{"id":"2bc74494-a116-48da-89bf-aef4f900026d","arxiv_id":"1908.11481","paper_version":2,"verdict":"CONDITIONAL","confidence":"MODERATE","novelty_score":5.0,"correctness_risk":"medium","formal_verification":"none","parameter_count":1,"one_line_summary":"The paper formulates LA SALT stochastic fluid equations whose mean field satisfies a closed Lie-Laplacian Navier-Stokes equation, and establishes well-posedness and fluctuation variance dynamics.","lead":"This paper defines a new class of stochastic fluid equations, LA SALT, in which the velocity field is transported by white noise plus its own ensemble-average velocity. Because the average velocity appears in the transport, the expected motion obeys a closed, regularized equation reminiscent of Navier-Stokes, and the paper proves local and conditional global well-posedness.","discovery_kind":"extension","skeptic_critique":{"model":"deepseek-v4-flash","headline":"Theorem 1's proof leaves a regularity gap: the cited stochastic-transport theorem requires C^{n+1} drift coefficients, but the LLNS solution E[u] only provides C^n under the stated assumptions.","rationale":"The reader's weakest_assumption correctly identifies Assumption 1 and the reliance on [41, Thm 3.3]. I focused on the latter because it is the less explicit and potentially more consequential gap: the proof of Theorem 1, which is the paper's central well-posedness result, is only a sketch and its stated regularity conditions appear internally inconsistent. The index mismatch between n = floor(m − d/2) and the requirement m > d/2 + n + 1 is a concrete signal that the proof needs either a corrected Sobolev index or a weaker-coefficient version of the external theorem. This does not disprove the LA SALT construction, but it means the central theorem is not yet justified as stated. The reader's verdict CONDITIONAL remains appropriate because the gap is likely fixable (e.g., by lowering the solution regularity or citing a sharper transport theorem), and the rest of the paper's algebraic derivation of LLNS is sound. I also noted the factor-of-2 error in Definition 1's weak formulation (the Itô correction should be −(1/2)Σ(L_ξ u, L_ξ φ), not −Σ(L_ξ u, L_ξ φ)), but that appears to be a typo that does not affect the strong-form equation (3.13) or the main conceptual claim; hence I did not make it the primary concern. The missing time integral in Eq. (3.9) is a side-result error, not load-bearing for the central theorem.","tokens_in":36887,"tokens_out":40323,"duration_ms":330637,"concrete_test":"Read Theorem 3.3 of [41] (and Theorem 3.1 of [42]) and determine the minimal Hölder/Sobolev regularity of the drift coefficient b required for well-posedness of the linear SPDE du = (b·∇u + (∇b)^T u + (1/2)Σ(ξ·∇)^2 u)dt − Σ(ξ·∇u + (∇ξ)^T u) dW in H^s. Then for s = n−1 (the claim of Theorem 1), check whether the available regularity v ∈ H^m with m > d/2 + 2 and m − d/2 possibly integer is sufficient. If C^{s+1} is required, recompute the largest attainable s from the stated assumptions and compare it with the H^{n−1} conclusion; if the result is lower, Theorem 1 needs stronger assumptions or a different proof.","verdict_should_be":"UNCHANGED","load_bearing_attack":"Section 3.3 reduces Theorem 1 to (i) Theorem 2 for the deterministic LLNS equation (3.15), giving v=E[u] in L∞_T H^m, and (ii) an external linear SPDE theorem ([41, Thm 3.3]) for the fluctuation equation (3.13) with drift coefficient v. The proof sketch states that this theorem needs v in L∞_T C^{n+1} and obtains it from m > d/2 + n + 1. But with n defined in Theorem 1 as floor(m − d/2), the inequality m > d/2 + n + 1 is impossible, since m − d/2 < floor(m − d/2) + 1. The natural fix is to shift the Sobolev index of the solution to N = n − 1, so that v ∈ C^{N+1} = C^n suffices; but this shift is never stated, and the theorem's conclusion u ∈ L^2_ω L∞_T H^{n−1} would then be H^{N−1} = H^{n−2}, one order lower than claimed. Alternatively, the external theorem might only require v ∈ C^n for H^n solutions, but that must be verified against [41, Thm 3.3]. The paper does not perform this verification, so Theorem 1 as stated is not established.","agreement_with_reader":"partial"},"referee_report":{"model":"deepseek-v4-flash","summary":"The paper introduces a class of stochastic fluid equations, termed Lagrangian averaged stochastic advection by Lie transport (LA SALT), in which the drift velocity of the stochastic transport vector field is the expectation E[u] of the velocity over noise realizations, while the noise terms are Stratonovich Lie transport along fixed divergence-free vector fields. The authors derive the Euler--Poincaré and Lie--Poisson formulations, show that the expectation field solves a closed deterministic equation with a Lie--Laplacian second-order operator (LLNS), compute local and integrated variance dynamics for fluctuations, and discuss examples including Euler, Burgers, Camassa--Holm, rigid body, and MHD. The paper also states well-posedness theorems: local existence for LA SALT Euler in Sobolev spaces in d=2,3, global existence in d=2, and global existence in d=3 for sufficiently large noise ellipticity constant. The main advertised regularity mechanism is that the mean-field nonlocality in probability space converts the otherwise conservative Euler dynamics into a regularized LLNS equation for the expectation.","tokens_in":37127,"tokens_out":8795,"duration_ms":84294,"significance":"If the results are correct, the paper provides an appealing geometric framework that connects stochastic fluid dynamics, mean-field (McKean--Vlasov) nonlinearity, and deterministic Navier--Stokes-type regularization. The derivation of the closed LLNS equation from the LA SALT system, the explicit link to the Constantin--Iyer stochastic Weber velocity representation for constant noise correlations, and the geometric conservation structure (Kelvin's theorem, Casimirs, helicity, enstrophy) are valuable and clearly presented. The variance equations in Section 2.3 are explicit and go beyond the usual statement of the mean-field closure. The paper also credits the earlier work [3] for the LLNS equation and the statistical Kelvin theorem, which is appropriate. The main weakness is that the well-posedness theorem for LA SALT Euler is only sketched and, as stated, contains a regularity-gap in the proof that the hypotheses of the cited stochastic transport theorem are satisfied. This is load-bearing for the claimed well-posedness and regularization result, so the theorem needs repair or reformulation.","major_comments":[{"comment":"The proof asserts that if u0∈H^m, f∈L2_T H^{m−1}, ξ∈C^{m+2}, and m>d/2+n+1, then Theorem 2 gives E[u]∈L∞_{T*} C^{n+1}_x with n=⌊m−d/2⌋. This inequality is impossible: by definition n≤m−d/2<n+1, so m<d/2+n+1. The same false inequality is used in the pressure estimate immediately below, where the paper requires m−1>n+d/2. Consequently the hypotheses of the cited linear stochastic transport theorem [41, Thm 3.3] are not verified, and Theorem 1 as stated is not established. The authors should either shift the Sobolev index n so that the required drift regularity follows, restate the theorem with the additional loss of regularity, or prove directly that the drift regularity demanded by [41, Thm 3.3] follows from Theorem 2.","section":"§3.3, Sketch of Proof of Theorem 1"},{"comment":"The displayed identity is dimensionally inconsistent: it equates ∫(E[ω_t])^2 dA with ∫(E[ω_0])^2 dA minus a term evaluated at time t, with no time integral. Integrating the differential identity d/dt ∫(E[ω])^2 = −Σ_k ∫(£_{ξ(k)}E[ω])^2 dA gives an extra ds integral, and the integrand should be evaluated at time s, not t. Therefore the sentence immediately after (3.9) that the magnitude |E[ω]| 'will decay to zero' does not follow from the displayed equation as written. The corrected identity should be stated and the decay conclusion re-derived; equation (3.10) can be recovered after this correction.","section":"§3.1, Eq. (3.9)"}],"minor_comments":[{"comment":"The continuity assertion for LA SALT states the solution map as taking values in L2_ω C_{T*} H^{n−1} ∩ L2_ω L∞_{T*} H^n, but Theorem 1 only provides u∈L2_ω L∞_{T*} H^{n−1}, weak continuity in H^{n−1}, and strong continuity in H^{n−2}; the regularity and continuity statements in the remark and the theorem should be made consistent.","section":"§3.3, Remark 7"},{"comment":"The definition of L_p T E writes L^p([0,T], B([0,T]), Ω; E) and then says 'Ω is the Lebesgue measure on R'; this should presumably be the Lebesgue measure on [0,T].","section":"§3.3, notation for L_p T E"},{"comment":"The display of the LLNS equation has unbalanced parentheses in the Lie--Laplacian term; the closing parenthesis of the second Lie derivative is missing.","section":"§1.1, Eq. (1.13)"},{"comment":"In the equation for the advected quantity a′, the expression '1/2 d|a′|_{L2}' is missing the square on the norm; it should be 1/2 d|a′|²_{L2}.","section":"§2.3, Eq. (2.14)"}],"recommendation":"major_revision","confidential_remarks":"The modeling and derivation portions of the paper are strong and likely of interest to the journal's readership. The decision hinges on whether the well-posedness theorem can be repaired: either by verifying the hypotheses of the external stochastic transport theorem with a corrected regularity index, or by reformulating Theorem 1 with the regularity loss that the proof can actually support. I would not require a fully detailed proof of Theorem 2, but the gap in Theorem 1 must be addressed before publication."},"author_rebuttal":null,"desk_editor":{"model":"deepseek-v4-flash","letter":"The paper does a real service to the SALT program. It takes the mean-field (McKean) idea that first appeared in Drivas–Holm [3] and the Constantin–Iyer representation and packages it into a general Euler–Poincaré/Lie–Poisson framework with advected quantities. The expectation closure is transparent: the drift is E[u] by construction, so the mean satisfies a deterministic equation, now with Lie–Laplacian 'dissipation'. The fluctuation equations and the variance dynamics (2.16) are a genuine addition—I don't know a prior reference for those. The examples (rigid body, Burgers, CH, MHD) illustrate the breadth. Given the prior work, the main new content is the general formulation plus the variance equations and the well-posedness theorems; that's a reasonable, incremental contribution.\n\nThe soft spots are real but not fatal. First, the proof of Theorem 1 as written has a regularity index gap. The text requires m > d/2 + n + 1 to get E[u] ∈ C^{n+1}, with n = floor(m − d/2). That inequality is impossible. Sobolev embedding from H^m gives at best C^n. So the cited external theorem [41, Thm 3.3], which needs drift in C^{n+1}, is not applicable at the stated regularity. The standard fix is to lower the solution index by one (N = n − 1), which weakens the theorem's conclusion. As it stands, Theorem 1 is not established at the claimed regularity.\n\nSecond, Eq. (3.9) in the 2D vorticity section is wrong: the time integral is missing. The correct identity is an ODE, d/dt ||E[ω]||² = −Σ_k ||£_{ξ(k)}E[ω]||². And the 'decay to zero' claim ignores conservation of the spatial mean of E[ω]; the L² norm decays to the mean, not to zero. This is a concrete error, but it's in an illustrative section and is easily fixed.\n\nThird, Assumption 1 (uniform ellipticity of the noise) is doing a lot of work. The paper's 'regularization mechanism' and the 3D global existence threshold both depend on it. Calling it 'minor' in the introduction undersells it; in data-driven SALT applications, such uniform ellipticity is a strong assumption. Worth flagging for the reader.\n\nThe algebraic core—the closure of the mean and the variance dynamics—looks solid, and the authors are appropriately transparent about the prior appearance of the LLNS equation. The paper deserves a serious referee, but the referees should ask for a corrected Theorem 1 proof and a fix of Eq. (3.9). I'd encourage engagement, not desk rejection.","headline":"LA SALT is a genuinely useful extension of the SALT mean-field construction, but the main well-posedness theorem has a regularity-index gap and the 2D vorticity identity in (3.9) is wrong as written.","tokens_in":37703,"tokens_out":7375,"would_cite":true,"duration_ms":59564,"reading_group":"maybe","serious_thinker":"yes","would_accept_peer_review":true},"rs_alignment":null,"lean_confirmation":null,"pith_extraction":{"msc":["35Q35","35R60","60H15","76D05","76B03"],"pacs":[],"model":"deepseek-v4-flash","headline":"The paper claims that replacing the drift velocity of a stochastic fluid by its expectation yields a closed, Navier–Stokes-type equation for the mean field, with a Lie–Laplacian dissipation that regularizes the expected solution.","keywords":["stochastic fluid dynamics","Lagrangian averaging","Lie transport","mean-field nonlinearity","Lie-Laplacian dissipation","well-posedness","stochastic Euler equations","Navier-Stokes equations"],"falsifier":"For a fixed smooth divergence-free initial datum, solve the LLNS equation (3.15) for a sequence of noise fields whose ellipticity constants tend to infinity; if a finite-time singularity appears for arbitrarily large $\\kappa$, the claimed global well-posedness of Theorem 2 for large noise is false. In 2D, measure the three terms in the vorticity variance identity (3.10): total enstrophy, mean enstrophy, and $\\sum_k\\int|\\mathcal{L}_{\\xi^{(k)}}E[\\omega]|^2\\,dA$. The paper predicts total enstrophy is conserved and the variance grows exactly at the rate given by that sum; any consistent departure from this identity would falsify the fluctuation-variance mechanism.","tokens_in":36650,"feed_emoji":"🌊","tokens_out":12150,"duration_ms":106709,"temperature":0.7,"pith_summary":"The paper sets out a class of stochastic fluid equations, LA SALT, in which the velocity is transported both by white-noise vector fields and by its own ensemble-average velocity. Replacing the drift by the expectation makes the equations nonlinear in the sense of McKean (the drift depends on the law of the solution), and the central discovery is that the expectation field then satisfies a closed Navier–Stokes-type equation whose dissipation is a Lie–Laplacian built from the noise fields. The stochastic model therefore keeps the geometric conservation laws of ideal fluids, such as Kelvin's circulation theorem and helicity, while the mean motion is regularized. The paper proves local well-posedness of the LA SALT Euler equations in two and three dimensions, global well-posedness in two dimensions, and global well-posedness in three dimensions when the noise fields are large enough to make the Lie–Laplacian uniformly elliptic. A sympathetic reader would care because this offers a mechanism, grounded in stochastic transport, by which unresolved fluctuations dissipate the mean while the ideal-fluid structure is preserved.","feed_headline":"Averaging noise gives Euler equations a Navier-Stokes smoothing","feed_subtitle":"When the transport velocity is replaced by its own ensemble average, the mean obeys a closed dissipative PDE.","key_machinery":"The central object is the LA SALT transport vector field $dX_t = E[u_t]\\,dt + \\sum_k \\xi^{(k)}(x)\\circ dW_t^{(k)}$, whose drift is the expectation of the SALT velocity instead of the velocity itself; this expectation is what makes the system nonlinear in the sense of McKean. The key identity that carries the argument is that the Itô correction for the mean produces the Lie–Laplacian operator $\\frac{1}{2}\\sum_k \\mathcal{L}_{\\xi^{(k)}}(\\mathcal{L}_{\\xi^{(k)}}\\cdot)$, which under Assumption 1, $\\kappa|y|^2 \\le \\frac12\\sum_k y_i\\xi^{(k)}_i\\xi^{(k)}_j y_j$ for all $x,y$, is a uniformly elliptic second-order operator in divergence form plus lower-order terms. This operator supplies the dissipation in the closed equation for $E[u]$, making Navier–Stokes-type energy estimates available; the bound $\\frac12\\sum_k ((\\mathcal{L}^T_{\\xi^{(k)}})^2 u, u)_{H^m} \\le -\\kappa'|\\nabla u|^2_{H^m} + C|u|^2_{H^m}$ is the workhorse of the well-posedness theorems. The fluctuation dynamics are then linear stochastic transport equations slaved to the mean, which is what lets the paper close the variance equations.","core_discovery":"The paper's central claim is that averaging the SALT transport velocity over noise realizations — using $dX_t = E[u_t]\\,dt + \\sum_k \\xi^{(k)}\\circ dW_t^{(k)}$ in place of $dx_t = u_t\\,dt + \\sum_k \\xi^{(k)}\\circ dW_t^{(k)}$ — converts the stochastic Euler equations into a system whose expectation $v = E[u]$ obeys the Lie-Laplacian Navier-Stokes equation $\\partial_t v + P\\mathcal{L}^T_v v = P\\frac12\\sum_k \\mathcal{L}^T_{\\xi^{(k)}}(\\mathcal{L}^T_{\\xi^{(k)}}v) + Pf$. The double Lie derivative acts as a dissipation operator; under Assumption 1 it is uniformly elliptic, so the mean equation behaves analytically like Navier-Stokes with viscosity replaced by noise geometry. Theorem 1 states that LA SALT Euler is locally well-posed in Sobolev spaces for $d=2,3$, globally well-posed for $d=2$, and globally well-posed for $d=3$ whenever the ellipticity constant $\\kappa$ exceeds a data-dependent threshold $\\kappa_*$. The paper also shows that the total enstrophy of the two-dimensional vorticity is a conserved Casimir while the enstrophy of the mean decays and the fluctuation variance grows at an explicit rate, and it extends the construction to rigid-body dynamics, Burgers, Camassa-Holm, and stratified magnetohydrodynamics.","pith_inferences":["Editorial: the 2D variance identity (3.10) is a directly testable quantitative prediction, so the LA SALT framework could serve as a numerical laboratory for fluctuation-dissipation balance in turbulent transport without fitting parameters.","Editorial: the large-$\\kappa$ global-existence threshold in 3D suggests a stochastic counterpart of the Navier–Stokes regularity problem in which the control parameter is the geometric strength of transport noise rather than molecular viscosity; the paper does not pursue this analogy.","Editorial: if the construction extends to the stratified MHD example as formulated, unresolved small-scale transport noise could act as an effective dissipation mechanism in geophysical or plasma models, replacing ad hoc viscosity terms; the paper leaves the physical testing of this open."],"forward_implications":["The expected velocity of a LA SALT fluid satisfies a closed Navier–Stokes-type PDE, so non-locality in probability space acts as a regularization mechanism without molecular viscosity.","LA SALT Euler is locally well-posed in Sobolev spaces for $d=2,3$, globally well-posed for $d=2$, and globally well-posed for $d=3$ whenever the noise ellipticity constant $\\kappa$ exceeds a data-dependent threshold.","The ideal-fluid conservation laws persist: Kelvin's circulation theorem holds along the stochastic flow, helicity is preserved in 3D, and total enstrophy in 2D is a conserved Casimir even while the mean enstrophy decays and the fluctuation variance grows.","The same construction applied to Burgers gives a viscous Burgers equation for the mean; applied to Camassa–Holm it yields a finite system of expectation-dependent SDEs for peakon parameters; applied to MHD it produces a stratified 3D system with Lie–Laplacian dissipation.","With constant noise fields equal to coordinate basis vectors, the Lie–Laplacian reduces to the ordinary Laplacian and the mean equation becomes the classical Navier–Stokes equation, recovering the stochastic Lagrangian representation of Navier–Stokes solutions."],"supporting_citations":[{"why":"Introduces the SALT equations (stochastic advection by Lie transport) that LA SALT averages.","marker":"[1]"},{"why":"Provides the stochastic Lagrange-to-Euler representation whose constant-noise case LA SALT recovers.","marker":"[2]"},{"why":"Is the earlier mean-field stochastic Euler construction and Lie-Laplacian Navier-Stokes relation that this paper extends.","marker":"[3]"},{"why":"Defines the mean-field nonlinearity (drift depending on the law) used in LA SALT.","marker":"[6]"},{"why":"Supplies the Euler-Poincaré and semidirect-product framework for fluids with advected quantities.","marker":"[7]"},{"why":"Provides the classical Navier-Stokes theory whose energy estimates the LLNS proof follows.","marker":"[39]"},{"why":"Provides elliptic and Sobolev estimates used for the pressure and the well-posedness arguments.","marker":"[40]"},{"why":"Contains the linear stochastic transport well-posedness theorem (Theorem 3.3) used to solve LA SALT for a given mean field.","marker":"[41]"}],"fun_headline_variants":["Averaging noise turns stochastic Euler into Navier-Stokes-like flow","Noise-averaged transport velocity yields dissipative fluid equations","Stochastic Euler averaged: closed dissipative mean dynamics","Lie-Laplacian from noise average regularizes fluid equations","When transport averages itself: Navier-Stokes emerges"],"cache_read_input_tokens":3200,"weakest_assumption_plain":"Everything rests on Assumption 1 (equation (3.16)): the fixed noise fields $\\xi^{(k)}$ must generate uniform ellipticity, meaning at every point and in every direction the quadratic form built from $\\xi^{(k)}$ is bounded below by $\\kappa|y|^2$; without it the Lie–Laplacian is not elliptic, the regularization mechanism fails, and the three-dimensional global-existence threshold $\\kappa > \\kappa_*$ is unavailable.","fun_headline_variants_meta":{"raw":{"variants":["Averaging noise turns stochastic Euler into Navier-Stokes-like flow","Noise-averaged transport velocity yields dissipative fluid equations","Stochastic Euler averaged: closed dissipative mean dynamics","Lie-Laplacian from noise average regularizes fluid equations","When transport averages itself: Navier-Stokes emerges"]},"model":"deepseek-v4-flash","effort":"low","cost_usd":0.000689,"raw_usage":{"total_tokens":3241,"prompt_tokens":1186,"completion_tokens":2055,"prompt_tokens_details":{"cached_tokens":384},"prompt_cache_hit_tokens":384,"prompt_cache_miss_tokens":802,"completion_tokens_details":{"reasoning_tokens":1972}},"tokens_in":802,"tokens_out":2055,"duration_ms":14220,"temperature":1.0,"reasoning_tokens":1972,"cache_read_input_tokens":384,"cache_creation_input_tokens":0},"cache_creation_input_tokens":0},"created_at":"2026-08-14T10:15:02.327721+00:00","model_set":{"reader":"deepseek-v4-flash"},"falsifier":"For a fixed smooth divergence-free initial datum, solve the LLNS equation (3.15) for a sequence of noise fields whose ellipticity constants tend to infinity; if a finite-time singularity appears for arbitrarily large $\\kappa$, the claimed global well-posedness of Theorem 2 for large noise is false. In 2D, measure the three terms in the vorticity variance identity (3.10): total enstrophy, mean enstrophy, and $\\sum_k\\int|\\mathcal{L}_{\\xi^{(k)}}E[\\omega]|^2\\,dA$. The paper predicts total enstrophy is conserved and the variance grows exactly at the rate given by that sum; any consistent departure from this identity would falsify the fluctuation-variance mechanism.","supporting_citations":[{"cited_title":null,"cited_arxiv_id":null,"evidence_quote":"Introduces the SALT equations (stochastic advection by Lie transport) that LA SALT averages."},{"cited_title":"A stochastic Lagrangi an representation of the three-dimensional incompressible Navier-Stokes equations","cited_arxiv_id":null,"evidence_quote":"Provides the stochastic Lagrange-to-Euler representation whose constant-noise case LA SALT recovers."},{"cited_title":"Circulation and ener gy theorem preserving stochastic fluids","cited_arxiv_id":null,"evidence_quote":"Is the earlier mean-field stochastic Euler construction and Lie-Laplacian Navier-Stokes relation that this paper extends."},{"cited_title":null,"cited_arxiv_id":null,"evidence_quote":"Defines the mean-field nonlinearity (drift depending on the law) used in LA SALT."},{"cited_title":"Holm, Jerrold E","cited_arxiv_id":null,"evidence_quote":"Supplies the Euler-Poincaré and semidirect-product framework for fluids with advected quantities."},{"cited_title":"The three-dimensional Navier–Stokes equa- tions: Classical theory , volume 157","cited_arxiv_id":null,"evidence_quote":"Provides the classical Navier-Stokes theory whose energy estimates the LLNS proof follows."},{"cited_title":"Mathematical Tools for the Study of the Incompressible Navi er-Stokes Equations and Related Models","cited_arxiv_id":null,"evidence_quote":"Provides elliptic and Sobolev estimates used for the pressure and the well-posedness arguments."},{"cited_title":"On d egenerate linear stochastic evolution equations driven by jump processes","cited_arxiv_id":null,"evidence_quote":"Contains the linear stochastic transport well-posedness theorem (Theorem 3.3) used to solve LA SALT for a given mean field."}],"review_version":1}