Propagation in the Fisher-KPP equation with Mixed Operator

Alexander Quaas; Bego\~na Barrios; Bryan Pichucho

Propagation in the Fisher-KPP equation with Mixed Operator

T0 review · reviewed 2026-08-05 · deepseek-v4-flash

Challenge this review Re-run · record.json Download PDF Read on arXiv ↗

Pith's one-line read The mixed local-nonlocal Fisher–KPP equation spreads exponentially at rate σ*=f′(0)/(N+2s) and has no nonconstant planar traveling waves.

desk verdict Credible and likely true theorem, but the key invasion lemma is asserted, not proved, so the paper needs a major gap-filled revision before I'd rely on it. read the letter →

arxiv 2508.21151 v1 pith:TGWP2CWJ submitted 2025-08-28 math.AP

Begoña Barrios , Bryan Pichucho , Alexander Quaas This is my paper · ORCID

classification math.AP MSC 35K5735R1135B4035C0747D06

keywords Fisher-KPPequationmixedlocal-nonlocaldiffusionfractionalLaplacianheatkernelestimatesexponentialspreadingratenonexistenceoftravelingwavescomparisonprincipleasymptoticpropagation

verification ladder T0 review T1 audit T2 compute T3 formal

The pith

A machine-rendered reading of the paper's core claim, the machinery that carries it, and where it could break.

The reading

The paper asks what happens to Fisher–KPP propagation when diffusion is both classical and fractional, via the mixed operator L = −Δ + (−Δ)^s. It establishes that the fractional part wins: for concave monostable nonlinearities there are no nonconstant planar traveling waves, and every nontrivial solution starting from a suitably decaying initial datum spreads exponentially, with a sharp rate σ* = f′(0)/(N+2s). This matters because mixed local-nonlocal diffusion models populations that combine ordinary Brownian movement with rare long jumps; the result says the long jumps set the invasion speed, while short-range diffusion only modifies the shape of the front.

What carries the argument

The load-bearing object is the mixed heat kernel H(t,z) = (4πt)^{-N/2} ∫ e^{-|z−y|²/(4t)} p^(s)(t,y) dy, the convolution of the Gaussian heat kernel with the fractional heat kernel. Its two-sided estimates (14)–(15) — which switch between Gaussian and fractional regimes and show the fractional tail dominates for t ≥ 1 — support the definition of mild solutions, the boundedness of the semigroup on weighted spaces Xγ, the comparison principles, and the iterative 'same profile with larger radius' lemma that produces the exponential spreading lower bound.

What would settle it

Compute, via the two-sided bounds (14)–(15), the convolution T_{t0} v0 of the mixed heat kernel with the initial profile v0 of Lemma 4.3 and check whether v(t0,x) ≥ a1 |x|^{−N−2s} on |x| ≥ r0 e^{σ t0}. If the Gaussian part of H suppresses the tail so much that this inequality fails for some t0 ≥ 1, the induction in Lemma 4.3 and the exponential spreading of Theorem 2.7(b) would not follow.

Watch

Extended reading notes

Core claim

The central discovery is that the heat kernel of L = −Δ + (−Δ)^s is, for large times, comparable to the fractional heat kernel alone: H(t,x) is the convolution of the Gaussian kernel with the fractional heat kernel, and for t ≥ 1 or |x| ≥ 1 it retains the same algebraic tail as the fractional kernel. Using this kernel, the paper proves Theorem 2.6: the only [0,1]-valued planar traveling waves of (5) are the constants 0 and 1. It also proves Theorem 2.7: if 0 ≤ u0 ≤ 1 and u0(x) ≤ C|x|^{-N−2s}, then for σ > σ* the solution tends to 0 uniformly in {|x| ≥ e^{σt}}, while for σ < σ* it tends to 1 uniformly in {|x| ≤ e^{σt}}. Thus the fractional Laplacian dictates the asymptotic exponential propaga

Load-bearing premise

The exponential lower bound rests on the assertion in Lemma 4.3 that the solution keeps the same algebraic tail shape, with radius multiplied by e^{σ t0}, after every time step t0; the paper states this follows from the comparison principle but does not display the kernel computation that would prove it.

Editorial extensions

If this is right

If the central claim is correct, every nontrivial solution with power-law-decaying initial data has level sets that advance like e^{σ* t}, not at the linear speed seen in classical KPP equations.
Ahead of the moving ball {|x| ≥ e^{σt}} with σ > σ*, the solution is uniformly close to 0; behind {|x| ≤ e^{σt}} with σ < σ*, it is uniformly close to 1, so the exponential rate is sharp.
No nonconstant planar traveling wave exists for the mixed operator, so the familiar constant-speed KPP front is absent whenever fractional diffusion is present alongside classical diffusion.
The sharp rate depends only on f′(0) and N+2s, not on the classical Laplacian's local smoothing, so the long-time invasion speed is governed by the Levy-jump component rather than by Brownian motion.
The same kernel-comparison mechanism explains the 'initial layer': even at early times the fractional tail of H can dominate and start the exponential acceleration before the Gaussian part has spread the mass locally.

Reading between the lines

Editorial extensions of the paper, not claims the author makes directly.

A direct numerical experiment with compactly supported initial data should show level sets of (5) growing like e^{σ* t} for a range of local-diffusion strengths, in stark contrast with the linear level-set growth of the classical KPP equation; the paper's theorem predicts the rate is independent of the local-diffusion coefficient.
The same kernel-comparison machinery likely extends to other concave monostable nonlinearities or to nonlocal kernels with stable-like tails, provided two-sided heat-kernel bounds analogous to (14)–(15) are available; the paper does not pursue that generalization.
If a second-order correction to the front location exists — for instance a logarithmic delay as in classical KPP — it would be invisible to the uniform statements in Theorem 2.7; detecting it would require tracking level sets with finer precision than e^{σt}.
The iterative step in Lemma 4.3 could be made fully explicit by a direct convolution estimate; until such a computation is displayed, the lower-bound exponential spreading rests on an asserted but plausible induction.

Share X Bluesky LinkedIn Reddit HN

Editorial analysis

A structured set of objections, weighed in public.

Desk editor's note, referee report, and a circularity audit.

Desk Editor's Note

read the letter

The paper extends the Cabré–Roquejoffre result to the mixed Fisher–KPP equation u_t – Δu + (–Δ)^s u = f(u), and I believe the headline claims are true: the fractional term dictates the spreading rate σ* = f'(0)/(N+2s) and excludes nonconstant traveling waves. That is a natural and useful result, and the paper is honest about what it is doing: it follows [8] and uses the Song–Vondracek kernel bounds as the external input.

What is genuinely new are the semigroup estimates in the weighted spaces Xγ (Propositions 3.4–3.5), and the clean representation of the mixed heat kernel as the convolution of the Gaussian and fractional kernels. The upper-bound side of the spreading theorem, Lemma 4.1, is sketched plausibly from (14). The paper also flags its own limitations in Remark 3.11, which I take as a sign of honesty.

The problem is Lemma 4.3. That lemma is the engine for Corollary 4.4, Lemma 4.5, Theorem 2.6, and the lower-bound half of Theorem 2.7, and its proof is one sentence: 'by applying the comparison principle given in Proposition 3.7 one can prove… profile preservation.' Proposition 3.7 requires v ∈ C^1([0,∞);Xγ) and v(t) ∈ Dγ(L) at each time. The mild solution with a piecewise power-law initial datum is not shown to have that regularity, and Remark 3.11 explicitly says Proposition 3.7 is not directly applicable to the needed solution class. For the purely fractional kernel, the profile-preservation step is a computation with a self-similar kernel; here the mixed kernel has two regimes in (14)–(15), so the computation cannot be taken verbatim from [8]. That computation is absent. Lemma 4.2, used for convergence to 1 behind the front, is also asserted with 'we omit them'. And Proposition 2.4 states H(t,x) ≤ C t^{1/(2s)} for all t>0, which fails as t→0: the Gaussian part gives sup_x H(t,·) ~ t^{-N/2}. These are fillable gaps, but they are load-bearing.

You asked for my take: the paper deserves a serious referee, and I would send it out, but the referee should be told to demand the missing kernel computation and a comparison theorem that genuinely covers the mild solutions. I would not cite it in this form. If the gap closes, the paper is a solid contribution for people working on mixed local/nonlocal reaction–diffusion models.

Circularity Check

0 steps flagged · score 0.0 of 10

No circular reasoning; the paper's dependence on [8] and [24] is external and load-bearing only through explicit kernel estimates.

full rationale

I find no circular step. The spreading rate sigma* = f'(0)/(N+2s) is taken from the linearized problem and the kernel estimates of [8,24], not fitted to the target conclusion; no parameter is calibrated to the result being predicted. The comparison principle (Proposition 3.7) is proved from a maximum principle for the mixed semigroup, and Theorems 2.6 and 2.7 reduce to [8]'s arguments plus the kernel bounds, which are external, stated results and therefore count as independent grounding. The one genuinely questionable passage is the proof of Lemma 4.3, which asserts that 'one can show that the solution preserves its profile over each time interval of length t0' and says this follows from Proposition 3.7, without supplying the kernel computation and without addressing the fact that Proposition 3.7 applies to regular X-gamma functions rather than the mild solutions used in the lemma. That is an omitted proof or correctness gap, not a circular reduction: the conclusion of Lemma 4.3 is not assumed as an input, and no equation is shown to be equivalent to itself by construction. Accordingly, the circularity score is 0, with the missing-support concern noted as a correctness risk rather than circularity.

Assumptions & free parameters 0 free parameters · 5 assumptions · 0 invented entities

No data-fitted parameters exist in this analytical paper: sigma* = f'(0)/(N+2s) is a function of the model data, not a fitted constant, and the kernel constants (B, C, c, C_gamma) are generic. No new entities are introduced: the mixed operator L, the spaces X_gamma, and the kernel H all come from prior literature [7, 8, 24]. The load-bearing imported facts are the kernel bounds of [24] and the strategy of [8], and the main unproved internal step is the Lemma 4.3 induction.

assumptions (5)

standard math Two-sided heat kernel bounds for the mixed operator (Proposition 2.5, imported from Song and Vondracek [24], equations (14)-(15)).
The whole of Section 4 leans on H(t,x) being comparable to the fractional kernel for t >= 1 and to the Gaussian or fractional kernel in the regimes of (15). The paper adopts this without proof; it is the external fact that encodes 'the fractional Laplacian dominates'.
standard math Two-sided bounds for the fractional heat kernel p(s) (equation (7)), from [8, 10, 11].
Used in Propositions 2.4-2.5 and Lemmas 4.1-4.3 to bound convolutions against |x|^{-N-2s} tails and to control the semigroup in X_gamma.
standard math The operator P = A + B_s is densely defined, positive, self-adjoint and generates a strongly continuous semigroup (Proposition 2.3, from Biagi et al. [7]).
Justifies the semigroup and mild-solution framework of Sections 2 and 3, including the representation u(t) = H(t,.) * u0.
ad hoc to paper The asserted profile-transport step of Lemma 4.3: after time t0 a power-law-tail initial profile maps to a profile of the same form with radius r1 >= r0 e^(sigma t0).
The paper asserts this is proved by the comparison principle, but no computation is given. Corollary 4.4, Lemma 4.5, Theorems 2.6 and 2.7 all depend on this step.
domain assumption Mild solutions with power-law initial data satisfy the regularity hypotheses needed to apply the comparison principles (C^1 into X_gamma, or classical solutions with |x|^{-N-2s} decay).
The paper flags after Lemma 3.8 and in Remark 3.11 that Proposition 3.7 does not directly cover such solutions and patches with Theorem 3.10, but Lemma 4.3 says it uses Proposition 3.7 on mild solutions. The gap is acknowledged, not closed.

how reviews work

0 comments

Cite this review

Pith. "Pith review of Propagation in the Fisher-KPP equation with Mixed Operator." pith.science (2026). https://pith.science/paper/TGWP2CWJ

@misc{pith2026250821151,
  author       = {Pith},
  title        = {Pith review of: Propagation in the Fisher-KPP equation with Mixed Operator},
  year         = {2026},
  howpublished = {\url{https://pith.science/paper/TGWP2CWJ}},
  note         = {Machine review of arXiv:2508.21151}
}

read the original abstract

Our investigation focuses on the asymptotic spreading behavior of the Fisher-KPP equation with a mixed local-nonlocal operator in the diffusion (see the work by X. Cabr\'e and J.-M. Roquejoffre, 2013, ref.[8]) to the setting of mixed diffusion, which involves both the classical and the fractional Laplacian in order to analyze the long-time dynamics of the equation. A key step in our approach involves the construction and detailed study of the heat kernel associated with the mixed operator, which we use to develop a theory of mild solutions and establish a comparison principle in suitable weighted function spaces. This framework allows us to rigorously establish the non-existence of traveling waves and characterize the large-time spreading rate of solutions. We show that the influence of the fractional Laplacian dominates over the classical Laplacian, especially in the initial layer, where it dictates the exponential propagation rate and the thickness of the solution tails.

Discussion (0). Continue with ORCID to comment.

Reference graph

Works this paper leans on

25 extracted references · 25 canonical work pages

[8]

Cabr´ e and J.-M

X. Cabr´ e and J.-M. Roquejoffre. The Influence of Fractional Diffusion in Fisher-KPP Equations . Communications in Mathematical Physics 320.3 (2013), pp. 679–722

work page 2013
[1]

D. G. Aronson and H. F. Weinberger. Multidimensional Nonlinear Diffusion Arising in Population Genetics. Advances in Mathematics 30.1 (1978), pp. 33–76

work page 1978
[2]

D. G. Aronson and H. F. Weinberger. Nonlinear diffusion in population genetics, combustion, and nerve pulse propagation . Partial Differential Equations and Related Topics. Berlin, Heidelberg: Springer, 1975, pp. 5–49

work page 1975
[3]

Ashok Kumar and N

K. Ashok Kumar and N. Biswas. Strict Faber–Krahn-type inequality for the mixed local–nonlocal operator under polarization . Proceedings of the Edinburgh Mathematical Society 68.2 (2025), pp. 506–525

work page 2025
[4]

Berestycki, F

H. Berestycki, F. Hamel, and L. Roques. Analysis of the periodically fragmented environment model: I – Species persistence . Journal of Mathematical Biology 51.1 (2005), pp. 75–113

work page 2005
[5]

Biagi, D

S. Biagi, D. Serena, V. Enrico, and E. Vecchi. Mixed local and nonlocal elliptic operators: regularity and maximum principles. Communications in Partial Differential Equations 47.3 (2022), pp. 585–629

work page 2022
[6]

Biagi, S

S. Biagi, S. Dipierro, E. Valdinoci, and E. Vecchi. A Faber-Krahn inequality for mixed local and nonlocal operators. Journal d’Analyse Math´ ematique 150.2 (2023), pp. 405–448

work page 2023
[7]

Biagi, F

S. Biagi, F. Punzo, and E. Vecchi. Global solutions to semilinear parabolic equations driven by mixed local–nonlocal operators. Bulletin of the London Mathematical Society 57.1 (2025), pp. 265–284

work page 2025

Show all 25 references

[9]

Cazenave and A

T. Cazenave and A. Haraux. An Introduction to Semilinear Evolution Equations . Clarendon Press, 1998

1998
[10]

Z.-Q. Chen, P. Kim, and T. Kumagai. Global Heat Kernel Estimates for Symmetric Jump Processes . Transactions of the American Mathematical Society 363.9 (2011), pp. 5021–5055

2011
[11]

Chen and T

Z.-Q. Chen and T. Kumagai. Heat kernel estimates for stable-like processes on d-sets . Stochastic Processes and their Applications 108.1 (2003), pp. 27–62

2003
[12]

Coulon Chalmin

A.-C. Coulon Chalmin. Fast propagation in reaction-diffusion equations with fractional diffusion. PhD thesis, University Toulouse 3, 2014, 1 vol. (175 p.)

2014
[13]

Dipierro, X

S. Dipierro, X. Su, E. Valdinoci, and J. Zhang. Qualitative properties of positive solutions of a mixed order nonlinear Schr¨ odinger equation. Discrete and Continuous Dynamical Systems 45.6 (2025), pp. 1948–2000

2025
[14]

Dipierro and E

S. Dipierro and E. Valdinoci. Description of an ecological niche for a mixed local/nonlocal dispersal: An evolution equation and a new Neumann condition arising from the superposition of Brownian and L´ evy processes. Physica A: Statistical Mechanics and its Applications 575 (...

2021
[15]

L. C. Evans. Partial differential equations . Vol. 19. American Mathematical Society, 2022

2022
[16]

R. A. Fisher. The Wave of Advance of Advantageous Genes . Annals of Eugenics 7.4 (1937), pp. 355–369

1937
[17]

Gonz´ alvez, F

I. Gonz´ alvez, F. Quir´ os, F. Soria, and Z. Vondraˇ cek.On the nonlocal heat equation for certain L´ evy operators and the uniqueness of positive solutions . arXiv:2504.04246 [math]. 2025

2025 arXiv
[18]

Grigoryan

A. Grigoryan. Heat Kernel and Analysis on Manifolds . American Mathematical Society
[19]

Hamel and L

F. Hamel and L. Roques. Fast propagation for KPP equations with slowly decaying initial conditions . Journal of Differential Equations 249.7 (2010), pp. 1726–1745

2010
[20]

Kolmogorov, I

A. Kolmogorov, I. Petrovsky, and N. Piskunoff. Etude de l’´ equation de La Diffusion Avec Croissance de La Quantit´ e de Mati` ere et Son Application ` a Un Probl` eme Biologique. Moscow Univ. Bull. Ser. Internat. Sect. A 1 (1937), p. 1. 20

1937
[21]

A. Pazy. Semigroups of Linear Operators and Applications to Partial Differential Equations . Springer Science & Business Media, 2012

2012
[22]

Roquejoffre, L

J.-M. Roquejoffre, L. Rossi, and V. Roussier-Michon. Sharp large time behaviour in N-dimensional Fisher-KPP equations. Discrete and Continuous Dynamical Systems 39.12 (2019), pp. 7265–7290

2019
[23]

Silvestre

L. Silvestre. Regularity of the obstacle problem for a fractional power of the laplace operator . Com- munications on Pure and Applied Mathematics 60.1 (2007), pp. 67–112

2007
[24]

Song and Z

R. Song and Z. Vondraˇ cek. Parabolic Harnack Inequality for the Mixture of Brownian Motion and Stable Process. Tohoku Mathematical Journal 59.1 (2007), pp. 1–19

2007
[25]

P. R. Stinga. User’s guide to the fractional Laplacian and the method of semigroups . Volume 2 Frac- tional Differential Equations. De Gruyter, 2019, pp. 235–266. 21

2019

Pith tools

Reviewed August 5, 2026 · model on record in the stance chip above.

[1] [8]

Cabr´ e and J.-M

X. Cabr´ e and J.-M. Roquejoffre. The Influence of Fractional Diffusion in Fisher-KPP Equations . Communications in Mathematical Physics 320.3 (2013), pp. 679–722

work page 2013

[2] [1]

D. G. Aronson and H. F. Weinberger. Multidimensional Nonlinear Diffusion Arising in Population Genetics. Advances in Mathematics 30.1 (1978), pp. 33–76

work page 1978

[3] [2]

D. G. Aronson and H. F. Weinberger. Nonlinear diffusion in population genetics, combustion, and nerve pulse propagation . Partial Differential Equations and Related Topics. Berlin, Heidelberg: Springer, 1975, pp. 5–49

work page 1975

[4] [3]

Ashok Kumar and N

K. Ashok Kumar and N. Biswas. Strict Faber–Krahn-type inequality for the mixed local–nonlocal operator under polarization . Proceedings of the Edinburgh Mathematical Society 68.2 (2025), pp. 506–525

work page 2025

[5] [4]

Berestycki, F

H. Berestycki, F. Hamel, and L. Roques. Analysis of the periodically fragmented environment model: I – Species persistence . Journal of Mathematical Biology 51.1 (2005), pp. 75–113

work page 2005

[6] [5]

Biagi, D

S. Biagi, D. Serena, V. Enrico, and E. Vecchi. Mixed local and nonlocal elliptic operators: regularity and maximum principles. Communications in Partial Differential Equations 47.3 (2022), pp. 585–629

work page 2022

[7] [6]

Biagi, S

S. Biagi, S. Dipierro, E. Valdinoci, and E. Vecchi. A Faber-Krahn inequality for mixed local and nonlocal operators. Journal d’Analyse Math´ ematique 150.2 (2023), pp. 405–448

work page 2023

[8] [7]

Biagi, F

S. Biagi, F. Punzo, and E. Vecchi. Global solutions to semilinear parabolic equations driven by mixed local–nonlocal operators. Bulletin of the London Mathematical Society 57.1 (2025), pp. 265–284

work page 2025