{"id":"eada4226-44f4-451c-8848-8b4adf1c1d5c","arxiv_id":"2507.16474","paper_version":2,"verdict":"CONDITIONAL","confidence":"MODERATE","novelty_score":7.0,"correctness_risk":"medium","formal_verification":"none","parameter_count":0,"one_line_summary":"Finite sums of Lamb dipoles in the half-plane, with ordered speeds and well-separated initial positions, are Lyapunov stable under the 2D Euler equations.","lead":"This paper proves that several Lamb dipoles, pairs of counter-rotating vortices that travel without changing shape, remain stable when placed together in a half-plane, provided faster dipoles start to the right of slower ones. It is an advance in the mathematical theory of fluid motion, showing that multi-vortex structures can persist for all time.","discovery_kind":"extension","skeptic_critique":{"model":"deepseek-v4-flash","headline":"Proposition 5.5 omits the proof of the estimate that closes the energy bootstrap; as written, Theorem A does not follow.","rationale":"I read the paper focusing on whether the bootstrap for Theorem A is actually closed. The weakest formal point in the written proof is Proposition 5.5, where the key bulk interaction estimate is omitted with the justification that it is 'somewhat simpler' than Proposition 5.3. Since this estimate is necessary to close (B3), the proof as written has a genuine gap. The reader's weakest assumption was Proposition 1.3 (the local coercivity of the Lamb dipole variational principle); I agree that this is a foundational dependency, but it is at least quoted from a specific source ([22], reformulating [1]). The omitted estimate in Proposition 5.5 is an internal gap: the paper itself flags that no proof is given. I suspect the estimate is true and fillable, but it is load-bearing. A conditional acceptance is therefore appropriate, requiring the authors to supply the missing argument in a revision. I partly agree with the reader: our concerns overlap, but the reader weighted Proposition 1.3 more heavily, while I see the omitted Prop 5.5 estimate as the immediate obstruction to the proof of Theorem A as written.","tokens_in":35114,"tokens_out":31922,"duration_ms":347057,"concrete_test":"Independently derive the omitted estimate in Proposition 5.5: decompose each ω_i into the central part and the error strips of §3.4.1. For the central contribution, use (2.15) and gap(t) ≥ c(D0 + (V_i−V_j)t) to get a time-integrable O(1/D0) bound. For the error-strip contribution, combine Lemma 4.4 (δ-smallness of errors) with the Lagrangian time-integrability argument of Lemma 5.4 to show the total is o(δ) uniformly in T. If the near-border error contribution can only be bounded pointwise by δ^{3/4} per unit time, without time decay, then the infinite time horizon in Proposition 5.5 cannot be closed and the bootstrap fails.","verdict_should_be":"UNCHANGED","load_bearing_attack":"The bootstrap proof of Theorem A closes the energy assumption (B3) in Proposition 5.5. For 1 ≤ i ≤ N−1, the proof bounds the flux terms, then reduces the remaining task to estimating the time integral of ∫ −∇ψ_i·(u_j ω_i) dx for j ≠ i, and states: 'Bounding this by o(δ) is somewhat simpler than the proof of Proposition 5.3 ... Therefore we omit the proof.' This is the unique place where an essential estimate for closing (B3) is asserted without proof. The asserted simplification is not evident: −∇ψ_i is the self-velocity u_i of the i-th piece, of order V_i on its support, so the integrand is not smaller than the corresponding impulse integrand; the separation bound |u_j| ≲ μ_j/gap(t)^2 only applies away from the border. Near the border, the error parts of ω_i and ω_j have no spatial separation, so their contribution requires an additional time-integrability argument (of the type supplied for the gain–error interaction in Lemma 5.4). Without that estimate, the bootstrap cannot be closed and Lemma 4.1 cannot be applied globally in time, so Theorem A is not established as written.","agreement_with_reader":"partial"},"referee_report":{"model":"deepseek-v4-flash","summary":"This paper proves Theorem A: for a fixed ordered family of Lamb dipoles with speeds κ1 > κ2 > ... > κN in the upper half-plane, sufficiently separated initial positions, and initial data close in L2 ∩ L1* to their superposition, the Euler evolution remains ε-close to a translated superposition for all positive times, with the shifts satisfying |p_i(t) - p̄_i - V̄_i t| ≤ C ε^{1/2}(1+t). The proof decomposes the solution using moving cutoffs, imposes bootstrap assumptions (B1)-(B3) on cumulative enstrophy, cumulative impulse, and individual energies, derives consequences (approximation by Lamb dipoles, smallness of gains and errors, smallness of interaction energy), and then closes the bootstrap via flux estimates and a Lagrangian estimate for gain-error interactions. The paper also states Theorem B, which separates out a Lamb dipole from a slower and less structured background, and sketches its proof by parallel bootstrap assumptions.","tokens_in":35359,"tokens_out":6627,"duration_ms":73046,"significance":"If fully established, the result is a substantial advance: it provides the first multi-soliton-type Lyapunov stability theorem for the 2D Euler equation in the half-plane, with quantitative orbital stability and shift estimates. The paper contains genuinely useful new ingredients, especially the interaction energy bounds of Proposition 2.4, the use of rearrangements in Theorem B, and the Lagrangian treatment of gain-error interactions in Lemma 5.4. The overall architecture of the proof is clear and the bootstrap is carefully arranged. However, the current manuscript contains an explicit omitted proof of a load-bearing estimate in Proposition 5.5 and only a sketched closing argument for Theorem B; these gaps prevent the results from being accepted as fully proved at present.","major_comments":[{"comment":"The proof of Proposition 5.5 does not establish a bound for the interaction term ∫_0^T ∫ −∇ψ_i·(u_j ω_i) dx dt, which appears in the energy formula (5.5) for i ≠ j. The manuscript states that bounding this term by o(δ) is 'somewhat simpler than the proof of Proposition 5.3' and then omits the proof. This term is essential for closing the energy bootstrap (B3), and Lemma 4.1, and hence Theorem A, depends on (B3) holding globally in time. The asserted additional smallness from the factor −∇ψ_i is not demonstrated: |∇ψ_i| is of order one on the support of ω_i, and near the border L_i(t) the gain and error parts can be arbitrarily close in space, so a separate time-integrability argument of the type used for Lemma 5.4 is required. Without this estimate, the bootstrap cannot be closed and Theorem A is not established as written.","section":"Section 5.4, Proposition 5.5"},{"comment":"Theorem B is stated as a theorem, but its proof closes the bootstrap assumptions (B2') and (B3') with a single sentence saying that the arguments are 'almost identical' to the rightmost dipole case of Theorem A. That case itself relies on the omitted estimate from Proposition 5.5, so the asserted parallel is not sufficient. In particular, the analogue of Lemma 5.4 for the slanted border L(t, x2) is not written out, and the energy closing for ω_r is not verified. The authors should either provide the complete proof of these closing steps or explicitly state Theorem B as a conditional result.","section":"Section 6.3, Closing (B2') and (B3')"},{"comment":"The proof of Theorem A relies crucially on Proposition 1.3, the coercive version of the variational characterization of a single Lamb dipole, whose proof is deferred to the preprint [22] by Choi, Jeong, and Yao. This proposition is used for each of the N pieces to obtain the approximation in Lemma 4.1. Because it is a load-bearing external input, the authors should either include a self-contained proof of Proposition 1.3 or a precise derivation from the published results in [1], and clarify whether [22] is available in refereed form.","section":"Section 1.2.1, Proposition 1.3"}],"minor_comments":[{"comment":"The proof of Lemma 4.4 contains the line 'using ∥ω_i,err∥_{L2} ≤ ∥ω_i,err∥_{L2} ≤ ε0', which is tautological and appears to be a typo; the intended bound is likely ∥ω_i,err∥_{L2} ≤ ∥ω_i,rem∥_{L2} ≤ ε0.","section":"Section 4.4, Lemma 4.4"},{"comment":"The notation 'o(δ)' in the proof of Proposition 5.5 is not quantified, although the bootstrap argument requires a specific improvement from δ to δ/100. The omitted estimate should be stated with an explicit power of δ, e.g. C δ^{1+α}, so that it can be absorbed for δ sufficiently small.","section":"Section 5.4, Proposition 5.5"},{"comment":"The Lagrangian decomposition ω≥i = ω≥i,∗ + ω≥i,gain is introduced before the flow map notation is fully explained; the definitions of Φ^{-1} and the statement that particles cross each L_i(t) only from right to left would be clearer if the relevant crossing lemma were stated immediately after the decomposition.","section":"Section 3.4.2"},{"comment":"The proof of Theorem 2.7 says that the sharp constant in the energy inequality is obtained by plugging in a Lamb dipole, which is correct only after the variational principle from [1] is invoked; the phrasing could be made more explicit that the equality case in (2.18) follows from the uniqueness part of the first statement.","section":"Section 2.3, Theorem 2.7"}],"recommendation":"major_revision","confidential_remarks":"The central theorem is likely correct and the strategy is promising, but the omitted estimate in Proposition 5.5 is a genuine load-bearing gap: without it the energy bootstrap does not close and Lemma 4.1 cannot be applied globally in time. I would not accept the paper until that estimate is supplied and the closing steps for Theorem B are written out. The dependence on the unpublished preprint [22] for Proposition 1.3 should also be clarified in the final version."},"author_rebuttal":null,"desk_editor":{"model":"deepseek-v4-flash","letter":"The paper proves Lyapunov stability for finite sums of Lamb dipoles with ordered speeds, a first multi-soliton result for 2D Euler. That is new and significant. The strategy is a bootstrap on enstrophy, impulse, and energy, with a Lagrangian analysis of gain-error interactions. The authors clearly explain the difficulties, the role of the ordering condition, and the contrast with point-vortex dynamics. The variational input from Abe-Choi is used correctly, and the decomposition into N pieces is natural. Lemma 5.4, the gain-error interaction estimate, is detailed and convincing; it is the heart of the paper. The interaction energy bounds in Proposition 2.4 are also useful and appear new.\n\nThat said, there is a genuine gap. Proposition 5.5, which closes the energy bootstrap (B3), needs a bound on the time integral of ∫ -∇ψ_i·(u_j ω_i) for j≠i. The paper says this is 'somewhat simpler than the proof of Proposition 5.3' and omits the proof. This is not a side detail: without it, Theorem A is not established as written. The stress-test concern is valid. Near the border, the error parts of ω_i and ω_j are not spatially separated, so the simplification is not immediate. The authors need to supply the argument or a clear reference. The proof of Theorem B is also sketched, with the closing of (B2') and (B3') deferred as parallel; given the slanted border and Vmax condition, the sketch should be expanded if the paper is to be accepted.\n\nThe reliance on prior single-dipole results is transparent and not circular; those results are published. The main theorem appears true, and the omission is likely fillable. But as written, the proof has a load-bearing gap.\n\nThis paper is for anyone working on stability of coherent structures in 2D Euler, and more generally on multi-soliton stability in fluid equations. It deserves a serious referee; the editor should send it out with a specific request to fill the omitted estimate in Proposition 5.5. Conditional acceptance is the right call.","headline":"A substantial multi-dipole stability theorem whose proof has a real gap: the energy bootstrap in Proposition 5.5 omits an essential estimate.","tokens_in":35888,"tokens_out":2422,"would_cite":true,"duration_ms":28145,"reading_group":"yes","serious_thinker":"yes","would_accept_peer_review":true},"rs_alignment":null,"lean_confirmation":null,"pith_extraction":{"msc":["76B47","35Q35","35B40"],"pacs":[],"model":"deepseek-v4-flash","headline":"A superposition of Lamb dipoles is Lyapunov stable for all time, provided the faster dipoles start to the right of the slower ones and the initial gaps are large enough.","keywords":["Lamb dipole","vortex stability","Euler equations","upper half-plane","variational principle","Lagrangian bootstrapping","multi-soliton stability","energy maximizer"],"falsifier":"A direct numerical simulation of two Lamb dipoles with ordered speeds and initial separation $D_0$ large, tracking the bootstrap quantities: if the enstrophy of the left piece $K_{\\le 1}(t)$ or its impulse $\\mu_{\\le 1}(t)$ ever increases by more than $C\\delta_0$ relative to its initial value, or if the $L^2\\cap L^1_*$ distance from $\\omega(t)$ to the best shifted two-dipole superposition exceeds $\\varepsilon$ for some $t>0$ while the initial data satisfy the hypotheses of Theorem A, the theorem would be false.","tokens_in":34916,"feed_emoji":"🌀","tokens_out":10302,"duration_ms":90633,"temperature":0.7,"pith_summary":"This paper proves that a finite collection of Lamb dipoles—self-propelled, compactly supported vortex pairs that translate without changing shape—is stable in the upper half-plane, provided the faster dipoles are initially to the right of the slower ones and the gaps between them are large enough. The result is Lyapunov stability in the $L^2\\cap L^1_*$ norm: initial data close to such a superposition stays $\\varepsilon$-close to a shifted superposition for all time, with each center tracking its linear motion up to an error of order $\\varepsilon^{1/2}(1+t)$. This matters because 2D Euler generically creates filaments and exchanges enstrophy, impulse, and energy between nearby structures, and the proof shows that speed ordering plus separation makes these exchanges too weak to break the configuration. A second theorem shows that a fast Lamb dipole can be separated out from an arbitrary slower left-side vorticity without detailed information about it.","feed_headline":"Ordered Lamb dipoles are stable for all time","feed_subtitle":"Faster dipoles placed ahead of slower ones keep a multi-vortex configuration close to its shifted sum forever.","key_machinery":"The load-bearing object is the Lamb dipole, defined by $\\omega_{\\mathrm{Lamb}}(x)=-\\frac{2c_L}{J_0(c_L)}J_1(c_L r)\\mathbf{1}_{[0,1]}(r)\\sin\\theta$ with $c_L$ the first positive zero of $J_1$; it is the unique maximizer of kinetic energy among nonnegative vorticities with $L^2$ norm bounded by $\\kappa$ and weighted $L^1$ impulse bounded by $\\mu$, up to horizontal translations. The proof uses a quantitative, local version of this maximizer statement (Proposition 1.3): if a vorticity has near-maximal energy and only slightly relaxed $L^2$ and impulse bounds, it must be close to a translate of the Lamb dipole. Around this, the authors build a decomposition of the solution into $N$ pieces separated by moving vertical borders placed at the midpoints of expected dipole centers, and a bootstrap that keeps each piece inside the coercive regime of Proposition 1.3. The most delicate element is the 'Gain–Error interaction' estimate (Lemma 5.4), which controls the exchange of impulse between vorticity that has crossed a border from right to left and the error layer just right of that border; because the borders separate linearly in time, each gained particle contributes significant impulse exchange only for a short interval, which is tracked in Lagrangian coordinates.","core_discovery":"The central claim is Theorem A: for any $N\\ge 1$ and any strictly ordered speeds (equivalently, strictly ordered $L^2$ norms) $\\kappa_1>\\kappa_2>\\cdots>\\kappa_N$, there exist thresholds $\\delta_0$ and $D_0$ such that every nonnegative initial vorticity that is $\\delta_0$-close in $L^2\\cap L^1_*$ to a superposition of $N$ Lamb dipoles with centers separated by more than $D_0$, and with $L^1\\cap L^\\infty$ and support-area bounds, evolves so that the solution remains $\\varepsilon$-close in $L^2\\cap L^1_*$ to a superposition of the same dipoles for all time, with centers $p_i(t)$ satisfying $|p_i(t)-\\bar p_i-\\bar V_i t|\\le C\\varepsilon^{1/2}(1+t)$. The dipoles need not have ordered amplitudes or radii; only their traveling speeds are ordered. The proof treats the case $N=1$ as known and reduces the general case to a bootstrap on the enstrophy, impulse, and energy of each of the $N$ pieces, using the variational characterization of a single Lamb dipole as the unique energy maximizer in its admissible class.","pith_inferences":["The proof suggests a sharp ordering rule for 2D Euler in the half-plane: a separated vortex configuration persists exactly when the spatial order matches the order by speed; the paper's remark that reversing the order breaks stability indicates this is a genuine threshold phenomenon.","The Lagrangian gain–error estimate may extend to countably many dipoles with summable $L^1$ norm, where the limiting measure could accumulate Dirac masses at the origin; the paper explicitly leaves this as a challenging open direction.","The sharp energy inequality $E[\\omega]\\le (1/\\sqrt{\\pi c_L})\\|\\omega\\|_{L^1_*}\\|\\omega\\|_{L^2}$, with equality only at Lamb dipoles, is an isoperimetric statement for the half-plane interaction energy that could be of independent use in rearrangement problems.","Because the bootstrap uses only $L^2$, $L^1_*$, and support-area bounds, replacing $L^2$ by $L^p$ with $p>1$ through the energy inequality of Remark 2.6 may yield stability for other orbitally stable traveling waves under the same ordering and separation hypotheses."],"forward_implications":["Theorem A gives the first Lyapunov stability result for multi-dipole solutions of 2D Euler in the half-plane, with a quantitative drift estimate for the centers.","The bootstrap shows that linear separation between ordered dipoles dominates the time-integrated effects of filamentation and lift-up, so the configuration remains coherent for all time.","As a corollary, the rescaled vorticity $t^2\\omega(t,tx)$ converges in the sense of measures to a finite sum of Dirac masses with explicitly predicted positions and strengths, giving concrete realizations of the general scattering picture for the half-plane.","The same strategy yields Theorem B, a separation result for a single fast Lamb dipole in front of an arbitrary slower vorticity; this requires no detailed information about the slower part beyond a rearrangement-velocity bound."],"supporting_citations":[{"why":"Supplies the variational characterization of a single Lamb dipole as the unique energy maximizer in the admissible class, the foundation of the coercivity used on each piece.","marker":"[1]"},{"why":"Supplies the local coercivity version (Proposition 2.6 there) stated here as Proposition 1.3, which turns near-maximal energy and relaxed bounds into $L^2\\cap L^1_*$ closeness to a translate.","marker":"[22]"},{"why":"Supplies the method for constructing the smooth function $H$ and the inductive shift estimate that defines $p_i(t)$ and controls its derivative.","marker":"[47]"},{"why":"Provides the point-vortex analogue of interaction velocity decay and linear separation, and the scattering framework for the rescaled vorticity that Theorem A realizes.","marker":"[42]"},{"why":"Provides the Yudovich global well-posedness theory used to define the solution for $L^1\\cap L^\\infty$ initial data.","marker":"[63]"}],"fun_headline_variants":["Lamb dipole sums stay stable if ordered and spaced","Faster dipoles ahead guarantee vortex stability","Ordered Lamb dipoles remain stable forever","Vortex sums of Lamb dipoles are Lyapunov stable","Spacing and speed order lock in Lamb dipole stability"],"cache_read_input_tokens":3200,"weakest_assumption_plain":"The proof depends on the local coercivity of the Lamb dipole as an energy maximizer: if a vorticity with near-maximal energy and only slightly relaxed $L^2$ and impulse bounds could be far from every translate of the Lamb dipole, the bootstrap for each piece would fail and the stability conclusion would not follow.","fun_headline_variants_meta":{"raw":{"variants":["Lamb dipole sums stay stable if ordered and spaced","Faster dipoles ahead guarantee vortex stability","Ordered Lamb dipoles remain stable forever","Vortex sums of Lamb dipoles are Lyapunov stable","Spacing and speed order lock in Lamb dipole stability"]},"model":"deepseek-v4-flash","effort":"low","cost_usd":0.000672,"raw_usage":{"total_tokens":3035,"prompt_tokens":892,"completion_tokens":2143,"prompt_tokens_details":{"cached_tokens":384},"prompt_cache_hit_tokens":384,"prompt_cache_miss_tokens":508,"completion_tokens_details":{"reasoning_tokens":2069}},"tokens_in":508,"tokens_out":2143,"duration_ms":15460,"temperature":1.0,"reasoning_tokens":2069,"cache_read_input_tokens":384,"cache_creation_input_tokens":0},"cache_creation_input_tokens":0},"created_at":"2026-08-06T15:08:16.796147+00:00","model_set":{"reader":"deepseek-v4-flash"},"falsifier":"A direct numerical simulation of two Lamb dipoles with ordered speeds and initial separation $D_0$ large, tracking the bootstrap quantities: if the enstrophy of the left piece $K_{\\le 1}(t)$ or its impulse $\\mu_{\\le 1}(t)$ ever increases by more than $C\\delta_0$ relative to its initial value, or if the $L^2\\cap L^1_*$ distance from $\\omega(t)$ to the best shifted two-dipole superposition exceeds $\\varepsilon$ for some $t>0$ while the initial data satisfy the hypotheses of Theorem A, the theorem would be false.","supporting_citations":[{"cited_title":"Abe and K","cited_arxiv_id":null,"evidence_quote":"Supplies the variational characterization of a single Lamb dipole as the unique energy maximizer in the admissible class, the foundation of the coercivity used on each piece."},{"cited_title":"Iftimie, M","cited_arxiv_id":null,"evidence_quote":"Provides the point-vortex analogue of interaction velocity decay and linear separation, and the scattering framework for the rescaled vorticity that Theorem A realizes."},{"cited_title":null,"cited_arxiv_id":null,"evidence_quote":"Provides the Yudovich global well-posedness theory used to define the solution for $L^1\\cap L^\\infty$ initial data."}],"review_version":1}