{"id":"37f11c12-19c7-4b36-bbad-3813ddabaf2b","arxiv_id":"2501.00705","paper_version":2,"verdict":"CONDITIONAL","confidence":"MODERATE","novelty_score":6.0,"correctness_risk":"medium","formal_verification":"none","parameter_count":2,"one_line_summary":"A linear Stokes flow with boundary-concentrated forcing exhibits non-vanishing viscous dissipation in the inviscid limit, showing the global criterion for anomalous dissipation does not detect convective cascade.","lead":"This paper constructs a linear Stokes flow inside a sphere whose average energy dissipation stays positive even as viscosity tends to zero, using forcing that concentrates at the boundary. It argues that the standard global criterion for anomalous dissipation cannot distinguish boundary-driven energy loss from the turbulent cascade.","discovery_kind":"extension","skeptic_critique":{"model":"deepseek-v4-flash","headline":"The proof of Theorem 3.6 imports a lower Gaussian kernel bound from a scalar Robin-operator result ([37]) and applies it to the matrix-valued Stokes semigroup with Navier-slip; this transfer is not justified and is likely false componentwise, so the boundary blow-up underpinning Theorem 3.9 is…","rationale":"The reader's weakest_assumption correctly identifies the transfer of the lower Gaussian kernel bound from scalar Robin operators to the vector Stokes semigroup as the most fragile point. My reading of the paper confirms this: Proposition 3.3 is a self-contained calculation for the Gaussian kernel H_t, and the only bridge to the real solution is the asserted inequality K_t \\geq H_t in Theorem 3.6. This bridge is load-bearing because the final anomalous-dissipation conclusion in Theorem 3.9 uses exactly the boundary trace lower bound produced by that comparison, combined with the interpolation inequality. I checked the change-of-variables in Proposition 3.3 and found no missing Jacobian factor: the kernel prefactor cancels with the change of variables before the Itô isometry, so the Gaussian lower-bound calculation is internally consistent. The weakness is confined to the kernel comparison. The cited reference [37] treats Robin scalar operators; the Stokes operator with Navier-slip has a matrix kernel, a Leray projection, and mixed boundary conditions involving tangential derivatives, so the pointwise lower bound is not justified and is unlikely to hold componentwise. A positive result could be recovered if the authors prove a suitable lower bound for the relevant component of the Stokes kernel, or find another mechanism for the boundary blow-up. Since this is a serious but potentially fixable gap, the existing CONDITIONAL verdict remains appropriate; no verdict change is needed, but the condition should explicitly require a proof or a valid citation for the kernel lower bound. Secondary issues, such as the asserted deterministic extension in Remark 3.11, are also unsupported, but they are less central because the main theorem is stochastic and the deterministic claim is presented as an extension rather than the core proof.","tokens_in":17251,"tokens_out":10543,"duration_ms":107352,"concrete_test":"Use the explicit vector-spherical-harmonic expansion of the Stokes operator with Navier-slip on B(0,R) to compute, or rigorously bound, the kernel component that acts on the azimuthal basis vector e_phi near the boundary. For x in the boundary and y near the boundary, check whether the relevant component of K_t(x,y) satisfies the lower bound c t^{-3/2} e^{-b|x-y|^2/t} e^{-wt}; if the component is negative on any open set, or decays faster than the Gaussian lower bound, the comparison in Theorem 3.6 fails and Theorem 3.9 lacks proof. A simpler preliminary test is to check positivity preservation: evolve a nonnegative divergence-free vector field supported near the boundary under the Stokes semigroup with Navier-slip; if any component becomes negative, the scalar positivity-based reasoning in [37] cannot transfer to this matrix kernel.","verdict_should_be":"UNCHANGED","load_bearing_attack":"The entire argument reduces to comparing the true Stokes semigroup kernel K_t(x,y) with the scalar heat kernel H_t(x,y)=c t^{-3/2}e^{-b|x-y|^2/t}e^{-wt} and asserting K_t \\geq H_t in Theorem 3.6. This comparison is what converts Proposition 3.3, which is a statement about the Gaussian kernel only, into a positive lower bound for the boundary trace of the actual stochastic Stokes solution. Without it, Theorem 3.6 does not follow, and neither does the interpolation argument in Theorem 3.9. The cited reference [37] proves lower Gaussian bounds for scalar Robin Laplacians, not for the Stokes operator with Navier-slip. The Stokes semigroup involves the Leray projection and a nonlocal pressure coupling; it is not positivity-preserving, and even in the whole-space case the Stokes kernel (Oseen tensor) has matrix components that change sign. A pointwise scalar lower bound such as K_t(x,y) \\geq H_t(x,y) is therefore not a consequence of [37] and is very likely false for all components. The paper provides no derivation, no spectral computation for the ball, and no a priori estimate that would justify the comparison. This is the single most load-bearing gap: every later step presupposes the boundary blow-up that only this unproved kernel transfer can supply.","agreement_with_reader":"agree"},"referee_report":{"model":"deepseek-v4-flash","summary":"The paper studies the linear stochastic Stokes equations with Navier-slip boundary conditions on the ball B(0,R) ⊂ R³ (system (2)), with zero initial data and zero deterministic forcing, and with a stochastic forcing g dW_t whose amplitude grows like dist(x, ∂D)^{-δ/2}. The main result, Theorem 3.9, states that for δ < (11−√41)/10 ≈ 0.45969 there is a divergence-free g such that the unique weak solution zν satisfies lim sup_{ν→0} ν E∫_0^T ∥∇zν∥²_{L²(D)} dt > 0. The proof combines a Gaussian kernel lower bound at the boundary (Theorem 3.6), an H^s regularity estimate (Proposition 3.8), and interpolation with H¹ (Theorem 3.9). Section 4 reports finite-difference simulations for a half-space and for a sphere, and the paper argues in Remarks 3.10–3.14 that this shows definition (1) is not a faithful measure of convective dissipation.","tokens_in":17495,"tokens_out":19413,"duration_ms":183043,"significance":"If the result were established, it would be a valuable contribution to the ongoing discussion of the correct definition of anomalous dissipation: it would give a non-stationary, linear example in a bounded domain, with forcing that is smooth away from the boundary and only in L², where boundary concentration alone produces a positive dissipation limit. This complements the stationary examples of Bedrossian et al. and sharpens the point that the global dissipation limsup does not isolate convective cascade. The paper is transparent about its limitations (Remark 2.5, Remark 3.14) and includes numerical evidence. The main obstacle is that several load-bearing technical steps are not justified in the manuscript; in particular, the concern raised about the transfer of scalar Robin kernel bounds to the Stokes semigroup is confirmed and is central.","major_comments":[{"comment":"Theorem 3.6 asserts that the kernel K_t of the Stokes semigroup e^{-tA_{NS}} on the ball satisfies the scalar lower Gaussian bound K_t ≥ H_t, citing [37]. Reference [37] treats scalar Robin operators; the paper provides no argument transferring this to the vector-valued Stokes operator with Leray projection and pressure coupling. Self-adjointness alone does not imply pointwise lower bounds, and the Stokes kernel is not componentwise nonnegative in general. The proof also states 'g ≥ 0 and K_t, H_t ≥ 0', but the forcing g in Proposition 3.3 is a vector field with sign-changing components, and no componentwise nonnegativity of K_t is established. Since this comparison is exactly what converts Proposition 3.3 into the boundary blow-up used in Theorem 3.9, this is a load-bearing gap that must be closed.","section":"§3, Theorem 3.6"},{"comment":"The vector field g defined in Proposition 3.3 is not divergence-free as written. The claimed calculation ∇·g = (1/(r sinθ)) ∂_φ((R-r)^{-δ/2}) = 0 computes the divergence of f(r)e_φ, not of the stated bφ(x) = (−√(x_1²+x_2²)e_1 + x_3e_2)/|x|; the latter has nonzero Cartesian divergence (even its constant-vector part, up to the radial factor, has ∂_1(−ρ/r) ≠ 0). Consequently g ∉ [L²_σ(D)]³ as stated, the semigroup e^{-ν(t−s)A_{NS}}g is not defined for this forcing unless the Leray projection is applied, and the subsequent boundary lower bounds would need to track the projection. Please correct the definition of the tangential vector field and re-verify the divergence-free claim.","section":"§3, Proposition 3.3"},{"comment":"The proof of the uniform bound sup_ν E∫_0^T ∥zν∥²_{H^γ(D)} uses only the L² semigroup estimate ∥e^{-νtA_{NS}}∥ ≤ M e^{νω_0 t} and then replaces ∥e^{-ν(t-s)A_{NS}}g∥_{H^γ} by ∥g∥_{H^γ}. No bound for the Stokes semigroup on H^γ(D), uniform in ν, is established or cited. The H^γ norm of the semigroup acting on g requires an estimate on the heat kernel or on the spectral multiplier, which is not supplied. This uniform bound is an input to interpolation inequality (7), so the step is currently unjustified.","section":"§3, Proposition 3.8"},{"comment":"The stochastic convolution is an L²(D)-valued process, but Proposition 3.3 and Theorem 3.6 apply Itô's isometry pointwise in x ∈ ∂D. Pointwise boundary values of a cylindrical-Wiener stochastic convolution are not a direct consequence of Itô's isometry; they require a spatially regular version (e.g., via the trace theorem for fixed ν > 0) and a justification that the two-point second moment formula matches the pointwise evaluation. As written, the computation treats the noise as though it were a scalar Wiener process for each x, which is not the definition used in (2)-(3).","section":"§3, Proposition 3.3 and Theorem 3.6"}],"minor_comments":[{"comment":"The simulations fix δ=0.75 and δ values up to 0.9, which lie outside the range δ < (11−√41)/10 of Theorem 3.9; the text should state more cautiously that the numerics are suggestive rather than confirmatory.","section":"§4"},{"comment":"In the computation of the radial integral, the gamma factor should be Γ((3−δ)/2), not Γ(3−δ); the displayed constant is therefore incorrect, although the positivity conclusion is unaffected.","section":"§3, Proposition 3.3"},{"comment":"The boundary condition n·∇fτ + αfτ = 0 is stated for f ∈ [H¹(D)]^d, but α ∈ L∞(∂D) and the trace of ∇fτ requires more regularity; the functional setting should be clarified.","section":"§2.1"},{"comment":"There are numerous notational and grammatical issues (e.g., 'existance', 'Navier Stokes' without hyphen, 'inquality'); a careful proofread is needed.","section":"Throughout"}],"recommendation":"major_revision","confidential_remarks":"The paper addresses a question of current interest and is likely within the scope of the journal. The submitted proof has several gaps, the most serious being the unjustified transfer of a scalar Robin kernel lower bound to the Stokes semigroup. I do not see evidence of circularity or misuse of fitted parameters; the forcing is explicitly constructed. The numerical section is preliminary but supportive. I recommend inviting a major revision, with the requirement that the revised version prove or precisely cite a lower Gaussian bound for the Navier-slip Stokes semigroup, correct the divergence-free forcing, and rigorously justify the H^γ semigroup bound and the pointwise stochastic-integral identities."},"author_rebuttal":null,"desk_editor":{"model":"deepseek-v4-flash","letter":"Colleague—\n\nThe paper has a genuinely interesting idea, but the main theorem currently rests on a kernel bound that is borrowed from the wrong setting and almost certainly false as stated. The stress-test note is on target.\n\nWhat's new: this is, as far as I know, the first non-stationary example of a flow satisfying the global anomalous-dissipation criterion without any nonlinearity. The authors take the linear Stokes problem on a ball with Navier-slip, drive it with an explicit L2 forcing g ~ dist^{-δ/2} that is smooth in the interior but singular at the boundary, and show limsup ν E∫|∇u|^2 > 0 for δ < 0.45969. The forcing is fixed and ν-independent; the mechanism is boundary-generated vorticity rather than vortex stretching. That is a legitimate conceptual correction: the global definition (1) does not isolate convective dissipation. The interpolation argument that converts an H^s bound plus a boundary-trace lower bound into a dissipation lower bound is clever and mostly clean.\n\nThe soft spots, in order. First and load-bearing: Theorem 3.6 asserts that the Stokes semigroup kernel K_t satisfies the lower Gaussian bound K_t ≥ H_t, citing [37], a paper on scalar Robin operators. The Stokes semigroup is a system with a nonlocal Leray projection; its kernel components change sign even in the whole space. A componentwise pointwise lower bound by a positive Gaussian is not a consequence of [37] and is very likely false. All the later steps—the boundary blow-up, the interpolation in Theorem 3.9, the conclusion—depend on that comparison. Without a proof of a suitable lower bound for the actual Stokes kernel, Theorem 3.6 is unsupported. This is not a cosmetic gap.\n\nSecond, Proposition 3.8 is sketched far too quickly: the uniform H^γ bound is derived from a semigroup estimate plus the fact that g∈H^γ, but the argument as written doesn't handle the boundary condition carefully. I think it's probably fixable, but it needs a real proof.\n\nThird, the deterministic extension in Remark 3.11 is asserted, not proven. The numerical simulations use δ=0.75, outside the theorem's range, so they illustrate the phenomenon but don't confirm the rigorous result.\n\nBottom line: the idea is good and probably right in spirit, and the paper deserves a serious referee. It should be sent out—the conceptual point matters to the turbulence and stochastic-PDE communities—but the referee report will have to demand a genuine lower bound for the Stokes kernel or a different route. I wouldn't cite the result in its current form.","headline":"The paper's conceptual point is real, but the proof currently rests on an unjustified and likely false lower Gaussian bound for the Stokes kernel.","tokens_in":18034,"tokens_out":4326,"would_cite":false,"duration_ms":40861,"reading_group":"maybe","serious_thinker":"yes","would_accept_peer_review":true},"rs_alignment":null,"lean_confirmation":null,"pith_extraction":{"msc":["35Q30","76D05","60H15","35B25"],"pacs":[],"model":"deepseek-v4-flash","headline":"This paper constructs a linear Stokes flow in a ball with Navier-slip walls whose dissipation rate stays positive in the zero-viscosity limit, showing the standard global test of anomalous dissipation can be satisfied without any…","keywords":["anomalous dissipation","Navier-slip boundary condition","Stokes equations","inviscid limit","boundary layer","stochastic PDE","heat kernel bounds","energy dissipation"],"falsifier":"Compute the boundary kinetic energy $\\nu^{\\delta/2}\\,\\mathbb{E}\\int_0^T \\|z_\\nu\\|^2_{L^2(\\partial D)}$ for $\\delta$ near $0.45$ on a ball with Navier-slip walls; if it tends to zero as $\\nu\\to 0$, the imported kernel lower bound fails at the boundary. Equivalently, check directly whether the Navier-slip Stokes kernel satisfies $K_t(x,y) \\ge c t^{-3/2} e^{-b|x-y|^2/t} e^{-wt}$ for boundary points $x$; a counterexample at any boundary point would break the comparison in Theorem 3.6.","tokens_in":17004,"feed_emoji":"🌊","tokens_out":10531,"duration_ms":91195,"temperature":0.7,"pith_summary":"The paper shows that the usual global test for anomalous dissipation—positivity of the limit of viscosity times integrated squared velocity gradient as viscosity goes to zero—can be met by a purely linear Stokes flow with no convection term. The construction works on the ball with Navier-slip boundary data, using an external force that grows like a power of distance to the wall and concentrates there. For the admissible singularity exponent $\\delta < (11-\\sqrt{41})/10 \\approx 0.45969$, the unique weak solution has a positive $\\limsup$ of the global dissipation in the inviscid limit. The authors conclude that this definition is not a faithful measure of convective dissipation, even for solutions that are not statistically stationary and even in the deterministic case where the stochastic forcing is replaced by a deterministic singular force.","feed_headline":"Linear Stokes flow shows anomalous dissipation without turbulence","feed_subtitle":"A sphere with slip walls keeps dissipating energy as viscosity vanishes, so the usual test cannot separate linear from turbulent mixing.","key_machinery":"The load-bearing object is the kernel $K_t$ of the Stokes semigroup with Navier-slip boundary conditions, together with the Gaussian lower bound $K_t(x,y) \\ge c t^{-3/2} e^{-b|x-y|^2/t} e^{-wt}$, imported from scalar Robin operator theory (reference [37]). This lower bound lets the authors compare the singular forcing against an explicit heat kernel and obtain a positive boundary-energy lower bound (Theorem 3.6). The second ingredient is an interpolation step: because the solution is uniformly bounded in $H^{s_\\delta}$ and the boundary blow-up sits on a measure-zero set, interpolation with $H^1$ turns the wall concentration into a positive bulk dissipation $\\limsup$, with the trace operator providing the link.","core_discovery":"The central result is Theorem 3.9: take zero initial data and zero bulk force, take $D = B(0,R)$, and take $g(x) = (R-|x|)^{-\\delta/2}$ times a divergence-free azimuthal vector field. For every $\\delta < (11-\\sqrt{41})/10 \\approx 0.45969$, the solution $z_\\nu$ of the linear Stokes problem satisfies $\\limsup_{\\nu\\to 0} \\nu \\, \\mathbb{E}\\int_0^T \\|\\nabla z_\\nu\\|^2_{L^2(D)} > 0$. The proof first shows that the boundary kinetic energy does not vanish: $\\nu^{\\delta/2}\\,\\mathbb{E}\\int_0^T \\|z_\\nu\\|^2_{L^2(\\partial D)}$ has positive $\\liminf$ (Theorem 3.6). Then a trace theorem and real interpolation between $H^{s_\\delta}$ and $H^1$ convert the wall blow-up into a positive $\\limsup$ for the bulk gradient norm (Theorem 3.9).","pith_inferences":["One can test whether the restriction $\\delta < 0.45969$ is purely technical: the simulations suggest the positive dissipation persists for all $\\delta\\in(0,1)$, so a proof for the full range would likely only require a sharper interpolation or kernel estimate.","If the same lower Gaussian kernel bound holds on more general $C^{1,1}$ or Lipschitz domains, then the construction should transfer directly, which would make the definitional failure independent of spherical symmetry.","The deterministic analogue with singular $f$ and $g=0$ should show the same boundary blow-up but at a possibly faster rate; the numerical comparison suggests the sign changes of the stochastic forcing slow the boundary energy growth, a mechanism that could be isolated by computing the boundary energy as a function of the noise amplitude."],"forward_implications":["If the construction is correct, the global criterion $\\limsup_{\\nu\\to0} \\nu \\mathbb{E}\\int \\|\\nabla u_\\nu\\|^2 > 0$ cannot by itself distinguish turbulent Navier-Stokes dynamics from a linear Stokes problem with boundary forcing.","A fixed $L^2$ forcing that is smooth in the interior and singular only on the boundary suffices to trigger the effect, because the boundary has Lebesgue measure zero.","The same analysis works for the linear heat equation, so the phenomenon does not depend on incompressibility or on the vector structure of the Stokes operator.","The paper proposes that anomalous dissipation be measured locally (for example by the Duchon-Robert dissipation measure), since the global and weak definitions are also satisfied by these linear examples."],"supporting_citations":[{"why":"Supplies the kernel representation and the lower Gaussian heat-kernel bound used to lower-bound the boundary energy in Theorem 3.6.","marker":"[37]"},{"why":"Shows statistically stationary stochastic heat solutions already satisfy the global dissipation inequality, the contrast the paper's non-stationary example extends.","marker":"[5]"},{"why":"Proves that under no-slip conditions bulk regularity prevents anomalous dissipation, the opposite boundary-condition case the paper compares against.","marker":"[12]"},{"why":"Provides the Sobolev trace and interpolation theory used to convert boundary energy blow-up into bulk dissipation in Theorem 3.9.","marker":"[1]"},{"why":"Provides the convolution inequality used to prove $g \\in H^s$, which fixes the allowed range of $\\delta$ in Remark 3.5.","marker":"[19]"},{"why":"Gives the Faedo-Galerkin existence framework for stochastic Navier-Stokes that motivates the Stokes semigroup construction in Section 2.2.","marker":"[16]"},{"why":"Establishes the spectral and boundary-condition properties of the Stokes operator with Navier-slip used to set up the semigroup.","marker":"[8]"}],"fun_headline_variants":["No turbulence needed: linear flow dissipates in a sphere","Anomalous dissipation from linear Stokes flow in a sphere","Sphere with slip walls shows energy loss as viscosity vanishes","Linear flow in a ball: inviscid limit still dissipates energy"],"cache_read_input_tokens":3200,"weakest_assumption_plain":"The proof assumes the Gaussian lower bound on the solution kernel, proven for a simpler scalar Robin problem, also holds for the vector-valued Stokes semigroup with Navier-slip boundary data, uniformly up to the boundary.","fun_headline_variants_meta":{"raw":{"variants":["No turbulence needed: linear flow dissipates in a sphere","Anomalous dissipation from linear Stokes flow in a sphere","Sphere with slip walls shows energy loss as viscosity vanishes","Linear flow in a ball: inviscid limit still dissipates energy"]},"model":"deepseek-v4-flash","effort":"low","cost_usd":0.000385,"raw_usage":{"total_tokens":1980,"prompt_tokens":836,"completion_tokens":1144,"prompt_tokens_details":{"cached_tokens":384},"prompt_cache_hit_tokens":384,"prompt_cache_miss_tokens":452,"completion_tokens_details":{"reasoning_tokens":1075}},"tokens_in":452,"tokens_out":1144,"duration_ms":8930,"temperature":1.0,"reasoning_tokens":1075,"cache_read_input_tokens":384,"cache_creation_input_tokens":0},"cache_creation_input_tokens":0},"created_at":"2026-08-10T22:45:10.165862+00:00","model_set":{"reader":"deepseek-v4-flash"},"falsifier":"Compute the boundary kinetic energy $\\nu^{\\delta/2}\\,\\mathbb{E}\\int_0^T \\|z_\\nu\\|^2_{L^2(\\partial D)}$ for $\\delta$ near $0.45$ on a ball with Navier-slip walls; if it tends to zero as $\\nu\\to 0$, the imported kernel lower bound fails at the boundary. Equivalently, check directly whether the Navier-slip Stokes kernel satisfies $K_t(x,y) \\ge c t^{-3/2} e^{-b|x-y|^2/t} e^{-wt}$ for boundary points $x$; a counterexample at any boundary point would break the comparison in Theorem 3.6.","supporting_citations":[{"cited_title":"3, 1195–1225","cited_arxiv_id":null,"evidence_quote":"Supplies the kernel representation and the lower Gaussian heat-kernel bound used to lower-bound the boundary energy in Theorem 3.6."},{"cited_title":"3, 1045–1075","cited_arxiv_id":null,"evidence_quote":"Shows statistically stationary stochastic heat solutions already satisfy the global dissipation inequality, the contrast the paper's non-stationary example extends."},{"cited_title":"5, 4785–4811","cited_arxiv_id":null,"evidence_quote":"Proves that under no-slip conditions bulk regularity prevents anomalous dissipation, the opposite boundary-condition case the paper compares against."},{"cited_title":null,"cited_arxiv_id":null,"evidence_quote":"Provides the Sobolev trace and interpolation theory used to convert boundary energy blow-up into bulk dissipation in Theorem 3.9."},{"cited_title":"2, 505–510","cited_arxiv_id":null,"evidence_quote":"Provides the convolution inequality used to prove $g \\in H^s$, which fixes the allowed range of $\\delta$ in Remark 3.5."},{"cited_title":"3, 367–391","cited_arxiv_id":null,"evidence_quote":"Gives the Faedo-Galerkin existence framework for stochastic Navier-Stokes that motivates the Stokes semigroup construction in Section 2.2."},{"cited_title":null,"cited_arxiv_id":null,"evidence_quote":"Establishes the spectral and boundary-condition properties of the Stokes operator with Navier-slip used to set up the semigroup."}],"review_version":1}