1: %\documentclass[preprint,floatfix,amsmath,amssymb]{revtex4}
2: \documentclass[twocolumn,floatfix,amsmath,amssymb]{revtex4}
3: %\documentclass[pre,preprint,showpacs,floatfix,amsmath,amssymb]{revtex4}
4: \usepackage{graphicx}
5: \usepackage{dcolumn}
6: \usepackage{color}
7:
8: \def\NOTE#1{{\textcolor{red}{\bf [#1]}}} % note
9: \def\ADD#1{{\textcolor{blue}{#1}}} % question
10: \def\AD#1{{\textcolor{magenta}{#1}}} % plug a value, a ref, ...
11: \def\DEL#1{{\textcolor{green}{ #1}}} % suggested deletion in text
12: \def\BB#1{{\textcolor{blue}{\bf #1}}}
13:
14: % table definitions
15: \def\hfq{\hfill\quad}
16: \def\cc#1{\hfq#1\hfq}
17: \def\tvi{\vrule height 12pt depth 5pt width 0pt}
18: \def\traithorizontal{\noalign{\hrule}}
19: \def\tv{\tvi\vrule}
20: %
21: \newcommand{\be}{\begin{equation}}
22: \newcommand{\ee}{\end{equation}}
23: \newcommand{\uh}{\hat{u}}
24: \def\eg{{\it e.g.}\ }
25: \def\etal{{\it et al.}}
26: \def\ie{{\it i.e.}\ }
27: \def\lhs{{\it l.h.s.}\ }
28: \def\op{{\it op. cit.}\ }
29: \def\resp{{\it resp.}\ }
30: \def\rhs{{\it r.h.s.}\ }
31: \def\rms{{\it r.m.s.}\ }
32: \def\viz{{\it viz.}\ }
33: \def\vs{{\it vs.}\ }
34: \def\al{Alfv\'en\ }
35: \def\els{Els\"asser variables\ }
36: \def\kol{Kolmogorov\ }
37: \def\nse{Navier-Stokes equations\ }
38: \def\u{{\bf u}} \def\v{{\bf v}} \def\x{{\bf x}}
39: \def\dv{\delta {\bf v}}
40: %
41: \newcommand{\curlv} {\nabla \times {\bf v}}
42: \newcommand{\ba}{\mathbf{a}} \newcommand{\bb}{\mathbf{b}}
43: \newcommand{\bA}{\mathbf{A}} \newcommand{\bB}{\mathbf{B}}
44: \newcommand{\Asz}{A_{s_z}}
45: \newcommand{\alp}{\alpha} \newcommand{\alpm}{\alpha^{-1}}
46: \newcommand{\alpmm}{\alpha^{-1}_m} \newcommand{\alpmv}{\alpha^{-1}_v}
47: \newcommand{\bc}{\mathbf{c}} \newcommand{\bd}{\mathbf{d}}
48: \newcommand{\bj}{\mathbf{j}} \newcommand{\bk}{\mathbf{k}}
49: \newcommand{\bom}{\mbox{\boldmath $\omega$}}
50: \newcommand{\bomp}{\mbox{\boldmath $\omega^+$} }
51: \newcommand{\bomm}{\mbox{\boldmath $\omega^-$} }
52: \newcommand{\bompm}{\mbox{\boldmath $\omega^{\pm}$} }
53: % bold face
54: \newcommand{\br}{\mathbf{r}}
55: \newcommand{\bu}{\mathbf{u}}
56: \newcommand{\bv}{\mathbf{v}}
57: \newcommand{\bw}{\mathbf{w}}
58: \newcommand{\bAs}{\mathbf{A_s}}
59: \newcommand{\bjs}{\mathbf{j_s}}
60: \newcommand{\bus}{\mathbf{u_s}}
61: \newcommand{\bBs}{\mathbf{B_s}}
62: \newcommand{\boms}{\mbox{\boldmath $\omega_s$}}
63: \newcommand{\bx}{\mathbf{x}} \newcommand{\bxp}{\mathbf{x^{\prime}}}
64: \newcommand{\bzp}{\mathbf{z^{+}}} \newcommand{\bzm}{\mathbf{z^{-}}}
65: \newcommand{\bzpm}{\mathbf{z^{\pm}}} \newcommand{\bzmp}{\mathbf{z^{\mp}}}
66: \newcommand{\ca}{{\rm a}} \newcommand{\caa}{{\rm aa}}
67: \newcommand{\cab}{{\rm a} b} \newcommand{\vba}{vb{\rm a}}
68: \newcommand{\K}{{\cal K}} \newcommand{\ud}{{\langle{u}^2 \rangle}}
69: \newcommand{\li}{\ell_{I}} \newcommand{\Rla}{R_{\lambda}}
70: \newcommand{\up}{{{\bf u}({\bf x})}} \newcommand{\R}{{\cal R}}
71: \newcommand{\dr}{{\partial_r}} \newcommand{\dt}{{\partial_t}}
72: \newcommand{\vg}{{{\bf v(x)} \cdot \nabla}}
73: %
74: %\topmargin -3pt
75:
76: \begin{document}
77: \title{Rapid directional alignment of velocity and magnetic field in
78: magnetohydrodynamic turbulence}
79:
80: \author{W. H. Matthaeus$^1$, A. Pouquet$^2$,
81: P. D. Mininni$^{2,3}$, P. Dmitruk$^1$, and B. Breech$^1$}
82: \affiliation{$1$ Bartol Research Institute and Department of Physics
83: and Astronomy,
84: University of Delaware, Newark DE 18716, U.S.A. \\
85: $2$ NCAR, P.O. Box 3000, Boulder, Colorado 80307-3000, U.S.A. \\
86: $3$ Departamento de F\'\i sica, Facultad de Ciencias Exactas y
87: Naturales, Universidad de Buenos Aires, Ciudad Universitaria, 1428
88: Buenos Aires, Argentina.
89: }
90: \date{\today}
91:
92: \begin{abstract}
93: We show that local directional alignment of the velocity
94: and magnetic field fluctuations occurs rapidly
95: in magnetohydrodynamics for a variety of parameters.
96: This is observed both in direct numerical simulations and in
97: solar wind data.
98: The phenomenon is due to an alignment between the magnetic field and either
99: pressure gradients or shear-associated kinetic energy gradients.
100: A similar alignment, of velocity and vorticity, occurs in the Navier Stokes fluid case.
101: This may be the most rapid and robust
102: relaxation process in turbulent flows, and leads to a local weakening of the nonlinear terms in the small scale vorticity and current structures where alignment takes place.
103: \end{abstract}
104:
105: %\pacs{47.27.ek; 47.27.Ak; 47.27.Jv; 47.27.Gs}
106: \maketitle
107: %\section{Introduction and Background}
108:
109: In magnetohydrodynamic (MHD) turbulence,
110: the fluctuating magnetic field $\bf b$
111: and velocity fluctuation $\bf v$
112: enter on nearly equal footing.
113: One consequence
114: is that the nonlinear MHD
115: equations are in effect linearized
116: when the fluctuation components of the magnetic field (in Alfv\'en
117: speed units) are everywhere
118: equal (or opposite) to the velocity field.
119: Such ``Alfv\'enic'' states have long been thought to be favored in
120: relaxation processes \cite{Woltjer58a}.
121: Near Alfv\'enic states are observed in the
122: solar wind plasma \cite{BelcherDavis71},
123: mostly in the inner heliosphere \cite{BavassanoEA82b}.
124: Global evolution towards the Alfv\'enic state,
125: or ``dynamic alignment,''
126: \cite{DobrowolnyEA80-prl},
127: %\cite{DobrowolnyEA80-prl,GrappinEA82,MattEA83,PouquetEA86},
128: when it occurs, requires many nonlinear eddy turnover times.
129: Here we describe a related, more rapid relaxation process,
130: in which local, directional near-alignment of $\bf v$ and $\bf b$
131: emerges in less than one turnover time,
132: for a wide variety of turbulence parameters. This process
133: need not be associated with global alignment, but rather
134: occurs independently in numerous localized patches.
135:
136: Dynamic alignment competes with other MHD
137: relaxation processes as shown in simulations \cite{TingEA86}
138: %\cite{TingEA86,StriblingMatt91}
139: where, for some parameters, it
140: does not occur, or is incompletely realized.
141: Solar wind observations also show that the degree of Alfv\'enicity
142: tends to decrease with increasing heliocentric distance, in apparent contradiction
143: to the dynamic alignment principle.
144: There are suggestions
145: that directional alignment (a necessary condition
146: for global dynamic alignment)
147: may be more ubiquitous.
148: When MHD relaxation is
149: described by a
150: constrained energy minimization principle \cite{TingEA86},
151: %\cite{TingEA86,StriblingMatt91}
152: cross helicity (Alfv\'enicity) and a magnetic invariant
153: (helicity in three dimensions -- 3D; or mean square flux function in two dimensions -- 2D) are held constant, while
154: energy is minimized. The emergent Euler-Lagrange equations
155: predict final states, in both 2D and 3D,
156: and for all parameters, in which
157: $\bf v$ and $\bf b$
158: are directionally aligned or anti-aligned.
159: This theory is reasonably
160: well confirmed by direct numerical simulations.
161: Evidently, in the long time limit
162: for decaying MHD turbulence, pointwise directional alignment is
163: obtained more easily than is the global Alfv\'enic state.
164: Below we show, using MHD numerical experiments,
165: that local directional alignment is even more robust,
166: occurs more rapidly, and appears locally, in patches.
167: The distribution of alignment angle
168: in the solar wind is shown to be consistent with this picture.
169: This rapid relaxation can be understood by an elementary examination of
170: the MHD equations. While these features appear not to have been fully recognized previously,
171: the situation is in fact analogous to the local emergence of Beltrami flows
172: \cite{PelzEA85} in hydrodynamics.
173:
174: %\section{MHD and Alfv\'enic states}
175: {\it MHD and Alfv\'enic states.}
176: In familiar dimensionless (``Alfv\'enic'') units, the
177: equations of incompressible MHD
178: are
179: \begin{eqnarray}
180: \frac{\partial {\bf v}}{\partial t} + {\bf v \cdot \nabla v} &=&
181: -{\bf \nabla} {\mathcal P} + {\bf j \times b} + \nu \nabla^2 {\bf v}
182: \label{eq:momentum} \\
183: \frac{\partial {\bf b}}{\partial t} + {\bf v \cdot \nabla b} &=&
184: {\bf b \cdot \nabla v} + \eta \nabla^2 {\bf b} ,
185: \label{eq:induc}
186: \end{eqnarray}
187: with ${\bf \nabla \cdot v} = {\bf \nabla \cdot b} = 0$. Here, ${\bf v}$
188: is the velocity field, and ${\bf b}$ is the magnetic field, related to
189: the electric current density ${\bf j}$ by ${\bf \nabla \times b} = {\bf j}$;
190: ${\mathcal P}$ is the pressure. The viscosity $\nu$ and magnetic
191: diffusivity $\eta$ define mechanical and magnetic
192: Reynolds numbers respectively as $R_V = LU/\nu$ and $R_M = LU/\eta$.
193: Here $U = \left<|{\bf v}|^2\right>^{1/2}$,
194: with $\langle \dots \rangle $ denoting a spatial average,
195: and $L$ is a length scale associated with
196: the large-scale flow (integral length scale).
197: The total energy $E = E_v + E_b = \langle |{\bf v} |^2 + |{\bf b}|^2 \rangle/2$,
198: the cross helicity $H_c = \langle {\bf v} \cdot {\bf b} \rangle$, and the
199: magnetic helicity $H_m = \langle {\bf a} \cdot {\bf b} \rangle$ are
200: ideal ($\nu = \eta = 0$) invariants in 3D. Here
201: ${\bf b} = \nabla \times {\bf a}$.
202: Dimensionless measures of the helicities
203: are $\sigma_c = 2H_c/E$ and $\sigma_m = (E_L-E_R)/E_b$,
204: where $E_L$ and $E_R$ are magnetic energy in left- and right-handed magnetic
205: polarizations, respectively.
206:
207:
208: %\section{Simulations}
209: {\it Simulations.}
210: We consider several sets of simulations (see Table I),
211: in which the MHD equations are integrated in a
212: spatially periodic domain of side $2\pi$,
213: using a second order Runge-Kutta method, and
214: either $2/3$-rule dealiased \cite{OrszagPatterson72},
215: or non-dealiased pseudospectral methods.
216: All runs freely decay in time, with no external forcing.
217:
218: The type labelled RAN are $128^3$ incompressible runs, with random broadband
219: initial conditions. Four cases are distinguished
220: by their values of
221: $\sigma_m$
222: and $\sigma_c $, spanning a range of possibilities for relaxation
223: starting from a fully random state.
224:
225: We also employ two other types of initial conditions in which helicities are controlled and the fields are more ordered.
226: OT runs are a generalization
227: of the 2D Orszag-Tang (OT) vortex \cite{OrszagTang79},
228: a standard large scale initial
229: condition for MHD turbulence.
230: In our OT case, initially energies $E_v=E_b=2$,
231: $\sigma_c \approx 0.4$, and $\sigma_m \approx 0$.
232: Another set or runs labelled ABC
233: consists of a parameterized large scale helical
234: flow, an uncorrelated and helical large scale magnetic field,
235: and added noise with energy spectra $\sim k^{-3}\exp [-2(k/k_0)]^2$ at $t=0$, with $k_0=N/6$
236: \cite{1536}.
237: These runs have $E_v = E_b = 0.5$, $\sigma_c \approx 1 \times 10^{-4}$, and
238: $\sigma_m \approx 0.5 $, while numerical resolution and Reynolds numbers
239: vary (see Table I).
240: Finally we analyze a small spatial region near a current
241: sheet in a very high Reynolds number ABC simulation, ABC4 in the table.
242:
243: \begin{table}
244: \caption{\label{table} Parameters in the MHD simulations shown in
245: the figures.
246: RAN, OT, and ABC are described in the text.
247: $N$ is the resolution, $\nu$ and $\eta$ are respectively the kinematic
248: viscosity and magnetic diffusivity, and $\sigma_c$ and $\sigma_m$ reffer
249: to the normalized cross and magnetic helicities defined in the text. For
250: the 2D run, $\sigma_m$ is based on the mean-square flux function.}
251: \begin{ruledtabular}
252: \begin{tabular}{ccccc}
253: Run & $N^3$&$\nu=\eta$ & $\sigma_c$ & $\sigma_m$ \\
254: \hline
255: RAN1 & $128^3$ & $2.5 \times 10^{-3}$ & 0 & 0 \\
256: RAN2 & $128^3$ & $2.5 \times 10^{-3}$ & 0.5 & 0 \\
257: RAN3 & $128^3$ & $2. 5 \times 10^{-3}$ & 0 & 0.5 \\
258: RAN4 & $128^3$ & $ 2.5 \times 10^{-3}$ & 0.5 & 0.5 \\
259: \hline
260: OT1 & $128^3$ & $5\times 10^{-3}$ & 0.4 & 0 \\
261: OT2 & $256^3$ & $1.5\times 10^{-3}$ & 0.4 & 0 \\
262: OT3 & $512^3$ & $7.5\times 10^{-4}$ & 0.4 & 0 \\
263: \hline
264: ABC1 & $128^3$ & $3\times 10^{-3}$ & 0 & 0.5 \\
265: ABC2 & $256^3$ & $1.25\times 10^{-3}$ & 0 & 0.5 \\
266: ABC3 & $512^3$ & $6\times 10^{-4}$ & 0 & 0.5 \\
267: ABC4 & $1536^3$ & $2 \times 10^{-4}$ & 0 & 0.5 \\
268: \hline
269: 2D & $1024^2 $ & $2.5 \times 10^{-4}$ & 0 & 0 \\
270: \end{tabular} \end{ruledtabular} \end{table}
271:
272: %\section{Probability density functions}
273: {\it Probability density functions.}
274: Our main diagnostics are
275: probability density functions (pdfs) of the
276: local cosine of the angle $\theta$ %$\theta_{vb}$
277: between $\bv$ and $\bb$
278: \begin{equation}
279: \cos{\theta} = \cos(\bv,\bb) = \frac{\bv \cdot \bb}{|\bv||\bb|}
280: %\cos{\theta_{vb}} = \cos(\bv,\bb) = \frac{\bv \cdot \bb}{|\bv||\bb|}
281: \end{equation}
282: which are computed for each run.
283:
284: The distribution function for RAN2
285: is shown Fig(\ref{fig1}), at times $t= 0, 0.5, 1.0$ and $2.0$.
286: These distributions are highly peaked near $\cos \theta \approx 1$,
287: %$\cos \theta_{v,b} \sim +1$,
288: much more so than would be needed to account for the cross helicity
289: which is initially $\sigma_c \approx 0.5$,
290: decaying to $\sigma_c = 0.24$ at $t=2.0$.
291: The more peaked curves are for the progressively later times.
292: The results for RAN1, having no helicities, are shown in Fig. \ref{fig2}.
293: Now, the distributions are
294: suppressed near $\cos \theta \approx 0$ and
295: strongly peaked near $\cos \theta \approx \pm 1$
296: %$\cos \theta_{v,b}= \pm 1$
297: indicating an enhanced probability of magnetic and velocity field
298: being very nearly aligned or antialigned.
299: Enhanced directional alignment occurs even when
300: the globally averaged
301: cross helicity is approximately zero.
302: We do not show results for RAN3 and RAN4, with $\sigma_m \approx 0.5$,
303: as the distributions
304: are almost indistinguishable
305: from the corresponding case with $\sigma_m \approx 0$.
306:
307: \begin{figure}[tbh]
308: \centerline{\includegraphics[width=7.5cm]{fig1}}
309: \caption{Pdfs of $\cos \theta$
310: for initial normalized cross helicity $\sigma_c = 0.5$ for Run RAN2.
311: Global normalized cross helicity is 0.24 at t=2.
312: Different lines are for different times (see text).
313: }
314: \label{fig1}
315: \end{figure}
316:
317: \begin{figure}[htb]
318: \centerline{\includegraphics[width=7.5cm]{fig2}}
319: \caption{Pdfs of $\cos \theta$ at times
320: t=0 (dotted), 0.5 (dash-dotted), 1.0 (dashed), 2 (solid)
321: in a 3D simulation
322: $\sigma_c \approx 0$ and $\sigma_m \approx 0$ (Run RAN1).
323: The initial distribution is flat.
324: }
325: \label{fig2}
326: \end{figure}
327:
328: %\begin{figure}
329: %\centerline{\includegraphics[width=6.5cm]{histOT_pdf1}}
330: %\caption{Pdfs of $\cos(\bv,\bb)$ in runs OT1 (solid), OT2 (dotted),
331: % and OT3 (dashed) at the peak of energy dissipation.}
332: %\label{fig3}
333: %\end{figure}
334:
335: \begin{figure}
336: \centerline{\includegraphics[width=7.5cm]{fig3}}
337: \caption{Pdfs of $\cos \theta$ in runs ABC1 (solid), ABC2 (dotted),
338: and ABC3 (dashed) at the peak of energy dissipation and ABC4
339: (dash-dotted).}
340: \label{fig4}
341: \end{figure}
342:
343: The pdfs in the OT runs (not shown)
344: %Fig (\ref{fig3}),
345: are asymmetric and strongly peaked at $\cos \theta \approx 1$, as in
346: the RAN2 and RAN4 cases.
347: For the ABC runs, with no net cross-helicity, the pdfs peak at
348: $\cos \theta \approx \pm 1$ after less than half a turnover time,
349: following the pattern of the RAN runs. Figure \ref{fig4} shows the
350: pdfs from the ABC runs at the peak of dissipation ($t \approx 4$) for
351: different Reynolds numbers.
352:
353: This local alignment process is fast in all cases, with substantial
354: and apparently nearly saturated alignment occurring
355: in less than one large
356: scale turnover time. As stated above, no clear dependence with the
357: Reynolds numbers is seen when we compare cases ABC1-4.
358:
359: When pdfs of $\cos \theta$ are computed in the vicinity of
360: a cluster of strong current sheets, or in regions of strong shear in
361: the magnetic field (run ABC4), an only slightly different result is
362: obtained (Fig. \ref{fig5}). Inside the current sheet, the magnetic
363: and velocity field are strongly antialigned (which gives the peak near
364: $-1$), and the pdf is linear. As larger subvolumes surrounding
365: the current sheet are considered,
366: or at later times when current sheets accumulate and interact,
367: and thus more current sheets with different
368: alignments are integrated in the sub-volume,
369: the pdf converges towards
370: the form seen in Fig. (\ref{fig4}) for the whole flow.
371:
372: \begin{figure}
373: \centerline{\includegraphics[width=7.5cm]{fig4}}
374: \caption{Pdfs of $\cos \theta$ in the vicinity of a current sheet
375: (sub-volume of $150^3$ grid points) in a $1536^3$ simulation with ABC plus noise
376: initial conditions (run ABC4).}
377: \label{fig5}
378: \end{figure}
379:
380: %\section{Physics of alignment}
381: {\it Physics of alignment.}
382: Why does local alignment take place in these simulations? And why is
383: it so fast? Manipulating the MHD equations,
384: Eqs. (\ref{eq:momentum}) and (\ref{eq:induc})
385: in the ideal case ($\nu = \eta = 0$), one finds
386: the equation for evolution
387: of the local cross helicity:
388: \begin{equation}
389: \frac{\partial (\bv \cdot \bb)}{\partial t} + \bv \cdot \nabla (\bv \cdot \bb)
390: = \bb \cdot \nabla \frac{\bv^2}{2} - \bb \cdot \nabla {\cal P} .
391: \label{eq:evolution}
392: \end{equation}
393: The terms on the left are the convective derivative, indicating that
394: $\bv \cdot \bb$ is advected by the velocity field. The terms on the
395: right are divergences: using that $\nabla \cdot {\bf b} = 0$,
396: and when integrated over volume with the proper
397: boundary conditions (e.g. periodic boundaries) they vanish. This expresses
398: the simple fact that the total cross helicity is an ideal invariant in
399: MHD.
400:
401: However, gradients of kinetic energy and pressure gradients affect the
402: {\it local} alignment
403: between the two fields. The first term on the right of Eq.
404: (\ref{eq:evolution}) shows that gradients in the kinetic energy
405: (e.g., shear) can change the alignment between the fields when they
406: are parallel to the magnetic field lines. Indeed, a magnetic field
407: line (which behaves as a material line as follows from Alfv\'en's
408: theorem) tends to be distorted by the shear, and aligned with the
409: velocity field locally. For a planar shear, this would be very similar
410: to what is called field-line stretching.
411: Pressure
412: gradients aligned along magnetic field lines, from the second term on the right, also change the alignment.
413: Where a pressure gradient is present, velocity goes from the region of
414: higher pressure to the region of lower pressure. If the pressure
415: gradient has a projection onto the magnetic field, the resulting
416: velocity field will be aligned as a result with the magnetic field.
417: Moreover, from dimensional analysis we can estimate the time for
418: the local alignment to take place as $\sim b_l/l$, where $b_l$ is the
419: amplitude of the magnetic fluctuations at scale $l$.
420:
421: Note that the induction equation is formally equivalent to the vorticity
422: equation in hydrodynamics. Consequently by the same reasoning, it can be shown that
423: the hydrodynamic
424: velocity and vorticity fields tend to align locally, as found
425: numerically in \cite{PelzEA85} for regions of low dissipation.
426: This, replacing ${\bf b}$ by the vorticity $\mbox{\boldmath $\omega$}= \nabla \times {\bf v}$
427: in Eq. (\ref{eq:evolution}), occurs
428: according to alignment of $\mbox{\boldmath $\omega$}$ with gradients of the kinetic energy and the pressure.
429:
430: \begin{figure}
431: \centerline{\includegraphics[width=6.5cm]{fig5}}
432: \caption{Pdf of $\cos(\bv,\bb)$ from 30 years of
433: Omni data (ISEE, IMP and other satellite data).
434: Also shown is the pdf of $\cos(\bv,\bb)$ from Ulysses spacecraft data
435: between 50 and 59 $^\circ$ heliospheric North latitude during a polar pass
436: in solar minimum conditions.
437: }
438: \label{fig6}
439: \end{figure}
440:
441: %\section{Solar wind observations}
442: {\it Solar wind observations.}
443: Using samples of spacecraft data we computed distributions
444: of the alignment angle for two interplanetary
445: datasets -- the Omni dataset at 1AU near Earth orbit
446: in the ecliptic plane, and a sample of Ulysses data from high heliographic latitude.
447: Fig. (\ref{fig6}) shows the results of these analyses.
448: The low latitude OMNI analysis is divided into intervals in which the
449: large scale interplanetary magnetic field is directed either away from or towards the sun.
450: The net Alfv\'enicity is outward at the higher latitude of the Ulysses sample.
451: In each of these cases, the pdfs of the {\it local} alignment are
452: consistent with the net cross helicity in each sample.
453:
454:
455: \begin{figure}
456: \centerline{\includegraphics[width=6.cm]{fig6}}
457: \caption{Cross-helicity density in a 2D incompressible MHD
458: simulation, showing areas that have values of $\cos(\bv,\bb)< -0.7$ (black),
459: $|\cos(\bv,\bb)|< 0.7$ (gray) and $\cos(\bv,\bb) > 0.7$ (white).
460: Areas having highly aligned or anti-aligned velocity and magnetic
461: field fluctuations dominate the picture. }
462: \label{fig8}
463: \end{figure}
464:
465: %\section{Discussion and conclusions}
466: {\it Discussion and conclusions.}
467: The characteristic pdfs of $\cos \theta$
468: described above cannot be explained as
469: a superposition of two uncorrelated Gaussian distributions for the
470: velocity and magnetic fields, although the pdfs of the velocity and
471: magnetic field themselves are Gaussian (but clearly correlated). Pdfs
472: computed from random broadband uncorrelated Gaussian-component
473: velocity and magnetic fields have a flat $f(\cos \theta)$ distribution.
474: For the coherent ABC flows, $\cos \theta$ peaks at $0$ initially,
475: while for the non-helical RAN1 and RAN3 flows
476: the distribution is initially flat.
477: All cases evolve towards the characteristic shape that is high-peaked
478: at $|\cos \theta| = 1$.
479: In contrast,
480: prior studies have shown that
481: the distribution of the induced {\it emf}, ${\bf v} \times {\bf b}$,
482: is accurately computed from the Gaussian statistics, for
483: both high and low cross helicity, in simulations and in solar wind data
484: \cite{BreechEA03}.
485: What apparently accounts for the difference is that the induced
486: {\it emf} does not correspond to a conserved quantity, while
487: the alignment angle is closely associated with the ideally conserved cross helicity.
488: The {\it emf} can be accounted for using Gaussian statistics, but alignment,
489: even of Gaussian fields, is a dynamical quantity constrained by
490: the local transport and conservation, as implied by Eq. (4).
491:
492: Note that Alfv\'en vortices \cite{PetviashviliPokhotelov92}, which are
493: coherent structures
494: predicted for reduced MHD,
495: have been recently observed in space plasmas \cite{SundkvistEA05};
496: %\cite{SundkvistEA05,AlexandrovaEA06}
497: the generalized Alfv\'en condition
498: obeyed by these vortices corresponds to a local directional alignment.
499: Evidently this type of robust alignment
500: process may be influential in a variety of space
501: and astrophysical plasmas in which turbulent relaxation operates,
502: as well as in the neutral fluid case.
503:
504: We conclude that
505: directional alignment is a rapid and robust process in turbulence.
506: The magnetic and velocity fields respond,
507: as described above, according to the local values of the
508: shear and pressure gradients, essentially independently of the conditions
509: at remote locations, leading to local
510: alignment or anti-alignment; Fig. \ref{fig8} illustrates this localization or patchiness
511: of the directional alignment, using a 2D MHD simulation (see also \cite{sparse}).
512: Since the alignment appears to be an essentially universal and rapid process,
513: it would not be surprising if the coherent small scale structures in turbulence %produced by MHD
514: are associated with it. Indeed, the case shown in Fig. \ref{fig5}
515: is such an example where current sheets are observed
516: to have maximum alignment between the velocity and the magnetic fields \cite{1536}; similarly the local
517: ${\bf v}$-$\mbox{\boldmath $\omega$}$ alignment may explain the slow return to full isotropy in fluid turbulence.
518:
519: %\begin{acknowledgments}
520: Research supported in part by NSF under ATM-0539995 and by NASA
521: under NASA NNG06GD47G.
522: Computer time was provided in part by NCAR. PDM is a member of
523: the Carrera del Investigador Cient\'{\i}fico of CONICET.
524: %\end{acknowledgments}
525:
526: %\bibliographystyle{abbrev}
527: %\bibliography {ag,refs_whm,mp,hl,qz}
528: \newcommand{\SortNoop}[1] {} \newcommand{\au} {{A}{U}\ } \newcommand{\AU}
529: {{A}{U}\ } \newcommand{\MHD} {{M}{H}{D}\ } \newcommand{\mhd} {{M}{H}{D}\ }
530: \newcommand{\RMHD} {{R}{M}{H}{D}\ } \newcommand{\rmhd} {{R}{M}{H}{D}\ }
531: \newcommand{\wkb} {{W}{K}{B}\ } \newcommand{\alfven} {{A}lfv\'en\ }
532: \newcommand{\Alfven} {{A}lfv\'en\ } \newcommand{\alfvenic} {{A}lfv\'enic\ }
533: \newcommand{\Alfvenic} {{A}lfv\'enic\ }
534: \begin{thebibliography}{10}
535:
536: \bibitem{Woltjer58a}
537: L.~Woltjer.
538: %\newblock On hydromagnetic equilibrium.
539: \newblock {\em Proc. Nat. Acad. Sci. USA}, {\bf 44}:833, 1958.
540:
541: \bibitem{BelcherDavis71}
542: J.~W. Belcher and L.~Davis~Jr.
543: %\newblock Large-amplitude \alfven waves in the interplanetary medium, 2.
544: \newblock {\em J.\ Geophys.\ Res.}, {\bf 76}:3534, 1971.
545:
546: \bibitem{BavassanoEA82b}
547: B.~Bavassano, M.~Dobrowolny, G.~Fanfoni, F.~Mariani, and N.~F. Ness.
548: %\newblock Statistical properties of \mhd fluctuations associated with
549: % high-speed streams from {H}elios-2 observations.
550: \newblock {\em Solar Phys.}, {\bf 78}:373, 1982;
551: D. A. Roberts, M. L. Goldstein, L. W. Klein and W. H. Matthaeus,
552: %Origin and Evolution of Fluctuations in the Solar Wind:
553: %{H}elios Observations and {H}elios-{V}oyager Comparisons}
554: {\it J. Geophys. Res.} {\bf 92}, 12\,023 (1987).
555:
556: \bibitem{DobrowolnyEA80-prl}
557: M.~Dobrowolny, A.~Mangeney, and P.~Veltri.
558: %\newblock Fully developed anisotropic hydromagnetic turbulence in
559: % interplanetary space.
560: \newblock {\em Phys.\ Rev.\ Lett.}, {\bf 45}:144, 1980;
561: %\bibitem{GrappinEA82}
562: R. Grappin, U. Frisch, J.~L\'eorat, and A.~Pouquet.
563: %\newblock \alfvenic fluctuations as asymptotic states of \mhd turbulence.
564: \newblock {\em Astron.\ Astrophys.}, {\bf 105}:6, 1982;
565: %\bibitem{MattEA83}
566: W.~H. Matthaeus, M. L. Goldsteinn, and D. Montgomery,
567: \newblock {\em Phys.\ Rev.\ Lett.}, {\bf 51}:1484, 1983.
568:
569: \bibitem{PouquetEA86}
570: A.~Pouquet, U.~Frisch, and M.~Meneguzzi.
571: %\newblock Growth of correlations in magnetohydrodynamic turbulence.
572: \newblock {\em Phys.\ Rev.\ A}, {\bf 33}:4266, 1986.
573:
574: \bibitem{TingEA86}
575: A.~C. Ting, W.~H. Matthaeus, and D. Montgomery.
576: %\newblock Turbulent relaxation processes in magnetohydrodynamics.
577: \newblock {\em Phys.\ Fluids}, {\bf 29}:3261, 1986;
578: %\bibitem{StriblingMatt91}
579: T.~Stribling and W.~H. Matthaeus.
580: %\newblock Relaxation processes in a low order three-dimensional
581: % magnetohydrodynamics model.
582: \newblock {\em Phys.\ Fluids B}, {\bf 3}:1848, 1991.
583:
584: \bibitem{PelzEA85}
585: R.~ Pelz et al. %, V.~{Yakhot}, S.~A. {Orszag}, L.~{Shtilman}, and E.~{Levich}.
586: %\newblock {Velocity-vorticity patterns in turbulent flow}. \newblock
587: {\em Phys. Rev. Lett.} {\bf 54}:2505, 1985;
588: H.K. Moffatt, {\it J. Fluid Mech.} {\bf 150}, 359 (1985);
589: A. Tsinober and E. Levich, {\it Phys. Lett.} {\bf 99A}, 321 (1983);
590: E. Levich, {\it Phys. Rep.} {\bf 151}, 129 (1987);
591: M. Farge, G. Pellegrino, and K. Schneider, {\it Phys. Rev. Lett.}
592: {\bf 87}, 054501 (2001)
593: %A. Alexakis, P.~Mininni \& A.~Pouquet.
594: %2006: ``Large scale flow effects, energy transfer, and self-similarity in turbulence,''
595: %arXiv:physics/0602148,
596: % {\it Phys. Rev. E} {\bf 74}:056320, 2006.
597:
598: \bibitem{OrszagPatterson72}
599: S.~A. Orszag and G.~S. Patterson.
600: %\newblock Numerical simulation of three dimensional homogeneous isotropic
601: % turbulence.
602: \newblock {\em Phys.\ Rev.\ Lett.}, {\bf 28}:76, 1972.
603:
604: \bibitem{OrszagTang79}
605: S.~A. Orszag and {C}-{M} Tang.
606: %\newblock Small-scale structure of two-dimensional magnetohydrodynamic
607: % turbulence.
608: \newblock {\em J.\ Fluid Mech.}, {\bf 90}:129, 1979.
609:
610: \bibitem{1536}
611: P. Mininni, A. Pouquet and D. Montgomery,
612: %``Small-scale structures in three-dimensional magnetohydrodynamic turbulence,''
613: {\it Phys. Rev. Lett.} {\bf 97}, 244503, 2006; P. Mininni and A. Pouquet, preprint (2007), see
614: http://arxiv.org/abs/0707.3620v1.
615:
616: \bibitem{BreechEA03}
617: B.~Breech, W.~H. Matthaeus, L.~J. Milano, and C.~W. Smith.
618: %\newblock Probability distributions of the induced electric field of the solar
619: % wind.
620: \newblock {\em J.\ Geophys.\ Res.}, {\bf 108}:1153, doi:1029/2002JA009529,
621: 2003.
622:
623: \bibitem{PetviashviliPokhotelov92}
624: V.I. Petviashvili and O. A. Pokhotelov, in
625: Solitary Waves in Plasmas and in the Atmosphere, Gordon and Breach, New York
626: (1992).
627:
628: \bibitem{SundkvistEA05}
629: D.~{Sundkvist} et al.,
630: %, V.~{Krasnoselskikh}, P.~K. {Shukla}, A.~{Vaivads},
631: % M.~{Andr{\'e}}, S.~{Buchert}, and H.~{R{\`e}me}.
632: %\newblock {In situ multi-satellite detection of coherent vortices as a
633: % manifestation of Alfv{\'e}nic turbulence}.
634: \newblock {\em \nat}, 436:825, 2005;
635: %\bibitem{AlexandrovaEA06}
636: O. Alexandrova et al.,
637: %, A. Mangeney, M. Maksimovic,
638: % N. Cornilleau-Wehrlin, J.-M. Bosqued, and M. Andr\'e,
639: \newblock {\em J. Geophys. Res.}, 111:A12208, doi:10.1029/
640: 2006JA011934, 2006.
641:
642: \bibitem{sparse}
643: M. Meneguzzi, H. Politano, A. Pouquet and M. Zolver, {\it J. Comp. Phys.} {\bf 123}, 32, 1996.
644:
645: \end{thebibliography}
646:
647: \end{document}
648:
649: \begin{figure}
650: \centerline{\includegraphics[width=7.5cm]{vb_hist_mhd3d_cases.eps}}
651: \caption{Pdfs of $\cos(\bv,\bb)$ at $t=1.0$ for four runs:
652: I: ($\sigma_c = 0$, $\sigma_m = 0$, as in Fig 1,
653: II: $\sigma_c = 0.4$, $\sigma_m = 0$
654: III:$\sigma_c = 0$, $\sigma_m = 0.4$
655: IV: $\sigma_c = 0.4$, $\sigma_m = 0.4$
656: }
657: \label{fig3}
658: \end{figure}
659: