1: \documentclass[12pt]{iopart}

2: \usepackage{graphicx}

3: \usepackage{color}

4: %\usepackage{iopams}

5: 

6: %-Bibliography-------------------------------------------

7: %\usepackage{natbib}

8: %\bibliographystyle{unsrt}

9: %--------------------------------------------------------

10: 

11: %-My commands--------------------------------------------

12: \newcommand{\vomega}{\mbox{\boldmath $ \omega $}}

13: \newcommand{\vOmega}{\mbox{\boldmath $ \Omega $}}

14: \newcommand{\vmu}{\mbox{\boldmath $ \mu $}}

15: %--------------------------------------------------------

16: 

17: %-Comments---------------------------------------------------------

18: \def\NOTE#1{{\textcolor{red}{\bf [#1]}}}  % note

19: \def\ADD#1{{\textcolor{blue}{#1}}}        % question

20: \def\AD#1{{\textcolor{magenta}{#1}}}      % plug a value, a ref, ...

21: \def\DEL#1{{\textcolor{green}{ #1}}}      % suggested deletion in text

22: \def\BB#1{{\textcolor{blue}{\bf #1}}}  

23: %------------------------------------------------------------------

24: 

25: %-To correct the margin - remove before submision------------------

26: %\topmargin 0pt

27: %------------------------------------------------------------------

28: 

29: \begin{document}

30: 

31: \title[Hydrodynamics and magnetohydrodynamics inside a rotating sphere]

32:   {Hydrodynamic and magnetohydrodynamic computations inside a rotating sphere}

33: \author{P D Mininni$^1$, D C Montgomery$^2$, and L Turner$^3$}

34: \address{$^1$ National Center for Atmospheric Research, P.O. Box 3000, 

35:   Boulder, CO 80307, USA}

36: \address{$^2$ Dept. of Physics and Astronomy, Dartmouth College, Hanover, 

37:   NH 03755, USA}

38: \address{$^3$ Dept. of Astronomy, Cornell University, Ithaca, NY 14853, USA}

39: \ead{mininni@ucar.edu}

40: 

41: \begin{abstract}

42: Numerical solutions of the incompressible magnetohydrodynamic (MHD) 

43: equations are reported for the interior of a rotating, perfectly-conducting, 

44: rigid spherical shell that is insulator-coated on the inside. A 

45: previously-reported spectral method is used which relies on a Galerkin 

46: expansion in Chandrasekhar-Kendall vector eigenfunctions of the curl. The 

47: new ingredient in this set of computations is the rigid rotation of the 

48: sphere. After a few purely hydrodynamic examples are sampled (spin down, 

49: Ekman pumping, inertial waves), attention is focused on selective 

50: decay and the MHD dynamo problem. In dynamo runs, prescribed 

51: mechanical forcing excites a persistent velocity field, usually turbulent 

52: at modest Reynolds numbers, which in turn amplifies a small seed magnetic 

53: field that is introduced. A wide variety of dynamo activity is observed, 

54: all at unit magnetic Prandtl number. The code lacks the resolution to 

55: probe high Reynolds numbers, but nevertheless interesting dynamo regimes 

56: turn out to be plentiful in those parts of parameter space in which the 

57: code is accurate. The key control parameters seem to be mechanical and 

58: magnetic Reynolds numbers, the Rossby and Ekman numbers (which in our 

59: computations are varied mostly by varying the rate of rotation of the 

60: sphere) and the amount of mechanical helicity injected. Magnetic energy 

61: levels and magnetic dipole behavior are exhibited which fluctuate strongly 

62: on a time scale of a few eddy turnover times. These seem to stabilize as 

63: the rotation rate is increased until the limit of the code resolution is 

64: reached.

65: \end{abstract}

66: \pacs{47.11.-j, 47.11.Kb, 91.25.Cw, 95.30.Qd}

67: \submitto{\NJP}

68: % Comment out if separate title page not required

69: \maketitle

70: 

71: \section{Introduction \label{sec:intro}}

72: In a previous paper, a spectral method for computing incompressible fluid 

73: and magnetohydrodynamic (MHD) behavior inside a sphere was introduced 

74: (Ref. \cite{Mininni06}, hereafter referred to as ``MM''). The emphasis in 

75: MM was on accurate computation of global flow patterns throughout the full 

76: sphere (including the origin) with special attention to the computation of 

77: dynamo action, whereby the kinetic energy of a turbulent conducting fluid 

78: may give rise to macroscopic magnetic fields. The computations were limited 

79: to moderate Reynolds numbers, with boundary conditions in which the normal 

80: components of the velocity field, magnetic field, vorticity, and electric 

81: current density were required to vanish at a rigid, spherical, 

82: insulator-lined, perfectly-conducting shell, and the three components of 

83: the velocity and magnetic field were regular at the origin. The feature 

84: previously lacking that we wish to explore in the present paper is that of 

85: uniform rotation of the sphere. We postpone to the future the investigation 

86: of insulating spherical shells which permit the magnetic field to penetrate 

87: into a non-conducting region outside \cite{Kono02}, though we remark later 

88: on some considerations relevant to this modification.

89: 

90: In Section \ref{sec:equations}, we formulate the equations to be solved for 

91: a uniform-density conducting fluid inside a sphere in a familiar set of 

92: dimensionless (``Alfvenic'') units. We refer to MM for background and 

93: such of the details as remain unchanged. The main changes reported here 

94: are: (1) the introduction of a Coriolis term in the equation of motion 

95: (the centrifugal term may be absorbed in the pressure for incompressible 

96: flow); and (2) the velocity field ${\bf v}$ (instead of the vorticity 

97: $\vomega$) and magnetic field ${\bf B}$ are expanded in orthonormal 

98: Chandrasekhar-Kendall (``C-K'') vector eigenfunctions of the curl 

99: \cite{Chandrasekhar57,Montgomery78,Turner83,Cantarella00,Mininni06}. The 

100: physical situation being simulated is again a perfectly conducting, 

101: mechanically impenetrable sphere coated on the inside with a thin layer 

102: of insulator, but now viewed from a coordinate system that is regarded as 

103: rigidly rotating with the bounding spherical shell. Unsurprisingly, the 

104: dynamical phenomena resulting are markedly different from, and richer 

105: than, they were in the case without rotation.

106: 

107: Section \ref{sec:hdresults} describes the results of some purely hydrodynamic 

108: runs (the code may be readily converted into a Navier-Stokes code by simply 

109: deleting the terms associated with the magnetic field). Included are 

110: examples \cite{Greenspan,Acheson} of: (i) spin down, or decay of  

111: relatively rotating kinetic energy due to the action of viscosity; (ii) 

112: Ekman pumping with flow patterns that result from rotating boundaries; 

113: (iii) internal waves, three-dimensional relatives of meteorological Rossby 

114: waves, that depend on the stabilization introduced by rotation for their 

115: oscillatory features; and (iv) some mechanically-forced runs with a finite 

116: angle between the symmetry axis of the forcing and the axis of rotation. 

117: This fourth case results in columnar vortices due to the effect of rotation 

118: \cite{Acheson,Zhang00} (note convection is absent in our present formulation).

119: 

120: Section \ref{sec:mhdresults} proceeds to a consideration of the MHD case 

121: with an emphasis on selective decay and the kinds of dynamo behavior 

122: we have been able to resolve. The spectral method we use involves 

123: inherently less resolution than some other methods in use, and we have 

124: been careful to study parameter regimes only where we can resolve the 

125: relevant length scales. Selective decay is observed to be somewhat 

126: arrested as the rotation rate is increased. A pleasant surprise has 

127: been the wide variety of dynamo behavior we have been able to resolve 

128: without the need to reach parameter regimes regarded as realistic for 

129: planetary dynamos 

130: \cite{Glatzmaier95,Glatzmaier96,Zhang00,Christensen01,Roberts01,Kono02}. 

131: In the summary, Section \ref{sec:future}, we describe briefly some plans 

132: we have for improving the resolution of the code by some pseudo-spectral 

133: modifications and some intended future diversification of the boundary 

134: conditions.

135: 

136: \section{Computational method \label{sec:equations}}

137: We begin from the magnetohydrodynamic (MHD) equation of motion in a 

138: rotating coordinate frame \cite{Moffatt,Kono02},

139: \begin{equation}

140: \frac{\partial {\bf v}}{\partial t} = {\bf v} \times \vomega + 

141:     {\bf j} \times {\bf B} - \nabla \left({\mathcal P} + \frac{v^2}{2} 

142:     \right) - 2 \vOmega \times {\bf v} + \nu \nabla^2 {\bf v} + {\bf f} , 

143: \label{eq:momentum}

144: \end{equation}

145: and the MHD induction equation,

146: \begin{equation}

147: \frac{\partial {\bf B}}{\partial t} = \nabla \times \left( {\bf v} 

148:     \times {\bf B} \right) + \eta \nabla^2 {\bf B} .

149: \label{eq:induction}

150: \end{equation}

151: In the dimensionless Alfvenic units \cite{Mininni06}, ${\bf v}$ is the 

152: vector velocity field, ${\bf B}$ is the magnetic induction, 

153: $ \vomega = \nabla \times {\bf v}$ is the vorticity, and 

154: ${\bf j} = \nabla \times {\bf B}$ is the electric current density. 

155: The generalized pressure is ${\mathcal P}$. The dimensionless viscosity, 

156: which, in the dimensionless variables, can be interpreted as the reciprocal 

157: of a mechanical Reynolds number, is $\nu$, and the magnetic diffusivity, 

158: which can be interpreted as the reciprocal of a magnetic Reynolds number, 

159: is $\eta$.  The vector field ${\bf f}$ is a solenoidal, externally-applied 

160: forcing field which is intended to mimic the presence of mechanical sources 

161: of excitation of ${\bf v}$. Equations (\ref{eq:momentum}) and 

162: (\ref{eq:induction}) are to be supplemented by the requirements that the 

163: divergences of both ${\bf v}$ and ${\bf B}$ must vanish everywhere. 

164: Dropping equation (\ref{eq:induction}) and the terms in equation 

165: (\ref{eq:momentum}) containing ${\bf j}$ and ${\bf B}$ leaves the forced 

166: Navier-Stokes equation, and dropping ${\bf f}$ leaves the unforced version 

167: of Navier-Stokes. $\Omega$ is the (constant) rotation speed of the 

168: coordinate system, understood to be attached to a rotating spherical 

169: shell that constitutes the boundary and is both mechanically impenetrable 

170: and perfectly conducting, with a thin layer of insulator on the inside 

171: surface.

172: 

173: The non-trivial boundary conditions imposed are that the normal components 

174: of ${\bf v}$, ${\bf B}$, ${\bf j}$, and $\vomega$ shall all vanish at the 

175: radius of a unit sphere centered at the origin. The three components of the 

176: fields ${\bf v}$ and ${\bf B}$ are also required to be regular at the origin. 

177: The vanishing of the normal components of ${\bf v}$ and $\vomega$ at the 

178: surface of the unit sphere are implied by, but do not imply, no-slip 

179: boundary conditions at that radius. Going further with an attempt to 

180: implement fully a set of no-slip boundary conditions raises unresolved 

181: paradoxes with respect to the pressure determination which we prefer not 

182: to confront here (see Refs. \cite{Kress00,Gallavotti,Mininni06} for a 

183: discussion of these), believing that their seriousness and intractability 

184: require consideration in the context of simpler situations than the 

185: present one.

186: 

187: The spectral technique implemented involves expanding ${\bf v}$ and ${\bf B}$ 

188: in terms of C-K functions (defined below):

189: \begin{equation}

190: {\bf v}({\bf r},t) = \sum_{qlm} \xi^v_{qlm}(t) \, {\bf J}_{qlm}({\bf r}) ,

191: \label{eq:expanv}

192: \end{equation}

193: and

194: \begin{equation}

195: {\bf B}({\bf r},t) = \sum_{qlm} \xi^B_{qlm}(t) \, {\bf J}_{qlm}({\bf r}) .

196: \label{eq:expanb}

197: \end{equation}

198: 

199: The C-K functions 

200: \cite{Chandrasekhar57,Montgomery78,Turner83,Cantarella00,Mininni06} 

201: ${\bf J}_i$ are defined by

202: \begin{equation}

203: {\bf J}_i = \lambda \nabla \times {\bf r} \psi_i + \nabla \times \left( 

204:     \nabla \times {\bf r} \psi_i \right) ,

205: \end{equation}

206: where we work with a set of spherical orthonormal unit vectors 

207: $(\hat{r},\hat{\theta},\hat{\phi})$ and the scalar function $\psi_i$ is a 

208: solution of the Helmholtz equation, $(\nabla^2 + \lambda^2) \psi_i = 0$. 

209: The explicit form of $\psi_i$ is

210: \begin{equation}

211: \psi_i (r, \theta, \phi) = C_{ql} \, j_l(|\lambda_{ql}| r) Y_{lm} 

212:     (\theta,\phi) ,

213: \label{eq:psi}

214: \end{equation}

215: where $j_l(|\lambda_{ql}| r)$ is a spherical Bessel function of the first 

216: kind which vanishes at $r=1$ and $Y_{lm}(\theta,\phi)$ is a spherical 

217: harmonic in the polar angle $\theta$ and the azimuthal angle $\phi$. The 

218: subindex $i$ is a shorthand notation for the three indices $(q,l,m)$, where 

219: $q$ indexes the successive values of $\lambda$ that make $j_l$ vanish at 

220: $r=1$ for each value of $l$; $q=1,2,3,\dots$ corresponds to the positive 

221: values of $\lambda$, and $q=-1,-2,-3,\dots$ indexes the negative values; 

222: finally $l=1,2,3,\dots$ and $m$ runs in integer steps from $-l$ to $+l$. 

223: The vectors ${\bf J}_i$ satisfy 

224: \begin{equation}

225: \nabla \times {\bf J}_i = \lambda_i {\bf J}_i ,

226: \end{equation}

227: and with the proper normalization constants are an orthonormal set that 

228: has been shown to be complete \cite{Cantarella00}. The integral relation 

229: expressing the orthogonality of the ${\bf J}_i$ is:

230: \begin{equation}

231: \int {\bf J}_{qlm} \cdot {\bf J}_{q',l',m'}^* dV = \delta_{q,q'} 

232:     \delta_{l,l'} \delta_{m,m'} ,

233: \end{equation}

234: where the asterisk denotes complex conjugate, and with the normalization 

235: constants given by:

236: \begin{equation}

237: C_{ql} =  \left|\lambda_{ql} \, j_{l+1}(|\lambda_{ql}|) \right|^{-1} 

238:     \left[ l(l+1) \right]^{-1/2} .

239: \label{eq:normalization}

240: \end{equation}

241: 

242: The scheme for solving equations (\ref{eq:momentum}) and (\ref{eq:induction}) 

243: is conceptually simple. We substitute the expansions (\ref{eq:expanv}) 

244: and (\ref{eq:expanb}) into equations (\ref{eq:momentum}) and 

245: (\ref{eq:induction}), utilize the fact that the ${\bf J}_i$ are 

246: eigenfunctions of the curl, and then take inner products one at a 

247: time with the individual ${\bf J}_i$. Their orthogonality enables us 

248: to pick off expressions  for the time derivatives of the time dependent 

249: expansion coefficients $\xi_i^v$ and $\xi_i^B$, and equations 

250: (\ref{eq:momentum}) and (\ref{eq:induction}) are thereby converted 

251: into a set of ordinary differential equations for the expansion 

252: coefficients. These appear as 

253: \begin{equation}

254: \frac{\partial \xi^v_i}{\partial t} = \sum_{j,k} A^i_{jk} \left( 

255:     \xi^v_j \xi^v_k - \xi^B_j \xi^B_k \right) + 

256:     2 \sum_j \vOmega \cdot {\bf O}^i_j - \nu \lambda_i^2 

257:     \xi^v_i + \xi^f_i ,

258: \label{eq:CK1}

259: \end{equation}

260: and

261: \begin{equation}

262: \frac{\partial \xi^B_i}{\partial t} = \sum_{j,k} B^i_{jk} \xi^v_j \xi^B_k

263:     - \eta \lambda_i^2 \xi^B_i ,

264: \label{eq:CK2}

265: \end{equation}

266: with the coupling coefficients defined as

267: \begin{equation}

268: A^i_{jk} = \lambda_k I^i_{jk} , \;\;\;\;\; B^i_{jk} = \lambda_i I^i_{jk} ,

269: \end{equation}

270: \begin{equation}

271: I^i_{jk} = \int {\bf J}_i^* \cdot {\bf J}_j \times {\bf J}_k dV , \;\;\;\;\; 

272: {\bf O}^i_j = \int {\bf J}_i^* \times {\bf J}_j dV .

273: \end{equation}

274: 

275: \begin{figure}

276: \begin{center}\includegraphics[width=15cm]{fig1.eps}

277: \end{center}

278: \caption{(a) Time histories of  the decay of mechanical energy for four 

279: hydrodynamic runs (H1 to H4) with identical initial velocity fields but 

280: different rotation rates $\Omega$. (b) Decay rates vs. the square root 

281: of $\Omega$ for the same four runs.}

282: \label{fig:spin}

283: \end{figure}

284: 

285: The infinite set of ordinary differential equations is truncated at some 

286: level above maximum values of $|q|$ and $l$, in the usual manner of a 

287: Galerkin approximation \cite{Canuto}. The evaluation of equations 

288: (\ref{eq:CK1}) and (\ref{eq:CK2}) and the storage of the resulting 

289: arrays of coupling coefficients in tables, are the most demanding 

290: numerical tasks of the problem. Once available, they do not have to 

291: be recomputed, and provide a method for verifying the ideal quadratic 

292: conservation laws with high accuracy \cite{Mininni06}. Also, since 

293: the method is purely spectral and fields are only computed in 

294: real space for visualization purposes, there is no numerical 

295: singularity in the center of the sphere.

296: 

297: The main drawback of the scheme, as with any wholly spectral one, is that 

298: the convolution sums in equations (\ref{eq:CK1}) and (\ref{eq:CK2}) grow 

299: rapidly with increasing maximum values of $|q|$ and $l$, and limit the 

300: resolution when compared to pseudospectral computations utilizing fast 

301: transforms (in practice, a resolution of $\max \{|q|\} = \max \{l\} = 9$ 

302: was used in all the runs). This limits us to modest Reynolds numbers 

303: (all our computations reported here have limited themselves to resolvable 

304: Reynolds numbers). Future plans include pseudospectral modifications to 

305: the evaluation of at least the angular parts of the nonlinear terms in 

306: equations (\ref{eq:momentum}) and (\ref{eq:induction}), as will be 

307: mentioned again in the final summary (Section \ref{sec:future}).

308: 

309: \section{Hydrodynamic examples \label{sec:hdresults}}

310: 

311: Some neutral-fluid effects (good introductions to all of which may be 

312: found in Refs. \cite{Greenspan,Acheson}) are treated first before proceeding 

313: to MHD. It is worth noticing here that although our boundary conditions 

314: are implied by, but do not imply no-slip velocities, several qualitative 

315: and some quantitative agreements are observed with previous experiments 

316: and theory. First we study simple problems in which initially relatively 

317: rotating fluids adjust themselves to rigid rotation with the spherical 

318: shell. We study these decays as functions of the rotation rate $\Omega$. 

319: The prediction \cite{Greenspan} is that the decay of the non-rigid body 

320: components should be exponential, with a decay rate that varies as 

321: $\Omega^{1/2}$. Figure \ref{fig:spin}(a) shows the time histories of the 

322: decay of mechanical energy for four runs with identical initial velocity 

323: fields limited to a few random low mode numbers (large spatial scales). 

324: The runs are delineated as runs H1, H2, H3, and H4. 

325: Specifically, we have as initial conditions a random superposition of 

326: modes with $|q|=1,2$, $l=1,2$, and all allowed values of $m$, with 

327: viscosity $\nu = 0.01$ and values of $\Omega$ of $0$, $1$, $4$, and 

328: $10$ respectively. Each curve is approximately exponential, and when they 

329: are fitted with exponentials, the decay rates plotted vs. the square root of 

330: $\Omega$ appear as in Figure \ref{fig:spin}(b), and are adjudged to be in 

331: satisfactory agreement with theory \cite{Greenspan}. The nonlinearities 

332: excite smaller spatial scales, and the decay process is progressively 

333: enhanced by increasing the rotation rate.

334: 

335: \begin{figure}

336: \begin{center}\includegraphics[width=16cm]{fig2.eps}

337: \end{center}

338: \caption{(a) Enstrophy for runs H1 to H4 as a function of time (same 

339: labels as in Figure \ref{fig:spin}). (b) Energy spectra at $t=1$ for 

340: the same runs, and (c) energy spectra at $t=6$. In all cases, solid 

341: lines are for $\Omega = 0$, dotted lines for $\Omega =1$, dashed lines 

342: for $\Omega =4$, and dash-dotted lines for $\Omega =10$.}

343: \label{fig:ekman}

344: \end{figure}

345: 

346: A second feature of rotating spherical flows observed in our code is the 

347: development of Ekman-like layers and the action of Ekman pumping 

348: \cite{Greenspan,Acheson} (see also \cite{Dormy98} for a detailed study 

349: in rotating spherical shells). The flow patterns are characterized by the 

350: development of interior vortical flows with some symmetries, and thin 

351: layers that separate the large vortices and also lie along the wall 

352: boundary layers. These have a characteristic thickness of the order of 

353: $\delta \sim E_K^{1/2} R$, where $R=1$ is the radius of the sphere, and 

354: the Ekman number is $E_K = \nu \Omega^{-1} L^{-2}$, with $L$ a 

355: characteristic scale of the flow. The ability of the code to compute 

356: these layers is limited by its resolution.  Realistic values of the 

357: Ekman number are, for planetary core regimes 

358: \cite{Moffatt,Zhang00,Kono02}, beyond our range. In all the runs 

359: presented here we will limit ourselves to cases where $\delta$ can 

360: be properly resolved with the number of modes used in the simulations. The 

361: presence of the Ekman layers is concomitant with the development of smaller 

362: spatial scales and hence of more rapid dissipation. Figure 

363: \ref{fig:ekman} illustrates this fact. Figure \ref{fig:ekman}(a) is the 

364: integrated squared vorticity, or enstrophy, for the four runs whose decay 

365: has just been seen to be exponential. The largest rotation rate corresponds 

366: to the curve with the highest early peak in enstrophy spectrum, the second 

367: highest with the second largest, and so on. Figure \ref{fig:ekman}(b) 

368: shows the energy spectra at $t=1$ for the four runs and Figure 

369: \ref{fig:ekman}(c) shows the same energy spectra at $t=6$. In every 

370: case, the flatter spectra, and hence the shorter wavelength dominances, 

371: correspond to the higher values of $\Omega$. This is the result of 

372: the formation of a thinner boundary layer as $\Omega$ is increased.

373: 

374: \begin{figure}

375: \begin{center}\includegraphics[width=15cm]{fig3.eps}

376: \end{center}

377: \caption{Above: Mechanical energy density and velocity field lines in run 

378: E1, at $t=0$ (left) and at $t=6$ (right). Below: idem for run E2. For 

379: convenience, energies and field lines are always shown in pairs, with 

380: energy densities on the left and field lines on the right. The field lines 

381: change color according to the distance integrated from the initial point, 

382: from red to yellow, blue, and magenta. The red, green, and blue arrows 

383: indicate respectively the $x$, $y$, and $z$ axis. $\vOmega$ is in the 

384: $z$ direction. In both cases, the energy density is symmetric with 

385: respect to the equator, while the flow itself is antisymmetric in run E1 

386: (above).}

387: \label{fig:3Dekman}

388: \end{figure}

389: 

390: \begin{figure}

391: \begin{center}\includegraphics[width=13cm]{fig4.eps}

392: \end{center}

393: \caption{Equatorial cross section of the velocity field for an inertial 

394: wave in the rotating sphere. The reference frame is fixed to the sphere. 

395: $\vOmega$ is in the $z$ direction. The axial velocity is indicated by 

396: the colors, while the radial and azimuthal velocities are indicated by 

397: the arrows. For this mode, the frequency is $\varpi \approx 49$.}

398: \label{fig:wave}

399: \end{figure}

400: 

401: The range of Ekman layer behavior we have been able to observe is very wide. 

402: We show in Figure \ref{fig:3Dekman} the results of two simulations (labeled 

403: E1 and E2) with initially random axisymmetric ($m=0$) velocity fields which 

404: are purely azimuthal. For both runs, the initial velocity is proportional to 

405: the difference between ${\bf J}_{q,l,0}$ and ${\bf J}_{-q,l,0}$, which has 

406: only an azimuthal component. For run E1, $q=1$ and $l=1$, and for run E2, 

407: $q=1$ and $l=2$. Both runs have $\nu=0.01$ and $\Omega=10$. The time evolution 

408: of the runs is similar to the evolution displayed in Figures \ref{fig:spin} 

409: and \ref{fig:ekman}. However, the axisymmetric initial conditions in runs 

410: E1 and E2 make visualization of flow patters easier. Figure \ref{fig:3Dekman} 

411: shows the initial and late-time flow patterns for these runs, using the 

412: VAPOR graphics package \cite{vapor} that will be repeatedly used throughout 

413: this paper for graphical demonstrations. The rotation generates poloidal 

414: components of the velocity field fast, and at late times different patterns 

415: are observed depending on the initial value of $l$. In run E1, at late 

416: times the flow displays a poloidal circulation on top of the initial 

417: toroidal field: the flow is directed towards the center of the sphere 

418: along the axis of rotation, and a return flow is observed in both 

419: hemispheres close to the wall. In other words, the flow can be 

420: described as the superposition of a toroidal differential rotation and a 

421: poloidal meridional circulation. This circulation is radially outward in 

422: the meridional plane, directed towards the poles close to the wall, and 

423: redirected toward the equatorial plane again as the flow gets close to 

424: the poles. Both hemispheres show the same pattern. In run E2 the pattern 

425: is more complex, and vertical velocities are observed in the vicinity of 

426: the axis of rotation, while high and at intermediate latitudes a poloidal 

427: circulation is generated.

428: 

429: \begin{figure}

430: \begin{center}\includegraphics[width=14cm]{fig5.eps}

431: \end{center}

432: \caption{(a) Magnetic energy $E_M$ (solid line), kinetic energy $E_V$ 

433: (dotted line), and magnetic helicity $H_M$ (dashed line) in the selective 

434: decay run S1. (b) Same quantities for run S2. The two power laws are 

435: indicated in the figures only as a reference.}

436: \label{fig:selective1}

437: \end{figure}

438: 

439: As a third hydrodynamic test of the code we demonstrate inertial wave 

440: motion in the small-amplitude limit. Equations (\ref{eq:CK1}) and 

441: (\ref{eq:CK2}) can be linearized in powers of a small departure from a 

442: uniform rotation velocity; then solutions can be sought which vary with 

443: time as $e^{i \varpi t}$. The resulting linear homogeneous algebraic 

444: system can be solved in a Galerkin approximation by expanding the velocity 

445: and vorticity in the C-K functions. An anti-Hermitian matrix results whose 

446: eigenfunctions can be found numerically and whose corresponding eigenvalues 

447: $i \varpi$ may be computed numerically in the process.  Then any one of 

448: the oscillatory modes can be loaded numerically into the ideal version 

449: of the code and run with the overall amplitudes chosen to be very small. 

450: The time evolution is accurately predicted by the computation of the 

451: single modes, which are standing waves. Figure \ref{fig:wave} shows four 

452: equatorial cross-sections of the sphere at different times with the axial 

453: velocity indicated by the color codes, and the radial and azimuthal 

454: velocities indicated by arrows ($\Omega = 10$ in this run). The times 

455: are chosen to be one quarter-period apart. The oscillation frequency 

456: $\varpi$ for the particular mode shown as obtained from the eigenvalue 

457: problem is in good agreement with the results obtained from the fully 

458: non-linear code ($\varpi \approx 49$).

459: 

460: In forced hydrodynamic simulations in the presence of strong rotation, 

461: we also observe the development of columnar structures in the flow, aligned 

462: with the axis of rotation. The discussion of these simulations will be 

463: left for the next section, where the connection between these columns and 

464: dynamo action will be considered.

465: 

466: \section{MHD and the dynamo \label{sec:mhdresults}}

467: 

468: \subsection{Selective decay in the sphere}

469: 

470: \begin{figure}

471: \begin{center}\includegraphics[width=8cm]{fig6.eps}

472: \end{center}

473: \caption{Relative helicity $H_M/E_M$ in selective decay runs S1 (solid) 

474: and S2 (dashed).}

475: \label{fig:selective2}

476: \end{figure}

477: 

478: Before passing to a discussion of the mechanically driven spherical dynamo, 

479: we present first the results of two tests of three-dimensional MHD 

480: ``selective decay'' as affected by the presence of rotation. Selective 

481: decay is a familiar turbulent decay process, usually incompressible, 

482: long studied in periodic geometry \cite{Matthaeus80,Ting86,Kinney95}, 

483: wherein one ideal invariant is cascaded to short wavelengths and 

484: dissipated while another remains locked into long wavelengths and is 

485: approximately conserved. The phenomenon, closely connected with inverse 

486: cascade processes for driven systems, leads toward a state in which 

487: the ratio of the two ideal invariants involved is minimized and which 

488: therefore is accessible to variational methods. In 3D MHD, a quantity 

489: that may be preferentially dissipated is the total energy $E = E_V+E_M$ 

490: (kinetic plus magnetic) while magnetic helicity $H_M$ may be 

491: approximately conserved. Under other circumstances, energy may be 

492: dissipated while cross helicity $K$ is approximately conserved, 

493: leading to the phenomenon of ``dynamic alignment,'' 

494: \cite{Grappin83,Pouquet86,Ghosh88} in which the velocity field and 

495: magnetic fields are highly correlated. The definitions of $H_M$ and 

496: $K$ are

497: \begin{equation}

498: H_M = \frac{1}{2} \int {\bf A} \cdot {\bf B} dV,

499: \end{equation}

500: \begin{equation}

501: K = \frac{1}{2} \int {\bf u} \cdot {\bf B} dV,

502: \end{equation}

503: where ${\bf A}$ is the vector potential whose curl is ${\bf B}$, and 

504: the integrals run over the entire volume of the fluid. What we are 

505: interested in demonstrating here is the effect that rotation has on 

506: the development of the selective decay of total energy relative to 

507: magnetic helicity inside a sphere. It will be useful for these and 

508: other purposes, to have a definition of the magnetic dipole moment:

509: \begin{equation}

510: \vmu = \frac{1}{2} \int {\bf r} \times {\bf j} \, dV \, .

511: \end{equation}

512: which is seen to be readily expressible in terms of the expansion 

513: coefficients for ${\bf B}$ \cite{Mininni06}.

514: 

515: \begin{figure}

516: \begin{center}\includegraphics[width=7.2cm]{fig7a.eps}

517:               \includegraphics[width=7.2cm]{fig7b.eps}

518: \end{center}

519: \caption{(a) Trace of the dipole moment on the surface of the unit sphere 

520: (above) and amplitude of the dipole moment as a function of time (below) 

521: for the selective decay run S1. (b) Same quantities for run S2.}

522: \label{fig:selec_dip}

523: \end{figure}

524: 

525: The initially excited modes for the two runs we will present 

526: (``S1'' and ``S2'') are those for $q=\pm 3$, $l=3$, and all possible 

527: values of $m$. The initial values chosen for the expansion coefficients 

528: are:

529: \begin{eqnarray}

530: & \xi^v_{\pm 3,3,0} = -u_0 , \;\; \xi^v_{\pm 3,3,0<m\le 3} = u_0 (1+i) , \\

531: & \xi^B_{3,3,0} = \frac{10}{6} \xi^B_{-3,3,0} = b_0 , \;\; 

532:   \xi^B_{3,3,0<m\le 3} = \frac{10}{6} \xi^B_{-3,3,0<m\le 3} = b_0 (1-i) , 

533: \end{eqnarray}

534: with $u_0$ and $b_0$ chosen so that at $t=0$, the magnetic and kinetic 

535: energies are $E_M=E_V \approx 0.5$, $K=0$, and $H_M \approx 0.034$. Some 

536: helicity cancellation occurs because of the two signs of $\lambda$ (or $q$). 

537: As a comparison, note that for the $q=3$, $l=3$ mode alone, $H_M/E_M$ 

538: is no more than about $0.072$ (this is the maximum value of $|H_M/E_M|$ 

539: if only modes with $|q|=3$, $l=3$, and one sign of $\lambda$ are excited).

540: In both runs, the magnetic diffusivity and kinematic viscosity are 

541: $\nu=\eta=0.006$; the Reynolds numbers are $R_e \approx R_m \approx 170$, 

542: based on the radius of the sphere. The two runs differ by the values of 

543: $\Omega$ chosen, which are $2$ and $12$, respectively. These mean that 

544: the Rossby and Ekman numbers of the two runs are, respectively, 

545: $R_O=U(\Omega R)^{-1}=0.5$, $E_K=0.003$ for S1 and $R_O=0.083$, 

546: $E_K=0.0005$ for S2. The decay of magnetic energy, kinetic energy, 

547: and magnetic helicity for Runs S1 and S2 are shown in Figure 

548: \ref{fig:selective1}. The behavior in Run S1 is not significantly different 

549: from the non-rotating case \cite{Mininni06}. Note that in both runs, the 

550: kinetic energy at late times is negligible, and that magnetic and kinetic 

551: energies decay faster than the magnetic helicity. However, the decay of 

552: all these quantities in S2 seems to be faster than in S1.

553: 

554: The relative helicity, $H_M/E_M$, is shown for Runs S1 and S2 in Figure 

555: \ref{fig:selective2}. It will be seen that the increased rate of rotation 

556: in S2 has somewhat arrested the selective decay, for reasons not totally 

557: understood. It may be that the rotation has resulted in sufficient 

558: two-dimensionalization of the flow \cite{Greenspan,Acheson,Zhang00} 

559: that the inherently three dimensional nature of the selective decay 

560: has been compromised. But from Figure \ref{fig:selective2} it can also 

561: be seen that the ratio of $H_M$ to $E_M$ has approached reasonably 

562: closely to its maximal value of 

563: $\min^{-1}\{|\lambda_{ql}|\} \approx 0.22$ (the maximum value of 

564: $|H_M/E_M|$ when only modes with $|q|=1$, $l=1$, and one sign of 

565: $\lambda$ are excited). The maximal value would indicate a total 

566: disappearance of the non-rotational kinetic energy (i.e., a rigid 

567: rotation), and all the magnetic energy in the largest-scale modes 

568: (smallest $|\lambda|$) allowed by the boundary conditions: a 

569: magnetized, ``frozen'' condition.

570: 

571: \begin{figure}

572: \begin{center}\includegraphics[width=15cm]{fig8.eps}

573: \end{center}

574: \caption{Above: kinetic energy density and velocity field lines (left) 

575: and magnetic energy density and magnetic field lines (right) in the initial 

576: conditions of selective decay runs S1 and S2. Middle: same fields at $t=14$ 

577: in run S1. Below: same fields at $t=14$ in run S2. Colors and labels are 

578: as in Figure \ref{fig:3Dekman}.}

579: \label{fig:3Dselective}

580: \end{figure}

581: 

582: The behavior of the dipole moment for the two runs is shown in Figure 

583: \ref{fig:selec_dip}. The solid lines are the traces of the projected 

584: direction of the dipole moment on the surface of the sphere as functions 

585: of the time, and the orientation is such that the axis of rotation points 

586: upward. In the lower parts of these two figures, the magnitude of the dipole 

587: moment is plotted vs. time. In both cases, it will be seen that the dipole's 

588: orientation initially wanders erratically near its initial position, and 

589: finally ends at a ``mid-latitude'' direction not far from where it began. 

590: This was something of a surprise to us, since we had expected it to line 

591: up with the axis of rotation or at least close to it. Figure 

592: \ref{fig:3Dselective} shows VAPOR plots with the energy densities, velocity 

593: and magnetic field line structure for the initial conditions in Runs S1 and 

594: S2, as well as the late stages of both runs ($t=14$) when the selective 

595: decay process has saturated and all the nonlinear terms are small, preventing 

596: any further evolution of the system except for dissipation. For strong 

597: rotation (run S2), the velocity field  is quasi-two dimensional and 

598: develops column-like structures at late times, while the magnetic field 

599: is highly anisotropic (although the dipole moment is not aligned with the 

600: axis of rotation). Note magnetic field lines in this case are aligned 

601: with the $z$ axis (the axis of rotation), and velocity field lines are 

602: mostly toroidal. Actually, the ratio $|v_\phi / v_z|$ at $t=14$ averaged 

603: over the whole volume for this run at $t=14$ is $\approx 13$.

604: 

605: \begin{table}

606: \caption{Dynamo runs: $q$ and $l$ give the scales where mechanical energy 

607: is injected by the forcing, $\Omega$ is the rotation rate, and $\nu$ and 

608: $\eta$ are respectively the kinematic viscosity and magnetic diffusivity. 

609: ``Helical'' indicates whether the forcing injects mechanical helicity, and 

610: ``Axisym.'' indicates whether the forcing is axisymmetric. $\alpha$ is the 

611: angle between $\vOmega$ and the $z$-axis (the axis of symmetry in 

612: axisymmetric forcings). Finally, $R_O$, $E_K$, and $R_e$ are respectively 

613: the Rossby, Ekman, and Reynolds numbers, all based on the radius of the 

614: sphere. A resolution of $\max \{|q|\} = \max \{l\} = 9$ was used in all 

615: the runs.}

616: \label{table:dynamo}

617: \begin{indented}

618: \item[]\begin{tabular}{@{}lllllllllll}

619: \br

620: Runs & $q$ & $l$ & $\Omega$ & $\nu=\eta$         & Helical & Axisym.

621:      & $\alpha$  & $R_O$    & $E_K$              & $R_e$ \\

622: \mr

623: D1   &  1  &  1  &    2     & $2\times 10^{-3}$  & No      & Yes 

624:      & $30^\circ$& $0.5$    & $1\times 10^{-3}$  & $500$ \\

625: D2   &  2  &  2  &    2     & $2\times 10^{-3}$  & No      & Yes  

626:      & $30^\circ$& $0.5$    & $1\times 10^{-3}$  & $500$ \\

627: D3   &  3  &  3  &    2     & $2\times 10^{-3}$  & No      & Yes  

628:      & $30^\circ$& $0.5$    & $1\times 10^{-3}$  & $500$ \\

629: D4   &  3  &  3  &    4     & $2\times 10^{-3}$  & No      & Yes  

630:      & $30^\circ$& $0.25$   & $5\times 10^{-4}$  & $500$ \\

631: D5   &  3  &  3  &    8     & $2\times 10^{-3}$  & No      & Yes  

632:      & $30^\circ$& $0.125$  & $2.5\times 10^{-4}$& $500$ \\

633: D6   &  3  &  3  &    8     & $4\times 10^{-3}$  & No      & Yes  

634:      & $30^\circ$& $0.125$  & $5\times 10^{-4}$  & $250$ \\

635: D7   &  3  &  3  &    16    & $4\times 10^{-3}$  & No      & Yes  

636:      & $30^\circ$& $0.0625$ & $2.5\times 10^{-4}$& $250$ \\

637: D8   &  3  &  3  &    8     & $2\times 10^{-3}$  & No      & Yes  

638:      & $30^\circ$& $0.125$  & $2.5\times 10^{-4}$& $500$ \\

639: D9   &  3  &  3  &    16    & $4\times 10^{-3}$  & No      & Yes  

640:      & $20^\circ$& $0.125$  & $5\times 10^{-4}$  & $250$ \\

641: D10  &  3  &  3  &    16    & $4\times 10^{-3}$  & No      & No  

642:      & $0^\circ$ & $0.125$  & $5\times 10^{-4}$  & $250$ \\

643: D11  &  3  &  3  &    16    & $4\times 10^{-3}$  & Yes     & No 

644:      & $0^\circ$ & $0.125$  & $5\times 10^{-4}$  & $250$ \\

645: D12  &  3  &  3  &    16    & $4\times 10^{-3}$  & Yes     & No 

646:      & $90^\circ$& $0.125$  & $5\times 10^{-4}$  & $250$ \\

647: \br

648: \end{tabular}

649: \end{indented}

650: \end{table}

651: 

652: \subsection{Dynamos}

653: We turn now to the case of the forced spherical dynamo computations, in 

654: which specified non-zero forcing functions ${\bf f}$ are added to the right 

655: hand side of equation (\ref{eq:momentum}) or (\ref{eq:CK1}) to provide 

656: a persistently-active, non-decaying velocity field. After a purely 

657: hydrodynamic run to reach a statistically steady state, very small 

658: magnetic fields are introduced to see if the velocity fields will cause 

659: them to amplify, and attention focuses on questions like the orientation 

660: of the resulting magnetic dipole moment relative to the axis of rotation 

661: and the dipole moment's magnitude. We are also interested in the kinetic 

662: and magnetic energy spectra that result, and how the eponymous 

663: dimensionless numbers of the rotating fluid (Reynolds, magnetic Reynolds, 

664: Rossby, Ekman) influence the magnetic quantities. These appear to be 

665: essential control parameters of the problem. The range of possibilities 

666: is clearly very wide, and we have not begun to explore the entire space 

667: of possible parameters. Rather, we content ourselves with showing samples 

668: of different behavior that have emerged for different combinations that 

669: lead to regimes which the code will resolve satisfactorily. 

670: Our efforts should not be compared with explorations of the space 

671: of parameters in realistic geodynamo simulations (see e.g. 

672: \cite{Christensen06}) but rather as an extension of dynamo simulations 

673: of incompressible MHD flows (often done using periodic boundary conditions 

674: \cite{Meneguzzi81,Brandenburg01,Ponty05,Mininni05b}) to include the effect 

675: of boundaries and rotation.

676: 

677: Several of the forcing functions used in the runs we will display are 

678: axisymmetric, but their axes of symmetry are not aligned with the axis of 

679: rotation. The resulting overall asymmetry quickly excites all the available 

680: retained modes, to some degree. The general form of the axisymmetric forcing 

681: function used is

682: \begin{equation}

683: {\bf f} = \sum_{ql} \xi^f_{q,l,0} \left( A {\bf J}_{q,l,0} - 

684:     B {\bf J}_{-q,l,0} \right) .

685: \label{eq:force1}

686: \end{equation}

687: For any value of $A$ and $B$, this superposition of C-K functions gives an 

688: axisymmetric forcing ${\bf f}$. For $A=B=1$, there will be no net helicity 

689: involved in the forcing (the curl of ${\bf f}$ is perpendicular to 

690: ${\bf f}$), and the only non-vanishing component of velocity that 

691: is forced is the $\phi$ component (the azimuthal component with respect 

692: to the axis of symmetry of the forcing, the $z$ axis). This forcing can 

693: be considered as a simple differential rotation, where the number of nodes 

694: in $v_\phi (r,\theta)$ is controlled by the values of $q$ and $l$. 

695: The axis of rotation is typically oriented at some specified angle $\alpha$ 

696: (often $30^\circ$) to the forcing function's axis of symmetry (the polar 

697: axis, in spherical coordinates). Thus the rotational motion and the forcing 

698: have no shared symmetry, and the resulting mechanical motion is totally 

699: asymmetrical.

700: 

701: \begin{figure}

702: \begin{center}\includegraphics[width=15cm]{fig9.eps}

703: \end{center}

704: \caption{Left: kinetic energy density and velocity field lines at late 

705: times in run D1, when the dynamo has saturated. Right: magnetic energy 

706: density and magnetic field lines at the same time. Colors and labels are 

707: as in Figure \ref{fig:3Dekman}.}

708: \label{fig:D13D}

709: \end{figure}

710: 

711: \begin{figure}

712: \begin{center}\includegraphics[width=7.2cm]{fig10a.eps}

713:               \includegraphics[width=7.2cm]{fig10b.eps}

714: \end{center}

715: \caption{Left: kinetic (dashed black line) and magnetic (solid blue line) 

716: energy as a function of time in run D1. Right: cosine of the angle between 

717: $\vOmega$ and $\vmu$ in the same run.}

718: \label{fig:D1evolution}

719: \end{figure}

720: 

721: Non-axisymmetric forcing functions are obtained by superposing C-K modes 

722: with $m \neq 0$, i.e.

723: \begin{equation}

724: {\bf f} = \sum_{qlm} \xi^f_{qlm} {\bf J}_{qlm} ,

725: \label{eq:force2}

726: \end{equation}

727: and when $\xi^f_{qlm} = \xi^f_{-q,l,m}$ the forcing is non-helical. In 

728: any other case the curl of ${\bf f}$ has a projection into ${\bf f}$, 

729: and the forcing injects mechanical helicity into the flow.

730: 

731: In an effort to systematize the runs we have done and the reasons we have 

732: done them, we have assembled the important parameters for the dynamo runs 

733: (labeled D1 through D12) in Table \ref{table:dynamo}. Listed in Table 

734: \ref{table:dynamo} are: the $q$ and $l$ values where the forcing was 

735: concentrated (determining its characteristic length scale); the rigid 

736: rotation rate; the kinematic viscosity (reciprocal Reynolds number, if 

737: the kinetic energy is close to unity) and magnetic diffusivity (this 

738: study is restricted to the magnetic Prandtl number $P_M = \nu/\eta = 1$ 

739: case); an indication of whether the forcing was axisymmetric or not; the 

740: angle between the axis of rotation and the axis of symmetry of the forcing, 

741: when the forcing function has an internal symmetry; the Rossby and Ekman 

742: numbers of the flow into which the seed magnetic field is introduced; and 

743: whether or not the forcing injected net mechanical helicity.

744: 

745: It is perhaps worthwhile to say a word about the motivation for the 

746: progression of runs shown in Table \ref{table:dynamo}. The first remark is 

747: that it seems to be relatively easy to excite a dynamo and generate a dipole 

748: moment, but relatively difficult to generate one that behaves according to 

749: our predispositions and hopes: a dipole moment with some alignment with the 

750: axis of rotation, and that reverses periodically or randomly with long times 

751: between reversals (compared with the turbulent turnover time). We have found 

752: wild oscillations in both magnitude and direction that seem to decrease with 

753: decreasing Rossby and Ekman numbers. Since the Reynolds numbers are limited 

754: by the resolution, the principal means of decreasing both the Rossby and 

755: Ekman numbers is by increasing the rotation rate $\Omega$. This, however, 

756: eventually decreases the thickness of the boundary layer below the resolution 

757: of the code, and beyond that point, the accuracy of the computations becomes 

758: suspect. The lowest Rossby and Ekman numbers that appear in the entries of 

759: Table \ref{table:dynamo} represent those below which the resolution 

760: limitations are encountered. It will be seen below that as we progress 

761: toward them, the dipolar behavior looks gradually more like what we might 

762: expect, the dipole moment gets stronger and more aligned with the axis of 

763: rotation, and the time between reversals gets larger.

764: 

765: We begin by showing some results for a weak dynamo situation, Run D1, 

766: where the magnetic energy $E_M$ remains always much smaller than the kinetic 

767: energy $E_V$. In the forcing function (\ref{eq:force1}), the driven modes 

768: have $q=l=1$, $A=B=1$, $\nu=\eta=0.002$, $\alpha=30^\circ$, and $\Omega=2$. 

769: The amplitude of the forced modes is $|\xi^f_{q,l,0}|= 0.4$. The two 

770: Reynolds numbers, $R_e$ and $R_m$, based on the measured r.m.s. velocity 

771: before the magnetic seed is introduced and the radius of the sphere, are 

772: both about 500. The Rossby number is $R_O=0.5$, and the Ekman number also 

773: based on the radius of the sphere is $E_K=1\times 10^{-3}$. Figure 

774: \ref{fig:D13D} shows the streamlines of the flow, magnetic field lines, and 

775: the energy densities at late times, once the dipole is established 

776: ($t\approx 80$). In the steady state, the magnetic dipole moment $|\vmu|$ 

777: is of the order of $0.001$ and the ratio of magnetic to kinetic energies 

778: is $E_M/E_V \approx 0.0005$.  The magnetic energy rises to a characteristic 

779: value and oscillates somewhat irregularly as shown in Figure 

780: \ref{fig:D1evolution}, while the cosine of $\gamma$ (the angle between 

781: the axis of rotation and the orientation of $\vmu$) oscillates with 

782: roughly the same periodicity (see also Figure \ref{fig:D1evolution}) 

783: with almost a $180^\circ$ variation in some cases, and with an average 

784: orientation almost perpendicular to $\vOmega$. In this weak case D1, 

785: the hydrodynamic flow remains laminar, stable, and almost time independent.

786: 

787: \begin{figure}

788: \begin{center}\includegraphics[width=14cm]{fig11.eps}

789: \end{center}

790: \caption{(a) Magnetic energy in dynamo runs D1 (solid), D2 (dotted), D3 

791: (dashed), D4 (dash-dotted) and D5 (dash-triple dotted). (b) Magnetic energy 

792: spectrum for the same runs at late times, after non-linear saturation of 

793: the dynamo takes place. The arrows on top indicate the scale where mechanical 

794: energy is injected in each run; from left to right: D1, D2, and D3-D5. The 

795: magnetic energy spectrum corresponding to Run D1 has been multiplied by a 

796: factor of 100.}

797: \label{fig:trend1}

798: \end{figure}

799: 

800: The global evolution of the system is similar to what we will show in the 

801: remaining runs. Once the magnetic field is introduced at $t=0$, and if 

802: $R_m$ is large enough, the magnetic field is amplified exponentially (this 

803: stage is often called the ``kinematic dynamo'' regime) until the Lorentz 

804: force modifies the flow and non-linear saturation is reached. At late 

805: times, an MHD state is reached in which magnetic energy is sustained 

806: against Ohmic dissipation by dynamo action. In run D1, the flow is 

807: reminiscent of the hydrodynamic flow previously described in Figure 

808: \ref{fig:3Dekman}. Although the forcing is axisymmetric and purely 

809: toroidal, rotation generates a poloidal circulation and as a result the 

810: flow points outwards in the equatorial plane, and inwards along the axis 

811: of rotation. Each hemisphere has mechanical helicity of opposite signs, 

812: while the net mechanical helicity of the system fluctuates around zero. 

813: The magnetic field seems to be sustained by an $\alpha$-$\Omega$ mechanism, 

814: where the differential rotation is sustained by the mechanical forcing 

815: and the $\alpha$-effect is given by the Ekman-like circulation. Magnetic 

816: energy is concentrated in the center of the sphere, where the flow has 

817: a stagnation point.

818: 

819: \begin{figure}

820: \begin{center}\includegraphics[width=5cm]{fig12a.eps}

821:               \includegraphics[width=5cm]{fig12b.eps}

822:               \includegraphics[width=5cm]{fig12c.eps}

823: \end{center}

824: \caption{Left: (a) Trace of the dipole moment on the surface of the unit 

825: sphere (above), amplitude of the dipole moment (middle), and cosine of the 

826: angle between the dipole moment and the axis of rotation as a function of 

827: time, for run D3. (b) Same quantities for run D5. (c) Same quantities for 

828: run D7.}

829: \label{fig:trend2}

830: \end{figure}

831: 

832: A considerably stronger dynamo than D1 is represented in Run D2, where 

833: in the excitation function (\ref{eq:force1}) we choose $q=l=2$ and 

834: $|\xi^f_{q,l,0}|=1.1$. Again, $A=B=1$, $\nu=\eta=0.002$, and $\Omega=2$. 

835: The magnetic moment rises from zero, and attains a typical magnitude 

836: of $|\vmu|\approx 0.5$, about 100 times larger than in D1. The Reynolds 

837: numbers are $R_e = R_m \approx 500$, and the ratio of magnetic to kinetic 

838: energy oscillates around $0.15$. The cosine of $\gamma$ (the angle 

839: between $\vmu$ and $\vOmega$) oscillates wildly in time, and the 

840: orientation of the dipole shows no preferred direction. The only 

841: difference between run D1 and D2 is the change in the forcing scale, and 

842: the result seems to indicate a separation of scales between the forcing 

843: and the largest scale in the system helps the dynamo, as indicated by 

844: the larger ratio $E_M/E_V$ in run D2, and as also reported before in 

845: simulations with periodic boundary conditions 

846: \cite{Brandenburg01,Gomez04,Mininni05a}. In all these runs, the largest 

847: available scale is fixed and given by the inverse of the smallest $|\lambda|$ 

848: (corresponding to $\lambda_{\pm1,1}$) and determined by the radius of the 

849: sphere ($R=1$), while the separation between this scale and the forcing 

850: scale is controlled by the values of $q$ and $l$ in the forcing function 

851: (see Table \ref{table:dynamo}). The largest the values of $|q|$ and $l$, 

852: the smallest the scale where mechanical energy is injected.

853: 

854: Several runs were done (see e.g. the runs D1 to D5 in Table 

855: \ref{table:dynamo}) in which the forcing was gradually moved to smaller 

856: scales (from $q=l=1$ in D1 to $q=l=3$ in D3), and in which the rotation 

857: rate was progressively increased (from $\Omega = 2$ in runs D1-D3 to 

858: $\Omega = 8$ in D5). As these changes were made, the amplitude of the 

859: forced modes $|\xi^f_{q,l,m}|$ in equation (\ref{eq:force1}) had to be 

860: increased in order to reach a statistically steady state with r.m.s. 

861: velocities of order one before the magnetic field was introduced (their 

862: amplitudes were $0.4$, $1.1$, $1.6$, $2.2$, and $3.6$, from run D1 to D5). 

863: The reason for this can be understood as follows: the Coriolis force in 

864: equation (\ref{eq:momentum}) acts as a restoring force that opposes the 

865: growth of perturbations. This is also the reason why this system can 

866: sustain waves, as was shown in Section \ref{sec:hdresults}. In all these 

867: runs, $A$, $B$, $\nu$, $\eta$, and the angle of inclination $\alpha$ were 

868: kept the same. As will be seen from Table 1, all the forcing was non-helical.

869: 

870: \begin{figure}

871: \begin{center}\includegraphics[width=8cm]{fig13.eps}

872: \end{center}

873: \caption{Magnetic energy in dynamo runs D1 (solid), D2 (dotted), D3 

874: (dashed), D5 (dash-dotted) and D7 (dash-triple dotted). Note the 

875: intermittent growth of magnetic energy at early times in Run D7.}

876: \label{fig:intermittency}

877: \end{figure}

878: 

879: All five runs are considered to have been able to resolve the Ekman layers 

880: that developed, but they would likely not have been resolved at higher 

881: values of $\Omega$. Figure \ref{fig:trend1} shows the general trend 

882: resulting from the smaller scale forcing and increased rotation. Figure 

883: \ref{fig:trend1}(a) shows a logarithmic-linear plot of the total magnetic 

884: energy versus time for runs D1 to D5. The  five runs showed increasingly 

885: large growth rate, a higher saturation level of $E_M/E_V$, and increasing 

886: $|\vmu|$. In Run D5, the ultimate ratio of $E_M$ to $E_V$ was about $0.4$ 

887: and $|\vmu|$ was close to unity. Figure \ref{fig:trend1}(b) shows magnetic 

888: energy spectra for runs D1 to D5, with decreasing Rossby number $R_O$ 

889: (based on the radius of the sphere) of $0.5$ (runs D1 to D3), $0.25$ 

890: (D4), and $0.125$ (D5). It will be seen that there develops a 

891: small excess of magnetic energy in scales larger than the forcing scale 

892: with decreasing Rossby numbers. Note the development of a ``bump'' 

893: in the magnetic energy spectrum at $\lambda \approx 9$ in runs D4 and 

894: D5.

895: 

896: \begin{figure}

897: \begin{center}\includegraphics[width=15cm]{fig14.eps}

898: \end{center}

899: \caption{Above: kinetic energy density and velocity field lines at late 

900: times in Run D5 (left), and magnetic energy density and magnetic field 

901: files at the same time for the same run. Below: same quantities for Run 

902: D7. Note the development of anisotropies in the presence of large 

903: $\vOmega$ in this run.}

904: \label{fig:D5D7}

905: \end{figure}

906: 

907: A second trend is indicated in Figure \ref{fig:trend2}: namely the 

908: orientation of the dipole moment onto the axis of rotation seems less 

909: erratic with decreasing Rossby and Ekman numbers. That is, there are more 

910: eddy turnover times (in units of $R/U$) between the near reversals 

911: as $R_O$ and $E_K$ are decreased, and the projection of the dipole 

912: moment on the unit sphere gets more localized around the two ``poles'' 

913: defined by the axis of rotation. We borrow here the term ``reversal'' 

914: from the palaeomagnetic record, where during a reversal the orientation 

915: of the dipole moment changes about $180^\circ$ and its amplitude 

916: decreases, in opposition to an ``excursion'' in which the direction and 

917: the orientation of the dipole moment changes in a short period of time 

918: without resulting in a full reversal \cite{Valet05}.

919: 

920: To verify this behavior, in Runs D6, D7, and D8 we successively 

921: decreased $R_O$ and $E_K$ while keeping the other parameters constant. 

922: At the present resolution, we could not decrease the values of $R_O$ 

923: and $E_K$ below the values for run D8 while keeping the boundary layer 

924: well resolved. In geodynamo simulations, a similar effect was 

925: reported, and it was noted that the behavior of the dipole moment was 

926: controlled by the amplitude of the Rossby number, independently of the 

927: values of the Ekman and Rayleigh numbers \cite{Christensen06}.

928: 

929: Figure \ref{fig:trend2} shows the trace of the dipole moment $\vmu$ 

930: on the surface of the unit sphere, its amplitude, and the cosine of the 

931: angle between $\vmu$ and $\vOmega$ for runs D3, D5, and D7. While in Run 

932: D3 the trace of $\vmu$ fills the entire surface of the sphere, as 

933: $R_O$ and $E_K$ are decreased $\vmu$ seems to fluctuate around two 

934: regions in opposite sides of the sphere. These regions get more localized 

935: with decreasing $R_O$ and $E_K$. Also, the time between excursions of 

936: $\vmu$ outside these regions gets larger, as shown by $\cos(\gamma)$. 

937: In Run D7, after a transient that finishes at $t \approx 80$, 

938: $\cos(\gamma)$ stays at $1$ or $-1$ for $\approx 20$ turnover times 

939: before changing sign rapidly. It is also worth noticing that as 

940: $R_O$ and $E_K$ decrease, the time it takes the system to develop a 

941: dipole moment of order one gets larger (see the evolution of $|\vmu|$ at 

942: early times in Figure \ref{fig:trend2}). This is also observed in the 

943: time evolution of the magnetic energy (see Figure \ref{fig:intermittency}). 

944: Instead of having an exponential growth of $E_M$ at early times as in Runs 

945: D1 to D5, Runs D6 to D8 show a more intermittent behavior: magnetic energy 

946: grows in rapid ``bursts'' and stays around that value until a new burst 

947: increases the magnetic energy again. In Run D7, this process saturates 

948: around $t\approx 80$ and no further change in the average value of $E_M$ 

949: is observed. It is possible that the slow-down during the kinematic 

950: dynamo regime is a consequence of a quasi-two-dimensionalization of the 

951: flow by the high rotation rate; this remains to be investigated fully.

952: 

953: \begin{figure}

954: \begin{center}\includegraphics[width=15cm]{fig15.eps}

955: \end{center}

956: \caption{Kinetic energy density (left), velocity field lines (middle), and 

957: view from top of the kinetic energy density superposed with velocity 

958: field lines (right) in run D10, before magnetic energy is introduced. Note 

959: the columnar structures in the velocity field aligned along $\vOmega$ 

960: (in the $z$ direction).}

961: \label{fig:columns}

962: \end{figure}

963: 

964: Figure \ref{fig:D5D7} shows visualizations of energy densities and 

965: field lines at late times in Runs D5 and D7. The development of 

966: anisotropies in the velocity and magnetic fields can be observed 

967: in run D7, which has the highest rotation rate attained of $\Omega=16$. 

968: Indeed, as $\Omega$ is increased the velocity field shows a tendency 

969: to develop columns, with mechanical energy concentrated in cylindrical 

970: structures aligned along $\vOmega$ and with a larger component of $v_\phi$ 

971: than of $v_z$. The velocity field in these column is helical, although in 

972: general the total mechanical helicity of the flow fluctuates in time 

973: around zero. These structures are observed before the magnetic field is 

974: introduced (although they persist as the magnetic energy grows) and seem 

975: to be the result of the Taylor-Proudman effect (see e.g. 

976: \cite{Taylor23,Davies72}). It is a trend observed through runs D5 to D10 

977: (see Figure \ref{fig:columns} for an example).

978: 

979: Run D9 experiments with lowering the angle between the forcing's axis 

980: of symmetry and the axis of rotation to $\alpha = 20^\circ$, and the 

981: dipole becomes more difficult to excite (as evidenced by a smaller growth 

982: rate in the kinematic dynamo regime). Indeed, dynamo action could not be 

983: excited below $15^\circ$ for the values of Reynolds numbers explored, 

984: as the resulting driven flow approaches axisymmetry.

985: 

986: \begin{figure}

987: \begin{center}\includegraphics[width=7.2cm]{fig16a.eps}

988:               \includegraphics[width=7.2cm]{fig16b.eps}

989: \end{center}

990: \caption{(a) Trace of the dipole moment on the surface of the unit sphere 

991: (above) and amplitude of the dipole moment as a function of time (below) 

992: for Run D11. (b) Same quantities for run D12.}

993: \label{fig:diphel}

994: \end{figure}

995: 

996: Run D10 is actually part of a set of experiments (Runs D10 to D12; again, 

997: see Table \ref{table:dynamo}) with forcing functions that are 

998: non-axisymmetric, and which may also inject mechanical helicity (Runs 

999: D11 and D12). Run D10, having no net mechanical helicity, shows no 

1000: big differences in the evolution of global quantities from the previously 

1001: discussed runs. The dipole develops but its orientation wanders randomly, 

1002: with some preferred orientation perpendicular to $\vOmega$. In this case, 

1003: axisymmetry is broken by the forcing directly instead of by a non-zero 

1004: angle between an axisymmetry forcing and the axis of rotation.

1005: 

1006: Runs D11 and D12 have non-axisymmetric forcing that injects mechanical 

1007: helicity. The forcing for these runs is given by equation 

1008: (\ref{eq:force2}) with coefficients

1009: \begin{equation}

1010: \xi^f_{3,3,0} = 5 \xi^f_{-3,3,0} = F_0 , \;\; 

1011: \xi^f_{3,3,0<m\le 3} = 5 \xi^f_{-3,3,0<m\le 3} = F_0 (1+i) ,

1012: \end{equation}

1013: with $F_0 = 1.7$. In the presence of net helicity, dynamo excitation 

1014: suddenly becomes much easier (as evidenced by a much larger growth rate 

1015: of magnetic energy during the kinematic regime), and the ultimate 

1016: saturation occurs at $E_M/E_V \approx 2$: more magnetic than kinetic, 

1017: with magnetic helicity, having sign opposite that of the mechanical 

1018: helicity inversely cascading to the large scales. As a result, the system 

1019: is dominated by a helical magnetic field at the largest available scale. 

1020: 

1021: An interesting qualitative argument from mean field theory 

1022: \cite{Steenbeck66,Krause} (which assumes large scale separations and 

1023: often some form of periodic boundary conditions, as do all 

1024: ``$\alpha$-effect'' calculations) can be seen to anticipate this result 

1025: as follows. From mean field theory , the induction equation for the mean 

1026: magnetic field $\overline{\bf B}$ (assuming there is no mean flow 

1027: $\overline{\bf U}$) is 

1028: \begin{equation}

1029: \frac{\partial \overline{\bf B}}{\partial t} = 

1030:     \alpha \nabla \times \overline{\bf B} + \beta \nabla^2 \overline{\bf B} .

1031: \label{eq:mean}

1032: \end{equation}

1033: Here, $\alpha$ is proportional to minus the kinetic helicity of the 

1034: flow \cite{Steenbeck66,Pouquet76,Krause}, and $\beta$ is a turbulent 

1035: magnetic diffusivity. Dotting equation (\ref{eq:mean}) with the mean 

1036: vector potential $\overline{\bf A}$ (such as 

1037: $\overline{\bf B} = \nabla \times \overline{\bf A}$) and integrating 

1038: over volume, an equation for the evolution of the mean magnetic helicity 

1039: $\overline{H_M}$ is obtained,

1040: \begin{equation}

1041: \frac{d \overline{H_M}}{d t} = 

1042:     \alpha \overline{E_M} + \beta \nabla^2 \overline{H_J} ,

1043: \end{equation}

1044: where $\overline{E_M}$ is the mean magnetic energy and $\overline{H_J}$ 

1045: is the mean current helicity. As a result, if magnetic diffusion is 

1046: neglected, the dynamo process injects into the mean (large) scales 

1047: magnetic helicity of opposite sign than the kinetic helicity, and in 

1048: the small scales magnetic helicity of the same sign. This effect has 

1049: been observed before in numerical dynamo simulations with periodic 

1050: boundary conditions \cite{Brandenburg01,Gomez04}. The large scale magnetic 

1051: helicity then inverse-cascades to the largest available scale in the 

1052: system, while the small scale magnetic helicity is transferred to smaller 

1053: scales where it is dissipated \cite{Alexakis06}. As a result, at late 

1054: times the system is dominated by a large scale magnetic field with 

1055: magnetic helicity of opposite sign to that of the kinetic helicity 

1056: injected by the forcing.

1057: 

1058: \begin{figure}

1059: \begin{center}\includegraphics[width=15cm]{fig17.eps}

1060: \end{center}

1061: \caption{Kinetic energy density and velocity field lines (left), and 

1062: magnetic energy density and magnetic field lines (right) at the same 

1063: time in the saturated state of run D12. The axes are aligned as in 

1064: Figure \ref{fig:diphel}.}

1065: \label{fig:D123D}

1066: \end{figure}

1067: 

1068: As seen in Figure \ref{fig:diphel}, the dipolar orientation in D11 

1069: seems to have a preference for being perpendicular to $\vOmega$. For 

1070: run D12, whose forcing function differs in its orientation to $\vOmega$ 

1071: by $90^\circ$, the dipole orientation seems to remain in a single 

1072: hemisphere (see Figure \ref{fig:diphel}) and its tip precesses about 

1073: $\vOmega$. Figure \ref{fig:D123D} shows energy density and field lines 

1074: in the saturated steady state of run D12. The axes are aligned as in 

1075: Figure \ref{fig:diphel}. As is often the case with helical flows, the 

1076: geometry of the flow is more complex than in the non-helical runs. These 

1077: runs in which a net sign of mechanical helicity is sustained by the 

1078: mechanical forcing were continued for several thousands turnover times, 

1079: and no reversals of the dipole moment were observed. The dipole moment 

1080: seems to fluctuate around a preferred orientation with only short 

1081: excursions of the dipole to the opposite hemisphere.

1082: 

1083: \section{Discussion and future directions \label{sec:future}}

1084: By solving the mechanically-forced MHD equations inside a rotating 

1085: conducting spherical boundary, we have found that a bewilderingly 

1086: wide variety of dynamo behavior is possible for a magnetic Prandtl 

1087: number of unity. The behavior is sensitive to mechanical and magnetic 

1088: Reynolds numbers, to rotation rate, and more indirectly to the Ekman 

1089: number. It is also sensitive to the geometry and strength of 

1090: the forcing functions and their relation to the axis of rotation. 

1091: We have only begun to explore this multidimensional parameter space. 

1092: Note we have not made much effort to tailor the forcing functions we 

1093: have chosen to model what the mechanical flows in planetary cores or 

1094: stellar convective regions might be. Rather, we have been exploring 

1095: dynamo behavior in the abstract, and feel somewhat overwhelmed by the 

1096: variety of dynamo behavior that has been found. In this light, 

1097: our attempt should be consider as an extension of dynamo simulations 

1098: in periodic boundary conditions 

1099: \cite{Meneguzzi81,Brandenburg01,Gomez04,Brandenburg05,Ponty05,Mininni05b} 

1100: to consider the effect of boundaries and rotation, and not be compared 

1101: with explorations of the space of parameters in realistic geodynamo 

1102: simulations \cite{Glatzmaier95,Zhang00,Roberts01,Kono02,Christensen06}.

1103: 

1104: What has become clear is that the wholly spectral methods we are using, 

1105: while accurate, do not scale well into the parameter regimes of planetary 

1106: and astrophysical dynamos, which involve many orders of magnitude between 

1107: the largest length scales in the flows and the Ekman or dissipation scales 

1108: that also play a role in the process. This limitation afflicts all numerical 

1109: attempts to explore planetary and stellar dynamos, particularly in view of 

1110: the low magnetic Prandtl numbers that are expected to prevail there, 

1111: in simulations \cite{Ponty05,Mininni05b}, and in liquid-sodium experiments 

1112: \cite{Gailitis01,Steiglitz01,Petrelis03,Sisan03,Spence05}. But the rapid 

1113: multiplication of the number of terms to compute in the convolution sums 

1114: in equations (\ref{eq:CK1}) and (\ref{eq:CK2}) provides a rather 

1115: intractable limitation on efforts to use our code without modification 

1116: at higher Reynolds numbers than the few hundred we have explored here. 

1117: What seems to be called for is an exploration of the possibilities of 

1118: using fast transforms (for spherical harmonics 

1119: \cite{Driscoll94,Mohlenkamp99} and possibly for spherical Bessel functions 

1120: \cite{Sharafeddin92,Cree93})to turn the code into a pseudospectral one 

1121: in which the nonlinear terms are computed in configuration space rather 

1122: than spectral space, as is commonly done for rectangular periodic 

1123: boundary conditions \cite{Canuto} which would increase available 

1124: resolution by many orders of magnitude. Our future investigations will 

1125: explore this possibility.

1126: 

1127: We also intend to replace the conducting boundary by an insulating but 

1128: mechanically-impenetrable one that will permit protrusion of the magnetic 

1129: field into the vacuum region surrounding the shell (see e.g. \cite{Kono02}). 

1130: It may be that forcing the magnetic field lines to return to the interior 

1131: each time they get near the shell that is now playing a dynamical role in 

1132: what we are seeing would be different in the case of an insulating shell. 

1133: This raises some conceptual difficulties, since the problem of matching 

1134: MHD fields on to vacuum electromagnetic ones has not been thoroughly 

1135: solved. For example, one approximation that has been used has been to 

1136: match magnetic fields at 

1137: a spherical surface to magnetostatic ones outside, which involve a magnetic 

1138: field that is derivable from a scalar potential and for which there is no 

1139: electric field. However, Maxwell's equations tell us that the tangential 

1140: electric field must be continuous at any interface, and there is no reason 

1141: why this tangential electric field should vanish or even be ``small'' 

1142: immediately inside an insulating boundary of a conducting magnetofluid. 

1143: The magnetostatic approximation may be the best we can do, but it 

1144: would be desirable to have more justification for it than we presently 

1145: have.

1146: 

1147: Finally, we need to devote attention to the geometry and strength of 

1148: the forcing functions that are being employed and to study the effect of 

1149: different forcing functions in dynamo action. Mechanical processes, 

1150: convective and otherwise, are believed to power the dynamo in cases of 

1151: geophysical or astrophysical interest. In the simulations discussed, we 

1152: have no explanation, for example, why the ``columns'' or columnar vortices 

1153: aligned along $\vOmega$ form in the velocity field in runs toward the 

1154: end of Table \ref{table:dynamo}. It is clear the physical effect that 

1155: is triggering their formation is the rotation alone, perhaps through a 

1156: two-dimensionalization of the flow via the Taylor-Proudman effect 

1157: \cite{Taylor23,Davies72,Acheson}. This is a different process than the 

1158: conventional explanation for the formation of columns in planetary interiors, 

1159: which involves both thermal convection and rotation \cite{Busse70}, since 

1160: thermal convection is completely absent in our incompressible MHD 

1161: formulation.

1162: 

1163: \ack

1164: Computer time was provided by NCAR. NSF grants CMG-0327888 at NCAR, 

1165: ATM-0327533 at Dartmouth, and AST-0507760 at Cornell supported this work 

1166: in part and are gratefully acknowledged.

1167: 

1168: %\bibliography{ms}

1169: 

1170: \section*{References}

1171: \begin{thebibliography}{50}

1172: 

1173: \bibitem{Mininni06}

1174: P.~D. Mininni and D.~C. Montgomery.

1175: \newblock Magnetohydrodynamic activity inside a sphere.

1176: \newblock {\em Phys.\ Fluids}, 18:116602, 2006.

1177: 

1178: \bibitem{Kono02}

1179: M.~Kono and P.~H. Roberts.

1180: \newblock Recent geodynamo simulations and observations of the geomagnetic

1181:   field.

1182: \newblock {\em Rev.\ Geophys.}, 40:1--53, 2002.

1183: 

1184: \bibitem{Chandrasekhar57}

1185: S.~Chandrasekhar and P.~C. Kendall.

1186: \newblock On force-free magnetic fields.

1187: \newblock {\em Astrophys.\ J.}, 126:457--460, 1957.

1188: 

1189: \bibitem{Montgomery78}

1190: D.~C. Montgomery, L.~Turner, and G.~Vahala.

1191: \newblock Three-dimensional magnetohydrodynamic turbulence in cylindrical

1192:   geometry.

1193: \newblock {\em Phys.\ Fluids}, 21:757--764, 1978.

1194: 

1195: \bibitem{Turner83}

1196: L.~Turner.

1197: \newblock Statistical mechanics of a bounded, ideal magnetofluid.

1198: \newblock {\em Ann.\ Phys.\ (NY)}, 149:58--161, 1983.

1199: 

1200: \bibitem{Cantarella00}

1201: J.~Cantarella, D.~DeTurck, H.~Gluck, and M.~Teytel.

1202: \newblock The spectrum of the curl operator on spherically symmetric domains.

1203: \newblock {\em Phys.\ Plasmas}, 7:2766--2775, 2000.

1204: 

1205: \bibitem{Greenspan}

1206: H.~P. Greenspan.

1207: \newblock {\em The theory of rotating fluids}.

1208: \newblock Cambridge Univ.\ Press., Cambridge, 1968.

1209: 

1210: \bibitem{Acheson}

1211: D.~J. Acheson.

1212: \newblock {\em Elementary fluid dynamics}.

1213: \newblock Clarendon Press, Oxford, 1990.

1214: 

1215: \bibitem{Zhang00}

1216: K.~Zhang and G.~Schubert.

1217: \newblock Magnetohydrodynamics in rapidly rotating spherical systems.

1218: \newblock {\em Annu.\ Rev.\ Fluid Mech.}, 32:409--443, 2000.

1219: 

1220: \bibitem{Glatzmaier95}

1221: G.~A. Glatzmaier and P.~H. Roberts.

1222: \newblock A three-dimensional self-consistent computer simulation of a

1223:   geomagnetic field reversal.

1224: \newblock {\em Nature}, 377:203--209, 1995.

1225: 

1226: \bibitem{Glatzmaier96}

1227: G.~A. Glatzmaier and P.~H. Roberts.

1228: \newblock Rotation and magnetism of earth's inner core.

1229: \newblock {\em Science}, 274:1887--1891, 1996.

1230: 

1231: \bibitem{Christensen01}

1232: U.~R. Christensen, J.~Aubert, P.~Cardin, E.~Dormy, S.~Gibbons, G.~A.

1233:   Glatzmaier, E.~Grote, Y.~Honkura, C.~Jones, M.~Kono, M.~Matsushima,

1234:   A.~Sakuraba, F.~Takahashi, A.~Tilgner, J.~Wicht, and K.~Zhang.

1235: \newblock A numerical dynamo benchmark.

1236: \newblock {\em Phys.\ of the Earth and Plan.\ Int.}, 128:25--34, 2001.

1237: 

1238: \bibitem{Roberts01}

1239: P.~H. Roberts and G.~A. Glatzmaier.

1240: \newblock The geodynamo, past, present and future.

1241: \newblock {\em Geophys.\ Astrophys.\ Fluid Dyn.}, 94:47--84, 2001.

1242: 

1243: \bibitem{Moffatt}

1244: H.~K. Moffatt.

1245: \newblock {\em Magnetic field generation in electrically conducting fluids}.

1246: \newblock Cambridge Univ. Press, Cambridge, 1978.

1247: 

1248: \bibitem{Kress00}

1249: B.~T. Kress and D.~C. Montgomery.

1250: \newblock Pressure determinations for incompressible fluids and magnetofluids.

1251: \newblock {\em J.\ Plasma Phys.}, 64:371--377, 2000.

1252: 

1253: \bibitem{Gallavotti}

1254: G.~Gallavotti.

1255: \newblock {\em Foundations of fluid dynamics}.

1256: \newblock Springer Verlag, Berlin, 2002.

1257: \newblock pp. 83--87 ff.

1258: 

1259: \bibitem{Canuto}

1260: C.~Canuto, Y.~Hussaini, A.~Quarteroni, and T.~Zang.

1261: \newblock {\em Spectral methods in fluid dynamics}.

1262: \newblock Springer Verlag, New York, 1988.

1263: 

1264: \bibitem{Dormy98}

1265: E.~Dormy, P.~Cardin, and D.~Jault.

1266: \newblock {M}{H}{D} flow in a slightly differentially rotating spherical shell,

1267:   with conducting inner core, in a dipolar magnetic field.

1268: \newblock {\em Phys.\ of the Earth and Plan.\ Int.}, 160:15--30, 1998.

1269: 

1270: \bibitem{vapor}

1271: J.~Clyne and M.~Rast.

1272: \newblock A prototype discovery environment for analyzing and visualizing

1273:   terascale turbulent fluid flow simulations.

1274: \newblock In R.~F. Erbacher, J.~C. Roberts, M.~T. Grohn, and K.~Borner,

1275:   editors, {\em Visualization and data analysis 2005}, pages 284--294,

1276:   Bellingham, Wash., 2005. SPIE.

1277: \newblock http://www.vapor.ucar.edu.

1278: 

1279: \bibitem{Matthaeus80}

1280: W.~H. Matthaeus and D.~Montgomery.

1281: \newblock Selective decay hypothesis at high mechanical and magnetic reynolds

1282:   numbers.

1283: \newblock {\em Ann.\ N.\ Y.\ Acad.\ Sci.}, 357:203--222, 1980.

1284: 

1285: \bibitem{Ting86}

1286: A.~C. Ting, W.~H. Matthaeus, and D.~Montgomery.

1287: \newblock Turbulent relaxation processes in magnetohydrodynamics.

1288: \newblock {\em Phys.\ Fluids}, 29:3261--3274, 1986.

1289: 

1290: \bibitem{Kinney95}

1291: R.~Kinney, J.~C. McWilliams, and T.~Tajima.

1292: \newblock Coherent structures and turbulent cascades in two-dimensional

1293:   incompressible magnetohydrodynamic turbulence.

1294: \newblock {\em Phys.\ Plasmas}, 2:3623--3639, 1995.

1295: 

1296: \bibitem{Grappin83}

1297: R.~Grappin, A.~Pouquet, and J.~L\'eorat.

1298: \newblock Dependence on correlation of {M}{H}{D} turbulence spectra.

1299: \newblock {\em Astron.\ Astrophys.}, 126:51--56, 1983.

1300: 

1301: \bibitem{Pouquet86}

1302: A.~Pouquet, M.~Meneguzzi, , and U.~Frisch.

1303: \newblock The growth of correlations in {M}{H}{D} turbulence.

1304: \newblock {\em Phys.\ Rev.\ A}, 33:4266--4276, 1986.

1305: 

1306: \bibitem{Ghosh88}

1307: S.~Ghosh, W.~H. Matthaeus, , and D.~C. Montgomery.

1308: \newblock The evolution of cross helicity in driven/dissipative two-dimensional

1309:   magnetohydrodynamics.

1310: \newblock {\em Phys.\ Fluids}, 31:2171--2184, 1988.

1311: 

1312: \bibitem{Christensen06}

1313: U.~R. Christensen and J.~Aubert.

1314: \newblock Scaling properties of convection-driven dynamos in rotating spherical

1315:   shells.

1316: \newblock {\em Geophys.\ J.\ Int.}, 166:97--114, 2006.

1317: 

1318: \bibitem{Meneguzzi81}

1319: M.~Meneguzzi, U.~Frisch, and A.~Pouquet.

1320: \newblock Helical and nonhelical turbulent dynamos.

1321: \newblock {\em Phys.\ Rev.\ Lett.}, 47:1060--1064, 1981.

1322: 

1323: \bibitem{Brandenburg01}

1324: A.~Brandenburg.

1325: \newblock The inverse cascade and nonlinear alpha-effect in simulations of

1326:   isotropic helical hydromagnetic turbulence.

1327: \newblock {\em Astrophys.\ J.}, 550:824--840, 2001.

1328: 

1329: \bibitem{Ponty05}

1330: Y.~Ponty, P.~D. Mininni, D.~C. Montgomery, J.-F. Pinton, H.~Politano, and

1331:   A.~Pouquet.

1332: \newblock Numerical study of dynamo action at low magnetic prandtl numbers.

1333: \newblock {\em Phys.\ Rev.\ Lett.}, 94:164502, 2005.

1334: 

1335: \bibitem{Mininni05b}

1336: P.~D. Mininni and D.~C. Montgomery.

1337: \newblock Low magnetic prandtl number dynamos with helical forcing.

1338: \newblock {\em Phys.\ Rev.\ E}, 72:056320, 2005.

1339: 

1340: \bibitem{Gomez04}

1341: D.~O. G\'omez and P.~D. Mininni.

1342: \newblock Direct numerical simulations of helical dynamo action: {M}{H}{D} and

1343:   beyond.

1344: \newblock {\em Nonlin.\ Proc.\ Geophys.}, 11:619--629, Dic 2004.

1345: 

1346: \bibitem{Mininni05a}

1347: P.~D. Mininni, D.~O. G\'omez, and S.~M. Mahajan.

1348: \newblock Direct simulations of helical {H}all-{M}{H}{D} turbulence and dynamo

1349:   action.

1350: \newblock {\em ApJ}, 619:1019--1027, 2005.

1351: 

1352: \bibitem{Valet05}

1353: J.-P. Valet, L.~Meynadier, and Y.~Guyodo.

1354: \newblock Geomagnetic dipole strength and reversal rate over the past two

1355:   million years.

1356: \newblock {\em Nature}, 435:802--805, 2005.

1357: 

1358: \bibitem{Taylor23}

1359: G.~I. Taylor.

1360: \newblock Experiments on the motion of solid bodies in rotating fluids.

1361: \newblock {\em Proc.\ Roy.\ Soc.\ A}, 104:213--218, 1923.

1362: 

1363: \bibitem{Davies72}

1364: P.~A. Davies.

1365: \newblock Experiments on taylor columns in rotating stratified flows.

1366: \newblock {\em J.\ Fluid Mech.}, 54:691--717, 1972.

1367: 

1368: \bibitem{Steenbeck66}

1369: M.~Steenbeck, F.~Krause, and K.-H. R{\"a}dler.

1370: \newblock Berechnung der mittleren {L}orentz-{F}eldstaerke $\overline{\bf v

1371:   \times b}$ fuer ein elektrisch leitendendes {M}edium in turbulenter, durch

1372:   {C}oriolis-{K}raefte beeinflu\ss ter {B}ewegung.

1373: \newblock {\em Z.\ Naturforsch.}, 21a:369--376, 1966.

1374: 

1375: \bibitem{Krause}

1376: F.~Krause and K.-H. Raedler.

1377: \newblock {\em Mean-field magnetohydrodynamics and dynamo theory}.

1378: \newblock Pergamon Press, New York, 1980.

1379: 

1380: \bibitem{Pouquet76}

1381: A.~Pouquet, U.~Frisch, and J.~L\'eorat.

1382: \newblock Strong {M}{H}{D} helical turbulence and the nonlinear dynamo effect.

1383: \newblock {\em J.\ Fluid Mech.}, 77:321--354, 1976.

1384: 

1385: \bibitem{Alexakis06}

1386: A.~Alexakis, P.~D. Mininni, and A.~Pouquet.

1387: \newblock On the inverse cascade of magnetic helicity.

1388: \newblock {\em Astrophys.\ J.}, 640:335--343, 2006.

1389: 

1390: \bibitem{Brandenburg05}

1391: A.~Brandenburg and K.~Subramanian.

1392: \newblock Astrophysical magnetic fields and nonlinear dynamo theory.

1393: \newblock {\em Phys.\ Rep.}, 417:1--209, 2005.

1394: 

1395: \bibitem{Gailitis01}

1396: A.~Gailitis, O.~Lielausis, E.~Platacis, S.~Dement'ev, A.~Cifersons, G.~Gerbeth,

1397:   T.~Gundrum, F.~Stefani, M.~Christen, and G.~Will.

1398: \newblock Magnetic field saturation in the riga dynamo experiment.

1399: \newblock {\em Phys.\ Rev.\ Lett.}, 86:003024, 2001.

1400: 

1401: \bibitem{Steiglitz01}

1402: R.~Steiglitz and U.~M\"uller.

1403: \newblock Experimental demonstration of a homogeneous two-scale dynamo.

1404: \newblock {\em Phys.\ Fluids}, 13:561--564, 2001.

1405: 

1406: \bibitem{Petrelis03}

1407: F.~P\'etr\'elis, M.~Bourgoin, L.~Mari\'e, J.~Burguete, A.~Chiffaudel,

1408:   F.~Daviaud, S.~Fauve, P.~Odier, and J.-F. Pinton.

1409: \newblock Nonlinear magnetic induction by helical motion in a liquid sodium

1410:   turbulent flow.

1411: \newblock {\em Phys.\ Rev.\ Lett.}, 90:174501, 2003.

1412: 

1413: \bibitem{Sisan03}

1414: D.~R. Sisan, W.~L. Shew, and D.~P. Lathrop.

1415: \newblock Lorentz force effects in magneto-turbulence.

1416: \newblock {\em Phys.\ Earth Plan.\ Int.}, 135:137--159, 2003.

1417: 

1418: \bibitem{Spence05}

1419: E.~J. Spence, M.~D. Nornberg, C.~M. Jacobson, R.~D. Kendrick, and C.~B. Forest.

1420: \newblock Observation of a turbulence-induced large scale magnetic field.

1421: \newblock {\em Phys.\ Rev.\ Lett.}, 96:055002, 2006.

1422: 

1423: \bibitem{Driscoll94}

1424: J.~R. Driscoll and D.~M.~Healy Jr.

1425: \newblock Computing {F}ourier transforms and convolutions on the 2-sphere.

1426: \newblock {\em Adv.\ Applied Math.}, 15:202--250, 1994.

1427: 

1428: \bibitem{Mohlenkamp99}

1429: M.~Mohlenkamp.

1430: \newblock A fast transform for spherical harmonics.

1431: \newblock {\em J.\ of Fourier Analysis and Appl.}, 5:159--184, 1999.

1432: 

1433: \bibitem{Sharafeddin92}

1434: O.~A. Sharafeddin, H.~F. Bowen, D.~J. Kouri, and D.~K. Hoffman.

1435: \newblock Numerical evaluation of spherical {B}essel transforms via fast

1436:   {F}ourier transforms.

1437: \newblock {\em J.\ Comp.\ Phys.}, 100:294--296, 1992.

1438: 

1439: \bibitem{Cree93}

1440: M.~J. Cree and P.~J. Bones.

1441: \newblock Algorithms to numerically evaluate the {H}ankel transform.

1442: \newblock {\em Computers Math.\ Applic.}, 26:1--12, 1993.

1443: 

1444: \bibitem{Busse70}

1445: F.~H. Busse.

1446: \newblock Thermal instabilities in rapidly rotating systems.

1447: \newblock {\em J.\ Fluid Mech.}, 44:441--460, 1970.

1448: 

1449: \end{thebibliography}

1450: 

1451: \end{document}

1452: 

1453: