1:
2: \documentclass[12pt, preprint]{aastex}
3:
4: %\documentclass[preprint2,longabstract]{aastex}
5:
6: %\documentclass[preprint2]{aastex}
7:
8: \newcommand{\vdag}{(v)^\dagger}
9:
10: \newcommand{\myemail}{Andria.Rogava@wis.kuleuven.be}
11:
12: \slugcomment{Astrophysical Journal, August 1, \textbf{664},
13: pp.??-??}
14:
15: \shorttitle{Over-reflection and generation of Gravito-Alfv\'en
16: waves}
17:
18: \shortauthors{Rogava et al.}
19:
20: \begin{document}
21:
22: \title{On over-reflection and generation of Gravito-Alfv\'en waves
23: in solar-type stars}
24:
25: \author{Andria Rogava\altaffilmark{1}}
26: \affil{Centre for Plasma Astrophysics, K.U.Leuven, Celestijnenlaan
27: 200B, 3001 Leuven, Belgium; Abdus Salam International Centre for
28: Theoretical Physics, Trieste I-34014, Italy.}
29: \email{Andria.Rogava@wis.kuleuven.be}
30:
31: \author{Grigol Gogoberidze\altaffilmark{1}}
32: \affil{Centre for Plasma Astrophysics, K.U.Leuven, Celestijnenlaan
33: 200B, 3001 Leuven, Belgium.}
34: \email{Grigol.Gogoberidze@wis.kuleuven.be }
35:
36: \author{Stefaan Poedts}
37: \affil{Centre for Plasma Astrophysics, K.U.Leuven, Celestijnenlaan
38: 200B, 3001 Leuven, Belgium.}
39: \email{Stefaan.Poedts@wis.kuleuven.be}
40:
41: \altaffiltext{1}{On leave from the Georgian National Astrophysical
42: Observatory, Kazbegi ave. $2^a$ Tbilisi-0160, Georgia.}
43:
44: \begin{abstract}
45: The dynamics of linear perturbations is studied in magnetized
46: plasma shear flows with a constant shearing rate and with
47: gravity-induced stratification. The general set of linearized
48: equations is derived and the two-dimensional case is considered in
49: detail. The \emph{Boussinesq approximation} is used in order to
50: examine relatively small-scale perturbations of low-frequency
51: modes: \emph{Gravito-Alfv\'en waves} (GAW) and \emph{Entropy Mode}
52: (EM) perturbations. It is shown that for flows with arbitrary
53: shearing rate there exists a finite time interval of
54: \emph{non-adiabatic} evolution of the perturbations. The
55: non-adiabatic behavior manifests itself in a twofold way, viz.\ by
56: the over-reflection of the GAWs and by the generation of GAWs from
57: EM perturbations. It is shown that these phenomena act as
58: efficient transformers of the equilibrium flow energy into the
59: energy of the perturbations for moderate and high shearing rate
60: solar plasma flows. Efficient generation of GAW by EM takes place
61: for shearing rates about an order of magnitude smaller than
62: necessary for development of a shear instability. The latter fact
63: could have important consequences for the problem of angular
64: momentum redistribution within the Sun and solar-type stars.
65: \end{abstract}
66:
67: \keywords{MHD --- Sun: magnetic fields --- Sun: rotation --- waves
68: --- stars: rotation}
69:
70: %----------------------------------------------------------------------
71: \section{Introduction}
72: %----------------------------------------------------------------------
73:
74: It is well-known that the most advanced stellar models, taking
75: into account rotation-induced hydrodynamic processes of meridional
76: circulation and shear mixing, are quite adequately describing the
77: structure of massive stars \citep{mm00}. For solar-type stars with
78: relatively wide convective regions, however, these models predict
79: \emph{large} rotation gradients and are in a notable contradiction
80: with recent observational helioseismology data \citep{ct05}. This
81: is probably due to the insufficiently high efficiency of the
82: common hydrodynamic instabilities that are characteristic to these
83: models: these instabilities can redistribute the angular momentum
84: between the different layers of the radiative region, but they are
85: \emph{not} strong enough to ensure the onset of a uniform rotation
86: regime.
87:
88: While the radiative interiors of solar-type stars rotate quite
89: uniformly, their convective zones rotate differentially. Mounting
90: observational evidence indicates that the radiative interior of
91: the Sun, for instance, rotates almost as a solid body, with a
92: quasi-flat seismic rotation profile \citep{c99,c03}. But the solar
93: convective zone exhibits a strong shearing in the latitudinal
94: direction, with its equatorial layers rotating about 1/3 faster
95: than the polar regions \citep{km03}. The transition between the
96: differential and the uniform rotation regimes takes place in a
97: relatively thin layer (with thickness $\le 0.05 R_{\bigodot}$),
98: called the \emph{tachocline}, where the radial shearing of the
99: rotation is particularly significant \citep{m05,p03}. It is often
100: argued that some kind of shear instability must be the source of
101: gravity waves at the base of the solar convection zone
102: \citep{k99}, either in the immediate vicinity or just inside the
103: tachocline. This instability can hardly be of a purely
104: hydrodynamical nature, because from various numerical studies of
105: the buoyant rise of magnetic flux from the bottom of the
106: convection zone to the surface of the sun it follows that the
107: tachocline is strongly magnetized with its poloidal magnetic field
108: about $10^4-10^5\;$G \citep{m05}.
109:
110: From this perspective it seems quite plausible to surmise that on
111: the time-scale of the order of the solar age or less not only
112: purely hydrodynamic internal gravity waves (IGWs) participate in
113: the process of the angular momentum redistribution within the
114: solar interior, but also, or rather, Gravito-MHD waves. This idea
115: is \emph{not} new. In the literature it has been argued
116: \citep{z97,kq97} that Gravito-MHD waves could be one of the most
117: promising candidates for the onset of the redistribution
118: processes. It was argued that a self-consistent model should
119: comprise a large-scale magnetic field in the Sun's interior, and
120: an accurate consideration of the Coriolis effects in the
121: convection zone and in the tachocline
122: \citep{tc98,tc03,tc04,tc05,ct05}. It was also argued that
123: turbulent stresses in the convection zone, via the agency of
124: Coriolis effects, induce a meridional circulation, causing the gas
125: from the convection zone to burrow downwards and generate both
126: horizontal and vertical velocity shear characterizing the
127: tachocline. However, the interior magnetic field confines the
128: burrowing and, hence, smooths and diminishes the shear, as
129: demanded by the observed structure of the tachocline. According to
130: \citet{ct05}, a decisive test of this \emph{qualitative} scenario,
131: after numerical refinements, would be its \emph{quantitative}
132: consistency with the observed interior angular velocity structure.
133:
134: %In another recent study, \citet{p06} considered the vertical
135: %propagation of magneto-acoustic gravity waves (MAGWs) in the
136: %approximation of a horizontal potential magnetic field. This study
137: %was based on a plane-parallel isothermal model for the solar
138: %atmosphere with gravity-induced stratification. It found that the
139: %cutoff frequency for the MAGWs, below which they can not
140: %propagate, does not depend on the magnetic field strength and is
141: %equal to the cutoff frequency for hydrodynamic internal gravity
142: %waves, viz.\ the so-called \emph{Lamb frequency}. It was also
143: %showed that the cutoff frequency is unaffected by the linear
144: %interaction between counter-propagating MAGWs, resulting from a
145: %nonuniform height distribution of the Alfv\'en velocity. The
146: %latter, in turn, leads to the reflection of propagating waves at
147: %relatively large heights.
148:
149: The aim of the present paper is to present the results of a
150: detailed study of the linear dynamics of perturbations in
151: gravitationally stratified magnetized shear flows. Bearing in mind
152: the highly probable importance of the Gravito-MHD waves for the
153: structure of the Sun and solar-type stars, we decided to pose and
154: study the following three interrelated issues: 1)~how the presence
155: of the nonuniform velocity field affects the propagation of the
156: waves through the stellar plasma; 2)~what kind of energy exchange
157: processes between the different collective modes and between the
158: modes and the ambient flow may happen; and 3)~what other
159: astrophysical consequences these processes could have.
160:
161: Originally, we develop the three-dimensional (3D) model, allowing
162: the ambient flow to have velocity shearing in both transversal
163: (prior to the gravity field) directions. Further on, the study is
164: focused by a number of simplifying assumptions, enabling the
165: solution of the equations. First of all, we consider only two
166: spatial dimensions (2D) bearing in mind that it is quite
167: straightforward to extend the analysis to the fully
168: three-dimensional case. Second, we assume a constant steady-state
169: (or stationary) shear flow, an assumption that allows us to use
170: the \emph{shearing sheet approximation} \citep{gl65}. And third,
171: we use the \emph{Boussinesq approximation}, and hence we study the
172: dynamics of relatively small-scale\footnote{with the vertical
173: wavelength, $\ell_z$, much smaller then stratification length
174: scale $z_0$.} perturbations of the low-frequency modes, viz.\ the
175: Gravito-Alfv\'en waves (GAW) and the Entropy Mode (EM)
176: perturbations.
177:
178: Our study leads to the conclusion that for \emph{arbitrary} shear
179: in the flow, there exists a finite interval of time in which the
180: evolution of the modes is highly non-adiabatic. Outside this
181: interval, the evolution is adiabatic and can be accurately
182: described by a WKB approximation. Therefore, we have an
183: asymptotic problem, quite common for various quantum-mechanical
184: applications, where one needs to obtain connection formulae in
185: \emph{turning point problems} \citep{ll77}. The relevant analysis
186: is done in this paper and it is found that the non-adiabatic
187: behavior of the considered modes manifests itself in the form of
188: two phenomena, viz.\ the \emph{over-reflection} of the GAW and the
189: \emph{generation} of the GAW by the EM perturbations. We show that
190: for flows with moderate and high shearing rates both these
191: processes are very effective converters of the equilibrium flow
192: energy into the energy of the waves.
193:
194: In the context of above-discussed astrophysical problems our
195: results imply that efficient generation of the GAW by the EM
196: perturbations takes place for shearing rates of about an order of
197: magnitude smaller then necessary for development of shear
198: instability.
199:
200: The remainder of the paper is organized in the following way: the
201: general mathematical and heuristic formalism is presented in the
202: Sec.~2. The over-reflection of GAW is studied in the Sec.~3; and
203: the second non-adiabatic process, viz.\ the generation of the GAW
204: by the EM perturbations, is studied in the Sec.~4. Finally, in
205: Sec.~5 we conclude with a summary and a brief discussion of our
206: results in the context of their possible importance for the
207: angular momentum redistribution within solar-type stars and the
208: solution of the problem of the uniform rotation of their radiative
209: zones.
210:
211: %----------------------------------------------------------------------
212: \section{Linear perturbations of sheared plasma flows}
213: %----------------------------------------------------------------------
214:
215: For the sake of generality, we derive the basic set of equations
216: for the case of 3D perturbations in a compressible,
217: gravitationally stratified MHD fluid. This formalism will be
218: presented in the next (2.1) subsection. Later on (subsection 2.2),
219: we will restrict the consideration to the somewhat simpler case of
220: 2D perturbations in an incompressible medium. Besides, the
221: \emph{Boussinesq approximation} will be used and the stationary
222: shear flow will be considered. %The study of the fully
223: %3D, compressible case will be presented in the subsequent
224: %publication \citep{r06}.
225:
226: %----------------------------------------------------------------------
227: \subsection{General theory}
228: %----------------------------------------------------------------------
229:
230: In our model, the geometry of the considered problem is simplified
231: in the following way: the equilibrium flow ${\bf {U}}$ is supposed
232: to be plane-parallel, to be directed along the $x$-axis, and to
233: have both a horizontal ($A_y$) and a vertical ($A_z$) shear:
234: $$
235: {\bf {U}} \equiv \left[ A_y y + A_z z, 0, 0 \right]. \eqno(1)
236: $$
237:
238: The uniform gravity ${\bf g}$ is assumed to be directed along the
239: negative direction of the $z$-axis:
240: $$
241: {\bf g} \equiv \left[ 0, 0, -g \right]. \eqno(2)
242: $$
243:
244: We consider a simplified model and assume that the equilibrium
245: magnetic field ${\bf B_0}$ is toroidal, parallel to $\bf {U}$ and
246: that it is possessing the gravity-induced vertical stratification:
247: $$
248: {\bf B_0} \equiv \left[ B(z),~0,~0~ \right]. \eqno(3)
249: $$
250:
251: The set of equations of one-fluid ideal magnetohydrodynamics
252: (MHD), governing the physics of the flow is:
253: $$
254: D_t \rho + \rho (\nabla \cdot \bf V)=0, \eqno(4)
255: $$
256: $$
257: \rho D_t{\bf V}= -{\nabla}{\left(P+{{B^2}\over{8{\pi}}}\right)}
258: +{{1}\over{4{\pi}}}({\bf B}\cdot{\nabla}){\bf B}+ \rho {\bf g},
259: \eqno(5)
260: $$
261: $$
262: D_t \bf B= (\bf B \cdot \nabla){\bf V} - \bf B (\nabla \cdot \bf
263: V), \eqno(6)
264: $$
265: $$
266: D_t S=0, \eqno(7)
267: $$
268: while
269: $$
270: \nabla \cdot \bf B = 0. \eqno(8)
271: $$
272:
273: In (4--7) the total (convective) time derivative operator is
274: denoted by: $D_t \equiv \partial_t + ({\bf V} \cdot \nabla)$.
275:
276: Instantaneous values of all physical variables are decomposed into
277: their regular (mean) and perturbational components:
278: $$
279: \rho \equiv \rho_0 + \varrho, \eqno(9)
280: $$
281: $$
282: P \equiv P_0 + p, \eqno(10)
283: $$
284: $$
285: S \equiv S_0 + s, \eqno(11c)
286: $$
287: $$
288: {\bf V} \equiv {\bf U} + {\bf u}, \eqno(12a)
289: $$
290: $$
291: {\bf B} \equiv {\bf B}_0 + {\bf b}. \eqno(12b)
292: $$
293:
294: It is straightforward to see that the horizontal background flow
295: does not affect the vertical hydromagnetic equilibrium of the
296: flowing MHD fluid, governed by the equation:
297: $$
298: \partial_z{\left(P_0 + {{B_0^2}\over{8\pi}} \right)} =-\rho_0g.
299: \eqno(13)
300: $$
301:
302: In general the variation of the density $d \rho$ is related with
303: variations of the pressure $dP$ and entropy $ds$ via the relation:
304: $$
305: d \rho = \mu ds + (1/C_s^2)dp, \eqno(14)
306: $$
307: where:
308: $$
309: \mu \equiv (\partial \rho/\partial s)_p, \eqno(15)
310: $$
311: $$
312: C_s^{2} \equiv (\partial P/\partial \rho)_s. \eqno(16)
313: $$
314:
315: Linearized set of equations governing the evolution of
316: perturbations within this flow can be written in the following way
317: [${\cal D}_t \equiv \partial_t + (A_y y + A_z z)\partial_x$]:
318: $$
319: {\cal D}_t \varrho + (\partial_z \rho_0) v_z + \rho_0(\nabla \cdot
320: {\bf u}) = 0, \eqno(17)
321: $$
322: $$
323: {\cal D}_t u_x = - {\cal A}u_y - Au_z
324: -{{1}\over{\rho_0}}\partial_xp
325: +{\left[{{(\partial_zB_0)}\over{4\pi}\rho_0}\right]}b_z, \eqno(18)
326: $$
327: $$
328: {\cal D}_t u_y = -{{1}\over{\rho_0}}\partial_yp
329: -{\left({{B_0}\over{4\pi
330: \rho_0}}\right)}(\partial_xb_y-\partial_yb_x), \eqno(19)
331: $$
332: $$
333: {\cal D}_t u_z = -{{1}\over{\rho_0}}\partial_zp +
334: {\left({{B_0}\over{4\pi}\rho_0}\right)}(\partial_xb_z-\partial_zb_x)
335: $$
336: $$
337: -{\left[{{(\partial_zB_0)}\over{4\pi}\rho_0}\right]}b_x
338: -{{g}\over{\rho_0}}\varrho, \eqno(20)
339: $$
340: $$
341: {\cal D}_t s + (\partial_z s) v_z = 0, \eqno(21)
342: $$
343: $$
344: {\cal D}_tb_y=B_0\partial_xu_y, \eqno(22)
345: $$
346: $$
347: {\cal D}_tb_z=B_0\partial_xu_z, \eqno(23)
348: $$
349: $$
350: \partial_xb_x + \partial_yb_y + \partial_zb_z = 0. \eqno(24)
351: $$
352: Due to the stratification some coefficients on the right hand
353: sides of Eqs. (17-23) depend on $z$. But when studying the
354: dynamics of small-scale perturbations (with $\ell_z \ll z_0$) one
355: can consider the mean components as constant. In this case the set
356: of first-order, partial differential equations (17-24) can be
357: reduced to the set of ordinary differential equations (ODEs) with
358: time-dependent coefficients if we look for solutions in the
359: following form:
360: $$
361: F({\bf r},t) \equiv {\hat F}[{\bf k}(t),t]e^{i{\left[({\bf k}
362: \cdot {\bf r})- \varphi({\bf k},t) \right]}}, \eqno(25)
363: $$
364: with
365: $$ {\varphi}({\bf k},t) \equiv {\left( {\bf U} \cdot
366: {\int}{\bf k}(t)dt\right)}, \eqno(26)
367: $$
368: where ${\bf k}(t)$ is a time-dependent wavenumber vector,
369: determined by the following set of equations \citep{lpl84, cc86,
370: mr99}:
371: $$
372: {\bf k}^{(1)}+{\cal S}^T{\cdot}~{\bf k}=0, \eqno(27)
373: $$
374: with ${\cal S}^T$ being the transposed \emph{shear matrix}:
375: $$
376: {\cal S}{\equiv}{\left(\matrix{U_{x,x}&U_{x,y}&U_{x,z}\cr
377: U_{y,x}&U_{y,y}&U_{y,z}\cr
378: U_{z,x}&U_{z,y}&U_{z,z}\cr}\right)}, \eqno(28)
379: $$
380: which, in our case has only two nonzero components: $A_y$ and
381: $A_z$:
382: $$
383: {\cal S}{\equiv}{\left(\matrix{0&A_y&A_z\cr
384: 0&0&0\cr
385: 0&0&0\cr}\right)}. \eqno(29)
386: $$
387: Due to Eq.~(27), transversal components of the ${\bf k}(t)$
388: acquire linear time-dependence:
389: $$
390: k_y(t) = k_y(0) - A_ytk_x, \eqno(30a)
391: $$
392: $$ k_z(t) = k_z(0) - A_ztk_x. \eqno(30b)
393: $$
394:
395: Therefore, we see that the wave-number vector ${\bf k}(t)$ of the
396: Spatial Fourier Harmonics (SFH) varies in time, i.e.\ in the
397: linear approximation there is a ``drift" of the SFH in the {\bf
398: k}-space. The physical reason of this ``drift" is related to the
399: fact that in the sheared flow the perturbations cannot have the
400: form of a simple plane wave due to the effect of the shearing
401: background on the wave crests.
402:
403: Applying the ansatz (25) to Eqs.~(17-23), we reduce the system to
404: the following set of first-order ODEs [hereafter we are using the
405: following notation $f^{(n)} \equiv \partial_t^{n}f$]:
406: $$
407: d^{(1)}=i \varepsilon v_z + v_x + {\cal K}_y(\tau) v_y + {\cal
408: K}_z(\tau) v_z, \eqno(31)
409: $$
410: $$
411: v_x^{(1)}= -{\cal R} v_y - R v_z - \sigma^2{\cal P} +
412: i{{\varepsilon}\over{2}} b_z, \eqno(32)
413: $$
414: $$
415: v_y^{(1)}= - \sigma^2{\cal K}_y(\tau) {\cal P} + b_y - {\cal
416: K}_y(\tau) b_x, \eqno(33)
417: $$
418: $$
419: v_z^{(1)}= - \sigma^2{\cal K}_z(\tau) {\cal P} + b_z - {\cal
420: K}_z(\tau) b_x -
421: $$
422: $$
423: i{{\varepsilon}\over{2}} b_x + i \alpha d, \eqno(34)
424: $$
425: $$
426: e^{(1)} = v_z, \eqno(35)
427: $$
428: $$
429: b_y^{(1)} = - v_y, \eqno(36)
430: $$
431: $$
432: b_z^{(1)} = - v_z, \eqno(37)
433: $$
434: while the Maxwell equation (24) reduces to the following algebraic
435: relation:
436: $$
437: b_x + {\cal K}_y(\tau) b_y + {\cal K}_z(\tau) b_z = 0. \eqno(38)
438: $$
439:
440: In these equations, the following dimensionless notation
441: \footnote{The Alfv\'en speed $C_A^2 = B_0^2/4 \pi \rho_0$ and the
442: Brunt-V\"ais\"al\"a frequency $N_{BV}^2=-g\partial_z
443: \rho_0/\rho_0$ are defined in the usual way.} is used for the
444: constants: $R_y \equiv A_y/k_xC_A$, $R_z \equiv A_z/k_xC_A$,
445: ${\cal K}_y(0) \equiv k_y(0)/k_x$, ${\cal K}_z(0) \equiv
446: k_z(0)/k_x$, $\varepsilon \equiv (k_z z_0)^{-1}$, $\sigma^2 \equiv
447: (C_s/C_A)^2$, $N^2 \equiv (N_{BV}/C_Ak_x)^2$. The dimensionless
448: variables appearing in the above set of equations are defined as:
449: $v_j \equiv {\hat u}_j/C_A$, $b_j \equiv i{\hat b}_j/B_0$, $d
450: \equiv i{\hat \varrho}/\rho_0$, ${\cal P} \equiv i{\hat
451: p}/\rho_0C_s^2$, $e \equiv -k_x {\hat s}/(\partial_zs_0)$, $\tau
452: \equiv k_xC_At$, ${\cal K}_y(\tau) \equiv {\cal K}_y(0) - {\cal
453: R}\tau$, ${\cal K}_z(\tau) \equiv {\cal K}_z(0) - R\tau$.
454:
455: Notice that this system is \emph{not} closed, because we have only
456: \emph{eight} equations for \emph{nine} variables (${\bf v}, {\bf b},
457: d, e, \cal P$). The closure of the set of equations is ensured by
458: the thermodynamic relation that follows from Eq.~(14) [$\alpha
459: \equiv g/k_xC_s^2$]:
460: $$
461: d = {\cal P} - i (N^2/\alpha) e. \eqno(39)
462: $$
463:
464: This system of equations describes the temporal evolution of
465: Gravito-MHD waves modified by the presence of a velocity shear in
466: the two planes transversal to the flow direction. The full
467: analysis of this set of equations is beyond the scope of the
468: present paper. Instead, in the next sections, we focus our study
469: on the relatively simple 2D and incompressible case. We will see
470: that even in this simplified case the presence of the velocity
471: shear brings a considerable novelty in the dynamics of
472: perturbations.
473:
474: %----------------------------------------------------------------------
475: \subsection{The incompressible, 2D limit}
476: %----------------------------------------------------------------------
477:
478: In order to simplify the mathematical aspects of the problem, we
479: make the following assumptions:
480:
481: \begin{itemize}
482: \item We consider the 2D case, restricting the problem to the
483: $x0z$ plan and implying:
484: $$
485: {\cal K}_y(0) = R_y =0. \eqno(40)
486: $$
487: Hereafter we write $R \equiv R_z (A \equiv A_z)$ In this case,
488: Eq.~(38) enables us to express the longitudinal ($x$-)component of
489: the magnetic field perturbation in terms of its transversal
490: ($z$-)component:
491: $$
492: b_x = - {\cal K}_z(\tau) b_z. \eqno(41)
493: $$
494:
495: \item We drop the $d^{(1)}$ term in the continuity equation, i.e.\ we
496: adopt the concept of so-called {\it dynamic incompressibility}
497: \citep{ll59}. This implies that the velocity of the flow is assumed
498: to be small enough in comparison with the speed of sound. As a
499: consequence, the density perturbations evoked by the pressure
500: perturbations are negligible and instead of the Eq.~(39) we will
501: have:
502: $$
503: d \simeq - i (N^2/\alpha) e. \eqno(42)
504: $$
505:
506: However, since the medium is thermally conducting its density
507: should vary also due to the temperature variation, and this effect
508: should be (and is) taken into account.
509: \end{itemize}
510:
511: In hydrodynamics, the second assumption is known as the
512: \emph{Boussinesq approximation}: in the conservation equations the
513: terms proportional to the density perturbation are retained if,
514: and only if, they produce a buoyancy force \footnote{The last term
515: in Eq.~(34).}. The Boussinesq approximation automatically implies
516: \citep{ll59,g82} that the vertical length scale of perturbations,
517: $\ell_z \sim k_z^{-1}$, is much smaller then the characteristic
518: stratification length scale $z_0$. This condition was already used
519: for the derivation of Eqs. (31-37). In terms of our dimensionless
520: notation, this condition transcribes into:
521: $$
522: \varepsilon \ll 1. \eqno(43)
523: $$
524: When this condition is satisfied, the above approach enables us to
525: study the dynamics of small-scale, low-frequency perturbations.
526: \footnote{for detailed analysis of the Boussinesq approximation in
527: stratified shear flows see \citet{d97} and \citet{dkp04}.}
528:
529: These assumptions make the basic set of equations much simpler.
530: Taking into account (41) and (42), we get [with ${\cal K}^2(\tau)
531: \equiv 1 + {\cal K}_z^2(\tau)$]:
532: $$
533: v_x + {\cal K}_z(\tau) v_z=0, \eqno(44)
534: $$
535: $$
536: v_x^{(1)}= - R v_z - \sigma^2{\cal P}, \eqno(45)
537: $$
538: $$
539: v_z^{(1)}= - \sigma^2{\cal K}_z(\tau) {\cal P} + {\cal K}^2(\tau)
540: b_z + N^2 e, \eqno(46)
541: $$
542: $$
543: e^{(1)} = v_z, \eqno(47)
544: $$
545: $$
546: b_z^{(1)} = - v_z. \eqno(48)
547: $$
548:
549: From the last two equations it is apparent that there is an
550: algebraic relation between the entropy perturbations and the
551: magnetic field perturbations, viz.\
552: $$
553: e + b_z = {\cal I}=const(\tau). \eqno(49)
554: $$
555:
556: Another, less obvious, relation between the perturbed quantities
557: may be found if one takes time derivative of (44) and takes into
558: account (45) and (46). The result can be used to express the
559: pressure perturbation as a function of the other variables, i.e.\
560: $$
561: \sigma^2{\cal P}={{-2R}\over{{\cal K}^2(\tau)}}v_y+{\beta}b_y
562: -{{W^2{\beta}}\over{{\cal K}^2(\tau)}}e. \eqno(50)
563: $$
564:
565: Having three algebraic relations, Eqs.~(44), (49), and (50), for
566: five variables, we can reduce the system to {\it one} second-order
567: ODE. For instance, for the variable $b(\tau)$, defined as $b(\tau)
568: \equiv {\cal K}(\tau)b_z$, we can derive the following explicit,
569: second-order inhomogeneous ODE:
570: $$
571: \frac{{\rm d^2} b}{{\rm d} \tau^2} + \left[1 + \frac{ N^2}{{\cal
572: K}^2(\tau)} - \frac{R^2}{{\cal K}^4(\tau)}\right] b =
573: \frac{N^2}{{\cal K}(\tau)} {\cal I}. \eqno(51)
574: $$
575:
576: This equation describes the dynamics of the so-called
577: \emph{continuous eigenspectrum} (see, {\it e.g.} \cite{bm02} and
578: the references therein). In order to take into consideration the
579: perturbations belonging to the \emph{discrete eigenspectrum} one
580: should specify some boundary conditions with respect to the
581: vertical spatial variable $z$. In the absence of an external
582: magnetic field, it is known \citep{dr81} that all the discrete
583: eigenmodes are stable if the Richardson number, defined as $Ri
584: \equiv N_{BV}^2/A^2 = N^2/R^2$, is greater then $1/4$.
585:
586: Equation~(51) describes two important physical phenomena related
587: to the dynamics of the continuous eigenspectrum, which will be
588: considered in detail in the two subsequent sections, viz.\ (a)~the
589: (over)-reflection of the GAW and (b)~the conversion of the EM into
590: tghe GAW. It will be shown that the EM perturbations can be a very
591: efficient source of GAW, even for relatively small shearing rates
592: ($Ri > 1/4$).
593:
594: %----------------------------------------------------------------------
595: \section{Over-reflection of Gravito-Alfv\'en waves}
596: %----------------------------------------------------------------------
597:
598: In order to describe the phenomenon of the \emph{over-reflection
599: }of the GAW in its pure form, let us study Eq.~(51) with ${\cal
600: I}=0$, yielding:
601: $$
602: \frac{{\rm d^2} b}{{\rm d} \tau^2} + \left[1 + \frac{ N^2}{{\cal
603: K}^2(\tau)} - \frac{R^2}{{\cal K}^4(\tau)}\right] b = 0. \eqno(52)
604: $$
605:
606: The standard methods of the analysis of similar equations
607: \citep{o74} is well-known in quantum mechanics \citep{ll77}. The
608: behavior of the solution is fully determined by the shear-modified
609: frequency:
610: $$
611: \omega^2(\tau) \equiv 1 + \frac{ N^2}{{\cal K}^2(\tau)} -
612: \frac{R^2}{{\cal K}^4(\tau)}. \eqno(53)
613: $$
614:
615: From the mathematical point of view, the (over)-reflection is
616: caused by the cavity of $\omega^2(\tau)$ that appears in the
617: vicinity of the point $\tau^\ast$, where $K_z(\tau^\ast)=0$.
618:
619: \begin{figure}
620: \plotone{f1.eps} \caption{Dependence of the transmission
621: coefficient $D(R)$ on the normalized shear parameter $R$ for
622: different values of the Brunt-V\"ais\"al\"a frequency: $N = 0$
623: (solid line), $N = 1$ (dashed line) and $N = 2$ (dash-dotted
624: line).}
625: \end{figure}
626:
627: In the areas where $K_z(\tau) \gg R$, the evolution of $b(\tau)$
628: is adiabatic [${\rm d}\omega(\tau)/{\rm d} \tau \ll
629: \omega^2(\tau)$] and, therefore, the approximate solution has the
630: form
631: $$
632: b(\tau) \approx A_+ b_+(\tau) + A_- b_-(\tau), \eqno(54)
633: $$
634: where
635: $$
636: b_\pm(\tau) = \frac{1}{\sqrt{\omega(\tau)}} e^{\mp i \int
637: \omega(\tau) d \tau} \eqno(55)
638: $$
639: are WKB solutions having positive and negative phase velocities
640: along the $x$-axis, respectively, while the $A_\pm$ can be treated
641: as the intensities of the corresponding types of perturbations.
642:
643: Let us assume that $A^L_\pm $ and $A^R_\pm $ are the WKB
644: amplitudes far on the left and right hand side from the point
645: $\tau^\ast$, respectively. It is well-known \citep{o74} that the
646: conservation of the Wronskian leads to the following invariant:
647: $$
648: |A_+^L|^2 - |A_-^L|^2 = |A_+^R|^2 -
649: |A_-^R|^2, \eqno(56)
650: $$
651: which physically corresponds to the conservation of the wave
652: action (see, {e.g.}, \citet{g04} and references therein). If
653: initially $A_-^L \equiv 0$, the refraction and transmission
654: coefficients may be defined in the usual way:
655: $$
656: Re=\frac{|A_-^R|^2}{|A_+^L|^2},\quad D=\frac{|A_+^R|^2}{|A_+^L|^2},
657: \eqno(57)
658: $$
659: while from Eq.~(56) we obviously have:
660: $$
661: 1 + Re = D. \eqno(58)
662: $$
663:
664: The dependence of the transmission coefficient $D(R)$ on $R$,
665: obtained through the numerical solution of Eq.~(52) for different
666: values of the Brunt-V\"ais\"al\"a frequency, viz.\ $N = 0$ (solid
667: line), $N = 1$ (dashed line) and $N = 2$ (dash-dotted line), is
668: presented on Fig.~1. The initial conditions for the numerical
669: solution are chosen from the WKB solutions (55) far on the
670: left-hand side of the resonant area ($K_z(0) \gg 1,R$). The
671: analysis leads to the conclusion that the amplification of the
672: energy density of the perturbations is always finite, but it can
673: become arbitrarily large with a proper increase of the shearing
674: rate. It can be seen that the velocity shear can cause a strong
675: amplification of the GAW, even for moderate values of the
676: normalized shearing rates.
677:
678: According to the numerical study of Eq.~(52), the amplitude of the
679: reflected wave exceeds the amplitude of the incident wave if:
680: $$
681: R^2 > 2. \eqno(59)
682: $$
683: This phenomenon, originally discovered by Miles for acoustic waves
684: \citep{m57}, is called \emph{over-reflection}.
685:
686: Bearing in mind that the energy of the mode is proportional to the
687: square of its amplitude, we can easily surmise that in the course
688: of its non-adiabatic evolution the perturbation energy is
689: increasing. According to Eqs. (56) and (58) the ratio of the total
690: `post-reflection' energy to the initial one equals $Re+D>1$. The
691: energy increases at the expense of the mean (shear) flow energy.
692: Consequently, over-reflection represents a quite efficient
693: mechanism for transferring the mean flow energy to perturbations.
694:
695: In the limit $R \rightarrow 0 $, the reflection coefficient becomes
696: exponentially small with respect to the parameter $-1/R$, i.e., $Re
697: \sim \exp(-1/R)$ \citep{g04}.
698:
699: %----------------------------------------------------------------------
700: \section{The generation of Gravito-Alfv\'en waves}
701: %----------------------------------------------------------------------
702:
703: Let us return to the analysis of the full, inhomogeneous equation
704: (51). When the condition of the adiabatic evolution ${\rm
705: d}\omega(\tau)/{\rm d} \tau \ll \omega^2(\tau)$ holds, then the
706: approximate solution of Eq.~(51) has the following form \citep{c94}:
707: $$
708: b(\tau) \simeq A_+ b_+(\tau) + A_- b_-(\tau) + b_E(\tau), \eqno(60)
709: $$
710: where the non-periodic solution $b_E(\tau)$ is given by:
711: $$
712: b_E(\tau) \simeq \frac{N^2}{{\cal K}(\tau)} \left( 1 +
713: \frac{N^2}{{\cal K}^2(\tau)} - \frac{R^2}{{\cal K}^4(\tau)}
714: \right)^{-1}{\cal I}. \eqno(61)
715: $$
716: and can be readily identified as the shear-modified \emph{Entropy
717: Mode} (EM) perturbation.
718:
719: \begin{figure}
720: \plotone{f2.eps} \caption{Temporal evolution of $b(\tau)$ for $N =
721: 1$ and $R=0.7$.}
722: \end{figure}
723:
724: It turns out that another physically important phenomenon that
725: takes place in the system under consideration for moderate and
726: high shearing rates, is the generation of the GAW by the EM
727: perturbations. An illustrative example is given by the Fig.~2,
728: where the numerical solution of Eq.~(51) for the case when $N = 1$
729: and $R=0.7$ for the variable $b(\tau)$ is shown. The initial
730: conditions were chosen in such a way that at the beginning (at
731: $K_z(0) = 35$) there exists only the EM perturbation. Namely, the
732: initial conditions were found by means of Eqs.~(60) and (61) with
733: $A_\pm=0$ and ${\cal I}=1$). The figure clearly demonstrates the
734: generation of the GAW while the EM passes through the interval of
735: non-adiabatic evolution, situated in the vicinity of the point
736: ${\cal K}_z(\tau^\ast)=0$.
737:
738: For the quantitative description of the wave generation process,
739: analogously to the reflection and transmission coefficients
740: defined in the previous section, we define the \emph{wave
741: generation coefficient} $G_\pm$ in the following way:
742: $$
743: G_\pm =\frac{|A_\pm^L|^2}{|{\cal I}|^2 N^2}. \eqno(62)
744: $$
745:
746: With this definition, $G_\pm$ is proportional to the ratio of the
747: generated wave energy to the energy of incident entropy mode
748: perturbation. From the symmetry of the problem, it is obvious that
749: $G_+ = G_- \equiv G$.
750:
751: In Figure~3, the dependence of the generation coefficient $G(R)$
752: on the normalized shear parameter $R$ for different values of the
753: Brunt-V\"ais\"al\"a frequency: $N = 1$ (solid line), $N = 2$
754: (dashed line) and $N = 3$ (dash-dotted line) is presented.
755:
756: \begin{figure}
757: \plotone{f3.eps} \caption{Dependence of the generation coefficient
758: $G(R)$ on the normalized shear parameter $R$ for different values
759: of the Brunt-V\"ais\"al\"a frequency: $N = 1$ (solid line), $N =
760: 2$ (dashed line) and $N = 3$ (dash-dotted line).}
761: \end{figure}
762:
763: \begin{figure}
764: \plotone{f4.eps} \caption{Dependence of the generation coefficient
765: $ G(Ri)$ on Richardson number $Ri$ for $N = 4$.}
766: \end{figure}
767:
768: %----------------------------------------------------------------------
769: \section{Discussion}
770: %----------------------------------------------------------------------
771:
772: In order to explain the near uniformity of the rotation profile in
773: the radiative region of the Sun and similar regions in solar-type
774: stars usually two different mechanisms are being proposed. The
775: first one implies the presence of travelling IGWs generated at the
776: base of the convective region \citep{ct05}, while the second one
777: is based on the presumed (and for the Sun observationally
778: justified) existence of the magnetic field in the radiative zone
779: \citep{gm98}.
780:
781: It seems reasonable to assume that the above-mentioned two
782: mechanisms are not necessarily alternatives, but could be
783: envisaged as complementary ones. In particular, we may argue that
784: if magnetic fields are indeed present inside the radiative zone
785: and the Gravito-MHD waves do pass through the zone, it leads to
786: intense energy exchange processes between the waves and the
787: stellar plasma. This circumstance, in its turn, might lead to the
788: more efficient redistribution of the angular momentum within these
789: stars, establishing quasi-flat rotation profiles. The presence of
790: the magnetic field implies that it is more reasonable to speak not
791: about purely hydrodynamical IGWs, but about Gravito-MHD waves in
792: general, and Gravito-Alfv\'en waves in particular.
793:
794: The evolution of the latter mode, present only in the magnetized
795: case and absent in a purely hydrodynamic medium, is studied in the
796: present paper. The theoretical background for studying of the
797: evolution of the full set of Gravito-MHD waves is also developed.
798:
799: From the general theory of IGWs, we know \citep{dr81} that all the
800: discrete eigenmodes are stable under the condition that $Ri >
801: 1/4$. Our study shows that the generation of the GAW by the EM
802: perturbations could provide an efficient mechanism for creating
803: the required flux of the GAW, even for relatively low shearing
804: rates, when the discrete eigenmodes are stable. Note, in this
805: context, that for the case presented in Fig.~2, for instance, the
806: Richardson number was about $Ri \approx 1.02$.
807:
808: In the magnetized case, we studied the dependence of the
809: generation coefficient $G(Ri)$ on the Richardson number $Ri$ for
810: the relatively weak external magnetic field ($N = 4$). The results
811: of this study are presented in Fig.~4. The results clearly
812: indicate that even though the generation coefficient falls rapidly
813: with increasing $Ri$, the efficient generation of the GAW by the
814: EM perturbations still takes place for $Ri < 1.25$.
815:
816: Obviously, our study has a simplified nature and it only indicates
817: at the possibly important role of \emph{non-modal} phenomena
818: related with GAW in the redistribution of the angular momentum and
819: the flattening of the rotation curves in solar-type stars. For
820: making this claim more convincing, we have to consider the
821: compressible, three-dimensional analogue of this problem and we
822: have to verify whether the full set of Gravito-MHD waves is as
823: efficient (or even more efficient) in exchanging energy with the
824: equilibrium flow. We know from previous studies that in the
825: absence of the gravity-induced stratification among the
826: shear-modified MHD waves, the fast magnetosonic waves are the most
827: efficient ones in exchanging energy with the equilibrium flow
828: \citep{prm99,rpm00}. It is of great interest to check whether the
829: fast Gravito-magnetosonic waves have the same quality and to
830: determine what their contribution in the angular momentum
831: redistribution could be. The study of the linear dynamics of all
832: Gravito-MHD waves is currently initiated and the results will be
833: published in a subsequent paper.
834:
835: Finally, one has to remember that the non-modal approach that has
836: been applied in the present paper, provides no information about
837: the spatial aspects of the shear-induced processes, because the
838: study of the Spatial Fourier Harmonics is confined to the phase
839: space of the wave number vectors ${\bf k}(t)$. In order to have a
840: clear idea about the spatial appearance of these processes, one
841: has to study them numerically (similarly to the uniform,
842: non-stratified flow study made by \citet{b01}) and to check how
843: the non-modal phenomena couple with the traditional rotational
844: instabilities.
845:
846: %----------------------------------------------------------------------
847: \acknowledgments
848: %----------------------------------------------------------------------
849: These results were obtained in the framework of the projects GOA
850: 2004/01 (K.U.Leuven), G.0304.07 (FWO-Vlaanderen) and C90203 (ESA
851: Prodex 8). Andria Rogava wishes to thank the \emph{Katholieke
852: Universiteit Leuven} (Leuven, Belgium) and the \emph{Abdus Salam
853: International Centre for Theoretical Physics}(Trieste, Italy) for
854: supporting him, in part, through a Senior Postdoctoral Fellowship
855: and Senior Associate Membership Award, respectively. The research
856: of Andria Rogava and Grigol Gogoberidze was supported in part by
857: the Georgian National Science Foundation grant GNSF/ST06/4-096.
858: The research of Grigol Gogoberidze was supported in part by the
859: INTAS grant 06-1000017-9258.
860:
861: %----------------------------------------------------------------------
862: \begin{thebibliography}{}
863: \bibitem[Balmforth \& Morrison (2002)]{bm02} Balmforth, N. J., \&
864: Morrison, P. J. 2002, Large-Scale Atmosphere-Ocean Dynamics II, J.
865: Norbury \& I. Roulstone (Cambridge: Cambridge Univ. Press)
866: %\bibitem[Barnes et al. (1998)]{b98} Barnes, G., MacGregor, K. B., \& Charbonneau, P. 1998, \apj, 498,
867: %L169
868: \bibitem[Bodo et al. (2001)]{b01} Bodo, G., Poedts, S., Rogava, A. D., \& Rossi, P. 2001, \aap, 374,
869: 337
870: %\bibitem[Bretherton (1966)]{b66} Bretherton F. P. 1966, Q. J. R.
871: %Meteorol. Soc., 92, 466
872: %\bibitem[Booker \& Bretherton (1967)]{bb67} Booker, J. R., \& Bretherton, F. P. 1967,
873: %J. Fluid Mech., 27, 513
874: \bibitem[Chagelishvili, Rogava \& Segal (1994)]{c94} Chagelishvili, G. D., Rogava, A. D., \&
875: Segal I. N. 1994, \pre, 50, 4283
876: \bibitem[Chaplin et al. (1999)]{c99} Chaplin, W. J. et al. 1999, \mnras, 308, 405
877: \bibitem[Charbonnel \& Talon (2005)]{ct05} Charbonnel, C., \& Talon, S. 2005, Science, 309,
878: 2189
879: \bibitem[Couvidat et al. (2003)]{c03} Couvidat, S. et al. 2003, \apj, 597, L77
880: \bibitem[Craik \& Criminale (1986)]{cc86} Craik, A. D. D., \& Criminale, W. O. 1986, Proc. R.
881: Soc. Lond. A, 406, 13
882: \bibitem[Didebulidze (1997)]{d97} Didebulidze, G. G. 1997, Phys. Lett. A, 235, 65
883: \bibitem[Didebulidze et. al (2004)]{dkp04} Didebulidze, G. G., Kafkalidis, J. F., \& Pataraya, A. D.
884: 2004, J. Atmosph. Sol.-Terr. Physics, 66, 715
885: \bibitem[Drazin \& Reid (1981)]{dr81} Drazin, P. G., \& Reid, W. H. 1981, Hydrodynamic
886: Stability (Cambridge: Cambridge Univ. Press)
887: %\bibitem[Eggenberger et al. (2005)]{emm05} Eggenberger P., Maeder, A., \& Meynet G. 2005, \aap, 440,
888: %L9
889: \bibitem[Gill (1982)]{g82} Gill, A. E. 1982, Atmosphere-Ocean Dynamics (New York: Academic
890: Press)
891: \bibitem[Gogoberidze et al. (2004)]{g04} Gogoberidze, G., Chagelishvili, G. D., Sagdeev, R.
892: Z., \& Lominadze, J. G. 2004, Phys. Plasmas, 11, 4672
893: \bibitem[Goldreich \& Lynden-Bell (1965)]{gl65} Goldreich, P., \& Lynden-Bell, D. 1965, \mnras, 130,
894: 125
895: \bibitem[Gough \& McIntyre (1998)]{gm98} Gough, D. 0., \& McIntyre, M. E. 1998, \nat, 394,
896: %755
897: \bibitem[Kim \& MacGregor (2003)]{km03} Kim, E-j., \& MacGregor, K. B. 2003, \apj, 588,
898: 645
899: \bibitem[Kumar \& Quataert (1997)]{kq97} Kumar, P., \& Quataert, E.J. 1997, \apj, 475, L143
900: \bibitem[Kumar, Talon \& Zahn (1999)]{k99} Kumar, P., Talon, S., \& Zahn, J.-P. 1999, \apj, 520, 859
901: \bibitem[Lagnado et al. (1984)]{lpl84} Lagnado, R. R., Phan-Thien, N., \& Leal, L. G. 1984,
902: Phys. Fluids, 27, 1094
903: \bibitem[Landau \& Lifshitz (1959)]{ll59} Landau, L. D., \& Lifschitz, E. M. 1959,
904: Fluid Mechanics (Oxford: Pergamon Press)
905: \bibitem[Landau \&
906: Lifshitz (1977)]{ll77} Landau, L. D., \& Lifschitz, E. M. 1977,
907: Quantum Mechanics (Non-Relativistic Theory) (Oxford: Pergamon
908: Press) p.304
909: %\bibitem[Litwick \& Goldreich (2001)]{lg01} Lithwick, Y. \& Goldreich, P. 2001, \apj 562, 279
910: \bibitem[Maeder \& Meynet (2000)]{mm00} Maeder, A., \& Meynet, G. 2000, \araa, 38,
911: 143
912: \bibitem[Mahajan \& Rogava (1999)]{mr99} Mahajan, S. M., \& Rogava, A. D. 1999, \apj, 518,
913: 814
914: \bibitem[Miesch (2005)]{m05} Miesch, M. S. 2005, Living Rev. in \solphys, 2,
915: lrsp-2005-1
916: %(\texttt{http://solarphysics.livingreviews.org/Articles/lrsp-2005-1/})
917: \bibitem[Miles, 1957]{m57} Miles, J. W. 1957, J. Acoust. Soc. Amer., 29, 226
918: \bibitem[Olver (1974)]{o74} Olver, F. W. J. 1974, Asymptotics and Special
919: Functions (New York: Academic Press) p.480
920: \bibitem[Petrovay (2003)]{p03} Petrovay, K. 2003, \solphys, 215, 17
921: %\bibitem[Petukhov (2006)]{p06} Petukhov, M. Yu. 2006, Astron. Lett 32, 418
922: \bibitem[Poedts et al. (1998)]{p98} Poedts, S., Rogava, A. D., \&
923: Mahajan, S. 1998, \apj, 505, 369
924: \bibitem[Poedts, Rogava \& Mahajan (1999)]{prm99} Poedts, S., Rogava, A. D., \& Mahajan, S. 1999,
925: \ssr, 87, 295
926: %\bibitem[Press (1981)]{p81} Press, W. H. 1981,
927: %\apj 245, 286
928: \bibitem[Rogava, Poedts \& Mahajan (2000)]{rpm00} Rogava, A. D., Poedts, S., \& Mahajan,
929: S. M. 2000, \aap, 354, 749
930: \bibitem[Rogava (2004)]{r04} Rogava, A. D. 2004, \apss, 293, 189
931: %\bibitem[Rogava et al. (2006)]{r06} Rogava, A. D. et al. 2006, (in
932: %preparation)
933: %\bibitem[Rogers \& Glatzmeier (2005)]{rg05} Rogers, T. M., \& Glatzmaier, G. A. 2005, \mnras, 364,
934: %1135
935: %\bibitem[Shergelashvili et al. (2005)]{s05} Shergelashvili, B. M., Rogava, A. D., \&
936: %Poedts, S. 2004, SOHO 15 Workshop - Coronal Heating, R.W. Walsh,
937: %J. Ireland, D. Danesy, \& B. Fleck, Paris: European Space Agency
938: %(ESA SP-575), 437
939: \bibitem[Talon \& Charbonnel (1998)]{tc98} Talon, S., \& Charbonnel, C. 1998, \aap, 335,
940: 959
941: \bibitem[Talon \& Charbonnel (2003)]{tc03} Talon, S., \& Charbonnel, C. 2003, \aap, 405,
942: 1025
943: \bibitem[Talon \& Charbonnel (2004)]{tc04} Talon, S., \& Charbonnel, C. 2004, \aap, 418,
944: 1051
945: \bibitem[Talon \& Charbonnel (2005)]{tc05} Talon, S., \& Charbonnel, C. 2005, \aap, 440,
946: 981
947: %\bibitem[Thompson et al. (2003)]{t03} Thompson, M. J., Christensen-Dalgrado, J., Miesch, M. S.,
948: %\& Toomre, J. 2003, \araa, 41, 599
949: %\bibitem[Wolff \& Mayr (2004)]{wm65} Wolff, C. L., \& Mayr, H. G. 2004, \apj, 606,
950: %L163
951: \bibitem[Zahn, Talon \& Matias (1997)]{z97} Zahn, J.-P., Talon, S., \& Matias, J. 1997, \aap, 322, 320
952: \end{thebibliography}
953: %----------------------------------------------------------------------
954:
955: \end{document}
956: