Reynolds-Dependent Intermittency in Turbulence

Updated 2 January 2026

Reynolds-Dependent Intermittency is defined by the control of rare, intense turbulent bursts by the Reynolds number, critical for understanding energy dissipation and flow transitions.
Multifractal, statistical, and information-theoretic methods reveal both logarithmic and power-law scaling behaviors that quantify deviations from Gaussianity in turbulent flows.
Distinct Re-dependent behaviors are observed in homogeneous, wall-bounded, and quantum turbulence, informing advanced modeling techniques and DNS validations.

Reynolds-Dependent Intermittency

Reynolds-dependent intermittency refers to the explicit control of the statistical occurrence, intensity, and spatial or temporal structure of intermittent events by the Reynolds number (Re) in hydrodynamic, magnetohydrodynamic, and related complex flows. Intermittency in turbulence describes rare, intense, and highly localized bursts of dissipative activity (e.g., velocity gradients, energy flux, or vorticity), and its Re-dependence is pivotal to understanding fully developed turbulence, transitional flows, and the crossover between laminar and turbulent regimes. Recent advances use multifractal, statistical, and information-theoretic approaches to quantify how turbulent intermittency evolves as Re is varied, encompassing both classical and quantum fluids, wall-bounded and homogeneous turbulence, as well as transitional regimes with coexisting laminar/turbulent domains.

1. Fundamental Definitions and Statistical Measures

Intermittency is characterized by deviations from Gaussianity or strict self-similarity, typically manifested in the heavy tails of PDFs (probability density functions) of velocity increments, dissipation rates, enstrophy, or related observables. Canonical diagnostics include:

Flatness/Kurtosis: The flatness $F(r) = S_4(r) / [S_2(r)]^2$ , where $S_n(r)$ are nth-order structure functions, quantifies the amplitude of “fat tails.” For dissipation, the kurtosis of velocity gradients or magnetic increments is widely used (Sarkar et al., 8 May 2025, Cuesta et al., 2022, Parashar et al., 2019).
Scaling Exponents/Structure Functions: Deviations in the scaling exponents $\zeta_n$ of $S_n(r) \sim r^{\zeta_n}$ from dimensional predictions indicate intermittency. Anomalous exponents $(\zeta_n - n/3)$ and the intermittency exponent $\mu$ are commonly reported (Cerbus et al., 2013, Biferale et al., 2014, Duan et al., 28 Jun 2025).
Shannon Entropy and KL-Divergence: Quantifies intermittency via information content and departure from a baseline (Gaussian random field–GRF) PDF (Sarkar et al., 8 May 2025).
Predator–Prey/Bistable Dynamics: Minimal stochastic models use Re-dependent dimensionless numbers to demarcate intermittent and non-intermittent regimes, even in non-continuum systems (1901.10567).

2. Homogeneous Isotropic Turbulence: Intermittency Growth with Reynolds Number

For forced isotropic Navier–Stokes turbulence, both empirical and theoretical studies show that small-scale intermittency strengthens monotonically with increasing Re, but the rate depends on the chosen metric.

Scaling Laws and Entropic Measures: The Kullback–Leibler divergence $D_{KL}$ grows only logarithmically with Taylor-Reynolds number ( $R_\lambda$ ):

$D_{KL}(f_\phi \Vert f_{GRF}) \approx -0.925 + 0.376 \ln R_\lambda$

over $R_\lambda$ from 10 to almost 600, for pseudodissipation $\phi$ (Sarkar et al., 8 May 2025).

Flatness/Kurtosis: Moment-based intermittency measures (e.g., the flatness of velocity increments) grow more rapidly with Re, typically as a weak power law (exponent $\approx 0.16\,$ – $\,0.20$ ), while information-theoretic measures reveal much slower, logarithmic growth (Cuesta et al., 2022, Parashar et al., 2019).
Saturation at Large Re: Ensemble-averaged mean-field statistics approach Kolmogorov–Obukhov scaling as $R_\lambda \rightarrow \infty$ , with finite- $R_\lambda$ corrections receding slowly (Yoffe et al., 2021, Bos et al., 2011).

3. Wall-Bounded and Inhomogeneous Flows: Inner and Outer Scaling

Intermittency in wall-bounded flows exhibits distinct Re-dependent behavior in "outer" and "near-wall" regions (Duan et al., 28 Jun 2025):

Region	Moment Scaling ( $\langle \epsilon^p \rangle$ )	Intermittency Persistence
Outer/log-layer	$\langle \epsilon^p \rangle^{1/p} \sim Re^\gamma_p$	Power-law ( $\gamma_p > 0$ from MF exponents)
Near-wall	$\langle \epsilon^p \rangle^{1/p} \to \mathrm{const.}$	Moments saturate, intermittency vanishes

Here, in the log-layer and center, multifractal scaling linked to anomalous exponents $\zeta_n$ persists and strengthens with $Re_\tau$ . Near the wall, all moments approach finite limits for $Re_\tau \to \infty$ , so small-scale intermittency effectively disappears, and extreme singularities become wall-localized, consistent with Onsager-type anomalous dissipation (Duan et al., 28 Jun 2025).

4. Transitional and Spatiotemporal Intermittency: Laminar–Turbulent Coexistence

In transitional regimes (e.g., pipe, channel, or boundary layer flows at moderate Re), intermittency manifests as alternating laminar and turbulent patches, with Re controlling their nucleation, growth, and statistics.

Classical Pipe Flow: The minimum spatial separation $L_{int}(Re)$ between turbulent puffs decreases linearly with Re, resulting in a monotonically increasing maximum turbulent fraction. At a critical Re (around 2550), the separation vanishes, and fully turbulent flow ensues (Samanta et al., 2010).
Prandtl–Tietjens Mechanism: Laminar–turbulent intermittency arises as a feedback between friction fluctuations and flow speed under a constant pressure drop, with the period, amplitude, and extent of cycles scaling predictably with Re. A refined model quantitatively reproduces experiments for $3000 \lesssim Re \lesssim 7000$ (Cerbus, 2021).
Predator–Prey Analogue: In many-body crystalline models, an effective Reynolds analog $\Theta = c\Phi/\gamma^2$ controls stochastic intermittency, with phase transitions between quiescent (crystalline), intermittent, and “melted” states as $\Theta$ varies (1901.10567).

5. Re-Dependence in Complex and Quantum Turbulence

The role of Reynolds-dependent intermittency extends into quantum, 2D, and astrophysical turbulence:

Quantum Turbulence in He II: Small-scale intermittency and vortex-line density statistics collapse when plotted against a quantum Reynolds number, $Re_\kappa = v_{rms} L_0 / \kappa$ , independent of temperature. The inter-vortex spacing scales as $\ell/L_0 \sim Re_\kappa^{-3/4}$ , and observed variations in intermittency are entirely attributable to variations in $Re_\kappa$ rather than intrinsic T-dependence (Polanco et al., 4 Jul 2025).
2D Turbulence: In soap-film experiments, the enstrophy-cascade intermittency exponent $\mu(R_\lambda)$ decreases rapidly as $\sim R_\lambda^{-0.47}$ for $R_\lambda \gtrsim 50$ , while the energy cascade exponent remains Re-independent—intermittency can vanish in some regimes as Re increases, confirming the importance of specific cascade dynamics (Cerbus et al., 2013).
Expanding Solar Wind: Effective $Re \sim R^{-2/3}$ controls the evolution of magnetic field intermittency, with kurtosis falling as a power law in heliocentric distance and scaling as $K \sim Re^{0.16-0.2}$ , consistent with hydrodynamic predictions (Parashar et al., 2019, Cuesta et al., 2022).

6. Modeling, Computation, and RANS Transition Modeling

In turbulence models and RANS transition closures, intermittency is rendered quantitatively Re-dependent:

Transition Intermittency Correlation: In bypass transition over flat plates, the streamwise evolution of intermittency $\gamma(x)$ is reconstructed as a function of the local Re and free-stream turbulence, with explicit parametric equations predicting the transition Re to within 10% in both DNS and wind-tunnel data (Gonzalez et al., 26 Feb 2025).
Stabilized RANS/SA Modeling: New $\gamma$ -equations for laminar–turbulent intermittency incorporate Re-number dependent correlations and guarantee robust prediction of separation-induced transition over $10^5 \lesssim Re \lesssim 10^7$ (D'Alessandro et al., 4 Aug 2025).
DNS and EDQNM Analyses: High-fidelity simulations and closure models (e.g., EDQNM) reveal that finite- $Re$ produces power-law corrections ( $\sim r^{0.04}$ in the inertial range for skewness), distinct from persistent multifractal intermittency effects in the limit $R_\lambda \rightarrow \infty$ (Bos et al., 2011).

7. Information-Theoretic and Multifractal Perspectives

Modern approaches dissect Reynolds-dependent intermittency by quantifying the information content of small-scale turbulence:

Information-Theory Scaling: Shannon entropy and KL divergence show clear crossovers in their $R_\lambda$ -scaling, with the excess intermittency over a Gaussian baseline growing only logarithmically at high Re, fundamentally slower than the growth of moments—a key guidance for multifractal cascade models (Sarkar et al., 8 May 2025).
Limitations and Ensemble-Dependence: In stationary, homogeneous turbulence, fluctuating, Re-dependent intermittency "averages out" at the level of ensemble means, preserving Kolmogorov–Obukhov scaling even as rare intense events grow more probable at larger Re (Yoffe et al., 2021).