{"id":"d56b6a2f-2bc9-4294-8cac-3103a9c306ff","arxiv_id":"2411.15054","paper_version":1,"verdict":"CONDITIONAL","confidence":"MODERATE","novelty_score":7.0,"correctness_risk":"medium","formal_verification":"none","parameter_count":0,"one_line_summary":"Tiny admissible perturbations of the top-hat kernel leave the 1D nonlocal Fisher-KPP dynamics regularly perturbed when D=O(1), but trigger O(1) changes, hump splitting, and secondary bifurcations when D is comparable to the perturbation size.","lead":"This paper asks whether the predictions of a nonlocal growth equation change when its 'top hat' interaction kernel is slightly modified. It finds that for normal diffusion the model is robust, but for weak diffusion tiny kernel changes create new multi-peaked patterns and extra bifurcations.","discovery_kind":"extension","skeptic_critique":{"model":"deepseek-v4-flash","headline":"The singular-existence theorem (Theorem 3) is proved only for the reduced core problem; matching of the exponentially small tail is deferred and fails for the two-hump profiles used in the bifurcation network, so the analytic claim overreaches its proof.","rationale":"The reader's weakest assumption is exactly the imported two-region structure and the informal matching of the exponentially small tail; my stress-test concurs and sharpens it. The paper's analytic proof of Theorem 3 consists of a Schauder fixed-point argument for the reduced core problem [NBVP] plus an asymptotic matching statement that is not carried out. The paper itself identifies a single-hump core as a sufficient condition for the matching and states that a sufficiently developed double-hump structure can make (67) fail. For the positive kernel perturbation φ_+, double-hump cores are present exactly on the parameter set where the secondary-bifurcation network is later claimed. Therefore the theorem's assertion of existence of a singularly perturbed periodic steady state on all of Ω-(φ) is not established by the proof given; what fills the gap is the numerical exploration of Section 4. This is a correctness risk in the central claim, not merely a matter of presentation. The test I propose would settle whether the [NBVP]-plus-tail representation is actually valid in the double-hump regime: if the full numerical solution satisfies the tail equation and matches the composite to the stated order, the concern is resolved and Theorem 3 can be accepted as a formal-asymptotic result supported by numerics. If not, the theorem needs a restriction, and the bifurcation-network conclusions rest entirely on unverified numerics. Either way, the reader's CONDITIONAL verdict remains appropriate; I do not see a reason to move to REJECT, because the regular regime, the linearized stability analysis, and the qualitative singular structure are credible and independently plausible, and the numerical evidence, while not provided as code or data, is detailed and internally consistent.","tokens_in":42589,"tokens_out":14260,"duration_ms":158321,"concrete_test":"Recompute the upper panel of Figure 7 with an independent continuation code (for example, Fourier spectral discretization with deflation and pseudo-arclength, checked against a finite-difference solver) for ε=0.01, λ=0.95, and at D≈1.4×10^{-6} compute the residual of the leading-order tail equation (220) using the full [FPP] solution. Then construct the matched composite from the [NBVP] core plus an explicitly constructed exponentially small tail and edge-layer solution. If the full solution differs from this composite by more than O(ε) in sup norm, or if φ*e^W − 1 is not strictly positive in the gap region, the ansatz (67) is invalid and Theorem 3 must be restricted to single-hump cores or replaced by a different existence proof.","verdict_should_be":"UNCHANGED","load_bearing_attack":"The central existence theorem in the singular regime is not actually proved for the parameter region where the paper later finds secondary bifurcations. In §3.2.1 the reduction to [NBVP] assumes the two-region structure (67): an O(1) core on [-a(λ), a(λ)] with Dirichlet conditions, an O(||φ||^{1/4}) edge layer, and an exponentially small tail. The Schauder argument solves only the core problem [NBVP]; whether a genuine periodic solution of [FPP] exists is deferred to Remark 1 and the end of §3, where only a sufficient condition (single-hump core) is stated and it is admitted that the form (67) may fail when the core becomes two-humped. For the positive perturbation φ_+, two-humped core profiles occur precisely for D < D* and a > a_c(D) (§3.2.3), and the Section 4 bifurcation network is built in that regime. Thus Theorem 3's claim of existence for every (λ,D) in Ω-(φ) is not supported by the analysis on the set where it matters; the 1-3-5 peak network is supplied by numerical continuation with no code, data, or error control. This is a genuine gap between the theorem as stated and the proof as written.","agreement_with_reader":"agree"},"referee_report":{"model":"deepseek-v4-flash","summary":"The paper analyses the 1D nonlocal Fisher-KPP equation with a top-hat kernel perturbed by an admissible kernel perturbation φ̄ of small norm ||φ̄||_1^m. It claims that for D of order 1 the dynamics and periodic steady states are regular perturbations of the top-hat case (Theorems 1 and 2), whereas for D of order ||φ̄||_1^m the problem is singularly perturbed, leading to O(1) changes in the profile and, for certain perturbations, a network of secondary bifurcations connecting one-, three- and five-peak branches (Theorem 3, Section 4). For two specific cosine perturbations φ_+ and φ_- the core reduction is solved explicitly via a scalar Sturm-Liouville problem, giving hump-splitting and spike-formation regimes, and numerical continuations of the full problem are reported. The paper also presents numerical evolution results indicating that the wave-selection mechanism near wavelength 1/2 is robust.","tokens_in":42766,"tokens_out":7494,"duration_ms":79148,"significance":"If the analytic claims are made to match the proofs, the paper gives a useful and concrete picture of when kernel perturbations are regular and when they become singular: the regular/singular dichotomy at D=O(1) versus D=O(||φ̄||) is a natural and nontrivial extension of the top-hat analysis. The explicit scalar reductions, the threshold values D*≈5.22×10^{-3} and a_c(D), the parabolic-cylinder condition (192), and the comparison of [NBVP] and [FPP] profiles in Figure 6 are concrete, falsifiable predictions. The logarithmic reformulation (217) for the numerical work is also sensible. However, several theorem statements currently go beyond what is proved, and the numerical evidence for the main bifurcation network is not accompanied by code, data, or error control.","major_comments":[{"comment":"Theorem 3 states that for each (λ,D) in Ω_-(φ) there is a positive periodic steady state of [FPP] which is a singular perturbation of the top-hat state. The proof, however, establishes only existence of a solution to the reduced core problem [NBVP] via Schauder's theorem (equations (102)-(104)), and the matching to the exponentially small tail is deferred: Remark 1 and the end of §3 state that the two-region form (67) may fail once the [NBVP] solution has a two-humped core. For the positive perturbation φ_+, two-humped cores occur precisely on {0<D<D*, a_c(D)<a<1/4} (§3.2.3), and this is exactly the regime used for the secondary-bifurcation network in Section 4. Thus Theorem 3's existence assertion is not supported by the proof on the parameter set where the paper's main singular-perturbation conclusion is used. The theorem should either be restricted to the single-hump regime or the matching argument must be supplied.","section":"§3.2.1, Theorem 3; Remark 1; end of §3"},{"comment":"Theorem 2 asserts existence and uniqueness of a positive periodic steady state in Ω_+(φ) together with an O(D^{-1}||φ̄||_1^m) sup-norm bound. What is actually constructed is the first-order correction F_p via the eigenfunction expansion (39)-(40), after which the text says only 'This formally confirms...'. No fixed-point, contraction, or implicit-function argument is given to show that the full nonlinear problem has a solution within o(||φ̄||_1^m) of the unperturbed state, nor is the claimed uniformity over Ω_+(φ) established. The theorem should be proved, or explicitly labelled as a formal asymptotic construction.","section":"§3.1, Eqs. (18)-(40), Theorem 2"},{"comment":"The claim that the singular principal branch is 'the natural continuation' of the regular branch (Theorem 3) is justified only by comparing the asymptotic forms (109) and (110) and invoking the Van Dyke matching principle. No rigorous error bounds or matching theorem are given, and in the two-hump regime the underlying ansatz (67) is admitted to fail (Remark 1; end of §3). The continuation statement should therefore be qualified as formal, or proved under an explicit single-hump condition.","section":"§3.2.1, Eqs. (107)-(110)"}],"minor_comments":[{"comment":"There are several typos that should be corrected: 'perurbation' (Section 1), 'arbitrarilly' (Section 1), 'exponentaialy' and 'principle' (end of Section 3), and 'ration' (Section 6).","section":"Throughout"},{"comment":"The manuscript repeatedly refers to (NM) (for example in Sections 1 and 3) but the reference list contains only [1]-[5]; the companion paper must be cited or its results summarised in a way that does not require the reader to have access to an unpublished manuscript.","section":"References"},{"comment":"The remark numbering is inconsistent: the text introduces 'Remark 1' at the start of §3.2.1 but later refers to 'Remark 3.1'; please harmonize the numbering.","section":"§3.2.1"},{"comment":"The numerical continuation results underlying Figures 7-12 are described in detail but no code, data, or grid-dependence checks are provided; for reproducibility, please include the numerical parameters (tolerances, grid sizes, continuation step sizes) and ideally deposit the code or the data.","section":"§4"},{"comment":"The labels in the bifurcation diagrams (for example '1133357', '1 3357', '1 357') are difficult to read; please redraw the figures with clearer branch labels and a legend explaining the peak counts.","section":"Figures 7 and 9"},{"comment":"The formula a_R(D,ε) = 1/4 - (1/4)(16πD ± 1)ε + O(ε^2) should be checked for a sign error: the definition of ̄D and the convention for the plus/minus sign should be stated in the caption or immediately before the formula.","section":"Eq. (221)"}],"recommendation":"major_revision","confidential_remarks":"The paper contains genuinely useful analytical and numerical work, but the central existence theorem (Theorem 3) is stated more strongly than what is proved, and Theorem 2 also lacks a nonlinear fixed-point argument. I do not think this warrants rejection, because the issues are fixable by restricting the theorems to the parameter regions where the proofs work, or by supplying the missing matching/contraction arguments, and by making the numerical continuation reproducible. Please also ask the authors to sort out the status of the (NM) reference before publication."},"author_rebuttal":null,"desk_editor":{"model":"deepseek-v4-flash","letter":"Punchline: the regular/singular dichotomy you identify is real and worth knowing—small kernel perturbations of the top hat are harmless when D is O(1) but change the periodic steady states at O(1) scale when D is O(||phi||). The paper gives a convincing picture of the singular regime, including a neat reduction of the nonlocal core problem to a scalar equation for cosine perturbations. The main caveat is that Theorem 3 overreaches: existence is proved only for the reduced core problem, and the two-scale matching that would turn that into genuine periodic steady states is verified only under a single-hump condition that fails in precisely the region where the interesting bifurcations occur.\n\nWhat is new and good: the regular/singular classification itself is clean and, as far as I know, new. The Schauder fixed-point argument in l^1 for the nonlocal Sturm–Liouville reduction is a solid idea, even if compressed. The explicit treatment of the two cosine perturbations is the best part: the monotonicity of G1, the WKB asymptotics, the a_c(D) hump-splitting boundary, and the Airy/Gaussian spike limits are concrete and checkable. The numerically discovered 1-3-5 peak cascade is striking and likely robust.\n\nSoft spots: (1) Theorem 3 claims a singularly perturbed positive periodic state for every (lambda,D) in Omega_minus, but the proof solves only the core problem [NBVP]; the exponentially small tail region is deferred to a remark, and the stated sufficient condition for matching (single-hump core) does not hold in the double-hump region used in Section 4. So the theorem as stated is not supported on the set where it matters. The paper should restate it as an existence result for the core problem plus a matching verification under the single-hump condition, and present the secondary-bifurcation network as a numerically supported conjecture. (2) Uniqueness for the negative perturbation is concluded from numerical continuation, not from the analysis. (3) The numerics are described carefully but no code, data, or error bars are provided; Figure 7 cannot be reproduced without significant effort. (4) The reference list is missing the companion papers (NM) and Van Dyke—sloppy for a paper that leans so heavily on its own series.\n\nWho it is for: applied mathematicians working on nonlocal reaction-diffusion equations and singular perturbation methods. The message that top-hat predictions are robust at O(1) diffusivity but not in the small-D regime is well worth knowing. With the theorem restated honestly and the references fixed, the paper deserves publication. I would send it to a serious referee, with a note that the gap between Theorem 3 and its proof needs careful checking.","headline":"A genuinely useful regular/singular dichotomy for kernel perturbations in nonlocal Fisher-KPP, but Theorem 3 is stated more broadly than what is actually proved and the most striking claims rest on unreproduced numerics.","tokens_in":43392,"tokens_out":3277,"would_cite":false,"duration_ms":34947,"reading_group":"maybe","serious_thinker":"yes","would_accept_peer_review":true},"rs_alignment":null,"lean_confirmation":null,"pith_extraction":{"msc":["35K57","35B32","35B25","35Q92"],"pacs":[],"model":"deepseek-v4-flash","headline":"A small kernel tweak stays benign at ordinary diffusion but rewires periodic states into one-, three- and five-peak branches at very small diffusion.","keywords":["nonlocal reaction-diffusion equation","Fisher-KPP equation","top hat kernel","kernel perturbation","regular perturbation","singular perturbation","secondary bifurcations","numerical continuation"],"falsifier":"Perform a higher-order matched-asymptotic calculation of the edge layer and exponentially small tail for one admissible kernel perturbation and check whether the Dirichlet condition at $x=\\pm a(\\lambda)$ survives at leading order; a leading-order correction leaking through the edge layer would invalidate the reduction to [NBVP] and the secondary-bifurcation network built on it.","tokens_in":42289,"feed_emoji":"🔀","tokens_out":12026,"duration_ms":101489,"temperature":0.7,"pith_summary":"The paper studies the one-dimensional nonlocal Fisher-KPP equation with the top hat kernel replaced by $\\phi_T + \\bar\\phi$, where $\\bar\\phi$ is a small admissible perturbation. It establishes a threshold: when the diffusivity $D$ is $O(1)$ relative to the perturbation size, the perturbed dynamics are a uniform regular perturbation of the top-hat dynamics; when $D$ becomes comparable to $\\|\\bar\\phi\\|_1^m$, the problem becomes singular, with $O(1)$ changes in the periodic steady states. For two representative symmetric kernel perturbations, the singular regime reveals a network of secondary bifurcations connecting one-, three- and five-peak periodic branches, driven by hump splitting in the exponentially small tail. The paper argues that the top-hat model's qualitative predictions are robust for moderate diffusion but cease to be so once diffusion is small, no matter how small the kernel perturbation is.","feed_headline":"Kernel perturbation triggers new wave branches at tiny diffusivity","feed_subtitle":"For D comparable to the perturbation size, steady states shift by O(1) and connect one-, three- and five-peak branches.","key_machinery":"The load-bearing machinery is the reduction of the singular regime to the nonlinear nonlocal boundary value problem [NBVP] on the core interval $[-a(\\lambda),a(\\lambda)]$: the steady profile satisfies $\\mathfrak u''+\\mathfrak u(\\mathfrak C-\\bar D^{-1}\\bar J(\\mathfrak u))=0$ with Dirichlet conditions at the edges, evenness and positivity, and normalization $\\int \\mathfrak u\\,dy=1$, where $\\mathfrak C$ is an unknown mass constant. The problem is put in one-to-one correspondence with a Sturm-Liouville eigenvalue problem [SL(I,a)] and a fixed-point equation in $\\ell^1$, so Schauder's theorem gives existence and a principal solution branch; WKB analysis in the limits $I\\to\\pm\\infty$, $a\\to0^+$ and $a\\to(1/4)^-$ yields the hump and delta-function structures and the critical curve $a_c(\\bar D)$. Numerically, the rescaled equation $W=\\log u$ is used to resolve the exponentially small tail where the secondary bifurcations nucleate.","core_discovery":"The central discovery is that robustness of the top-hat dynamics depends on the relative size of $D$ and the kernel perturbation. For admissible $\\bar\\phi\\in K(\\mathbb R)$ with small $\\|\\bar\\phi\\|_1^m$ and $D=O(1)$, the equilibrium states, dispersion relation, neutral curve and the unique even positive periodic steady state in each tongue are all regular perturbations, with displacement of order $D^{-1}\\|\\bar\\phi\\|_1^m$. When $D=O(\\|\\bar\\phi\\|_1^m)$, the paper constructs a reduced nonlinear nonlocal boundary value problem on the core region $[-a(\\lambda),a(\\lambda)]$, $a(\\lambda)=(\\lambda-1/2)/2$, with Dirichlet conditions and an integral normalization, and proves via a Schauder fixed-point argument that at least one $O(1)$ singularly perturbed periodic steady state exists on a principal branch. For the perturbation $\\epsilon\\cos(2\\pi x)$ concentrated at the centre of the kernel support, the principal branch undergoes hump splitting and, for wavelengths $\\lambda\\in(3/4,\\lambda_0(\\epsilon))$, secondary fold bifurcations connect the one-peak branch to three- and five-peak branches; for the inverted perturbation concentrated at the edges, no secondary bifurcations occur on the main parameter range and the core focuses to a single Gaussian-type spike. Direct numerical solution of the full boundary value problem confirms the predicted bifurcation network and the stability exchanges.","pith_inferences":["A testable extension: the threshold $\\lambda=3/4$ for secondary bifurcations comes from requiring the kernel support to span four spikes; repeating the numerical continuation with other admissible kernels should shift that threshold monotonically with support width.","The authors leave implicit that the $O(1)$ changes driving the bifurcations occur in the exponentially small tail, invisible in $u$ but visible in $\\log u$; observational probes of such patterns should measure log-density or spectral quantities rather than raw density.","A likely generalisation of the positive/negative classification is that any symmetric perturbation concentrating nonlocal weight near the origin promotes hump splitting, while concentrating weight near the kernel edges suppresses it; this can be checked with a one-parameter family of shifted symmetric perturbations.","The regular-to-singular switch at $D\\sim\\|\\bar\\phi\\|_1^m$ suggests a scaling principle: however small the kernel's fine structure, it becomes controlling once the diffusion length is comparable to the kernel's $L^1$ magnitude."],"forward_implications":["The linearised stability of $u=0$ is identical to the top-hat case for every admissible perturbation, and the neutral curve for $u=1$ moves by at most $O(\\|\\bar\\phi\\|_1^m)$, so the stability conjectures [P1] and [P2] from part 1 remain valid uniformly in $D>0$ for small perturbations.","Crossing into the boundary region $\\Omega_-=\\Omega_1\\cap\\{D=O(\\|\\bar\\phi\\|_1^m)\\}$, every admissible perturbation produces at least one singularly perturbed positive periodic steady state of $O(1)$ size, with a principal branch that continues the regular branch from $\\Omega_+$.","For the positive cosine-type perturbation with small $\\epsilon$, the one-peak branch develops two humps as $\\bar D$ decreases and, for wavelengths $\\lambda\\in(3/4,\\lambda_0(\\epsilon))$, folds and connects to three- and five-peak branches, creating a window with coexisting stable states of wavelength $\\lambda$ and $\\lambda/3$.","For the negative cosine-type perturbation the principal branch has no secondary bifurcations on the main parameter range, and the core profile focuses to a single central Gaussian-type spike as $\\bar D\\to0$, preserving the top-hat evolutionary mechanism that selects wavelength close to $1/2$.","Direct evolution simulations show the front dynamics remains close to the unperturbed case even when $D=O(\\epsilon)$; the complex bifurcation network is engaged only when initial data select wavelengths in $(3/4,\\lambda_0(\\epsilon))$."],"supporting_citations":[{"why":"Supplies the unperturbed top-hat Cauchy problem (IBVP), the periodic steady states $F_p$, the small-$D$ asymptotics (41)-(43) and the conjectures [P1]/[P2] whose robustness this paper tests.","marker":"[5]"},{"why":"Provides the Sturm-Liouville theory used to define the principal eigenfunction and eigenvalue of the auxiliary problem [SL(I,a)].","marker":"[1]"},{"why":"Supplies the spectral theory of weakly perturbed self-adjoint operators used to solve the inhomogeneous regular-perturbation problem via eigenfunction expansions.","marker":"[2]"},{"why":"Provides the Schauder fixed-point theorem used to prove existence of at least one fixed point and hence a principal solution branch in the singular regime.","marker":"[3]"},{"why":"Underpins the weakly nonlinear analysis in Appendix A that gives the initial small-amplitude periodic states used to start the numerical continuation.","marker":"[4]"}],"fun_headline_variants":["Perturbed kernel shifts stability as diffusivity shrinks","Top-hat kernel tweak spawns peak-branch folds","Small kernel change, big wave pattern shifts","At tiny D, one-peak waves split into five"],"cache_read_input_tokens":3200,"weakest_assumption_plain":"The singular-perturbation conclusions rest on the assumption that the perturbed periodic steady state keeps the same two-region shape as in the top-hat case: an $O(1)$ core on $[-a(\\lambda),a(\\lambda)]$ with Dirichlet conditions at the edges, an exponentially small tail outside, and a thin edge layer of thickness $O(\\|\\bar\\phi\\|_1^{1/4})$ matching the two regions, and this matching is verified only informally.","fun_headline_variants_meta":{"raw":{"variants":["Perturbed kernel shifts stability as diffusivity shrinks","Top-hat kernel tweak spawns peak-branch folds","Small kernel change, big wave pattern shifts","At tiny D, one-peak waves split into five"]},"model":"deepseek-v4-flash","effort":"low","cost_usd":0.000244,"raw_usage":{"total_tokens":1661,"prompt_tokens":1203,"completion_tokens":458,"prompt_tokens_details":{"cached_tokens":384},"prompt_cache_hit_tokens":384,"prompt_cache_miss_tokens":819,"completion_tokens_details":{"reasoning_tokens":394}},"tokens_in":819,"tokens_out":458,"duration_ms":5364,"temperature":1.0,"reasoning_tokens":394,"cache_read_input_tokens":384,"cache_creation_input_tokens":0},"cache_creation_input_tokens":0},"created_at":"2026-08-12T14:33:45.110608+00:00","model_set":{"reader":"deepseek-v4-flash"},"falsifier":"Perform a higher-order matched-asymptotic calculation of the edge layer and exponentially small tail for one admissible kernel perturbation and check whether the Dirichlet condition at $x=\\pm a(\\lambda)$ survives at leading order; a leading-order correction leaking through the edge layer would invalidate the reduction to [NBVP] and the secondary-bifurcation network built on it.","supporting_citations":[{"cited_title":null,"cited_arxiv_id":null,"evidence_quote":"Supplies the unperturbed top-hat Cauchy problem (IBVP), the periodic steady states $F_p$, the small-$D$ asymptotics (41)-(43) and the conjectures [P1]/[P2] whose robustness this paper tests."},{"cited_title":null,"cited_arxiv_id":null,"evidence_quote":"Supplies the spectral theory of weakly perturbed self-adjoint operators used to solve the inhomogeneous regular-perturbation problem via eigenfunction expansions."},{"cited_title":null,"cited_arxiv_id":null,"evidence_quote":"Provides the Schauder fixed-point theorem used to prove existence of at least one fixed point and hence a principal solution branch in the singular regime."},{"cited_title":"V Kantorovich and G","cited_arxiv_id":null,"evidence_quote":"Underpins the weakly nonlinear analysis in Appendix A that gives the initial small-amplitude periodic states used to start the numerical continuation."}],"review_version":1}