betaFeedback

WorksPhysics and Astronomy

Intermittent fluctuations determine the nature of chaos in turbulence

Abstract

We adapt recent ideas for many-body chaos in nonlinear, Hamiltonian fluids [Murugan et al., Phys. Rev. Lett. 127, 124501 (2021)] to revisit the question of the Reynolds number Re dependence of the Lyapunov exponent λ∝Reαλ\propto{\rm Re}^α in fully developed turbulence. The use of such decorrelators allow us to investigate the interplay of the competing effects of viscous dissipation and nonlinearity. We obtain a precise value of α=0.59±0.04α= 0.59 \pm 0.04 and show that departure from the Kolmogorov mean field result λ∝Reλ\propto \sqrt{\rm Re} is a consequence of the intermittent fluctuations in the velocity-gradient tensor. The robustness of our results are further confirmed in a local, dynamical systems model for turbulence.

Fully developed, incompressible turbulence is perhaps the most celebrated example of a chaotic system. In contrast to other examples of classical, many-body systems showing chaotic behaviour, turbulent flows have the distinction of being central across natural world. Unsurprisingly, therefore, physicists working in problems ranging from astrophysics, atmospheric sciences, oceanography, and of course fluid dynamics have to factor in the underlying chaotic nature of such flows [1]. This ubiquity is not surprising: After all the underlying Navier-Stokes equation, in all such systems, lead to solutions which are turbulent in the limit of small viscosities commonly seen in most fluids [2].

While mathematical and engineering tools remain indispensable, the ideas of statistical physics provide the basis for much of our understanding of turbulence, especially in its most general form where the flow is homogeneous and isotropic. In particular, chaos along with the other fingerprint of fully developed turbulence, namely intermittency, and their dependence inter alia the Reynolds number Re\mathrm{Re}, forms a major challenge in a complete understanding of turbulence.

The degree of chaos is usually quantified by the (largest) Lyapunov exponent λ\lambda of the flow. By using arguments tracing back to Kolmogorov’s seminal work from 1941, Ruelle showed that λ∝Re\lambda\propto\sqrt{\mathrm{Re}} [3]. This result is a consequence of associating the (largest) Lyapunov exponent with the inverse of the smallest time-scale of the flow τη∝1/Re\tau_\eta\propto1/\sqrt{\mathrm{Re}}, obtained most simply from phenomenological arguments, and thence the λ\lambda-scaling. More generally, assuming a Hölder exponent hh characterising the flow, it is easy to show λ∝Re1−h1+h\lambda\propto\mathrm{Re}^{\frac{1-h}{1+h}} [4]; indeed taking the monofractal, Kolmogorov limit [5] of a unique h=1/3h = 1/3 recovers the Ruelle scaling.

The result λ∝Re\lambda\propto\sqrt{\mathrm{Re}} implicitly assumes a unique small length and time-scale characteristic of a monofractal flow.

However, the observational, experimental, and numerical evidence against this is overwhelming. Indeed, a modern rationalisation of turbulence rests on the multifractal approach [6] developed by Frisch and Parisi [7]. Adapting this model, which allows for a spread of Hölder exponents hh in the flow, leads to the revised scaling obtained by Crisanti et al. λ∝Reα\lambda\propto\mathrm{Re}^{\alpha}, with α≈0.459\alpha\approx0.459 [4].

This scaling exponent α≲0.5\alpha\lesssim0.5 has been challenged in recent years with data from different experiments and direct numerical simulations. For example, Berera and Ho [8] have recently reported α=0.53\alpha= 0.53. This lack of consensus between the theoretical estimate and measurements — as well as the lack of agreement between different simulations [4, 8, 9, 10, 11, 12, 13] — suggests that the origins of chaos in turbulence is far from settled.

But is there a way to address the microscopic origin of α\alpha from the Navier-Stokes equation itself and connect it to the intermittent aspect of fully developed turbulence? And what role does the non-locality (in length scale) of interactions — a defining feature of the Navier-Stokes equation — play?

In this paper we show that this is indeed possible through recently constructed ideas of decorrelators [14] — largely limited to Hamiltonian [15, 16, 17] systems in the context ergodicity, thermalisation and chaos — and adapt them for driven-dissipative, non-equilibrium systems such as turbulence. Consequently, by using a suitable mix of theory and numerical simulations, we uncover the competing effects of the viscous and nonlinear terms to show exactly the chaos and the scaling λ∼Reα\lambda\sim\mathrm{Re}^{\alpha} emerges along with what determining the value of α=0.59±0.04\alpha= 0.59 \pm0.04 and why this is so. We also confirm the robustness of this result by using a dynamical systems approach to turbulence with the additional advantage of exploiting its nearest and next-nearest neighbour interactions to investigate short-time effects of locality which are absent in the Navier-Stokes equation.

We begin with the three-dimensional, incompressible, unit-density Navier-Stokes equation, driven by an external force to a non-equilibrium statistically steady state. The force f\mathbf{f} varies to ensure a constant energy injection [18] J=⟨f⋅u⟩≡⟨ε⟩J = \langle\mathbf{f}\cdot\mathbf{u}\rangle\equiv\langle\varepsilon\rangle, where the mean energy dissipation rate ⟨ε‾⟩\langle\overline{\varepsilon}\rangle is the time ε‾\overline{\phantom{\varepsilon}} and space ⟨⋅⟩\langle\cdot\rangle average of the spatio-temporally fluctuating dissipation ε(x,t)\varepsilon(\mathbf{x},t) characteristic of such non-equilibrium steady states of fully developed turbulence with a statistically steady velocity field usteady\mathbf{u}_{\mathrm{steady}}. This follows from the familiar, instantaneous (spatially-averaged) kinetic energy EE balance:

dEdt=−⟨ε⟩+J\frac{dE}{dt}=-\langle\varepsilon\rangle+J

with a mean kinetic energy E0=E‾E_0=\overline{E}.

We now construct two, nearly identical, velocity fields u0A≡usteady\mathbf{u}_0^A\equiv\mathbf{u}_{\mathrm{steady}} and u0B=u0A+δu0\mathbf{u}_0^B=\mathbf{u}_0^A+\delta\mathbf{u}_0, where the form of δu0\delta\mathbf{u}_0 ensures that system B remains incompressible. We choose the amplitude of perturbation ϵ0\epsilon_0 sufficiently small (see Appendix A) and localise it in space at x0\mathbf{x}_0; the precise choice of x0\mathbf{x}_0 is irrelevant and can be chosen to be the center of our domain.

We choose these fields — u0A\mathbf{u}_0^A and u0B\mathbf{u}_0^B — as initial conditions t=0t=0 for the Navier-Stokes equation, evolve them independently in time, and calculate the velocity difference δu(x,t)≡uB(x,t)−uA(x,t)\delta\mathbf{u}(\mathbf{x},t)\equiv\mathbf{u}^B(\mathbf{x},t)-\mathbf{u}^A(\mathbf{x},t) for t>0t>0. From such velocity differences, we are able to construct the decorrelator ϕ(x,t)≡12∣δu(x,t)∣2\phi(\mathbf{x},t)\equiv\frac{1}{2}|\delta\mathbf{u}(\mathbf{x},t)|^2 [17, 19]. It is easy to show, starting from the Navier-Stokes equation written separately for systems A and B, that the decorrelator follows an evolution (written in component form and sum over repeated indices assumed)

∂ϕ(x,t)∂t=∂iWi−δuiSijδuj+νδui∇2δui(1)\frac{\partial\phi(\mathbf{x},t)}{\partial t}=\partial_iW_i-\delta u_iS_{ij}\delta u_j+\nu\delta u_i\nabla^2\delta u_i \tag*{(1)}

with the strain-rate tensor Sij=12(∂iujA+∂juiA)S_{ij}=\frac{1}{2}(\partial_i u_j^A+\partial_j u_i^A), Wi=uiBϕ+δui∂jk∫d3x′G(x,x′)[ujAδuk+ukAδuj+δujδuk]W_i=u_i^B\phi+\delta u_i\partial_{jk}\int d^3x'G(\mathbf{x},\mathbf{x}')[u_j^A\delta u_k+u_k^A\delta u_j+\delta u_j\delta u_k], where G(x,x′)G(\mathbf{x},\mathbf{x}') is the Green’s function obtained from the Poisson equation for pressure, and ν\nu the coefficient of kinematic viscosity (see Appendix B).

A further simplification exploits the statistical isotropy and homogeneity of fully developed turbulence to construct, over the volume VV, the spatially integrated decorrelator Φ(t)≡1V∫dxϕ(x,t)=⟨ϕ⟩\Phi(t)\equiv\frac{1}{V}\int d\mathbf{x}\phi(\mathbf{x},t)=\langle\phi\rangle. This leads to a simple cancellation of the divergence term in Eq. (1) and a resulting evolution equation for the integrated decorrelator:

Φ˙=dΦdt=βS+βη(2)\dot{\Phi}=\frac{d\Phi}{dt}=\beta_S+\beta_\eta \tag*{(2)}

where βS=−⟨δu⋅S⋅δu⟩\beta_S=-\langle\delta\mathbf{u}\cdot\mathbf{S}\cdot\delta\mathbf{u}\rangle and βη=ν⟨δu⋅∇2δu⟩\beta_\eta=\nu\langle\delta\mathbf{u}\cdot\nabla^2\delta\mathbf{u}\rangle as the contributions from the strain and dissipative terms, respectively. There is one additional remark to make. In the derivations of Eqs. (1)–(2) we have neglected the forcing terms on systems A and B. This is perfectly valid at short times but in the long time t→∞t\to\infty limit it plays a critical role as we shall see below.

At short times we conjecture an exponential growth of the decorrelator: Φ˙/Φ≡λ\dot{\Phi}/\Phi\equiv\lambda. We confirm this numerically by integrating Eq. (2) with an initial perturbation field δu\delta\mathbf{u} and measurements of the strain-rate tensor S\mathbf{S} drawn from direct numerical simulations (DNSs) of the Navier-Stokes equation (Appendix A).

In Figure 1 we show a representative plot of Φ˙/Φ\dot{\Phi}/\Phi vs time for Re=178.5\mathrm{Re}=178.5 displaying a clear plateau λ\lambda demarcated

Representative plots of the normalised spatially averaged decorrelator and its contributions

Figure 1. Representative plots of the normalised spatially averaged decorrelator Φ˙/Φ\dot{\Phi}/\Phi, βS/Φ\beta_S/\Phi and βη/Φ\beta_\eta/\Phi vs time for Re=178.5\mathrm{Re}=178.5. The pair of vertical dashed lines indicate the short-time exponential-growth phase. (Inset) Probability density function of the n^i2\hat{n}_i^2 direction cosines of u\mathbf{u} (at short times) showing a preferential growth of the difference field along the compressional i=3i=3 direction.

by a pair of vertical dashed lines. While this is not very surprising, given that turbulence is chaotic, it is far from obvious how to trace its origins in the structure of Eq. (2) and in particular how the competing effects of the strain and viscous terms conspire to give such exponential-growth phases. In Figure 1 we plot βS\beta_S and βη\beta_\eta (compensated by Φ\Phi) and find strong evidence of their short-time exponential growth — the plateau between the vertical dashed lines — corresponding to Lyapunov exponents λS\lambda_S and λη\lambda_\eta, respectively. Furthermore, the dissipative effect of the viscous term λη∼O(1/τη)<0\lambda_\eta\sim O(1/\tau_\eta)<0 is compensated by the effect of the strain term λS>∣λη∣\lambda_S>|\lambda_\eta|, leading to an overall positive Lyapunov exponent λ=λS+λη\lambda=\lambda_S+\lambda_\eta.

How do we understand the origins of λS\lambda_S which leads to a compensation of the negative λη\lambda_\eta and hence a positive λ\lambda? In particular, how does the strain field S\mathbf{S} ensure the exponential amplification of δu\delta\mathbf{u} at short times and thence Φ∼eλt\Phi\sim e^{\lambda t}? The answer lies in decomposing the strain term in the eigenbasis of the strain tensor δu⋅S⋅δu=∑i=13n^i2γi∣δu∣2\delta\mathbf{u}\cdot\mathbf{S}\cdot\delta\mathbf{u}=\sum_{i=1}^{3}\hat{n}_i^2\gamma_i|\delta\mathbf{u}|^2, with n^i\hat{n}_i direction cosines of u\mathbf{u} (along the eigendirections) and γi\gamma_i the eigenvalues. The relative distributions for the extensional (i=1i=1; γ1>0\gamma_1>0), intermediate (i=2i=2; γ2≈0\gamma_2\approx0) and compressional directions (i=3i=3; γ3<0\gamma_3<0), shown in the inset of Figure 1, underlines a bias for the compressional direction. Thus, clearly the dominant contribution in this expansion must come from the statistics of the compressional eigenvalue.

At such short times, a careful analysis of βS\beta_S shows (see Appendix B) that βS∼⟨γ3⟩+γ3std\beta_S\sim\langle\gamma_3\rangle+\gamma_3^{\mathrm{std}}. And hence, the Reynolds number dependence of λS\lambda_S must be dominated by the competing effects of the mean ⟨γ3⟩\langle\gamma_3\rangle eigenvalue and their intermittency-induced fluctuations γ3std\gamma_3^{\mathrm{std}}.

In the long time limit, βS→0\beta_S \to0, all eigenvalues are sampled equally (see Appendix B, Figure 5(c)) leading to ⟨δu⋅S⋅δu⟩∝∑i=13γi=0\langle\delta\mathbf{u}\cdot\mathbf{S}\cdot\delta\mathbf{u}\rangle\propto\sum_{i=1}^{3}\gamma_i=0 because of incompressibility. We see from Figure 1 that this is indeed the case as the compensated βS\beta_S term rapidly falls to 0 for t≳25t\gtrsim25.

Plots and distributions for the decorrelator in the Navier–Stokes equation and GOY shell model

Figure 5. (a) Plots of βn\beta_n, at late times, for different Reynolds numbers, showing a saturation to −2⟨ε⟩-2\langle\varepsilon\rangle (black dotted, horizontal line). (b) A plot of the probability distribution of components of δu\delta u (for Re=358.2\mathrm{Re}=358.2), for the Navier-Stokes equation, along the three eigen vectors of the strain rate tensor at late times once the decorrelator saturates. (c) The distribution of the compressional eigenvalues for different values of Re\mathrm{Re}. Both the peak and the spread of the distribution scales with Re\mathrm{Re}. (d) The evolution of ϕn\phi_n vs knk_n, for the GOY shell model, on a loglog scale at different times; the late time shows an asymptotic scaling corresponding to kn−2/3k_n^{-2/3} indicated by the dashed line.

Does the viscous term βη\beta_\eta also go to zero similarly at late times leading to Φ˙=0\dot{\Phi}=0? At such late times it is easy to show that (see Appendix B) βη→−2⟨ε⟩\beta_\eta\to-2\langle\varepsilon\rangle. Indeed, in Figure 1 we do see a clear saturation of the normalised viscous term to a value consistent with −2⟨ε⟩≈1.6-2\langle\varepsilon\rangle\approx1.6 (corresponding to the simulations from which the rate of strain matrix is drawn).

A trivial consequence of this is that within the framework of Eq. (2) Φ˙=−2⟨ε⟩\dot{\Phi}=-2\langle\varepsilon\rangle (clearly seen in Figure 1) and suggesting that the decorrelator never saturates. However this is a contradiction: By definition, the decorrelator is defined as the spatial average of the velocity difference of systems A and B, and hence at long times Φ→2E0\Phi\to2E_0 since the cross-correlator ⟨uA⋅uB⟩\langle\mathbf{u}_A\cdot\mathbf{u}_B\rangle vanishes as the fields decorrelate as t→∞t\to\infty.

This contradiction is resolved by recalling that the derivation of Eq. (2) neglects the effective energy injection. While at short times, this is zero, at long times this contribution is no longer irrelevant. In fact we can show (see Appendix B) that ⟨δf⋅δu⟩=2⟨ε⟩\langle\delta\mathbf{f}\cdot\delta\mathbf{u}\rangle=2\langle\varepsilon\rangle as t→∞t\to\infty and thus compensates the viscous contribution −2⟨ε⟩-2\langle\varepsilon\rangle (see Appendix B, Figure 5(a)) leading to Φ˙=0\dot{\Phi}=0.

How consistent are these ideas in actual measurements of the decorrelator in DNSs of the Navier–Stokes equation? We check for this by solving the incompressible Navier–Stokes equation on a 2π2\pi triply-periodic domain by using a pseudospectral method with a large-scale constant energy I=⟨ε⟩\mathcal{I}=\langle\varepsilon\rangle [18] injection scheme to drive

the flow and maintain it in non equilibrium statistically steady state with a mean energy E0E_0; see Appendix A for a detailed summary of the numerical scheme and the parameters used. By using the same protocol described earlier, we calculate the decorrelator Φ\Phi for several different values of the Reynolds number Re\mathrm{Re}. In Figure 2 we show representative snapshots at t=0t=0 (left inset) and t=12.5t=12.5 (right inset), corresponding to the initial and the full exponential-growth phases, respectively. (Ref. [20] shows the full evolution of ∣δu∣2\lvert\delta\mathbf{u}\rvert^2.)

Semilog plots of $\Phi$ vs $t$ from DNSs for different $\mathrm{Re}$, with insets showing snapshots of the square of the difference field

Figure 2. Semilog plots of Φ\Phi vs tt from DNSs for different Re\mathrm{Re} showing an exponential regime followed by saturation at a value $2E_0$ (dashed horizontal line). The insets show two snapshots at (left) t=0t=0 and (right) t=12.5t=12.5 of the square of the difference field ∣δu∣2\lvert\delta\mathbf{u}\rvert^2 for Re=178.5\mathrm{Re}=178.5. (See Ref. [20] for the full evolution of $\lvert\delta\mathbf{u}\rvert^2.)

These difference fields allow us to construct the spatially averaged decorrelator Φ\Phi which we plot, on a semilog scale, in the main panel of Figure 2 as a function of time. We see a convincing exponential regime from which we can extract the Lyapunov exponent λDNS\lambda^{\mathrm{DNS}}, followed by a saturation at a value approximately equal to 2E02E_0, indicated by the dashed horizontal line, as our theory suggests. (We have confirmed from our data that the onset of the saturation of the decorrelators (see Figure 1 and Figure 2) is of the order of the inverse of the Lyapunov exponent.)

It is important to observe that even at very short times, before the exponential growth phase, our results from the Navier–Stokes equation summarised in Figure 2 do not suggest an initial power-law growth which has been seen in many local models of chaotic systems[21, 22]. Could this — as well as the somewhat self-similar evolution of the difference field (see Figure 2, insets) — be a consequence of the essential non-local nature of the equation?

To test this as well as the robustness of our conclusions of an exponentially growing decorrelator, we use a simpler, phenomenological (cascade) model for turbulence, inspired from dynamical systems, namely the Gledzer–

Ohkitani-Yamada (GOY) shell model [22, 23, 24]. In such shell models (Appendix A), the velocities unu_n, associated with a scalar, exponentially growing, wave-number kn=k02nk_n=k_0 2^n, are treated as complex, dynamical variables, with the shells numbers being integers 1≤n≤N1\le n\le N (see Appendix A for further details).

In the statistically steady state (Appendix A), let the velocity field be denoted as unAu_n^{\mathrm{A}}. We now construct a second field unB=unA ∀n≠npu_n^{\mathrm{B}}=u_n^{\mathrm{A}}\ \forall n\ne n_p and unpB=(1+ϵ0)unpAu_{n_p}^{\mathrm{B}}=(1+\epsilon_0)u_{n_p}^{\mathrm{A}}. Our choice of np=16n_p=16 ensures that the tiny perturbation (with ϵ0=10−3\epsilon_0=10^{-3}) are always added at wavenumbers just before the onset of the dissipation scales for all the Reynolds number being considered. (We have checked that our results are robust to the specific choice of npn_p or variations in ϵ0\epsilon_0.)

We now use the nearly-identical fields unAu_n^{\mathrm{A}} and unBu_n^{\mathrm{B}} as initial conditions and evolve the GOY shell model independently for systems A and B and hence construct the shell model decorrelator ϕn(t)≡⟨∣unA−unB∣2/2⟩\phi_n(t)\equiv\langle|u_n^{\mathrm{A}}-u_n^{\mathrm{B}}|^2/2\rangle. The ⟨⋅⟩\langle\cdot\rangle denote an ensemble averaging over independent realisations of the initial field unAu_n^{\mathrm{A}}. While it is straightforward to write the evolution equation for ϕn\phi_n (Appendix C), we find this is neither revealing nor analytically tractable. Instead, we choose to work with the numerically constructed ϕn\phi_n and assess its behaviour from our simulations.

Given the local nature of interactions in the shell model[21], it is reasonable to expect that decorrelators such as ϕn\phi_n may well have a self-similar short time growth

followed by an eventual exponential rise with a saturation as t→∞t\to\infty. In the inset of Figure 3 we show a loglog plot of the temporal evolution of ϕn\phi_n for a small initial time window 0≤t≲0.50\le t\lesssim0.5 with a clear power-law range for n<npn<n_p. This power-law eventually (t≳0.1t\gtrsim0.1) gives way to the exponential-growth phase of decorrelators. The Lyapunov exponent characterising such shell-by-shell decorrelators ϕn\phi_n are of course nn-dependent. Hence we find it useful to construct the nn-independent Lyapunov exponent λGOY\lambda^{\mathrm{GOY}} via the integrated decorrelator ΦGOY=∑1Nϕn\Phi^{\mathrm{GOY}}=\sum_{1}^{N}\phi_n. In the main panel of Figure 3 we show a semilog plot of the temporal evolution of ΦGOY\Phi^{\mathrm{GOY}} for different values of Re showing, yet again, a clear exponential growth, with a shell model Lyapunov exponent λGOY\lambda^{\mathrm{GOY}}, followed by a saturation to values roughly corresponding to 2E02E_0, with E0E_0 the statistically steady state energy in our shell model.

Semilog plot of the shell model decorrelator $\Phi^{\mathrm{GOY}}$ vs $t$ for different $\mathrm{Re}$

Figure 3. Semilog plot of the shell model decorrelator ΦGOY\Phi^{\mathrm{GOY}} vs tt for different Re\mathrm{Re} showing an exponential growth. (Inset) Loglog plot of the shell-specific decorrelator ϕn\phi_n for different shells (the arrow indicates the direction n=1n=1 to n=16n=16 shells) at very early times shows a power-law growth.

We can now summarise all of these ideas in answering the question of how consistent are the various (largest) Lyapunov exponent — λ\lambda, λDNS\lambda^{\mathrm{DNS}}, and λGOY\lambda^{\mathrm{GOY}} — with each other given the different ways and indeed the different models from which they are derived? In Figure 4 we show a composite plot of λ\lambda, λDNS\lambda^{\mathrm{DNS}} and λGOY\lambda^{\mathrm{GOY}} versus the Reynolds number Re. For measurements from the DNSs and the GOY shell model simulations, we estimate the Lyapunov exponents from a linear fit in the exponential-growth phase of the decorrelators (as shown in Figure 2 and Figure 3); the errors on such fits yield the errorbars on λDNS\lambda^{\mathrm{DNS}} and λGOY\lambda^{\mathrm{GOY}}. We find our errorbars smaller than the symbols sizes in Figure 4. The errorbars for the exponent γ3\gamma_3 are extracted likewise and are of the same order as those on λDNS\lambda^{\mathrm{DNS}} and λGOY\lambda^{\mathrm{GOY}}. The exponent λ\lambda is obtained differently. We use plots of Φ/Φ∞\Phi/\Phi_\infty, such as the one shown in Figure 1, to extract λ\lambda and its errorbar as the mean and standard deviation of the plateau, respectively.

Loglog plot of $\lambda$, $\lambda^{\mathrm{DNS}}$, $\gamma_3^{\mathrm{std}}$, and $-\langle\gamma_3\rangle$ versus Reynolds number

Figure 4. A loglog plot of λ\lambda, λDNS\lambda^{\mathrm{DNS}}, γ3std\gamma_3^{\mathrm{std}} and −⟨γ3⟩-\langle\gamma_3\rangle, obtained from various DNSs versus the Reynolds number Re (DNS) (bottom X axis), as well as the Lyapunov exponent obtained from the shell model λGOY\lambda^{\mathrm{GOY}} (rescaled by a constant factor to overlap with the DNS data since the Reynolds number ranges are different) with the corresponding shell model Reynolds numbers Re⁡(GOY)\operatorname{Re}(\mathrm{GOY}) (top X axis). (Inset) A loglog plot of the probability density function P\mathcal{P} for λη\lambda_\eta showing a power-law ∣λη∣−4|\lambda_\eta|^{-4} tail.

We find remarkable consistency (within error bars) between the different measurements of the Lyapunov exponents and indeed across the different models, as well as the range of Reynolds number between the DNSs (bottom axis) and shell models (top axis), that we use. All the Lyapunov exponents seem to follow a universal scaling: λDNS=λGOY=λ∼Re⁡α\lambda^{\mathrm{DNS}}=\lambda^{\mathrm{GOY}}=\lambda\sim\operatorname{Re}^{\alpha}, with α=0.59±0.04\alpha=0.59\pm0.04. Remarkably, combining the shell models and DNS data we show in Figure 4 that our reported scaling holds for nearly 7 decades in Reynolds number.

The theory developed earlier suggests a connection between the Lyapunov exponent and the statistics of γ3\gamma_3 with λS∼⟨γ3⟩+γ3std\lambda_S\sim\langle\gamma_3\rangle+\gamma_3^{\mathrm{std}} (Appendix B). The departure of the measured Lyapunov exponents (Figure 4) from a Kolmogorov-like mean field result λ∼⟨γ3⟩∼Re⁡\lambda\sim\langle\gamma_3\rangle\sim\sqrt{\operatorname{Re}}, is a clear indication that the dynamics must stem from intermittent fluctuations and the dominant scaling in the statistics of γ3\gamma_3 is due to γ3std∼Re⁡α\gamma_3^{\mathrm{std}}\sim\operatorname{Re}^{\alpha}. In Figure 4 we show a plot of γ3std\gamma_3^{\mathrm{std}} from our DNSs and find a near perfect agreement with the Lyapunov exponent confirming nearly 7 decades in Reynolds number, the robustness of this theory: λ=λDNS=γ3std∼Re⁡0.59\lambda=\lambda^{\mathrm{DNS}}=\gamma_3^{\mathrm{std}}\sim\operatorname{Re}^{0.59}. Furthermore, we confirm that the mean eigenvalue −⟨γ3⟩∼Re⁡-\langle\gamma_3\rangle\sim\sqrt{\operatorname{Re}} (Figure 4) consistent with the simpler scaling analysis, not accounting for fluctuations, and at odds with the measured Lyapunov exponent.

In this work we have shown how recent ideas of decorrelators [14, 15, 16, 17, 25] provide a microscopic way to understand the origins of chaos in fully developed turbulence. Remarkably, these ideas are just as robust for simpler, cascade models of turbulence where locality [21] ensures an initial scale-invariant growth of the decorrelator. We find signatures of these fluctuations in one further aspect of the strain and viscous terms in Eq. (2) which connects with the idea of intermittency. The theory developed by us involve the mean exponents: λS\lambda_S and λη\lambda_\eta (and hence λ\lambda) are obtained from the spatial integrals over the corresponding terms. However, it is possible to measure the local exponents λS(x)≡δu⋅S⋅δu∣δu∣2\lambda_S(x) \equiv\frac{\delta\mathbf{u}\cdot\mathbf{S}\cdot\delta\mathbf{u}}{|\delta\mathbf{u}|^2} and λη(x)≡νδu⋅∇2δu∣δu∣2\lambda_\eta(x) \equiv\frac{\nu\delta\mathbf{u}\cdot\nabla^2\delta\mathbf{u}}{|\delta\mathbf{u}|^2} in the exponential-growth phase and thence their probability distribution functions P\mathcal{P}. These distribution are found to have exponential (for λS(x)\lambda_S(x)) or power-law (for λη(x)\lambda_\eta(x); see inset of Figure 4) tails strongly suggestive of intermittency and its connections with the question of chaos in turbulence. We do not explore these ideas, and in particular the origins of the power-law tails, any further in this work as well as its implications for Lagrangian chaos [26, 27], but leave it for a more detailed study of the statistics of βS\beta_S and βη\beta_\eta in the future.

Acknowledgments

We acknowledge Anupam Kundu, Siddhartha Mukherjee, Sthitadhi Roy, Sibaram Ruidas and Akash Sarkar for discussions on related issues. A.B. acknowledges the Long Term Visiting Students’ Program (LTSVP) of the ICTS-TIFR which enabled this collaboration. SB acknowledges funding from the Swarna Jayanti fellowship of SERB-DST (India) Grant No. SB/SJF/2021-22/12 and DST, Government of India (Nano mission), under Project No. DST/NM/TUE/QM-10/2019 (C)/7. S.S.R. acknowledges SERB-DST (India) projects STR/2021/000023 and CRG/2021/002766 for financial support. This research was supported in part by the International Centre for Theoretical Sciences (ICTS) for the programs - Field Theory and Turbulence ( (code:ICTS/ftt2023/12), Indo-French workshop on Classical and quantum dynamics in out of equilibrium systems (code: ICTS/ifwcqm2024/12) and 10th Indian Statistical Physics Community Meeting (code: ICTS/10thISPCM2025/04). The simulations were performed on the ICTS clusters Mario, Tetris, and Contra. The authors acknowledge the support of the DAE, Government of India, under projects nos. 12-R&D-TFR-5.10-1100 and RTI4001.

Direct Numerical Simulations

The Navier-Stokes Equation

We perform direct numerical simulations (DNSs), by using a pseudospectral method, of the three-dimensional (3D), incompressible, unit density Navier Stokes equation, with the pressure field PP

∂u∂t+u⋅∇u=ν∇2u−∇P+f.(3)\frac{\partial\mathbf{u}}{\partial t}+\mathbf{u}\cdot\nabla\mathbf{u}=\nu\nabla^{2}\mathbf{u}-\nabla P+\mathbf{f}. \tag*{(3)}

on a periodic box of size L=2πL=2\pi with N3N^{3} collocation points. We choose N=256N=256 and N=512N=512 (to check for numerical convergence) and vary the coefficient viscosity 10−3≤ν≤32×10−310^{-3}\leq\nu\leq32\times10^{-3} to obtain Reynolds numbers 50≤Re⁡≤140050\leq\operatorname{Re}\leq1400. We use a second-order Adams-Bashforth for time-marching with a time step δt=5×10−4\delta t=5\times10^{-4} (for N=256N=256) and δt=4×10−4\delta t=4\times10^{-4} (for N=512N=512).

We initialise the flow with a random initial condition such that the initial energy spectrum spectrum E(k)=A0k2exp⁡(−k2/2k02)E(k)=A_{0}k^{2}\exp(-k^{2}/2k_{0}^{2}) with A0=k0=1A_{0}=k_{0}=1. The system is driven to a non-equilibrium steady state (NESS) through a large-scale forcing with a constant energy injection at large scales corresponding to wavenumber 1≤kforce≤21\leq k_{\mathrm{force}}\leq2.

This statistically steady velocity field u\mathbf{u} is taken as the initial field for system A: u0A≡u\mathbf{u}_{0}^{A}\equiv\mathbf{u} and further define the initial condition for system B via u0B=u0A+δu0\mathbf{u}_{0}^{B}=\mathbf{u}_{0}^{A}+\delta\mathbf{u}_{0}. We use a Gaussian perturbation δu0(x)=ϵ0exp⁡(−(x−x0)2/2σ02)x^\delta\mathbf{u}_{0}(\mathbf{x})=\epsilon_{0}\exp\left(-(\mathbf{x}-\mathbf{x}_{0})^{2}/2\sigma_{0}^{2}\right)\hat{\mathbf{x}} with ϵ0=10−5\epsilon_{0}=10^{-5}, σ0=4dx\sigma_{0}=4dx and x0=(π,π,π)\mathbf{x}_{0}=(\pi,\pi,\pi) at the center of the domain and which also satisfies the incompressibility constraint.

The nearly identical fields u0A\mathbf{u}_{0}^{A} and u0B\mathbf{u}_{0}^{B} are then used as initial conditions for the Navier-Stokes equation and evolved as before. The decorrelator is calculated from the evolution of such twin simulations as ϕ(x,t)=∣uB−uA∣2/2\phi(\mathbf{x},t)=|\mathbf{u}^{B}-\mathbf{u}^{A}|^{2}/2 and thence Φ(t)=∫dxϕ(x,t)/V\Phi(t)=\int dx\phi(\mathbf{x},t)/\mathcal{V}. The decorrelators are then used to estimate the Lyapunov exponents (obtained by fitting the initial exponential growth phase) and its dependence on the Reynolds number.

The Gledzer-Ohkitani-Yamada (GOY) Model

The Gledzer-Ohkitani-Yamada or the GOY shell model

ddtun=ikn[un+2un+1−14un+1un−1−18un−1un−2]∗−νkn2un+f(4)\frac{d}{dt}u_{n}=\mathrm{i}k_{n}\left[u_{n+2}u_{n+1}-\frac{1}{4}u_{n+1}u_{n-1}-\frac{1}{8}u_{n-1}u_{n-2}\right]^{*}-\nu k_{n}^{2}u_{n}+f \tag*{(4)}

is one of several cascade models which mimic Navier-Stokes turbulence and are designed to achieve very high Reynolds number. In such models, the velocities unu_{n} (with boundary conditions u−1=u0=uN+1=uN+2=0u_{-1}=u_{0}=u_{N+1}=u_{N+2}=0) associated with a scalar, exponentially growing, wave-number kn=k02nk_{n}=k_{0}2^{n}, are complex, dynamical variables. The shell numbers range from 1≤n≤N1\leq n\leq N; in our simulations we choose N=22N=22 and the coefficient of viscosity varies as 2×10−7≤ν≤64×10−72\times10^{-7}\leq\nu\leq64\times10^{-7} yielding Reynolds numbers in the range 8.7×106≤Re⁡≤2.6×1088.7\times10^{6}\leq\operatorname{Re}\leq2.6\times10^{8}; the shell model, Reynolds number is simply defined as Re⁡=∣urms∣/k0ν\operatorname{Re}=|u_{\mathrm{rms}}|/k_{0}\nu. The equation is integrated numerically by using a second-order Runge-Kutta scheme with a time step δt=2×10−5\delta t=2\times10^{-5}. The constant amplitude forcing f=0.1f=0.1, applied on the n=2n=2 shell and drives the system to a statistically steady state [28].

As with the strategy for the Navier-Stokes equation, we first obtain a non-equilibrium statistically steady state with an energy spectrum E(kn)=∣un∣2/kn∼kn−5/3E(k_{n})=|u_{n}|^{2}/k_{n}\sim k_{n}^{-5/3} over an inertial range of shells whose extent is determined by Re. We use this steady state to define the initial field for system A — un,0Au_{n,0}^{A} and construct the initial condition for system B via un,0B=(1+ϵ0)δn,npun,0Au_{n,0}^{B}=(1+\epsilon_{0})\delta_{n,n_{p}}u_{n,0}^{A}. In all our simulations we have used ϵ0=0.001\epsilon_{0}=0.001 and np=16n_{p}=16. Numerically it has been seen that the shell wise decorrelators first show a power-law rise later followed by the exponential-growth regime. The initial power law largely depends on the distance from the perturbed shell. Therefore, we add the perturbation in a fixed shell np=16n_{p}=16 for all values of Re.

Derivation of the decorrelator from the Navier-Stokes equation and its limits

Evolution equation for ϕ(r,t)\phi(r,t) and Φ\Phi

In this section we provide the derivation of the equation of motion of the decorrelator for the Navier-Stokes equation and in particular estimate the long and short time limits of βS\beta_{S} and βη\beta_{\eta}.

We begin by recalling that the pressure term in the Navier–Stokes equation can be rewritten as a Poisson equation

∇2P=−∂jui∂iuj(5)\nabla^2 P = -\partial_j u_i \partial_i u_j \tag*{(5)}

by exploiting the incompressibility constraint. This Poisson equation underlines the non-local nature of turbulence.

It is useful, in what follows, to have a Green’s function representation

P(x)=−∫dx′G(x,x′)[∂jui∂iuj]′(6)P(\mathbf{x}) = -\int\mathrm{d}\mathbf{x}' G(\mathbf{x},\mathbf{x}') [\partial_j u_i \partial_i u_j]' \tag*{(6)}

where the ′' denotes the quantity evaluated at x′\mathbf{x}'.

This allows us to construct the evolution equation of the solenoidal velocity perturbation δu=uB−uA\delta\mathbf{u} = \mathbf{u}^{\mathrm{B}}-\mathbf{u}^{\mathrm{A}} in component form

∂δui∂t=−δuj∂juiA−ujB∂jδui−∂i(∫dx′G(x,x′)[∂j2(ujAδuk+δujukA+δujδuk)]′)+ν∇2δui+δuiδfi(7)\frac{\partial\delta u_i}{\partial t} = -\delta u_j \partial_j u_i^{\mathrm{A}} - u_j^{\mathrm{B}}\partial_j \delta u_i - \partial_i \left(\int\mathrm{d}\mathbf{x}'G(\mathbf{x},\mathbf{x}')[\partial_j^2(u_j^{\mathrm{A}}\delta u_k+\delta u_j u_k^{\mathrm{A}}+\delta u_j\delta u_k)]'\right) + \nu\nabla^2\delta u_i+\delta u_i\delta f_i \tag*{(7)}

and hence, via a dot product of δu\delta\mathbf{u} with Eq. (B-3), the equation for the decorrelator ϕ(r,t)\phi(\mathbf{r},t):

∂ϕ∂t=−δuiSijδuj−∂iWi+νδui∇2δui+δuiδfi.(8)\frac{\partial\phi}{\partial t}=-\delta u_i S_{ij}\delta u_j-\partial_iW_i+\nu\delta u_i\nabla^2\delta u_i+\delta u_i\delta f_i. \tag*{(8)}

Here, Sij=12(∂juiA+∂iujA)S_{ij}=\frac{1}{2}(\partial_j u_i^{\mathrm{A}}+\partial_i u_j^{\mathrm{A}}) is the strain-rate tensor and Wi=uiBϕ+δui∫dx′G(x,x′)[∂jk2(ujAδuk+δujukA+δujδuk)]′W_i=u_i^{\mathrm{B}}\phi+\delta u_i\int\mathrm{d}\mathbf{x}'G(\mathbf{x},\mathbf{x}') [\partial_{jk}^2(u_j^{\mathrm{A}}\delta u_k+\delta u_j u_k^{\mathrm{A}}+\delta u_j\delta u_k)]'.

A spatial integration with Φ=∫dxϕ\Phi=\int\mathrm{d}\mathbf{x}\phi now leads to the cancellation of the divergence term and hence

dΦdt=−⟨δuiSijδuj⟩+ν⟨δui∇2δui⟩+⟨δuiδfi⟩≡βS+βη+⟨δuiδfi⟩(9)\frac{\mathrm{d}\Phi}{\mathrm{d}t}=-\langle\delta u_iS_{ij}\delta u_j\rangle+\nu\langle\delta u_i\nabla^2\delta u_i\rangle+\langle\delta u_i\delta f_i\rangle\equiv\beta_S+\beta_\eta+\langle\delta u_i\delta f_i\rangle \tag*{(9)}

the evolution equation for our spatially integrated decorrelator as discussed in the main text of the manuscript.

Short and long-time asymptotics of SS

At early times, ⟨δuiδfi⟩=0\langle\delta u_i\delta f_i\rangle=0 (since the forcing terms are nearly identical), the pre saturation exponential growth of the decorrelator with the Lyapunov exponent λ=λDNS∼λS\lambda=\lambda^{\mathrm{DNS}}\sim\lambda_S is controlled by the short time asymptotics of βS\beta_S. Hence, in what follows, we drop the contribution from βη\beta_\eta, to simplify notation.

We now expand δuiSijδuj\delta u_iS_{ij}\delta u_j in the eigenbasis of the strain-rate tensor, with eigenvalues γi\gamma_i and direction cosines cos⁡θi\cos\theta_i, to obtain

∂∂tϕ=−ϕ(x,t)Γ(x,t)(10)\frac{\partial}{\partial t}\phi=-\phi(\mathbf{x},t)\Gamma(\mathbf{x},t) \tag*{(10)}

with Γ(x,t)≡(∑i=13γi(x,t)cos⁡2θi(x,t))≈γ3(x,t)cos⁡2θ3(x,t)\Gamma(\mathbf{x},t)\equiv\left(\sum_{i=1}^{3}\gamma_i(\mathbf{x},t)\cos^2\theta_i(\mathbf{x},t)\right)\approx\gamma_3(\mathbf{x},t)\cos^2\theta_3(\mathbf{x},t) dominated by the compressional eigendirection as discussed in the main text.

We choose an integrating factor eΘ(x,t)e^{\Theta(\mathbf{x},t)}, such that ∂tΘ(x,t)=Γ(x,t)\partial_t\Theta(\mathbf{x},t)=\Gamma(\mathbf{x},t), to obtain the form of the spatially-resolved decorrelator at time tt

ϕ(x,t)=ϕ(x,0)eΘ(x,0)−Θ(x,t)(11)\phi(\mathbf{x},t)=\phi(\mathbf{x},0)e^{\Theta(\mathbf{x},0)-\Theta(\mathbf{x},t)} \tag*{(11)}

We are now able to spatially integrate Eq. (B-7) and, keeping in mind that the initially localised perturbation ϕ(x,0)\phi(\mathbf{x},0) is independent of eΘ(x,0)−Θ(x,t)e^{\Theta(\mathbf{x},0)-\Theta(\mathbf{x},t)}, we obtain

Φ(t)=Φ(0)⟨eΘ(x,0)−Θ(x,t)⟩(12)\Phi(t)=\Phi(0)\langle e^{\Theta(\mathbf{x},0)-\Theta(\mathbf{x},t)}\rangle \tag*{(12)}

where the angular brackets ⟨⋅⟩≡∫dx(⋅)\langle\cdot\rangle\equiv\int\mathrm{d}\mathbf{x}(\cdot).

Since we are interested in the short-time asymptotics of βS\beta_S, it is possible to Taylor expand the exponential and, keeping in mind the mapping of Θ\Theta to Γ\Gamma, we obtain

Φ(t)≲Φ(0)exp⁡[∑p=1∞(−1)pp!⟨ΔΘ(x,t)p⟩](13)\Phi(t)\lesssim\Phi(0)\exp\left[\sum_{p=1}^{\infty}\frac{(-1)^p}{p!}\langle\Delta\Theta(\mathbf{x},t)^p\rangle\right] \tag*{(13)}

where ΔΘ(x,t)=Θ(x,t)−Θ(x,0)\Delta\Theta(x,t)=\Theta(x,t)-\Theta(x,0). Considering terms up to O(ΔΘ2)\mathcal{O}(\Delta\Theta^{2}), to account for corrections stemming from intermittent fluctuations [8],

Φ(t)≲Φ(0)exp⁡[−⟨ΔΘ⟩+12⟨ΔΘ2⟩](14)\Phi(t)\lesssim\Phi(0)\exp\left[-\langle\Delta\Theta\rangle+\frac{1}{2}\langle\Delta\Theta^{2}\rangle\right] \tag*{(14)}

Now all that remains is the evaluation of ⟨ΔΘ⟩\langle\Delta\Theta\rangle and ⟨ΔΘ2⟩\langle\Delta\Theta^{2}\rangle. From the choice of Θ\Theta it is easy to see (by interchanging time and space integrals ⟨⋅⟩≡∫dx(⋅)\langle\cdot\rangle\equiv\int dx(\cdot) suitably) that ⟨ΔΘ⟩∝⟨γ3⟩∼Re\langle\Delta\Theta\rangle\propto\langle\gamma_{3}\rangle\sim\sqrt{\mathrm{Re}} (see Figure 4 of the main text). The determination of ⟨ΔΘ2⟩\langle\Delta\Theta^{2}\rangle requires a further Ansatz of an approximate delta-correlation in time, at least over time-scales of the order associated with strain-rate, to yield ⟨ΔΘ2⟩∝⟨γ3std⟩∼Re0.59\langle\Delta\Theta^{2}\rangle\propto\langle\gamma_{3}^{\mathrm{std}}\rangle\sim\mathrm{Re}^{0.59} (see Figure 4 of the main text as well as Figure 5(c) which shows that indeed the distribution of γ3\gamma_{3} changes with Re\mathrm{Re}) which, in the large Reynolds number limit dominates over the Re\sqrt{\mathrm{Re}} of the mean exponent. Hence λ=λDNS∼λS∼Reα\lambda=\lambda^{\mathrm{DNS}}\sim\lambda_{S}\sim\mathrm{Re}^{\alpha} with α=0.59±0.04\alpha=0.59\pm0.04.

This then completes the proof of how the intermittency induced fluctuations determines the Lyapunov exponent for fully-developed turbulence.

At long times, beyond the exponential-growth phase when the systems have started to completely decorrelate the forcing term can not be neglected. Thus as t→∞t\to\infty, we estimate ⟨δuiδfi⟩=⟨uiBfiB⟩+⟨uiAfiA⟩−⟨(uiBfiA+uiAfiB)⟩→2⟨ε⟩\langle\delta u_i\delta f_i\rangle=\langle u_i^Bf_i^B\rangle+\langle u_i^Af_i^A\rangle-\langle(u_i^Bf_i^A+u_i^Af_i^B)\rangle\to2\langle\varepsilon\rangle. As we shall see below, this exactly compensates the contribution of the viscous term βη\beta_\eta at long times.

Short and long-time asymptotics of βη\beta_\eta

The viscous term can be expanded as

βη=ν⟨δui∇2δui⟩=−ν⟨∂jδui∂jδui⟩=−ν⟨(∂iujB∂iujB+∂iujA∂iujA−2∂iujB∂iujA)⟩=−4νTr⁡(S2)+2ν⟨∂iujB∂iujA⟩=2⟨ε⟩+2ν⟨∂iujB∂iujA⟩.(15)\begin{aligned} \beta_\eta&=\nu\langle\delta u_i\nabla^2\delta u_i\rangle=-\nu\langle\partial_j\delta u_i\partial_j\delta u_i\rangle\\ &=-\nu\langle(\partial_i u_j^B\partial_i u_j^B+\partial_i u_j^A\partial_i u_j^A-2\partial_i u_j^B\partial_i u_j^A)\rangle\\ &=-4\nu\operatorname{Tr}(S^2)+2\nu\langle\partial_i u_j^B\partial_i u_j^A\rangle=2\langle\varepsilon\rangle+2\nu\langle\partial_i u_j^B\partial_i u_j^A\rangle. \tag*{(15)} \end{aligned}

We find at short times, in the exponential-growth phase, this terms scales as βη=O(1/τη)\beta_\eta=\mathcal{O}(1/\tau_\eta), where τη=ν/⟨ε⟩\tau_\eta=\sqrt{\nu/\langle\varepsilon\rangle} corresponds to the smallest Kolmogorov time scales. However, at long times t→∞t\to\infty, systems A and B decorrelate leading to ⟨∂iujA∂iujB⟩→0\langle\partial_i u_j^A\partial_i u_j^B\rangle\to0 and hence βη∼−2⟨ε⟩\beta_\eta\sim-2\langle\varepsilon\rangle (see Figure 5(a) as well as the main text). This is easily understandable because, on average, dissipation is balanced by the energy injection derived above.

Decorrelator in the shell model for turbulence

It is possible to construct an evolution equation for the decorrelator in our shell model but, as we shall see, its structure is less transparent than what we have seen for the Navier-Stokes equation.

By subtracting the GOY model equations for systems A and B we obtain the evolution of the shell-by-shell velocity difference δun=unB−unA\delta u_n=u_n^B-u_n^A.

ddtδun=ikn[(δun+2un+1A+δun+2un+1A+δun+2δun+1)−ϵλ2(δun+1un−1A+δun+1un−1A+δun+1δun−1)+ϵ−1λ2(δun−2un−1A+δun−2un−1A+δun−2δun−1)∗]−νkn2δun.(16)\frac{d}{dt}\delta u_n=\mathrm{i}k_n\left[(\delta u_{n+2}u_{n+1}^A+\delta u_{n+2}u_{n+1}^A+\delta u_{n+2}\delta u_{n+1})-\frac{\epsilon}{\lambda^2}(\delta u_{n+1}u_{n-1}^A+\delta u_{n+1}u_{n-1}^A+\delta u_{n+1}\delta u_{n-1})+\frac{\epsilon-1}{\lambda^2}(\delta u_{n-2}u_{n-1}^A+\delta u_{n-2}u_{n-1}^A+\delta u_{n-2}\delta u_{n-1})^{*}\right]-\nu k_n^2\delta u_n. \tag*{(16)}

From this by suitably multiplying with the complex conjugate field δun∗\delta u_n^*, the decorrelator obeys

ddtϕn=kn2[(δun+2un+1Aδun+δun+2un+1Aδun+δun+2δun+1δun)−ϵλ2(δun+1δunun−1A+δun+1δunun−1A+δun+1δunδun−1)+ϵ−1λ2(δun−2un−1Aδun+δun−2un−1Aδun+δun−2δun−1δun)∗]+c.c.−2νkn2ϕn(17)\frac{d}{dt}\phi_n=\frac{k_n}{2}\left[(\delta u_{n+2}u_{n+1}^A\delta u_n+\delta u_{n+2}u_{n+1}^A\delta u_n+\delta u_{n+2}\delta u_{n+1}\delta u_n)-\frac{\epsilon}{\lambda^2}(\delta u_{n+1}\delta u_nu_{n-1}^A+\delta u_{n+1}\delta u_nu_{n-1}^A+\delta u_{n+1}\delta u_n\delta u_{n-1})+\frac{\epsilon-1}{\lambda^2}(\delta u_{n-2}u_{n-1}^A\delta u_n+\delta u_{n-2}u_{n-1}^A\delta u_n+\delta u_{n-2}\delta u_{n-1}\delta u_n)^{*}\right]+\mathrm{c.c.}-2\nu k_n^2\phi_n \tag*{(17)}

A summation over all shells leads to simplifications in the evolution of the GOY model decorrelator ΦGOY\Phi^{\mathrm{GOY}}

ddtΦGOY=ι2∑nkn[ϵδunun+1Aδun+2+(1−ϵ)δunδun+1un+2A−unAδun+1δun+2]∗+c.c.−2ν∑nkn2ϕn(18)\frac{d}{dt}\Phi^{\mathrm{GOY}}=\frac{\iota}{2}\sum_n k_n\left[\epsilon\delta u_nu_{n+1}^A\delta u_{n+2}+(1-\epsilon)\delta u_n\delta u_{n+1}u_{n+2}^A-u_n^A\delta u_{n+1}\delta u_{n+2}\right]^*+\mathrm{c.c.}-2\nu\sum_n k_n^2\phi_n \tag*{(18)}

Further progress is only possible through a numerical solution of the decorrelator equation as we report in the main text of the manuscript. The shell-wise decorrelator at very short times show a power-law growth with a wavenumber-dependent scaling exponent. In particular at long times when the decorrelator saturates, we find, unsurprisingly, ∣δun∣2∼∣un∣2∼kn−2/3|\delta u_n|^2 \sim|u_n|^2 \sim k_n^{-2/3} as shown in Figure 5(d).

References

  1. [1]R. Pandit, P. Perlekar, and S. S. Ray, Pramana 73, 157 (2009).
  2. [2]U. Frisch, Turbulence: The Legacy of A. N. Kolmogorov (Cambridge University Press, 1995).
  3. [3]D. Ruelle, Physics Letters A 72, 81 (1979).
  4. [4]A. Crisanti, M. H. Jensen, A. Vulpiani, and G. Paladin, Phys. Rev. Lett. 70, 166 (1993).
  5. [5]A. N. Kolmogorov, Dokl. Akad. Nauk SSSR 30, 299–303 (1941).
  6. [6]R. Benzi, G. Paladin, G. Parisi, and A. Vulpiani, Journal of Physics A: Mathematical and General 17, 3521 (1984).
  7. [7]U. Frisch and G. Parisi, Proceedings of the International School of Physics Enrico Fermi, Course LXXXVIII, Varenna, 1985 (1985).
  8. [8]A. Berera and R. D. J. G. Ho, Phys. Rev. Lett. 120, 024101 (2018).
  9. [9]M. Yamada and K. Ohkitani, Journal of the Physical Society of Japan 56, 4210 (1987).
  10. [10]S. Mukherjee, J. Schalkwijk, and H. J. Jonker, Journal of the Atmospheric Sciences 73, 2715 (2016).
  11. [11]G. Boffetta and S. Musacchio, Phys. Rev. Lett. 119, 054102 (2017).
  12. [12]P. Mohan, N. Fitzsimmons, and R. D. Moser, Phys. Rev. Fluids 2, 114606 (2017).
  13. [13]J. Ge, J. Rolland, and J. C. Vassilicos, Journal of Fluid Mechanics 977, A17 (2023).
  14. [14]A. Das, S. Chakrabarty, A. Dhar, A. Kundu, D. A. Huse, R. Moessner, S. S. Ray, and S. Bhattacharjee, Phys. Rev. Lett. 121, 024101 (2018).
  15. [15]T. Bilitewski, S. Bhattacharjee, and R. Moessner, Phys. Rev. B 103, 174302 (2021).
  16. [16]T. Bilitewski, S. Bhattacharjee, and R. Moessner, Phys. Rev. Lett. 121, 250602 (2018).
  17. [17]S. D. Murugan, D. Kumar, S. Bhattacharjee, and S. S. Ray, Phys. Rev. Lett. 127, 124501 (2021).
  18. [18]A. G. Lamorgese, D. A. Caughey, and S. B. Pope, Physics of Fluids 17, 015106 (2004).
  19. [19]S. Mukherjee, R. K. Singh, M. James, and S. S. Ray, Nature Physics 19, 891 (2023).
  20. [20]Supplemental video (2025), spatio-temporal evolution of an initially localised perturbation.
  21. [21]L. Biferale, Annual Review of Fluid Mechanics 35, 441 (2003).
  22. [22]P. D. Ditlevsen, Turbulence and Shell Models (Cambridge University Press, Cambridge, UK, 2010).DOI
  23. [23]E. B. Gledzer, Soviet Physics Doklady 18, 216 (1973).
  24. [24]K. Ohkitani and M. Yamada, Progress of Theoretical Physics 81, 329 (1989).
  25. [25]S. H. Shenker and D. Stanford, Journal of High Energy Physics 2014 (2014).
  26. [26]J. Bec, L. Biferale, G. Boffetta, M. Cencini, S. Musacchio, and F. Toschi, Physics of Fluids 18, 091702 (2006).
  27. [27]S. S. Ray, Phys. Rev. Fluids 3, 072601 (2018).
  28. [28]S. S. Ray, D. Mitra, and R. Pandit, New Journal of Physics 10, 033003 (2008).

Paper details

vlnc:W_pqmhjc4p
Version
v1 · latest
Published
Fresh — added to the Hub within the last 30 days
Added
License
CC BY 4.0
Usage

views · 8

downloads · 0

A view counts once per browser per day, a download each time the PDF is requested.

  • Physics and Astronomy

Not recorded for this paper

View statement
Looking for related papers…
Contents