{"id":"d5b7703e-984b-4a02-a31d-cc503ee050e9","arxiv_id":"2509.10211","paper_version":1,"verdict":"ACCEPT","confidence":"MODERATE","novelty_score":8.0,"correctness_risk":"low","formal_verification":"none","parameter_count":0,"one_line_summary":"Random turbulent velocity fields cause generic scalar fields to dissipate energy and instantly regularize to sharp fractional regularity, with particle dispersion obeying t^{1/(1-alpha)}.","lead":"Mathematicians proved that rough random velocity fields that model turbulence make scalar fields lose energy at a positive rate while instantly becoming smoother. The results turn heuristic turbulence laws, including Richardson's particle-spreading law, into rigorous theorems with explicit exponents.","discovery_kind":"first_principles","skeptic_critique":{"model":"deepseek-v4-flash","headline":"No significant objection identified: the central regularity, sharpness, and dissipation claims are internally consistent; the main hypothesis (Assumption 4.1) is exactly verified for the Kraichnan noise in the diffusive regime.","rationale":"The reader's weakest_assumption identifies Assumption 4.1 as the load-bearing premise, and I agree that it is the most sensitive input to the proof of Theorem 1.3. However, for the Kraichnan model, which is the stated target of the main theorems, Assumption 4.1 is exactly equivalent to the diffusive-regime condition eta > 1 - d/(4 alpha^2), as shown in Corollary A.2. Thus the assumption is not an unverified or extraneous restriction for the central claim. I found no internal inconsistency in the proof chain: the generic anomalous dissipation result (Theorem 3.1) is supported by a standard dichotomy argument; the anomalous regularization proof reduces the PDE to a radial one via random rotations and then establishes uniform-in-kappa Holder estimates for the two-point correlation; the sharpness of the H^{1-alpha} threshold follows from the contrapositive of the critical regularity criterion in Proposition 3.10 combined with the Besov characterization; and the Duchon-Robert-type formula is derived from a carefully tracked limit of the dissipation measure, with the nontrivial Lemma 5.6 providing the distributional limit. The paper explicitly notes where the proofs fail on the torus and where the degenerate-parabolic theory is incomplete in general, which further supports the reliability of the claims for the isotropic full-space setting. The apparent S^{d-2} typo in Lemma 2.18 does not affect any of the main statements, since all applications use S^{d-1} averages and the equivalence is otherwise standard. Overall, the reader's ACCEPT verdict with moderate confidence is appropriate; my pass does not identify a concern that would warrant changing the verdict.","tokens_in":83800,"tokens_out":44323,"duration_ms":339900,"concrete_test":"As a verification pass, symbolically re-derive Lemma 4.7 from (4.18) and (4.6)-(4.8), and run a numerical spot-check of Proposition 4.2: solve the radial PDE (4.15) with kappa = 10^{-8} and coefficients satisfying Assumption 4.1 for beta at threshold and beta + 0.01, using adaptive mesh refinement near r = 0; verify the seminorm bound (4.10) with exponent 2 - 2 alpha - delta is uniform in kappa and degrades only as beta approaches the threshold.","verdict_should_be":"UNCHANGED","load_bearing_attack":"I do not find a load-bearing concern. The paper's main theorems depend on Assumption 4.1, but for the Kraichnan model this assumption is not an extra restriction: Corollary A.2 shows it holds precisely when eta > 1 - d/(4 alpha^2), which is exactly the stated diffusive regime. The degenerate-parabolic argument in Proposition 4.2 is the delicate part; I checked the change of variables (4.18), the monotonicity lemma (Lemma 4.7), the boundary-term cancellation in Lemma 4.8, and the interpolation/iteration in Sections 4.1-4.2, and the estimates do close with constants independent of kappa. The sharpness statements (1.10) follow from the dichotomy theorem and the Besov characterization; the limit identities in Theorem 5.1 are supported by the explicit Lemma 5.6. The only issue I noticed is the apparent typo 'S^{d-2}' in Lemma 2.18, where the unit sphere in R^d should presumably be S^{d-1}; the main theorems all use S^{d-1} averages, and the equivalence in Lemma 2.18 is otherwise standard, so this is cosmetic rather than load-bearing. The paper states its limitations candidly in Remark 1.7, Remark 1.4, and the discussion around Assumption 4.1.","agreement_with_reader":"partial"},"referee_report":{"model":"deepseek-v4-flash","summary":"The paper studies the stochastic transport equation dθ + ◦dW·∇θ = 0 and the stochastic continuity equation driven by a Gaussian velocity field that is white in time, space-homogeneous and isotropic, and α-Hölder in space, with the Kraichnan model as the main example. The main results are: (i) a dichotomy (Theorem 3.1) under which either all L² initial data conserve mean energy or every non-zero datum dissipates mean energy continuously in time; for Kraichnan noise this identifies the diffusive regime η > 1 − d/(4α²) as the dissipative one; (ii) anomalous Sobolev regularization (Theorem 1.3): every L² initial datum gives a solution lying in H^{1−α−δ}_x, with time-integrated and pointwise-in-time bounds, plus sharpness in the sense that the H^{1−α} norm integrated over any time interval is infinite for non-zero data; (iii) in the incompressible case, an explicit Duchon–Robert-type formula for the expected dissipation measure in terms of spherical averages of increments (Theorem 5.1); and (iv) Richardson's law (Theorem 1.5): for Dirac initial data, E[Var(μ_t)] is asymptotically K_Ric t^{1/(1−α)} in the diffusive regime. The technical core is a uniform-in-viscosity regularity estimate for the degenerate parabolic PDE satisfied by the two-point self-correlation function, after a reduction to radially symmetric data.","tokens_in":84050,"tokens_out":10351,"duration_ms":96121,"significance":"If the results hold, this is a substantial contribution to the mathematical theory of the Kraichnan model and to the rigorous understanding of anomalous dissipation and anomalous regularization in stochastic transport. The paper is proof-heavy and largely self-contained: the Wiener chaos representation, the duality with the continuity equation, the degenerate-parabolic estimates of Section 4, and the sharpness statements are all presented in detail. The proportionality constants in the dissipation formula and in Richardson's law are explicit rather than fitted, and the sharp regularity exponent is matched by a matching lower bound. The main structural assumption, Assumption 4.1, is shown in Corollary A.2 to be equivalent for Kraichnan noise to the diffusive regime η > 1 − d/(4α²), so it is not an extra restriction for the flagship application. The limitations of the argument, in particular the use of radial symmetry and the failure of the proof on the torus, are stated candidly in Remarks 1.4 and 1.7. In my reading, the central claims are internally consistent, and the stress-test concern about the load-bearing role of isotropy is real but already acknowledged in the paper.","major_comments":[],"minor_comments":[{"comment":"The integration in the spherical-average equivalence is written against σ(dẑ) on S^{d−2}; since all later statements and the definition of the noise live on R^d, the sphere should be S^{d−1}. The equivalence is otherwise standard, so this is a typo rather than a mathematical issue, but it should be corrected.","section":"Lemma 2.18, Eqs. (2.24)–(2.25)"},{"comment":"The statement says 'θ be the unique solution to (2.11)', but (2.11) is the smoothed-noise approximation and the dissipation measure D[θ] is defined for the inviscid equation. The intended reference is the inviscid stochastic transport equation, so the equation number should be corrected.","section":"Theorem 5.1 and first paragraph of Section 5"},{"comment":"The formal computation contains a mismatched bracket in the line '≲ E[∥θ̃^κ_T∥²_{L²_x}] + 2κ∫_0^T ∥∇θ̃^κ_t∥²_{L²_x}] dt = ∥θ_0∥²_{L²_x}': the expectation is missing on the dissipative term and there is an extra closing bracket. This is only a heuristic passage, but as written it is confusing.","section":"Section 1.1, heuristic after Eq. (1.3)"},{"comment":"The reduction to non-negative Fourier transforms is described in one sentence ('one can verify that Lemma 2.12, (2.17), and (2.23) still hold'). Since this is the bridge from the PDE back to the SPDE, a short explanation of why the Itô computation survives for complex-valued solutions with non-negative Fourier transform would improve readability.","section":"Theorem 4.13, proof of part (1)"},{"comment":"The reduction to radially symmetric correlation functions is clearly the reason the torus is excluded, as Remark 1.7 explains. It would be helpful to state explicitly at the beginning of Section 4.1 that Step 4 is the only place where the full rotation group SO(d) is used.","section":"Section 4.1, Step 4 and Remark 1.7"}],"recommendation":"minor_revision","confidential_remarks":"This is a strong paper whose central claims appear sound and whose limitations are honestly described. The local typos and the missing detail in Theorem 4.13 should be fixed in a minor revision; I do not think another full review is needed."},"author_rebuttal":null,"desk_editor":{"model":"deepseek-v4-flash","letter":"This is a serious paper and I think the main theorems hold up. It proves, for the Kraichnan model in the fully diffusive regime, three things that previously were formal or only partially known: anomalous dissipation is generic and continuous in time, solutions starting from L^2 data instantly gain H^{1-alpha-} and L^infty regularity with the expected sharp threshold, and the stochastic continuity equation has short-time variance growing like t^{1/(1-alpha)} with an explicit constant. The Duchon-Robert-type formula (Theorem 1.2) is also new and cleanly links the dissipation measure to spherical increments of the scalar.\n\nThe genuinely new ingredient is the regularity theory for the degenerate parabolic PDE satisfied by the two-point correlation function. The reduction to radial data via randomized rotations is clever, and the change of variables together with the monotonicity lemma close the estimate uniformly in the viscosity. I checked the delicate endpoint computations and the boundary-term cancellations; they work. No free parameters are fitted: K_Ric and the Duchon-Robert constant are computed from the covariance, which is the right standard.\n\nSoft spots are minor. Assumption 4.1 (exact leading isotropic scaling with beta > (2alpha-1)/(d-1)) is load-bearing, but for Kraichnan noise it is exactly equivalent to eta > 1 - d/(4alpha^2), so it is the diffusive regime, not an extra restriction. The paper leans on [LJR02] and [Kry08] for existence and parabolic regularity; both are standard and appropriate. A few \"standard approximation\" steps in the SPDE identities could be expanded, but they are not suspicious. Lemma 2.18 has an apparent typo: S^{d-2} should be S^{d-1}; the proof averages on the unit sphere, so this is cosmetic. The paper is candid about torus limitations (Remark 1.7) and about the uncertain state of general degenerate parabolic theory (Remark 1.4).\n\nI do not share any lingering suspicion that this is a formal paper dressed up as rigorous. The estimates close with constants independent of the viscosity, and the sharpness statements follow from the dichotomy theorem plus the Besov characterization. It is long, but the proof-heavy parts are detailed enough to referee. I would send it to a serious referee; my own inclination is accept after minor revision, mainly to fix the sphere typo and possibly add a sentence justifying one or two approximation passages. If you need a reference for anomalous regularization in Kraichnan models, this is now the one to cite.","headline":"Rigorous confirmation of three pillars of scalar turbulence in the Kraichnan model, with a genuinely new degenerate-parabolic regularity argument; accept after minor revision.","tokens_in":84611,"tokens_out":2257,"would_cite":true,"duration_ms":285458,"reading_group":"maybe","serious_thinker":"yes","would_accept_peer_review":true},"rs_alignment":null,"lean_confirmation":null,"pith_extraction":{"msc":["76M35","76F25","60H15"],"pacs":[],"model":"deepseek-v4-flash","headline":"This paper proves that in the Kraichnan model of isotropic turbulence, every $L^2$ initial datum of an advected scalar instantly gains sharp fractional Sobolev regularity, and that the same covariance structure fixes the exact rate of…","keywords":["Kraichnan model","anomalous dissipation","anomalous regularization","stochastic transport equation","Richardson's law","degenerate parabolic PDE","isotropic Gaussian noise","self-correlation function"],"falsifier":"Find one nonzero $L^2$ initial datum for the Kraichnan noise in the diffusive regime for which the time integral of $\\mathbb{E}\\|\\theta_r\\|^2_{H^{1-\\alpha}_x}$ is finite on some interval; this directly contradicts the sharpness claim (1.10) and would break the claimed correspondence between dissipation and the $H^{1-\\alpha}$ threshold.","tokens_in":2244,"feed_emoji":"🌪","tokens_out":5351,"duration_ms":91211,"temperature":0.7,"pith_summary":"This paper proves that the formal turbulent phenomena long associated with the Kraichnan model are genuine, mathematically provable effects. Advecting a passive scalar by a Gaussian, white-in-time, spatially $\\alpha$-Hölder velocity field, the authors show that whenever particle trajectories become spontaneously stochastic (the diffusive regime), every nonzero $L^2$ initial datum dissipates mean energy continuously in time. At the same time, the solution instantly jumps into the fractional Sobolev space $H^{1-\\alpha-\\delta}_x$ for any small $\\delta$, with explicit bounds, and this regularity is sharp: no solution can be more regular. In the incompressible case they derive an explicit Duchon--Robert-type formula for the dissipation measure, and for the stochastic continuity equation started at a Dirac delta they prove Richardson's law $\\mathbb{E}[\\operatorname{Var}(\\mu_t)] \\sim K_{\\mathrm{Ric}}\\, t^{1/(1-\\alpha)}$.","feed_headline":"Rough noise smooths turbulent scalars—only so far","feed_subtitle":"For the Kraichnan model, every $L^2$ scalar jumps into fractional smoothness at exactly the energy-dissipation threshold.","key_machinery":"The load-bearing object is the two-point self-correlation function $F^\\kappa_t(z) = \\mathbb{E}[\\int \\tilde\\theta^\\kappa_t(x+z)\\,\\tilde\\theta^\\kappa_t(x)\\,dx]$, which solves the degenerate parabolic PDE $\\partial_t F^\\kappa = (1-\\kappa)Q : D^2_z F^\\kappa + \\kappa C(0) : D^2_z F^\\kappa$. Because the noise is isotropic, $F^\\kappa$ is radial and the PDE reduces to a one-dimensional problem in $r$; the proof changes variables to $\\xi(r)$ and shows that the derivative $\\partial_\\xi g$ is uniformly bounded, which translates into $C^{2-2\\alpha-\\delta}$ regularity of $F$ near the origin and, via Lemma 2.15, into $H^{1-\\alpha-\\delta}$ Sobolev regularity of the random scalar. The reduction to radial initial data is achieved by randomly rotating the initial condition with Haar measure on $SO(d)$, and the Neumann boundary condition $\\partial_r f(t,0)=0$ cancels the boundary terms in the energy estimates.","core_discovery":"The central discovery is that the degenerate parabolic PDE satisfied by the two-point self-correlation function of the solution, $\\partial_t F = Q : D^2_z F$ with $Q(z) \\sim |z|^{2\\alpha}$, admits a sharp $C^{2-2\\alpha-\\delta}_x$ regularity theory once the covariance is isotropic, with longitudinal and normal parts scaling as $c r^{2\\alpha}$ and $\\beta c r^{2\\alpha}$ and with $\\beta > (2\\alpha-1)/(d-1)$. Through the correspondence between Hölder regularity of $F$ and Sobolev regularity of the random scalar, this yields instantaneous $H^{1-\\alpha-\\delta}_x$ regularity for every $L^2$ initial datum of the stochastic transport equation, uniformly in the vanishing-diffusivity approximation, and the gain is sharp: the time integral of the $H^{1-\\alpha}_x$ norm is infinite for every nonzero initial datum. In the divergence-free case the dissipation measure equals an explicit spherical average of second-order increments at the critical scale, and for the continuity equation the expected variance from a Dirac delta grows exactly as $K_{\\mathrm{Ric}}\\, t^{1/(1-\\alpha)}$.","pith_inferences":["If the sharp regularity theory extends to the divergence-form analogue of the degenerate PDE, the same anomalous regularization should hold for the stochastic continuity equation in the full diffusive regime, not only in the incompressible case.","The threshold $\\beta = (2\\alpha-1)/(d-1)$ suggests a testable phase transition: for noises with $\\beta$ below this value, or with non-isotropic covariance, the generic-dissipation dichotomy should fail; the paper's own remarks indicate torus and strongly anisotropic counterexamples.","Because the Richardson constant $K_{\\mathrm{Ric}}$ vanishes as the compressibility approaches the threshold $1 - d/(4\\alpha^2)$, the particle-dispersion law should degenerate continuously at the phase boundary; a numerical study of $\\mathbb{E}[\\operatorname{Var}(\\mu_t)]/t^{1/(1-\\alpha)}$ near the threshold would test the predicted rate.","The explicit dissipation formula (1.5) points toward direct experimental falsification: measuring the angular average of second-order scalar increments at small scales and comparing its $\\varepsilon\\to 0$ limit with the measured energy dissipation would test the Kraichnan mechanism in real turbulent flows."],"forward_implications":["Every nonzero $L^2$ initial datum in the diffusive regime dissipates mean energy continuously at every time; energy conservation cannot occur for any nonzero datum.","Solutions starting from $L^2$ data are instantaneously in $H^{1-\\alpha-\\delta}_x$ and $L^\\infty_x$, while the time integral of their $H^{1-\\alpha}_x$ norm is infinite, so the regularization threshold is exactly the dissipation threshold.","In the incompressible Kraichnan model the dissipation measure is given by the explicit increment formula (1.5), tying the energy loss to the angular average of increments at scale $\\varepsilon$.","The stochastic continuity equation from a Dirac delta satisfies $\\mathbb{E}[\\operatorname{Var}(\\mu_t)] \\asymp t^{1/(1-\\alpha)}$ in the diffusive regime, giving Richardson's law; the upper bound holds for every homogeneous isotropic noise.","The degenerate parabolic equation $\\partial_t F = Q : D^2_z F$ has sharp $C^{2-2\\alpha-\\delta}$ spatial regularity for the selection criterion of being a vanishing-diffusivity limit of self-correlations; this theorem is of independent interest beyond the stochastic setting."],"supporting_citations":[{"why":"Supplies the Wiener-chaos solution theory for the stochastic transport and continuity equations, the classification of diffusive regimes, and the lemma linking energy conservation to flows of maps.","marker":"[LJR02]"},{"why":"Links anomalous dissipation to non-uniqueness of particle trajectories, the conceptual basis for connecting dissipation with spontaneous stochasticity.","marker":"[BGK98]"},{"why":"The original Richardson dispersion law that Theorem 1.5 rigorously reproduces with exponent $1/(1-\\alpha)$.","marker":"[Ric26]"},{"why":"Provides the incompressible torus result and the random-rotation technique used in Step 4 to reduce to radially symmetric initial data.","marker":"[Row24]"},{"why":"Previous negative-integer Sobolev regularization result for the incompressible Kraichnan model that the present work extends to the full diffusive regime.","marker":"[CM23]"},{"why":"Strengthened regularization for the incompressible Kraichnan model, the baseline the present paper improves with pointwise-in-time estimates and sharpness.","marker":"[GGM24]"},{"why":"The Duchon--Robert formula for Euler turbulence that the dissipation measure representation (1.5) mirrors and extends to the Kraichnan setting.","marker":"[DR00]"},{"why":"Provides counterexamples and the state of the art for degenerate parabolic regularity, the obstacle that the new PDE theorem 4.13 overcomes.","marker":"[CS84]"},{"why":"Establishes Harnack-type estimates with unspecified Hölder exponents for the degenerate parabolic class, the quantitative sharpness gap filled by the present regularity result.","marker":"[CS87]"}],"fun_headline_variants":["Rough noise instantly smooths turbulent scalars","Noise-induced jump to fractional smoothness","Roughness dictates regularity and dissipation","Turbulent scalars jump to fractional smoothness instantly","Anomalous dissipation linked to roughness"],"cache_read_input_tokens":86784,"weakest_assumption_plain":"The proofs lean on the covariance of the noise being exactly isotropic, with longitudinal and normal components scaling as $c r^{2\\alpha}$ and $\\beta c r^{2\\alpha}$ at small scales with $\\beta > (2\\alpha-1)/(d-1)$; purely qualitative $\\alpha$-Hölder regularity would not suffice, and the results are stated to fail on the torus (Remark 1.7).","fun_headline_variants_meta":{"raw":{"variants":["Rough noise instantly smooths turbulent scalars","Noise-induced jump to fractional smoothness","Roughness dictates regularity and dissipation","Turbulent scalars jump to fractional smoothness instantly","Anomalous dissipation linked to roughness"]},"model":"deepseek-v4-flash","effort":"low","cost_usd":0.001238,"raw_usage":{"total_tokens":5472,"prompt_tokens":1099,"completion_tokens":4373,"prompt_tokens_details":{"cached_tokens":1024},"prompt_cache_hit_tokens":1024,"prompt_cache_miss_tokens":75,"completion_tokens_details":{"reasoning_tokens":4305}},"tokens_in":75,"tokens_out":4373,"duration_ms":278465,"temperature":1.0,"reasoning_tokens":4305,"cache_read_input_tokens":1024,"cache_creation_input_tokens":0},"cache_creation_input_tokens":0},"created_at":"2026-08-15T15:56:36.599215+00:00","model_set":{"reader":"deepseek-v4-flash"},"falsifier":"Find one nonzero $L^2$ initial datum for the Kraichnan noise in the diffusive regime for which the time integral of $\\mathbb{E}\\|\\theta_r\\|^2_{H^{1-\\alpha}_x}$ is finite on some interval; this directly contradicts the sharpness claim (1.10) and would break the claimed correspondence between dissipation and the $H^{1-\\alpha}$ threshold.","supporting_citations":[],"review_version":1}