1:
2: %% \documentclass[12pt,preprint]{aastex}
3: %\documentclass[twocolumn,showpacs,preprintnumbers,amsmath,amssymb]{revtex4}
4: %\documentclass[twocolumn,showpacs,preprintnumbers,amsmath,amssymb]{aastex}
5: %% \documentclass[manuscript]{aastex}
6: \documentclass[twocolumn,showpacs,preprintnumbers,amsmath,amssymb]{revtex4}
7: % \documentclass[preprint2]{aastex}
8: %\documentclass[twocolumn]{aastex}
9: %% \documentclass[preprint2,longabstract]{aastex}
10: %% You can insert a short comment on the title page using the command below.
11: \usepackage{graphicx}% Include figure files REVTEX
12: \usepackage{dcolumn}% Align table columns on decimal point REVTEX
13: \topmargin -15pt %REVTEX
14:
15: \newcommand{\beq}{\begin{equation}}
16: \newcommand{\eeq}{\end{equation}}
17: %\slugcomment{????????????????}
18:
19: %% If you wish, you may supply running head information, although
20: %% this information may be modified by the editorial offices.
21: %% The left head contains a list of authors,
22: %% usually a maximum of three (otherwise use et al.). The right
23: %% head is a modified title of up to roughly 44 characters.
24: %% Running heads will not print in the manuscript style.
25:
26: %APJ%%\shorttitle{Nonlocal MHD turbulence}
27: %APJ%%%\shortauthors{Alexandros Alexakis}
28:
29: %% This is the end of the preamble. Indicate the beginning of the
30: %% paper itself with \begin{document}.
31:
32: \begin{document}
33:
34: %% LaTeX will automatically break titles if they run longer than
35: %% one line. However, you may use \\ to force a line break if
36: %% you desire.
37:
38: \title{Nonlocal phenomenology for anisotropic MHD turbulence}
39:
40: %% Use \author, \affil, and the \and command to format
41: %% author and affiliation information.
42: %% Note that \email has replaced the old \authoremail command
43: %% from AASTeX v4.0. You can use \email to mark an email address
44: %% anywhere in the paper, not just in the front matter.
45: %% As in the title, use \\ to force line breaks.
46:
47: \author{A. Alexakis}
48: %%%APJ%%%%\affil{Laboratoire Cassiop\'ee, Observatoire de la C\^ote d'Azur, BP 4229, Nice Cedex 04, France}
49: \affiliation{Laboratoire Cassiop\'ee, Observatoire de la C\^ote d'Azur, BP 4229, Nice Cedex 04, France}
50: %%%APJ%%%%\email{alexakis@obs-nice.fr, aalexakis@gmail.com}
51:
52: %% Notice that each of these authors has alternate affiliations, which
53: %% are identified by the \altaffilmark after each name. Specify alternate
54: %% affiliation information with \altaffiltext, with one command per each
55: %% affiliation.
56:
57: \begin{abstract}
58: A non-local cascade model for anisotropic MHD turbulence in the presence of a
59: guiding magnetic field is proposed. The model takes into
60: account that (a) energy cascades in an anisotropic manner
61: and as a result a different estimate for the cascade rate
62: in the direction parallel and perpendicular to the guiding field is made.
63: (b) the interactions that result in the cascade are between different
64: scales. Eddies with wave numbers $k_\|$ and $k_\perp$ interact with eddies with
65: wave numbers $q_\|,q_\perp$ such that a resonance condition between the wave numbers
66: $q_\|,q_\perp$ and $k_\|,k_\perp$ holds.
67: As a consequence energy from the eddy with wave numbers $k_\|$
68: and $k_\perp$ cascades due to interactions with eddies
69: located in the resonant manifold whose wavenumbers are determined by:
70: $q_\|\simeq \epsilon^{{1}/{3}}k_\perp^{2/3}/B$, $q_\perp=k_\perp$ and
71: energy will cascade along the lines
72: $k_\|\sim C+k_\perp^{2/3} \epsilon^{1/3}/B_0$.
73: For a uniform energy injection rate in the parallel direction the resulting energy spectrum is
74: $E(k_\|,k_\perp)\simeq \epsilon^{2/3}k_\|^{-1}k_\perp^{-5/3}$. For a
75: general forcing however the model suggests a non-universal behavior.
76: The connections with previous models, numerical simulations and weak turbulence theory are discussed.
77: \end{abstract}
78:
79: %% Keywords should appear after the \end{abstract} command. The uncommented
80: %% example has been keyed in ApJ style. See the instructions to authors
81: %% for the journal to which you are submitting your paper to determine
82: %% what keyword punctuation is appropriate.
83:
84: \keywords{MHD,guiding field Anisotropic energy spectrum}
85:
86: %% From the front matter, we move on to the body of the paper.
87: %% In the first two sections, notice the use of the natbib \citep
88: %% and \citet commands to identify citations. The citations are
89: %% tied to the reference list via symbolic KEYs. The KEY corresponds
90: %% to the KEY in the \bibitem in the reference list below. We have
91: %% chosen the first three characters of the first author's name plus
92: %% the last two numeral of the year of publication as our KEY for
93: %% each reference.
94:
95: \maketitle % REVTEX
96:
97:
98: \section{Introduction}
99:
100: Magnetic fields are met very often in astrophysics; interstellar
101: medium, accretion discs, the interior of stars and planets. In most
102: of these cases the magnetic fields are strong enough to play a
103: dynamical role in the evolution of the involved astronomical objects
104: \citep{Zeldovich}. Furthermore, the fluid and magnetic Reynolds
105: numbers are large enough so that a large number of scales are exited
106: and coupled together, making it very difficult to calculate the
107: evolution of these systems even with the power of present day
108: computers. As a result a turbulence theory that models the behavior
109: of the small unresolved scales is in need. The simplest set of equations
110: that describes the evolution of the flow and the magnetic field when the
111: two are coupled together are the magneto-hydro-dynamic (MHD)
112: equations that in the Els\"asser formulation are written as
113: \begin{equation}
114: \label{MHD}
115: \partial_t {\bf z}^\pm = \pm {\bf B} \cdot \nabla {\bf z}^\pm - {\bf z}^\mp \cdot \nabla {\bf z}^\pm -\nabla P + \nu \nabla^2 {\bf z}^\pm
116: \end{equation}
117: where ${\bf z}^\pm={\bf u\pm b}$ with ${\bf u}$ the velocity,
118: ${\bf b}$ the magnetic field, $\nu$ the molecular viscosity
119: assumed here equal to the magnetic diffusivity $\eta=\nu$, $B$ is a
120: uniform magnetic field, and incompressibility $\nabla \cdot z^\pm=0$
121: has been assumed.
122:
123: For zero viscosity the above equations conserve three quadratic
124: invariants the magnetic Helicity (that we are not going to be
125: concerned with in the present work) and the two energies
126: $E^{\pm}=\int (z^\pm)^2 dx^3$. The question then arises in the limit
127: of infinite Reynolds number is there a physical process under which
128: the two energies cascade to sufficiently small scales so that
129: they can be dissipated?
130:
131: %\section{Previous Models}
132:
133: For hydrodynamic turbulence a description of such a process
134: exists and was given by
135: \citet{Kolmogorov1941} (K41). In his phenomenological description the energy
136: $z_l^2$ at a scale $l$ interacts with similar size eddies and
137: cascades in a timescale $l/z_l$. As a result in a
138: statistically steady state the energy cascades in a scale
139: independent way at a rate $\epsilon \simeq z_l^3/l$ that leads to
140: the prediction $z_l\sim l^{1/3}$ or in terms of the 1-D energy
141: spectrum $E(k)\sim k^{-5/3}$. Since the phenomenological
142: description of the energy cascade in hydrodynamic turbulence
143: %by {Kolmogorov}
144: there have been attempts to derive similar results for
145: MHD flows. However, non-trivial difficulties arise when a mean
146: magnetic field is present.
147: %
148: %
149: First the MHD equations are no longer
150: isotropic resulting in an anisotropic energy flux and energy spectrum.
151: Simple dimensional arguments can not be used
152: to estimate the degree of anisotropy that is a dimensionless quantity.
153: %
154: %
155: Second the MHD equations are
156: no longer scale invariant, as a result simple power law behavior of
157: the energy spectrum is expected only in the small or large B limit
158: that scale similarity is recovered.
159: %
160: %
161: Finally, it is not clear that
162: interactions of similar size eddies (local interactions)
163: dominate the cascade. Different size eddies could
164: play an important in cascading the energy.
165:
166:
167: The first model for MHD turbulence was proposed by \citet{Iroshnikov1963} and by
168: \citet{Kraichnan1965} (IK). The IK-model assumes isotropy and that the
169: time scale of the interactions of two wave packets of size $l$ is
170: given by the Alfven-time scale $\tau_{_A}\sim l/B$. The energy
171: cascade due to a single collision is given by $\Delta z^2 \sim
172: (z^3/l) \tau_{_A}$. The number of random collisions that would be
173: required then to cascade the energy is going to be $N\sim (z^2 /\Delta
174: z^2)^2$. As a result the energy will cascade in a rate $\epsilon
175: \sim E/N \tau_{_A} \sim z^4/(Bl)$ and therefore $z_l \sim (\epsilon
176: B)^{1/4}l^{1/4}$. The resulting 1-D energy spectrum is then given
177: by $E(k)\sim (\epsilon B)^{1/2} k^{-3/2}$.
178: %
179: %
180: The assumption of
181: isotropy however has been criticized in the literature and
182: anisotropic models have been proposed for the energy spectrum.
183: \citet{Goldreich1995} (GS) proposed that in strong turbulence the
184: cascade happens for eddies such that the Alfven time scale
185: $\tau_{_A}\sim Bk_\|$ is of the same order with the non-linear time
186: scale $\tau_{_{NL}}\sim z k_\perp$ (so called critical balance
187: relation), where $k_\|$ and $k_\perp$ are the parallel and
188: perpendicular to the mean magnetic field wavenumbers respectively.
189: Repeating the Kolmogorov arguments then one ends up with the
190: energy spectrum
191: $E(k_\|,k_\perp)\sim k_\perp^{-5/3}$ with
192: the parallel and perpendicular wave numbers following the relation
193: $k_\|\sim k_\perp^{2/3}$.
194: %
195: %
196: A generalization of this result was proposed by \citet{Galtier2005} where
197: the ratio of the two time scales $\tau_{_A}/\tau_{_{NL}}$ was kept
198: fixed but not necessarily of order one, in an attempt to model MHD
199: turbulence both in the weak and the strong limit.
200: %
201: %
202: \citet{Bhatta2001} (BN) repeated the IK-model arguments replacing
203: the nonlinear time scale by $\tau_{_{NL}}\sim l_\perp/z_l$ and the
204: Alfven time scale by $\tau_{_{A}}\sim l_\|/B$.
205: Further assuming that the
206: cascade is happening only in the $k_\perp$ direction
207: obtained the energy spectrum $E(k)\sim (\epsilon B)^{1/2} k_\perp^{-2}k_\|^{-1}$.
208: %
209: %
210: Finally \citet{Zhou2004} (ZMD) suggested using as
211: time scale the one given by the inverse average of the Alfven and
212: nonliner time scale
213: $\tau^{-1}=(\tau_{_A})^{-1}+\tau_{_{NL}}^{-1}$ to obtain a smooth
214: transition from the K41 to the IK and the anisotropic BN result
215: depending on the amplitude of the guiding field.
216:
217: Although this large variety of models exists the agreement with
218: observations \citep{Goldstein1995} and with the results of numerical
219: simulations
220: \citep{Muller2005,Ng2003,Maron2001,Cho2000,Biskamp2000,Ng1996}
221: %\citep{Muller2005,Ng2003,Maron2001,Cho2000,Biskamp2000,Ng1997,Ng1996}
222: is only partially satisfactory and seems to be case dependent.
223: Furthermore, all these models assume locality of interactions (i.e.
224: only similar size eddies interact and only one
225: length scale is needed in each direction in the phenomenological description).
226: Locality of interactions however has been shown to be in question by
227: both theoretical arguments and analysis of data in numerical
228: simulations even in the isotropic case
229: \citep{Alexakis2005,Debliquy2005,Yousef2007}. Furthermore the
230: rigorous result for weak turbulence \citep{Galtier2000,Galtier2002}
231: has shown that only modes in the resonant manifold $k_\|=0$ are
232: responsible for the energy cascade. It seams reasonable therefore
233: that non-locality is an essential ingredient of MHD turbulence that
234: needs to be taken in to account in a model. %%
235: %%
236: In addition we expect that the energy will not cascade isotropically
237: so not only the amplitude of the energy cascade rate is of
238: importance but also the direction.
239: %%
240: With these two points in mind (anisotropy and non-locality)
241: we try to construct a non-local
242: model, for the energy cascade.
243:
244: %\section{New Model}
245:
246: To begin with the new model let us consider a MHD flow
247: in a statistically steady turbulent state forced at
248: large scales, in the presence of mean magnetic field $B$.
249: We will denote the
250: two 2-D energy spectra as $E^+(k_\perp, k_\|)$ and $E^-(k_\perp,
251: k_\|)$ where the total energy is given by $E_{_T}^\pm \equiv \int
252: E^\pm dk_\perp dk_\|$.
253: %
254: %Here we have used $k$ and $q$ because we want
255: %to allow interactions between different scales ($k \ne q$).
256: %
257: For simplicity assume $E^-(k_\perp, k_\|)\sim E^+(k_\perp, k_\|)$ (i.e.
258: negligible cross helicity) and drop the $\pm$ indexes leaving the case
259: $E^-\ll E^+$ to be investigated in the future. To shorten the
260: notation we will write $E_k=E(k_\perp, k_\|)$.
261: %and $E_q=E(q_\perp, q_\|)$
262: The index $k$ denotes that $E_k$ depends on the
263: wavenumbers $k_\|$ and $k_\perp$.
264:
265: Let us now consider two eddies of different scales $z^+_k$ and $z^-_q$ interacting.
266: Let us assume that the $z^+_k$ eddy has wave numbers $\sim k_\perp,k_\|$ and
267: the $z^-_q$ eddy has wave numbers $\sim q_\perp,q_\|$.
268: Here we will focus on the cascade of the energy of the $z^+_k$ eddy,
269: the cascade of the $z^-_q$ can be obtained by changing the indexes $k,q$ and $\pm$.
270: From the form of the non-linear term we expect by dimensional analysis
271: %argument
272: that the rate of energy cascade of the $z^+_k$ eddy will be
273: %$\partial_t \langle (z_k)^2 \rangle \sim \langle z_k z_q \nabla z_k\rangle$
274: \begin{eqnarray}
275: \mathcal{E}(k) & \sim (z_k^+)^2 (z_q^-) q \qquad \qquad \qquad \nonumber \\
276: & \sim [k_\perp k_\| E_k ]
277: [ q_\perp q_\| E_q]^{1/2} q
278: \end{eqnarray}
279: and similar for the energy cascade rate of the $z^-_q$ eddy.
280: Note that in such an interaction
281: the energy will not cascade isotropically but it will depend on the value of ${\bf q}$.
282: In a interaction of the two eddies the energy of the $z^+_k$ eddy will move from the wavenumber
283: ${\bf k}$ to the wave number ${\bf k+q}$.
284: If $q_\|\ll q_\perp$ then most of the cascade will be in the $q_\perp$ direction.
285: As a result we need to define separately the rate energy cascades
286: to larger $k_\perp$: $\mathcal{E}_\perp (k)$ and the rate energy cascades to larger
287: $k_\|$: $\mathcal{E}_\| (k)$ as:
288: \beq
289: \label{Eperp}
290: \mathcal{E}_\perp \sim [ k_\perp k_\| E_k ] [ q_\perp q_\| E_q ]^{1/2} q_\perp
291: \eeq
292: \beq
293: \label{Epara}
294: \mathcal{E}_\| \sim [ k_\perp k_\| E_k ] [ q_\perp q_\| E_q ]^{1/2} q_\|.
295: \eeq
296: %
297: Note that in writing the equations above we have not taken in to account possible
298: scale dependent correlations between the two fields that could reduce the energy
299: cascade. Such an effect has been taken into account by \citet{Boldyrev2005} based
300: on the GS model and could be incorporated in the present model.
301: However we will not make such an attempt here since we want to present the model
302: in its simplest form.
303: Equations \ref{Eperp} and \ref{Epara} express the rate energy cascades in the absence
304: of a mean magnetic field and are valid only when $|q|<|k|$ because small eddies
305: although they have a stronger shear rate $z_qq$ decorrelate making them less effective in cascading the energy.
306: However in the presence of guiding field not all wave numbers $q$
307: are as effective in cascading the energy $E_k$.
308: Because the two eddies $z^+_k$,$z^-_q$ travel in opposite directions the time they will interact
309: will be the Alfven time $\tau_{_A} \sim [q_\| B_0]^{-1}$ and the time needed to cascade the
310: energy will be $\tau_{_{NL}} \sim (z^-|q|)^{-1} \sim [ |q| \sqrt{q_\perp q_\| E_q}]^{-1}$.
311: %
312: %
313: Therefore, from all the available wavenumbers only
314: the wave numbers with $\tau_{_A} \gtrsim \tau_{_{NL}}$ will be effective
315: in cascading the energy.
316: This restriction leads to:
317: \beq
318: \label{qpara1}
319: q_\| B \lesssim q_\perp \sqrt{q_\perp q_\| E_q}
320: \eeq
321: where we used the approximation $|q| \simeq q_\perp$ as a first order approximation
322: of $|q|$ for large $B$.
323: This relation looks very similar to the critical balance relation of
324: the GS model. However in this case the relation \ref{qpara1} gives
325: the wave numbers that the eddy $z^+_k$ will interact with and does not
326: restrict the location of $z^+_k$ in spectral space.
327: We are going to refer to the set of wavenumbers that satisfy the relation above as
328: the resonant manifold and
329: we are going to use this relation as an equality because from the allowed wavenumbers $q_\|$
330: the ones closer to $k_\|$ will be more effective in cascading the energy.
331: Here as in the equations \ref{Eperp} and \ref{Epara}, we assume that $q_\|<k_\|$.
332: Finally since the mean magnetic field does not directly effect the $\perp$ direction we assume that
333: similar size $k_\perp$ and $q_\perp$ are the most effective to cascade the energy
334: $q_\perp \sim k_\perp$ ({i.e. locality in the $k_\perp$ direction}).
335: So equation \ref{qpara1} is written as
336: \beq
337: \label{qpara}
338: q_\|\simeq \frac{ k_\perp^3 E_q}{B_0^2}
339: \eeq
340: We are now ready to impose the constant energy flux condition that would lead to a stationery
341: spectrum. Because the cascade is anisotropic, constant energy flux now reads:
342: \beq
343: \label{flux}
344: \partial_{k_\perp} \mathcal{E}_\perp + \partial_{k_\|} \mathcal{E}_\| =0
345: \eeq
346: The equations \ref{Eperp},\ref{Epara},\ref{qpara},\ref{flux} form the basic equations of our model.
347: It is worth noting, that in this model the cascade of energy decreases by the introduction of the guiding field
348: not because the individual interactions weakened but because the number of modes that are able to cascade the energy
349: decreases due to the resonance condition \ref{qpara}.
350: A sketch of the mechanisms involved in the model is shown in figure \ref{fig1}.
351: %%\begin{figure}
352: %%\epsscale{.80}
353: %%\plotone{FIG1.eps}
354: %%\caption{An ilustration of the non local model. A eddie at $k_\|,k_|perp$ interacts
355: %%with an eddie at the resonant manifold $q_\|,k_\perp$ cascading the energy in the direction of the arrow.
356: %%\label{fig1}}
357: %%\end{figure}
358: \begin{figure}
359: \includegraphics[width=8cm]{FIG1c.eps}
360: %\includegraphics{FIG1.eps}
361: \caption{An illustration of the non local model. A eddy at $(k_\|,k_\perp)$ interacts (non-locally)
362: with an eddy at the resonant manifold $(q_\|(k_\perp),k_\perp)$ cascading the energy in the direction of the arrow.
363: \label{fig1}}
364: \end{figure}
365:
366: %\section{Energy spectra}
367:
368: First let us concider the weak turbulence limit that is obtained
369: in the limit $B\to \infty$.
370: For large $B$ based on equation \ref{qpara} the resonant manifold becomes very thin $q_\|/k_\perp \ll 1$.
371: Furthermore because $\mathcal{E}_\|/\mathcal{E}_\perp \sim q_\|/k_\perp \ll 1$
372: we can neglect the cascade in the parallel direction.
373: If also $E_k$ is non-singular at $k_\|=0$ we have that $E_q\simeq E(k_\perp,0)$.
374: Substituting $q_\|$ from \ref{qpara} in \ref{Eperp} and imposing the constant flux condition \ref{flux}
375: we obtain:
376: \beq
377: \mathcal{E}_\perp \sim k_\perp^4 k_\| E_k E(k_\perp, 0 )/B=\epsilon(k_\|).
378: \eeq
379: Note that the spectrum $E(k_\perp, k_\|)$ depends on the energy of the resonant
380: manifold $E(k_\perp, 0 )$ just like the weak turbulence result and
381: unlike what the BN local theory for weak turbulence predicts.
382: If the energy spectrum for $E(k_\perp, k_\|)$ and for the resonant manifold $E(k_\perp, 0)$
383: scale like $k_\perp^n$ and $k_\perp^m$ respectively then we end up with
384: the weak turbulence prediction \citep{Galtier2000}:
385: \beq
386: \label{WC}
387: m+n=-4
388: \eeq
389: Assuming that the two spectra are
390: smooth around $k_\|=0$ (just like the weak turbulence theory
391: needs to assume) we obtain:
392: \beq
393: \label{Wspec}
394: E^\pm \sim k_\perp^{-2} \sqrt{B\epsilon(k_\|)/k_\|}.
395: \eeq
396:
397: As we decrease the value of $B$ we need to take in to account that the energy
398: cascade in the $\|$ direction is non zero. In this case energy does not cascade
399: in the $\perp$ direction but cascades along the lines that
400: are tangent to the direction of $\mathcal{E}$ and
401: satisfy $dk_\|/dk_\perp=\mathcal{E}_\perp/\mathcal{E}_\|=q_\|/k_\perp$ or
402: \beq
403: \label{lines}
404: k_\|=\int_0^{k_\perp} \frac{q_\|(k_\perp')}{k_\perp'}dk_\perp'+C.
405: \eeq
406: Let $\lambda$ be the length along such a curve; then we can move
407: to a new coordinate system given by ($\lambda,C$).
408: In this new coordinate system the constant flux relation to first order in $q_\|$ reads
409: \beq
410: \frac{d|\mathcal{E}|}{d\lambda}=\frac{d}{d\lambda}\left( [ k_\perp k_\| E_k ] [ q_\perp q_\| E_q ]^{1/2} |q| \right)=0
411: \eeq
412: or
413: \beq
414: \label{energyGS}
415: [ k_\perp k_\| E_k ] [ k_\perp q_\| E_q ]^{1/2} k_\perp = \epsilon(C)
416: \eeq
417: where
418: %$\epsilon(C)$ is independent of $\lambda$ but could depend on principle on
419: %$C$ and
420: only terms up to order $q_\|$ are kept.
421: Letting $k_\|\to q_\|$ we obtain the equation for the resonant manifold.
422: \beq
423: [ k_\perp q_\| E_q ] [ k_\perp q_\| E_q ]^{1/2} k_\perp = \epsilon(0) \equiv \epsilon_0.
424: \eeq
425: Substituting $q_\|$ from \ref{qpara} and solving for $E_q$ we obtain
426: \beq
427: \label{EnR}
428: E_q =E(q_\|(k_\perp),k_\perp)= \epsilon_0^{1/3} B k_\perp^{-7/3}.
429: \eeq
430: The equations \ref{qpara} and \ref{lines} then gives us
431: \beq
432: q_\|= k_\perp^{2/3} \epsilon_0^{1/3}/B \,\, \mathrm{and}\,\, k_\|=\frac{3}{2}k_\perp^{2/3} \epsilon_0^{1/3}/B+C.
433: \eeq
434: Returning to the equation for the energy energy \ref{energyGS} we get
435: \beq
436: \label{spec1}
437: E_k = \frac{ \epsilon(C) \epsilon_0^{-1/3}}{ k_\perp^{5/3} k_\|}
438: \eeq
439: where $C$ is given by equation \ref{lines} and
440: the predicted spectra \ref{spec1},\ref{spec2} are valid in the range $q_\|< k_\| <\infty$.
441: For smaller values of $k_\|$ the condition $k_\|< q_\|$ that we initially assumed is not satisfied.
442: The energy of the modes inside the resonant manifold is given by \ref{EnR} and no singularity
443: at $k_\|=0$ exists.
444:
445: In the special case that $\epsilon(C)$ is a constant $\epsilon(C)=\epsilon_0$
446: that corresponds to a uniform injection rate per unit of wavenumber ($k_\|$)
447: at the large scales $k_\perp \to 0$
448: the spectrum reduces to
449: \beq
450: \label{spec2}
451: E_k = \frac{ \epsilon_0^{2/3} }{ k_\perp^{-5/3} k_\|}
452: \eeq
453: but in general the spectrum will depend on the way energy is injected in the system.
454: The non-universality that the model suggests is due to the fact that we assumed
455: that the energy cascades in a deterministic way only along the lines in the ($k_\|,k_\perp$) plane
456: given by \ref{lines}. In reality energy will not cascade strictly along the lines \ref{lines}
457: but there is going to be some exchange of energy between lines that could bring the energy spectrum in the form
458: \ref{spec2}. However, if and how fast a universal spectrum can be obtained in MHD
459: is not an easy question to answer.
460: This question is related to the return to isotropy of an
461: anisotropicaly forced flow in hydrodynamic turbulence,
462: %Recent work has shown a slower
463: %return to isotropy that K41 predictions would suggest
464: that is still an open question.
465: If indeed in MHD turbulence in the presence of a guiding field there is a universal
466: spectrum this is expected to happen at smaller scales than in hydrodynamic turbulence
467: because nonlinear interactions are weaker.
468:
469: So far we concerned ourselves with only large values of $B$.
470: In principle we could extend our results to smaller values of $B$
471: with out making some of the approximations used to arrive at the results \ref{spec1},\ref{spec2}.
472: However, such procedure leads to more complex equations that prevent us from deriving
473: the energy spectrum in a compact form and we do not make such an attempt at present.
474:
475: It is worth emphasizing the similarities the current model has with GS model.
476: Both models emphasize the role
477: of the manifold $k_\|\simeq k_\perp^{2/3}\epsilon_0^{1/3}/B$,
478: obtained by the resonance condition \ref{qpara} in out model
479: or the critical balance condition in the GS model.
480: However in this model
481: the cascade is not restricted in this manifold, but instead all modes in the ($k_\|,k_\perp$) plane cascade
482: due to non-local interactions with the modes in this manifold.
483:
484: An other point we need to emphasize is that taking $B\to \infty$ does not reduce
485: the predicted spectra in \ref{spec1} to the weak turbulence limit \ref{Wspec}.
486: This is due to the two different limiting procedures followed.
487: In the first case (eq.\ref{Wspec}) first the limit $B\to\infty$ was taken and then the 1-D flux was determined,
488: while in the second case \ref{spec1} we first obtained the 2-D energy flux and then the limit $B\to\infty$ was taken.
489: Note however that the condition \ref{WC} is satisfied in both cases and the spectrum
490: is also smooth at $k_\|=q_\|$ because although the resonant manifold scales
491: like $k_\perp^{-7/3}$ while the rest of the spectrum scales like $k_\perp^{-5/3}k_\|^{-1}$
492: the resonant manifold widens as $k_\perp$ increases.
493: It is possible as we discuss bellow that in different (numerical) setups
494: either of the two limiting procedures
495: can be valid and different spectra could be obtained.
496:
497: In numerical simulations a finite discrete number of modes is kept.
498: Based on this model
499: the cascade rate is reduced in the presence of a mean magnetic field
500: not because the individual interactions themselves are weakened but because
501: the number of modes that that interact effectively is reduced.
502: If the modes in a numerical simulation with the smallest non-zero wavenumber $k_1=2\pi/L$
503: (where $L$ is the box height) is larger than the resonant manifold
504: ( $k_1>q_\|$ ) then if $B$ is further increased
505: the scaling with $B$ of the energy dissipation rate
506: will be lost and the spectrum exponents could change,
507: since the number of modes in the
508: resonant manifold already have taken their minimum value (i.e. the number of
509: modes that have $k_\|=0$).
510: Furthermore a difference in the energy spectrum exponents can be expected
511: in numerical simulations if the modes inside the resonant manifold are not forced.
512: The sensitivity of the model to the way the system is forced could in part
513: explain the disagreement in the measured spectrum exponents.
514:
515:
516:
517: %% The equation environment wil produce a numbered display equation.
518:
519:
520: %\acknowledgments
521:
522: The author is grateful for the support
523: he received from
524: the Observatoire de la C\^ote d'Azur.
525:
526:
527:
528: \begin{thebibliography}{}
529: \bibitem[Alexakis, Mininni \& Pouquet (2005)]{Alexakis2005}
530: Alexakis A., Mininni P.D. \& Pouquet A.,
531: 2005, Phys. Rev. E 72, 046301
532: % title = "Shell to shell energy transfer in {M}{H}{D}. {I}. {S}teady
533: % state turbulence",
534:
535:
536: \bibitem[Boldyrev (2005)]{Boldyrev2005}
537: Boldyrev, S.,
538: 2001, \apj, 626, L37-L40.
539:
540:
541: \bibitem[Bhattacharjee \& Ng (2001)]{Bhatta2001}
542: Bhattacharjee, A., Ng, C. S.,
543: 2001, \apj, 548, 318-322
544: % Random Scattering and Anisotropic Turbulence of Shear Alfvén Wave Packets
545:
546: \bibitem[Biskamp \& M{\"u}ller (2000)]{Biskamp2000}
547: Biskamp, D. \& M{\"u}ller, W.C.,
548: 2000 Phys. Plasmas, 7, 4889
549:
550: \bibitem[Cho \& Vishniac (2000)]{Cho2000}
551: Cho J. \& Vishniac E.T., 2000 \apj 539, 273
552:
553: \bibitem[Debliquy, Verma \& Carati (2005)]{Debliquy2005}
554: Debliquy O.,Verma M.K. \& Carati D.,
555: 2005, Phys. Plasmas 12, 042309
556: %Energy fluxes and shell-to-shell transfers in three-dimensional
557: % decaying magnetohydrodynamic turbulence",
558:
559: \bibitem[Galtier et al.(2005)]{Galtier2005}
560: Galtier, S., Pouquet, A., \& Mangeney A.,
561: 2005, Phys. Plasmas 12, 092310
562:
563: \bibitem[Galtier et al.(2002)]{Galtier2002}
564: Galtier, S.; Nazarenko, S. V.; Newell, A. C.; Pouquet, A.,
565: 2002, \apj, 564, L49-L52.
566:
567: \bibitem[Galtier et al.(2000)]{Galtier2000}
568: Galtier, S.; Nazarenko, S. V.; Newell, A. C.; Pouquet, A.
569: 2000, J. Plasma Phys., 63, 447.
570: %A Weak Turbulence Theory for Incompressible Magnetohydrodynamics
571:
572: \bibitem[Goldreich \& Sridhar (1995)]{Goldreich1995}
573: Goldreich, P., Sridhar, S. 1995, \apj, 438, 763.
574: %TOWARD A THEORY OF INTERSTELLAR TURBULENCE .2. STRONG ALFVENIC TURBULENCE
575:
576: \bibitem[Goldstein, Roberts \& Matthaeus]{Goldstein1995}
577: Goldstein, M.L., Roberts D.A., \& Matthaeus W.H., 1995,
578: Annu. Rev. Astron. Astrophys. 33, 283.
579:
580: \bibitem[Iroshnikov (1963)]{Iroshnikov1963}
581: Iroshnikov, P. 1963, Soviet Astron., 7, 566.
582:
583: \bibitem[Kolmogorov (1941)]{Kolmogorov1941}
584: Kolmogorov, A.N. 1941,
585: Dokl. Akad. Nauk SSSR 30,299
586:
587: \bibitem[Kraichnan (1965)]{Kraichnan1965}
588: Kraichnan, R. 1965, Phys. Fluids, 8, 1385
589:
590: \bibitem[Maron \& Goldreich (2001)]{Maron2001}
591: Maron J., Goldreich P., 2001 \apj 554, 1175
592:
593: \bibitem[M{\"u}ller \& Grappin (2005)]{Muller2005}
594: M{\"u}ller W.C., Grappin, R., 2005, Phys. Rev. Lett. 95, 114502
595:
596:
597: \bibitem[Ng \& Bhattacharjee (1996)]{Ng1996}
598: Ng, C.S., Bhattacharjee, A. 1996, \apj, 465, 845.
599:
600: %\bibitem[Ng, Bhattacharjee (1997)]{Ng1997}
601: % Ng, C. S., Bhattacharjee, A. 1997, Phys. Plasmas, 4, 605
602:
603: \bibitem[Ng, Bhattacharjee \& Germaschewski (2003)]{Ng2003}
604: Ng, C.S., Bhattacharjee, A., Germaschewski K., Galtier, S. 2003, Phys. Plasmas, 10, 1954
605:
606: %\bibitem[Sridhar \& Goldreich (1994)]{Sridhar1994}
607: %Sridhar, S., Goldreich, P. 1994, \apj, 432, 612
608:
609: \bibitem[Yousef, Rincon \& Schekochihin (2007)]{Yousef2007}
610: Yousef T.A., Rincon F., \& Schekochihin A.A., 2007, J. Fluid Mech. 575, 111.
611: % title = "Exact scaling laws and the local structure of isotropic
612: % magnetohydrodynamic turbulence",
613:
614: \bibitem[Zeldovich, Ruzmaikin \& Sokoloff (1990)]{Zeldovich}
615: Zeldovich Ya. B., Ruzmaikin A.A., Sokoloff D.D.
616: ``Magnetic Fields in Astrophysics" 1990 Gordon \& Breach Science Pub.
617:
618:
619: \bibitem[Zhou, Matthaeus \& Dmitruk (2004)]{Zhou2004}
620: Zhou Y., Matthaeus W.H., Dmitruk P. 2004
621: Rev. Mod. Phys., 76, 1015-1035
622: %Colloquium: Magnetohydrodynamic turbulence and time scales in astrophysical and space plasmas
623:
624:
625: %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%5
626: %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
627:
628:
629: \end{thebibliography}
630:
631:
632: %% Use the figure environment and \plotone or \plottwo to include
633: %% figures and captions in your electronic submission.
634: %% To embed the sample graphics in
635: %% the file, uncomment the \plotone, \plottwo, and
636: %% \includegraphics commands
637: %%
638: %% If you need a layout that cannot be achieved with \plotone or
639: %% \plottwo, you can invoke the graphicx package directly with the
640: %% \includegraphics command or use \plotfiddle. For more information,
641: %% please see the tutorial on "Using Electronic Art with AASTeX" in the
642: %% documentation section at the AASTeX Web site,
643: %% http://www.journals.uchicago.edu/AAS/AASTeX.
644: %%
645: %% The examples below also include sample markup for submission of
646: %% supplemental electronic materials. As always, be sure to check
647: %% the instructions to authors for the journal you are submitting to
648: %% for specific submissions guidelines as they vary from
649: %% journal to journal.
650:
651: %% This example uses \plotone to include an EPS file scaled to
652: %% 80% of its natural size with \epsscale. Its caption
653: %% has been written to indicate that additional figure parts will be
654: %% available in the electronic journal.
655:
656:
657: \end{document}
658:
659: %%
660: %% End of file `sample.tex'.
661: