1: %\documentclass[aps,preprint,groupedaddress,amsmath,amssymb,floatfix]{revtex4}
2: \documentclass[pre,twocolumn,amsmath,amssymb,floatfix]{revtex4}
3: %\documentclass[preprint,preprintnumbers,amsmath,amssymb]{revtex4}
4:
5: \usepackage{graphicx}% Include figure files
6: \usepackage{dcolumn} % Align table columns on decimal point
7: \usepackage{bm} % Bold math
8: \usepackage{color}
9:
10: % COLOR %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
11: \def\NOTE#1{{\textcolor{red}{\bf [#1]}}} % note
12: \def\DEL#1{{\textcolor{red}{[#1]}}} % suggested deletions
13: \def\ADDA#1{{\textcolor{blue}{#1}}} % addition (Referee 1)
14: \def\ADDB#1{{\textcolor{green}{#1}}} % addition (Referee 2)
15: \def\AD#1{{\textcolor{magenta}{#1}}} % change
16: %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
17:
18: %\topmargin -3pt
19: \begin{document}
20:
21: \title{Effect of helicity and rotation on the free decay of turbulent flows}
22: \author{T. Teitelbaum$^{1}$ and P.D. Mininni $^{1,2}$}
23: \affiliation{$^1$ Departamento de F\'\i sica, Facultad de Ciencias Exactas y
24: Naturales, Universidad de Buenos Aires, Ciudad Universitaria, 1428
25: Buenos Aires, Argentina. \\
26: $^2$ NCAR, P.O. Box 3000, Boulder, Colorado 80307-3000, U.S.A.}
27: \date{\today}
28:
29: \begin{abstract}
30: The self-similar decay of energy in a turbulent flow is studied in
31: direct numerical simulations with and without rotation. Two initial
32: conditions are considered: one non-helical (mirror-symmetric), and
33: one with maximal helicity. The results show that, while in the absence
34: of rotation the energy in the helical and non-helical cases decays with
35: the same rate, in rotating flows the helicity content has a major impact on
36: the decay rate. These differences are associated with differences in the
37: energy and helicity cascades when rotation is present. Properties of the structures that arise in the flow at late times in each time are also discussed.
38: \end{abstract}
39: \maketitle
40:
41: %\section{Introduction}
42:
43: Turbulence is ubiquitous in nature, and many turbulent flows are also
44: rotating. The effect of rotation becomes important when the Rossby number
45: (the ratio of the convective to the Coriolis acceleration) is sufficiently
46: small. Mid-latitude synoptic scales in the atmosphere, stellar convective
47: regions, and turbomachinery are examples of such flows. Helicity (alignment of the velocity and the vorticity) is also important
48: for many processes in astrophysical, geophysical and engineering flows.
49: As an example, helical flows were proposed as the reason for the
50: stability of rotating convective thunderstorms \cite{Lilly86}.
51:
52: Studies of isotropic and homogeneous helical turbulence
53: \cite{group1,Kraichnan73} showed that both the helicity
54: and the energy are transferred toward smaller scales with
55: constant fluxes. Moreover, it was observed that the scaling of the
56: energy (Kolmogorov's law) was unchanged by the presence of helicity.
57: As a result, helicity is expected to globally arrest the
58: energy decay, but not to change its self-similar decay rate. Non-helical
59: rotating turbulence has been studied in detail (see e.g.,
60: \cite{group2} and references therein), but perhaps because
61: of the similarities between the helical and non-helical cases in
62: non-rotating flows, not much attention has been paid to helical rotating
63: turbulence.
64:
65: The lack of detailed studies of rotating helical flows is remarkable
66: considering the relevance of both helicity and rotation in many
67: astrophysical and geophysical processes. In this work, we study the
68: effect of rotation and of helicity in the self-similar decay of energy
69: in turbulent flows. Even in isotropic and homogeneous turbulence, the
70: law for the decay rate of energy is a matter of debate \cite{group5}.
71: It is known that it depends on properties of the infrared energy spectrum
72: (i.e., the spectrum at scales larger than the energy containing scale),
73: and may depend on other statistical properties of the initial conditions.
74: As a result, in this letter we will consider only two flows with the same
75: infrarred spectrum and the same energy decay rate in the absence of
76: rotation, and study how the presence of helicity and rotation changes
77: their decay. New decay laws are found for helical rotating flows in numerical simulations.
78: The results are then interpreted in terms of how the
79: energy and helicity cascades are modified when rotation is present,
80: and a phenomenological theory that is in agreement with the simulations
81: is finally discussed.
82:
83: %\section{Numerical simulations}
84:
85: The numerical simulations solve the Navier-Stokes equations for an
86: incompressible fluid in a rotating frame,
87: \begin{equation}
88: \partial_t {\bf u} + \mbox{\boldmath $\omega$} \times
89: {\bf u} + 2 \mbox{\boldmath $\Omega$} \times {\bf u} =
90: - \nabla {\cal P} + \nu \nabla^2 {\bf u} ,
91: \label{eq:momentum}
92: \end{equation}
93: where ${\bf u}$ is the velocity field ($\nabla \cdot {\bf u} =0$),
94: $\mbox{\boldmath $\omega$} = \nabla \times {\bf u}$ is the vorticity,
95: ${\cal P}$ is the total pressure (modified by the centrifugal term),
96: and $\nu$ is the kinematic viscosity. We chose the rotation axis to be
97: in the $z$ direction, $\mbox{\boldmath $\Omega$} = \Omega \hat{z}$,
98: with $\Omega$ the rotation frequency. Our integration domain is a periodic
99: box of length $2\pi$. Two sets of runs were done at resolutions of $256^3$
100: (set A) and $512^3$ grid points (set B) using a pseudo spectral code. The parameters for all the runs
101: are listed in Table \ref{table:runs}.
102:
103: \begin{table}
104: \caption{\label{table:runs}Parameters used in the simulations: $\nu$ is
105: the kinematic viscosity, $\Omega$ is the rotation rate, $Re$ is the
106: Reynolds number, $Ro$ is the Rossby number, $Ro^{\omega}$ is the
107: micro-Rossby number, $E_k$ is the Ekman number, and $h$ is the relative
108: helicity of the initial conditions. The values of $Re$, $Ro$, $Ro^{\omega}$,
109: and $E_k$ are given at the time of the peak of dissipation $t^*$.}
110: \begin{ruledtabular}
111: \begin{tabular}{ccccccccc}
112: Run&$\nu$&$\Omega$&$Re$&$Ro$&$Ro^\omega$&$E_k$ &$h$ \\
113: \hline
114: A1 &$1.5\times 10^{-3}$& $0$ & $450$ & $-$ & $-$ &
115: $-$ &$0$ \\
116: A2 &$1.5\times 10^{-3}$& $0$ & $600$ & $-$ & $-$ &
117: $-$ &$0.95$ \\
118: A3 &$1.5\times 10^{-3}$& $4$ & $550$ & $0.12$ & $1.28$ &
119: $2.2\times 10^{-4}$ &$0$ \\
120: A4 &$1.5\times 10^{-3}$& $4$ & $830$ & $0.083$ & $0.8$ &
121: $1.0\times 10^{-4}$ &$0.95$ \\
122: B1 &$7\times 10^{-4}$& $4$ & $1100$ & $0.12$ & $1.82$ &
123: $1.1\times 10^{-4}$ &$0$ \\
124: %B2 &$7\times 10^{-4}$& $4$ & $?$ & $?$ & $$ &
125: % $?$ &$0.4$ \\
126: B2 &$7\times 10^{-4}$& $4$ & $1750$ & $0.083$ & $1.15$ &
127: $4.7\times 10^{-5}$ &$0.95$ \\
128: \end{tabular}
129: \end{ruledtabular}
130: \end{table}
131:
132: To simulate systems with different amount of relative initial helicity
133: ($h = H/\left<|{\bf u}||\mbox{\boldmath $\omega$}|\right>$ where
134: $H=\left<{\bf u}\cdot\mbox{\boldmath $\omega$}\right>$ is the flow net
135: helicity), two flows were considered as initial conditions: the
136: Taylor-Green (TG) flow \cite{Taylor37}, and the Arn'old-Beltrami-Childress
137: (ABC) flow \cite{Childress}. The TG flow is non-helical, and has zero
138: energy in the $k_z=0$ mode, whose amplification observed in the rotating
139: cases (see below) is thus only due to a cascade process. The TG flow was
140: chosen for its importance in hydrodynamics; it was originally motivated
141: as an initial condition which leads to rapid development of small spatial scales.
142: It also mimics the von K\'arm\'an flows between two counter-rotating
143: disks used in several experiments. % \NOTE{of rotating turbulence \cite{Simand00}}.
144: The ABC flow is an eigenfunction of the curl operator and as a result
145: has maximum helicity. It was used as a paradigmatic example to study helical
146: flows in the atmosphere \cite{Lilly86}. Both flows develop, after a
147: short time, an infrared energy spectrum proportional to $k^2$; this is
148: important to ensure we can compare the decay rates of both flows.
149: %\NOTE{equal foot}.
150:
151: The simulations were started using a superposition of these flows
152: from wavenumbers $k=4$ to $14$. Runs with zero relative helicity
153: have a supperposition of TG flows and runs with $h\approx 0.95$ have
154: a supperposition of ABC flows. %and run B2 with $h\approx 0.4$ has a
155: %mixture of both.
156: The initial energy spectrum from $k=4$ to $14$ was
157: proportional to $k^{-4}$, leaving enough spectral space
158: to allow for direct and inverse transfer of energy. All the runs were
159: extended for over 40 turnover times,
160: and the dissipative range was properly resolved. Times in
161: Table \ref{table:runs} and in the figures are expressed in units of the
162: turnover time at $t=0$, $T=L/U$, where $L\approx 2\pi/k_0 = 2\pi/4$
163: is the initial integral scale of the flow and $U\approx 1$ is the initial
164: r.m.s. velocity.
165:
166: Several Reynolds, Rossby, and Ekman numbers can be defined for the runs.
167: We will consider here the Reynolds number based on the integral scale
168: $Re = UL/\nu$, and the accompanying Rossby number
169: $Ro = U/(2 \Omega L)$. The integral scale is defined as
170: $L = 2\pi \int{E(k) k^{-1} dk}/E$ where $E(k)$ is the isotropic energy
171: spectrum and $E$ is the total energy. The Ekman number is then defined
172: as $E_k=Ro/Re$. At $t=0$ runs in set A have $Re = 1047$ and runs in set
173: B have $Re = 2244$. At the same time the Rossby number is $Ro = 0.08$
174: for runs A3, A4, and B1-B2; the initial Ekman number is then $E_k = 7.6 \times 10^{-5}$
175: in runs A3 and A4, and $E_k = 3.6 \times 10^{-5}$ in runs B1-B2. Values of the
176: controlling parameters for each run at the time of maximum dissipation $t^*$
177: are given in Table \ref{table:runs}.
178:
179: It is also convenient to introduce a micro-scale
180: Rossby number as $Ro^{\omega} = \omega/(2 \Omega)$ \cite{Jacquin90}.
181: This number can be interpreted as the ratio of the convective to the
182: Coriolis acceleration at the Taylor scale, a scale characteristic of
183: the turbulent inertial range. In order for anisotropies to develop in the
184: simulations, the Rossby number $Ro$ must be small enough for the rotation
185: to affect the free decaying turbulence, but the micro-Rossby
186: $Ro^{\omega}$ must be larger than one for scrambling effects of inertial
187: waves not to completely damp the nonlinear terms (which leads to a pure
188: exponencial viscous decay) \cite{Cambon97}. We are
189: thought interested in simulations with moderate rotation rates to ensure
190: $Ro \lesssim 1$ and with $Re$ large enough to have enough scale
191: separation between the energy-containing and the dissipative scale. Note
192: in Table \ref{table:runs} how $Ro$ and $Ro^{\omega}$ are one order of
193: magnitude apart in all runs at time $t^*$. This difference
194: is sustained in time in the runs during the self-similar energy decay.
195:
196: %\section{\label{sec:evol}Time evolution without rotation}
197:
198: \begin{figure}.
199: \includegraphics[width=8cm]{fig1}
200: \caption{Energy evolution for runs A1 (solid), A2 (dotted),
201: B1 (dashed) and B2 (dot-dashed). Runs A1 and A2 have $\Omega=0$ and decay with the same rate
202: independently of the helicity content. When rotation is turned on the decay rates differ.
203: Inset: enstrophy evolution for runs A1 and A2.}
204: \label{fig:time1} \end{figure}
205:
206: Figure \ref{fig:time1} shows the time history of the energy and enstrophy
207: in runs A1, A2, B1-B2. In all runs there is a self-similar decay after
208: the time $t^*$. Runs A3 and A4 (not shown) decay
209: respectively as runs B1 and B2, although the time span of the self-similar
210: stage is shorter. In both runs with $\Omega=0$, the energy decays as
211: $\sim t^{-2}$. As will be discussed later, this
212: exponent corresponds to the decay of a flow with constant integral length
213: \cite{Biferale03}.
214:
215: The presence of helicity in run A2 doesn't seem to affect the self-similar
216: decay of turbulence, as predicted in Ref. \cite{Kraichnan73}. This is in
217: good agreement with the fact that helicity doesn't change the spectral
218: index of the energy in non-rotating turbulence \cite{group1}.
219: However, the self-similar decay in run A2 starts at a later time, as also
220: previously reported in Ref. \cite{Morinishi01}. This is associated with the
221: slow-down in the generation of small-scales in helical flows
222: \cite{Lesieur}, which results in a longer time to reach the maximum of
223: dissipation in run A2 (see the inset in Fig. \ref{fig:time1}).
224:
225: %\section{\label{sec:transfer}Time evolution with rotation}
226:
227: In runs with rotation (A3, A4, B1 and B2), a transient is also observed
228: before $t^*$. Then, self-similar decays with different power
229: laws for the energy are found in all runs. Runs A3 and B1 have a decay
230: near $t^{-1}$ \cite{group3,bellet}, while the runs with maximum helicity
231: (A4 and B2) follow a decay slightly faster than $t^{-1/3}$. % A run
232: %with intermediate relative helicity (run B2) lies between these two
233: %extreme cases.
234: Contrary to the non-rotating case, we can appreciate now
235: how the presence of initial helicity changes the self-similar energy decay.
236:
237:
238: %{\bf The time evolution of the enstrophy for runs B1 and
239: %B2 is also shown in the inset of Fig. \ref{fig:time1} where wee see again
240: %how the maximum of enstrophy is reached at differents times for each flow.}
241: %{\bf [esta ultima oracion sacarla o dejarla pero no poner nada en le inset]}
242:
243: \begin{figure}
244: \includegraphics[width=8cm,height=4cm]{fig2a}
245: \includegraphics[width=8cm,height=4cm]{fig2b}
246: \caption{(a) Isotropic energy spectrum for run B2 at times $t \approx 2$ (thin), $t \approx 12$ (thick) and $t \approx 30$ (dot-dashed). At early times the spectrum follows a $\sim k^{-2.5}$ power-law (slope shown as a reference) and becames later dominated by the piling up of energy in the large scales. Inset: Isotropic energy spectrum for run B2. A $k^{-2}$ slope is plotted for reference.
247: (b) Energy (thin) and helicity (thick) normalized fluxes for run B2 at times $t \approx 5$ (solid) and $t \approx 13$ (dashed) respectively. %Observe how a direct helicity cascade dominates at early times and the coexistence of an inverse energy and a direct helicity flux for later times.
248: Inset: Energy flux for run B1 at times $t \approx 2$ (solid) and $t \approx 11$ (dashed).}
249: \label{fig:espectros} \end{figure}
250:
251: The time evolution is accompanied by a change in the shape of the energy
252: spectrum (see Fig. \ref{fig:espectros}(a)). While the runs without rotation
253: show at $t^*$ and during the self-similar decay an
254: spectrum consistent with Kolmogorov scaling in both the energy and the
255: helicity, runs with rotation show different scaling laws. Run B1 (which
256: has a larger scale separation than run A3) shows an isotropic energy
257: spectrum compatible with $k^{-2}$ (the anisotropic energy spectrum is
258: also compatible with a $k_\perp^{-2}$ scaling) \cite{muller,MininniPress}. On the
259: other hand, the energy spectrum in run B2 is slightly steeper than in
260: run B1, while the helicity spectrum in this run is shallower than the
261: energy spectrum. The product of both spectra is consistent with a
262: $k^{-4}$ scaling law, as predicted using phenomenology for rotating
263: helical flows in \cite{MininniPress}.
264:
265: In the runs with rotation, a change in the small-scale spectrum is
266: observed at $t \approx 20$. As energy piles up at the largest available
267: scale in the box and column-like structures form in the velocity, the
268: small-scale energy spectrum becomes steeper. This is accompanied by a
269: decrease in the decay rate of the energy (see Fig. \ref{fig:time1}). This
270: process is reminiscent of the change observed in the free-decay of
271: two-dimensional turbulence, when the coalescence of large-scale vortices
272: at late times leads to a steeper energy spectrum and a change in the
273: self-similar decay \cite{McWilliams84}
274:
275: %In run B2 the isotropic energy spectrum also develops an early power law behavoir with exponent $-2$ ($t \sim 2$) which changes to $-2.5$ some turn over times later ($t \approx 10$) toghether with a pile up of energy at large scales as for the TG case (citas). At time $t^* \approx 20$ the largest scales dominates and a steeper spectrum is observed (not shown) ($ \approx -2.7$). The predominance of large structures is observed in both B1 and B2 \NOTE{eventualmente B2} runs and can also be evidenciated by the slow down of the energy decay observed in Figure \ref{fig:timeer} for $t>20$ (solid) and $t>10$ (dashed). In section "anisotropy and strucutres" we will find more evidence of this coalescence of ... and we will also visualize the flows and actually observe the presence of few big structures (column-shaped vortical structures) for late times.
276: %Later on, the spectrum is a power law with exponent $< \approx 3$ and the largest scales are still dominant.
277:
278: %{\bf Flux:}
279:
280: %\begin{figure}
281: %\includegraphics[width=8cm]{flujos_1.eps}
282: %\caption{Energy (thinn) and helicity (thick) normalized fluxes for run B2 at times $t \approx 5$ (solid) and $t \approx 13$ (dashed) respectively. Observe how a direct helicity cascade dominates at early times and the coexistence of both, inverse energy and direct helicity fluxes for later times. Inset:Energy flux for run B1 at times $t \approx 1.7$ (solid) and $t \approx 11$ (dashed).}
283: %\label{fig:flujos} \end{figure}
284:
285: Runs A1 and A2 developt direct energy fluxes toward small-scales
286: while the rotating non-helical runs (A3 and B1) show both a direct and an
287: inverse energy cascade (see the inset of Fig. \ref{fig:espectros} (b)).
288: In the helical runs (Fig. \ref{fig:espectros} (b)), at $t \approx 2$ we observe maximum flux
289: of energy and helicity toward smaller scales, evidenciating both energy
290: and helicity have a direct cascade. However, the helicity flux is larger
291: than the direct energy flux. Later, an inverse cascade of energy can be
292: clearly identified from the negative energy flux at large scales. At
293: $t \approx 13$, the coexistence of both an inverse cascade of energy and
294: direct cascades of energy and helicity is observed.
295:
296:
297: \begin{figure}
298: \includegraphics[width=8cm]{fig3}
299: \caption{Ratio of the energy in modes with $k_{||}=0$ to the total energy in runs B1 (dashed) and B2 (solid). Despite the different initial conditions, the curves grow monotonically to 1, showing a transfer of energy to perpendicular modes. Inset: time evolution of the Shebalin angle for the same runs.}
300: \label{fig:ani1} \end{figure}
301:
302: %In all runs, as the time of saturation of the energy at the largest
303: %available scale and of coalescence of columnar structures is reached
304: %($t \approx 20$), the inverse energy flux at the largest scales
305: %becomes positive bouncing back to smaller scales.
306:
307: %\section{\label{sec:anis} anisotropy and structures}
308:
309: %\begin{figure}
310: %\includegraphics[width=8cm]{ani2.eps}
311: %\caption{Evolution of the ratio of perpendicular to parallel integral scales for runs B1 (dashed) and B2 (solid). The %inset shows the time evolution of the parallel integral scale for the same runs.}
312: %\label{fig:ani2} \end{figure}
313:
314: Although Figs. \ref{fig:espectros} (a) and (b) show the
315: isotropic spectra and fluxes, the flows in the rotating case are
316: anisotropic. Most of the energy in the spectra is in modes perpendicular
317: to the axis of rotation, and the anisotropic spectum $E(k_\perp)$ and
318: flux $\Pi(k_\perp)$ look similar to the ones previously discussed.
319: We present instead some global indications of the development of anisotropies.
320: Figure \ref{fig:ani1} shows the ratio of the energy
321: in perpendicular modes (i.e., modes with $k_{||}=0$) to the total energy
322: in runs B1 and B2. Differences at early times are due to differemt
323: initial conditions. However, it can be seen that both curves grow
324: monotonically to a value near 1, indicating the flows evolve toward
325: anisotropic states as the energy is transferred to perpendicular modes
326: \cite{group2}.
327:
328: %Note also that due to the fourier descomposition of the initial flows used (TG and ABC) $E_{\perp}/E$ starts at $0$ for B1 and at aproximately $2/3$ for B2.
329:
330: A measure of anisotropies in the small-scale fluctuations is given
331: by the Shebalin angle, defined as
332: \begin{equation}
333: \tan^2\theta=2 \frac {\sum_{k=1}^{k_{max}}{k_{\perp}^2E(k_{\perp})}}{\sum_{k=1}^{k_{max}}{k_{\parallel}^2E(k_{\parallel})}},
334: \label{eq:shebalin}
335: \end{equation}
336: The evolution of this angle in runs B1 and B2 can be seen in the inset of
337: Fig. \ref{fig:ani1}. It grows monotonically although in run
338: B2 it reaches a maximum at $t \approx 12$ and then seems to saturate. A
339: similar behavior is observed for $L_{\parallel}$ (not shown) which again grows monotically in
340: B1 but reaches a maximum in B2 at aproximately the same time. In all cases, $\tan^2\theta>>2$ which corresponds to anisotropic flows.
341:
342: The increase of the correlation lengths, together with the growth of
343: $E(k_{||}=0)/E$ and of $\tan^2\theta$ speaks of anisotropization
344: of the flows %as time evolves and the inverse cascade of energy develops.
345: This tendency towards two-dimensionalization is confirmed by exploring the flows in real space. Fig. \ref{fig:columnas} shows visualizations of the r.m.s. vorticity
346: with superimposed velocity field lines for runs A3 and A4 at late times
347: ($t \approx 45$). In both cases, a strong anisotropy is observed
348: with large scale column-like structures in the vorticity (similar
349: structures are observed if energy density is visualized
350: instead). However, a significant difference in the geometry of the flow in the columns
351: is observed between runs A3 and A4 . While the flow in
352: the columns of the helical run is strongly helical %\NOTE{h=0.9;hr=1 }
353: (the flow goes either up or down in the entire column, thus giving helical
354: velocity field lines), the columns of the non-helical run have no net
355: helicity (the flow goes up and down inside the same column). This indicates
356: that even at late times, the properties of the emerging structures in
357: rotating flows depend on the initial helicity content.
358:
359:
360: \begin{figure}
361: \includegraphics[width=8cm]{fig4}
362: \caption{(Color online) Visualizations of the r.m.s. vorticity for $t \approx 45$ with superimposed velocity field lines for runs A3 (right) and A4 (left). Cuts of $v_z$ and $H$ on a plane perpendicular to \mbox{\boldmath $\Omega$} are shown. The color table gives the amplitudes normalized to the maximum of each quantity.}
363: \label{fig:columnas} \end{figure}
364:
365:
366:
367:
368: %\section{\label{sec:scaling}Phenomenology}
369:
370: The results presented so far show a novel and distinct evolution in the
371: free decay of helical and non-helical flows when rotation is present.
372: These differences can be understood in terms of a simple phenomenological
373: theory. For the non-rotating cases we make use of the classical Kolmogorov
374: phenomenology which leads to the well known energy spectrum
375: $E(k) \sim \epsilon^{2/3}k^{-5/3}$, where $\epsilon$ represents the energy
376: dissipation rate. From the Navier-Stokes equation it is easy to show that for a freely-decaying flow
377: $dE/dt \sim E^{3/2}/L$ which leads to a self-similar decay law
378: $E(t) \sim t^{-2}$ if the integral scale is assumed constant \cite{Biferale03,group5}.
379:
380: % we have seen that an energy cascade dominates the flux over the inertial range. Considering the energy flux in the inertial range $\epsilon$ is slowed down by $\epsilon \sim \frac{u_l^2}{t_l} (\frac{t_{\Omega}}{t_l}) $ where $u_l$ is the velocity at the scale $l$, $\tau_l \sim l/u_l$ is the eddy turnover time at the same scale and $\tau_{\Omega} \sim 1/\Omega$ is the characteristic time of inertial waves. This expression takes into account the slowing-down of transfer to small scales due to three-wave interactions(referencias:Iroshnikov and Kraichnan, [25]y[26] de "scale interactions and scaling laws in rotating flows at moderate..").
381:
382: In the case of rotating non-helical flows, it is often assumed that
383: $E(k) \sim \epsilon^{1/2}\Omega^{1/2}k^{-2}$ (for simplicity, we use the isotropic
384: wavenumber $k$, although the arguments here can be easily generalized to the
385: anisotropic case replacing $k$ by $k_\perp$). This spectrum was observed
386: in simulations \cite{muller}, and obtained from closures \cite{bellet}. It can
387: also be derived from phenomenological arguments assuming the inertial
388: waves slow down the energy cascade \cite{Zhou95}. Using this spectrum it
389: follows that $dE/dt \sim (E/L)^{2}$ resulting in a decay law
390: $E(t) \sim t^{-1}$.
391:
392: The novel case of rotating helical flows differs from the previous two cases.
393: The direct transfer is dominated by the helicity cascade. Writting the
394: helicity flux as $\delta \sim h_l/(\Omega \tau_l^2)$ where $h_l$ is the
395: helicity at the scale $l$ and $\tau_l$ the eddie turnover time, leads
396: to spectra $E(k) \sim k^{-n}$ and $H(k) \sim k^{n-4}$, where $n=5/2$ for the
397: case of maximum helicity \cite{MininniPress}. In this case,
398: dimensional analysis leads to $E(t) \sim \epsilon^{1/4}\Omega^{5/4}k^{-5/2}$
399: and to $E(t) \sim t^{-1/3}$. Note that for flows with initial helicity
400: between zero and the maximum, the decay rate is between $-2$ and $-1/3$.
401: A run with relative helicity of $ h \approx 0.4$ was done to verify this.
402:
403: %-> increase of the correlation lengths along the axis of rotation associated with a two-dimensionalization process (citar experimentos y simulaciones que aparecen en paper de Bourouiba & P Bartello (pag 140). Fijarme si los papers que tengo muestran esta longitud paralela).
404:
405: %-> Fijarse: se forman ciclones o anti-ciclones (alineacion del vector vorticidad con el eje de rotacion. Rompimiento de simetria?leer Metais & Lesieur (1994))
406:
407: %\section{Conclusions}
408: %- Las estructuras observadas se generan para rotacion no muy grade.
409: %- 2-dimensionalizacion del flujo
410: %- Fenomenologia se condice con las simulaciones
411:
412: % It is interesting to point out that
413: %although both the energy and the helicity direct cascade, the direct
414: %helicity flux is expected to dominate the direct range as the Rossby
415: %number $Ro$ is decreased \cite{Mininni08}. It has been observed in
416: %simulations of forced rotating helical turbulence that the sum of the
417: %difference between the direct and inverse energy and helicity fluxes
418: %remains constant independly of $Ro$, meaning that the amount of direct
419: %helicity flux is associated with the energy flux to smaller wave-lenths.
420: %Increasing the rotation rate $\Omega$ more energy transfer to small
421: %scales would developt and the helicity direct flux would become more
422: %dominant.
423:
424: The results indicate that helical and non-helical rotating flows differ in their scaling laws and decay rates. This is different than the case of non-rotating flows where helicity doesn't change the energy scaling. The observed differences indicate that helical rotating flows must be studied in detail, and open new directions in the study of rotating turbulence. This can also be the starting point to elucidate the actual role helicity plays in the decay of flows in nature where rotation is often relevant \cite{Lilly86}.
425:
426:
427: %\begin{acknowledgments}
428:
429: {\it Computer time provided by NCAR and CECAR. The authors acknowledge support from grant UBACYT X468/08. PDM is a member of the Carrera del Investigador Cient\'{\i}fico of CONICET. Flow visualizations were done using VAPOR \cite{vapor}.}
430:
431: %\end{acknowledgments}
432:
433: %\bibliographystyle{acm}
434: %\bibliography{ms}
435:
436: \begin{thebibliography}{10}
437:
438: \bibitem{Lilly86}
439: {D.K. Lilly, Atm. Sc. {\bf 40}, 126 (1986).}
440:
441: \bibitem{group1}
442: {V. Borue and S.A. Orszag, Phys. Rev. E {\bf 55}, 7005 (1997); Q. Chen, S. Chen, and G.L. Eyink, Phys. Fluids {\bf 15}, 361 (2003); P.~D. Mininni, A. Alexakis, and A. Pouquet, Phys.\ Rev.\ E {\bf74}, 016303 (2006).}
443:
444: \bibitem{group2}
445: {C. Cambon and L. Jacquin, J. Fluid Mech. {\bf 202}, 295 (1989); F. Waleffe, Phys.\ Fluids A {\bf5}, 677 (1993); C. Cambon, N.N. Mansour, and F.S. Godeferd, J. Fluid Mech. {\bf 337}, 303 (1997).}
446:
447: \bibitem{group5}
448: {P.G. Saffman, Phys. of Fluids {\bf 10}, 1349 (1967); U. Frisch, {\it Turbulence: The legacy of A.N. Kolmogorov} (Cambridge Univ.\ Press, Cambridge, 1995); L. Skrbek and S.R. Stalp, Phys. Fluids {\bf 12}, 1997 (2000); P.A. Davidson,{\it Turbulence} (Oxford University Press, 2004).}
449:
450: \bibitem{Taylor37}
451: {G.I. Taylor and A.E. Green, Proc.\ Roy.\ Soc.\ Lond.\ Ser.\ A {\bf158}, 895 (1937), 499.}
452:
453: \bibitem{Childress}
454: {S. Childress and A.D. Gilbert, ({\it Stretch, Twist, Fold: The fast dynamo}(Springer-Verlag Berlin, 1995).}
455:
456: \bibitem{Jacquin90}
457: {L. Jacquin {\it et al}, J.\ Fluid Mech. {\bf220}, 1 (1990).}
458:
459: \bibitem{Cambon97}
460: {C. Cambon, N.N. Mansour, and F.S. Godeferd, J. Fluid Mech. {\bf 337}, 303 (1997).}
461:
462: \bibitem{Biferale03}
463: {L. Biferale {\it et al}, Phys. Fluids {\bf 15}, 2105 (2003).}
464:
465: \bibitem{Kraichnan73}
466: {R.H. Kraichnan, J. Fluid Mech. {\bf 59}, 745 (1973).}
467:
468: \bibitem{Morinishi01}
469: {Y. Morinishi, K. Nakabayashi, and S. Ren, JSME Int. J. Ser. B {\bf44}, 410 (2001).}
470:
471: \bibitem{Lesieur}
472: {J.C. Andr\'e and M. Lesieur, J. Fluid Mech. {\bf 81}, 187 (1977).}
473:
474: \bibitem{group3}
475: {P.A. Davidson, P.J. Staplehurst, and S.B. Dalziel, J. Fluid Mech. {\bf 557}, 135 (2006); C. Morize and F. Moisy, Phys. Fluids {\bf 18}, 065107 (2006).}
476:
477: \bibitem{bellet}
478: {F. Bellet {\it et al}, J. Fluid Mech. {\bf 562}, 83 (2006).}
479:
480: %\bibitem{group4}
481: %{P.D. Mininni and A. Pouquet, Phys. Rev. Lett. {\bf 99}, 254502 (2007);}
482:
483: \bibitem{Mininni07}
484: {P.D. Mininni and A. Pouquet, Phys.\ Rev.\ Lett. {\bf 99}, 254502 (2007).}
485:
486: \bibitem{McWilliams84}
487: {J.C. Mc{W}illiams, J.\ Fluid Mech. {\bf146}, 21 (1984).}
488:
489: \bibitem{muller}
490: {W.C. M\"uller and M. Thiele, Europhy. Lett. {\bf 77}, 34003 (2007);P.D. Mininni, A. Alexakis, and A. Pouquet, Phys. Fluids (in press) arXiv:0802.3714v1 (2008).}
491:
492: \bibitem{MininniPress}
493: {P.D. Mininni and A. Pouquet, (submitted to Phys. Rev. E), arXiv:0707.3620v1 (2008).}
494:
495: \bibitem{Zhou95}
496: {Y. Zhou, Phys. Fluids {\bf 7}, 2092 (1995).}
497:
498: %\bibitem{Mininni08}
499: %{P.D. Mininni, A. Alexakis, and A. Pouquet, Phys.\ Rev.\ E {\bf77}, 036306 (2008).}
500:
501: \bibitem{vapor}
502: {J. Clyne {\it et. al.}, New J. Phys. {\bf 9}, 301 (2007).}
503:
504: %\bibitem{Chen031}
505: %{Q. Chen {\bf et. al.}, Phys. Rev. Lett., {\bf 90}, 214503 (2003).}
506:
507: %\bibitem{Davidson1}
508: %{P.A. Davidson,P.J. Staplehurst, and S.B. Dalziel, J. Fluid Mech., {\bf 557}, 135 (2006).}
509:
510: %\bibitem{Davidson2}
511: %{P.J. Staplehurst,P.A. Davidson, and S.B. Dalziel, J. Fluid Mech., {\bf 598}, 81 (2008).}
512:
513: %\bibitem{Moisy}
514: %{F. Moisy,C. Morize and M. Rabaud,Conference on turbulence and interactions TI2006}
515:
516: %\bibitem{Bellet}
517: %{F. Bellet, {\bf et al}, J. Fluid Mech., {\bf 562}, 83 (2008).}
518:
519: %\bibitem{Brachet83}
520: %{M. Brachet, {\bf et al}, J. Fluid Mech., {\bf 130}, 411 (1983).}
521:
522: %\bibitem{Mininni07}
523: %{P.~D. Mininni and A. Pouquet,Phys.\ Rev.\ Lett., {\bf99}, 254502 (2007).}
524:
525: %\bibitem{Mininni08}
526: %{P.~D. Mininni, A. Alexakis ,and A. Pouquet,Phys.\ Rev.\ E, {\bf77}, 036306 (2008).}
527:
528: %\bibitem{Simand00}
529: %{C.Simand, F.Chill\`a, and J.-F.Pinton, Europhys.\ Lett.,\ {\bf49}, 336 (2000).}
530:
531: %\bibitem{McWilliams84}
532: %{J.~C. Mc{W}illiams, J.\ Fluid Mech., {\bf146}, 21 (1984).}
533:
534: %\bibitem{Zhou95}
535: %{Y. Zhou, Phys. Fluids {\bf 7}, 2092 (1995).}
536:
537: %\bibitem{Kraichnan73}
538: %{R.~H. Kraichnan and J. Fluid Mech., {\bf 59}, 745 (1973).}
539:
540: %\bibitem{Chen03}
541: %{Q. Chen, S. Chen, and G.L. Eyink, Physics of fluids, {\bf 15}, 361 (2003).}
542:
543: %\bibitem{Borue97}
544: %{V. Borue and S.A. Orszag, J.Phys.Rev.E, {\bf 55}, 7005 (1997).}
545:
546: %\bibitem{Mininni06}
547: %{P.~D. Mininni, A. Alexakis, and A. Pouquet,Phys.\ Rev.\ E {\bf74}, 016303 (2006).}
548:
549: \end{thebibliography}
550:
551: \end{document}
552: