Population switching under a time-varying environment

Michael Assaf; Tom Israeli

REVIEW 2 major objections 7 minor 51 references

Population switching under a time-varying environment

T0 review · 2 major / 7 minor · reviewed 2026-08-14 · deepseek-v4-flash

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

Pith's one-line read A changing environment exponentially alters the odds of population establishment.

desk verdict New WKB results for switching under time-varying environments, but the temporary-perturbation exponent rests on an unproved path ansatz that needs testing or proof. read the letter →

arxiv 1908.09028 v1 pith:YV6GKBGE submitted 2019-08-23 cond-mat.stat-mech

Tom Israeli , Michael Assaf This is my paper · ORCID

classification cond-mat.stat-mech

keywords populationswitchingestablishmentprobabilitymeantimeWKBapproximationAlleeeffecttime-varyingenvironmentstochasticdynamicssemi-classicalaction

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

This paper analyzes a stochastic population in a pre-established state, subject to a strong Allee effect and constant immigration, and asks how a deterministically time-varying environment changes the rare event of establishment. It claims that under a temporary favorable change, the action governing the exponential establishment probability is reduced by an explicit piecewise-path formula, $S = S_0 - E_p T - \int_{q_p^1}^{q_p^2} [p_0(q) - p_p(q;E_p)]\,dq$, and that under a weak periodic perturbation the mean establishment time is exponentially shortened by the frequency-dependent correction of Eq. (32). These results give ecologists and cell biologists a way to compute, within exponential accuracy, how pulse-like or seasonal environmental changes influence switching rates. The paper verifies the formulas against a modified Gillespie algorithm with explicitly time-dependent rates.

What carries the argument

The machinery is the WKB Hamilton–Jacobi formulation of the master equation, with Hamiltonian $H(q,p,t) = [\lambda(q,t)-\mu(q,t)e^{-p}](e^p-1)$, whose zero-energy heteroclinic trajectory between the fixed points $(q_1,0)$ and $(q_2,0)$ carries the optimal switching path, and the action $S = \int p\,dq - H\,dt$ sets the exponential rate. For the temporary perturbation, the new object is the constant-energy perturbed trajectory $p_p(q;E_p)$, whose energy is fixed by requiring the perturbed segment to last exactly the perturbation duration $T$, via Eq. (20). For the periodic case, the key step is expanding the action to first order in the amplitude and minimizing over the initial phase $t_0$, which yields the closed-form correction of Eq. (32).

What would settle it

Numerically minimize the action functional of Eq. (14) over all smooth heteroclinic paths for a temporary perturbation, without assuming the piecewise constant-energy form, and compare the minimizing action with Eq. (21); any discrepancy beyond the WKB accuracy would falsify the concatenation assumption and hence the predicted establishment probability.

Watch

Extended reading notes

Core claim

The central discovery is that the semi-classical WKB approach, previously used for extinction under time-modulated environments, extends to population switching, with the action computed along a heteroclinic path of an explicitly time-dependent Hamiltonian. For a temporary environmental change, the optimal path is the concatenation of the unperturbed zero-energy trajectory before and after the perturbation and a constant-energy trajectory of energy $E_p$ during it; the action is $S = S_0 - E_p T - \int_{q_p^1}^{q_p^2}[p_0(q) - p_p(q; E_p)]\,dq$, with $E_p$ fixed by the duration constraint of Eq. (20). For a weak sinusoidal perturbation, the leading correction to the action, after minimizing over the phase of the perturbation, is $\Delta S = -\varepsilon \pi \sqrt{1-\delta^2}\,\mathrm{csch}(\pi\omega/\delta)\,\sinh[(\omega/\delta)\arcsin(\delta)]$, so the establishment time decreases exponentially with the perturbation amplitude in a frequency-dependent way. The paper verifies the exponential scaling of both results numerically.

Load-bearing premise

The temporary-perturbation formula assumes that the optimal switching path is exactly the unperturbed zero-energy trajectory before and after the change joined to a constant-energy trajectory during the change, with the energy chosen only so that the path lasts the perturbation duration; the paper does not prove that no other path gives a smaller action.

Editorial extensions

If this is right

A temporary favorable change of magnitude $F$ and duration $T$ exponentially raises the establishment probability, with the exponent given by Eq. (21); longer or stronger perturbations monotonically reduce the action.
Close to the bifurcation limit, the action depends only on the product $FT$ (as long as $FT < 2$), so pulses with the same duration–magnitude product have the same effect on the establishment probability.
A weak periodic perturbation exponentially reduces the mean establishment time; the correction is linear in the amplitude and is largest when the perturbation is slow compared to the relaxation rate.
In the adiabatic limit, the shape of the periodic perturbation, sinusoidal versus square wave, does not change the leading-order exponent; only the amplitude matters.
The same theoretical framework applies to phenotypic switching in gene regulatory networks under extrinsic fluctuations, where switching times depend sensitively on the magnitude, frequency, and duration of environmental variations.

Reading between the lines

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

I infer that the piecewise constant-energy ansatz generalizes to a sequence of temporary switches: the action would accumulate terms $-E_i T_i$ plus pairwise area corrections, providing a route to model randomly switching environments by averaging over switch protocols.
Equation (32)'s factor $\mathrm{csch}(\pi\omega/\delta)$ suggests a cutoff frequency $\omega \sim \delta$ beyond which periodic forcing becomes exponentially ineffective; a microbial evolution experiment with controlled periodic stressors could test whether the measured switching rate saturates to the unperturbed rate at high frequency.
The minimization over the initial phase $t_0$ in the periodic case is analogous to a stochastic-resonance condition: the optimal escape attempt locks to the phase of maximal perturbation, which may be observable in time-resolved switching statistics accumulated over many cycles.
The near-bifurcation result that only the product $FT$ matters hints at a broader principle for weak-noise escape: the leading exponential effect of a temporary change may be governed by the time-integrated perturbation, a conjecture worth testing in other escape models.

Share X Bluesky LinkedIn Reddit HN

Signed reviews

No signed human review yet.

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

Two things you should know before reading: the paper is a legitimate extension of the extinction WKB machinery to establishment (upward switching) under temporary and periodic environmental changes; and the headline temporary-perturbation formula Eq. (21) depends on an optimal-path ansatz that is asserted, not derived. If you work on stochastic population dynamics, the adiabatic and periodic results are worth your time; the temporary case is plausible but not pinned down.

What's actually new: to my knowledge this is the first systematic treatment of establishment probability and mean time under time-dependent environments in this model. The periodic linear correction Eq. (32) is a clean application of existing theory, and the adiabatic result Eq. (36) with its prefactor is a solid, testable prediction. The near-bifurcation reduction Eq. (25) is a nice simplification showing a scaling collapse on FT. The simulations use a modified Gillespie with time-dependent rates and the code description is adequate. All of this is credible.

The soft spot is exactly where the stress-test note lands. In Sec IV A the authors say 'It turns out' the optimal path is the concatenation of the H0=0 trajectory, a constant-energy Hp=Ep trajectory, and the H0=0 trajectory again, with Ep set by the duration T. They never rule out a path that completes the switch during the perturbation, so the post-perturbation segment is unnecessary. That is not a prefactor worry; it changes the exponent. The paper doesn't give a variational argument or a check. The numerical agreement in Fig. 3 confirms the exponential scaling, but the theory is multiplied by an arbitrary prefactor, so it can't distinguish between Eq. (21) and another exponent in the tested range. The authors themselves note the theory fails when the action becomes O(1), which is exactly when a crossing-during-perturbation path might compete. So the central temporary result is under-supported.

The periodic and adiabatic parts don't have this problem; they are standard and the simulations support them without hand-tuned prefactors in the adiabatic case (Fig. 7). The paper cites its own prior work heavily, but that's appropriate because the method is an extension rather than a borrowed result.

Bottom line: this deserves a serious referee. I'd send it to review with a request to either justify the ansatz (even a simple upper bound on alternative actions) or test it numerically by simulating the actual path probabilities. The paper is honest about its limitations and the physics community in population dynamics would cite the periodic and adiabatic results even if the temporary case turns out to need revision.

Referee Report

2 major / 7 minor

Summary. This paper studies noise-induced establishment in a stochastic Verhulst-Allee model with immigration and a strong Allee threshold. In the WKB limit N≫1, it computes the leading-order action for switching from the metastable pre-established state to the established state. For a temporary environmental change, the authors propose a piecewise optimal path constructed from the unperturbed zero-energy trajectory before and after the perturbation and a constant-energy trajectory H_p=E_p during the perturbation; the energy is fixed by the duration condition Eq. (20), and the action is S=S0-E_pT-∫_{q_p1}^{q_p2}(p0-pp)dq (Eq. 21). Near bifurcation this reduces to Eq. (25), which depends only on the product FT. For a periodic perturbation they derive the linear-in-amplitude correction Eq. (32), an adiabatic expression with prefactor Eq. (36), and a square-wave analogue Eq. (39). The analytical exponents are compared with Gillespie simulations for time-dependent rates, using arbitrary prefactors for plotting.

Significance. If the temporary-perturbation ansatz is correct, the paper is a useful contribution: it gives explicit closed-form expressions for the exponential change of establishment probability and mean establishment time under two types of environmental variability, identifies scaling laws (FT dependence near bifurcation, amplitude-dominated adiabatic limit, exponential suppression with frequency), and extends earlier extinction work to switching. The derivation is self-contained from the master equation, the analytical exponents contain no fitted parameters, and the simulations independently implement a time-dependent Gillespie algorithm. The main caveat is that the central temporary-perturbation result rests on an unproved optimal-path ansatz; until that is justified, the significance of Eq. (21) is conditional.

major comments (2)

[Sec. IV A, Eqs. (18)-(21)] The optimality of the piecewise path is asserted, not proved. The text says "It turns out that..." and then constructs the path by matching the H0=0 trajectory before and after the perturbation with a single Hp=Ep trajectory; continuity and conservation produce a candidate, not a minimizer of the action in Eq. (14). In particular, the paper never compares against paths that reach q2 during the favorable window, so that the post-perturbation segment is absent and the final time is free, nor against paths carrying nonzero H0 energy before the perturbation. Because Eq. (15) makes the EP exponentially sensitive to S, any lower-action competitor changes the predicted exponent, not just a prefactor. This is not a pedantic issue: at larger F or T, where the action reduction is large and S can approach O(1), the excluded alternatives are precisely the ones that might compete. Please add a variational or transversality argument, or a numerical minimization of Eq. (14) over path families, or state a regime in which such alternatives are provably subdominant.
[Sec. IV A, Eqs. (19)-(20)] The manuscript does not state the domain of (F,T) for which the duration equation has a solution and for which q_p1 and q_p2 lie in the physical interval (q1,q2). The near-bifurcation limit gives the condition FT<2, but the full Eq. (20) is not analyzed. Since the central formula (21) is meaningful only when such an intersection exists, the domain of validity of the temporary-perturbation result is currently undefined; the numerical failures at large F,T are attributed to WKB breakdown without separating the two possible causes, namely absence of an admissible solution versus action becoming O(1). Please provide this analysis or at least an explicit statement of the assumed parameter range.

minor comments (7)

[Eq. (8)] As written, σ(t)=1-δ tanh(δt/2) describes a trajectory that starts at q2=1+δ and ends at q1=1-δ, opposite to the stated boundary conditions in the same paragraph. The intended optimal path from q1 to q2 requires σ(t)=1+δ tanh(δt/2). The phase in Eq. (31) suggests the authors used the plus sign in the actual integration; please correct Eq. (8) and the time parameterization.
[Sec. IV A.1, Eqs. (24)-(25)] The symbol T is used both for the actual perturbation duration and for the rescaled duration T/δ. Please use a distinct symbol, e.g. \mathcal T, in Eqs. (24)-(25) and in the caption of Fig. 4, because the condition FT<2 is easily misread as a condition on the physical duration.
[Sec. IV B.2] In the paragraph introducing the adiabatic approximation, "w≪δ" should read "ω≪δ".
[Fig. 2 caption] The word "crushed lines" should be "dashed lines" or another standard term.
[Sec. IV A] The multiplicative temporary perturbation is dismissed with "similar results (not shown)", although the paper claims to treat both additive and multiplicative variability. Please include the analog of Eqs. (18)-(21) for the multiplicative case, or state explicitly that the calculation is identical in form with the corresponding λp and μp.
[Abstract and Sec. IV A] For the temporary perturbation the paper computes only the change in EP; the MTE is asserted in footnote [46] to be almost unaffected. Please make this explicit in Sec. IV A when the temporary protocol is introduced, since the abstract mentions both EP and MTE.
[Figs. 3-5] The comparisons use linear vertical scales together with arbitrary multiplicative prefactors. Since the analytical claims are for exponents only, plotting N^{-1}\log ΔP or \log τ would make the agreement and its limits more transparent, especially where the prefactor is not predicted.

Circularity Check

0 steps flagged · score 0.0 of 10

No circularity: the action and rate formulas are derived from the stated master equation and verified against Gillespie simulations, with no fitted parameter renamed as a prediction.

full rationale

The paper's central derivations are self-contained. The unperturbed action S0, the WKB action integral (14), and the rate formulas (10), (21), (32), (36), and (39) are obtained by starting from the explicit master equation, applying a WKB ansatz, and evaluating standard variational/perturbative integrals; no simulation data are used to fix any parameter that then appears as a prediction. The temporary-perturbation action (21) is constructed by integrating Eq. (14) along a piecewise path: the unperturbed zero-energy trajectory outside [t0,t0+T] and a perturbed constant-energy trajectory inside it, with Ep determined by the duration constraint (20). This is a consistency equation, not a fit to the establishment probability being predicted. The periodic-perturbation correction (32) is the closed-form evaluation of the first-order perturbation integral (30), and the adiabatic result (36) uses the independently derived unperturbed rate with an instantaneous parameter, as in the standard adiabatic approximation. The arbitrary multiplicative prefactors used in the figures are explicitly acknowledged as outside exponential accuracy and do not enter the analytical exponential predictions. The paper cites prior work, including the authors' own, for standard WKB and periodic-perturbation formulas, but the load-bearing derivations are reproduced in the manuscript, so these self-citations are not load-bearing reductions. The piecewise-optimality assumption for the temporary perturbation is asserted rather than proven, but that is a correctness/rigor limitation, not a circularity: the resulting action is not equivalent by construction to its inputs. Hence the appropriate circularity score is 0.

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

The central claim rests on the standard WKB large-N approximation, the modeling choice of an established state at infinity, and a specific unproved ansatz for the optimal path under temporary perturbations. No free parameters are fitted to data, and no new entities are introduced.

assumptions (5)

domain assumption The master equation can be solved asymptotically for N >> 1 using the WKB ansatz π(n) = A exp[-N S(q,t)], yielding a Hamilton-Jacobi equation (Sec. III).
The semi-classical approximation is standard in this field but is an assumption about the form of the solution; it is not derived from the master equation.
domain assumption The established state is taken at infinity; a finite established state would not affect the leading-order MTE (footnote 41).
This relies on Ref. [19] and is invoked to replace the finite threshold with a barrier at infinity. If the finite state affected the action, the MTE would change.
ad hoc to paper In the temporary perturbation, the optimal path is the concatenation of the unperturbed zero-energy trajectory and the perturbed constant-energy trajectory, with energy E_p determined by the duration constraint (Eqs. 19-21).
This ansatz is asserted in Sec. IV A ('It turns out...') but not proved to be the global minimizer of the action. The central formula (21) depends on it.
domain assumption The system is in a quasi-stationary distribution before the perturbation, with exponentially small switching probability PBP ≈ t/τ (Sec. IV A).
Standard in metastable dynamics; the calculation of ΔP assumes the pre-perturbation state is quasi-stationary.
standard math For periodic perturbations, the action is minimized over the phase t0 of the perturbation (Sec. IV B).
This is a known result from Refs. [22,23,47], used to select the optimal escape path.

how reviews work

0 comments

Cite this review

Pith. "Pith review of Population switching under a time-varying environment." pith.science (2026). https://pith.science/paper/YV6GKBGE

@misc{pith2026190809028,
  author       = {Pith},
  title        = {Pith review of: Population switching under a time-varying environment},
  year         = {2026},
  howpublished = {\url{https://pith.science/paper/YV6GKBGE}},
  note         = {Machine review of arXiv:1908.09028}
}

read the original abstract

We examine the switching dynamics of a stochastic population subjected to a deterministically time-varying environment. Our approach is demonstrated in the realm of ecology on a problem of population establishment. Here, by assuming a constant immigration pressure along with a strong Allee effect, at the deterministic level one obtains a critical population size beyond which the system experiences establishment. Notably the latter has been shown to be strongly influenced by the interplay between demographic and environmental noise. We consider two prototypical examples for environmental variations: a temporary environmental change, and a periodically-varying environment. By employing a semi-classical approximation we compute, within exponential accuracy, the change in the establishment probability and mean establishment time of the population, due to the environmental variability. Our analytical results are verified by using a modified Gillespie algorithm which accounts for explicitly time-dependent reaction rates. Finally, our theoretical approach can also be useful in studying switching dynamics in gene regulatory networks under extrinsic variations.

Figures

Figures reproduced from arXiv: 1908.09028 by the authors.

**Figure 3.** The change in the establishment probability (EP), [PITH_FULL_IMAGE:figures/full_fig_p004_3.png] view at source ↗

**Figure 4.** The action S as function of the product F T close to the bifurcation limit. Here we compare between the bifurcation result (25) (dashed line), the full analytical solution (21), and numerical Monte-Carlo simulations with S = − log(∆P)/N (symbols). The parameters are N = 3000, δ = 0.2, T = 0.15 and 0 ≤ F ≤ 2.7, while ∆P is multiplied by a constant prefactor of 2. where ω = 2π/T is the angular frequency of the pertu… view at source ↗

Figures from the paper (3 more)

**Figure 5.** Figure 5: The MTE as function of the perturbation amplitude. [PITH_FULL_IMAGE:figures/full_fig_p006_5.png]

**Figure 7.** Figure 7: The MTE as function of the perturbation amplitude [PITH_FULL_IMAGE:figures/full_fig_p007_7.png]

**Figure 8.** Figure 8: The MTE versus the sinusoidal perturbation ampli [PITH_FULL_IMAGE:figures/full_fig_p008_8.png]

Discussion (0). Continue with ORCID to comment.

Reference graph

Works this paper leans on

51 extracted references · 41 canonical work pages

[1]

Here, for simplicity, we only consider the case of additive perturbation, whereas the multiplicative case can be treated in a similar manner

Bifurcation Limit The above results drastically simplify close to the bifur- cation limit, for δ≪ 1, where the stable and unstable ﬁxed 5 points, q1 = 1−δ and q2 = 1 +δ, become close. Here, for simplicity, we only consider the case of additive perturbation, whereas the multiplicative case can be treated in a similar manner. In the following, it is conveni...

work page
[2]

Weak perturbation - Linear Theory Here we assume that the perturbation amplitude is small, ε≪ 1. In this limit, it can be shown that the action (29) satisﬁesS(t0) =S0 + ∆S(t0) [22, 23, 31, 47], where S0 is the action along the unperturbed optimal path{q0(t−t0),p 0(t− 6 Figure 5. The MTE as function of the perturbation amplitude. Shown are Monte-Carlo simu...

work page
[3]

In this limit, the mean rate of establishment, ¯res, reads [31] ¯res = ω 2π 2π/ω∫ 0 res(t′)dt′, (34) where res(t) denotes the instantaneous establishment rate

Adiabatic Approximation We now compute the MTE in the adiabatic limit, w≪ δ, using a diﬀerent approach, which allows ﬁnding the pre- exponential correction to the MTE as well. In this limit, the mean rate of establishment, ¯res, reads [31] ¯res = ω 2π 2π/ω∫ 0 res(t′)dt′, (34) where res(t) denotes the instantaneous establishment rate. Since in this limit t...

work page 2016
[4]

C. W. Gardiner et al., Handbook of stochastic methods, Vol. 3 (springer Berlin, 1985)

1985
[5]

N. G. Van Kampen, Stochastic processes in physics and chem- istry, Vol. 1 (Elsevier, 1992)

work page 1992
[6]

S. R. Beissinger and D. R. McCullough, Population viability analysis (University of Chicago Press, 2002)

work page 2002
[7]

M. S. Bartlett, Stochastic population models; in ecology and epidemiology, Tech. Rep. (1960)

work page 1960
[8]

I. A. Hanski, O. E. Gaggiotti, and O. F. Gaggiotti, Ecology, genetics and evolution of metapopulations (Academic Press, 2004)

work page 2004

Show all 51 references

[9]

Assaf and B

M. Assaf and B. Meerson, Physical review letters100, 058105 (2008)

2008
[10]

Meerson and P

B. Meerson and P. V. Sasorov, Physical Review E 78, 060103 (2008)

2008
[11]

Shahrezaei and P

V. Shahrezaei and P. S. Swain, Proceedings of the National Academy of Sciences 105, 17256 (2008)

2008
[12]

Escudero and A

C. Escudero and A. Kamenev, Physical Review E 79, 041149 (2009)

2009
[13]

Assaf, E

M. Assaf, E. Roberts, and Z. Luthey-Schulten, Physical re- view letters 106, 248102 (2011)

2011
[14]

Biancalani and M

T. Biancalani and M. Assaf, Physical review letters 115, 208101 (2015)

2015
[15]

Sagi and M

C. Sagi and M. Assaf, Journal of Statistical Mechanics: The- ory and Experiment 2019, 063403 (2019)

2019
[16]

Mehta, R

P. Mehta, R. Mukhopadhyay, and N. S. Wingreen, Physical biology 5, 026005 (2008)

2008
[17]

J. H. Brown and A. Kodric-Brown, Ecology 58, 445 (1977). 9

1977
[18]

Hanski, Metapopulation ecology (Oxford University Press, 1999)

I. Hanski, Metapopulation ecology (Oxford University Press, 1999)

1999
[19]

P. A. Stephens, W. J. Sutherland, and R. P. Freckleton, Oikos , 185 (1999)

1999
[20]

Courchamp, L

F. Courchamp, L. Berec, and J. Gascoigne, Allee eﬀects in ecology and conservation (Oxford University Press, 2008)

2008
[21]

Meerson and O

B. Meerson and O. Ovaskainen, Physical Review E 88, 012124 (2013)

2013
[22]

Be’er, M

S. Be’er, M. Assaf, and B. Meerson, Physical Review E 91, 062126 (2015)

2015
[23]

M´ endez, M

V. M´ endez, M. Assaf, A. Mas´ o-Puigdellosas, D. Campos, and W. Horsthemke, Physical Review E 99, 022101 (2019)

2019
[24]

Assaf and B

M. Assaf and B. Meerson, Physical Review E 81, 021116 (2010)

2010
[25]

Dykman, H

M. Dykman, H. Rabitz, V. Smelyanskiy, and B. Vugmeister, Physical review letters 79, 1178 (1997)

1997
[26]

Dykman, B

M. Dykman, B. Golding, L. McCann, V. Smelyanskiy, D. Luchinsky, R. Mannella, and P. McClintock, Chaos 11, 587 (2001)

2001
[27]

M. B. Elowitz, A. J. Levine, E. D. Siggia, and P. S. Swain, Science 297, 1183 (2002)

2002
[28]

Taniguchi, P

Y. Taniguchi, P. J. Choi, G.-W. Li, H. Chen, M. Babu, J. Hearn, A. Emili, and X. S. Xie, Science 329, 533 (2010)

2010
[29]

Assaf, E

M. Assaf, E. Roberts, Z. Luthey-Schulten, and N. Golden- feld, Physical review letters 111, 058102 (2013)

2013
[30]

Roberts, S

E. Roberts, S. Be’er, C. Bohrer, R. Sharma, and M. Assaf, Physical Review E 92, 062717 (2015)

2015
[31]

Lande, The American Naturalist 142, 911 (1993)

R. Lande, The American Naturalist 142, 911 (1993)

1993
[32]

Kamenev, B

A. Kamenev, B. Meerson, and B. Shklovskii, Physical review letters 101, 268103 (2008)

2008
[33]

E. Y. Levine and B. Meerson, Physical Review E 87, 032127 (2013)

2013
[34]

Assaf, A

M. Assaf, A. Kamenev, and B. Meerson, Physical Review E 78, 041123 (2008)

2008
[35]

Assaf, A

M. Assaf, A. Kamenev, and B. Meerson, Physical Review E 79, 011127 (2009)

2009
[36]

Baca¨ eret al

N. Baca¨ eret al. , Journal of mathematical biology 68, 533 (2014)

2014
[37]

Billings and E

L. Billings and E. Forgoston, Ricerche di Matematica , 1 (2017)

2017
[38]

Vilk and M

O. Vilk and M. Assaf, Physical Review E 97, 062114 (2018)

2018
[39]

Be’er and M

S. Be’er and M. Assaf, Physical Review E 97, 020302 (2018)

2018
[40]

Dykman, E

M. Dykman, E. Mori, J. Ross, and P. Hunt, The Journal of chemical physics 100, 5735 (1994)

1994
[41]

D. A. Kessler and N. M. Shnerb, Journal of Statistical Physics 127, 861 (2007)

2007
[42]

D. T. Gillespie, Journal of computational physics 22, 403 (1976)

1976
[43]

D. F. Anderson, The Journal of chemical physics 127, 214107 (2007)

2007
[44]

However, if the pop- ulation size at the established state is suﬃciently large [19], this term has no eﬀect in the leading order on the MTE

In order to have a ﬁnite established state, a nonlinear cubic term has to be added to the death rate. However, if the pop- ulation size at the established state is suﬃciently large [19], this term has no eﬀect in the leading order on the MTE
[45]

Assaf and B

M. Assaf and B. Meerson, Journal of Physics A: Mathemat- ical and Theoretical 50, 263001 (2017)

2017
[46]

Dykman and M

M. Dykman and M. Krivoglaz, Sov. Phys. JETP 50, 30 (1979)

1979
[47]

Moss and P

F. Moss and P. V. McClintock, Noise in nonlinear dynamical systems, Vol. 2 (Cambridge University Press, 1989)

1989
[48]

M. I. Freidlin and A. D. Wentzell, in Random perturbations of dynamical systems (Springer, 1998) pp. 15–43

1998
[49]

In this scenario, while ∆ P turns out to be signiﬁcant, the MTE is almost unaﬀected by the temporary perturbation, as its duration is ﬁnite, and much shorter than the MTE
[50]

Escudero and J

C. Escudero and J. ´A. Rodr´ ıguez, Physical Review E 77, 011130 (2008)

2008
[51]

Volfson, J

D. Volfson, J. Marciniak, W. J. Blake, N. Ostroﬀ, L. S. Tsim- ring, and J. Hasty, Nature 439, 861 (2006)

2006

Pith tools

Reviewed August 14, 2026 · model on record in the stance chip above.

[1] [1]

Here, for simplicity, we only consider the case of additive perturbation, whereas the multiplicative case can be treated in a similar manner

Bifurcation Limit The above results drastically simplify close to the bifur- cation limit, for δ≪ 1, where the stable and unstable ﬁxed 5 points, q1 = 1−δ and q2 = 1 +δ, become close. Here, for simplicity, we only consider the case of additive perturbation, whereas the multiplicative case can be treated in a similar manner. In the following, it is conveni...

work page

[2] [2]

Weak perturbation - Linear Theory Here we assume that the perturbation amplitude is small, ε≪ 1. In this limit, it can be shown that the action (29) satisﬁesS(t0) =S0 + ∆S(t0) [22, 23, 31, 47], where S0 is the action along the unperturbed optimal path{q0(t−t0),p 0(t− 6 Figure 5. The MTE as function of the perturbation amplitude. Shown are Monte-Carlo simu...

work page

[3] [3]

In this limit, the mean rate of establishment, ¯res, reads [31] ¯res = ω 2π 2π/ω∫ 0 res(t′)dt′, (34) where res(t) denotes the instantaneous establishment rate

Adiabatic Approximation We now compute the MTE in the adiabatic limit, w≪ δ, using a diﬀerent approach, which allows ﬁnding the pre- exponential correction to the MTE as well. In this limit, the mean rate of establishment, ¯res, reads [31] ¯res = ω 2π 2π/ω∫ 0 res(t′)dt′, (34) where res(t) denotes the instantaneous establishment rate. Since in this limit t...

work page 2016

[4] [4]

C. W. Gardiner et al., Handbook of stochastic methods, Vol. 3 (springer Berlin, 1985)

1985

[5] [5]

N. G. Van Kampen, Stochastic processes in physics and chem- istry, Vol. 1 (Elsevier, 1992)

work page 1992

[6] [6]

S. R. Beissinger and D. R. McCullough, Population viability analysis (University of Chicago Press, 2002)

work page 2002

[7] [7]

M. S. Bartlett, Stochastic population models; in ecology and epidemiology, Tech. Rep. (1960)

work page 1960

[8] [8]

I. A. Hanski, O. E. Gaggiotti, and O. F. Gaggiotti, Ecology, genetics and evolution of metapopulations (Academic Press, 2004)

work page 2004