1: %\documentclass{emulateapj}

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

3: %\usepackage{graphics}

4: 

5: \newcommand\bb[1] {   \mbox{\boldmath{$#1$}}  }

6: \newcommand\del{\bb{\nabla}}

7: \newcommand\bcdot{\bb{\cdot}}

8: \newcommand\btimes{\bb{\times}}

9: \newcommand\vv{\bb{v}}

10: \newcommand\B{\bb{B}}

11: \newcommand\BV{Brunt-V\"{a}is\"{a}l\"{a} }

12: \newcommand\iw{ i \omega }

13: \newcommand\kva{ \bb{k\cdot v_A}  }

14: \newcommand\kb{ \bb{k\cdot b}  }

15: \newcommand\kkz { \left( \frac{k}{k_Z}\right)^2\>}

16: \newcommand\dd{\partial}

17: \newcommand\etal{{\it et al.}}  

18: \newcommand\eg{{\it e.g.}}

19: \newcommand\etc{{\it etc.}}

20: \newcommand\ie{{\it i.e.}}

21: \newcommand\beq{ \begin{equation} }

22: \newcommand\eeq{ \end{equation} }

23: \newcommand\wtilde{\widetilde}

24: \newcommand{\schwz}{ {\dd  \ln P\rho ^{-\gamma} \over \dd Z}}

25: \newcommand{\schwR} { {\dd  \ln P\rho ^{-\gamma} \over \dd R} }

26: \newcommand{\balbz}{ {\dd  \ln T \over \dd Z}}

27: \newcommand{\balbR} { {\dd  \ln T \over \dd R} }

28: \newcommand\rsun { \,\mathrm{R_{\sun}} }

29: 

30: \begin{document}

31: 

32: \title{The Origin of Solar Activity in the Tachocline}

33: \author{Kyle P. Parfrey and Kristen Menou}

34: \affil{Department of Astronomy, Columbia University,}

35: \affil{550 West 120th Street, New York, NY 10027, U.S.A.}

36: 

37: 

38: \begin{abstract}

39: Solar active regions, produced by the emergence of tubes of strong

40: magnetic field in the photosphere, are restricted to within $35\degr$

41: of the solar equator. The nature of the dynamo processes that create

42: and renew these fields, and are therefore responsible for solar

43: magnetic phenomena, are not well understood. We analyze the

44: magneto-rotational stability of the solar tachocline for general field

45: geometry. This thin region of strong radial and latitudinal

46: differential rotation, between the radiative and convective zones, is

47: unstable at latitudes above $37\degr$, yet is stable closer to the

48: equator. We propose that small-scale magneto-rotational turbulence

49: prevents coherent magnetic dynamo action in the tachocline except in

50: the vicinity of the equator, thus explaining the latitudinal

51: restriction of active regions. Tying the magnetic dynamo to the

52: tachocline elucidates the physical conditions and processes relevant

53: to solar magnetism.

54: 

55: \end{abstract}

56: 

57: \keywords{hydrodynamics --- instabilities --- MHD --- Sun: activity

58:   --- Sun: magnetic fields --- Sun: rotation}

59: 

60: 

61: %\maketitle

62: 

63: 

64: \section{INTRODUCTION}

65: Magnetic fields are a leading actor throughout a star's life---from

66: accreting protostar to degenerate dwarf, main sequence star to

67: magnetar (Mestel 1999). They shape jets in some supernovae and drive a

68: pulsar's radiative engine. Yet we still lack a coherent theory of how

69: even our own star generates and sustains its field, and so are

70: ignorant of the wellspring of much of this evolutionary richness

71: (Parker 1979). Since the 1950s, much effort has been expended on

72: models which locate the dynamo

73: in the convective zone. There are several known difficulties with this

74: scenario (\eg, Parker 1975; Petrovay 2000). Recently, attention has

75: turned to the lower convective zone and the tachocline, its boundary

76: with the stably stratified radiative zone, in part because

77: stratification can anchor the strong emerging fields (Parker 1975;

78: Petrovay 2000; Dikpati \& Gilman 2005; Browning et al. 2006; Hughes,

79: Rosner, \& Weiss 2007). Dynamical considerations suggest the presence

80: of a large-scale magnetic field in the upper radiative zone (Gough \&

81: McIntyre 1998), of plausible strength $\sim 1\; \mathrm{G}$. 

82: 

83: 

84: Here, we propose that the solar magnetic dynamo is intimately tied to the

85: tachocline, where latitudinal differential rotation amplifies the

86: toroidal field and, together with radial differential rotation, controls the

87: limiting role of the magneto-rotational instability (MRI). This

88: instability creates small-scale turbulence which tends to prevent

89: large-scale, ordered field growth, that leads to surface emergence and

90: sunspots.

91: 

92: The hydrodynamic (\eg, Charbonneau, Dikpati \& Gilman 1999; Dikpati \&

93: Gilman 2001; Arlt, Sule \& R\"{u}diger 2005) and magneto-hydrodynamic

94: (\eg, Gilman \& Fox 1997, Miesch, Gilman \& Dikpati 2006; Miesch 2007)

95: stability of the tachocline have been the subject of extensive

96: study. Here, we consider stability with respect to the diffusive MRI,

97: which has not been previously examined. Naively, one would expect that

98: since $\dd \Omega/\dd\theta > 0$ the tachocline would be stable to the

99: MRI, in the presence of the strong radial entropy stratification that

100: suppresses fluid motions in the radial direction. However, we show

101: that the inclusion of thermal diffusion enables the radial movement of

102: fluid on small scales, allowing the negative radial gradient of

103: angular velocity to be tapped, which results in instability at large

104: heliolatitudes.

105: 

106: 

107: 

108: \section{STABILITY ANALYSIS}

109: 

110: The tachocline is a boundary layer imposed on the solid-body-rotating

111: radiative zone by the convective zone, whose global angular momentum

112: is extracted by the magnetic solar wind (\eg, Spiegel \& Zahn 1992;

113: Hughes et al. 2007). Differential rotation in the tachocline, created

114: by external convective torques, is not necessarily reduced by

115: instabilities operating within it. The

116: rotational gradients result from the different means of energy and

117: momentum transport in the radiative and convective zones, so that a

118: tachoclinic dynamo could in principle feed on the central nuclear

119: energy source. Using helioseismic constraints (Thompson et al. 2003),

120: we consider a thin shell confined to the tachocline with differential

121: rotation given by an angular velocity profile

122: \beq

123: \Omega(r,\theta) = \Omega_{rad} + \frac{r - r_t + w}{2w}(1 -

124: \alpha_2\cos^2\theta - \alpha_4\cos^4\theta)\,\delta\Omega_{eq},

125: \label{omprof}

126: \eeq

127: where $\Omega_{rad}=2.69\times 10^{-6}\;\mathrm{rad\; s^{-1}}$ is

128: the angular velocity of the radiative zone, $\delta\Omega_{eq} =

129: 1.08\times 10^{-7}\;\mathrm{rad\; s^{-1}}$ is the differential in

130: angular velocity across the tachocline at the equator, $r_t = 0.7

131: \rsun$ is the

132: radius of the tachocline's mid-point, and $2w = 0.02 \rsun$ is its

133: width (Basu \& Antia 2001). We use both spherical co-ordinates

134: $\left(r,\phi,\theta\right)$, where $\theta$ is the co-latitudinal

135: angle measured from the pole, and cylindrical co-ordinates

136: $\left(R,\phi,Z\right)$ in the following analysis. This model implies

137: $q\equiv \dd\ln \Omega/\dd\ln r = -10.2$ near the pole, $q = +1.4$ at

138: the equator and $q=0$ at $\theta = 62\degr$. Calculations are

139: performed at $r = r_t$. The constants $\alpha_2 = 3.56$ and $\alpha_4

140: = 4.21$ are derived from GONG data (Basu \& Antia 2001).

141: 

142: We do not have direct knowledge of the radial structure of the entropy

143: stratification in the tachocline, because it is not resolved by

144: helioseismology. Using density and pressure profiles from a standard

145: solar model (Bahcall, Serenelli, \& Basu 2005), we calculate an

146: approximate stratification profile for the tachocline, and vary the

147: stratification strength in our stability analysis within the range of

148: plausible values. We find that our results are not strongly dependent

149: on the assumed level of stratification.

150: 

151: 

152: The MRI is a local, linear, weak-field instability present in systems

153: with radially decreasing angular velocity (Balbus \& Hawley 1991,

154: 1998). The most general form (Acheson 1978; Menou, Balbus \& Spruit

155: 2004) includes the effects of entropy stratification, viscosity $\nu$,

156: and thermal and magnetic diffusivities ($\xi$ and $\eta$

157: respectively). Thermal diffusion allows a small plasma element to

158: radiate photons and entropically equilibrate with its surroundings

159: which reduces the effective stratification. Stability is determined by

160: the rotational gradients (including their signs) and not by the

161: presence of (direction-independent) shear, making the MRI

162: qualitatively different from shear instabilities.

163: 

164: Necessary and sufficient linear stability criteria for the adiabatic

165: MRI ($\nu = \xi = \eta =0$) can be derived via a generalized

166: axisymmetric Solberg-H\o iland analysis (Balbus 1995):

167: \begin{eqnarray}

168: N_r^2 + \left(r\sin^2\theta\frac{\partial}{\partial r} + \sin\theta

169: \cos\theta \frac{\partial}{\partial \theta}\right)\Omega^2 > 0 ,

170: \label{balb1}\\

171: N_r^2\cos\theta \frac{\partial\Omega^2}{\partial \theta} > 0,\label{balb2}

172: \end{eqnarray}

173: where $N_r$ is the radial component of the \BV (buoyancy)

174: frequency. In the limit of strong thermal diffusion the plasma is able

175: to come to entropic equilibrium with its surroundings arbitrarily

176: quickly, so that the stratification is effectively removed. Therefore

177: by setting $N_r$ to a vanishingly small constant we can use the

178: adiabatic criteria to 

179: determine magneto-rotational stability in the idealized case of

180: infinite thermal diffusivity. Calculated using our rotation profile

181: and $N_r \rightarrow 0$, Eqs. (2) \&

182: (3) indicate that the tachocline can be magneto-rotationally

183: unstable for all $\theta < 62\degr$ (Fig. 1). We vary $r_t$ and $w$

184: independently and find that the extent of the unstable region is

185: insensitive to details of the radial structure. The linear radial

186: profile of angular velocity assumed in our tachocline model is that

187: which is least magneto-rotationally unstable, because any non-linear

188: profile with the same angular velocities at the boundaries must have a

189: greater radial angular velocity gradient at some point. For example,

190: we find the same region of instability for an exponential profile

191: (Basu \& Antia 2001).

192: 

193: These stability criteria (with $N_r^2 \rightarrow 0$) determine the

194: maximum extent of the instability, because they are inclusive of every

195: relevant local length scale, field strength, and geometry, subject to

196: several assumptions (see Balbus 1995). In particular the magnetic

197: field must be ``weak'', \ie  having an Alfv\'{e}n

198: velocity less than both the rotational velocity and the sound

199: speed. To investigate whether instability exists for physically

200: reasonable scales and fields, and to confirm that it is not stabilized

201: by a realistic entropy stratification, we numerically solve the

202: triple-diffusive dispersion relation (Menou et al. 2004) for

203: axisymmetric modes of frequency $\omega$:

204: %\begin{eqnarray}

205: \beq

206: {\wtilde\omega_{b+v}}^4 {\omega_{e}}\, {k^2\over k_Z^2} + {\wtilde\omega_{b+v}}^2 {\omega_{b}}

207: \left[ {1\over \gamma \rho}\, \left({\cal D} P\right)\, {\cal D} \ln

208: P\rho^{-\gamma}\right] \nonumber\\

209: + {\wtilde\omega_{b}}^2 {\omega_{e}} \left[

210: {1\over R^3}\, {\cal D} (R^4\Omega^2) \right] - 4 \Omega^2 (\kva)^2 {\omega_{e}}=

211: 0,

212: \eeq

213: %\end{eqnarray}

214: where

215: \begin{eqnarray}

216: \bb{v_A} = { \bb{B}/\sqrt{4\pi\rho}}, \qquad

217: k^2 = k_R^2 + k_Z^2, {\wtilde\omega_{b+v}}^2=\omega_b \omega_v -(\kva)^2,

218: \qquad {\wtilde\omega_{b}}^2=\omega_b^2 -(\kva)^2, \nonumber

219: \end{eqnarray}

220: \begin{eqnarray}

221: \omega_b=\omega+i \eta k^2, \qquad  \omega_v=\omega+i \nu k^2,

222: \omega_e=\omega+ \frac{\gamma -1}{\gamma}{i  T\over  P} \chi k^2,

223: \qquad {\cal D} \equiv \left( \frac{k_R}{k_Z}\frac{\dd}{\dd Z}

224: -\frac{\dd}{\dd R}\right) . \nonumber

225: \end{eqnarray}

226: 

227: For each magnetic field configuration, we solve for two million

228: wavevectors distributed in $k$ phase space (guaranteeing convergence

229: with respect to phase space resolution) and choose the fastest growing

230: mode which, under exponential growth, should quickly

231: dominate. Following an established procedure (Menou \& Le Mer 2006), we ensure

232: that the modes comfortably fit within the tachocline, allowing their

233: vector components to be positive or negative. The density and

234: diffusivity values adopted are {standard:} $\rho =

235: 0.2\;\mathrm{g\;cm^{-3}}$, $\nu = 23.6\;\mathrm{cm^2\;s^{-1}}$, $\eta

236: = 596\;\mathrm{cm^2\;s^{-1}}$, $\xi = 1.2\times 10^7\;\mathrm{cm^2\;s^{-1}}$.

237: 

238: 

239: \section{RESULTS}

240: 

241: We find growth rates on the order of the rotation frequency for

242: $|\bb{B}|=1 \;\mathrm{G}$ and $N_r^2/\Omega_{rad}^2 = 5\times 10^4$

243: (see Fig. 1), appropriate values for the strongly stratified lower

244: tachocline (\eg, Gough \& McIntyre 1998). The diffusive instability

245: extends to $\theta \approx 53\degr$. For this value of $N_r$, radial

246: fields are relatively insensitive to field strength; at

247: $870\;\mathrm{G}$ the growth rate peaks at a value

248: $\sigma_{max}/\Omega_{rad} \sim 2\times 10^{-3}$ while stabilization

249: occurs at $880\;\mathrm{G}$ and above. With latitudinal fields, the

250: growth rates decrease more rapidly as the field strength is increased:

251: the peak value of $\sigma_{max}/\Omega_{rad} \sim 10^{-3}$ at

252: $50\;\mathrm{G}$ and $\sim 10^{-6}$ at $200\;\mathrm{G}$. Instability

253: persists for stratifications up to and including that present at the

254: top of the radiative zone, $N_r^2/\Omega_{rad}^2 \approx 1.6\times

255: 10^5$. Radial fields give higher growth rates than latitudinal ones,

256: with mixed geometries falling in between. As expected (Menou \& Le Mer

257: 2006), at very small field values (\ie, $\sim 10^{-2}\;\mathrm{G}$)

258: diffusion causes the MRI modes to have smaller growth rates than the

259: hydrodynamical modes. Our survey indicates that a large region of the

260: parameter space is magneto-rotationally unstable for $\theta \lesssim

261: 53\degr$, in the manner shown in Fig. 1, for field strengths from a

262: small fraction of a Gauss to several tens or hundreds of Gauss and for

263: a range of stratifications.

264: 

265: Diffusive instability does not extend as far in co-latitude as in the

266: maximally unstable {case discussed above;} the radial gradient

267: of angular velocity becomes less negative as the point of inflexion,

268: $\theta=62\degr$, is approached from the pole and its destabilizing

269: influence is more easily counteracted by the positive latitudinal

270: gradient ($\dd \Omega/\dd \theta > 0$) and stratification ($N_r^2 >

271: 0$). For definiteness, we only show values of

272: $\sigma_{max}/\Omega_{rad}$ larger than $10^{-3}$. As we now discuss,

273: very small values may allow linear winding of the field before it is

274: disrupted by significant MRI exponential growth, possibly introducing

275: other instabilities.

276: 

277: Surveying relevant processes, it appears that two other weak-field

278: instabilities could be active in the tachocline (\eg, Acheson 1978;

279: Spruit 1999). Both should be secondary to the exponentially-growing

280: MRI when it is present with a sufficiently large growth rate, however,

281: since they initially rely on linear amplification of a toroidal field

282: by azimuthal stretching of the poloidal

283: component. Tayler instability is of a pinch or interchange type, to

284: which a toroidal field is most unstable by $m=0$ or $m=1$

285: perturbations (Tayler 1973). Stability is determined solely by the values

286: of $\dd B_{\phi}/\dd \theta$ and $\theta$ (\eg, Spruit 1999). We test the

287: Tayler stability of toroidal fields created by azimuthal winding of

288: four distinct poloidal field configurations, having poloidal

289: components $\bb{B}_p$ strictly along $\bb{\hat{r}}$,

290: $\bb{\hat{\theta}}$, $\bb{\hat{R}}$ and $\bb{\hat{Z}}$, using (Spruit 1999) 

291: \beq 

292: B_{\phi} = N\, \bb{\hat{n}} \cdot \bb{B}_p , 

293: \eeq where $N$

294: is the number of windings and $\bb{\hat{n}} \equiv

295: \nabla\Omega/|\nabla\Omega|$ is calculated from Eq.~(1). We find that

296: the instability should be largely contained within $62\degr$ or less

297: of the pole, although the $62\degr \lesssim \theta \lesssim 68\degr$

298: region could be unstable for some field geometries (Fig. 1). This

299: indicates a limited relevance of the Tayler process for the

300: tachocline, consistent with the general expectation that this process

301: should be active primarily in polar regions. For instance, if the

302: tachocline's rotational structure is related to magnetic stresses, the

303: weakness of $|\dd\Omega/\dd \theta|$ relative to $|\dd\Omega/\dd r|$

304: suggests that $B_{\theta} \gg B_r$, because the rates at which Maxwell

305: stresses transport angular momentum radially and latitudinally are

306: proportional to $B_r B_{\phi}$ and $B_{\theta} B_{\phi}$ respectively

307: (Spruit 2002). If so, Tayler instability would only be possible for

308: $\theta < 62\degr$, largely within the region where the exponentially

309: growing MRI dominates. Unlike the MRI, we find that the Tayler

310: instability depends strongly on the magnetic field geometry. In the

311: absence of direct constraints on $\bb{B}$ in the tachocline, we

312: conclude that there may be a role for the Tayler instability in the

313: transitional latitudes between MRI-dominance and active region

314: formation, $50\degr \lesssim \theta \lesssim 60\degr$.

315: 

316: A second weak-field instability occurs if $ p \equiv \dd \ln B/\dd \ln

317: r <0$, when gas is supported against gravity by magnetic

318: pressure. Given a sufficiently strong radial gradient, free energy can

319: be liberated by field-line buckling (\eg, Parker 1955; Acheson

320: 1978). As with other weak-field instabilities, stratification (reduced by thermal

321: diffusion) will be stabilizing. In the tachocline, at the equator, the

322: instability criterion is $|p|B^2 \gtrsim 10^{11} \;\mathrm{G^2}$

323: (Acheson 1978; Spruit 1999); assuming a non-singular $p \sim -1$

324: value, diffusive buoyancy instability would require $|\bb B | \gtrsim

325: 10^{5} \;\mathrm{G}$, which is also the field strength at which

326: large-scale adiabatic buoyant instability has been shown to cause flux

327: emergence resulting in active region formation (\eg, D'Silva \&

328: Choudhuri 1993; Sch\"{u}ssler et al. 1994). Therefore the small-scale

329: diffusive buoyant instability appears to be of little relevance.

330: 

331: 

332: To summarize, we infer that the tachocline is magneto-rotationally

333: unstable for $\theta \lesssim 53\degr$ and stable closer to the

334: equator. In the unstable region we expect the small-scale turbulence produced by the

335: diffusive MRI, which is a narrow salt-finger-type instability

336: (Korycansky 1991; Menou et al. 2004), to prevent the ordered growth of

337: magnetic field on large scales. We surmise that in the

338: magneto-rotationally  stable region, latitudinal differential rotation

339: is free to wind the poloidal field in the azimuthal direction,

340: creating a linearly amplified toroidal component as in the original

341: Babcock-Leighton dynamo (Babcock 1961; Leighton 1964). When the

342: toroidal field reaches a strength of order $10^{5} \;\mathrm{G}$ it is

343: subject to buoyant instability and emerges. The Babcock-Leighton model

344: is consistent with the dearth of active regions within $8\degr$ of the

345: equator, since $\dd \Omega/\dd \theta \rightarrow 0$ as the equator is

346: approached (giving no field winding), and can account for the phase

347: dependence of the flux-emergence latitude (Babcock 1961; Dikpati \&

348: Charbonneau 1999).

349: 

350: 

351: \section{DISCUSSION \& CONCLUSION}

352: 

353: Strong radial shear exists in both equatorial and polar regions of the

354: tachocline, yet active regions are confined to $\theta \gtrsim

355: 55\degr$. At approximately this co-latitude the negative radial

356: differential rotation is just strong enough to overcome stratification

357: and trigger the diffusive MRI, while closer to the equator the

358: rotational gradient is of the wrong sign. The $\theta = 55\degr$ point

359: does not appear to be a critical co-latitude for any other weak-field

360: instability, which is simple circumstantial evidence that the

361: diffusive MRI is a discriminant for the observed latitudinal

362: cutoff. We have confirmed this expectation with a rigorous stability

363: theorem and a detailed numerical stability analysis.

364: 

365: 

366: 

367: Our study is chiefly limited by its linear character, whereby it is

368: not possible fully to determine the effect of the saturated MRI on the

369: tachocline. The instability will likely generate an effective

370: anisotropic ``turbulent viscosity'', which could be important in explaining the

371: tachocline's radial thinness. Interestingly, observational constraints

372: on the tachocline's width imply a latitudinal thickness variation

373: consistent with two separate processes in action at high and low

374: latitudes, as in our scenario (Basu \& Antia 2001). 

375: 

376: 

377: The identification of the tachocline as a critical region for solar

378: dynamo action creates opportunities to further our understanding of

379: solar magnetism. A particularly interesting question concerns the

380: complex relationship that may exist between the tachoclinic generation

381: of magnetic fields, which emerge to form active regions that impact

382: the solar magnetic wind, and the effect that this same wind has in

383: modifying the differential rotation in the tachocline by applying

384: torque at the solar surface.

385: 

386: 

387: The authors thank H. Spruit for insightful discussions on solar

388: magnetism and the dynamo problem, and the anonymous referee for

389: comments that improved the manuscript.

390: 

391: 

392: \clearpage

393: 

394: \begin{thebibliography}{}

395: 

396: \bibitem[Acheson(1978)]{acheson} Acheson, D. J. 1978, Philos. Trans. R. Soc. London A, 289, 459-500

397: \bibitem[Arlt, Sule \& R\"{u}diger(2005)]{arlt2005} Arlt, R., Sule, A., \& R\"{u}diger, G. 2005, A\&A, 441, 1171-1175

398: \bibitem[Babcock(1961)]{babcock} Babcock, H. W. 1961, \apj, 133, 572-587

399: \bibitem[Bahcall, Senerelli, \& Basu(2005)]{bahcall} Bahcall, J. N., Serenelli, A. M., \& Basu, S. 2005, \apj, 621, L85-L88

400: \bibitem[Balbus(1995)]{balbus95} Balbus, S. A. 1995, \apj, 453, 380-383

401: \bibitem[Balbus \& Hawley(1991)]{balbus91} Balbus, S. A., \& Hawley, J. F. 1991, ApJ, 376, 214-222

402: \bibitem[Balbus \& Hawley(1998)]{balbus98} Balbus, S. A., \& Hawley, J. F. 1998, Rev. Mod. Phys., 70, 1-53

403: \bibitem[Basu \& Antia(2001)]{basu} Basu, S., \& Antia, H. M. 2001, MNRAS, 324, 498-508

404: \bibitem[Browning et al.(2006)]{browning} Browning, M. K., Miesch, M. S., Brun, A. S., \& Toomre, J. 2006, \apj, 648, L157-L160

405: \bibitem[Charbonneau, Dikpati \& Gilman(1999)]{charbonneau} Charbonneau, P., Dikpati, M., \& Gilman, P. A. 1999, \apj, 526, 523-537

406: \bibitem[Dikpati \& Charbonneau(1999)]{dikpati99} Dikpati, M., \& Charbonneau, P. 1999, \apj, 518, 508-520

407: \bibitem[Dikpati \& Gilman(2001)]{dikpati01} Dikpati, M., \& Gilman, P. A. 2001, \apj, 551, 536-564

408: \bibitem[Dikpati \& Gilman(2005)]{dikpati05} Dikpati, M., \& Gilman, P. A. 2005, ApJ, 635, L193-L196

409: \bibitem[D'Silva \& Choudhuri(1993)]{dsilva} D'Silva, S., \& Choudhuri, A. R. 1993, A\&A, 272, 621-633

410: \bibitem[Gilman \& Fox(1997)]{gilman} Gilman, P. A., \& Fox, P. A. 1997, \apj, 484, 439-454 

411: \bibitem[Gough \& McIntyre(1998)]{gough} Gough, D. O., \& McIntyre, M. E. 1998, Nature, 394, 755-757

412: \bibitem[Hughes, Rosner, \& Weiss(2007)]{hughes} Hughes, D., Rosner, R., \& Weiss, N. 2007, The Solar Tachocline (Cambridge: Cambridge Univ. Press)

413: \bibitem[Korycansky(1991)]{korycansky} Korycansky, D. G. 1991, \apj, 381, 515-525

414: \bibitem[Leighton(1964)]{leighton} Leighton, R. B. 1964, \apj, 140, 1547-1562

415: \bibitem[Menou, Balbus \& Spruit(2004)]{menou04} Menou, K., Balbus, S. A., \& Spruit, H. C. 2004, \apj, 607, 564-574

416: \bibitem[Menou \& Le Mer(2006)]{menou06} Menou, K., \& Le Mer, J. 2006, \apj, 650, 1208-1215

417: \bibitem[Mestel(1999)]{mestel} Mestel, L. 1999, Stellar Magnetism (Oxford: Clarendon Press)

418: \bibitem[Miesch(2007)]{miesch07} Miesch, M. S. 2007, \apj, 658, L131-L134

419: \bibitem[Miesch, Gilman, \& Dikpati(2006)]{miesch06} Miesch, M. S., Gilman, P. A., \& Dikpati, M. 2006, \apjs, 168, 337-361

420: \bibitem[Parker(1995)]{parker55} Parker, E. N. 1955, \apj, 121, 491-507

421: \bibitem[Parker(1975)]{parker75} Parker, E. N. 1975, ApJ, 198, 205-209

422: \bibitem[Parker(1979)]{parker79} Parker, E. N. 1979, Cosmical Magnetic Fields (New York: Oxford Univ. Press)

423: \bibitem[Petrovay(2000)]{petrovay} Petrovay, K. 2000, in The Solar Cycle and Terrestrial Climate (ESA Publ. SP), vol. 463, pp. 3-14 (http://www.arxiv.org/abs/astro-ph/0010096)

424: \bibitem[Sch\"{u}ssler et al.(1994)]{schussler} Sch\"{u}ssler, M., Caligari, P., Ferriz-Mas, A., \& Moreno-Insertis, F. 1994, A\&A, 281, L69-L72 

425: \bibitem[Spiegel \& Zahn(1992)]{spiegel} Spiegel, E. A., \& Zahn, J. -P. 1992, A\&A, 265, 106-114

426: \bibitem[Spruit(1999)]{spruit99} Spruit, H. C. 1999, A\&A, 349, 189-202

427: \bibitem[Spruit(2002)]{spruit02} Spruit, H. C. 2002, A\&A, 381, 923-932

428: \bibitem[Tayler(1973)]{tayler} Tayler, R. J. 1973, \mnras, 161, 365-380

429: \bibitem[Thompson et al.(2003)]{thompson} Thompson, M. J., Christensen-Dalsgaard, J., Miesch, M. S., \& Toomre, J. 2003, ARA\&A, 41, 599-643

430: \end{thebibliography}

431: 

432: \clearpage

433: \begin{figure}[h!]

434: \centering%

435: %\includegraphics[scale=0.85]{colorplot.eps}

436: \plotone{f1}

437: \caption{Growth rate of the fastest growing MRI mode in units of the

438: rotation rate for radial ($B_r$, black solid curve), latitudinal

439: ($B_{\theta}$, dashed), and equally radial and latitudinal fields

440: ($B_r + B_{\theta}$, dot dashed). Values $|\bb B

441: |=1 \;\mathrm{G}$ and

442: $N_r^2/\Omega_{rad}^2=5\times 10^4$ are used in all cases. The

443: hatching indicates the maximally unstable region, as determined by the

444: Balbus stability criteria. The arrows above the plot delimit the areas

445: potentially subject to Tayler instability, for radial and latitudinal

446: fields. Active regions are found to the right of the red vertical long-dashed

447: line.}

448: \label{fig1}

449: 

450: 

451: \end{figure}

452: 

453: 

454: 

455: 

456: 

457: 

458: \end{document}

459: 

460: 

461: 

462: 

463: 

464: 

465: 

466: