Time discretization of a semi-discrete scheme for 3D Chemotaxis-Navier-Stokes system driven by transport noise

Boris Jidjou Moghomye; Erika Hausenblas; Paul Andre Razafimandimby

REVIEW 2 major objections 5 minor 23 references

Time discretization of a semi-discrete scheme for 3D Chemotaxis-Navier-Stokes system driven by transport noise

T0 review · 2 major / 5 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 For the spatially truncated 3D chemotaxis-Navier-Stokes system with transport noise, a semi-implicit Euler scheme is well posed and converges in law to a martingale solution when γ² < ε/484.

desk verdict The paper targets a genuinely new stochastic chemotaxis-fluid system, but the central velocity energy estimate contains an index error in the noise term, so the existence proof as written does not go through. read the letter →

arxiv 2506.03658 v1 pith:UNKUHD2V submitted 2025-06-04 math.NA cs.NAmath.AP

Erika Hausenblas , Boris Jidjou Moghomye , Paul Andre Razafimandimby This is my paper

classification math.NAcs.NAmath.AP MSC 35R6035Q3560H1576M3586A05

keywords StochasticNavier-StokesequationsChemotaxissystemmartingalesolutionsweaktimediscretizationsemi-implicitEulerschemetransportnoiseGalerkintruncation

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

Chemotaxis—organisms moving toward a chemical signal—happens in a fluid, and this paper treats a stochastic version of the coupled chemotaxis-Navier-Stokes equations in three dimensions, with noise carried by the flow and by the chemical concentration. The paper’s aim is an existence theorem for the spatially Galerkin-truncated version of this system: under Assumption 2.1 and the small-noise condition $\gamma^2<\varepsilon/484$, for every truncation level $m$ the system has a martingale solution. The proof is numerical in spirit: it writes a semi-implicit Euler scheme in time, proves each time step is solvable, obtains uniform a priori bounds, and shows the laws of the interpolated discrete solutions are tight and pass to a limit that satisfies the original equations. The conclusion is that the semi-discrete model is nonempty and is consistently approximated by a concrete time-stepping algorithm; uniqueness of the limiting solution is left open.

What carries the argument

The load-bearing object is the semi-implicit Euler scheme (3.1): given the previous state, the new velocity is found from an implicit Stokes-type equation with the previous transport term, and the new concentration and density are found from implicit Neumann-diffusion equations with explicit couplings, using the $C^2$ truncation $\Phi_m$ of the identity and the cutoff $F_m(x)=\min(1,m/x)$ to tame the chemotaxis and noise terms. The noise term in the velocity equation is a transport term; after testing with the solution it vanishes identically (line (3.23)), so the discrete energy estimate for $u$ needs no smallness condition. The condition $\gamma^2<\varepsilon/484$ enters through Lemma 3.5(4): the discrete energy inequality for the chemical gradient has coefficient $3\varepsilon/2-24\gamma^2$, which must stay positive. The scheme is then extended to continuous-time interpolants, whose uniform bounds feed the tightness and the Skorokhod-type limit passage.

What would settle it

Supply the omitted proof of Lemma 4.1: the asserted bounds, for instance $\mathbb{E}\sup_{0\le s\le T}\|\hat u_N(s)\|_{0,2}^{2p}\le R_m$ for $p\in[1,2]$, must follow from Lemma 3.5 with a constant $R_m$ independent of $N$; a direct numerical evaluation for a fixed small $m$ and increasing $N$ showing any of these sup or integral quantities growing would break the tightness Lemma 4.5 and with it the existence theorem.

Watch

Extended reading notes

Core claim

Proposition 2.3 is the central claim: once $\gamma^2<\varepsilon/484$, problem (1.2) admits a martingale solution for every fixed $m\in\mathbb{N}$, in the sense of Definition 2.2—a solution on an enlarged probability space with its own Wiener processes, satisfying the integral form of the equations. The construction goes through the time-discrete scheme (3.1): a semi-implicit Euler step in which the velocity, chemical concentration, and organism density are solved with implicit diffusion and explicit nonlinear couplings. The paper proves pathwise solvability of each step by a Leray-Schauder fixed-point argument, $\mathcal{F}_{t_\ell}$-measurability of the discrete variables, uniform-in-$N$ estimates for all three components (Lemmas 3.5 and 3.6), tightness of the laws of the continuous and piecewise-constant interpolants (Lemma 4.5), and passage to the limit on a new probability space via a Skorokhod-type representation, with the error terms shown to vanish (Lemmas 4.7–4.10). Along the way the transport structure makes the velocity-noise term vanish in the energy identity, and the smallness assumption keeps the chemical energy inequality coercive.

Load-bearing premise

The load-bearing premise is that the uniform interpolant estimates in Lemma 4.1 are valid as stated; the paper says their proof is "very similar to [?, Proof of Lemma 4.1]" and omits it, even though those bounds are exactly what turn the discrete estimates into the tightness and integrability needed to construct the limit solution.

Editorial extensions

If this is right

For each fixed Galerkin level $m$, the truncated system (1.2) has at least one martingale solution, so the finite-dimensional model one would simulate is mathematically nonempty.
The semi-implicit Euler scheme (3.1) is well posed step by step, pathwise measurable, and its solutions obey uniform-in-$N$ bounds, so it can be implemented without the time step causing blow-up.
The laws of the interpolated discrete processes converge as $N\to\infty$ to the law of a martingale solution, making the time discretization consistent with the stochastic PDE in the sense of convergence in law.
The explicit threshold $\gamma^2<\varepsilon/484$ is the regime in which the argument closes; outside it, the energy estimate for the chemical gradient fails and the paper gives no existence statement.
Uniqueness is not established: the solution is a martingale solution on an enlarged probability space, so pathwise uniqueness and strong solutions remain open.

Reading between the lines

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

Because the Galerkin level $m$ is fixed throughout, the same time-discretization machinery should transfer to other finite-dimensional truncations of fluid models with transport noise; the smallness condition is tied to the chemical diffusion, not to the fluid viscosity.
If a positivity-preserving spatial discretization of the chemotaxis part were supplied, the uniform estimates derived here are exactly what would be needed to push convergence from the time-discrete to a fully discrete scheme, which the paper names as its intended application.
Should uniqueness of martingale solutions ever be proved, the convergence-in-law result would strengthen to a statement about the numerical scheme selecting the unique solution, removing the enlarged-probability-space caveat.

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

Quick take: the paper is a competent attempt at an interesting target, but the central energy estimate for the velocity equation has an index error that is load-bearing. The reader's conditional verdict is too generous; the proof as written fails.

What's new: the target system — a 3D chemotaxis-Navier-Stokes system with transport noise in both the velocity and chemical equations, reduced to a Galerkin truncation — is new. The semi-implicit Euler scheme and the fixed-point construction for well-posedness are standard but correctly adapted. The authors are honest that uniqueness is open and that m→∞ convergence is not proved.

The problem is in the proof of Lemma 3.5, equation (3.23). The discrete noise term is written as 2α Σ ΔW_k ∫ ∂_j u^{i-1}_k u^i_j dx. That expression does vanish by integration by parts because div u^i = 0. But it is not the term that arises from f(v)z = (∇v)z under Assumption 2.1(i). The actual term is 2α Σ ΔW_k ∫ ∂_k u^{i-1}_j u^i_j dx, which does not vanish pathwise. So the pathwise energy inequality (3.22)_1 is unsupported. At best one could try to close an expectation bound with an extra α²hE‖∇u‖² term, requiring a smallness condition on α that Proposition 2.3 does not state.

This is not cosmetic. The estimate (3.22)_1 is used in Lemma 3.6, then in Lemma 4.1, Lemma 4.3, and the entire tightness and passage-to-limit argument. Without it, the existence statement for fixed m does not follow from the given proof.

The missing proofs of Lemmas 4.1 and 4.2 (one with an unresolved '[?]' reference) are real but secondary; they would matter even if the energy estimate were fixed.

Bottom line: the paper deserves a serious referee because the issue is subtle and the framework is otherwise standard, but the current version should not be accepted. The authors need to correct the noise term, prove or properly cite Lemmas 4.1 and 4.2, and check whether the conclusion requires an additional smallness condition on the velocity noise intensity. If they can do that, the result would be credible.

I would send it back as a major revision, not desk reject. I wouldn't cite it in its current form.

Referee Report

2 major / 5 minor

Summary. The paper proposes a semi-implicit Euler time discretization for a Galerkin-truncated three-dimensional chemotaxis--Navier--Stokes system with Stratonovich transport noise in the velocity equation and multiplicative noise in the chemoattractant equation. For each fixed Galerkin truncation index m, the authors aim to prove existence of a martingale solution to the semi-discrete system (1.2) by establishing well-posedness of the discrete scheme, deriving uniform a priori estimates, proving tightness of suitable interpolants, and passing to the limit as the time step tends to zero. The main result, Proposition 2.3, is stated under the small-noise condition η² < ε/484.

Significance. If the result is correct, it gives a rigorous numerical-analysis route to martingale solutions for a biologically motivated stochastic fluid-chemotaxis model, and the paper contains a substantial amount of useful technical work: an explicit semi-implicit scheme, a Leray--Schauder fixed-point argument for the implicit step, a measurable-selection argument for the discrete solutions, detailed uniform estimates for the chemoattractant and organism densities, and a Jakubowski--Skorokhod compactness framework. However, the central velocity estimate contains an index error that invalidates the claimed cancellation of the transport-noise term, and two interpolation lemmas that are load-bearing for tightness and limit identification are stated without proof, one with an unresolved placeholder citation. The significance is therefore conditional until these points are repaired.

major comments (2)

[§3, Lemma 3.5, Eq. (3.23)] The cancellation of the discrete transport-noise term is based on an index swap. By Assumption 2.1(i), f(v)z = (∇v)z = (z·∇)v, so the noise term tested against 2u^i equals 2α Σ_{k,j} Δ^iW_k ∫ (∂_k u^{i-1}_j) u^i_j dx. Equation (3.23), however, computes 2α Σ_{j,k} Δ^iW_k ∫ (∂_j u^{i-1}_k) u^i_j dx; after integration by parts this displayed expression vanishes because Σ_j ∂_j u^i_j = 0, but the correct expression yields -2α ∫ u^{i-1}·(Δ^iW·∇)u^i dx, which is not identically zero. As a consequence the pathwise estimate (3.22)_1 is unsupported. That estimate underpins Lemma 3.6, Lemma 4.1, Lemma 4.3, and the tightness and passage-to-limit argument, so this is a load-bearing gap. The same erroneous cancellation reappears in the proof of Lemma 3.6, where the noise term is declared to vanish because ∇·(u^{ℓ+j}-u^ℓ)=0. The authors should either correct the discretization so that the cancellation is valid, or replace (3.22)_1 by a genuinely estimated bound, which will likely require an additional smallness condition on α and a stochastic Grönwall argument; Proposition 2.3 as stated contains no smallness condition on α.
[§4, Lemmas 4.1 and 4.2] Lemma 4.1 is stated with its proof omitted and attributed to “[?, Proof of Lemma 4.1]”, but no such reference appears in the bibliography. Lemma 4.2 is also stated without proof; the text says “We also state the following proof”, which appears to mean “lemma”, and no argument is supplied. These lemmas are load-bearing: Lemma 4.1 provides the uniform L^p and integrability estimates used for tightness in Lemma 4.5, and Lemma 4.2 provides the vanishing of the interpolant errors used to identify the limit processes in (4.8)–(4.9). Since Lemma 4.1 also depends on Lemma 3.5, whose velocity estimate is in doubt, the convergence argument is incomplete. The paper should supply complete proofs or precise statements from an available reference.

minor comments (5)

[§3, proof of Lemma 3.5] The inequality ‖∇² c^{i-1}‖² ≤ ‖Δ c^{i-1}‖² is stated for convex O without proof or reference; since it is used in a key estimate, a short justification or a citation to a standard elliptic regularity result would be helpful.
[§4, before Lemma 4.2] The sentence “We also state the following proof, because its uses Lemma 3.5 and very similar to [16, Proposition 3.11]” is ungrammatical and should be rewritten, for example as “We also state the following lemma, whose proof uses Lemma 3.5 and is very similar to [16, Proposition 3.11].”
[Assumption 2.1, observation after the assumption] The example g_k(x) = 1_\bar{O}(x)e_k on O and 0 on ∂O is not W^{1,∞} as written, because the indicator function of an open set is discontinuous at the boundary; this example should be replaced by a smooth cutoff construction.
[Definition 2.2] In item (3) of Definition 2.2 the notation “\bar{P}^m-a.s.” is used, while the probability measure is elsewhere denoted by “\bar{P}”; the notation should be harmonized.
[Throughout] The unresolved placeholder “[?]” in the proof of Lemma 4.1 should be corrected before resubmission; it is not merely a stylistic issue because the proof is omitted.

Circularity Check

0 steps flagged · score 0.0 of 10

No circularity: the existence proof proceeds by a semi-implicit Euler scheme and standard compactness arguments, without fitting parameters or importing the target result.

full rationale

The central claim, Proposition 2.3, asserts existence of a martingale solution for the fixed-Galerkin system (1.2). The proof introduces an independent time-discrete scheme (3.1), proves well-posedness by fixed-point arguments, derives uniform estimates in Lemma 3.5, establishes tightness of the interpolants, and passes to the limit. None of these steps defines the target solution in terms of the approximating sequence in a way that makes the conclusion an input by construction. The semi-implicit scheme is standard and its limiting equation is the original system (1.2), not a reformulation of the existence statement. Citations to prior work, including [16] by one of the authors, are used for auxiliary compactness and passage-to-the-limit arguments or measurable selection results; they do not assert Proposition 2.3 and are not equivalent to it. The omitted proofs of Lemmas 4.1 and 4.2 and the possible algebraic issue in (3.23) concerning the transport-noise cancellation are correctness risks or proof gaps, not circularity: even if they invalidate the estimates, they do not make the conclusion identical to the assumptions. The paper is self-contained in the sense required for a non-circular mathematical existence proof, and there is no fitted parameter, no renamed known result, and no self-citation chain that forces the claimed martingale solution.

Assumptions & free parameters 2 free parameters · 5 assumptions · 0 invented entities

No data-fitting is involved. The only adjusted quantities are structural parameters of the scheme (m, ε_m, F_m, θ_m) and the smallness restriction on the noise intensity; none is tuned to observations. The central theorem rests on Assumption 2.1, convex-domain elliptic regularity, and standard stochastic compactness results. The paper introduces no new physical entities.

free parameters (2)

Galerkin truncation index m = arbitrary fixed integer m
The theorem is stated for each fixed m; all estimates depend on m, but m is not fitted to data.
Cutoff parameters ε_m = 8/m and F_m(x)=min(1,m/x) = ε_m = 8/m; F_m threshold m
Introduced by hand to keep θ_m+ε_m positive and to damp the noise term in the chemical equation; they are structural choices, not fitted constants.

assumptions (5)

domain assumption Assumption 2.1(i)-(iv): transport noise fields g_k are W^{1,∞}, divergence-free, vanish on ∂O, with covariance G(x,x)=I_{R^3} and f(v)z=(∇v)z.
Defines the Stratonovich transport noise and the Itô correction structure used throughout the estimates.
domain assumption Initial data regularity: u0∈H, n0^m,c0^m∈H^2.
Assumption (2.1); needed for the first step of the induction and for elliptic estimates involving A1.
domain assumption Small noise intensity condition: ₝² < ε/484.
Assumption in Proposition 2.3; used in Lemma 3.5, estimate (3.22)_4, to absorb stochastic integral terms.
domain assumption O is a bounded convex domain, with Stokes eigenbasis in H^2∩V and Neumann Laplacian regularity.
Convexity gives the Hessian bound (3.25); the eigenbasis defines H_m; elliptic regularity from [13] is applied.
standard math Standard existence and compactness results: Lax-Milgram, Leray-Schauder fixed point theorem, BDG inequality, Jakubowski-Skorokhod representation.
Used in Lemmas 3.2-3.4 and Section 4; external results from [13, 12, 14, 15].

how reviews work

0 comments

Cite this review

Pith. "Pith review of Time discretization of a semi-discrete scheme for 3D Chemotaxis-Navier-Stokes system driven by transport noise." pith.science (2026). https://pith.science/paper/UNKUHD2V

@misc{pith2026250603658,
  author       = {Pith},
  title        = {Pith review of: Time discretization of a semi-discrete scheme for 3D Chemotaxis-Navier-Stokes system driven by transport noise},
  year         = {2026},
  howpublished = {\url{https://pith.science/paper/UNKUHD2V}},
  note         = {Machine review of arXiv:2506.03658}
}

read the original abstract

This work is devoted to the convergence of a time-discrete numerical scheme of a semi-discretization model arising from biology, consisting of a chemotaxis equation coupled with a Galerkin approximation of Navier-Stokes system driven by transport noise in a three-dimensional bounded and convex domain. We propose a semi-implicit Euler numerical scheme approximating the infinite dimensional model, for which we study the well-posedness and derive some uniform estimates for the discrete variables

Discussion (0). Sign in to comment.

Reference graph

Works this paper leans on

23 extracted references · 23 canonical work pages

[1]

Ba nas, Z

L. Ba nas, Z. Brzez ´niak, M. Neklyudov, A. Prohl, Stochastic ferromagnetism: Analysis and Computation. De Gruyter Studies in Mathematics, Berlin, 2014

work page 2014
[2]

Ba nas, Z

L. Ba nas, Z. Brzez ´niak, M. Neklyudov, A. Prohl, A convergent finite-element-based discretization of the stochastic Lan- dau–Lifshitz–Gilbert equation.IMA J. Numer. Anal.34(2014) 502–549

work page 2014
[3]

Bensoussan, R

A. Bensoussan, R. Temam, Equations stochastiques du type Navier–Stokes.J. Funct. Anal.13(1973) 195–222

work page 1973
[4]

Billingsley, Convergence of probability measures.John Wiley and Sons, Chicago, 2013

P. Billingsley, Convergence of probability measures.John Wiley and Sons, Chicago, 2013

work page 2013
[5]

Blanchet, J

A. Blanchet, J. A. Carrillo, N. Masmoudi. Infinite time aggregation for the critical Patlak-Keller-Segel Model inR 2 . Comm. on Pure and Appl. Math.LXI(2008) 1449-1481

work page 2008
[6]

Brze´ zniak, E

Z. Brze´ zniak, E. Motyl, Existence of a martingale solution of the stochastic Navier–Stokes equations in unbounded 2D and 3D domains.J. Differ. Equ.254(2013) 1627–1685

work page 2013
[7]

Caraballo, J

T. Caraballo, J. Real, P. E. Kloeden, Unique strong solutions and V-attractors of a three dimensional system of globally modified Navier–Stokes equations.Adv. Nonlinear Stud.6(2006) 411–436

work page 2006
[8]

Debussche, N

A. Debussche, N. Glatt-Holtz, R. Temam, Local martingale and pathwise solutions for an abstract fluids model.Physica D240(2011) 1123–1144

work page 2011

Show all 23 references

[9]

De Bouard, A

A. De Bouard, A. Debussche, A semi-discrete scheme for the stochastic nonlinear Schrödinger equation.Numer. Math. 96(2004) 733-770

2004
[10]

Flandoli, D

F. Flandoli, D. Gatarek, Martingale and stationary solutions for stochastic Navier-Stokes equations.Probab. Theory Re- lated Fields102(1995) 367-391

1995
[11]

Glatt-Holtz, R

N. Glatt-Holtz, R. Temam, C. Wang, Time discrete approximation of weak solutions to stochastic equations of geophysical fluid dynamics and applications.Chin. Ann. Math.38B(2017) 425–472

2017
[12]

Granas, J

A. Granas, J. Dugundji, Fixed Point Theory.Springer, New York, 2003

2003
[13]

Grisvard, Elliptic problems in nonsmooth domains.Pitman, Boston, 1995

P. Grisvard, Elliptic problems in nonsmooth domains.Pitman, Boston, 1995

1995
[14]

Ichikawa, Stability of semilinear stochastic evolution equations.J

A. Ichikawa, Stability of semilinear stochastic evolution equations.J. Math. Anal. Appl.1(1982) 12-44

1982
[15]

Jakubowski, The almost sure Skorokhod representation for subsequences in nonmetric spaces.Theory Probab

A. Jakubowski, The almost sure Skorokhod representation for subsequences in nonmetric spaces.Theory Probab. Appl. 42(1998) 167–174

1998
[16]

P. A. Razafimandimby, Grade-two fluids on non-smooth domain driven by multiplicative noise Existence, uniqueness and regularity.J. Differential Equations,263(2017) 3027–3089

2017
[17]

P. A. Razafimandimby, M. Sango, Existence and large time behavior for a stochastic model ofmodified magnetohydrody- namic equations.Z. Angew. Math. Phys.66(2015) 2197-2235

2015
[18]

Simon, Compact sets in the spaceL p p0; T;Bq.Ann

J. Simon, Compact sets in the spaceL p p0; T;Bq.Ann. Mat. Pure Appl.146(1987) 65-96

1987
[19]

Simon, Sobolcv, Besov and Nikolskii fractional Spaces Imbeddings and comparisons for vector valued spaces on an interval.Ann

J. Simon, Sobolcv, Besov and Nikolskii fractional Spaces Imbeddings and comparisons for vector valued spaces on an interval.Ann. Mat. Pura Appl.157(1990) 117-148

1990
[20]

M. E. Taylor, Partial Differential Equation I Basic Theory, Second ed. ,Applied Mathematical sciences, Vol 115, Springer, Carolina, 2011

2011
[21]

N. N. Vakhania, V . I. Tarieladze, S. A. Chobanyan, Probability Distributions on Banach spaces,D. Reidel Publishing Company, Holland, 1987

1987
[22]

J. Zhai, T. Zhang, 2D stochastic Chemotaxis-Navier-Stokes system,J. Math. Pures Appl.138(2020) 307-355

2020
[23]

Zhang, B

L. Zhang, B. Liu, Global martingale weak solution for the three-dimensional stochastic chemotatix-Navier-Stokes system with Lévy processes.J. Funct. Anal.286(2024) 110337

2024

Pith tools

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

[1] [1]

Ba nas, Z

L. Ba nas, Z. Brzez ´niak, M. Neklyudov, A. Prohl, Stochastic ferromagnetism: Analysis and Computation. De Gruyter Studies in Mathematics, Berlin, 2014

work page 2014

[2] [2]

Ba nas, Z

L. Ba nas, Z. Brzez ´niak, M. Neklyudov, A. Prohl, A convergent finite-element-based discretization of the stochastic Lan- dau–Lifshitz–Gilbert equation.IMA J. Numer. Anal.34(2014) 502–549

work page 2014

[3] [3]

Bensoussan, R

A. Bensoussan, R. Temam, Equations stochastiques du type Navier–Stokes.J. Funct. Anal.13(1973) 195–222

work page 1973

[4] [4]

Billingsley, Convergence of probability measures.John Wiley and Sons, Chicago, 2013

P. Billingsley, Convergence of probability measures.John Wiley and Sons, Chicago, 2013

work page 2013

[5] [5]

Blanchet, J

A. Blanchet, J. A. Carrillo, N. Masmoudi. Infinite time aggregation for the critical Patlak-Keller-Segel Model inR 2 . Comm. on Pure and Appl. Math.LXI(2008) 1449-1481

work page 2008

[6] [6]

Brze´ zniak, E

Z. Brze´ zniak, E. Motyl, Existence of a martingale solution of the stochastic Navier–Stokes equations in unbounded 2D and 3D domains.J. Differ. Equ.254(2013) 1627–1685

work page 2013

[7] [7]

Caraballo, J

T. Caraballo, J. Real, P. E. Kloeden, Unique strong solutions and V-attractors of a three dimensional system of globally modified Navier–Stokes equations.Adv. Nonlinear Stud.6(2006) 411–436

work page 2006

[8] [8]

Debussche, N

A. Debussche, N. Glatt-Holtz, R. Temam, Local martingale and pathwise solutions for an abstract fluids model.Physica D240(2011) 1123–1144

work page 2011