arxiv: 2605.04048 · v2 · submitted 2026-05-05 · 🧮 math.PR · q-bio.PE

Recognition: 2 theorem links

· Lean Theorem

Catastrophe-dispersion models in random and varying environments across generations

F\'abio P. Machado, Lucas R. de Lima

Pith reviewed 2026-05-08 18:18 UTC · model grok-4.3

classification 🧮 math.PR q-bio.PE

keywords branching processesvarying environmentsrandom environmentscatastrophe modelsdispersal mechanismssurvival probabilityextinction criterionlog-mean process

0 comments

The pith

Survival and extinction in catastrophe-dispersion branching processes are governed entirely by the log-mean process, exactly as in classical theory.

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

This paper examines branching processes where the number of offspring emerges implicitly from a cycle of colony growth, catastrophic reduction, and structured dispersal rather than being specified directly. Parameters controlling growth, survival, and dispersal are permitted to vary deterministically or randomly from one generation to the next, producing processes in varying and random environments. The central result is that whether the population survives or becomes extinct depends only on the associated log-mean process, the same criterion that governs the classical Galton-Watson case. The claim is established uniformly for four qualitatively different dispersal mechanisms, with explicit thresholds derived for Poissonian growth paired with binomial survival and with illustrations using Yule-Simon growth.

Core claim

We show that survival and extinction are governed entirely by the associated log-mean process, exactly as in the classical theory. The paper treats four qualitatively different dispersal mechanisms and establishes a universal ordering of the induced offspring means. For Poissonian growth with binomial survival, explicit thresholds are obtained that determine extinction or survival uniformly over all four mechanisms.

What carries the argument

The log-mean process: the sequence of logarithms of the mean offspring numbers induced each generation by the growth-catastrophe-dispersal cycle.

If this is right

The extinction criterion applies uniformly across all four dispersal mechanisms considered.
Explicit survival thresholds computed for Poissonian growth and binomial survival work for every dispersal mechanism.
A universal ordering holds among the offspring means induced by the four dispersal mechanisms.
The classical log-mean criterion extends directly to branching processes whose parameters vary randomly or deterministically across generations.

Where Pith is reading between the lines

These are editorial extensions of the paper, not claims the author makes directly.

Detailed mechanics of dispersal may matter less for long-term persistence predictions than accurate measurement of the average logarithmic growth rate.
The same reduction might apply to other implicitly defined offspring distributions arising in ecological or biological models.
Simplified forecasting of population fate in fluctuating habitats could be achieved by tracking only the log-mean sequence rather than full distributions.

Load-bearing premise

The growth-catastrophe-dispersal cycle must induce an offspring distribution that preserves the independent branching property across generations even when parameters vary.

What would settle it

A concrete counter-example would be a simulation or explicit construction in which the expected value of the log of the mean offspring number is positive yet the process becomes extinct with probability one (or the converse case).

read the original abstract

We study a class of branching processes in which the offspring distribution is not specified directly but is induced by a cycle of internal colony growth, catastrophic reduction and structured dispersal. The parameters governing growth, survival and dispersal are allowed to vary deterministically or randomly from one generation to the next, giving rise to branching processes in varying and random environments with implicitly defined offspring laws. We show that survival and extinction are governed entirely by the associated log-mean process, exactly as in the classical theory. The paper treats four qualitatively different dispersal mechanisms and establishes a universal ordering of the induced offspring means. For Poissonian growth with binomial survival, explicit thresholds are obtained that determine extinction or survival uniformly over all four mechanisms. A series of ecologically motivated examples with Yule-Simon growth illustrates the versatility of the framework.

Editorial analysis

A structured set of objections, weighed in public.

Desk editor's note, referee report, simulated authors' rebuttal, and a circularity audit. Tearing a paper down is the easy half of reading it; the pith above is the substance, this is the friction.

Desk Editor's Note private letter to a colleague

This paper induces offspring laws in branching processes from explicit growth-catastrophe-dispersal cycles in varying environments, keeps the classical log-mean extinction criterion, and adds a universal ordering of means across four dispersal types plus explicit thresholds for the Poisson-binomial case.

read the letter

This paper sets up branching processes where the offspring distribution is not chosen directly but emerges from a cycle of colony growth, catastrophe, and structured dispersal, with growth, survival, and dispersal parameters allowed to change deterministically or randomly each generation. Survival and extinction are still governed by the log-mean of the induced offspring means, the same criterion as in the classical theory for varying environments. It also establishes a universal ordering of those means across four qualitatively different dispersal mechanisms and gives explicit thresholds that decide extinction or survival uniformly for all four when growth is Poisson and survival is binomial. Yule-Simon growth examples show the setup can handle other distributions as well. The construction is set up so each individual still branches independently given the environment sequence, which lets the standard results apply without extra work. The ordering result is new relative to the cited varying-environment literature and could be useful for comparing dispersal strategies without re-deriving the whole model each time. The thresholds being mechanism-independent is a practical plus if the derivations hold. The main thing to verify is that the cycle really produces a well-defined offspring law that preserves the independent branching property under varying parameters; the abstract and stress-test indicate the setup is designed to do exactly that, with no obvious hidden dependence or correlation introduced. Without the full proofs it is impossible to rule out small gaps in the steps, but nothing in the described framework suggests a load-bearing flaw. This is aimed at researchers in stochastic processes and population biology who want to embed more ecological detail into branching models while retaining classical tools. It is not a broad advance but has precise, checkable results that are worth having. I would send it to peer review.

Referee Report

2 major / 3 minor

Summary. The manuscript develops a class of branching processes in random and varying environments in which the offspring law is not given directly but is induced by a deterministic or random cycle of colony growth, catastrophic reduction, and one of four structured dispersal mechanisms. The central claim is that survival versus extinction is governed exactly by the associated log-mean process, as in the classical theory; a universal ordering of the induced means across the four dispersal mechanisms is established, and explicit extinction thresholds are derived for the Poissonian-growth/binomial-survival case. Ecologically motivated examples with Yule-Simon growth are used to illustrate the framework.

Significance. If the derivations hold, the work supplies a clean reduction of a family of ecologically structured models to the classical log-mean criterion while preserving the independent branching property. The universal ordering of dispersal mechanisms and the explicit, parameter-explicit thresholds for the Poisson-binomial case are concrete, falsifiable contributions that can be checked against simulation or data. The approach avoids ad-hoc fitting and keeps the number of free parameters minimal, which is a strength for applications in population biology.

major comments (2)

[§3.2] §3.2, the construction of the induced offspring distribution: the argument that each individual’s reproduction remains independent conditional on the environment sequence is stated but not expanded; an explicit verification that the joint offspring count factors into a product of marginals (even when dispersal kernels are spatially structured) is needed to justify direct application of the classical log-mean extinction criterion.
[Theorem 4.2] Theorem 4.2 (Poisson-binomial thresholds): the explicit survival threshold is given in terms of the log-mean, but the derivation of the mean itself for the four dispersal mechanisms is only sketched; the step that equates the mean under random dispersal to the deterministic case should be written out to confirm that no extra variance term appears.

minor comments (3)

[§2] The notation for the log-mean process (denoted variously as m_n, Λ_n, or log μ_n in different sections) should be unified and introduced once in §2.
[Figure 2] Figure 2 (Yule-Simon examples): the caption does not indicate which dispersal mechanism corresponds to each curve; adding a legend or explicit labels would improve readability.
A short remark on the relation to existing literature on branching processes in random environments (e.g., the work of Athreya and Karlin) would help situate the contribution.

Simulated Author's Rebuttal

2 responses · 0 unresolved

We are grateful to the referee for the thorough review and valuable comments. The suggestions will help improve the clarity of the manuscript. We address each major comment below and will incorporate the necessary expansions in the revised version.

read point-by-point responses

Referee: [§3.2] §3.2, the construction of the induced offspring distribution: the argument that each individual’s reproduction remains independent conditional on the environment sequence is stated but not expanded; an explicit verification that the joint offspring count factors into a product of marginals (even when dispersal kernels are spatially structured) is needed to justify direct application of the classical log-mean extinction criterion.

Authors: We agree that an explicit verification strengthens the foundation for applying the classical theory. In the revised manuscript, we will expand the discussion in §3.2 to derive the joint distribution explicitly. Conditional on the environment sequence, each individual's offspring count is determined by independent growth, catastrophe survival, and dispersal steps. For spatially structured kernels, the independence follows from the fact that dispersal decisions and survivals are i.i.d. across individuals, allowing the joint probability to factor as the product of marginals. This confirms the branching property holds, justifying the log-mean criterion. revision: yes
Referee: [Theorem 4.2] Theorem 4.2 (Poisson-binomial thresholds): the explicit survival threshold is given in terms of the log-mean, but the derivation of the mean itself for the four dispersal mechanisms is only sketched; the step that equates the mean under random dispersal to the deterministic case should be written out to confirm that no extra variance term appears.

Authors: We agree that the derivation should be expanded for clarity. In the revised version, we will write out the computation of the induced means for all four dispersal mechanisms in detail. For the random dispersal case, the mean is derived by averaging over the random dispersal choice; by linearity of expectation, this equals the mean obtained under the corresponding deterministic dispersal, with no extra variance contribution to the first moment. This confirms the threshold depends only on the log-mean without additional terms from randomness in dispersal. revision: yes

Circularity Check

0 steps flagged

No significant circularity; classical criterion applied to induced offspring laws

full rationale

The paper defines a branching process whose offspring distribution is induced by an explicit growth-catastrophe-dispersal cycle whose parameters may vary across generations. It then invokes the standard log-mean extinction criterion from classical branching-process theory, which is an external result independent of the present construction. The log-mean is computed directly from the growth, survival and dispersal parameters, but the survival/extinction threshold itself is not redefined, fitted, or derived from the paper's own outputs. The four dispersal mechanisms are shown to produce ordered means by direct calculation from the model equations; no self-citation chain, ansatz smuggling, or renaming of known results is used to establish the central claim. The derivation therefore remains self-contained against the external benchmark of classical branching-process theory.

Axiom & Free-Parameter Ledger

0 free parameters · 2 axioms · 0 invented entities

The framework rests on standard branching-process axioms and the assumption that the cycle induces a valid offspring distribution; no free parameters are fitted to data and no new entities are postulated.

axioms (2)

domain assumption Each individual reproduces independently according to the offspring law induced by the growth-catastrophe-dispersal cycle.
Invoked to justify applying classical branching-process extinction criteria to the log-mean process.
standard math The log-mean process governs extinction probability in the same way as in Galton-Watson processes in varying environments.
Relies on existing theorems for branching processes in random environments; cited as 'exactly as in the classical theory'.

pith-pipeline@v0.9.0 · 5433 in / 1457 out tokens · 55807 ms · 2026-05-08T18:18:34.716165+00:00 · methodology

Review history (2 revisions) →

discussion (0)

Reference graph

Works this paper leans on

18 extracted references · 15 canonical work pages

[1]

Amaya, V

J. Amaya, V. V. Junior, F. P. Machado, and A. Rold´ an-Correa. Growth models under uniform catastrophes. Preprint, available at arXiv:2602.06649, 2025

work page arXiv 2025
[2]

J. R. Artalejo, A. Economou, and M. J. Lopez-Herrero. Evaluating growth measures in an immigration process subject to binomial and geometric catastrophes.Math. Biosci. Eng., 4(4):573–594, 2007. ISSN 1547-1063,1551-0018. doi: 10.3934/mbe.2007.4.573. URL https://doi.org/10.3934/mbe.2007.4.573

work page doi:10.3934/mbe.2007.4.573 2007
[3]

K. B. Athreya and S. Karlin. On branching processes with random environments. I. Extinction probabilities.Ann. Math. Statist., 42:1499–1520, 1971. ISSN 0003-4851. doi: 10.1214/aoms/1177693150. URLhttps://doi.org/10.1214/aoms/1177693150

work page doi:10.1214/aoms/1177693150 1971
[4]

P. J. Brockwell. The extinction time of a general birth and death process with catastrophes.J. Appl. Probab., 23(4):851–858, 1986. ISSN 0021-9002,1475-6072. doi: 10.1017/s0021900200118649. URLhttps://doi.org/10.1017/s0021900200118649. 16 LUCAS R. DE LIMA AND F ´ABIO P. MACHADO

work page doi:10.1017/s0021900200118649 1986
[5]

P. J. Brockwell, J. Gani, and S. I. Resnick. Birth, immigration and catastrophe processes. Adv. in Appl. Probab., 14(4):709–731, 1982. ISSN 0001-8678,1475-6064. doi: 10.2307/1427020. URLhttps://doi.org/10.2307/1427020

work page doi:10.2307/1427020 1982
[6]

Duque, V

F. Duque, V. V. Junior, F. P. Machado, and A. Rold´ an-Correa. Extinction time in growth models subject to binomial catastrophes.J. Stat. Mech. Theory Exp., 10:Paper No. 103501, 14, 2023. ISSN 1742-5468. doi: 10.1088/1742-5468/acf8bc. URLhttps: //doi.org/10.1088/1742-5468/acf8bc

work page doi:10.1088/1742-5468/acf8bc 2023
[7]

P. Jagers. Galton-Watson processes in varying environments.J. Appl. Probability, 11:174– 178, 1974. ISSN 0021-9002,1475-6072. doi: 10.2307/3212594. URLhttps://doi.org/10. 2307/3212594

work page doi:10.2307/3212594 1974
[8]

V. V. Junior, F. P. Machado, and A. Rold´ an-Correa. Dispersion as a survival strategy.J. Stat. Phys., 164(4):937–951, 2016. ISSN 0022-4715,1572-9613. doi: 10.1007/s10955-016-1571-3. URLhttps://doi.org/10.1007/s10955-016-1571-3

work page doi:10.1007/s10955-016-1571-3 2016
[9]

V. V. Junior, F. P. Machado, and A. Rold´ an-Correa. Evaluating dispersion strategies in growth models subject to geometric catastrophes.J. Stat. Phys., 183(2):Paper No. 30, 15, 2021. ISSN 0022-4715,1572-9613. doi: 10.1007/s10955-021-02759-5. URLhttps: //doi.org/10.1007/s10955-021-02759-5

work page doi:10.1007/s10955-021-02759-5 2021
[10]

V. V. Junior, F. P. Machado, and A. Rold´ an-Correa. Extinction time in growth models subject to geometric catastrophes.J. Stat. Mech. Theory Exp., 4:Paper No. 043501, 18,
[11]

doi: 10.1088/1742-5468/acc72e

ISSN 1742-5468. doi: 10.1088/1742-5468/acc72e. URLhttps://doi.org/10.1088/ 1742-5468/acc72e

work page doi:10.1088/1742-5468/acc72e
[12]

Kapodistria, T

S. Kapodistria, T. Phung-Duc, and J. Resing. Linear birth/immigration-death process with binomial catastrophes.Probab. Engrg. Inform. Sci., 30(1):79–111, 2016. ISSN 0269-9648,1469-8951. doi: 10.1017/S0269964815000297. URLhttps://doi.org/10.1017/ S0269964815000297

work page doi:10.1017/s0269964815000297 2016
[13]

Kersting and V

G. Kersting and V. Vatutin.Discrete time branching processes in random environment. John Wiley & Sons, 2017

2017
[14]

Lanchier.Stochastic interacting systems in life and social sciences, volume 5 ofDe Gruyter Series in Probability and Stochastics

N. Lanchier.Stochastic interacting systems in life and social sciences, volume 5 ofDe Gruyter Series in Probability and Stochastics. De Gruyter, Berlin, 2024. ISBN 978-3-11-079175-4; 978- 3-11-079188-4; 978-3-11-079193-8

2024
[15]

F. P. Machado, A. Rold´ an-Correa, and R. B. Schinazi. Colonization and collapse.ALEA Lat. Am. J. Probab. Math. Stat., 14(2):719–731, 2017. ISSN 1980-0436. doi: 10.30757/alea.v14-34. URLhttps://doi.org/10.30757/alea.v14-34

work page doi:10.30757/alea.v14-34 2017
[16]

F. P. Machado, A. Rold´ an-Correa, and V. V. Junior. Colonization and collapse on homogeneous trees.J. Stat. Phys., 173(5):1386–1407, 2018. ISSN 0022-4715,1572-9613. doi: 10.1007/s10955-018-2161-3. URLhttps://doi.org/10.1007/s10955-018-2161-3

work page doi:10.1007/s10955-018-2161-3 2018
[17]

R. B. Schinazi. Does random dispersion help survival?J. Stat. Phys., 159(1):101– 107, 2015. ISSN 0022-4715,1572-9613. doi: 10.1007/s10955-014-1173-x. URLhttps: //doi.org/10.1007/s10955-014-1173-x

work page doi:10.1007/s10955-014-1173-x 2015
[18]

W. L. Smith and W. E. Wilkinson. On branching processes in random environments.Ann. Math. Statist., 40:814–827, 1969. ISSN 0003-4851. doi: 10.1214/aoms/1177697589. URL https://doi.org/10.1214/aoms/1177697589. Department of Mathematics, Institute of Mathematical Sciences and Computation (ICMC), University of S ˜ao Paulo, A v. Trabalhador S ˜ao-Carlense, 40...

work page doi:10.1214/aoms/1177697589 1969