Intermittency and non-universality of pair dispersion in isothermal compressible turbulence

Through direct numerical simulations of two-dimensional isothermal compressible turbulence, this study reveals that while pair-dispersion halving-time statistics exhibit universal multifractal scaling, doubling-time statistics are non-universal and depend on the nature of the stirring force and Mach number, with significant implications for astrophysical gas transport and mixing.

The Gist

Imagine you are watching two tiny, invisible dancers (let's call them "particles") floating in a chaotic, swirling storm. In a calm, smooth river, you could easily predict how far apart they would drift over time. But in a storm, things get messy.

This paper is about studying those two dancers in a very specific, extreme kind of storm: compressible turbulence. Think of this not as water flowing, but as a gas (like the air in a giant cloud or the atmosphere of a star) that can be squashed and stretched. The researchers wanted to see if the old rules of how things drift apart still apply when the gas is being squeezed into shockwaves.

Here is the breakdown of their discovery, using some everyday analogies:

1. The Old Rule vs. The New Reality

For a long time, scientists believed in Richardson's Law. Imagine two friends holding hands in a crowded dance hall. If the music gets wilder (turbulence), they get pushed apart. The old rule said: "The faster the music, the faster they separate, and the time it takes to double their distance follows a simple, predictable pattern."

The researchers asked: Does this rule work if the dance hall is made of squishy, compressible gas where people can suddenly get crushed into tight corners (shocks)?

2. The Two Ways to Measure Time

To test this, they didn't just watch the dancers drift apart. They looked at two specific scenarios:

• The "Doubling" Time: How long does it take for the dancers to move farther apart (double their distance)?
• The "Halving" Time: How long does it take for the dancers to get closer together (halve their distance)?

In a smooth, non-compressible fluid, these two times are like mirror images of each other. If you know one, you know the other. But in this squishy, compressible gas, they are totally different.

3. The Big Surprise: "One-Way Streets"

The researchers found that the gas behaves like a city with one-way streets and sudden traffic jams.

• The "Halving" (Getting Closer): This is like the dancers getting sucked into a whirlpool or a traffic jam (a shockwave). The study found that getting closer is universal. No matter how the storm is started (whether you push the gas from the side or squeeze it from the top), the time it takes to get closer follows a strict, predictable rule. It's like gravity: it always pulls things down the same way.
• The "Doubling" (Getting Farther): This is where it gets weird. The time it takes to drift apart depends entirely on how the storm was created.
  • If the storm is driven by swirling forces (like a spinning fan), the dancers drift apart based on the "swirliness" of the gas.
  • If the storm is driven by squeezing forces (like a piston), the dancers drift apart in a completely different way that doesn't follow any known rulebook.

The Analogy: Imagine trying to predict how fast two cars will separate on a highway.

• Halving: If they are both trying to merge into a single lane (getting closer), they will always do it at the same speed, regardless of the car model.
• Doubling: If they are trying to speed up and separate, their speed depends entirely on who is driving the traffic. If the traffic cop is spinning (swirling force), they separate one way. If the traffic cop is honking and pushing (squeezing force), they separate another way.

4. The "Shockwave" Factor

In this gas, there are invisible "walls" called shocks. These are like sudden, invisible walls where the gas gets squashed incredibly hard.

• When the dancers get closer, they are usually being pulled into these shockwalls. The study found that these walls are like thin, one-dimensional lines (like a tightrope).
• When the dancers get farther, they are often running away from these walls or sliding along them. The study found that in some cases, they are actually running along thin, stretched-out ribbons of fast-spinning gas (vorticity) that form near the shocks.

5. Why Does This Matter?

You might ask, "Who cares about two particles in a computer simulation?"

This matters because most of the universe is made of this squishy, compressible gas.

• Star Formation: Stars are born in giant clouds of gas. If we use the old, simple rules to predict how these clouds mix and swirl, we might be wrong.
• Mixing: Understanding how gas mixes in these clouds is crucial for understanding how stars are born and how chemicals spread through space.

The Takeaway

The paper concludes that the universe is more complex than we thought.

1. Getting closer to a shockwave is predictable and follows a universal law.
2. Getting farther apart is chaotic and depends on exactly how the turbulence was started.
3. The old "Richardson's Law" needs a major update to work in the real, squishy universe.

In short: In a compressible storm, getting closer is easy to predict, but getting farther away is a mystery that depends on the specific nature of the storm itself.

Technical Summary

1. Problem Statement

The paper addresses the generalization of Richardson's law of pair dispersion to compressible turbulence.

• Background: In incompressible turbulence, Richardson's law states that the mean square separation of two Lagrangian particles grows as $\langle R^2(t) \rangle \sim t^3$. This is linked to the inertial range dynamics where nonlinear interactions dominate viscous dissipation.
• The Gap: While Richardson's law is well-studied in incompressible flows, its validity and the nature of intermittency corrections in compressible turbulence (prevalent in astrophysical systems like molecular clouds and star formation) remain poorly understood.
• Specific Challenge: Compressible flows introduce shocks and density fluctuations, leading to distinct behaviors for particle separation (divergence) and convergence (collapse). The authors investigate whether the multifractal models used for incompressible turbulence (which treat forward and backward time symmetrically) apply to compressible flows, specifically examining the statistics of doubling times (separation increasing) and halving times (separation decreasing).

2. Methodology

The authors employed Direct Numerical Simulations (DNS) of forced, isothermal, compressible turbulence in two dimensions.

• Governing Equations: The 2D Navier-Stokes equations for an isothermal ideal gas ($P = c_s^2 \rho$), including shear viscosity and a localized bulk viscosity (shock viscosity) to resolve shocks without excessive numerical dissipation.
• Simulation Setup:
  • Domain: 2D doubly-periodic square ($L=2\pi$) with $N^2 = 4096^2$ collocation points.
  • Forcing: Large-scale random white-in-time forcing at wavenumber $k_f = 3$.
  • Flow Types: Four distinct runs were performed by varying the nature of the forcing and the Mach number ($Ma$):
    • S1, S2: Solenoidal (divergence-free) forcing. $S1$ is transonic ($Ma \approx 1$), $S2$ is supersonic ($Ma \approx 2.5$).
    • C1, C2: Irrotational (compressive) forcing. $C1$ is transonic ($Ma \approx 0.75$), $C2$ is supersonic ($Ma \approx 2.0$).
  • Particle Tracking: $n = N^2$ Lagrangian tracers were seeded uniformly. Their positions were tracked to calculate separation statistics.
• Analysis Techniques:
  • Helmholtz Decomposition: Velocity field $\mathbf{u}$ was decomposed into solenoidal ($\mathbf{u}_s$) and irrotational ($\mathbf{u}_c$) components.
  • Structure Functions: Calculated equal-time structure functions $S_p(r)$ for total, solenoidal, and irrotational velocities to determine scaling exponents $\zeta_p, \zeta^s_p, \zeta^c_p$.
  • Exit-Time Statistics: Instead of relying solely on $\langle R^2(t) \rangle$ (which showed limited scaling ranges), the authors calculated exit times:
    • Doubling Time ($\tau_D$): Time for separation $R$ to reach $3R/2$.
    • Halving Time ($\tau_H$): Time for separation $R$ to reach $3R/4$.
  • Scaling Exponents: Extracted dynamic exponents $\chi^D_p$ and $\chi^H_p$ from the scaling of negative moments: $\langle \tau^{-p} \rangle \sim R^{-\chi_p}$.
  • Multifractal Analysis: Used the Legendre transform relationship between dynamic exponents $\chi_p$ and the multifractal spectrum $D(h)$ to identify the fractal dimensions of the structures dominating the dispersion.

3. Key Contributions and Results

A. Breakdown of Richardson's Law Universality

The study confirms that while a power-law growth of $\langle R^2(t) \rangle$ exists, the scaling regime is too narrow for robust quantitative measurement of the Richardson exponent. Consequently, the exit-time approach revealed deeper insights:

• Dynamic Multiscaling: Both doubling and halving times exhibit nonlinear scaling exponents $\chi_p$ as a function of order $p$, confirming dynamic multiscaling (intermittency) in compressible turbulence.
• Asymmetry: Crucially, $\chi^D_p \neq \chi^H_p$. The statistics of particle separation (doubling) and convergence (halving) are fundamentally different, violating the symmetry assumed in standard multifractal models for incompressible flows.

B. The Halving-Time Exponents ($\chi^H_p$) are Universal

• The exponents for halving times (particles coming together) satisfy the standard multifractal bridge relation:
  $$\chi^H_p = p - \zeta_p$$
  where $\zeta_p$ are the scaling exponents of the total velocity structure functions.
• Universality: This relation holds regardless of whether the forcing is solenoidal or irrotational, and is independent of the Mach number.
• Physical Interpretation: The multifractal spectrum $D_H(h)$ indicates that halving times are dominated by shocks (1D structures with local Hölder exponent $h \to 0$).

C. The Doubling-Time Exponents ($\chi^D_p$) are Non-Universal

The behavior of doubling times (particles moving apart) depends heavily on the forcing mechanism and Mach number:

• Solenoidal Forcing (S1, S2):
  • The exponents satisfy a modified bridge relation: $\chi^D_p = p - \zeta^s_p$, where $\zeta^s_p$ are the exponents of the solenoidal velocity component.
  • This implies that the expansion of particle pairs is driven primarily by the solenoidal (vortical) part of the flow.
  • The exponents depend strongly on the Mach number.
• Irrotational Forcing (C1, C2):
  • The exponents do not satisfy the relation $\chi^D_p = p - \zeta^s_p$ (since $\zeta^s_p$ is negligible here) nor the standard relation $\chi^D_p = p - \zeta_p$.
  • They are independent of the Mach number.
  • This indicates that in irrotational flows, the expansion is not solely determined by the solenoidal component but is significantly influenced by regions where $\nabla \cdot \mathbf{u} > 0$ (expansion zones).

D. Multifractal Spectrum Analysis

• Halving Times: The spectrum $D_H(h)$ converges to 1 as $h \to 0$, confirming that shocks (1D structures) control the convergence of particles.
• Doubling Times: The spectrum $D_D(h)$ is non-universal.
  • In supersonic solenoidal flows ($S2$), $D_D(h) \approx 1$ for $h \approx 0.4$, suggesting the dominant structures are thin, elongated patches of high vorticity near shocks, rather than the shocks themselves.
  • In transonic solenoidal flows ($S1$), $D_D(h)$ does not reach 1, indicating more space-filling structures.

4. Significance and Implications

• Generalization of Richardson's Law: The paper demonstrates that the classical view of pair dispersion (based on a single dynamic exponent and symmetric time evolution) is insufficient for compressible turbulence. One must distinguish between forward (doubling) and backward (halving) time evolution.
• Astrophysical Relevance: Since most astrophysical turbulence (e.g., in molecular clouds) is compressible and supersonic, models of mixing, star formation, and chemical transport in these environments must be revised. The "effective viscosity" concept used in incompressible mixing fails here because doubling and halving have different scaling properties.
• Theoretical Limitations: The findings suggest that current multifractal models of turbulence, which assume a single set of scaling exponents for the velocity field, cannot fully capture the statistics of pair dispersion in compressible flows. A new theoretical framework is required to explain the non-universality of the doubling-time exponents.
• 3D Extrapolation: The authors speculate that these qualitative findings (non-universality, distinct doubling/halving statistics, and the role of shocks/vorticity patches) will also hold in three-dimensional compressible turbulence.

In summary, this work provides a rigorous numerical characterization of pair dispersion in compressible turbulence, revealing a complex, non-universal landscape where the nature of the forcing and the Mach number dictate the statistical laws of particle separation, fundamentally differing from the incompressible paradigm.