1: %&LaTeX

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

3: \def\etal{{\it et al.\ }}

4: \begin{document}

5: \title{Generation of Flows in the Solar Chromosphere Due to 

6: Magnetofluid Coupling}

7: \author{Swadesh M. Mahajan\altaffilmark{1}}

8: \affil{Institute for Fusion Studies, The University of Texas at  Austin, 

9: Austin, Texas 78712}

10: \author{Komunela I. Nikol'skaya\altaffilmark{2}}

11: \affil{Institute of Terrestrial Magnetism, Ionosphere and Radio Wave 

12: Propagation, Troitsk of Moscow Region, 142092, Russia}

13: \author{Nana L. Shatashvili\altaffilmark{\,3,0} and Zensho 

14: Yoshida\altaffilmark{4}}

15: \affil{Graduate School of Frontier Sciences and High Temperature 

16: Plasma Center, The University of Tokyo, Tokyo 113--0033, Japan}

17: \altaffiltext{0}{\small {\it Permanent address:} \ Plasma Physics 

18: Department, Tbilisi State University, Tbilisi 380028, Georgia} 

19: \altaffiltext{1}{\small Electronic mail: \ mahajan@mail.utexas.edu}

20: \altaffiltext{2}{\small Electronic mail: \ knikol@izmiran.troitsk.ru}

21: \altaffiltext{3}{\small Electronic mail: \ 

22: nana@plasma.q.t.u-tokyo.ac.jp\hskip 0.3cm nanas@iberiapac.ge}

23: \altaffiltext{4}{\small Electronic mail: \ yoshida@k.u-tokyo.ac.jp}

24: 

25: \clearpage

26: 

27: \begin{abstract}

28:  

29: It is shown that a generalized magneto--Bernoulli mechanism can 

30: effectively generate high velocity flows in the Solar chromosphere by 

31: transforming the plasma pressure energy into kinetic energy. It is 

32: found that at reasonable heights and for realistic plasma parameters, 

33: there is a precipitous pressure fall accompanied by  a sharp 

34: amplification of the flow speed.  

35: \end{abstract}

36: 

37: \keywords{Sun: atmosphere --- Sun: chromosphere --- Sun: corona --- 

38: Sun: magnetic fields --- Sun: transition region}

39: 

40: \clearpage

41: 

42: \section{Introduction}

43: \label{sec:intro} 

44: \bigskip

45: 

46: Recent observations, strongly fortified by immensely improved measuring 

47: and interpretive capabilities, have convincingly demonstrated that the 

48: solar corona is a highly dynamic arena replete with multiple--scale 

49: spatiotemporal structures (Aschwanden \etal 2001a).  A major new advance is the 

50: discovery that strong flows are found everywhere --- in the subcoronal 

51: (chromosphere) as well as in the coronal  regions  (see e.g. 

52: (Schrijver \etal 1999; Winebarger, LeLuca and Golub 2001; Wilhelm 

53: 2001; Aschwanden \etal 2001a; Aschwanden \etal 2001b; Seaton \etal 

54: 2001; Winebarger \etal 2002) and references therein). Equally important is the growing

55: belief and realization that the plasma flows may complement 

56: the abilities of the magnetic field in the creation of the 

57: amazing richness observed in the coronal structures.

58:  

59: Interestingly enough, even before the observational mandate, theoretical 

60: efforts in harnessing the plasma flows to solve some of the riddles of 

61: solar physics had already begun. In particular the dynamics of flow--based 

62: structure creation and heating was the subject of Mahajan \etal (1999, 2001); 

63: in this model the flows provided the basic material as well as  energy for 

64: the primary heating of the coronal loops. A systematic treatment of loop models

65: that include flows was also developed by Orlando, Peres and Serio (1995a, 1995b), 

66: and by Mahajan \etal (1999, 2001).

67: 

68: If flows are to play an important and essential role in determining the 

69: dynamics and structure of the solar corona, we must immediately face the 

70: problem of finding sources and mechanisms for the creation of these flows. 

71: Catastrophic models of flow production in which the magnetic energy is 

72: suddenly converted into bulk kinetic energy (and thermal energy) are 

73: rather well--known; various forms of magnetic reconnection (flares, micro 

74: and nanoflares) schemes permeate the literature (see e.g.~(Wilhelm 

75: 2001; Christopoilou, Georgakilas and Koutchmy 2001)

76: for chromosphere up--flow generations). A few 

77: other  mechanism of this genre also exist: Uchida \etal (2001) proposed 

78: that the major part of the supply of energy and mass to the active regions 

79: of the corona may come from a dynamical leakage of magnetic twists produced in 

80: the subphotospheric convection layer; Ohsaki, Shatashvili, Yoshida and Mahajan (2001,2002) 

81: have shown how a slowly evolving closed structure 

82: (modelled as a double--Beltrami two--fluid equilibrium)  may experience, 

83: under appropriate conditions,  a sudden loss of equilibrium with the 

84: initial magnetic energy appearing as the mass flow energy. Another   

85: mechanism, based on loop interactions and fragmentations and  explaining 

86: the formation of loop threads, was given in Sakai and Furusawa (2002). These 

87: mechanisms, though extremely interesting, are an unlikely source to account 

88: for  the observed ubiquity of plasma flows.  One should, perhaps, look for 

89: relatively gentler, more widespread, and  steadier mechanisms.

90: 

91: Before exploring new avenues for flow generation, and deciding which 

92: region of the solar atmosphere should be the first target for 

93: investigation, we seek some guidance from phenomenology. Based on 

94: estimates of energy fluxes required to heat the chromosphere and the 

95: corona, Goodman (2001) has shown  that the  mechanism which transports 

96: mechanical energy from the convection zone to the chromosphere 

97: (to sustain its heating rate) could also supply the energy needed 

98: to heat the corona, and accelerate the solar wind (SW). The coronal

99: heating problem, in this context, is shifted to the problem of the

100: dynamic energization of the chromosphere. In the latter process the 

101: role of flows is found to be critical as warranted by the following 

102: observations made in soft X--rays and extreme ultraviolet (EUV) wavelengths, 

103: and recent findings from the {\it Transition Region and Coronal Explorer 

104: (TRACE)}: the overdensity of coronal loops, the chromospheric up--flows 

105: of heated plasma, and the localization of the heating function in the 

106: lower corona (e.g. Schrijver, \etal 1999; Aschwanden \etal 2001a; 

107: Aschwanden 2001b) and references therein). The

108: main message then, is that the coronal heating problem may only be solved

109: by including processes (including the flow dynamics) in the chromosphere 

110: and the transition region. The challenge, therefore, is to develop a 

111: semi--steady state theory of flow generation in the chromosphere.

112: 

113: There are only two obvious energy sources that could power flow generation 

114: in the chromosphere: the magnetic field, and the thermal pressure of the 

115: plasma. We have already mentioned a few examples of the magnetically 

116: driven transient but sudden flow--generation. Looking for a more quiescent pathway, 

117: we shall now concentrate on the latter source. To convert thermal energy into kinetic 

118: energy, some variant of the Bernoulli mechanism (the existence of larger kinetic energy

119: in regions of lower pressure) must be invoked. We will soon show that  the 

120: double--Beltrami--Bernoulli states accessible to a two--fluid  system in 

121: which the velocity field is formally treated at par with the magnetic 

122: field (Mahajan and Yoshida 1998), can readily provide the necessary framework for this conversion.

123: 

124: \section{Model}

125: \label{sec:Model}

126: 

127: In astrophysics (particularly  in the physics of the solar 

128: atmosphere), it is useful to assign at least two distinct connotations

129: to the plasma ``flow'': 1)~the flow is a primary object whose dynamics bears 

130: critically on the phenomena under investigation. The problems of the formation 

131: and the original heating of the coronal structure, the creation of channels for 

132: particle escape, for instance, fall in this category, 2)~The flow is a secondary 

133: feature of the system, possibly created as a by-product (e.g. see 

134: Ohsaki \etal (2002)) and/or used to drive or suppress an instability. 

135: Since the generation of flows 

136: which will eventually create the coronal loops is the theme of this effort, the 

137: flows  here are  fundamental. 

138: 

139: As pointed out earlier, our theoretical model is based on a rather simple 

140: application of the magnetofluid theory developed over the last few 

141: years. We plan to restrict ourselves to almost steady state considerations 

142: (for a steady and continuous supply of plasma flows emerging from  the 

143: chromosphere). Very near the photospheric surface, the influence of 

144: neutrals and ionization processes (and processes of flux emergence etc.) 

145: would not permit a quasi--equilibrium approach. A little farther distance 

146: ($\Delta r$) from the surface, however, we expect that the quasi--equilibrium 

147: two--fluid model will capture the essential physics of flow generation. 

148: 

149: >From recent observational data (see e.g. Goodman (2000); Aschwanden 

150: \etal (2001a); Socas-Navarro and Almeida (2002) and 

151: references therein) we could obtain the following average plasma density 

152: and temperature at $\Delta r\sim (500-2000)\,$km: $n\sim 

153: (10^{15}-10^{11})\,{\rm cm}^{-3}; \ T\sim 1\,$eV (for simplicity we will assume equal 

154: electron and ion temperatures). The information about the magnetic field 

155: is a little harder to extract because of the low sensitivity and lack of 

156: high spatial resolution of the measurements coupled with the inhomogeneity 

157: and co--existence of small-- and large--scale structures with 

158: different temperatures in  nearby regions. At these distances we have 

159: different values for the network and for the internetwork fields. The network 

160: plasmas have typically short--scale fields in the range $B_0\sim (700-1500)\,$G, 

161: have more or less uniform density and will be prone to explosive/eruptive 

162: analysis of the kind carried out in Ohsaki \etal (2002). 

163: The internetwork fields, on the other hand, are generally smaller (with 

164: some exceptions (Socas-Navarro and Almeida (2002)) --- $B_o\sim 500\,$G, and are embedded in 

165: larger--scale plasma structures with inhomogeneous densities. The theory 

166: of steady creation of flows in the lower chromosphere  will be based on 

167: these latter objects.

168: 

169: \section{Generation of flows in lower chromosphere}

170: \label{sec:flows}

171: 

172: To illustrate the basic physics of flow creation, we deal with the 

173: simplest two--fluid equilibria with $T=\rm const$ leading to \ 

174: $n^{-1}\nabla p \to \,T\nabla \,\ln\,n$. The generalization to a homentropic 

175: fluid with  $p=\rm const\cdot n^{\gamma}$ is straightforward and is done in the numerical work.

176: 

177: The dimensionless equations describing the model equilibrium can be read 

178: off from Mahajan \etal (2001)

179: \begin{equation}

180: \frac{1}{n}\nabla\times {\bf b\times b}+\nabla\left(\frac{r_{A0}}{r}

181: -\beta_0\ln\ n-\frac{V^2}{2}\right)+ {\bf V\times (\nabla\times

182: V})=0, \label{eq:DB-1}

183: \end{equation}

184: \begin{equation}

185: \nabla\times\left({\bf V}-\frac{\alpha_0}{n}\nabla \times {\bf b}

186: \right)\times {\bf b}=0, \label{eq:DB-2}

187: \end{equation}

188: \begin{equation}

189: \nabla \cdot (n{\bf V})=0, \label{eq:cont}

190: \end{equation}

191: \begin{equation}

192: \nabla\cdot{\bf b}=0, \label{eq:b}

193: \end{equation}

194: where the notation is standard with the following normalizations: the 

195: density $n$ to $n_0$ prevalent at about $500\,$km  (and farther) from the 

196: solar surface (this is where the plasma is created), the magnetic field to 

197: the ambient field strength at the same distance, and velocities to the 

198: Alfv\'en velocity $V_{A0}$. The parameters $r_{A0}=GM_\odot/V_{A0}^2R_{\odot}=2

199: \beta_0/r_{c0}, \ 

200: \alpha_0=\lambda_{i0}/R_{\odot}, \ \beta_0=c_{s0}^2/V_{A0}^2$ are defined 

201: with $n_0, \ T_0, \ B_0$. Here $c_{s0}$ is a sound speed, $R_\odot$ is the 

202: solar radius, $r_c=GM_\odot/2c_{s0}^2R_{\odot}$, and 

203: $\lambda_{i0}=c/\omega_{i0}$ is the collisionless skin depth.  

204: 

205: The double--Beltrami solutions, expressing the simple physics that the 

206: electrons follow the field lines, while the ions, due to their inertia, 

207: follow the field lines modified by the fluid vorticity, are contained in 

208: the pair,  

209: \begin{equation}

210: {\bf b}+\alpha_0 \nabla\times {\bf V}=d\ n\ {\bf V}, \qquad \qquad

211: {\bf b}=a\ n\ \left[{\bf V}-\frac{\alpha_0}{n}\,\nabla\times {\bf

212: b}\right], \label{eq:DB-3} 

213: \end{equation}

214: where $a$ and $d$ are dimensionless constants related to the two ideal

215: invariants, the magnetic and the generalized helicities (Mahajan and 

216: Yoshida 1998; Mahajan \etal 2001),

217: \begin{equation}

218: h_1=\int ({\bf A}\cdot {\bf b})\ d^3x , \label{eq:h1}

219: \end{equation}

220: \begin{equation}

221: h_2=\int ({\bf A}+{\bf V})\cdot ({\bf b}+\nabla\times {\bf V})d^3x. 

222: \label{eq:h2}

223: \end{equation}

224: On substituting (6)--(7) into (1)--(2), one obtains the Bernoulli condition

225: \begin{equation}

226: \nabla\left(\frac{2\beta_0r_{c0}}{r}-\beta_0\ln\,n-\frac{V^2}{2}\right)=0,

227: \label{eq:bernoulli}

228: \end{equation}

229: relating the density with the flow kinetic energy, and solar gravity. 

230: Equations~(\ref{eq:DB-1}), (\ref{eq:DB-3}), and (\ref{eq:bernoulli}) 

231: represent a close system, and may be easily manipulated to yield~($g(r)=r_{c0}/r$)

232: \begin{equation}

233: \frac{\alpha_0^2}{n}\nabla\times\nabla\times{\bf V}+\alpha_0\

234: \nabla\times \left[\left(\frac{1}{a\,n}-d\right)n\,{\bf 

235: V}\right]+\left(1-\frac{d}{a}\right){\bf V}=0, \label{eq:CC-1}

236: \end{equation}

237: \begin{equation}

238: \alpha_0^2\nabla\times\left(\frac{1}{n}\nabla\times{\bf 

239: b}\right)+\alpha_0\ \nabla \times \left[\left(\frac{1}{a\,n}-d\right){\bf 

240: b}\right]+\left(1-\frac{d}{a}\right){\bf b}=0,\label{eq:CC-2}

241: \end{equation}

242: \begin{equation}

243: n=\exp\left(-\left[2g_0-\frac{V^2_0}{2\beta_0}-2g+\frac{V^2}{2\beta_0}

244: \right]\right), \label{eq:density}

245: \end{equation}

246: a set ready to be solved for the density, the velocity and the magnetic 

247: field. We must point out that this time--independent set is not suitable 

248: for studying chromospheric heating processes (primary heating at lower 

249: heights). The main thrust of this paper is to uncover mechanisms which 

250: create flows in the  chromosphere --- flows that will supply matter 

251: and energy  needed to create the coronal structures, and provide their 

252: primary heating. The creation and heating problem, of course, requires a 

253: fully time--dependent treatment (Mahajan \etal 2001).

254: 

255: We have to resort to numerical methods to obtain detailed solutions for 

256: the coupled nonlinear system~(\ref{eq:CC-1}), (\ref{eq:CC-2}), and 

257: (\ref{eq:density}). We have carried out a 1D simulation (the relevant 

258: dimension being the height ``Z" from the center of the Sun; $Z_0=R_{\odot}+\Delta r$

259: is the surface at which the boundary conditions are applied) for a variety of 

260: boundary conditions. The boundary surface is so chosen that at this height $Z_0$ 

261: the ionization is expected to be complete. For estimates of $\Delta r$ given 

262: earlier, the relevant heights lie within 

263: $[(1+0.7\cdot 10^{-3})-(1+2.8\cdot 10^{-3})]\,R_{\odot}$.

264: 

265: The simulation results are presented in Figs.~1--2. These are the plots of 

266: various physical quantities as  functions of the height. The first figure 

267: consists of three frames (a--b, c--d, and e--f) each consisting of two 

268: pictures --- one for the density and the magnetic field and the other for 

269: the velocity field. The parameters defining  different frames are: 

270: $n_0\sim 10^{14}{\rm cm}^{-3}$, \ $B_0\sim 500\,$G, \  $V_{A0}\sim 110\,$km/s implying  

271: $\beta_0 \sim 0.01\ll 1$ and $r_{c0}=900$ for a--b; $n_0\sim 10^{16}{\rm cm}^{-3}$, 

272: \ $B_0\sim 1000\,$G, \ $V_{A0}\sim 22\,$km/s implying $\beta_0\sim 0.2< 1$ and 

273: $r_{c0}=950$ for c--d; and $n_0\sim 10^{17}{\rm cm}^{-3}$ and $B_0\sim 

274: 1500\,$G, \ $V_{A0}\sim 11\,$km/s, \ $\beta_0 \sim 1$ and $r_{c0}=1000$ for the frame e--f. 

275: In each frame there are three sets of curves labelled by $\alpha_0$ (1--2--3 

276: corresponding respectively to $\alpha_0 =10^{-5},\, 10^{-3},\, 10^{-1}$), the 

277: measure of the strength of the two--fluid Hall currents.

278: 

279: For all our runs the temperature at the boundary was taken to be $\sim 

280: 1\,$eV ($c_{s0}\sim 10\,$km/s), and the boundary conditions, $|b_0|=1, 

281: \quad V_0=1\,$km/s (with $V_{x0}=V_{y0}=V_{z0}$) were imposed. Notice that we 

282: begin with just a small residual flow speed. The choice, \ $d\sim a \sim 

283: 100$ \  and \ $(a-d)/a^2\sim 10^{-6}$ \  for the parameters characterizing the 

284: double Beltrami state, reflects the physical constraint that we are 

285: dealing with a sub--Alfv\'enic flow with a very small $\alpha_0$ (Mahajan 

286: \etal 1999). We must admit that the values of $\alpha_0$ chosen for the 

287: simulation are much  larger than their actual values ($\sim 10^{-8}$ for 

288: corona and $\sim 10^{-11}$ for chromosphere); our present code cannot resolve 

289: the equivalent short lengths, though, we hope to do better in future. We believe, 

290: however, that the nature of the final results is properly captured by these 

291: artificial values of $\alpha_0$.

292: 

293: The most remarkable result of the simulation is that for small and 

294: realistic values of $\alpha_0$ (curves labelled 1), there exists some 

295: height where the density begins to drop precipitously  with a 

296: corresponding sharp rise in the flow speed. The effect is even stronger 

297: for the low beta (a--b are the lowest beta frames) plasmas. It is also 

298: obvious that at very short distances, the stratification is practically 

299: due to gravity, but as we approach the velocity ``blow--up'' height, the 

300: self--consistent magneto--Bernoulli processes take over and control the 

301: density (and hence the velocity) stratification.

302: 

303: An examination of the Bernoulli condition~(\ref{eq:density}) readily 

304: yields an indirect estimate for the height at which the observed 

305: shock formation may take place. For a low--beta plasma, the sharp fall in 

306: density is expected  to occur where the local flow kinetic energy exceeds 

307: the kinetic energy specified at the  boundary (this is true for all  $\alpha_0$), i.e, 

308: \begin{equation}

309: |{\bf V}|^2- V_0^2 > 2\,\beta_0. \label{eq:criter}

310: \end{equation}

311: For the current simulation, at  $\beta_0 =0.01$, it occurs approximately 

312: at $|{\bf V}|^2 > 0.02$ or at $|V|\sim 0.14$. This analytically--predicted 

313: value is very close to the simulation result (see Fig.~2(b)). Simulation 

314: results also confirm that the velocity blow--up distance depends mainly on 

315: $\beta_0$, and that the final velocity is greater for greater $\beta_0$ 

316: (Fig.~2(a)). The data presented in Fig.~1  and Fig.~2 corresponds to a uniform 

317: temperature plasma. For this case, the variations in plasma pressure are 

318: entirely due to the variations in density. Since the magnetic energy remains

319: practically uniform over the distance, sharp decrease in density with a 

320: corresponding sharp rise in the flow--speed (flow energy changes are of the 

321: order of $n^{-1/2}$) is nothing but the expression of the commonly understood 

322: Bernoulli effect. We must emphasize that the general results remain unchanged  

323: in our extensive simulations in which the temperature is allowed to vary 

324: (but we have to use a homentropic equation of state to analytically derive 

325: the Beltrami states. The final parameters, naturally, depend upon the 

326: adiabaticity index $\gamma $).  

327: 

328: Taking into account the fact that the fiducial height $Z_0$ is different 

329: for different cases (larger for smaller $\beta_0$), one expects  that the 

330: plots for the ``blow--up'' distance will approximately match one other, 

331: and there exists some minimum distance (possibly of the order $500\,$km or so) 

332: from the photosphere below which steady up--flows can not be generated. This 

333: expectation is simply supported by the numerical results given in  

334: Fig.~2(a). Observations clearly indicate that there exists a narrow layer 

335: above the photosphere where no spicules, mottles etc. are seen.

336: 

337: To check whether the generated flows are predominantly radial or somewhat 

338: more isotropic (to explain the observational constraints) we studied in 

339: detail the relatively large $\beta_0$ case (fixing $\beta_0$ is quite 

340: difficult due to complications like ionization) and found that the flows 

341: tend to be mostly radial only for large  $\alpha_0$ (see, for example, 

342: plots labeled 2 and 3 in Fig.~1(b,d,f)). The situation could change 

343: considerably when we deal with a more inclusive time--dependent dynamical 

344: model with dissipation. Plasma heating, then, could result from the 

345: dissipation of the perpendicular energy so that at larger distances, the 

346: flows would have larger radial components. Heating would also keep 

347: $\beta({\bf r},t)$ large at upper heights shifting the velocity blow--up 

348: distance further or eliminating it all together; we know from Fig.~1 that 

349: as $\beta_0$ goes up, the density fall(velocity amplification) becomes  

350: smoother. These issues will be dealt with later in a more detailed work.

351: 

352: Note, that if one were to ignore the flow term in (\ref{eq:bernoulli}) (a 

353: totally wrong assumption commonly used in many studies), we will end up 

354: finding essentially radial flows. The magnitude of these flows, however, 

355: remains small; there is no region of sharp rise (\ref{eq:criter}), and 

356: the  generated flows achieve reasonable energies at heights typically 10 

357: times greater than the heights at which the correct Bernoulli condition 

358: would do the trick. 

359: 

360: \section{Conclusions and Summary}

361: \label{sec:conclusions} 

362: 

363: We have shown a  possible pathway  for a steady generation of flows in the 

364: quasi--equilibrium stage established through the interaction of the emerging

365: magnetic fluxes with the existing cold solar chromosphere (when an ionized   

366: $\sim 1\,$eV plasma is trapped in internetwork structures). The suggested 

367: mechanism is a straightforward application of the recently--developed  

368: magnetofluid model (Mahajan and Yoshida 1998; Mahajan \etal 1999, 

369: 2001); a generalized Bernoulli mechanism

370: (a necessary condition for the double--Beltrami magnetofluid equilibrium) 

371: allows the pressure energy to be very effectively transformed to flow kinetic energy 

372: as the plasma moves away from the sun. We find that at reasonable heights 

373: and for realistic plasma parameters, there is a precipitous pressure fall 

374: with a sharp amplification of the flow speed. In the presence of 

375: dissipation, these flows are likely to play a fundamental role in the 

376: heating of the inner and upper chromosphere, although our explicit purpose 

377: in this paper was to create  a steady source of matter and energy for the 

378: formation and primary heating of the corona. Our preliminary results agree 

379: with the observation data, and lend promise to attempts, based on the 

380: existence of subcoronal flows, to tackle unresolved problems like the 

381: coronal heating and origin of the solar wind. 

382:  

383: \section*{ACKNOWLEDGEMENTS}

384: 

385: %\acknowledgments

386: 

387: Authors acknowledge the help of Dr. K.I. Sigua for preparing the 

388: simulation  pictures and thank Dr. E. Marsch for interesting and useful discussions.

389: All the authors thank Abdus Salam International Centre for Theoretical Physics, 

390: Trieste, Italy. The work of SMM was supported by USDOE Contract No.~DE--FG03--96ER--54366. 

391: Work of KIN was supported by a Grant No.~02--02--16199a 

392: of Russia Fund of Fundamental Research 

393: (RFFR).  The work of NLS was supported by ISTC Project G--633. 

394: 

395: \clearpage

396: 

397: \begin{thebibliography}{99}

398:     

399:  \bibitem[Aschwanden {\it et al.} 2001a]{A1} Aschwanden, M.J., Poland 

400: A.I., and Rabin D.M. 2001a, Ann. Rev. Astron. Astrophys., 39, 175.

401: 

402: \bibitem[Aschwanden 2001b]{A2} Aschwanden, M.J. 2001b, \apj, 560, 

403: 1035. 

404: 

405: \bibitem[Christopoilou, Georgakilas, and Koutchmy 2001]{koutchmy} 

406: Christopoulou, E.B., Georgakilas, A.A., and Koutchmy, S. 2001, Solar 

407: Phys., 199, 61.

408: 

409: \bibitem[Goodman 2001]{goodman} Goodman, M.L. 2001, Space Sci. Rev., 

410: 95, 79.

411: 

412: \bibitem[Goodman 2000]{goodman2} Goodman, M.L. 2000, \apj, 533, 501.

413: 

414: \bibitem[Mahajan {\it et al.} 1999]{MMNS1} Mahajan, S.M., Miklaszewski, 

415: R., Nikol'skaya, K.I., and Shatashvili N.L., February 1999, Preprint 

416: IFSR \#857, {\it The University of Texas, Austin}, 67 pages. 

417: 

418: \bibitem[Mahajan {\it et al.} 2001]{MMNS2}  Mahajan, S.M., Miklaszewski, 

419: R., Nikol'skaya, K.I., and Shatashvili, N.L. 2001, Phys. Plasmas, 8, 

420: 1340. 

421: 

422: \bibitem[Mahajan and Yoshida 1998]{MY-1} Mahajan, S.M., and Yoshida, 

423: Z. 1998, Phys. Rev. Lett., 81, 4863. 

424: 

425: \bibitem[Nikol'skaya and Valchuk 1998]{nv} Nikol'skaya, K.I., and 

426: Valchuk, T.E. 1998, Geomagnetizm and Aeronomy, 38, No.~2, 14.

427: 

428: \bibitem[Ohsaki {\it et al.} 2001]{osym1} Ohsaki, S., Shatashvili, N.L., 

429: Yoshida, Z., and Mahajan, S.M. 2001, \apj, 559, L61.

430: 

431: \bibitem[Ohsaki {\it et al.} 2002]{osym2} Ohsaki, S., Shatashvili, N.L., 

432: Yoshida, Z., and Mahajan, S.M. 2002, \apj, 570.

433: 

434: \bibitem[Orlando, Peres and Serio 1995a]{flows1} Orlando, S., Peres, 

435: G., and Serio, S. 1995a, Astrophys. and Astron., 294, 861.

436: 

437: \bibitem[Orlando, Peres and Serio 1995b]{flows2} Orlando, S., Peres, 

438: G., and Serio, S. 1995b, Astrophys. and Astron., 300, 549.

439: 

440: \bibitem[Sakai and Furusawa 2002]{sakai1} Sakai, J.I., and Furusawa, 

441: K. 2002, \apj, 564, 1048.

442: 

443: \bibitem[Schrijver {\it et al.} 1999]{schrijver}  Schrijver, C.J., 

444: Title, A.M., Berger, T.E., Fletcher, L., Hurlburt, N.E., Nightingale, 

445: R.W., Shine, R.A., Tarbell, T.D., Wolfson, J., Golub, L., Bookbinder, 

446: J.A., Deluca, E.E., McMullen, R.A., Warren, H.P., Kankelborg, C.C., 

447: Handy B.N., and DePontieu, B. 1999, Solar Phys., 187, 261. 

448: 

449: \bibitem[Seaton {\it et al.} 2001]{ami1} Seaton, D.B., Winebarger, A.R., 

450: DeLuca, E.E., Golub, L., and Reeves, K.K. 2001, \apj, 563, L173.

451: 

452: \bibitem[Socas--Navarro and Sanchez Almeida 2002]{sa} Socas--Navarro, 

453: H., and Sanchez Almeida, J. 2002, \apj, 565, 1323.

454: 

455: \bibitem[Uchida {\it et al.} 2001]{uchida} Uchida Y., Miyagoshi, T., Yabiku 

456: T., Cable S., and Hirose S. 2001, Publ. Astron. Soc. Japan, 53, 331. 

457: 

458: \bibitem[Wilhelm 2001]{wilhelm} Wilhelm, K. 2001, Astrophys. and 

459: Astronomy, 360, 351.

460: 

461: \bibitem[Winebarger, DeLuca and Golub 2001]{golub} Winebarger, A.M., 

462: DeLuca, E.E., and Golub, L. 2001, \apj, 553, L81.

463: 

464: \bibitem[Winebarger {\it et al.} 2002]{ami2} Winebarger, A.R., Warren, 

465: H., Van Ballagooijen, A., DeLuca E.E., and Golub, L. 2002, \apj, 567, 

466: L89.

467: 

468: \end{thebibliography}

469: 

470: \clearpage

471:  

472: 

473: \figcaption[f1.eps]{Plots of density, magnetic fields and velocity 

474: versus height for values of $\alpha_0$ and $\beta_0$. Sub--figures (a) and 

475: (b) are for $\beta_0=0.01, \ r_{c0}=900$; (c) and (d) are for 

476: $\beta_0=0.2, \ r_{c0}=950$; (e) and (f) are for $\beta_0 = 1, \ 

477: r_{c0}=1000$. The numbers 1, 2, 3 represent  $\alpha_0 =10^{-5},\, 

478: 10^{-3}\, 10^{-1}$ respectively. $V_y$ is not displayed since its 

479: behavior is practically similar to  $V_x$. The velocity blow--up is 

480: controlled by $\beta_0$. For a bigger (unrealistic)  $\alpha_0$ there is a 

481: splitting of the velocity components --- at the end the radial component 

482: is dominant. Magnetic field energy does not change much on these distances.}

483: 

484: \figcaption[f2.eps]{The ``blow--up" distance (a) and velocity (b)

485: versus $\alpha_0$. The smaller the $\beta_0$, the smaller the   

486: ``blow--up" distance and smaller the velocity at ``blow--up" 

487: (compare with (\ref{eq:criter})). For fixed $\beta_0$, the process is less 

488: sensitive to changes in $\alpha_0$.}

489: 

490: 

491: \end{document}

492: 

493: 

494: 

495: 

496: 

497: %%%%%%%%%%%%%%%%%%%% Fig. 1

498: \begin{figure}

499:  \plotone{f1.eps}

500:      \caption{Plots of density, magnetic fields and velocity 

501: versus height for values of $\alpha_0$ and $\beta_0$. Sub--figures (a) and 

502: (b) are for $\beta_0=0.01, \ r_{c0}=900$; (c) and (d) are for 

503: $\beta_0=0.2, \ r_{c0}=950$; (e) and (f) are for $\beta_0 = 1, \ 

504: r_{c0}=1000$. The numbers 1, 2, 3 represent  $\alpha_0 =10^{-5},\, 

505: 10^{-3}\, 10^{-1}$ respectively. $V_y$ is not displayed since its 

506: behavior is practically similar to  $V_x$. The velocity blow--up is 

507: controlled by $\beta_0$. For a bigger (unrealistic)  $\alpha_0$ there is a 

508: splitting of the velocity components --- at the end the radial component 

509: is dominant. Magnetic field energy does not change much on these distances.

510: \label{fig1}}

511: \end{figure}

512: 

513: \clearpage

514: 

515: %%%%%%%%%%%%%%%%%%%% Fig. 2

516: \begin{figure}

517:     \plotone{f2.eps}

518:     \caption{The ``blow--up" distance (a) and velocity (b)

519: versus $\alpha_0$. The smaller the $\beta_0$, the smaller the   

520: ``blow--up" distance and smaller the velocity at ``blow--up" 

521: (compare with (\ref{eq:criter})). For fixed $\beta_0$, the process is less 

522: sensitive to changes in $\alpha_0$.

523: \label{fig2}}

524: \end{figure}

525: