{"id":"2723cf3c-e6bc-4e82-9aa0-ae5b825beb38","arxiv_id":"2509.11521","paper_version":2,"verdict":"CONDITIONAL","confidence":"MODERATE","novelty_score":6.0,"correctness_risk":"medium","formal_verification":"none","parameter_count":0,"one_line_summary":"For Fisher-KPP equations with a piecewise-constant shifting environment, the paper derives the exact logarithmic correction to the front location, extending Bramson's correction to growing domains and moving habitat boundaries.","lead":"This paper derives sharp formulas for how the invasion front of a spreading population is delayed by a shifting habitat boundary, including a logarithmic correction in time. It extends classical results on the Fisher-KPP equation to growing domains and to environments whose boundary moves with a logarithmic shift.","discovery_kind":"new_application","skeptic_critique":{"model":"deepseek-v4-flash","headline":"Theorem 1.3's printed boundary condition (1.13) cannot imply the speed lower bound (ζ^{-1})'(t)>cλ+2δ used in Lemma 4.1; the theorem is overbroad, though the main applications satisfy the stronger condition.","rationale":"The paper's central program is a Bramson-type logarithmic correction for shifting environments. The most load-bearing step is Theorem 1.3, since Theorems 1.6–1.8 rest on it. The reader identified that Lemma 4.1's use of (1.13) is invalid: the printed upper bound ζ'≤1/cλ+ε0 gives an inverse slope lower bound below cλ, not above cλ+2δ. This is confirmed by the paper's own comment before the β=2 proof, where it says (1.13) 'does not hold' even though the printed inequality is satisfied; the authors clearly intended a stronger 'boundary outruns the front' condition. The gap is real but localized: the applications have inverse boundary derivative β−η/(t+1)→β>cλ in the supercritical/critical/non-pulling cases, so the stronger hypothesis is met, and the paper provides a separate argument for β=2. The proof of Theorem 1.7 is omitted, but it is a routine variant of Theorem 1.6. I therefore do not see a reason to change the reader's conditional verdict, and I do not regard the paper as refuted. A careful revision should correct (1.13) and supply the missing Theorem 1.7 proof.","tokens_in":29458,"tokens_out":25386,"duration_ms":296232,"concrete_test":"Take R=1, λ=1 (so cλ=2) and let ζ(x)=x/0.6 for large x (with ζ=0 on (-∞,0]), so ζ'=1.666..., which is admissible in (1.13) with ε0=1.2. Then (ζ^{-1})'(t)=0.6<2, directly refuting the claim that (1.13) gives (ζ^{-1})'(t)>2+2δ for any δ>0. This settles that the printed assumption is insufficient. Then verify that replacing (1.13) by ζ'≤1/(cλ+2δ) makes Lemma 4.1's estimate (4.4) valid, and check that for X(t)=βt−η log(t+1) with β>cλ the inverse slope is β−η/(t+1), which satisfies this stronger bound for large t (and the β=2 case is handled separately).","verdict_should_be":"UNCHANGED","load_bearing_attack":"Lemma 4.1 states: 'we use cλ + 2δ < (ζ^{-1})'(t) ≤ 1/ε0 for t≫1 (by (1.13))'. But (1.13) says ε0 ≤ ζ'(x) ≤ 1/cλ + ε0, so (ζ^{-1})'(t)=1/ζ'(x) lies in [1/(1/cλ+ε0), 1/ε0]. The lower endpoint cλ/(1+cλ ε0) is strictly less than cλ, so (1.13) does not force (ζ^{-1})' > cλ+2δ; it even allows boundary speed < cλ (take ζ'∈(1/cλ,1/cλ+ε0]). In that case the KPP front can catch the expanding boundary, and the comparison argument leading to (4.2) has no basis. Thus Theorem 1.3 is not proved from its printed assumptions. This is not cosmetic: the proof replaces the boundary condition by the normalized limit (1.16) only because the boundary is assumed to outrun the front. The applications in Theorems 1.6–1.8 either have ζ^{-1}=X(t) with derivative β>cλ, or are treated separately for β=2, so the main formulas are not contradicted; the fix is to strengthen (1.13) to, e.g., ζ'≤1/(cλ+2δ) for large x, and to note the applications satisfy it. The omitted proof of Theorem 1.7 is a secondary gap.","agreement_with_reader":"agree"},"referee_report":{"model":"deepseek-v4-flash","summary":"The paper studies the precise long-time location and profile convergence of solutions to Fisher-KPP equations in shifting environments. It first extends Bramson's logarithmic-correction theory to a KPP equation posed in a growing domain with a moving boundary (Theorem 1.3), and then applies this framework to the scalar equation u_t = u_xx + u(1 - a χ_{(-∞,X(t)]} - u) with X(t) = βt - η log(t+1). The main results, Theorems 1.6-1.8, give the exact logarithmic delay of the front in the supercritical-pulling, critical, and non-pulling regimes, together with convergence to a traveling-wave profile after subtracting the corrected front position. The proofs are PDE-based, combining heat-kernel estimates, Bramson-type comparison arguments, and gluing of super/subsolutions across the shifting discontinuity.","tokens_in":29927,"tokens_out":21215,"duration_ms":227773,"significance":"If correct, this is a substantial contribution: it extends Bramson's classical logarithmic correction to a class of shifting environments and identifies the precise dependence of the correction on the logarithmic drift η of the shifting boundary. The explicit formulas, e.g. (1.30), (1.9) and (1.31), are falsifiable and should be of interest to both the PDE and mathematical-biology communities. The paper also gives a clean PDE route using Dirichlet heat-kernel estimates in a moving half-line. The strongest feature is the parameter-free derivation of the log-correction coefficient from the linear heat-kernel exponent rather than by ansatz. However, two load-bearing issues need to be addressed: the printed hypothesis of Theorem 1.3 is weaker than the boundary-speed condition actually used in its proof, and the proof of Theorem 1.7 is omitted.","major_comments":[{"comment":"The proof of Lemma 4.1(i) uses the bound \"c_λ + 2δ < (ζ^{-1})'(t) ≤ 1/ε_0 for t≫1 (by (1.13))\", but this is not a consequence of (1.13). From (1.13), ε_0 ≤ ζ'(x) ≤ 1/c_λ + ε_0, so (ζ^{-1})'(t) = 1/ζ'(ζ^{-1}(t)) lies in [c_λ/(1+c_λ ε_0), 1/ε_0]. The lower endpoint is strictly less than c_λ, so (1.13) permits a boundary speed below c_λ. In that case the front, moving at speed c_λ, overtakes the boundary, and the conclusion of Theorem 1.3 cannot hold as stated: the boundary condition would force u(t,ζ^{-1}(t)) to resemble Φ at a large negative argument (near B), while (1.16) with boundary speed v < c_λ forces the normalized boundary value to decay to zero. The proof therefore requires a stronger condition, for example ζ'(x) ≤ 1/(c_λ + 2δ) for large x, equivalently (ζ^{-1})'(t) ≥ c_λ + 2δ. The applications in Theorems 1.6-1.8 satisfy this because β > c_λ in the relevant regimes (with β=2 tre","section":"§4, Lemma 4.1 and Theorem 1.3"},{"comment":"Theorem 1.7, a main result, is not proved: the proof says \"we can repeat the proof of Theorem 1.6, except to replace m_{λ,q}(t) by \\tilde m_q(t) ... We omit the detailed proof here.\" The critical case λ = √(1-a) is precisely where Lemma 4.1(ii)/4.3(ii) and the three branches of (1.9) (q < -2, q = -2, q > -2) must be checked. One also needs to verify that Lemmas 5.2 and 5.3 apply at β = 2(√a+√(1-a)) for all real η. Please include the proof, or at least a detailed sketch that explicitly handles the q-threshold cases and the role of the O(1) constants in Lemma 4.1(ii)/4.3(ii).","section":"§5.3, Proof of Theorem 1.7"}],"minor_comments":[{"comment":"In the proof of Lemma 5.3, \\tilde φ is defined as e^{Rt}φ_{β,η} with R = 1-a, but the comparison with ψ that follows requires the factor e^t used in Lemma 5.2; with e^{(1-a)t} the two sides differ by e^{-a t} and the gluing inequalities (5.17)-(5.22) would not hold. This appears to be a typo (\"as in the proof of Lemma 5.2\" supports that), but it should be corrected explicitly.","section":"§5.3, Lemma 5.3"},{"comment":"The compactness argument contains the displayed inequality \"Φ_{min,R}(x-c_min t + C_2) ≤ u_∞ ≤ Φ_{min,R}(x-c_min t + C_2)\", with the same constant C_2 on both sides. It should be C_1 ≤ u_∞ ≤ C_2.","section":"§4, Proof of Theorem 1.3(ii)"},{"comment":"The proof refers to \"Lemma 1.1\" and \"Lemma 1.2\"; these should be Theorems 1.1 and 1.2.","section":"§4, Lemma 4.1"},{"comment":"In formula (5.3), the factor t_0^{βη/2 - 1} appears. The change of variables in (A.1) yields t_0^{1 - βη/2} times a constant; since t_0 is fixed and can be absorbed into C, this is not a mathematical obstruction, but the displayed formula is misleading and should be corrected.","section":"§5.1, Lemma 5.1"},{"comment":"There are numerous typos and OCR-style errors: \"recdueces\", \"givev\", \"ormtain\", \"nammer\", \"neighhorbood\", \"bXη\", \"Remark 5.4\" referring to u_2 instead of \\bar u_2, and \"d/dt A(t) = ... for t < 0\" in Lemma 5.2 where t > 0 is clearly intended. A careful proofreading pass is needed.","section":"Throughout"}],"recommendation":"major_revision","confidential_remarks":"The hypothesis gap in Theorem 1.3 is the most serious issue. It is localized and clearly fixable by strengthening (1.13) to the boundary-speed condition actually used in the proof, and the main applications satisfy that stronger condition. The omitted proof of Theorem 1.7 should be supplied rather than deferred. I therefore recommend major revision rather than rejection: the central formulas appear sound and significant, but the printed statement of the key theorem and one of the main proofs need work."},"author_rebuttal":null,"desk_editor":{"model":"deepseek-v4-flash","letter":"Short version: the main formulas in Theorems 1.6 and 1.8 are probably correct, and the eta-dependence of the log coefficient is a genuine step beyond speed-level results. But there is a real gap in Theorem 1.3. The printed assumption (1.13) does not imply the lower bound (zeta^{-1})'(t) > c_lambda + 2 delta that Lemma 4.1 uses; it even permits boundary speed below c_lambda. That is not a fatal flaw for the applications—the beta > c_lambda regime used in Theorem 1.6 satisfies the stronger condition, and the beta=2 case is treated separately—but the theorem is overbroad as stated and the proof needs repair (e.g., replace (1.13) with a condition that forces the inverse boundary speed above c_lambda+2 delta). The omitted proof of Theorem 1.7 (\"We omit the detailed proof here\") is a second gap; the critical-pulling case is central to the paper's table, so it should be supplied or at least reduced carefully.\n\nWhat is genuinely good: the coefficient (3/2 - sqrt(a) eta)/lambda_* in (1.30) is new, and it emerges from a coherent mechanism—the Dirichlet heat-kernel exponent in Lemma 5.1 combined with Bramson's m(t). The lower and upper estimates at X(t) in Lemmas 5.2 and 5.3 have consistent exponents, and the gluing construction is credible. The growing-domain theorem is a substantial extension of Bramson, even if its assumptions need tightening. The citation pattern looks fair: prior speed results [24,31] and Bramson are properly credited.\n\nBottom line: this paper deserves a serious referee. I'd send it out, with a request that the authors fix the hypothesis of Theorem 1.3 and fill the proof of Theorem 1.7.","headline":"Sharp eta-dependent log corrections for shifting environments look right, but Theorem 1.3's printed boundary assumption doesn't imply the boundary-speed bound used in its proof, and Theorem 1.7's proof is omitted.","tokens_in":30336,"tokens_out":2826,"would_cite":true,"duration_ms":34208,"reading_group":"maybe","serious_thinker":"yes","would_accept_peer_review":true},"rs_alignment":null,"lean_confirmation":null,"pith_extraction":{"msc":["35B40","35K57","92D25"],"pacs":[],"model":"deepseek-v4-flash","headline":"The paper determines the exact logarithmic delay of transition fronts in a Fisher-KPP equation with a shifting environment, extending the homogeneous-space log correction to moving boundaries that drift logarithmically.","keywords":["logarithmic delay","shifting environment","Fisher-KPP equation","reaction-diffusion equations","traveling wave","growing domain","front position","supercritical pulling"],"falsifier":"Take a=0.5, β=2.5, η=1, solve (1.22) numerically with compact initial data, and measure ξ_b(t). If ξ_b(t)−c_* t + (1/λ_*)(3/2−√aη) log t does not stay bounded as t grows past 10^4, the claimed O(1) precision fails. Separately, a constant boundary ζ'(x)=1/c_λ+ε0 satisfies (1.13) but gives (ζ^{-1})'(t)<c_λ, violating the proof's Lemma 4.1 requirement.","tokens_in":29385,"feed_emoji":"🌊","tokens_out":6980,"duration_ms":75792,"temperature":0.7,"pith_summary":"The paper studies a population front in an environment whose growth rate is higher ahead of a moving boundary and lower behind it, with the boundary following βt−ηlog(t+1). It claims that for compactly supported initial data, the front position is known up to O(1): it is a linear speed c_* t corrected by a logarithmic term whose coefficient depends explicitly on the boundary's log-drift η and on the pulling regime. In the supercritical pulling case, the correction is −(1/λ_*)(3/2−√a η) log t, with λ_* determined by β and the habitat contrast a. The paper also proves convergence to the corresponding traveling wave after this subtraction. If correct, this turns a previously order-o(t) picture of shifting-environment spreading into an exact asymptotic profile.","feed_headline":"Log-time delay of shifting fronts solved exactly","feed_subtitle":"A moving habitat edge gives a precisely known logarithmic correction, pinning the front position to O(1).","key_machinery":"The central object is a KPP equation posed in a growing domain Ω_ζ={t>ζ(x)}, whose boundary data mimic the shifting discontinuity. The proof constructs super- and subsolutions by gluing a heat-kernel solution on the fast-moving side x>X(t) to the nonlinear KPP solution on the slow side, using heat-kernel estimates and a boundary matching condition at the interface. The load-bearing identity is the relation between the effective exponent λ_*=β/2−√a and the boundary's log-slope η, which converts the boundary drift into the coefficient √aη inside the logarithmic correction.","core_discovery":"For 0<a<1, the solution of u_t=u_xx+u(1−aχ_{(−∞,X(t)]}−u), with X(t)=βt−η log(t+1) and compactly supported initial data, approaches the traveling wave Φ_{λ,1−a}(x−m(t)) with m(t) specified regime by regime. In the supercritical pulling range 2<β<2(√a+√(1−a)), m(t)=c_* t − (1/λ_*)(3/2−√a η) log t + O(1), where c_*=λ_*+(1−a)/λ_* and λ_*=β/2−√a. At the critical boundary β=2(√a+√(1−a)), the front follows the critical-speed formula with q=−3/2+η√a, including the log-log correction when q=−2. Far beyond the pulling threshold, the correction is the homogeneous minimal-front value −3/(2√(1−a)) log t, independent of η.","pith_inferences":["If the formula is correct, the logarithmic correction is continuously tunable by η, so a small logarithmic lag of the habitat edge changes the front's O(log t) position; this could be tested by level-set measurements in numerical simulations.","The paper's growing-domain reformulation suggests a general principle: for piecewise-constant shifting environments, the exponent selection and the log correction are governed by the boundary's log-slope, not just its linear speed; similar explicit formulas may hold for other monostable reactions.","Because the coefficient can change sign, there should be a critical η_*≈3/(2√a) at which the front neither lags nor advances logarithmically relative to c_* t; locating this crossover numerically would be a sharp test.","The proof gap in the stated Theorem 1.3 hints that the theorem likely needs a stronger boundary-slope hypothesis; the applications to X(t)=βt−η log(t+1) satisfy it, but the general theorem as printed may fail for slow-growing domains."],"forward_implications":["In the supercritical pulling regime, the front's logarithmic delay coefficient is −(1/λ_*)(3/2−√aη); a positive η can shrink or even reverse the delay, while negative η deepens it.","At the critical value β=2(√a+√(1−a)), the logarithmic correction crosses over through the q=−2 case, producing an additional log-log factor in the front position.","For β>2(√a+√(1−a)), the moving boundary is irrelevant to the correction: the front is the homogeneous minimal front with the classical 3/(2λ_min) log t delay.","In every regime, after subtracting the sharp front position, the solution converges locally uniformly to the corresponding traveling wave profile.","The same formula applies for β=2 with η<1/2, where the boundary is only marginally faster than the minimal speed."],"fun_headline_variants":["Exact log-time correction for shifting fronts","Shifting environments give precise front delay","Logarithmic front delay pinned down in shifting habitats","Fisher-KPP fronts in moving boundaries: log correction solved","Precise log shift for spreading fronts in changing habitats"],"cache_read_input_tokens":2304,"weakest_assumption_plain":"The growing-domain theorem relies on the boundary's inverse speed being strictly larger than c_λ+2δ eventually, but assumption (1.13) only guarantees a weaker bound, so the theorem as stated depends on an unstated stronger slope condition.","fun_headline_variants_meta":{"raw":{"variants":["Exact log-time correction for shifting fronts","Shifting environments give precise front delay","Logarithmic front delay pinned down in shifting habitats","Fisher-KPP fronts in moving boundaries: log correction solved","Precise log shift for spreading fronts in changing habitats"]},"model":"deepseek-v4-flash","effort":"low","cost_usd":0.000172,"raw_usage":{"total_tokens":1075,"prompt_tokens":674,"completion_tokens":401,"prompt_tokens_details":{"cached_tokens":256},"prompt_cache_hit_tokens":256,"prompt_cache_miss_tokens":418,"completion_tokens_details":{"reasoning_tokens":328}},"tokens_in":418,"tokens_out":401,"duration_ms":5589,"temperature":1.0,"reasoning_tokens":328,"cache_read_input_tokens":256,"cache_creation_input_tokens":0},"cache_creation_input_tokens":0},"created_at":"2026-08-04T16:47:26.949394+00:00","model_set":{"reader":"deepseek-v4-flash"},"falsifier":"Take a=0.5, β=2.5, η=1, solve (1.22) numerically with compact initial data, and measure ξ_b(t). If ξ_b(t)−c_* t + (1/λ_*)(3/2−√aη) log t does not stay bounded as t grows past 10^4, the claimed O(1) precision fails. Separately, a constant boundary ζ'(x)=1/c_λ+ε0 satisfies (1.13) but gives (ζ^{-1})'(t)<c_λ, violating the proof's Lemma 4.1 requirement.","supporting_citations":[],"review_version":1}