pith. sign in

arxiv: 2409.19378 · v2 · submitted 2024-09-28 · 🧬 q-bio.PE · cond-mat.dis-nn· cond-mat.stat-mech

Stochastic quasi-cycles as a simple explanation for the time evolution of the Cape Rodney-Okakari Point Marine ecological reserve

C\'esar Parra-Rojas , Duccio Fanelli , Alan J. McKane This is my paper

Pith reviewed 2026-05-23 21:14 UTC · model grok-4.3

classification 🧬 q-bio.PE cond-mat.dis-nncond-mat.stat-mech
keywords stochastic quasi-cyclesmarine ecological reservepopulation oscillationsmaximum likelihoodtime series analysisintertidal communityecological modeling
0
0 comments X

The pith

Stochastic quasi-cycles from intrinsic noise explain the population oscillations in the Cape Rodney-Okakari Point Marine reserve.

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

The authors show that stochastic quasi-cycles provide a simple explanation for the cyclic time evolution of species abundances in the Cape Rodney-Okakari Point Marine reserve, based on more than 20 years of data. Previous deterministic descriptions needed ad hoc mechanisms to match the observations. By applying a maximum likelihood method to interpolate stochastic trajectories, the paper demonstrates that fluctuations around a stable fixed point can account for the patterns. This is significant because it offers a way to identify stochastic quasi-cycles using only a single trajectory, which is often the case in ecological field studies.

Core claim

Following a maximum likelihood approach to interpolate individual stochastic trajectories, we here propose quasi-cycles as an alternative and simpler mechanism to explain the oscillations observed in the population numbers of the ecosystem. From a general standpoint, we also show that it is possible to return conclusive evidence on the existence of stochastic quasi-cycles, without resorting to global fitting strategies which necessitate handling a large collection of independent replicas of the dynamics.

What carries the argument

Maximum likelihood interpolation of individual stochastic trajectories around a stable fixed point

If this is right

  • The observed oscillations can be reproduced without ad hoc deterministic mechanisms.
  • Conclusive evidence for quasi-cycles can be obtained from a single trajectory.
  • This avoids the need for global fitting over many independent replicas.

Where Pith is reading between the lines

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

  • Similar single-trajectory methods could be used in other ecological studies to test for stochastic origins of cycles.
  • Management of marine reserves might benefit from considering intrinsic noise as a driver of apparent cycles.
  • Extensions could involve predicting how changes in noise strength affect cycle visibility.

Load-bearing premise

The observed cycles arise from intrinsic stochasticity around a stable fixed point rather than deterministic oscillatory mechanisms, and that maximum likelihood interpolation on a single trajectory can distinguish these cases.

What would settle it

If a deterministic model with oscillatory mechanisms fits the data better under maximum likelihood than the stochastic model with a stable fixed point.

Figures

Figures reproduced from arXiv: 2409.19378 by Alan J. McKane, C\'esar Parra-Rojas, Duccio Fanelli.

Figure 1
Figure 1. Figure 1: FIG. 1. Stochastic realization from the Lotka–Volterra model (5) (continuous line), with its cor [PITH_FULL_IMAGE:figures/full_fig_p007_1.png] view at source ↗
Figure 2
Figure 2. Figure 2: FIG. 2. Distribution of the parameters obtained for the data from the Lotka–Volterra model, after [PITH_FULL_IMAGE:figures/full_fig_p012_2.png] view at source ↗
Figure 3
Figure 3. Figure 3: FIG. 3. Reconstruction of the Lotka–Volterra time series by using the parameters obtained from the [PITH_FULL_IMAGE:figures/full_fig_p013_3.png] view at source ↗
Figure 4
Figure 4. Figure 4: FIG. 4. Reconstruction of the time series obtained from the processed CR-OPM data, shown in [PITH_FULL_IMAGE:figures/full_fig_p015_4.png] view at source ↗
Figure 5
Figure 5. Figure 5: FIG. 5. Distribution of the parameters obtained for the data from the Lotka–Volterra model, after [PITH_FULL_IMAGE:figures/full_fig_p023_5.png] view at source ↗
Figure 6
Figure 6. Figure 6: FIG. 6. Reconstruction of the Lotka–Volterra time series, using the parameters obtained from the [PITH_FULL_IMAGE:figures/full_fig_p024_6.png] view at source ↗
Figure 7
Figure 7. Figure 7: FIG. 7. Processed and normalized data from [3]. [PITH_FULL_IMAGE:figures/full_fig_p025_7.png] view at source ↗
read the original abstract

The dataset collected at the Cape Rodney-Okakari Point Marine (CR-OPM) reserve on the North Island of New Zealand is rather unique. It describes the cyclic time evolution of a rocky intertidal community, with the relative abundances of the various coastal species that have been meticulously monitored for more than 20 years. Past theoretical studies, anchored on a deterministic description, required invoking ad hoc mechanisms to reproduce the observed dynamical paths. Following a maximum likelihood approach to interpolate individual stochastic trajectories, we here propose quasi-cycles as an alternative and simpler mechanism to explain the oscillations observed in the population numbers of the ecosystem. From a general standpoint, we also show that it is possible to return conclusive evidence on the existence of stochastic quasi-cycles, without resorting to global fitting strategies which necessitate handling a large collection of independent replicas of the dynamics, a possibility that is often precluded in real life applications.

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.

Referee Report

3 major / 3 minor

Summary. The manuscript proposes stochastic quasi-cycles—arising from intrinsic noise around a stable fixed point in a stochastic dynamical model—as a simpler explanation for the multi-decadal population oscillations observed in the Cape Rodney-Okakari Point Marine reserve dataset. Using maximum-likelihood interpolation of a single trajectory, the authors claim this approach yields conclusive evidence for quasi-cycles without requiring global fitting across multiple independent replicas, in contrast to prior deterministic models that invoked ad hoc mechanisms.

Significance. If validated, the result would offer a parsimonious stochastic mechanism for ecological oscillations and a practical inference method for single long time series. The single-trajectory MLE strategy, if shown to be identifiable and superior to deterministic alternatives, would be a useful addition to the quantitative ecology toolkit; however, the manuscript provides no explicit model-comparison statistics or identifiability analysis to support the claim.

major comments (3)
  1. [Results / Model comparison] The central claim that quasi-cycles are preferred requires explicit comparison of the stochastic model likelihood against a deterministic limit-cycle or weakly damped oscillator alternative, yet no such comparison (likelihood ratio, AIC, or posterior odds) is reported in the results or methods sections.
  2. [Methods] The assertion that single-trajectory MLE suffices to distinguish a stable focus plus noise from a deterministic oscillatory mechanism is load-bearing but untested; the likelihood surface for a single noisy trajectory is typically non-identifiable between these cases without additional constraints or replica data (Methods, MLE procedure).
  3. [Abstract / Results] Abstract and results claim 'conclusive evidence' from the fitted parameters, but no goodness-of-fit metrics, parameter values, or spectrum comparison against the data are shown to substantiate that the quasi-cycle interpretation reproduces the observed frequencies better than ad hoc deterministic alternatives.
minor comments (3)
  1. [Methods] Notation for the stochastic model (drift and diffusion terms) should be defined explicitly with reference to the underlying master equation or Langevin equation.
  2. [Figures] Figure captions should include the number of data points, time span, and species included in the fit.
  3. [Discussion] The manuscript would benefit from a short discussion of how the chosen stochastic ansatz avoids the parameter proliferation that plagued earlier deterministic models.

Simulated Author's Rebuttal

3 responses · 0 unresolved

We thank the referee for the detailed and constructive report. The comments highlight important gaps in model comparison, identifiability testing, and quantitative substantiation of the claims. We address each point below and will incorporate revisions to strengthen the manuscript.

read point-by-point responses

  1. Referee: [Results / Model comparison] The central claim that quasi-cycles are preferred requires explicit comparison of the stochastic model likelihood against a deterministic limit-cycle or weakly damped oscillator alternative, yet no such comparison (likelihood ratio, AIC, or posterior odds) is reported in the results or methods sections.

    Authors: We agree that explicit quantitative model comparison is necessary to support the preference for the stochastic quasi-cycle model. In the revised manuscript we will add likelihood-ratio tests and AIC comparisons between the stochastic model and deterministic alternatives (limit-cycle and weakly damped oscillator formulations) in both the Results and Methods sections. revision: yes

  2. Referee: [Methods] The assertion that single-trajectory MLE suffices to distinguish a stable focus plus noise from a deterministic oscillatory mechanism is load-bearing but untested; the likelihood surface for a single noisy trajectory is typically non-identifiable between these cases without additional constraints or replica data (Methods, MLE procedure).

    Authors: The referee correctly notes that identifiability between a stable focus plus noise and deterministic oscillations is not automatically guaranteed for a single trajectory. While our fitted parameters on the CR-OPM series produce dynamics consistent with quasi-cycles, we acknowledge that a dedicated identifiability analysis (e.g., via profile likelihood or simulation-based recovery tests) is absent. We will add such an analysis to the Methods section in revision. revision: yes

  3. Referee: [Abstract / Results] Abstract and results claim 'conclusive evidence' from the fitted parameters, but no goodness-of-fit metrics, parameter values, or spectrum comparison against the data are shown to substantiate that the quasi-cycle interpretation reproduces the observed frequencies better than ad hoc deterministic alternatives.

    Authors: We will revise the Abstract and Results to report the maximum-likelihood parameter values, quantitative goodness-of-fit measures (e.g., residual spectra and periodogram comparisons), and direct frequency matching between the fitted quasi-cycle model and the data. This will replace the current qualitative claim of 'conclusive evidence' with explicit numerical support. revision: yes

Circularity Check

0 steps flagged

No significant circularity; derivation self-contained against external benchmarks

full rationale

The abstract presents a maximum-likelihood interpolation of single stochastic trajectories as a method to evidence quasi-cycles without global fitting or multiple replicas. No equations, fitted parameters renamed as predictions, self-citations, or ansatzes are exhibited that would reduce any claimed result to its own inputs by construction. The central claim is methodological (single-trajectory MLE suffices for conclusive evidence) and is offered as an alternative to prior deterministic models; it does not rely on a load-bearing self-citation chain or definitional equivalence. This is the normal case of an independent proposal whose validity can be checked externally via model comparison or additional data.

Axiom & Free-Parameter Ledger

0 free parameters · 0 axioms · 0 invented entities

Review performed on abstract only; no explicit free parameters, axioms, or invented entities are stated in the provided text. The modeling framework implicitly assumes standard stochastic population dynamics (master equation or Langevin description) whose details are not supplied.

pith-pipeline@v0.9.0 · 5712 in / 1156 out tokens · 19563 ms · 2026-05-23T21:14:59.377884+00:00 · methodology

discussion (0)

Sign in with ORCID, Apple, or X to comment. Anyone can read and Pith papers without signing in.

Reference graph

Works this paper leans on

30 extracted references · 30 canonical work pages

  1. [1]

    In this case, the state vector is given by x ≡ (x, y), where x and y represent the population densities of prey and predator, respectively

    An illustrative example: The Lotka–Volterra model As an illustration of the above machinery, we select a simple Lotka–Volterra model de- scribing predator-prey dynamics (see Appendices for details about the mathematical set- ting). In this case, the state vector is given by x ≡ (x, y), where x and y represent the population densities of prey and predator,...

    work page 1975

  2. [2]

    It should be however remarked that it is not in general straightforward to access a sufficiently large collection of independent experimental realizations of the dynamics

    the power spectrum of fluctuations as collected from different replicas of the experimen- tal time series was compared with the corresponding theoretical profile, as obtained from the governing master equation, under the linear noise approximation. It should be however remarked that it is not in general straightforward to access a sufficiently large colle...

    work page

  3. [3]

    Of the 251 total entries in the interpolated dataset, we find that more than half of them are incorrectly normalized

    max( B, A) + M + R ≤ 100. Of the 251 total entries in the interpolated dataset, we find that more than half of them are incorrectly normalized. The situation is fine for the first of the conditions above: the latest data point with this problem comes just after Dec 1995—when a change in the procedure used to sample the population numbers was introduced [3...

    work page 1995

  4. [4]

    Then b0 = B/100 − bA and a0 = A/100 − bA

    For correctly normalized data we take bA = (B + A + M + R)/100 − 1. Then b0 = B/100 − bA and a0 = A/100 − bA

    work page

  5. [5]

    We truncate the values to bA = 0 and then fill the remaining ‘unoccupied’ space with rock by forcing r = 1 − (B + A + M)/100

    For points violating the first condition, the “excess” over 100 is negative, which would result in bA < 0. We truncate the values to bA = 0 and then fill the remaining ‘unoccupied’ space with rock by forcing r = 1 − (B + A + M)/100

    work page

  6. [6]

    • If not, we force r = 1 − [max(B, A) + M]/100

    For points violating the second condition, there are two cases: • If max(B, A) + M > 100, we set r = 0 and force m = 1 − max(B, A)/100. • If not, we force r = 1 − [max(B, A) + M]/100. After this, we assume that whichever population is smaller between B and A is fully concentrated in bA; that is, if B > A (resp. A > B ), then bA = A/100, b0 = (B − A)/100 a...

    work page

  7. [7]

    Marine reserves: the need for systems,

    William J. Ballantine and Timothy.J. Langlois, “Marine reserves: the need for systems,” Hydrobiologia 606, 35–44 (2008)

    work page 2008

  8. [8]

    Fifty years on: Lessons from marine reserves in new zealand and principles for a worldwide network,

    William J. Ballantine, “Fifty years on: Lessons from marine reserves in new zealand and principles for a worldwide network,” Biological Conservation 176, 297–307 (2014)

    work page 2014

  9. [9]

    Species fluctuations sustained by a cyclic succession at the edge of chaos,

    Elisa Beninc` a, Bill Ballantine, Stephen P. Ellner, and Jef Huisman, “Species fluctuations sustained by a cyclic succession at the edge of chaos,” Proceedings of the National Academy of Sciences 112, 6389–6394 (2015)

    work page 2015

  10. [10]

    Van Kampen, Stochastic Processes in Physics and Chemistry , North-Holland Personal Library (Elsevier Science, 2011)

    N.G. Van Kampen, Stochastic Processes in Physics and Chemistry , North-Holland Personal Library (Elsevier Science, 2011)

    work page 2011

  11. [11]

    Gardiner, Stochastic Methods: A Handbook for the Natural and Social Sciences , Springer Series in Synergetics (Springer Berlin Heidelberg, 2009)

    C. Gardiner, Stochastic Methods: A Handbook for the Natural and Social Sciences , Springer Series in Synergetics (Springer Berlin Heidelberg, 2009). 22 0.98 0.99 1.00 β 0 100 200 300 400 4.38 4.40 4.42 4.44 4.46 4.48 4.50 γ 0 20 40 60 80 195 200 205 210 N 0.0 0.1 0.2 0.3 0.4 0.5 0.950 0.952 0.954 0.956 0.958 0.960 β 0 100 200 300 4.320 4.325 4.330 4.335 4...

    work page 2009

  12. [12]

    Predator-prey cycles from resonant amplification of de- mographic stochasticity,

    A. J. McKane and T. J. Newman, “Predator-prey cycles from resonant amplification of de- mographic stochasticity,” Phys. Rev. Lett. 94, 218102 (2005)

    work page 2005

  13. [13]

    Intertangled stochastic motifs in networks of excitatory-inhibitory units,

    Clement Zankoc, Duccio Fanelli, Francesco Ginelli, and Roberto Livi, “Intertangled stochastic motifs in networks of excitatory-inhibitory units,” Phys. Rev. E 96, 022308 (2017)

    work page 2017

  14. [14]

    Noise- driven neuromorphic tuned amplifier,

    Duccio Fanelli, Francesco Ginelli, Roberto Livi, Niccol´ o Zagli, and Clement Zankoc, “Noise- driven neuromorphic tuned amplifier,” Phys. Rev. E 96, 062313 (2017)

    work page 2017

  15. [15]

    Intrinsic noise and two-dimensional maps: Quasicycles, quasiperiodicity, and chaos,

    C´ esar Parra-Rojas, Joseph D. Challenger, Duccio Fanelli, and Alan J. McKane, “Intrinsic noise and two-dimensional maps: Quasicycles, quasiperiodicity, and chaos,” Phys. Rev. E 90, 032135 (2014)

    work page 2014

  16. [16]

    Stochastic oscillations in models of epidemics on a network of cities,

    G. Rozhnova, A. Nunes, and A. J. McKane, “Stochastic oscillations in models of epidemics on a network of cities,” Phys. Rev. E 84, 051919 (2011)

    work page 2011

  17. [17]

    Enhanced stochastic oscillations in autocatalytic reactions,

    Thierry Dauxois, Francesca Di Patti, Duccio Fanelli, and Alan J. McKane, “Enhanced stochastic oscillations in autocatalytic reactions,” Phys. Rev. E 79, 036112 (2009)

    work page 2009

  18. [18]

    Deterministic inference for stochastic systems using multiple shoot- ing and a linear noise approximation for the transition probabilities,

    C. Zimmer and S. Sahle, “Deterministic inference for stochastic systems using multiple shoot- ing and a linear noise approximation for the transition probabilities,” IET Systems Biology 9, 181–192 (2015). 23 FIG. 6. Reconstruction of the Lotka–Volterra time series, using the parameters obtained from the estimation with a sampling interval of 10 timesteps ...

    work page 2015

  19. [19]

    Reconstructing the hidden states in time course data of stochastic mod- els,

    Christoph Zimmer, “Reconstructing the hidden states in time course data of stochastic mod- els,” Mathematical Biosciences 269, 117 – 129 (2015)

    work page 2015

  20. [20]

    Stochastic pattern formation and spontaneous polarisation: the linear noise approximation and beyond,

    Alan J McKane, Tommaso Biancalani, and Tim Rogers, “Stochastic pattern formation and spontaneous polarisation: the linear noise approximation and beyond,” Bulletin of mathemat- ical biology 76, 895–921 (2014)

    work page 2014

  21. [21]

    Stochastic formulation of ecological models and their applications,

    Andrew J. Black and Alan J. McKane, “Stochastic formulation of ecological models and their applications,” Trends in Ecology and Evolution 27, 337–345 (2012)

    work page 2012

  22. [22]

    Exact stochastic simulation of coupled chemical reactions,

    Daniel T Gillespie, “Exact stochastic simulation of coupled chemical reactions,” The Journal of Physical Chemistry 81, 2340–2361 (1977). 24 0.0 0.2 0.4 0.6 0.8 b0 0.0 0.2 0.4 0.6 0.8 bA 0.00 0.01 0.02 0.03 0.04 a0 0 1000 2000 3000 4000 t [days] 0.0 0.2 0.4 0.6 m FIG. 7. Processed and normalized data from [3]

    work page 1977

  23. [23]

    Gaussian process ap- proximations for fast inference from infectious disease data,

    Elizabeth Buckingham-Jeffery, Valerie Isham, and Thomas House, “Gaussian process ap- proximations for fast inference from infectious disease data,” Mathematical biosciences 301, 111–120 (2018)

    work page 2018

  24. [24]

    Differential evolution—a simple and efficient heuristic for global optimization over continuous spaces,

    Rainer Storn and Kenneth Price, “Differential evolution—a simple and efficient heuristic for global optimization over continuous spaces,” Journal of global optimization 11, 341–359 (1997)

    work page 1997

  25. [25]

    SciPy: Open source scientific tools for Python,

    Eric Jones, Travis Oliphant, Pearu Peterson, et al. , “SciPy: Open source scientific tools for Python,” (2001–), [Online; accessed October 1, 2024]

    work page 2001

  26. [26]

    Robust, coherent, and synchronized circadian clock-controlled oscillations along Anabaena 25 filaments,

    Rinat Arbel-Goren, Valentina Buonfiglio, Francesca Di Patti, Sergio Camargo, Anna Zhit- nitsky, Ana Valladares, Enrique Flores, Antonia Herrero, Duccio Fanelli, and Joel Stavans, “Robust, coherent, and synchronized circadian clock-controlled oscillations along Anabaena 25 filaments,” eLife 10, e64348 (2021)

    work page 2021

  27. [27]

    Stochastic turing patterns in a synthetic bacterial popu- lation,

    David Karig, K. Michael Martini, Ting Lu, Nicholas A. DeLateur, Nigel Golden- feld, and Ron Weiss, “Stochastic turing patterns in a synthetic bacterial popu- lation,” Proceedings of the National Academy of Sciences 115, 6572–6577 (2018), https://www.pnas.org/doi/pdf/10.1073/pnas.1720770115

    work page doi:10.1073/pnas.1720770115 2018

  28. [28]

    Robust stochastic turing patterns in the development of a one- dimensional cyanobacterial organism,

    Francesca Di Patti, Laura Lavacchi, Rinat Arbel-Goren, Leora Schein-Lubomirsky, Duccio Fanelli, and Joel Stavans, “Robust stochastic turing patterns in the development of a one- dimensional cyanobacterial organism,” PLoS Biology 16, e2004877 (2018)

    work page 2018

  29. [29]

    Stochastic turing pat- terns of trichomes in arabidopsis leaves,

    Francesca Di Patti, Yamel Ugartechea Chirino, Rinat Arbel-Goren, Tom Sharon, Aaron Castillo, Elena Alvarezˆ a-Buylla, Duccio Fanelli, and Joel Stavans, “Stochastic turing pat- terns of trichomes in arabidopsis leaves,” Proceedings of the National Academy of Sciences 120, e2309616120 (2023), https://www.pnas.org/doi/pdf/10.1073/pnas.2309616120

    work page doi:10.1073/pnas.2309616120 2023

  30. [30]

    Structural and practical identifiability analysis of partially observed dynamical models by exploiting the profile likelihood,

    Andreas Raue, Clemens Kreutz, Thomas Maiwald, Julie Bachmann, Marcel Schilling, Ursula Klingm¨ uller, and Jens Timmer, “Structural and practical identifiability analysis of partially observed dynamical models by exploiting the profile likelihood,” Bioinformatics 25, 1923–1929 (2009). 26

    work page 1923