pith. sign in

arxiv: 2505.05770 · v2 · submitted 2025-05-09 · 🧮 math.DS · math.PR

On the structure of complex spectra and eigenfunctions of transfer and Koopman operators

Matheus M Castro , Gary Froyland This is my paper

Pith reviewed 2026-05-22 17:13 UTC · model grok-4.3

classification 🧮 math.DS math.PR
keywords transfer operatorsKoopman operatorseigenspectrumeigenfunctionsrotational dynamicssmall noisezero-noise limitcyclic motion
0
0 comments X

The pith

In a canonical model of state-dependent rotational dynamics with small noise, the eigenfunctions of transfer operators localize on cycles of distinct periods.

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

The paper provides precise characterisations of the eigenspectrum and eigenfunctions of transfer operators for rotational dynamics where speed varies with position in state space. It examines the behaviour of these quantities as noise strength approaches zero, including explicit quadratic and linear responses. Theorems on the localisation of eigenfunction support directly produce algorithms that identify the existence and state-space positions of approximately cyclic motions having different periods. This structure clarifies how complex eigenvalues encode rotational features and remains useful for detection even when moderate noise is present. Numerical verification confirms that extracted periods and locations stay stable in the linear response regime.

Core claim

For a canonical model of state-dependent rotational dynamics under small additive noise, the transfer operator admits an eigenspectrum whose complex eigenvalues correspond to the underlying rotation rates; the associated eigenfunctions concentrate their support on the state-space regions where the local rotation period matches the eigenvalue argument. In the zero-noise limit these eigenvalues and eigenfunctions admit explicit quadratic and linear expansions. The resulting support-localisation theorems furnish simple algorithms that recover both the periods and the locations of approximately cyclic motion from the eigendata.

What carries the argument

Support localisation of the eigenfunctions of the transfer operator, which concentrates mass on regions of constant local rotation period.

If this is right

  • Periods of approximately cyclic motion are recovered from the arguments of the complex eigenvalues of the transfer operator.
  • Spatial locations of those cycles are recovered from the supports of the corresponding eigenfunctions.
  • Cycle information extracted from the eigendata remains insensitive to noise level inside the linear response regime.
  • The same localisation properties extend to the Koopman operator and clarify the structure of complex spectra in noisy rotational systems.

Where Pith is reading between the lines

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

  • The localisation mechanism may be used to design data-driven cycle detectors that operate directly on noisy time series without explicit model fitting.
  • Similar support concentration could appear in other operator-based analyses of coherent structures when rotation speed varies across state space.
  • The quadratic and linear response formulas supply explicit error bounds that could guide the choice of noise level in numerical approximations of Koopman operators.

Load-bearing premise

A canonical model with small noise captures the essential features of state-dependent rotational dynamics and the derived localisation mechanisms apply to other systems.

What would settle it

A numerical simulation of the canonical model in which the supports of the computed eigenfunctions fail to concentrate on the predicted constant-period regions as noise strength is reduced to zero would falsify the localisation claim.

Figures

Figures reproduced from arXiv: 2505.05770 by Gary Froyland, Matheus M Castro.

Figure 1
Figure 1. Figure 1: Left: Magnitude of the leading complex eigenvector of the normalised [PITH_FULL_IMAGE:figures/full_fig_p002_1.png] view at source ↗
Figure 2
Figure 2. Figure 2: Streamlines of the cylinder rotation model with three bands of circular [PITH_FULL_IMAGE:figures/full_fig_p003_2.png] view at source ↗
Figure 3
Figure 3. Figure 3: Left: The small coloured disks indicate the spectrum λ1,ε (for δ = ε = 0.1) of the 33 × 33 matrix Dk,αWε = Dk,β,LWε, associated to the model from [PITH_FULL_IMAGE:figures/full_fig_p004_3.png] view at source ↗
Figure 4
Figure 4. Figure 4: Details of leading eigenfunctions of Pε on M, constructed from the eigenvectors in [PITH_FULL_IMAGE:figures/full_fig_p027_4.png] view at source ↗
Figure 5
Figure 5. Figure 5: Left: The small coloured disks indicate the response of the spectrum λˆ (ℓ) 1 , ℓ = 1, . . . , 33 of the 33 × 33 matrix Pˆ 1,β,L, associated to the model from [PITH_FULL_IMAGE:figures/full_fig_p028_5.png] view at source ↗
read the original abstract

Complex eigenspectra of transfer and Koopman operators describe rotational motion in dynamical systems. A particularly relevant situation in applications is when the rotation speed depends on the state-space position of the dynamics. We consider a canonical model of such dynamics in the presence of small noise, and provide precise characterisations of the eigenspectrum and eigenfunctions of the corresponding transfer operators. Further, we study the limiting behaviour of the eigenspectrum and eigenfunctions in the zero-noise limit, including their quadratic and linear response. Our results clarify the structure of transfer and Koopman operator eigenspectra, and provide new interpretations relevant to applications. Our theorems on support localisation of the eigenfunctions yield simple algorithms to detect the existence and state-space location of approximately cyclic motion with distinct periods. Our numerical results verify that information on the cycle periods and their locations determined by the operator eigendata is insensitive to noise level in the linear response regime. We believe that the dynamic mechanisms underlying the eigendata and their properties apply rather broadly and enhance our understanding of approximate cycle detection in dynamical systems with operator methods.

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

1 major / 2 minor

Summary. The manuscript studies complex eigenspectra and eigenfunctions of transfer and Koopman operators for a canonical model of state-dependent rotational dynamics subject to small additive noise. It derives precise characterizations of the spectrum and eigenfunctions, analyzes their zero-noise limit including linear and quadratic response, and proves support-localization theorems for the eigenfunctions. These localization results yield simple algorithms for detecting the existence and state-space location of approximately cyclic motion with distinct periods. Numerical experiments show that period and location information extracted from the eigendata is insensitive to noise level in the linear-response regime. The authors conjecture that the underlying mechanisms apply broadly beyond the canonical model.

Significance. If the characterizations and localization theorems hold rigorously, the work clarifies the structure of complex spectra for transfer/Koopman operators in rotational systems with position-dependent speeds and supplies practical, noise-robust algorithms for cycle detection. The zero-noise limit analysis and explicit response results add theoretical value, while the numerical verification of noise insensitivity supports applicability in the linear regime. The broad-applicability claim, however, rests on the canonical model capturing essential features without direct comparisons to other noise structures or rotation laws.

major comments (1)
  1. [Abstract and §1] Abstract (final sentence) and §1: the assertion that the dynamic mechanisms 'apply rather broadly' is load-bearing for the claim that the support-localization theorems enhance cycle detection in general dynamical systems, yet no comparisons to non-additive noise, non-smooth rotation speeds, or higher-dimensional state dependence are provided; if the proofs rely on explicit solvability or uniform ellipticity of the chosen Fokker-Planck operator, the algorithms may not transfer.
minor comments (2)
  1. [§3] §3 (model definition): clarify whether the rotation speed function is assumed C^2 or merely continuous, as this affects the regularity of the eigenfunctions in the zero-noise limit.
  2. [Numerical section] Numerical section: report the discretization scheme for the transfer operator and the number of Monte-Carlo realizations used to estimate the linear-response regime; without these details the insensitivity claim is difficult to reproduce.

Simulated Author's Rebuttal

1 responses · 0 unresolved

We are grateful to the referee for their careful reading and constructive comments. We address the major comment below.

read point-by-point responses

  1. Referee: [Abstract and §1] Abstract (final sentence) and §1: the assertion that the dynamic mechanisms 'apply rather broadly' is load-bearing for the claim that the support-localization theorems enhance cycle detection in general dynamical systems, yet no comparisons to non-additive noise, non-smooth rotation speeds, or higher-dimensional state dependence are provided; if the proofs rely on explicit solvability or uniform ellipticity of the chosen Fokker-Planck operator, the algorithms may not transfer.

    Authors: We thank the referee for this observation. Our analysis is performed on a canonical model that allows for explicit solvability of the associated Fokker-Planck equation under additive noise, which indeed relies on the uniform ellipticity provided by the diffusion term. The support localization results are derived specifically for this model. We do not provide comparisons to other noise structures or rotation laws because the focus is on deriving precise characterizations and algorithms for this representative case. Nevertheless, the mechanisms identified, such as the concentration of eigenfunctions on cyclic regions in the zero-noise limit, arise from the general structure of transfer operators for rotational dynamics with position-dependent speeds and small noise. We will revise the abstract and Section 1 to qualify the statement on broad applicability, framing it as a conjecture based on the canonical model's ability to capture key features of state-dependent rotations. This will clarify that the algorithms are validated for the considered class of systems, with potential extensions left for future work. revision: partial

Circularity Check

0 steps flagged

Derivation is self-contained mathematical analysis with no circular reductions

full rationale

The paper derives characterizations of eigenspectra and eigenfunctions for transfer/Koopman operators on a specified canonical model with small noise, including zero-noise limits, quadratic/linear response, and support-localization theorems that yield detection algorithms. These results follow from direct analysis of the model's Fokker-Planck structure and explicit solvability properties rather than any fitted parameters, self-definitional constructs, or load-bearing self-citations. The abstract and theorems treat the canonical model as the object of study; broad applicability is stated as a belief without being used to close any derivation loop. No step reduces an output to an input by construction.

Axiom & Free-Parameter Ledger

0 free parameters · 2 axioms · 0 invented entities

The central claim rests on the existence and suitability of a canonical model for state-dependent rotational dynamics with small noise, plus standard assumptions from dynamical systems and operator theory that allow precise spectral analysis.

axioms (2)
  • domain assumption A canonical model exists that captures state-dependent rotation speeds in the presence of small noise.
    Invoked at the start of the abstract to enable the characterisations of eigenspectrum and eigenfunctions.
  • domain assumption The dynamic mechanisms underlying the eigendata apply rather broadly beyond the specific model.
    Stated in the final sentence of the abstract as a belief supporting application relevance.

pith-pipeline@v0.9.0 · 5717 in / 1480 out tokens · 65158 ms · 2026-05-22T17:13:03.504350+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

25 extracted references · 25 canonical work pages

  1. [1]

    Alves, V´ ıtor Ara´ ujo, and Carlos H

    Jos´ e F. Alves, V´ ıtor Ara´ ujo, and Carlos H. V´ asquez. Stochastic stability of non-uniformly hyperbolic diffeomorphisms. Stoch. Dyn., 7(3):299–333, 2007

    work page 2007

  2. [2]

    Bernat Bassols Cornudella, Matheus Manzatto de Castro, and Jeroen S. W. Lamb. Conditioned stochastic stability of equilibrium states on uniformly expanding repellers, arXiv:2405.01343, 2024

    work page arXiv 2024

  3. [3]

    Fourier approximation of the statistical properties of Anosov maps on tori

    Harry Crimmins and Gary Froyland. Fourier approximation of the statistical properties of Anosov maps on tori. Nonlinearity, 33(11):6244, 2020

    work page 2020

  4. [4]

    Set oriented numerical methods in space mission design

    Michael Dellnilz and Oliver Junge. Set oriented numerical methods in space mission design. In Elsevier Astrodynamics Series, volume 1, pages 127–IV. 2006

    work page 2006

  5. [5]

    On the isolated spectrum of the Perron-Frobenius operator

    Michael Dellnitz, Gary Froyland, and Stefan Sertl. On the isolated spectrum of the Perron-Frobenius operator. Nonlinearity, 13(4):1171, 2000

    work page 2000

  6. [6]

    On the approximation of complicated dynamical behavior

    Michael Dellnitz and Oliver Junge. On the approximation of complicated dynamical behavior. SIAM J. Numer. Anal. , 36(2):491–515, 1999

    work page 1999

  7. [7]

    Springer, 2015

    Tanja Eisner, B´ alint Farkas, Markus Haase, and Rainer Nagel.Operator theoretic aspects of ergodic theory, volume 272. Springer, 2015

    work page 2015

  8. [8]

    Characterization of the Lorenz attractor by unstable periodic orbits

    Valter Franceschini, Claudio Giberti, and Zhi Ming Zheng. Characterization of the Lorenz attractor by unstable periodic orbits. Nonlinearity, 6(2):251–258, 1993

    work page 1993

  9. [9]

    Spectral analysis of climate dynamics with operator-theoretic approaches

    Gary Froyland, Dimitrios Giannakis, Benjamin R Lintner, Maxwell Pike, and Joanna Slawinska. Spectral analysis of climate dynamics with operator-theoretic approaches. Nature communications, 12(1):6570, 2021

    work page 2021

  10. [10]

    Revealing trends and persistent cycles of non-autonomous systems with autonomous operator-theoretic techniques.Nature Communications, 15(1):4268, 2024

    Gary Froyland, Dimitrios Giannakis, Edoardo Luna, and Joanna Slawinska. Revealing trends and persistent cycles of non-autonomous systems with autonomous operator-theoretic techniques.Nature Communications, 15(1):4268, 2024

    work page 2024

  11. [11]

    Almost-invariant sets and invariant manifolds—connecting probabilistic and geometric descriptions of coherent structures in flows

    Gary Froyland and Kathrin Padberg. Almost-invariant sets and invariant manifolds—connecting probabilistic and geometric descriptions of coherent structures in flows. Phys. D, 238(16):1507–1523, 2009

    work page 2009

  12. [12]

    Detection of coherent oceanic structures via transfer operators

    Gary Froyland, Kathrin Padberg, Matthew H England, and Anne Marie Treguier. Detection of coherent oceanic structures via transfer operators. Physical review letters, 98(22):224503, 2007

    work page 2007

  13. [13]

    How well-connected is the surface of the global ocean? Chaos: An Interdisciplinary Journal of Nonlinear Science , 24(3), 2014

    Gary Froyland, Robyn M Stuart, and Erik van Sebille. How well-connected is the surface of the global ocean? Chaos: An Interdisciplinary Journal of Nonlinear Science , 24(3), 2014

    work page 2014

  14. [14]

    Horn and Charles R

    Roger A. Horn and Charles R. Johnson. Matrix analysis. Cambridge University Press, Cambridge, second edition, 2013

    work page 2013

  15. [15]

    Perturbation theory for linear operators

    Tosio Kato. Perturbation theory for linear operators . Classics in Mathematics. Springer-Verlag, Berlin, 1995. Reprint of the 1980 edition. 29

    work page 1995

  16. [16]

    Stability of the spectrum for transfer operators

    Gerhard Keller and Carlangelo Liverani. Stability of the spectrum for transfer operators. Annali della Scuola Normale Superiore di Pisa-Classe di Scienze , 28(1):141–152, 1999

    work page 1999

  17. [17]

    Stochastic stability of differential equations , volume 66 of Stochastic Modelling and Applied Probability

    Rafail Khasminskii. Stochastic stability of differential equations , volume 66 of Stochastic Modelling and Applied Probability . Springer, Heidelberg, second edition, 2012. With contributions by G. N. Milstein and M. B. Nevelson

    work page 2012

  18. [18]

    Some theorems on small random perturbations of dynamical systems

    Yuri Kifer. Some theorems on small random perturbations of dynamical systems. Uspehi Mat. Nauk, 29(3(177)):205–206, 1974

    work page 1974

  19. [19]

    John M. Lee. Introduction to smooth manifolds , volume 218 of Graduate Texts in Mathematics . Springer, New York, second edition, 2013

    work page 2013

  20. [20]

    Comparison of systems with complex behavior.Physica D: Non- linear Phenomena, 197(1-2):101–133, 2004

    Igor Mezi´ c and Andrzej Banaszuk. Comparison of systems with complex behavior.Physica D: Non- linear Phenomena, 197(1-2):101–133, 2004

    work page 2004

  21. [21]

    Lagrangian geography of the deep gulf of mexico

    Philippe Miron, Francisco J Beron-Vera, Maria J Olascoaga, Gary Froyland, P P´ erez-Brunius, and J Sheinbaum. Lagrangian geography of the deep gulf of mexico. Journal of physical oceanography, 49(1):269–290, 2019

    work page 2019

  22. [22]

    Spectral analysis of nonlinear flows

    Clarence W Rowley, Igor Mezi´ c, Shervin Bagheri, Philipp Schlatter, and Dan S Henningson. Spectral analysis of nonlinear flows. Journal of Fluid Mechanics , 641:115–127, 2009

    work page 2009

  23. [23]

    Springer, 2001

    Ch Sch¨ utte, Wilhelm Huisinga, and Peter Deuflhard.Transfer operator approach to conformational dynamics in biomolecular systems . Springer, 2001

    work page 2001

  24. [24]

    Stochastic stability of hyperbolic attractors

    Lai-Sang Young. Stochastic stability of hyperbolic attractors. Ergodic Theory Dynam. Systems , 6(2):311–319, 1986

    work page 1986

  25. [25]

    Eigenvalues of several tridiagonal matrices

    Wen-Chyuan Yueh. Eigenvalues of several tridiagonal matrices. Appl. Math. E-Notes, 5:66–74, 2005. 30

    work page 2005