{"id":"fef0578a-3da4-411b-a72d-d53232319b4d","arxiv_id":"2607.07541","paper_version":1,"verdict":"ACCEPT","confidence":"MODERATE","novelty_score":7.0,"correctness_risk":"unknown","formal_verification":"none","parameter_count":3,"one_line_summary":"Non-trivial m-fold doubly-connected stationary vortex patches are proven to exist for the quasi-geostrophic shallow-water equations via bifurcation from annuli.","lead":"This paper proves the existence of non-trivial stationary vortex patches for the quasi-geostrophic shallow-water equations using bifurcation theory. A generalist might read it to understand how mathematical analysis reveals stable fluid structures relevant to atmospheric and oceanic dynamics.","discovery_kind":"unclear","skeptic_critique":{"model":"glm-5.2","headline":"Transversality for the b-bifurcation rests on a delicate leading-order cancellation; the sign argument is the most fragile link.","rationale":"The reader correctly identified the transversality condition as the weakest link. Having examined the argument in detail, I find that the concern is real but localized: the b-bifurcation transversality depends on a chain of inequalities involving β(λ), Bessel function products, and the Wronskian identity, where the key step is the lower bound (2.39) on β(λ). The λ-bifurcation transversality is comparatively straightforward. The logical structure of the paper is sound: the Crandall-Rabinowitz framework is correctly applied, the Fredholm property is properly established, and the kernel characterization is clean. The asymptotic expansions of the roots (Lemmas 2.4 and 2.5) are derived through careful analysis, and the key identity (2.40) connecting the root location to the function Φ is well-motivated. The paper also provides independent support through numerical simulations (Figures 2-5) that validate the analytical predictions. The rigidity results (Theorem 1.2) follow from established methods. The overall verdict of ACCEPT with MODERATE confidence is appropriate: the result is a genuine advance, the proof strategy is sound, but the dense asymptotic computations — particularly the transversality sign argument for the b-bifurcation — would benefit from independent verification. No adjustment to the verdict is needed.","tokens_in":44303,"tokens_out":1041,"duration_ms":712695,"concrete_test":"Independently verify the inequality ϖ(1/(2x)) > 2x(1-x) used in the proof of (2.39), where ϖ(x) = (2x-1+e^{-2x})/(2x²). Specifically, for x = λI₀(λ)K₁(λ) ∈ (1/2, 1), check numerically whether ϖ(1/(2x)) - 2x(1-x) > 0 holds for all λ > 0, with particular attention to the regime λ → 0⁺ (where x → 1⁻ and both sides → 0) and λ → ∞ (where x → 1/2⁺). If the inequality fails for any λ, the bound β(λ) > 1/(2λI₀K₁) is not established, and the sign of τ(λ) requires an alternative argument.","verdict_should_be":"UNCHANGED","load_bearing_attack":"The reader correctly identifies the transversality condition (Propositions 2.4(iv) and 2.6(iv)) as the load-bearing assumption. I focus on the b-bifurcation case (Proposition 2.4(iv)), which is the more delicate of the two. The transversality expression T_{m,λ} is shown to satisfy T_{m,λ} ~ -τ(λ)/(4m) + O(1/m³), where τ(λ) is a scalar function of β(λ) and the Bessel function products. The non-vanishing of τ(λ) is established by the chain of inequalities: (i) the Wronskian identity (2.44) gives λI₁K₀ < 1/2 < λI₀K₁, (ii) the lower bound β(λ) > 1/(2λI₀K₁) from (2.39) is combined with this to get λI₁K₀ - 2λβI₀K₁ < -1/2 < 0, and (iii) this yields τ(λ) ≤ e^{-2β}(λI₁K₀ - 2λβI₀K₁) < 0. The critical step is (ii): the bound β(λ) > 1/(2λI₀(λ)K₁(λ)) is derived via the function ϖ(x) = 1 - Φ(x), where Φ is defined in (2.41). The argument shows ϖ(1/(2λI₀K₁)) > ϖ(β(λ)) by establishing ϖ(1/(2x)) > 2x(1-x) and using the Wronskian. This chain of inequalities is the unique point where a sign error or an insufficiently tight bound would propagate to invalidate the transversality. The λ-bifurcation transversality (Proposition 2.6(iv)) is more robust: the leading-order matrix is diag(1, b) with b ∈ (0,1), giving T_{m,b} ~ -b(1-b²)^{5/2}/(4√(m log m)) → -∞, where the sign is unambiguous. So the concern is specifically about the b-bifurcation transversality sign argument.","agreement_with_reader":"agree"},"referee_report":{"model":"glm-5.2","summary":"This paper proves the existence of m-fold doubly-connected stationary vortex patches for the quasi-geostrophic shallow-water (QGSW) equations via Crandall–Rabinowitz bifurcation from annuli. Two bifurcation regimes are analyzed: one with the inner radius b as the bifurcation parameter (fixed λ), and one with the inverse Rossby radius λ as the parameter (fixed b). The core technical difficulty lies in the spectral analysis of the linearized operator, which involves modified Bessel functions depending on both the order and the argument as large parameters. The paper also proves rigidity results for simply-connected V-states (Theorem 1.2), showing that stationary simply-connected patches must be discs, and that sufficiently fast rotation forces radial symmetry.","tokens_in":44654,"tokens_out":3704,"duration_ms":130696,"significance":"The results are novel: stationary doubly-connected patches for QGSW do not exist in the Euler limit (λ = 0), making this a genuinely QGSW phenomenon. The b-bifurcation result is analogous to Gómez-Serrano's gSQG result [23] but requires substantially more delicate analysis due to the non-homogeneity of the QGSW kernel. The λ-bifurcation—using the model parameter itself as the bifurcation variable—is a new idea. The asymptotic laws (1–b_{m,λ} ~ β(λ)/m and λ_{m,b} ~ 2/√((1–b²)m log m)) are explicitly characterized. The rigidity results complete a natural classification table (Tables 1–2). The transversality verifications, particularly for the b-bifurcation, involve nontrivial sign arguments using Wronskian identities and monotonicity of Bessel function products; these are the most technically demanding parts of the paper and are carried out carefully.","major_comments":[{"comment":"Proposition 2.4(iv), transversality for the b-bifurcation: I have carefully checked the chain of inequalities establishing τ(λ) < 0. The argument proceeds as follows: (a) the Wronskian identity (2.44) gives λI₁K₀ < 1/2 < λI₀K₁; (b) the lower bound β(λ) > 1/(2λI₀K₁) from (2.39) is derived via the function ϖ(x) = 1−Φ(x), using the elementary inequality ϖ(1/(2x)) > 2x(1−x) and the Wronskian to conclude ϖ(1/(2λI₀K₁)) > ϖ(β(λ)); (c) combining (a) and (b) yields λI₁K₀ − 2λβI₀K₁ < −1/2 < 0; (d) since the second term −λI₁K₀(1−2λβI₀K₁)² in τ(λ) is non-positive, one obtains τ(λ) ≤ e^{−2β}(λI₁K₀ − 2λβI₀K₁) < 0. This argument is correct and the bound is in fact robust—it does not require tight asymptotics on β(λ), only the strict inequality from (2.39). The transversality for the λ-bifurcation (Proposition 2.6(iv)) is more straightforward, with T_{m,b} ~ −b(1−b²)^{5/2}/(4√(m log m)) → −∞. No load-ba","section":null},{"comment":"Remark 2.4 / Proposition 2.5: The λ-bifurcation result is stated on (0, λ_max) rather than (0, ∞). The authors acknowledge this limitation and explain the difficulty (uniform control of D_{n,b} for large λ). While this does not affect the main theorem as stated (Theorem 1.1(ii) only requires λ_{m,b} → 0), it would strengthen the paper if the authors could at least sketch how the uniform convergence (2.7) combined with the large-argument asymptotics (A.10) could be used to exclude zeros in [λ̄, ∞) for n large, even if the full argument is deferred. At minimum, the authors should clarify in the statement of Theorem 1.1(ii) that the result holds for λ in a neighborhood of 0 (which is what the asymptotic λ_{m,b} → 0 implies), so that readers are not confused about the scope.","section":null}],"minor_comments":[{"comment":"Section title 2.1.2: 'linearzation' should be 'linearization'.","section":null},{"comment":"Appendix A title: 'Formular on modified Bessel functions' should be 'Formulas on modified Bessel functions' or 'Formulae on modified Bessel functions'.","section":null},{"comment":"The word 'asymtotic' appears multiple times (e.g., in the proof of Lemma 2.5, before equation (2.50); in the text before Lemma 2.4). Should be 'asymptotic'.","section":null},{"comment":"Section 2.3, proof of Proposition 2.6: 'transersality' should be 'transversality'.","section":null},{"comment":"The proof of the positivity of α(λ) in Lemma 2.5 (pages 23–25) is quite lengthy. While the argument is correct, it would benefit from a brief summary at the beginning stating the strategy (reducing to showing Ψ(x) > 0, then to h(x) > 0, then to the quadratic Q_x analysis), so the reader can follow the logical structure more easily.","section":null},{"comment":"Equation (2.11): the decomposition of D_n into D_∞ plus correction terms is used repeatedly but is only displayed inline. Giving it an equation number and referencing it explicitly would improve readability.","section":null},{"comment":"Figures 2–5 are referenced but not visible in the manuscript text provided. The authors should ensure these are properly included and captioned in the final version.","section":null},{"comment":"In the proof of Theorem 1.2(ii), the function x ↦ I₁(x)/x is stated to be increasing on (0,∞) 'from (A.1)'. A one-line justification (e.g., differentiating the power series) would be helpful.","section":null},{"comment":"The reference [34] is cited for derivative bounds on I_nK_n (equation (2.24)) and for the regularity of the functional with respect to λ. Since [34] appears to be a memoir by two of the authors, it would be appropriate to state the relevant results explicitly rather than referring the reader to page numbers, at least for the key estimate (2.24).","section":null},{"comment":"Corollary 3.1: the estimate involves f_λ^{-1}(Ω), but the domain of f_λ is stated as (0,∞) → (0, 1/2). It should be clarified that f_λ is applied to the spatial variable (i.e., f_λ(x) = I₁(λx)K₁(λx)), not to λ itself, to avoid confusion with the notation elsewhere in the paper.","section":null}],"recommendation":"minor_revision","confidential_remarks":"The stress-test concern about the b-bifurcation transversality (Proposition 2.4(iv)) does not, on careful reading, identify an actual error. The sign argument for τ(λ) < 0 is correct: the key bound β(λ) > 1/(2λI₀K₁) follows from a clean and verifiable chain of inequalities, and the subsequent deduction that τ(λ) < 0 uses only that the second term in τ(λ) is non-positive. The argument is robust in the sense that it does not depend on tight asymptotics. The λ-bifurcation transversality is even more straightforward. I recommend minor revision for presentation improvements only."},"author_rebuttal":{"model":"glm-5.2","summary":"We thank the referee for the careful reading and the positive assessment. The referee raises two points: (1) a verification of the transversality argument in Proposition 2.4(iv), which the referee confirms is correct, and (2) a request regarding the lambda-bifurcation result (Proposition 2.5 / Theorem 1.1(ii)) to either sketch how zeros of D_{n,b} could be excluded for large lambda, or at minimum clarify the scope of Theorem 1.1(ii) in its statement. We address both below.","responses":[{"response":"We thank the referee for the careful verification of the transversality argument in Proposition 2.4(iv). The referee's reconstruction of the proof is accurate: the Wronskian identity (2.44) gives the key inequality (2.58), the lower bound (2.39) on beta(lambda) is derived via the function varpi and the elementary inequality varpi(1/(2x)) > 2x(1-x), and combining these yields tau(lambda) <= e^{-2beta}(lambda I_1 K_0 - 2 lambda beta I_0 K_1) < -1/2 e^{-2beta} < 0. We agree that the bound is robust in that it does not require tight asymptotics on beta(lambda), only the strict inequality from (2.39). No revision is needed for this point.","revision_made":"no","referee_comment":"Proposition 2.4(iv), transversality for the b-bifurcation: The referee has carefully checked the chain of inequalities establishing tau(lambda) < 0 and confirms the argument is correct and robust."},{"response":"We agree with the referee that the statement of Theorem 1.1(ii) should be clarified to make the scope explicit. The result indeed produces bifurcation points lambda_{m,b} converging to 0, so the theorem holds for lambda in a neighborhood of 0. We will add a clarifying sentence to the statement of Theorem 1.1(ii) making this explicit, so that readers are not confused about the scope. Regarding the suggestion to sketch how zeros could be excluded for large lambda: we agree that the natural approach would combine the uniform convergence (2.7) with the large-argument asymptotics (A.10). The key observation is that D_{infty,b}(lambda) < 0 for all lambda > 0 by (2.9), so for any fixed compact [0, Lambda], uniform convergence already excludes zeros for n large. The difficulty, as noted in Remark 2.4, is obtaining uniform control on [Lambda, infinity): the large-argument asymptotics (A.10) show that I_n(lambda) K_n(lambda) ~ 1/(2 lambda) for lambda >> n, but the regime where lambda is comparable to n requires more refined uniform asymptotics (Debye-type expansions) that would substantially lengthen the paper. We will expand Remark 2.4 to sketch this strategy and explain the remaining difficulty more precisely, while keeping the full argument for future work. This is a partial revision: the clarification in the theorem statement is straightforward, and the expanded remark provides the requested sketch, but the complete proof of non-existence of zeros for large lambda is deferred.","revision_made":"partial","referee_comment":"Remark 2.4 / Proposition 2.5: The lambda-bifurcation result is stated on (0, lambda_max) rather than (0, infinity). The referee requests either a sketch of how uniform convergence (2.7) combined with large-argument asymptotics (A.10) could exclude zeros in [lambda_bar, infinity) for n large, or at minimum a clarification in the statement of Theorem 1.1(ii) that the result holds for lambda in a neighborhood of 0."}],"tokens_in":44428,"tokens_out":827,"duration_ms":134415,"standing_objections":[]},"desk_editor":{"model":"glm-5.2","letter":"This paper proves the existence of non-trivial stationary doubly-connected vortex patches for the QGSW equations, via Crandall–Rabinowitz bifurcation from annuli, in two regimes: fixing λ and varying the inner radius b, and fixing b and varying λ. It also proves rigidity of simply-connected V-states (disc symmetry for Ω ≤ 0 or Ω large enough). The stationary doubly-connected QGSW result is genuinely new — the analogous gSQG result is due to Gómez-Serrano, and the QGSW rotating (non-stationary) case is in Roulley's prior work, but the stationary case requires different and harder analysis because the QGSW kernel is non-homogeneous and the Bessel function structure is more rigid than power-law kernels. The rigidity result (Theorem 1.2) is a clean extension of the Gómez-Serrano–Park–Shi–Yao framework and provides good motivation for the doubly-connected construction. The two bifurcation regimes are complementary and both are carried out in full: existence, uniqueness, and simplicity of zeros of the determinant, Fredholmness, and transversality. The Bessel function analysis in Section 2 and Appendix A is extensive and clearly the bulk of the technical work. The asymptotic expansion of the roots b_{m,λ} ~ β(λ)/m (Lemma 2.4) is the key new ingredient and is derived carefully — the obstruction from Taylor expansion failing (because derivatives of Λ_n grow polynomially in n) is real, and the resolution via the integral representation and the function Φ is the right approach. The λ-bifurcation transversality (Proposition 2.6(iv)) is clean: the leading-order matrix is diag(1, b), giving T_{m,b} ~ -b(1-b²)^{5/2}/(4√(m log m)) with unambiguous sign. The b-bifurcation transversality (Proposition 2.4(iv)) is the delicate point, as the reader and stress-test both flag. The argument reduces to showing τ(λ) < 0, where τ involves a cancellation at leading order. The sign is established through a chain: the Wronskian identity gives λI₁K₀ < 1/2 < λI₀K₁, the lower bound β(λ) > 1/(2λI₀K₁) comes from comparing ϖ(1/(2x)) > 2x(1-x) with ϖ(β(λ)) = 2λ²I₀K₀I₁K₁, and then λI₁K₀ - 2λβI₀K₁ < 1/2 - 1 < 0. I checked this chain and it holds — the inequality ϖ(1/(2x)) > 2x(1-x) follows from e^{-1/x} > 0, the Wronskian is standard, and ϖ is strictly decreasing because Φ is strictly increasing. The second term in τ(λ) is non-positive, so the upper bound is valid. This is tight but correct. The soft spots are minor relative to the whole. The asymptotic computations are dense and a referee should verify them line by line — particularly the O(1/m) remainder control in the b-bifurcation transversality, where the leading term is O(1/m) and one needs the remainder to be genuinely lower order. The restriction to λ < λ_max in the λ-bifurcation (Proposition 2.5) is acknowledged by the authors as a technical simplification; the numerical evidence suggests it's unnecessary, but the proof as written only covers bounded λ. This is a minor gap that doesn't affect the main claims. The paper is for researchers in PDE methods for geophysical fluid dynamics, specifically those working on vortex patch existence and symmetry. It deserves a serious referee who can independently verify the Bessel function asymptotics. I recommend sending it out for review.","headline":"Letter to colleague","tokens_in":45163,"tokens_out":2353,"would_cite":true,"duration_ms":155300,"reading_group":"no","serious_thinker":"yes","would_accept_peer_review":true},"rs_alignment":null,"lean_confirmation":null,"pith_extraction":{"msc":[],"pacs":[],"model":"glm-5.2","headline":"Stationary vortex patches exist for shallow-water flows","keywords":[],"falsifier":"A numerical computation at a specific (λ, m) pair showing that the transversality expression T_{m,λ} or T_{m,b} is zero or has the wrong sign, contradicting the asymptotic positivity/negativity derived in Propositions 2.4(iv) and 2.6(iv).","tokens_in":44562,"feed_emoji":"🌀","tokens_out":1191,"duration_ms":154945,"temperature":0.7,"pith_summary":"The paper proves that the quasi-geostrophic shallow-water (QGSW) equations — a model for large-scale atmospheric and oceanic circulation that includes Coriolis effects via a parameter λ (the inverse Rossby radius) — admit non-trivial stationary vortex patches with doubly-connected (annular) geometry and m-fold symmetry. The authors establish this through a bifurcation analysis from the trivial annular equilibrium, using either the inner radius of the annulus or the inverse Rossby radius as the bifurcation parameter. In the first regime, for any fixed λ, they show that m-fold symmetric stationary patches bifurcate from thin annuli whose inner radius approaches 1 as 1 - b ~ β(λ)/m. In the second regime, for any fixed annulus geometry, they show that stationary patches bifurcate at values of λ tending to zero at the rate λ ~ 2/√((1-b²)m log m). The central technical difficulty is that the spectral analysis of the linearized operator involves modified Bessel functions depending simultaneously on a large order and a varying argument, which defeats standard Taylor expansion and requires delicate asymptotic summation of all Taylor orders. The authors also prove that simply-connected stationary patches must be discs, confirming that the doubly-connected framework is necessary for non-trivial stationary solutions.","feed_headline":"Stationary vortex patches found for shallow-water flows","feed_subtitle":"Coriolis effects enable non-trivial steady vortex rings impossible in classical Euler, bifurcating from annuli at precise asymptotic rates.","key_machinery":"The Crandall-Rabinowitz bifurcation theorem applied to a contour dynamics functional linearized at annular equilibria, with the linearized operator being a 2×2 Fourier multiplier matrix M_n(λ,b) whose determinant D_n(λ,b) involves modified Bessel functions I_n and K_n. The proof hinges on: (1) uniform convergence of D_n to a limiting profile D_∞ that is strictly negative on (0,1), forcing zeros to accumulate at the boundary; (2) a resummed Taylor expansion overcoming the fact that all derivatives of Λ_n grow polynomially in n; (3) a transversality check requiring the first non-vanishing term in an asymptotic expansion of a scalar product, where the leading-order term vanishes and higher-corO","core_discovery":"The determinant D_n(λ,b), whose zeros are the candidate bifurcation points, changes sign inside (0,1) for each parameter due to the QGSW Green kernel's non-homogeneous structure — a phenomenon absent in the Euler limit (λ=0) where the corresponding frequencies never vanish. This sign change, combined with the Crandall-Rabinowitz transversality condition (verified through higher-order asymptotic corrections beyond the leading term), produces non-trivial stationary doubly-connected patches. The asymptotic laws 1 - b_{m,λ} ~ β(λ)/m and λ_{m,b} ~ 2/√((1-b²)m log m) characterize the bifurcation loci precisely.","pith_inferences":["The function β(λ) being strictly decreasing with an explicit lower bound suggests a continuous family of stationary patches interpolating between different rotation regimes, parameterized by the Coriolis strength.","The second bifurcation regime (varying λ) could potentially be extended to λ → ∞ by combining the uniform convergence of D_n with large-argument asymptotics of Bessel functions, which the authors note numerically but leave open.","The resummation technique for Taylor expansions of Λ_n — where all orders contribute at the same scale — may be applicable to other bifurcation problems involving special functions with simultaneously large order and argument.","The quantitative bound on patch deformation (Corollary 3.1) approaching the disc as angular velocity approaches the critical value suggests a continuous deformation path from non-trivial to trivial patches, hinting at possible global bifurcation structure beyond the local result."],"forward_implications":["The existence of stationary doubly-connected patches for QGSW but not for Euler (λ=0) shows that the Coriolis force qualitatively changes the solution landscape, creating stationary equilibria inaccessible in the non-rotating limit.","The asymptotic law λ_{m,b} ~ 2/√((1-b²)m log m) means that as symmetry increases, the QGSW model must approach Euler to support stationary patches — the patches live in a narrow corridor between the two models.","The rigidity result for simply-connected patches (stationary ones must be discs) combined with the existence result for doubly-connected patches establishes a sharp topological dichotomy: topology of the domain determines whether non-trivial stationary solutions exist.","The analytic regularity of the bifurcated patch boundaries (via existing results for uniformly rotating solutions) means these stationary patches have smooth, explicitly characterizable geometry near the annulus."],"fun_headline_variants":["Steady vortex rings emerge in quasi-geostrophic shallow water","Bifurcation yields steady vortex patches in shallow-water flows","Nontrivial steady vortex patches for quasi-geostrophic shallow water","Annular vortex patches bifurcate in quasi-geostrophic shallow water","Doubly-connected steady vortices in shallow-water flows"],"cache_read_input_tokens":0,"weakest_assumption_plain":"The transversality condition for the Crandall-Rabinowitz theorem is verified by extracting the first non-vanishing term in an asymptotic expansion of a scalar product involving the derivative of the linearized matrix. If the leading-order asymptotic of this expression were computed incorrectly, or if the O(1/m) remainder terms dominated the leading term, the bifurcation argument would fail.","fun_headline_variants_meta":{"raw":{"variants":["Steady vortex rings emerge in quasi-geostrophic shallow water","Bifurcation yields steady vortex patches in shallow-water flows","Nontrivial steady vortex patches for quasi-geostrophic shallow water","Annular vortex patches bifurcate in quasi-geostrophic shallow water","Doubly-connected steady vortices in shallow-water flows","Steady annular vortices via bifurcation in shallow-water flows"]},"model":"glm-5.2","effort":"high","cost_usd":0.0,"raw_usage":{"total_tokens":1142,"prompt_tokens":503,"completion_tokens":639,"prompt_tokens_details":null},"tokens_in":503,"tokens_out":639,"duration_ms":24271,"temperature":1.0,"reasoning_tokens":539,"cache_read_input_tokens":0,"cache_creation_input_tokens":0},"cache_creation_input_tokens":0},"created_at":"2026-07-09T07:24:27.695587+00:00","model_set":{"reader":"glm-5.2"},"falsifier":"A numerical computation at a specific (λ, m) pair showing that the transversality expression T_{m,λ} or T_{m,b} is zero or has the wrong sign, contradicting the asymptotic positivity/negativity derived in Propositions 2.4(iv) and 2.6(iv).","supporting_citations":[],"review_version":1}