pith. sign in

arxiv: 2510.11220 · v3 · submitted 2025-10-13 · ⚛️ physics.flu-dyn · nlin.CD· physics.comp-ph

A paradox of the Navier-Stokes turbulence

Shijie Qin , Kun Xu , Shijun Liao This is my paper

Pith reviewed 2026-05-18 07:50 UTC · model grok-4.3

classification ⚛️ physics.flu-dyn nlin.CDphysics.comp-ph
keywords Navier-Stokes equationsturbulenceRayleigh-Bénard convectiondirect numerical simulationnumerical noiseparadoxvortical flowzonal flow
0
0 comments X

The pith

Simulations of Navier-Stokes turbulence show flow type alternating between vortical and zonal states as time-steps shrink, implying small disturbances cannot be neglected.

A machine-rendered reading of the paper's core claim, the machinery that carries it, and where it could break.

The paper runs direct numerical simulations of two-dimensional Rayleigh-Bénard convection governed by the Navier-Stokes equations using double precision and a range of time-steps. It finds that the final flow settles into either vortical or zonal type, with completely different statistics, and these two outcomes alternate as the time-step is made smaller. The alternation indicates that numerical noise linked to the time-step size strongly shapes the turbulence. This outcome contradicts the standard assumption in the Navier-Stokes model that all small physical or artificial disturbances can be ignored. The contradiction produces a logical paradox for using the equations to describe turbulence.

Core claim

In direct numerical simulations of two-dimensional Rayleigh-Bénard convection governed by the Navier-Stokes equations, the final flow type switches between vortical flow and zonal flow as the time-step is reduced, with the two types having completely different statistics. The time-step has a close relationship with numerical noise, which therefore exerts a huge influence on the final flow type and statistics of the simulated turbulence. This clearly indicates that small disturbances have significant influences on the NS turbulence, which therefore should not be neglected, leading to a logical paradox for the NS turbulence.

What carries the argument

The alternation between vortical and zonal flow types that occurs when successively smaller time-steps are used in double-precision DNS of Rayleigh-Bénard convection.

Load-bearing premise

The observed switching of flow types with smaller time-steps stems from physical-scale small disturbances rather than from the chosen numerical scheme, grid, or specific chaotic structure of this convection setup.

What would settle it

Repeating the simulations at even smaller time-steps or with quadruple-precision arithmetic and checking whether the alternation between vortical and zonal states continues or eventually stops.

Figures

Figures reproduced from arXiv: 2510.11220 by Kun Xu, Shijie Qin, Shijun Liao.

Figure 1
Figure 1. Figure 1: FIG. 1. Schematic of 2D turbulent RBC tilted at an angle [PITH_FULL_IMAGE:figures/full_fig_p001_1.png] view at source ↗
Figure 2
Figure 2. Figure 2: FIG. 2 [PITH_FULL_IMAGE:figures/full_fig_p002_2.png] view at source ↗
Figure 3
Figure 3. Figure 3: FIG. 3. Final flow type of the tilted turbulent RBC versus [PITH_FULL_IMAGE:figures/full_fig_p002_3.png] view at source ↗
Figure 4
Figure 4. Figure 4: FIG. 4. Final flow type of the turbulent RBC tilted at [PITH_FULL_IMAGE:figures/full_fig_p003_4.png] view at source ↗
read the original abstract

The Navier-Stokes (NS) equations as a turbulence model have been widely applied in lots of fields. The NS equations contain such a fundamental assumption that all small physical/artificial disturbances could be neglected. Is this assumption correct? In this paper a two-dimensional Rayleigh-B\'{e}nard convection governed by the NS equations is predicted by traditional direct numerical simulation (DNS) using double precision arithmetic and a range of different time-steps. It is found that the final flow type tends either to vortical flow or zonal flow, whose statistics are completely different. Notably, these two flow types frequently alternate as the time-step is reduced to a very small value, suggesting that the time-step corresponding to each turbulent flow type should be densely distributed. Thus, stochastic numerical noise exerts a huge influence on the final flow type and statistics of numerically simulated NS turbulence because the time-step has a close relationship with numerical noise. This clearly indicates that small disturbances have significant influences on the NS turbulence, which therefore should not be neglected. This leads to a logical paradox for the NS turbulence, which is a great challenge for us, although a paradox often leads to some significant breakthroughs.

Editorial analysis

A structured set of objections, weighed in public.

Desk editor's note, referee report, simulated authors' rebuttal, and a circularity audit. Tearing a paper down is the easy half of reading it; the pith above is the substance, this is the friction.

Referee Report

3 major / 2 minor

Summary. The paper reports direct numerical simulations of 2D Rayleigh-Bénard convection governed by the Navier-Stokes equations in double precision, using a range of time steps. It finds that the long-time flow alternates between vortical and zonal states with qualitatively different statistics as the time step is successively reduced, and interprets the dense alternation as evidence that stochastic numerical noise (tied to the time step) exerts a dominant influence. The authors conclude that small disturbances cannot be neglected in the NS model, producing a logical paradox for NS-based turbulence.

Significance. If the reported alternation were shown to survive the continuum limit and to be independent of discretization details, the result would challenge a core modeling assumption in fluid turbulence and could motivate new approaches to handling infinitesimal perturbations. The present evidence, however, rests on a single numerical setup without controls that would establish this independence.

major comments (3)
  1. [Numerical experiments and discussion of time-step dependence] The central inference that time-step reduction proxies physical-scale disturbances (rather than discrete truncation or attractor-selection effects) is load-bearing for the paradox claim, yet the manuscript provides no tests with alternate time-marching schemes, spatial discretizations, or floating-point precisions to demonstrate that the vortical/zonal switching persists outside the chosen finite-difference/volume implementation.
  2. [Results on flow-type alternation] The assertion that the two flow types are densely distributed in the small-time-step limit is used to argue that noise cannot be neglected, but the manuscript does not report the zero-time-step (or fixed small-step) ensemble statistics or any convergence diagnostic that would confirm the alternation is not an artifact of the discrete dynamical system.
  3. [Conclusion and paradox statement] The logical step from observed numerical sensitivity to a paradox for the continuous NS equations requires an explicit argument that the same sensitivity would appear in the inviscid or continuum limit; the current text equates numerical noise with physical disturbances without deriving or testing this equivalence.
minor comments (2)
  1. [Flow classification] Clarify the precise definition of 'vortical' versus 'zonal' flow (e.g., via a quantitative diagnostic such as the ratio of kinetic energy in different Fourier modes) so that the statistical distinction is reproducible.
  2. [Simulation setup] Add a brief statement on the Rayleigh number, Prandtl number, and domain aspect ratio used in the DNS, as these parameters determine the attractor structure and are essential for assessing generality.

Simulated Author's Rebuttal

3 responses · 1 unresolved

We thank the referee for the thorough review and valuable suggestions. Below we address each major comment in detail, outlining our responses and planned revisions to the manuscript.

read point-by-point responses

  1. Referee: The central inference that time-step reduction proxies physical-scale disturbances (rather than discrete truncation or attractor-selection effects) is load-bearing for the paradox claim, yet the manuscript provides no tests with alternate time-marching schemes, spatial discretizations, or floating-point precisions to demonstrate that the vortical/zonal switching persists outside the chosen finite-difference/volume implementation.

    Authors: We recognize the importance of verifying that the alternation is not specific to our numerical implementation. Our simulations use a common finite-volume discretization in double precision, as is standard in many turbulence studies. The observed dependence on time step indicates that the final state is highly sensitive to small numerical perturbations. We will revise the manuscript to include a more detailed discussion of the numerical method and argue that similar sensitivity is expected in other schemes due to the universal presence of round-off and truncation errors in finite-precision arithmetic. Additionally, we plan to perform a limited test with an alternative time integration scheme to support this. revision: partial

  2. Referee: The assertion that the two flow types are densely distributed in the small-time-step limit is used to argue that noise cannot be neglected, but the manuscript does not report the zero-time-step (or fixed small-step) ensemble statistics or any convergence diagnostic that would confirm the alternation is not an artifact of the discrete dynamical system.

    Authors: The paper shows that as the time step is successively reduced, the final flow type alternates between vortical and zonal states. This alternation with finer time steps supports the idea of dense distribution in the small time-step limit. To strengthen this, we will add ensemble simulations at a fixed small time step and report the statistics of the resulting flow types, along with a diagnostic showing how the proportion of each state changes with decreasing time step. revision: yes

  3. Referee: The logical step from observed numerical sensitivity to a paradox for the continuous NS equations requires an explicit argument that the same sensitivity would appear in the inviscid or continuum limit; the current text equates numerical noise with physical disturbances without deriving or testing this equivalence.

    Authors: We believe the numerical results point to a deeper issue in the NS equations themselves, as the time-stepping approaches the continuous case while still exhibiting strong sensitivity. Numerical noise from finite time steps and precision can be seen as analogous to physical small disturbances that are inevitably present. We will expand the discussion in the conclusion to provide a clearer rationale for this equivalence, emphasizing that the paradox arises because the NS model neglects disturbances that, even when infinitesimal, can determine the turbulent state. A complete analytical treatment of the continuum limit is a significant undertaking that goes beyond the scope of this numerical study. revision: partial

standing simulated objections not resolved
  • A full analytical or numerical confirmation of the paradox directly in the continuum limit of the Navier-Stokes equations, independent of any discretization.

Circularity Check

0 steps flagged

No significant circularity; conclusions rest on explicit numerical experiments

full rationale

The paper's argument proceeds from direct numerical simulations of 2D Rayleigh-Bénard convection in double precision, varying the time-step as an external control parameter and observing alternation between vortical and zonal flow types with distinct statistics. This observation is then interpreted as evidence that stochastic numerical noise (linked to time-step size) exerts large influence, implying small disturbances cannot be neglected and yielding a logical paradox for the NS equations. No derivation chain reduces a claimed prediction or first-principles result to its own inputs by construction: the time-step is not a fitted quantity, there is no self-definitional mapping of outputs back to inputs, and no load-bearing self-citation or ansatz is invoked to force the conclusion. The central claim is an interpretive step from simulation data rather than a tautological re-expression of the method itself, rendering the paper self-contained against its own numerical benchmarks.

Axiom & Free-Parameter Ledger

0 free parameters · 1 axioms · 0 invented entities

The central claim rests on the modeling premise that the Navier-Stokes equations are intended to neglect all small physical or artificial disturbances; the numerical experiments are used to test that premise directly.

axioms (1)
  • domain assumption All small physical/artificial disturbances could be neglected in the Navier-Stokes equations as a turbulence model
    Stated explicitly in the abstract as the fundamental assumption under test.

pith-pipeline@v0.9.0 · 5734 in / 1421 out tokens · 27347 ms · 2026-05-18T07:50:09.997702+00:00 · methodology

discussion (0)

Sign in with ORCID, Apple, or X to comment. Anyone can read and Pith papers without signing in.

Reference graph

Works this paper leans on

28 extracted references · 28 canonical work pages · 1 internal anchor

  1. [1]

    A paradox of the Navier-Stokes turbulence

    illustrated that the maximum Lyapunov exponent of a chaotic system might alternate its sign as time-step de- creases to a very small value. It should be emphasized that initial condition has physical meanings, but numer- ical noise is “artificial” and thus “artificially” uncertain since it is strongly dependent of numerical algorithms that might have a la...

    work page internal anchor Pith review Pith/arXiv arXiv 2025

  2. [2]

    artificial

    The corresponding spatial resolution is sufficiently fine that the horizontal (maximum) grid spacing is less than the Kolmogorov scale, as previously verified by Pope [23]. Besides, we emphasize that the so-called Courant-Friedrichs-Lewy (CFL) condition, i.e. Courant number<1, is always satisfied for all simulations given by different time-step ∆tand vari...

    work page

  3. [3]

    the horizontal (maximum) grid spacing is less than the Kolmogorov scale

    work page

  4. [4]

    artificial

    the so-called Courant-Friedrichs-Lewy (CFL) con- dition, i.e. Courant number<1. We check all of our DNS results very carefully and it is found that all of our DNS results reported in this letter satisfy the above-mentioned two conditions. Thus, from the traditional viewpoint of DNS, all of our DNS results should be reliable and can be used as benchmark so...

    work page

  5. [5]

    Poincar´ e, Acta Math.13, A3 (1890)

    H. Poincar´ e, Acta Math.13, A3 (1890)

    work page

  6. [6]

    E. N. Lorenz, J. Atmos. Sci.20, 130 (1963)

    work page 1963

  7. [7]

    R. G. Deissler, Phys. Fluids29, 1453 (1986)

    work page 1986

  8. [8]

    Boffetta and S

    G. Boffetta and S. Musacchio, Phys. Rev. Lett.119, 054102 (2017)

    work page 2017

  9. [9]

    Berera and R

    A. Berera and R. D. J. G. Ho, Phys. Rev. Lett.120, 024101 (2018)

    work page 2018

  10. [10]

    J. Ge, J. Rolland, and J. C. Vassilicos, J. Fluid Mech. 977, A17 (2023)

    work page 2023

  11. [11]

    E. N. Lorenz, Tellus A: Dynamic Meteorology and Oceanography58A, 549 (2006)

    work page 2006

  12. [12]

    Rayleigh, Philos

    L. Rayleigh, Philos. Mag.32, 529 (1916)

    work page 1916

  13. [13]

    Wei and K

    P. Wei and K. Xia, J. Fluid Mech.720, 140 (2013)

    work page 2013

  14. [14]

    Shishkina and S

    O. Shishkina and S. Horn, J. Fluid Mech.790, R3 (2016)

    work page 2016

  15. [15]

    Subramanian, O

    P. Subramanian, O. Brausch, K. E. Daniels, E. Boden- schatz, T. M. Schneider, and W. Pesch, J. Fluid Mech. 794, 719 (2016)

    work page 2016

  16. [16]

    Zwirner and O

    L. Zwirner and O. Shishkina, J. Fluid Mech.850, 984 (2018)

    work page 2018

  17. [17]

    B. Wang, Q. Zhou, and C. Sun, Sci. Adv.6, eaaz8239 (2020)

    work page 2020

  18. [18]

    Q. Wang, S. Xia, B. Wang, D. Sun, Q. Zhou, and Z. Wan, J. Fluid Mech.849, 355 (2018)

    work page 2018

  19. [19]

    Q. Wang, Z. Wan, R. Yan, and D. Sun, Phys. Rev. Fluids 3, 113503 (2018)

    work page 2018

  20. [20]

    Saltzman, J

    B. Saltzman, J. Atmos. Sci.19, 329 (1962)

    work page 1962

  21. [21]

    J. H. Curry, J. R. Herring, J. Loncaric, and S. A. Orszag, J. Fluid Mech.147, 1 (1984)

    work page 1984

  22. [22]

    S. A. Orszag, J. Fluid Mech.41, 363 (1970)

    work page 1970

  23. [23]

    Z.-S. She, E. Jackson, and S. A. Orszag, Nature344, 226 (1990)

    work page 1990

  24. [24]

    Moin and K

    P. Moin and K. Mahesh, Annu. Rev. Fluid Mech.30, 539 (1998)

    work page 1998

  25. [25]

    Scardovelli and S

    R. Scardovelli and S. Zaleski, Annu. Rev. Fluid Mech. 31, 567 (1999)

    work page 1999

  26. [26]

    Huang, L

    J. Huang, L. Duan, and M. M. Choudhari, J. Fluid Mech. 937, A3 (2022)

    work page 2022

  27. [27]

    S. B. Pope,Turbulent Flows(IOP Publishing, 2001)

    work page 2001

  28. [28]

    L. D. Landau and E. M. Lifshitz,Course of Theoreti- cal Physics: Fluid Mechanics (Vol. 6)(Addision-Wesley, Reading, 1959)

    work page 1959