{"id":"17b90990-ea51-4a04-81cc-30500804fe8b","arxiv_id":"2506.23365","paper_version":1,"verdict":"CONDITIONAL","confidence":"MODERATE","novelty_score":6.0,"correctness_risk":"medium","formal_verification":"none","parameter_count":0,"one_line_summary":"Assuming a geometric directional derivative remains controlled, smooth approximations converge to Yudovich-type weak solutions of the 2D density-dependent Euler equations, and such solutions are unique.","lead":"This paper proves stability and uniqueness results for density-dependent incompressible Euler equations in 2D, under a priori control of a geometric quantity: the derivative of the velocity along the direction perpendicular to the density gradient. The results are conditional, because the paper gives no example of data where this geometric control is known to hold.","discovery_kind":"extension","skeptic_critique":{"model":"deepseek-v4-flash","headline":"The uniqueness class may force Lipschitz regularity wherever |grad rho| is bounded below; genuinely sub-Lipschitz solutions in Theorem 2.9 are not exhibited and may not exist, so the 'Yudovich theory' claim is not established.","rationale":"I read the paper as a conditional framework: under a geometric a priori bound, it proves convergence to Yudovich-type solutions and uniqueness inside a class satisfying an additional bound. The proofs in Sections 3-5 are internally plausible; I found no contradiction in the estimates. The reader's weakest assumption focused on the absence of known instances of (15)/(16). That concern is only partially accurate: a simple family of axisymmetric stationary flows (rho = rho(r), u = Ω(r)e_θ) with Ω_n approximating a Hölder profile shows (15) can hold nontrivially for non-Lipschitz limits. However, those examples have unbounded vorticity and do not enter the uniqueness class. The more serious issue is structural: in the uniqueness class, condition (16) combined with bounded vorticity has a local linear-algebra consequence that forces Lipschitz regularity wherever the density gradient is nondegenerate. Thus the theorem's Yudovich-type content is unproven: no sub-Lipschitz solution satisfying (16) with nonconstant density is exhibited, and the existence theorem does not produce one, since (16) is not closed under the weak limit. I therefore maintain the conditional verdict: the mathematical implications are likely correct, but the central conceptual claim that geometric regularity recovers a genuine Yudovich theory for variable density requires an example or a proof that the uniqueness class has nontrivial sub-Lipschitz members.","tokens_in":31518,"tokens_out":30759,"duration_ms":364361,"concrete_test":"Test the non-vacuity of the uniqueness class by an analytic check: prove or disprove that for divergence-free u ∈ L∞ with ω ∈ L∞, X = grad^perp rho ∈ L∞, and ∂_X u ∈ L∞, the condition |grad rho| ≥ c > 0 on an open set implies u ∈ W^{1,∞} on that set (write S in terms of S X and X). If the implication holds, then exhibit a non-Lipschitz u with ω ∈ L∞ and a transported density rho such that all non-Lipschitz points lie in {grad rho = 0} and (16) holds; if no such configuration can satisfy the transport equation ∂_t X + u·grad X = ∂_X u, Theorem 2.9's class contains only Lipschitz solutions and the central claim should be reframed.","verdict_should_be":"UNCHANGED","load_bearing_attack":"Let S be the symmetric part of grad u. In 2D, for X = grad^perp rho, one has ∂_X u = S X + (1/2) ω X^perp. Under (16) and ω ∈ L∞, both S X and the second term are bounded. Because S is symmetric traceless, the map S ↦ S X is invertible with inverse norm comparable to 1/|X| wherever X ≠ 0. Hence, on every open set where |grad rho| ≥ c > 0, S is bounded, so grad u is bounded and u is Lipschitz there. Consequently, any non-Lipschitz solution lying in the uniqueness class of Theorem 2.9 can only have its non-Lipschitz part concentrated on the critical set {grad rho = 0}, where X = 0 and (16) gives no information. In particular, for densities whose gradient is nondegenerate on the support of the non-Lipschitz part, the assumptions of Theorem 2.9 reduce to u ∈ W^{1,∞}; the theorem then says nothing beyond classical Lipschitz uniqueness. This does not invalidate the conditional proof, but it undercuts the claim that (16) is a genuine Yudovich-type (sub-Lipschitz) mechanism. Moreover, Theorem 2.3 only produces ∂_X u as a Radon measure, not in L^1_T L∞, so no solution of the uniqueness class is guaranteed to be constructed; Remarks 2.10 and 3.1 acknowledge this gap.","agreement_with_reader":"partial"},"referee_report":{"model":"deepseek-v4-flash","summary":"The paper studies the 2D density-dependent incompressible Euler system with non-vacuum, W^{1,\\infty} density and L^2 velocity data. Its central object is the geometric quantity \\partial_X u, where X=\\nabla^\\perp\\rho. The main results are conditional: Theorem 2.3 shows that if a sequence of smooth approximate solutions satisfies the a priori bound (15), sup_n\\int_0^T \\|\\partial_{X_n}u_n\\|_{L^\\infty}dt<\\infty, then a subsequence converges to a weak solution with u\\in L^\\infty_T(L^2\\cap L^\\infty), \\nabla\\rho\\in L^\\infty, \\omega\\in L^\\infty_T(L^{p_0}) and \\nabla\\Pi\\in L^\\infty_T(L^2). Under the stronger assumption \\omega_0\\in L^\\infty, Theorem 2.8 upgrades the construction to Zygmund velocity and bounded pressure gradient. Theorem 2.9 proves uniqueness among such solutions that additionally satisfy (16), \\int_0^T\\|\\partial_X u\\|_{L^\\infty}dt<\\infty. The paper is explicit in Remark 2.7 that no construction of data or solutions verifying (15) is known, and Remarks 2.10 and 3.1 acknowledge that (16) is not obtained from (15) by weak compactness.","tokens_in":31812,"tokens_out":10667,"duration_ms":112848,"significance":"If the geometric control (15)/(16) is ever satisfied by genuinely non-Lipschitz solutions, the conditional stability and uniqueness results would be a meaningful step toward a Yudovich-type theory for non-homogeneous Euler. The proofs are detailed, follow standard Yudovich, transport and elliptic arguments, and make clever use of the cancellation div X=0 for X=\\nabla^\\perp\\rho; the paper also honestly isolates the quantities it cannot control. The main limitation is that the central theorem is conditional on an a priori bound that the paper admits it cannot derive or even instantiate, and the uniqueness class may reduce to Lipschitz regularity on sets where the density gradient is non-degenerate. These issues affect the advertised interpretation more than the internal correctness of the conditional proofs.","major_comments":[{"comment":"Theorem 2.3 defines X_n:= \\nabla\\rho_n, but every other statement and every proof, including equation (22) and Lemma 4.2, uses X_n:=\\nabla^\\perp\\rho_n. As printed, assumption (15) is not the quantity that appears in the vorticity equation, the transport equation for X_n, or the geometric regularity discussion. The statement must be corrected to X_n:=\\nabla^\\perp\\rho_n, or the proofs must be reconciled with the stated definition.","section":"Theorem 2.3"},{"comment":"The manuscript is fully transparent that no construction of initial data or approximate solutions satisfying (15) is known, and Lemma 4.2 yields only \\partial_X u\\in M([0,T];L^\\infty), not the L^1_TL^\\infty condition (16) needed for the uniqueness class. Consequently Theorem 2.3 does not produce any solution in the uniqueness class of Theorem 2.9, and the existence of nontrivial, genuinely non-Lipschitz instances of the theory remains open. This is an admitted, load-bearing gap for the advertised 'Yudovich theory' claim, and it should be either filled by an explicit example or family of examples, or stated prominently as an open condition in the abstract and introduction.","section":"Remark 2.7 and Lemma 4.2"},{"comment":"Under (16) and \\omega\\in L^\\infty, on any open set where |\\nabla\\rho|\\ge c>0 the symmetric traceless part S of \\nabla u satisfies S X = \\partial_X u - \\tfrac12 \\omega X^\\perp. Since the right-hand side is bounded by assumption (16) and bounded vorticity, S is bounded with norm comparable to C/|\\nabla\\rho| on that set; hence \\nabla u is bounded and u is Lipschitz there. Thus the uniqueness class of Theorem 2.9 can be non-Lipschitz only on the critical set \\{\\nabla\\rho=0\\}. This does not invalidate the conditional proof, but it substantially weakens the interpretation of (16) as a genuine Yudovich-type sub-Lipschitz mechanism. The paper should discuss this structure and, ideally, provide an example or a discussion of whether non-Lipschitz solutions in this class can actually exist.","section":"Section 5.2, Theorem 2.9"}],"minor_comments":[{"comment":"In the statement of Proposition 5.3, the vorticity is written as \\omega=\\partial_1u_1-\\partial_2u_1; it should be \\partial_1u_2-\\partial_2u_1, as in equation (5).","section":"Proposition 5.3"},{"comment":"The notation M([0,T];L^\\infty) for Radon measures in time is used without being defined; a one-line definition would improve readability.","section":"Lemma 4.2"},{"comment":"The abstract states that uniqueness 'improves previous uniqueness results for regular solutions', but the role of the a priori geometric condition (16) is not mentioned there. Adding one sentence about the conditional nature of the results would align the abstract with the body of the paper.","section":"Abstract and Introduction"}],"recommendation":"major_revision","confidential_remarks":"The paper is honest about its conditional nature, which is a genuine strength of the exposition. The main obstacle to acceptance is not a detected mathematical error but the fact that the central statement is an a priori conditional theorem whose hypotheses are not shown to be satisfiable, combined with a uniqueness class that may force Lipschitz behaviour wherever the density gradient is non-degenerate. The typo in Theorem 2.3 should be corrected, and the claims about a 'Yudovich theory' should be carefully recalibrated. I would not recommend rejection, since the techniques and conditional results are likely useful, but the paper needs substantial revision before it can be accepted."},"author_rebuttal":null,"desk_editor":{"model":"deepseek-v4-flash","letter":"Dear colleague,\n\nShort version: this is a careful, honest paper that proves conditional stability and uniqueness for the 2D density-dependent Euler equations under a geometric bound on the directional derivative of u along X = ∇⊥ρ. The proofs follow standard Yudovich/transport/elliptic lines and look plausible. But the key assumption (15)/(16) is not derived from the equations, no example of data satisfying it is known, and the uniqueness class may not contain genuinely non-Lipschitz solutions except where ∇ρ vanishes. So the advertised 'Yudovich theory' is not actually established beyond the Lipschitz realm.\n\nWhat's new: Theorem 2.3 shows that a sequence of smooth approximations with uniform control of ∂_{X_n}u_n converges to a weak solution with u ∈ L∞(L2∩L∞), ∇ρ ∈ L∞, ω ∈ L∞(Lp0), ∇Π ∈ L∞(L2). Theorem 2.9 gives uniqueness for solutions with ∂_X u ∈ L1_T(L∞). The exposition is clean and the paper states its limitations explicitly, notably Remark 2.7.\n\nWhere it's soft: the main soft spot is conceptual. From the identity ∂_X u = S X + (1/2)ω X⊥ with S the symmetric part of ∇u, the bound (16) plus ω ∈ L∞ forces S X bounded. Since S ↦ SX is invertible with norm ∼ 1/|X| where X ≠ 0, u is Lipschitz on every open set where |∇ρ| ≥ c > 0. Hence any non-Lipschitz solution in the uniqueness class has to be concentrated on the critical set {∇ρ = 0}. For densities with non-degenerate gradient, assumption (16) reduces to Lipschitz uniqueness, which is already known. So (16) is not a Yudovich-type sub-Lipschitz mechanism unless the density gradient degenerates exactly where u fails to be Lipschitz. The paper doesn't construct such examples. Also, Theorem 2.3 only yields ∂_X u as a Radon measure, not in L1_T(L∞), so the existence argument doesn't produce solutions in the uniqueness class; the paper acknowledges this in Remarks 2.10 and 3.1.\n\nThe math seems sound to me. I didn't verify every estimate, but the structure is standard and the honesty is disarming. The citation pattern is fine; the main framework comes from the author's own prior work.\n\nWho it's for: people working on non-homogeneous incompressible fluids who want a clean conditional statement and a transparent roadmap. It deserves a serious referee – the conditional results are meaningful, and a referee could either find a way to discharge (15) or sharpen the obstruction. But I wouldn't cite it as evidence of Yudovich uniqueness for variable density, because that's not what it shows.\n\nRecommendation: send it to a competent referee. It's a good paper of its kind, but the framing should probably be toned down to 'conditional stability/uniqueness under a non-verifiable geometric assumption'.","headline":"Conditional stability and uniqueness are clean and honestly presented, but the geometric assumption has no known nontrivial instances and the uniqueness class collapses to Lipschitz wherever |grad rho| is bounded below, so the 'Yudovich theory' framing overreaches.","tokens_in":32389,"tokens_out":3871,"would_cite":false,"duration_ms":39241,"reading_group":"maybe","serious_thinker":"yes","would_accept_peer_review":true},"rs_alignment":null,"lean_confirmation":null,"pith_extraction":{"msc":["35Q31","35R05","76B03","35A02"],"pacs":[],"model":"deepseek-v4-flash","headline":"Controlling one directional derivative recovers Yudovich stability and uniqueness for density-dependent incompressible Euler in two dimensions.","keywords":["incompressible Euler equations","density variations","Yudovich theory","geometric regularity","directional derivative","stability and convergence","uniqueness","two-dimensional Euler"],"falsifier":"Compute $\\sup_n \\int_0^T\\|\\partial_{X_n}u_n\\|_{L^\\infty}dt$ for the standard mollified regularisation of a non-Lipschitz initial vorticity $\\omega_0\\in L^{p_0}\\cap L^\\infty$; if the quantity diverges for some datum, Theorem 2.3 has no conclusion for that datum. Alternatively, exhibit two weak solutions with the same initial data, both satisfying (16), that differ on a set of positive measure; that would disprove Theorem 2.9.","tokens_in":31194,"feed_emoji":"🌊","tokens_out":9099,"duration_ms":95244,"temperature":0.7,"pith_summary":"The paper's aim is to bring Yudovich's weak-solution theory for the two-dimensional incompressible Euler equations to the variable-density case, where the velocity need not be Lipschitz. It proves two conditional statements. The first is a stability result: if a sequence of smooth approximate solutions obeys a uniform bound on the time-integrated $L^\\infty$ norm of the single directional derivative $\\partial_X u$ of velocity along the vector field $X := \\nabla^\\perp\\rho$ tangent to density level sets, then a subsequence converges to a weak solution of the density-dependent Euler system with bounded density gradient, velocity in $L^2 \\cap L^\\infty$, vorticity in $L^{p_0}$, and pressure gradient in $L^2$. The second is a uniqueness result: among such weak solutions, at most one also has a finite time-integrated $L^\\infty$ norm of $\\partial_X u$. The point is that one geometric direction can stand in for the full Lipschitz control that the classical Yudovich argument uses; the paper notes explicitly that no non-Lipschitz initial data are known to satisfy the required geometric control.","feed_headline":"One derivative bound gives Yudovich theory for density-dependent Euler","feed_subtitle":"If that derivative stays integrable along density contours, non-Lipschitz velocity solutions exist and are unique.","key_machinery":"The load-bearing object is the directional derivative $\\partial_X u := (X\\cdot\\nabla)u$ with $X := \\nabla^\\perp\\rho$, treated as an indivisible geometric quantity rather than bounded through $X$ times $\\nabla u$. It appears as the source in the transport equations for $X$ and for the momentum vorticity $\\eta := \\operatorname{curl}(\\rho u)$: $\\partial_t X+u\\cdot\\nabla X=\\partial_X u$ and $\\partial_t\\eta+u\\cdot\\nabla\\eta=\\partial_X u\\cdot u$. Since $\\partial_X\\rho=0$ and $\\operatorname{div} X=0$, this term can be manipulated and estimated without a Lipschitz bound on $u$. The pressure gradient is recovered from the elliptic equation $-\\operatorname{div}(\\rho^{-1}\\nabla\\Pi)=\\operatorname{div}((u\\cdot\\nabla)u)$, and uniqueness is closed by Yudovich's $L^2$ stability estimate with the $p\\to\\infty$ differential-inequality argument.","core_discovery":"The central claim is that Yudovich theory survives in the non-homogeneous setting once a specific geometric quantity is controlled. With initial density in $W^{1,\\infty}$ bounded away from zero, initial velocity in $L^2$ divergence-free, and initial vorticity in $L^{p_0}$ for some $p_0\\in(2,4]$, Theorem 2.3 says that under assumption (15), namely $\\sup_n \\int_0^T \\|\\partial_{X_n} u_n\\|_{L^\\infty}\\,dt<\\infty$ for smooth approximations, a subsequence of approximations converges to a weak solution of (2) satisfying the bounds stated in items (i)-(iv). Theorem 2.8 adds that if $\\omega_0\\in L^\\infty$, the limit velocity is Zygmund continuous, hence log-Lipschitz, and the pressure gradient is bounded. Theorem 2.9 states that at most one Yudovich-type solution additionally satisfies (16), namely $\\int_0^T\\|\\partial_X u\\|_{L^\\infty}\\,dt<\\infty$. This uniqueness requires less smoothness than earlier uniqueness results for regular solutions.","pith_inferences":["Editorial extension: the theorems carry no nontrivial instance until someone produces non-Lipschitz initial data satisfying (15); the most direct next step is to look for such data in a striated-regularity class, where the directional derivative is controlled by geometry rather than by pointwise differentiability.","Editorial extension: because the density is assumed $W^{1,\\infty}$, discontinuous densities such as vortex patches lie outside the theory; whether a weak interpretation of $\\partial_X u$ across a jump interface can still enforce uniqueness is an open problem the paper does not address.","Editorial extension: the condition $p_0\\le 4$ is used only in the pressure estimate of Proposition 5.5, so a different pressure argument should remove it; this is a testable extension rather than a claim of the paper.","Editorial extension: the stronger integrability assumption (17) is the only hypothesis in the paper that transfers the geometric control to the limit solution, which suggests monitoring $\\int_0^T\\|\\partial_{X_n}u_n\\|_{L^\\infty}^{p^*}\\,dt$ as a concrete criterion for regularisation schemes."],"forward_implications":["If assumption (15) holds up to time $T$, Theorem 2.3 yields a Yudovich-type weak solution of the density-dependent Euler system up to $T$; if the control holds for $T=+\\infty$, the construction gives a global-in-time solution.","Under the stronger initial condition $\\omega_0\\in L^\\infty$, Theorem 2.8 upgrades the limit velocity to the Zygmund class, and therefore to log-Lipschitz regularity, so the velocity admits a unique flow.","Theorem 2.9 confines the class of Yudovich-type solutions: at most one solution can carry a finite $\\int_0^T\\|\\partial_X u\\|_{L^\\infty}dt$, and this uniqueness holds without the $L^1_T(L^\\infty)$ gradient control that earlier results required.","Replacing (15) by the stronger assumption (17), with $\\partial_{X_n}u_n$ bounded in $L^{p^*}_T(L^\\infty)$ for some $p^*>1$, makes the limit quantity $\\partial_X u$ itself lie in $L^1_T(L^\\infty)$, closing the gap between the stability and uniqueness hypotheses."],"supporting_citations":[{"why":"Introduces the geometric quantity $\\partial_X u$ and the transport equations for $X$ and $\\eta$ on which the whole argument is built.","marker":"[21]"},{"why":"Provides the $L^p$ well-posedness theory and Lemma 2, the elliptic estimate used for the pressure gradient.","marker":"[17]"},{"why":"Gives local well-posedness for smooth regularised data, producing the approximate solutions that the stability theorem starts from.","marker":"[18]"},{"why":"Yudovich's 1963 theorem for homogeneous Euler, the model that this paper extends to variable density.","marker":"[30]"},{"why":"Supplies the template for $L^2$ stability estimates, the $p\\to\\infty$ uniqueness argument, and the flow theory for log-Lipschitz velocities.","marker":"[24]"},{"why":"Provides the Besov characterisation of the Zygmund class, the embedding $Z\\subset LL$, and the unique-flow result for log-Lipschitz vector fields.","marker":"[3]"},{"why":"Supplies the renormalised-solution theorem used to divide the momentum equation by $\\rho$ and derive the vorticity formulation.","marker":"[22]"}],"fun_headline_variants":["Yudovich for density-dependent Euler from one derivative bound","Derivative along density contours yields Yudovich Euler","Non-homogeneous Euler: Yudovich under geometric control","One directional derivative bound gives Euler Yudovich uniqueness","Geometric regularity extends Yudovich to variable-density Euler"],"cache_read_input_tokens":3200,"weakest_assumption_plain":"The theorems collapse if the a priori bound on $\\int_0^T\\|\\partial_{X_n}u_n\\|_{L^\\infty}dt$ (and its limit analogue $\\int_0^T\\|\\partial_X u\\|_{L^\\infty}dt$) is never achievable for genuinely non-Lipschitz data, since the paper states in Remark 2.7 that no construction of such data or solutions is known; a separate structural restriction is the $W^{1,\\infty}$ regularity of the density, which excludes discontinuous densities such as vortex patches.","fun_headline_variants_meta":{"raw":{"variants":["Yudovich for density-dependent Euler from one derivative bound","Derivative along density contours yields Yudovich Euler","Non-homogeneous Euler: Yudovich under geometric control","One directional derivative bound gives Euler Yudovich uniqueness","Geometric regularity extends Yudovich to variable-density Euler"]},"model":"deepseek-v4-flash","effort":"low","cost_usd":0.000346,"raw_usage":{"total_tokens":1938,"prompt_tokens":1030,"completion_tokens":908,"prompt_tokens_details":{"cached_tokens":384},"prompt_cache_hit_tokens":384,"prompt_cache_miss_tokens":646,"completion_tokens_details":{"reasoning_tokens":828}},"tokens_in":646,"tokens_out":908,"duration_ms":9494,"temperature":1.0,"reasoning_tokens":828,"cache_read_input_tokens":384,"cache_creation_input_tokens":0},"cache_creation_input_tokens":0},"created_at":"2026-08-06T21:45:38.408417+00:00","model_set":{"reader":"deepseek-v4-flash"},"falsifier":"Compute $\\sup_n \\int_0^T\\|\\partial_{X_n}u_n\\|_{L^\\infty}dt$ for the standard mollified regularisation of a non-Lipschitz initial vorticity $\\omega_0\\in L^{p_0}\\cap L^\\infty$; if the quantity diverges for some datum, Theorem 2.3 has no conclusion for that datum. Alternatively, exhibit two weak solutions with the same initial data, both satisfying (16), that differ on a set of positive measure; that would disprove Theorem 2.9.","supporting_citations":[{"cited_title":"Danchin: On the well-posedness of the incompressible density-dependent Euler equations in the Lp framework","cited_arxiv_id":null,"evidence_quote":"Provides the $L^p$ well-posedness theory and Lemma 2, the elliptic estimate used for the pressure gradient."},{"cited_title":"Danchin, F","cited_arxiv_id":null,"evidence_quote":"Gives local well-posedness for smooth regularised data, producing the approximate solutions that the stability theorem starts from."},{"cited_title":null,"cited_arxiv_id":null,"evidence_quote":"Yudovich's 1963 theorem for homogeneous Euler, the model that this paper extends to variable density."},{"cited_title":"Fourier analysis and nonlinear partial differential equa- tions","cited_arxiv_id":null,"evidence_quote":"Provides the Besov characterisation of the Zygmund class, the embedding $Z\\subset LL$, and the unique-flow result for log-Lipschitz vector fields."},{"cited_title":"Singular limits in thermodynamics of viscous fluids","cited_arxiv_id":null,"evidence_quote":"Supplies the renormalised-solution theorem used to divide the momentum equation by $\\rho$ and derive the vorticity formulation."}],"review_version":1}