Suemoto:1990uq}

1412: {Suemoto} Z., {Hiei} E.,  {Nakagomi} Y., 1990, \solphys, 127, 11

1413: 

1414: \bibitem[\protect\citeauthoryear{{V{\"o}gler} et al.}{{V{\"o}gler} et

1415:   al.}{2005}]{Vogler:2005fj}

1416: {V{\"o}gler} A., {Shelyag} S., {Sch{\"u}ssler} M., et al., 2005, \aap, 429,

1417:   335

1418: 

1419: \bibitem[\protect\citeauthoryear{{Wedemeyer} et al.}{{Wedemeyer} et

1420:   al.}{2004}]{Wedemeyer:2004}

1421: {Wedemeyer} S., {Freytag} B., {Steffen} M., {Ludwig} H.-G.,  {Holweger} H.,

1422:   2004, \aap, 414, 1121

1423: 

1424: \end{thebibliography}

1425: 

1426: \clearpage

1427: 

1428: \begin{figure}

1429:   \includegraphics[width=9.5cm]{f1a.ps}

1430:   \includegraphics[width=9.5cm]{f1b.ps}

1431:  \caption{\label{fig:initsetup}Left panel: Temperatures (solid) and densities (dashed). Right panel: Gas (solid) and 

1432:           magnetic (dashed) pressures. All as a function of height in the initial model. 

1433:           Minimum and maximum values are shown in grey, average values are shown in black. The

1434:           photosphere is situated at $z\approx0$~Mm. The models extent in height is from $-1.4$ to 

1435:           $14$~Mm (top 4 Mm not shown here).}

1436:  \end{figure}

1437: 

1438: \clearpage

1439: 

1440: \begin{figure}

1441:   \includegraphics[width=9.5cm]{f2.ps}

1442:   \caption{\label{fig:initbox}The computational box from the convection zone to the corona. 

1443:            The color table is the temperature.

1444:            The magnetic flux tube coming from the bottom 

1445:            boundary (green lines). The initial (weak) background magnetic field

1446:            lines (red lines). 

1447:            The {\em z} axis runs from the corona ($z=14$~Mm top boundary) 

1448:            to the convection zone ($z=-1.4$~Mm bottom boundary). We center

1449:            our attention to 4 different heights shown in the figure as partial

1450:            planes with the

1451:            temperature (grey-scale coded); at heights $z=10$~km (photosphere),

1452:            $z=234$~km (reverse granulation), $z=458$~km (photosphere-chromosphere)

1453:            and $z=906$~km (mid-chromosphere).

1454:            The black lines at the left and right side show the 

1455:            heights of the four layers.}

1456: \end{figure}

1457: 

1458: \clearpage

1459: 

1460: \begin{figure}

1461: \begin{center}

1462: \includegraphics[width=7.5cm]{f3a.ps}

1463: \includegraphics[width=7.5cm]{f3b.ps}

1464: \includegraphics[width=7.5cm]{f3c.ps}

1465: \includegraphics[width=7.5cm]{f3d.ps}

1466: \includegraphics[width=7.5cm]{f3e.ps}

1467: \includegraphics[width=7.5cm]{f3f.ps}

1468: \end{center}

1469: \caption{\label{fig:intphot} The continuum intensity shown at 

1470: times 600~s, before the flux tube goes through 

1471: the photosphere, 

1472: at time 1100~s, when the flux tube is close to the

1473: photosphere, 

1474: at time 1210~s, when the big cells are cooling by

1475: expansion,

1476: at time 1660~s, when the tube is crossing the photosphere,

1477:  at time 2200~s, when the structure of the cells appears to have 

1478: returned back to normal and at time 3220~s, from the left to

1479: right and top to bottom respectively. The red contours correspond to

1480: positive $B_z$ and green contours to negative $B_z$ at $z=10$~km.

1481: An example of a collapsed granule is marked with a blue line.}

1482: \end{figure}

1483: 

1484: \clearpage

1485: 

1486: \begin{figure}

1487: \begin{center}

1488: 	\includegraphics[width=7.5cm]{f4a.ps}

1489: 	\includegraphics[width=7.5cm]{f4b.ps}

1490: 	\includegraphics[width=7.5cm]{f4c.ps}

1491: 	\includegraphics[width=7.5cm]{f4d.ps}

1492: 	\includegraphics[width=7.5cm]{f4e.ps}

1493: 	\includegraphics[width=7.5cm]{f4f.ps}

1494:  \end{center}

1495:  \caption{\label{fig:fieldphot} 

1496:  $B_z$ (longitudinal ``magnetograms'') in the photosphere, at $z=10$~km,

1497:  shown with grey-scale intensity at three different times in the simulation in the left panels. 

1498:  $B_h$ (horizontal magnetic field) in the photosphere at the same height and same instants shown in the 

1499:  right panels. The grey-scale goes from -1111.5~G to 1376.3~G. The times chosen are 

1500:   at time $1460$~s, when the tube just has started to cross the photosphere (top panels); 

1501:  at time $1710$~s, as the tube is well into the photosphere (middle panels); 

1502:  and at time $3220$~s (bottom panels).}

1503: \end{figure}

1504: 

1505: \clearpage

1506: 

1507: \begin{figure}

1508: \begin{center}

1509: \includegraphics[width=7.5cm]{f5a.ps}

1510: \includegraphics[width=7.5cm]{f5b.ps}

1511:   \end{center}

1512:   \caption{\label{fig:3dpho} 3D image of the simulation at time $t=1700$~s.

1513:   Temperature at the photosphere shown with grey-scale intensity, the red

1514:   lines are the magnetic field lines of the tube, temperatures in 

1515:   the 

1516:   corona shown with green-violet scale color (green is the transition

1517:   region) and with orange-scale is the magnetic field strength. 

1518:   View from above (left panel) and from below (right panel).}

1519: \end{figure}

1520: 

1521: \clearpage

1522: 

1523: \begin{figure}

1524: \includegraphics[width=7.5cm]{f6.ps}	

1525:   \caption{\label{fig:fluxret} Mean magnetic flux per unit area

1526:   perpendicular to the tube as function of time. Region confined between the

1527:   $z=10$~km and $z=254$~km in black color

1528:   and below the photosphere in red color for the runs A1 (dash-dot line),

1529:   A2 (dashed line), A3 (dotted line) and A4 (solid line).}

1530: \end{figure}

1531: 

1532: \clearpage

1533: 

1534: \begin{figure}

1535: \begin{center}

1536: \includegraphics[width=7.5cm]{f7a.ps}

1537: \includegraphics[width=7.5cm]{f7b.ps}

1538:   \end{center}

1539:   \caption{\label{fig:3drev} 3D image of the simulation at time $t=1900$~s.

1540:   Temperature at the height ($z=234$~km) shown

1541:   with grey-scale intensity. The rest of the color scale is the

1542:   same as in figure~\ref{fig:3dpho}. 

1543:   View from above (left panel) and from below (right panel).}

1544: \end{figure}

1545: 

1546: \clearpage

1547: 

1548: \begin{figure}

1549: \begin{center}

1550: \includegraphics[width=7.5cm]{f8a.ps}

1551: \includegraphics[width=7.5cm]{f8b.ps}

1552: \includegraphics[width=7.5cm]{f8c.ps}

1553: \includegraphics[width=7.5cm]{f8d.ps}

1554: \includegraphics[width=7.5cm]{f8e.ps}

1555: \includegraphics[width=7.5cm]{f8f.ps}

1556: \includegraphics[width=7.5cm]{f8g.ps}

1557: \includegraphics[width=7.5cm]{f8h.ps}

1558:   \end{center}

1559:   \caption{\label{fig:revgran} Continuum intensity at 170~nm (top row),

1560:   vertical velocity (second row), 

1561:   divergence of the velocity (third row) and 

1562:   the vertical vorticity (bottom row) at $z=234$~km 

1563:   at time 1620~s (left column) and 3220~s (right column).}

1564: \end{figure}

1565: 

1566: \clearpage

1567: 

1568: \begin{figure}

1569: \begin{center}

1570: 	\includegraphics[width=7.5cm]{f9a.ps}

1571: 	\includegraphics[width=7.5cm]{f9b.ps}

1572: 	\includegraphics[width=7.5cm]{f9c.ps}

1573: 	\includegraphics[width=7.5cm]{f9d.ps}

1574: 	\includegraphics[width=7.5cm]{f9e.ps}

1575: 	\includegraphics[width=7.5cm]{f9f.ps}

1576:  \end{center}

1577:  \caption{\label{fig:fieldlwch} $B_z$ 

1578:  at $z=234$~km (left panels) and $B_h$ (right panels).

1579:  The grey-scale goes from -560.3~G to 659.5~G. 

1580:  The field is shown at times 

1581:  1460~s (top panels) when the tube start to cross the 

1582:  reverse granulation layer, 

1583:  at time 2110~s when the tube is crossing the reverse granulation (middle panels) and

1584:   at time 3220~s when some rest of the flux returns to the reverse granulation (bottom panels)}

1585: \end{figure}

1586: 

1587: \clearpage

1588: 

1589: \begin{figure}

1590: \begin{center}

1591: \includegraphics[width=7.5cm]{f10.ps}	

1592:   \end{center}

1593:   \caption{\label{fig:fluxretlow} Variation in time of the mean magnetic 

1594:   flux by area  as defined in equation \ref{eq:phi2} from the photosphere ($z=10$~km) to the

1595:   height $z=234$~km with black color and from the height  

1596:   $z=234$~km to the height $z=458$~km with red

1597:   color for the runs A1 (dash-dot line), A2 (dashed line), A3 (dotted

1598:   line) and A4 (solid line).}

1599: \end{figure}

1600: 

1601: \clearpage

1602: 

1603: \begin{figure}

1604: \begin{center}

1605: \includegraphics[width=7.5cm]{f11a.ps}

1606: \includegraphics[width=7.5cm]{f11b.ps}

1607:   \end{center}

1608:   \caption{\label{fig:3dcro} 3D image of the simulation at time $t=1980$~s.

1609:   Temperature at $z=458$~km shown with

1610:   grey-scale intensity. The rest of the color scale is the

1611:   same as in figure~\ref{fig:3dpho}. 

1612:   View from above (left panel) and from below (right panel).}

1613: \end{figure}

1614: 

1615: \clearpage

1616: 

1617: \begin{figure}

1618: \begin{center}

1619: 	\includegraphics[width=7.5cm]{f12a.ps}

1620: 	\includegraphics[width=7.5cm]{f12b.ps}

1621: 	\includegraphics[width=7.5cm]{f12c.ps}

1622: 	\includegraphics[width=7.5cm]{f12d.ps}

1623:  \end{center}

1624:  \caption{\label{fig:fieldcr} $B_z$ at $z=458$~km (left panels)

1625:  and $B_h$ (right panels) at times 

1626:  2110~s when the tube is crossing the layer $458$~km height

1627:  (top panels)

1628:  and at time 3220~s (bottom panels). 

1629:  The grey-scale range is [-217,284] G for all panels.}

1630: \end{figure}

1631: 

1632: \clearpage

1633: 

1634: \begin{figure}

1635: \begin{center}

1636: \includegraphics[width=7.5cm]{f13a.ps}

1637: \includegraphics[width=7.5cm]{f13b.ps}

1638: \includegraphics[width=7.5cm]{f13c.ps}

1639: \includegraphics[width=7.5cm]{f13d.ps}

1640:   \end{center}

1641:   \caption{\label{fig:tgcr} Continuum intensity to the power

1642:   of 0.4 (equivalent to a gamma of 0.4 for an image) to better bring out

1643:   the structure,

1644:   at 160~nm (top row) and 130~nm (bottom row)

1645:   at time 1620~s when the tube starts to

1646:   enter the chromosphere (left column)

1647:   and at 3220~s when the tube has crossed the height $z=458$~km

1648:   (right column)}

1649: \end{figure}

1650: 

1651: \clearpage

1652: 

1653: \begin{figure}

1654: \begin{center}

1655: \includegraphics[width=7.5cm]{f14a.ps}

1656: \includegraphics[width=7.5cm]{f14b.ps}

1657: \includegraphics[width=7.5cm]{f14c.ps}

1658: \includegraphics[width=7.5cm]{f14d.ps}

1659: \includegraphics[width=7.5cm]{f14e.ps}

1660: \includegraphics[width=7.5cm]{f14f.ps}

1661:   \end{center}

1662:   \caption{\label{fig:divucr} 

1663:   Vertical velocity (top row), 

1664:   divergence of the velocity (middle row) and 

1665:   the vertical vorticity (bottom row) at $z=458$~km 

1666:   at time 1620~s, when the flux tube is close to the 

1667:   layer at height $z=458$~km (left column), 

1668:   and at time 3220~s, when the tube has crossed the layer $458$~km height

1669:   (right column).}

1670: \end{figure}

1671: 

1672: \clearpage

1673: 

1674: \begin{figure}

1675: \includegraphics[width=7.5cm]{f15.ps}	

1676:   \caption{\label{fig:fluxretup} Mean magnetic flux per unit area

1677:   perpendicular to the tube as function of time. Region confined between

1678:   $z=234$~km to $z=458$~km in red color and

1679:   from $z=458$~km to $z=901$~km in black color for the

1680:   runs A1 (dash-dot line), A2 (dashed line), A3 (dotted line) and A4

1681:   (solid line).}

1682: \end{figure}

1683: 

1684: \clearpage

1685: 

1686: \begin{figure}

1687: \begin{center}

1688: 	\includegraphics[width=7.5cm]{f16a.ps}

1689: 	\includegraphics[width=7.5cm]{f16b.ps}

1690: 	\includegraphics[width=7.5cm]{f16c.ps}

1691: 	\includegraphics[width=7.5cm]{f16d.ps}

1692: 	\includegraphics[width=7.5cm]{f16e.ps}

1693: 	\includegraphics[width=7.5cm]{f16f.ps}

1694:  \end{center}

1695:  \caption{\label{fig:fieldupcr} $B_z$ at $z=906$~km 

1696:   (left panels), and $B_h$ (right panels) at times 600~s, before the flux

1697:  tube starts to cross the chromosphere (top panels),

1698:  at time 2800~s when the tube is crossing the photosphere (middle panels) 

1699:  and at time 3220~s (bottom panels). 

1700:  The grey-scale goes from -60.46~G to 65.91~G}

1701: \end{figure}

1702: 

1703: \clearpage

1704: 

1705: \begin{figure}

1706: \begin{center}

1707: \includegraphics[width=7.5cm]{f17a.ps}

1708: \includegraphics[width=7.5cm]{f17b.ps}

1709: \includegraphics[width=7.5cm]{f17c.ps}

1710: \includegraphics[width=7.5cm]{f17d.ps}

1711: \includegraphics[width=7.5cm]{f17e.ps}

1712: \includegraphics[width=7.5cm]{f17f.ps}

1713:   \end{center}

1714:   \caption{\label{fig:tgupcr} Continuum intensity of the C~{\sc i} to the 

1715:   power of 0.4 (equivalent to a gamma of 0.4 for an image) to better

1716:   bring out the structure at 109.9~nm

1717:   at time 1620~s (top left) and

1718:   at time 3220~s (top right). The synthetic Ca II H line intensity 

1719:  at time 600~s, before the flux tube goes through 

1720:   the chromosphere, 

1721:   at time 1620~s when the tube starts to cross, 

1722:   at time 2110~s as the tube is crossing,

1723:   and at 3220~s when the tube has crossed the chromosphere, 

1724:   from the left to the right and top to bottom respectively.}

1725: \end{figure}

1726: 

1727: \clearpage

1728: 

1729: \begin{figure}

1730: \begin{center}

1731: \includegraphics[width=7.5cm]{f18a.ps}

1732: \includegraphics[width=7.5cm]{f18b.ps}

1733: \includegraphics[width=7.5cm]{f18c.ps}

1734: \includegraphics[width=7.5cm]{f18d.ps}

1735:   \end{center}

1736:   \caption{\label{fig:uzupcr} Vertical velocity at $z=906$~km 

1737:   at times 600~s, before the flux tube goes through the chromosphere (top left), 

1738:   at time 2110~s, when

1739:   the flux tube is closed to the chromosphere (top right), 

1740:   at time 2800~s, when

1741:   the big cells are cooling through expansion (bottom left),

1742:   and at time 3220~s, after the tube has passed through the

1743:   chromosphere (bottom right).}

1744: \end{figure}

1745: 

1746: \clearpage

1747: 

1748: \begin{figure}

1749: \begin{center}

1750: \includegraphics[width=7.5cm]{f19a.ps}

1751: \includegraphics[width=7.5cm]{f19b.ps}

1752:   \end{center}

1753:   \caption{\label{fig:tgxz} Temperature shown with grey-scale in a

1754:   vertical plane at 3220~s for the simulations A4 (left) and for the simulation

1755:   B1 at time 3370~s (right). Note the different scales.}

1756: \end{figure}

1757: 

1758: \clearpage

1759: 

1760: \begin{figure}

1761: \begin{center}

1762: \includegraphics[width=7.5cm]{f20a.ps}

1763: \includegraphics[width=7.5cm]{f20b.ps}

1764: \includegraphics[width=16cm]{f20c.ps}

1765:   \end{center}

1766:   \caption{\label{fig:caxz} The synthetic Ca II H line Hinode intensity 

1767:    to the power of 0.25 (equivalent to a 

1768:    gamma of 0.25 in an image) is plotted to better

1769:   bring out the structure of the bubble from simulation A4 

1770:   at time 600~s, before the flux tube goes through 

1771:   the chromosphere (top left), 

1772:   and at 3220~s when it has crossed the chromosphere (top right), 

1773:   and from the simulation

1774:   B1 at time 3370~s (bottom)}

1775: \end{figure}

1776: 

1777: \clearpage

1778: 

1779: \begin{figure}

1780: \begin{center}

1781: \includegraphics[width=13.5cm]{f21a.ps}

1782: \includegraphics[width=13.5cm]{f21b.ps}

1783: \includegraphics[width=13.5cm]{f21c.ps}

1784:   \end{center}

1785:   \caption{\label{fig:tg_bubble} Vertical cuts of the logarithm of the temperature (top),

1786:   vertical velocity (middle) and the logarithm of the gas pressure (bottom) from simulation 

1787:   B1 at time $t=2580$~s and $y=0$~Mm. Contours of plasma $\beta$ are overplotted the temperature.}

1788: \end{figure}

1789:  

1790: \clearpage

1791: 

1792: \begin{figure}

1793: \begin{center}

1794: \includegraphics[width=13.5cm]{f22a.ps}

1795: \includegraphics[width=13.5cm]{f22b.ps}

1796: \includegraphics[width=13.5cm]{f22c.ps}

1797:   \end{center}

1798:   \caption{\label{fig:tgxt} Temperature (top), vertical velocity

1799:   (middle) and logarithm of the pressure (bottom) at a height of

1800:   1.1 Mm as function of position

1801:   in $x$ and time for a vertical cut at $y=0$~Mm for the simulation B1.}

1802: \end{figure}

1803: 

1804: \clearpage

1805: 

1806: \begin{deluxetable}{cccccc}

1807: \tablecaption{\label{tab:runs} Simulation description}

1808: \tablehead{

1809: \colhead{Name} & \colhead{Twist $\lambda$}  & \colhead{B$_0$ [G]} & \colhead{Size} & \colhead{Comment} & \colhead{Time [s]}

1810: }

1811: \startdata

1812: A1 & 0.1 & 4484 & $8\times 4\times 16$~Mm$^{3}$ & Flux tube in $y$ direction & 2200 \\

1813: A2 & 0.2 & 3363 & $8\times 4\times 16$~Mm$^{3}$ & Flux tube in $y$ direction & 2100 \\

1814: A3 & 0.3 & 4484 & $8\times 4\times 16$~Mm$^{3}$ & Flux tube in $y$ direction & 2500 \\

1815: A4 & 0.6 & 4484 & $8\times 4\times 16$~Mm$^{3}$ & Flux tube in $y$ direction & 3200 \\

1816: B1 & 0.0 & 1121 & $16\times 8\times 16$~Mm$^{3}$ & Flux sheet in $y$ direction. & 4500 \\

1817: \enddata

1818: \end{deluxetable}

1819: 

1820: \clearpage

1821: 

1822: \begin{deluxetable}{lc}

1823: \tablecaption{\label{tab:events} time evolution}

1824: \tablehead{

1825: \colhead{Process} & \colhead{Time[s]}

1826: }

1827: \startdata

1828: Expansion of granular cells in the photosphere & 900 \\

1829: Expansion of the reverse granulation & 920  \\ 

1830: Cooling of the center of granular cells in the photosphere & 1160 \\

1831: Tube crosses the photosphere & 1300  \\

1832: Tube crosses the chromospheric height forming reverse granulation & 1380 \\ 

1833: Magnetic field moved to the intergranular lanes in the photosphere & 1550 \\

1834: Reverse granulation cooling with expansion& 1560 \\  

1835: Tube crosses the layer $z=450$~km & 1660 \\

1836: The cells return to ``normal'' size in the photosphere & 2100 \\ 

1837: Upper chromosphere vertical expansion and cooling & 2900 \\ 

1838: \enddata

1839: \end{deluxetable}

1840: 

1841: \end{document}

1842: