{"id":"ac665e58-91d8-4a4f-aeff-cc339128411b","arxiv_id":"2507.10390","paper_version":1,"verdict":"CONDITIONAL","confidence":"MODERATE","novelty_score":7.0,"correctness_risk":"medium","formal_verification":"none","parameter_count":0,"one_line_summary":"Small-amplitude plane internal waves in the 2D inviscid Boussinesq system are proven to be spectrally unstable via a rigorous Floquet-Bloch analysis.","lead":"This paper proves that small waves traveling along density layers inside a fluid are unstable: a tiny disturbance can grow. It gives the first rigorous mathematical confirmation of a long-studied ocean wave splitting mechanism.","discovery_kind":"first_principles","skeptic_critique":{"model":"deepseek-v4-flash","headline":"Equations (3.18)-(3.19) in Proposition 3.3 carry an extra factor ϵ that makes the off-diagonal entries of the 2x2 matrix O(ϵ^2), invalidating the claimed ±ϵ√e(µ) splitting in Theorem 2 as written.","rationale":"The paper's central claim is Theorem 1, which rests on Theorem 2's explicit eigenvalue splitting. The reader's L^2-vs-Bloch concern is real and should be fixed by rewording, but it does not block the mathematical construction of a growing Bloch wave. The extra ϵ in (3.18)-(3.19) is a more immediate obstacle: taken literally, the perturbation matrix becomes diagonal to O(ϵ^2), and the O(ϵ) splitting that powers the whole instability argument disappears. The surrounding computation clearly shows the ϵ should cancel, and (2.36) confirms the intended O(1) values, so the issue is repairable rather than fatal. I therefore maintain the reader's CONDITIONAL verdict, with the additional explicit requirement that the authors correct (3.18)-(3.19). I also note the introduction's claim that e(µ) is strictly positive in both regimes conflicts with (2.40) for the negative-τ branch, another minor correction. Because the central mathematical result survives both fixes, the verdict should remain conditional on revision.","tokens_in":27013,"tokens_out":16907,"duration_ms":193831,"concrete_test":"Recompute β1−γ1w from the final line of Section 5: starting from the identity (β1−γ1w)ϵ = ϵ/8 (k⊥·µ)(|k+µ|−|k|)(|k+µ|+|k|−|µ|)(N²+|µ|²)/(N|k||µ||k+µ|), divide both sides by ϵ and substitute the result into b1(µ)=(β1−γ1w)/ι00(µ,0). If the resulting b1 equals the expression in (2.36), the ϵ in (3.18) is a typo and the proof is repairable; if it does not, the off-diagonal scaling is O(ϵ^2) and Theorem 2's eigenvalue splitting is unsupported.","verdict_should_be":"UNCHANGED","load_bearing_attack":"In Proposition 3.3, the displayed formulas (3.18) and (3.19) state β1(µ)−γ1(µ)w(µ) = ϵ/8 · ... and β0(µ)−γ1(µ)w(µ) = −ϵ/8 · ... . But β1, β0, γ1 are defined in (3.16) as ϵ-independent Taylor coefficients, and Section 5's derivation computes (β1−γ1w)ϵ = ϵ/8 · ..., so the ϵ belongs on the left-hand side and cancels. As printed, the ratios b1(µ)=(β1−γ1w)/ι00 and b0(µ)=(β0−γ1w)/ι11 used in the proof of Theorem 2 become O(ϵ), so the off-diagonal entries ib1ϵ, ib0ϵ in (3.20) are O(ϵ^2) rather than O(ϵ). The eigenvalue formula λ± = λ+ + iO(ϵ^2) ± ϵ√e(µ)+O(ϵ) in (2.37) would then not follow. The final expressions for b1,b0 in (2.36) are O(1) and match the corrected identities, confirming that (3.18)-(3.19) contain a typo; nevertheless, a reader verifying the proof cannot reproduce Theorem 2 without repairing this step.","agreement_with_reader":"disagree"},"referee_report":{"model":"deepseek-v4-flash","summary":"The paper studies the two-dimensional inviscid Boussinesq system and proves a modulational instability result for small-amplitude internal plane waves. The authors linearize around an exact plane-wave solution, use Floquet-Bloch decomposition, restrict to the invariant subspace of harmonics of the primary wavevector, and show that for resonant Floquet parameters a purely imaginary double eigenvalue of the unperturbed operator splits into a pair with nonzero real part. The growth rate is computed explicitly as ϵ√e(µ) with an explicit function e(µ), and its small- and large-Floquet-parameter asymptotics are given. The paper also compares the resulting rate with the physical literature on parametric subharmonic instability and includes a Mathematica code for the algebraic computations.","tokens_in":27246,"tokens_out":10546,"duration_ms":129729,"significance":"If the result is correct, this is the first rigorous treatment of inviscid parametric subharmonic instability for internal gravity waves, a mechanism that is widely used in oceanography and experimental fluid mechanics. The explicit, parameter-free formula for the first-order growth rate and the verification of both small- and large-Floquet-parameter regimes are concrete and falsifiable predictions. The analytic setup, including the characterization of the resonant set in Lemma 2.5 and the spectral-isolation argument in Proposition 2.7, is a nontrivial adaptation of recent Stokes-wave techniques. The availability of machine-checkable code for the entanglement coefficients is a further strength. However, as detailed below, two load-bearing points need correction: a displayed algebra error in Proposition 3.3 that temporarily breaks the proof of Theorem 2, and an overstatement in Theorem 1 about the nature of the instability (spectral versus L²-eigenvalue instability).","major_comments":[{"comment":"Equations (3.18) and (3.19) contain an extra factor ϵ on the right-hand side. The coefficients β1, β0, γ1 are ϵ-independent Taylor coefficients defined in (3.16), and the derivation in Section 5 (the displayed chain ending with 'proving formula (3.18)') actually computes (β1−γ1w)ϵ = ϵ/8·..., i.e. β1−γ1w = 1/8·... . As printed, the quantities b1 and b0 defined in the proof of Theorem 2 become O(ϵ), the off-diagonal entries in (3.20) become O(ϵ²), and the eigenvalue splitting ±ϵ√e(µ) in (2.37) does not follow. The final formulas (2.36) confirm that the intended identities are O(1). This is a local but load-bearing typo; it must be corrected and the surrounding computation reconciled.","section":"Proposition 3.3, Eqs. (3.18)-(3.19)"},{"comment":"Theorem 1 claims that the linearized operator Lϵ on L²(R²) has an unstable eigenvalue. What is actually constructed is a Bloch eigenvalue of L_{µ,ϵ} on L²(T²); the corresponding spacetime function h(t,x)=e^{λt}e^{iµx}v(x) is not square-integrable over R², as the text itself notes after (2.6)-(2.7). Thus the result establishes spectral instability (Re σ(L_{µ,ϵ})>0 for some µ, equivalently the spectrum of Lϵ intersects the right half-plane) rather than existence of an L² eigenfunction. This is a standard accepted notion in modulational-instability theory, but Theorem 1 and the abstract should state this explicitly so that the claim is not overread as an L²-instability result.","section":"Theorem 1 and Eqs. (2.6)-(2.7)"},{"comment":"Lemma 3.1 is the main perturbative tool and is stated without proof. Because L0 is unbounded and the introduction emphasizes that the perturbation is not bounded, the analyticity of the spectral projectors P_{µ,ϵ} and the existence of the contour integral for small ϵ require justification. On the invariant subspace H¹_k this is straightforward: in view of (4.15), the jets L^{±k}_1 act on modes nk through the constants k⊥·µ, so L1 is actually bounded on H¹_k. Please add a proof or a precise Kato-type statement verifying the hypotheses, so that the reduction to the 2×2 matrix representation is fully justified.","section":"Lemma 3.1"}],"minor_comments":[{"comment":"The reversibility identities in (4.19) appear to be missing complex conjugation or an overline; as typeset, the first identity is tautological. Please check the intended relations.","section":"Lemma 4.6, Eq. (4.19)"},{"comment":"For ℓ=−1 the leading coefficient displayed after (B.7) contains the factor m²−2n²; a parenthetical noting that this factor is nonzero for nonzero integers would remove an apparent edge case.","section":"Appendix B, Step 2 near 0"},{"comment":"The notation r(ϵⁿ) is defined in the introduction, but in (2.37) the expression ±ϵ√e(µ)+r(ϵ) could be confused with the earlier real-valued remainders r(ϵ²); a brief reminder or a different symbol would improve readability.","section":"General notation"}],"recommendation":"major_revision","confidential_remarks":"The paper is a serious contribution and the final formulas in Section 5 and Theorem 2 are mutually consistent once the typo in Proposition 3.3 is repaired. The spectral-instability framing issue in Theorem 1 is a matter of precise statement rather than a fatal flaw, but it should be fixed before publication. I see no reason to doubt the novelty or the fit with the journal's scope."},"author_rebuttal":null,"desk_editor":{"model":"deepseek-v4-flash","letter":"This paper is worth reading: it gives the first rigorous proof of PSI-type instability for inviscid internal waves in the 2D Boussinesq system, and the proof strategy is genuinely novel. The main theorem establishes a positive-real-part spectral point for the linearization around a small-amplitude plane wave, with an explicit first-order growth rate. The resonance-set analysis (Lemma 2.5) and the spectral-isolation argument (Prop 2.7) are careful, and the comparison with the physical literature in Section 6 reads as a sensible consistency check, not a fitted input. The paper also ships a Mathematica script for the entanglement coefficients, which helps reproducibility.\n\nThe soft spots are mostly presentation-level, but one sits in a load-bearing place. In Proposition 3.3, the displayed formulas (3.18)-(3.19) carry an extra factor of epsilon on the right-hand side. The Section 5 derivation actually computes (beta1 - gamma1 w) epsilon, which cancels to an O(1) numerator; as printed, b1 and b0 become O(epsilon), so the off-diagonal entries in (3.20) are O(epsilon^2) and the eigenvalue splitting in (2.37) does not follow. This is clearly a typo — the final expressions for b1,b0 in (2.36) are O(1) and match the corrected identities — but a referee will need the authors to fix it. Also, the introduction claims e(mu) is strictly positive in both regimes |mu|<<1 and |mu|>>1, yet (2.40) is negative on the phi- branch near zero; the claim should be restricted to the phi+ branch, which is enough for the theorem. On the functional-analytic side, Theorem 1 states an eigenvalue of L_epsilon when the proof actually produces a Floquet–Bloch eigenvalue with a non-L^2 Bloch wave; this is standard in the modulational-instability literature, but it should be stated explicitly to avoid overclaiming.\n\nNone of these affect the central result: the existence of a positive-real-part spectral point and the explicit growth rate stand. The paper is a real contribution and deserves a serious referee. I would send it to review, and I would cite it once the typos are corrected.","headline":"First rigorous PSI-type instability for inviscid internal waves, with a proof that mostly holds up; the main caveat is a typo in Proposition 3.3 that needs fixing.","tokens_in":27850,"tokens_out":3561,"would_cite":true,"duration_ms":35529,"reading_group":"yes","serious_thinker":"yes","would_accept_peer_review":true},"rs_alignment":null,"lean_confirmation":null,"pith_extraction":{"msc":["35Q35","76E20","76B70"],"pacs":[],"model":"deepseek-v4-flash","headline":"The paper proves that small-amplitude internal gravity waves in an inviscid stably stratified Boussinesq fluid are linearly unstable, with an explicit linear-in-amplitude growth rate.","keywords":["Boussinesq equations","internal gravity waves","spectral instability","modulational instability","parametric subharmonic instability","Floquet-Bloch decomposition","inviscid stratified fluid"],"falsifier":"Direct numerical diagonalization of the restricted operator $L_{\\mu,\\epsilon}|_{H^1_k}$ for a concrete case, say $k=(1,1)$, small $\\epsilon$, and $\\mu$ on the resonant branch $R_k$, should reproduce an unstable eigenvalue whose real part equals $\\epsilon\\sqrt{e(\\mu)}$ at leading order; finding no such eigenvalue, or a leading rate that disagrees, would refute the theorem.","tokens_in":26747,"feed_emoji":"🌊","tokens_out":6574,"duration_ms":64620,"temperature":0.7,"pith_summary":"This paper proves that small-amplitude internal gravity waves in a two-dimensional inviscid, stably stratified Boussinesq fluid are linearly unstable: the linearization about such a wave has an eigenvalue with strictly positive real part. This is the first rigorous derivation of the Parametric Subharmonic Instability (PSI) in the inviscid setting, in which an initially excited primary wave transfers energy to two subharmonic waves of lower frequency. The result matters for oceanography because viscous effects are often negligible there, so the inviscid instability is the physically relevant one. The proof reduces the problem to a two-by-two matrix whose eigenvalues have an explicit first-order growth rate, linear in the wave amplitude.","feed_headline":"Inviscid internal gravity waves are provably unstable","feed_subtitle":"A rigorous proof shows a small primary wave seeds two subharmonic waves: the ocean's parametric subharmonic instability.","key_machinery":"The central object is the resonant set $R_k$, the curve of Floquet parameters $\\mu$ for which the unperturbed linearized operator $L_{\\mu,0}$ has a double eigenvalue $\\lambda_+ = \\lambda^-_k(\\mu) = \\lambda^+_0(\\mu)$ on the invariant subspace $H^1_k$ of functions whose Fourier coefficients are supported on wavevectors proportional to the primary wavevector $k$. Kato's similarity transformation and spectral projectors show that this double eigenvalue persists as an isolated pair for $\\epsilon>0$, reducing the problem to the eigenvalues of a $2\\times 2$ matrix. The product of the off-diagonal entries of that matrix yields the explicit growth-rate function $e(\\mu)$ in equation (2.38), whose sign controls the instability.","core_discovery":"The paper's central claim is that for every wavevector $k=(m,n)$ with $m,n>0$ and every sufficiently small amplitude $\\epsilon$, the operator obtained by linearizing the two-dimensional inviscid Boussinesq equations in vorticity-stream form around the plane wave solution has at least one eigenvalue with strictly positive real part. The sharper statement is that, restricting to the invariant subspace of functions supported on wavevectors proportional to $k$ and choosing the Floquet parameter $\\mu$ on the resonant set defined by $\\Omega(k)-\\Omega(k+\\mu)=\\Omega(\\mu)$, the perturbed eigenvalues take the form $\\lambda_+ + iO(\\epsilon^2) \\pm \\epsilon\\sqrt{e(\\mu)} + O(\\epsilon)$, where $e(\\mu)$ is given explicitly and is positive in the small-$|\\mu|$ and large-$|\\mu|$ limits. Because $e(\\mu)>0$, the real part of one of the eigenvalues is positive, an instability whose leading growth rate is linear in the amplitude $\\epsilon$.","pith_inferences":["One testable extension is to compute the spectrum of the truncated operator on $H^1_k$ numerically for a fixed $k$ and small $\\epsilon$, and check that the leading growth rate matches $\\epsilon\\sqrt{e(\\mu)}$; agreement to order $\\epsilon$ would confirm the sharpness of the formula.","Because the unstable Bloch mode is not square-integrable over $\\mathbb{R}^2$, the established instability is spectral; whether nonlinear effects convert it into actual norm growth for localized data remains an open question that the paper does not address.","The structure of the proof suggests that the resonance condition, rather than the specific form of $\\Omega$, drives the instability, so an analogous result may hold for more general stable density profiles with a homogeneous dispersion relation.","The relation in Proposition 6.1 between $e(\\mu)$ and the physical interaction coefficients $I_\\pm(\\mu)$ indicates that the rigorous growth rate coincides with the weakly-nonlinear triad expansion in the small-Floquet limit, which could be used to calibrate reduced models of internal-wave energy transfer."],"forward_implications":["If the theorem is right, every sufficiently small-amplitude internal plane wave in the inviscid two-dimensional Boussinesq system is modulationally unstable, with growth rate proportional to amplitude.","The explicit formula for $e(\\mu)$ makes it possible to predict, for a given primary wave vector $k$, which resonant subharmonic pairs grow fastest.","The instability persists all along each resonant branch, from small to large Floquet parameters, as the asymptotic formulas (2.40) and (2.41) show positivity in both limits.","The method extends the rigorous modulational-instability toolkit to a non-Hamiltonian, reversible system with an unbounded perturbation, where classical perturbation theory requiring boundedness does not apply.","The result rigorously confirms the physical PSI picture: a primary wave of frequency $\\Omega(k)$ transfers energy to two secondary waves whose frequencies are, to leading order, half of $\\Omega(k)$."],"supporting_citations":[{"why":"The prior viscous construction of approximate eigenvalues (quasi-modes) that the present inviscid analysis extends and contrasts with.","marker":"[15]"},{"why":"The physical description of PSI and TRI, including the resonance conditions and the oceanographic motivation, as well as the notation used in Section 6.","marker":"[23]"},{"why":"The source of the spectral-projector and Kato similarity-transformation approach that the paper adapts to the Boussinesq setting.","marker":"[10]"},{"why":"The origin of the entanglement-coefficient machinery generalized in Section 4 to compute the Taylor expansion of the reduced operator.","marker":"[9]"},{"why":"The standard reference for Kato's similarity transformation on which Lemma 3.1, the isolation of the perturbed spectral subspace, is based.","marker":"[33]"},{"why":"Used for the property that Bloch conjugation shifts the symbol of a pseudodifferential operator, which underlies the derivation of $L_{\\mu,\\epsilon}$.","marker":"[39]"},{"why":"Supplies the vorticity-stream Hamiltonian formulation of the Boussinesq system used to write the equations in the form (1.4).","marker":"[1]"}],"fun_headline_variants":["Proof: inviscid internal waves go unstable via PSI","Rigorous proof of parametric subharmonic instability","Math proof shows internal waves split into subharmonics","First rigorous proof of ocean wave instability"],"cache_read_input_tokens":3200,"weakest_assumption_plain":"The linear instability is a spectral point of the Floquet-Bloch operator on the torus, and the growing solution $e^{\\lambda t} e^{i\\mu x} v(x)$ is not square-integrable over $\\mathbb{R}^2$, so the theorem does not provide an $L^2$ eigenfunction of the operator on the whole plane.","fun_headline_variants_meta":{"raw":{"variants":["Proof: inviscid internal waves go unstable via PSI","Rigorous proof of parametric subharmonic instability","Math proof shows internal waves split into subharmonics","First rigorous proof of ocean wave instability"]},"model":"deepseek-v4-flash","effort":"low","cost_usd":0.000218,"raw_usage":{"total_tokens":1452,"prompt_tokens":967,"completion_tokens":485,"prompt_tokens_details":{"cached_tokens":384},"prompt_cache_hit_tokens":384,"prompt_cache_miss_tokens":583,"completion_tokens_details":{"reasoning_tokens":424}},"tokens_in":583,"tokens_out":485,"duration_ms":6727,"temperature":1.0,"reasoning_tokens":424,"cache_read_input_tokens":384,"cache_creation_input_tokens":0},"cache_creation_input_tokens":0},"created_at":"2026-08-06T17:34:12.182416+00:00","model_set":{"reader":"deepseek-v4-flash"},"falsifier":"Direct numerical diagonalization of the restricted operator $L_{\\mu,\\epsilon}|_{H^1_k}$ for a concrete case, say $k=(1,1)$, small $\\epsilon$, and $\\mu$ on the resonant branch $R_k$, should reproduce an unstable eigenvalue whose real part equals $\\epsilon\\sqrt{e(\\mu)}$ at leading order; finding no such eigenvalue, or a leading rate that disagrees, would refute the theorem.","supporting_citations":[{"cited_title":"Bianchini and T","cited_arxiv_id":null,"evidence_quote":"The prior viscous construction of approximate eigenvalues (quasi-modes) that the present inviscid analysis extends and contrasts with."},{"cited_title":"Dauxois, S","cited_arxiv_id":null,"evidence_quote":"The physical description of PSI and TRI, including the resonance conditions and the oceanographic motivation, as well as the notation used in Section 6."},{"cited_title":"Berti, A","cited_arxiv_id":null,"evidence_quote":"The source of the spectral-projector and Kato similarity-transformation approach that the paper adapts to the Boussinesq setting."},{"cited_title":"Berti, L","cited_arxiv_id":null,"evidence_quote":"The origin of the entanglement-coefficient machinery generalized in Section 4 to compute the Taylor expansion of the reduced operator."},{"cited_title":"Kato.Perturbation theory for linear operators","cited_arxiv_id":null,"evidence_quote":"The standard reference for Kato's similarity transformation on which Lemma 3.1, the isolation of the perturbed spectral subspace, is based."},{"cited_title":null,"cited_arxiv_id":null,"evidence_quote":"Used for the property that Bloch conjugation shifts the symbol of a pseudodifferential operator, which underlies the derivation of $L_{\\mu,\\epsilon}$."},{"cited_title":null,"cited_arxiv_id":null,"evidence_quote":"Supplies the vorticity-stream Hamiltonian formulation of the Boussinesq system used to write the equations in the form (1.4)."}],"review_version":1}