Drift, diffusion and divergence

Laurette S. Tuckerman

REVIEW 2 major objections 4 minor 23 references

Drift, diffusion and divergence

T0 review · 2 major / 4 minor · reviewed 2026-08-07 · deepseek-v4-flash

Challenge this review Re-run · record.json Download PDF Read on arXiv ↗

Pith's one-line read Axisymmetric Taylor-Couette flow at high Reynolds number shows Brownian axial drift of the Taylor-vortex stack, with the diffusion coefficient diverging as the number of rolls approaches a critical value around ten.

desk verdict A competent, readable summary of Feldmann & Avila's Brownian-drift result, but it is a perspective with no original content and a sloppy abstract that distorts the direction of the threshold. read the letter →

arxiv 2506.00621 v1 pith:J3VLLZFS submitted 2025-05-31 physics.flu-dyn

Laurette S. Tuckerman This is my paper

classification physics.flu-dyn PACS 47.27.-i47.27.Cn

keywords Taylor-CouetteflowaxisymmetricturbulenceBrownianmotionvortexdriftdiffusioncoefficientaspect-ratiothresholdlarge-scalestructuresRayleigh-Benardanalogy

verification ladder T0 review T1 audit T2 compute T3 formal

The pith

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

The reading

This paper presents and interprets a recent long-time numerical study of axisymmetric Taylor-Couette flow at high Reynolds number. The study finds that the axial phase of the Taylor-vortex stack carries out Brownian motion, so that the variance of the phase grows linearly in time, and that the effective diffusion coefficient diverges following a power law as the number of rolls approaches a critical value of about ten from above. For shorter systems the rolls become quasi-stationary after an initial adjustment. The paper argues that this matters beyond one flow because roll-like structures that redistribute an imposed gradient appear in many driven systems, and diffusive drift of the same type has been seen in convection and shear flows.

What carries the argument

The central object is the axial phase of the Taylor-vortex stack: a periodic chain of toroidal rolls whose number is fixed during a simulation and whose only remaining continuous motion is a collective shift along the cylinder axis. That phase is treated as a diffusing coordinate, with transport coefficient $D$ defined by the long-time growth of the mean-square phase displacement, and the paper's central observation is the power-law divergence of $D$ as the roll count approaches $\Gamma_c$ from above. Supporting machinery is the analogy between axisymmetric Taylor-Couette flow and two-dimensional Rayleigh-Benard convection, through which different pairs of Reynolds numbers that map to the same Rayleigh number are found to produce the same measured diffusion coefficient.

What would settle it

Run the same axisymmetric configuration at $\Gamma = 10$ for a duration that is an order of magnitude longer and compare the phase variance versus time: if the inferred diffusion coefficient $D = \lim_{t\to\infty} \langle(\Delta z)^2\rangle/(2t)$ changes with run length, or the variance departs from linear growth at long times, then the Brownian law and the divergence near $\Gamma_c$ are finite-time artifacts rather than properties of the flow.

Watch

Extended reading notes

Core claim

The central claim, attributed to the simulations under review, is that in long-time axisymmetric simulations of high-Reynolds-number Taylor-Couette flow the only remaining degree of freedom of the vortex stack, its axial phase, performs genuine diffusive drift. The phase variance grows linearly in time, and the measured diffusion coefficient obeys a power-law divergence as the aspect ratio (equivalently the number of rolls) is reduced toward a critical value $\Gamma_c$ around 10. Below that threshold, as in simulations with eight rolls, the stack settles into quasi-stationary motion with only weak chaotic jiggling about a fixed location. The paper presents this threshold as a sharp dynamical transition whose physical origin is not yet explained, and it reports that the drift persists, though weaker, when the axial flux is set to zero, showing that the motion is not simply advection by a net axial flow.

Load-bearing premise

The reported diffusion coefficients are assumed to be true long-time properties of the axisymmetric flow and not artifacts of how long or how long axially the simulations run; the paper also depends on the hypothesis that axisymmetric dynamics capture the mean transport of the full three-dimensional turbulent flow.

Editorial extensions

If this is right

The vortex stack possesses a collective Brownian mode, and the sharp threshold near ten rolls marks a dynamical transition that any theory of turbulent large-scale structure must explain.
The same diffusive drift observed in convection and shear flows suggests a possible universal length threshold and power-law exponent for drift in roll-carrying systems.
The analogy between axisymmetric Taylor-Couette flow and two-dimensional convection, supported by matching diffusion coefficients across parameter pairs, makes axisymmetric simulations a cheap route to at least some mean transport properties of three-dimensional turbulence.
The axial-flux boundary condition must be specified when reporting phase diffusion coefficients, because the two valid periodic choices give different drift magnitudes.
Near the threshold, the mobility of the large-scale structure changes sharply with aspect ratio, so geometry exerts a strong influence on the dynamics.

Reading between the lines

Editorial extensions of the paper, not claims the author makes directly.

If the divergence is a genuine critical phenomenon, the same power-law exponent should appear for roll drift in other driven systems such as Rayleigh-Benard convection, which would turn the threshold into a universality test that the present simulations do not yet provide.
The supposedly stationary eight-roll state could still diffuse, but with an exponentially small coefficient that is invisible on the simulated time scale, so the threshold might reflect a separation of time scales rather than an exact loss of mobility.
A Green-Kubo relation should connect the phase diffusion coefficient to the time autocorrelation of the net axial force exerted by turbulent fluctuations on the stack; computing that correlation in the existing data would provide a cheaper, independent check of the reported divergence.
A direct three-dimensional simulation at the same Reynolds numbers and aspect ratio could test whether the azimuthally averaged large-scale roll phase also diffuses, thus deciding whether axisymmetric simulations are a valid reduced model or whether the Brownian drift is an artifact of imposing axisymmetry.

Share X Bluesky LinkedIn Reddit HN

Editorial analysis

A structured set of objections, weighed in public.

Desk editor's note, referee report, and a circularity audit.

Desk Editor's Note

read the letter

Dear colleague,

Quick take: this is a well-written perspective on Feldmann & Avila's recent simulation result, not a research paper. It adds no new equations, data, or derivations. Its value is as a clear summary of the Brownian-drift result and a list of open questions. If you haven't read the original, this is a useful entry point; if you have, there is not much here.

What the paper does well: it correctly identifies the key physics—the Taylor-vortex stack has only one remaining degree of freedom (axial phase), and that phase can diffuse. It places the result in context: the Eckhardt–Doering–Whitehead axisymmetric analogy, the Veronis mapping to Rayleigh–Bénard, and the distinction between zero axial flux and zero pressure gradient. The reproduced figure is helpful, and the description of the phase dynamics for different aspect ratios matches the original abstract.

The soft spots are modest but real. First, the abstract states that the diffusion constant 'diverges as the number of rolls is reduced below a critical value.' That is imprecise at best, because below Γ_c the rolls are quasi-stationary; the divergence occurs as the threshold is approached from above. The body is clear, so this is an abstract-level wording problem, but it could mislead a casual reader. Second, the commentary repeats the divergence claim and the power-law behavior without interrogating the evidence. There is no exponent, no fit range, no uncertainty, and no comment on whether the variance growth is truly asymptotic. That is a legitimate concern about the underlying paper, and a perspective that aims to convey the significance of a sharp transition should at least acknowledge that the divergence is inferred from finite-time simulations.

The open questions in section 3 are sensible: why the threshold exists, whether it is generic, and whether the axisymmetric result carries over to three-dimensional turbulence. The paper does not overreach beyond those questions.

Bottom line: a decent commentary, not a research contribution. A serious editor for a letters-style journal would likely desk-reject it for lack of original content; a perspectives venue might accept it after fixing the abstract. I would not spend referee time on it. If you need a quick introduction to this line of work, it is worth a read.

Referee Report

2 major / 4 minor

Summary. This is a perspectives/commentary article by L. S. Tuckerman describing a recent study by Feldmann & Avila (2025) of long-time axisymmetric Taylor-Couette flow. The paper reports that the axial phase of the Taylor-vortex stack performs Brownian drift, with the phase variance growing linearly in time, and that the effective diffusion coefficient diverges following a power law as the aspect ratio (number of rolls) approaches a critical value Gamma_c ~ 10 from above. For a shorter system (Gamma = 8) the rolls become quasi-stationary. The article then frames three open questions: the origin of the sharp threshold, the generality of the drift to other driven flows, and the relevance of the axisymmetric results to fully three-dimensional turbulence. The paper contains no new mathematical derivations or simulations; its content is a summary and contextualization of the cited work.

Significance. If the reported effect is real, it identifies a collective slow mode in a turbulent system and a sharp dynamical transition in the transport coefficient of that mode, which would be a notable result for the fluid dynamics community. The value of this manuscript is in conveying that result to a broader readership and in articulating the open questions that follow. The author is appropriately explicit that the threshold's physical origin is unknown and that the extension to three-dimensional turbulence is an open question. However, the quantitative content of the paper is entirely inherited from Feldmann & Avila (2025); there is no independent verification, no code or data, and no reproduction of the original statistical diagnostics. For a perspectives article, this is acceptable only if the description of the inherited result is precise and appropriately qualified, which is where the manuscript currently has weaknesses.

major comments (2)

[Abstract and Section 2] The abstract states that the diffusion constant 'diverges as the number of rolls is reduced below a critical value,' but Section 2 states that the divergence occurs as the threshold axial length Gamma_c is approached from above, with Gamma = 8 (below the threshold) being quasi-stationary. The direction of the threshold is part of the central phenomenon, and the abstract wording predicts the opposite behavior. Please rephrase to say, for example, 'as the number of rolls is reduced toward a critical value from above,' or otherwise make the abstract consistent with the body.
[Section 2 and Figure 1] The central quantitative claims—linear growth of the phase variance and a power-law divergence of the diffusion coefficient—are reported without any supporting diagnostics from Feldmann & Avila (2025), such as the range of time over which Var(z_R) ~ 2Dt was fitted, the fitted exponent, the statistical uncertainty, or a demonstration that D is stable when the simulation window is extended. This matters because the Gamma = 10 trajectory shown in Figure 1 is dominated by a long unidirectional excursion, precisely the situation in which a finite-window variance estimate can mimic a large diffusion coefficient. Since these numbers are the load-bearing content of the commentary, the author should either quote the original fit statistics or add an explicit sentence stating that the asymptotic, window-converged character of the diffusion coefficient is established in the original paper and is not independently verified here. As written, the paper overstates the strength of the evidence for its central claim.

minor comments (4)

[Introduction] The word 'protypical' in the first sentence of the Introduction should be 'prototypical.'
[Section 2] In the sentence 'the rolls quickly becomes quasi-stationary,' subject and verb do not agree; it should be 'the rolls quickly become quasi-stationary.'
[Figure 1] The x-axis label '0 2t in d2/ν' appears to contain a formatting artifact; the label should likely read 't in d^2/ν' with tick labels at 0 and 2, or similar. Please check the typesetting of the figure axes.
[Section 3] The first sentence of Section 3, 'Several questions are raised by this paper,' could usefully specify that the questions are raised by the results of Feldmann & Avila (2025), not by the present commentary itself, to avoid ambiguity about attribution.

Circularity Check

0 steps flagged · score 0.0 of 10

No circularity: the commentary imports all quantitative claims from cited simulations and derives nothing locally.

full rationale

This is a perspectives/commentary article, not a derivation. Every load-bearing quantitative statement—Brownian axial drift of the Taylor-vortex stack, linear growth of phase variance, divergence of the effective diffusion coefficient as the aspect ratio approaches Gamma_c = 10 from above, and quasi-stationary behaviour at Gamma = 8—is explicitly attributed to Feldmann & Avila (2025), an external simulation study. The paper itself performs no fits, defines no new quantities in terms of the target result, and makes no prediction that is forced by construction. The threshold Gamma_c = 10 is reported as an imported value from the cited simulations, not fitted or derived here. The only self-citation (Edwards et al., 1991, cited for the consequences of axial-flux versus pressure-gradient boundary conditions when the flow is not reflection symmetric) is illustrative and does not carry the argument. The paper also explicitly flags open questions in Section 3, including the unknown origin of the threshold and the applicability of axisymmetric results to three-dimensional turbulence. Accordingly, there are no circular steps; the correct finding is a non-finding of circularity.

Assumptions & free parameters 1 free parameters · 4 assumptions · 0 invented entities

The ledger is nearly empty because this is a commentary: it introduces no free parameters of its own, postulates no entities, and performs no derivation. The only fitted quantity it relies on is the threshold Gamma_c about 10, which it imports from the cited simulations. The axioms are background results and the accuracy assumption for the summarized work.

free parameters (1)

Critical aspect ratio Gamma_c = about 10 (parameters of Feldmann and Avila 2025)
The commentary's central reported phenomenon, the divergence of the drift diffusion coefficient, is keyed to this threshold value imported from the cited simulations; it is not derived in this paper.

assumptions (4)

domain assumption The long-time axisymmetric simulations of Feldmann and Avila (2025) and their measured diffusion coefficients are reported accurately.
The entire content of the paper is a summary of this cited work (section 2); no independent verification, code, or data is offered.
domain assumption Axisymmetric dynamics capture the mean (axisymmetric) properties of three-dimensional turbulent Taylor-Couette flow, following Eckhardt, Doering and Whitehead (2020).
Invoked in section 2 to argue that the axisymmetric problem is a model for turbulent mean transport and that the drift finding has generality; the paper itself flags this as an open question in section 3.
domain assumption The Veronis (1970) analogy between rotating and stratified fluids underlies the Reynolds-to-Rayleigh mapping used to compare diffusion coefficients across parameter pairs.
Invoked in section 2 when reporting that matching Rayleigh number gives matching drift diffusion coefficients for different (Re_S, R_Omega) pairs.
standard math The Navier-Stokes equations govern Taylor-Couette flow in the regime simulated.
Background physical law stated in section 1 via Taylor (1923) agreement; standard for the field.

how reviews work

0 comments

Cite this review

Pith. "Pith review of Drift, diffusion and divergence." pith.science (2026). https://pith.science/paper/J3VLLZFS

@misc{pith2026250600621,
  author       = {Pith},
  title        = {Pith review of: Drift, diffusion and divergence},
  year         = {2026},
  howpublished = {\url{https://pith.science/paper/J3VLLZFS}},
  note         = {Machine review of arXiv:2506.00621}
}

read the original abstract

Turbulent Taylor-Couette flow displays traces of axisymmetric Taylor vortices even at high Reynolds numbers. With this motivation, Feldmann & Avila (2025) carry out long-time numerical simulations of axisymmetric high-Reynolds-number Taylor-Couette flow. They find that the Taylor vortices, using the only degree of freedom that remains available to them, carry out Brownian motion in the axial direction, with a diffusion constant that diverges as the number of rolls is reduced below a critical value.

Figures

Figures reproduced from arXiv: 2506.00621 by the authors.

Discussion (0). Sign in to comment.

Reference graph

Works this paper leans on

23 extracted references · 23 canonical work pages

[1]

& Busse, F.H

Akinaga, T., Generalis, S.C. & Busse, F.H. 2018 Tertiary and quaternary states in the Taylor-Couette system. Chaos Solit.\ Fractals\/ 109 , 107--117

work page 2018
[2]

& Lopez, J.M

Altmeyer, S., Do, Y., Marques, F. & Lopez, J.M. 2012 Symmetry-breaking Hopf bifurcations to 1-, 2-, and 3-tori in small-aspect-ratio counterrotating Taylor-Couette flow. Phys.\ Rev.\ E\/ 86 , 046316

work page 2012
[3]

& Swinney, H.L

Andereck, C.D., Liu, S.S. & Swinney, H.L. 1986 Flow regimes in a circular C ouette system with independently rotating cylinders. J.\ Fluid Mech.\/ 164 , 155--183

work page 1986
[4]

& Iooss, G

Chossat, P. & Iooss, G. 1994 The Couette-Taylor Problem\/ . Springer-Verlag

work page 1994
[5]

1965 Transition in circular C ouette flow

Coles, D. 1965 Transition in circular C ouette flow. J.\ Fluid Mech.\/ 21 , 385--425

work page 1965
[6]

& Altmeyer, S

Deguchi, K. & Altmeyer, S. 2013 Fully nonlinear mode competitions of nearly bicritical spiral or Taylor vortices in Taylor-Couette flow. Phys.\ Rev.\ E\/ 87 , 043017

work page 2013
[7]

2007 Direct numerical simulation of turbulent Taylor--Couette flow

Dong, S. 2007 Direct numerical simulation of turbulent Taylor--Couette flow. J.\ Fluid Mech.\/ 587 , 373--393

work page 2007
[8]

& Zahn, J.-P

Dubrulle, B., Dauchot, O., Daviaud, F., Longaretti, P.-Y., Richard, D. & Zahn, J.-P. 2005 Stability and turbulent transport in Taylor - Couette flow from analysis of experimental data . Phys.\ Fluids\/ 17 , 095103

work page 2005

Show all 23 references

[9]

& Whitehead, J.P

Eckhardt, B., Doering, C.R. & Whitehead, J.P. 2020 Exact relations between Rayleigh-B \'e nard and rotating plane Couette flow in two dimensions. J.\ Fluid Mech.\/ 903 , R4

2020
[10]

& Tuckerman, L.S

Edwards, W., Tagg, R., Dornblaser, B., Swinney, H.L. & Tuckerman, L.S. 1991 Periodic traveling waves with nonperiodic pressure. Eur. J. Mech. B/Fluids\/ 10 , 205--210

1991
[11]

& Avila, M

Feldmann, D. & Avila, M. 2025 Taylor rolls on tour: Slow drift of turbulent large-scale structures in flows with continuous symmetries. J.\ Fluid Mech.\/ 1008 , R1

2025
[12]

& Mutabazi, I

Goharzadeh, A. & Mutabazi, I. 2001 Experimental characterization of intermittency regimes in the C ouette- Taylor system. Eur.\ Phys.\ J.\ B\/ 19 , 157--162

2001
[13]

& Sun, C

Grossmann, S., Lohse, D. & Sun, C. 2016 High- Reynolds number Taylor-Couette turbulence. Ann.\ Rev.\ Fluid Mech.\/ 48 , 53--80

2016
[14]

G., Van Der Veen, R.C., Sun, C

Huisman, S. G., Van Der Veen, R.C., Sun, C. & Lohse, D. 2014 Multiple states in highly turbulent Taylor--Couette flow. Nat.\ Comm.\/ 5 , 3820

2014
[15]

& Eckhardt, B

Kreilos, T., Zammert, S. & Eckhardt, B. 2014 Comoving frames and symmetry-related motions in parallel shear flows. J.\ Fluid Mech.\/ 751 , 685--697

2014
[16]

& Swinney, H.L

Lathrop, D.P., Fineberg, J. & Swinney, H.L. 1992 Transition to shear-driven turbulence in Couette-Taylor flow. Phys.\ Rev.\ A\/ 46 , 6390

1992
[17]

& Hof, B

Lemoult, G., Shi, L., Avila, K., Jalikop, S.V., Avila, M. & Hof, B. 2016 Directed percolation phase transition to sustained turbulence in Couette flow. Nature Physics\/ 12 , 254--258

2016
[18]

& van Saarloos, W

Prigent, A., Gregoire, G., Chat\'e, H., Dauchot, O. & van Saarloos, W. 2002 Large-scale finite-wavelength modulation within turbulent shear flows. Phys.\ Rev.\ Lett.\/ 89 , 014501

2002
[19]

& Hof, B

Shi, L., Avila, M. & Hof, B. 2013 Scale invariance at the onset of turbulence in Couette flow. Phys.\ Rev.\ Lett.\/ 110 , 204502

2013
[20]

1923 Stability of a viscous fluid contained between two rotating cylinders

Taylor, G. 1923 Stability of a viscous fluid contained between two rotating cylinders. Phil.\ Trans.\ Roy.\ Soc.\ A\/ 223 , 289--343

1923
[21]

1970 The analogy between rotating and stratified fluids

Veronis, G. 1970 The analogy between rotating and stratified fluids. Ann.\ Rev.\ Fluid Mech.\/ 2 , 37--66

1970
[22]

& Nagata, M

Weisshaar, E., Busse, F. & Nagata, M. 1991 Twist vortices and their instabilities in the Taylor--Couette system. J.\ Fluid Mech.\/ 226 , 549--564

1991
[23]

& Xia, K.-Q

Xi, H.-D., Zhou, Q. & Xia, K.-Q. 2006 Azimuthal motion of the mean wind in turbulent thermal convection. Phys.\ Rev.\ E\/ 73 , 056312

2006

Pith tools

Reviewed August 7, 2026 · model on record in the stance chip above.

[1] [1]

& Busse, F.H

Akinaga, T., Generalis, S.C. & Busse, F.H. 2018 Tertiary and quaternary states in the Taylor-Couette system. Chaos Solit.\ Fractals\/ 109 , 107--117

work page 2018

[2] [2]

& Lopez, J.M

Altmeyer, S., Do, Y., Marques, F. & Lopez, J.M. 2012 Symmetry-breaking Hopf bifurcations to 1-, 2-, and 3-tori in small-aspect-ratio counterrotating Taylor-Couette flow. Phys.\ Rev.\ E\/ 86 , 046316

work page 2012

[3] [3]

& Swinney, H.L

Andereck, C.D., Liu, S.S. & Swinney, H.L. 1986 Flow regimes in a circular C ouette system with independently rotating cylinders. J.\ Fluid Mech.\/ 164 , 155--183

work page 1986

[4] [4]

& Iooss, G

Chossat, P. & Iooss, G. 1994 The Couette-Taylor Problem\/ . Springer-Verlag

work page 1994

[5] [5]

1965 Transition in circular C ouette flow

Coles, D. 1965 Transition in circular C ouette flow. J.\ Fluid Mech.\/ 21 , 385--425

work page 1965

[6] [6]

& Altmeyer, S

Deguchi, K. & Altmeyer, S. 2013 Fully nonlinear mode competitions of nearly bicritical spiral or Taylor vortices in Taylor-Couette flow. Phys.\ Rev.\ E\/ 87 , 043017

work page 2013

[7] [7]

2007 Direct numerical simulation of turbulent Taylor--Couette flow

Dong, S. 2007 Direct numerical simulation of turbulent Taylor--Couette flow. J.\ Fluid Mech.\/ 587 , 373--393

work page 2007

[8] [8]

& Zahn, J.-P

Dubrulle, B., Dauchot, O., Daviaud, F., Longaretti, P.-Y., Richard, D. & Zahn, J.-P. 2005 Stability and turbulent transport in Taylor - Couette flow from analysis of experimental data . Phys.\ Fluids\/ 17 , 095103

work page 2005