betaFeedback

WorksMathematics

A limit theorem for deviation of ergodic integrals on a class of flows on the two-torus

Abstract

We study the fluctuations of ergodic integrals for locally Hamiltonian flows on the two-torus with a single non-degenerate saddle (Arnol'd flows). We sample a flow from a suitably random family, together with a base point pp chosen uniformly at random on that flow's domain. For a smooth observable ff, we show that the centered Birkhoff integral ∫0tf(ϕsp) ds−tfˉ\int_{0}^t f(ϕ_s p)\,ds-t\bar f, where fˉ\bar f is the spatial mean of ff, normalized by (log⁡t)2(\log t)^2, converges in distribution to the product of two independent random variables. Using special representation of locally Hamiltonian flows, we reduce the problem to studying ergodic sums of functions with logarithmic singularities over circle rotations. Then, we prove a Cauchy limit theorem for such ergodic sums, which is a result of independent interest.

Introduction and main results

Arnol'd flows with a single saddle

Smooth area-preserving flows on surfaces are an extensively studied class of dynamical systems. They are low-dimensional examples of systems with non-trivial ergodic, mixing, and spectral properties.

Let M′\mathcal{M}' be a compact, connected, oriented smooth surface equipped with a smooth non-degenerate area form ω\omega. Any smooth closed 1-form λ\lambda on M′\mathcal{M}' determines a unique smooth vector field W:M′→TM′W:\mathcal{M}' \to T\mathcal{M}', given by ıW(ω)=λ\imath_W(\omega)=\lambda, where ı\imath denotes the contraction operator. The vector field WW (or equivalently λ\lambda) generates an ω\omega-preserving flow φR=(φt)t∈R\varphi_{\mathbb{R}}=(\varphi_t)_{t\in\mathbb{R}}. These are known as locally Hamiltonian flows. Indeed, locally there exists a smooth real-valued function HH such that λ=dH\lambda=\mathrm{d}H. In local coordinates (x,y)(x,y) for which ω=dx∧dy\omega=\mathrm{d}x\wedge\mathrm{d}y, the vector field reads W=x˙∂x+y˙∂yW=\dot{x}\partial_x+\dot{y}\partial_y where x˙=∂H∂y\dot{x}=\frac{\partial H}{\partial y}, y˙=−∂H∂x\dot{y}=-\frac{\partial H}{\partial x}.

Arnol'd flows, first studied by Arnold in [1], are among the most fundamental examples of locally Hamiltonian flows. In this setting, the underlying manifold is the two-torus T2\mathbb{T}^2, and we assume that WW has exactly two zeros: a non-degenerate (det⁡DW≠0\det DW\ne0) saddle and a center lying inside a homoclinic loop homologous to zero (see Figure 1). This loop divides T2\mathbb{T}^2 into two φR\varphi_{\mathbb{R}}-invariant components: the one bounded by the saddle loop is foliated by periodic orbits, and we denote the closure of its complement by M\mathcal{M}. We refer to the flow (φt∣M)t∈R(\varphi_t|_{\mathcal{M}})_{t\in\mathbb{R}} as an Arnol'd flow, and we also denote it by φR\varphi_{\mathbb{R}}.

An Arnol'd flow in $\mathbb{T}^2$

Figure 1. An Arnol'd flow in T2\mathbb{T}^2.

Arnol'd flows are mixing for every irrational frequency (see § 1.3). This was first shown by Sinai–Khanin [24] for Lebesgue-almost every irrational frequency and was later improved by Kočergin [Koc3; Koc4] to all irrational frequencies.

In particular, for such flows we have equidistribution of typical trajectories: for any integrable function ff, the Birkhoff integrals satisfy 1t∫0tf(ϕsp) ds→1ω(M)∫Mf dω\frac{1}{t}\int_{0}^{t} f(\phi_s p)\,\mathrm{d}s \to\frac{1}{\omega(\mathcal{M})}\int_{\mathcal{M}} f\,\mathrm{d}\omega pointwise almost-surely as t→∞t \to\infty, where ω(M)\omega(\mathcal{M}) is the area of M\mathcal{M}. This raises the question of the nature of its fluctuations: how large are the deviations of the ergodic averages and how are they asymptotically distributed?

More precisely, we study the centered Birkhoff integral

∫0tf(ϕsp) ds−tω(M)∫Mf dω,(1)\int_{0}^{t} f(\phi_s p)\,\mathrm{d}s-\frac{t}{\omega(\mathcal{M})}\int_{\mathcal{M}} f\,\mathrm{d}\omega, \tag*{(1)}

as t→∞t \to\infty. We show that, provided ff is smooth enough and the initial point pp is chosen uniformly in M\mathcal{M} (with respect to ω\omega), the appropriate scaling of the fluctuations is (log⁡t)2(\log t)^2, for suitable random families of flows (made precise in Definition 1.2). After this normalization, (1.1) converges in distribution to the product of a Cauchy random variable and a random variable quantifying the difference between the spatial average of ff and its value at the saddle.

As will be shown, this question is tightly related to limit theorems for ergodic sums over rotations, in particular, for functions with (asymmetric) logarithmic singularities.

Supension flows

Let TT be a measure-preserving endomorphism on an interval [a,b)[a,b) and let φ:[a,b)→R+∪{∞}\varphi:[a,b)\to\mathbb{R}_{+}\cup\{\infty\}, satisfying inf⁡x∈[a,b)φ(x)>0\inf_{x\in[a,b)}\varphi(x)>0. The suspension flow (Tt)t∈R(T_t)_{t\in\mathbb{R}} over TT built under the roof function φ\varphi is a flow defined on the space

X={(x,s)∈[a,b)×R:0≤s<φ(x)}X=\{(x,s)\in[a,b)\times\mathbb{R}:0\le s<\varphi(x)\}

with the identification (x,φ(x))∼(Tx,0)(x,\varphi(x))\sim(Tx,0) at the points where φ(x)<∞\varphi(x)<\infty, and acts by vertical translation

Tt(x,s)=(x,s+t)T_t(x,s)=(x,s+t)

followed by the canonical identifications whenever the second coordinate exits the interval [0,φ(x))[0,\varphi(x)).

Let T\mathbb{T} denote the torus, which we identify with [0,1)[0,1). For every α∈R\alpha\in\mathbb{R}, the rotation map Rα:T→TR_{\alpha}:\mathbb{T}\to\mathbb{T} is defined by Rαx=x+α mod⁡1R_{\alpha}x=x+\alpha\ \operatorname{mod} 1 for every x∈Tx\in\mathbb{T}.

Arnol'd flows as suspension flows

Consider T2\mathbb{T}^{2} equipped with a smooth non-degenerate area form ω\omega. Let ϕR\phi_{\mathrm{R}} be an Arnol'd flow on M⊂T2\mathcal{M}\subset\mathbb{T}^{2} generated by a smooth vector field WW. It is a classical result (see [23, 20] and the references therein) that ϕR\phi_{\mathrm{R}} is measure-theoretically isomorphic to the suspension flow (Tt)(T_t) over a rotation RαR_{\alpha} on T=[0,1)\mathbb{T}=[0,1), with roof function φ\varphi with logarithmic singularities and inf⁡Tφ>0\inf_{\mathbb{T}}\varphi>0; via a map Φ:X→M\Phi:X\to\mathcal{M} that is measure-preserving up to a linear rescaling of the xx-coordinate.1

We call Σ:=Φ(T×{0})⊂M\Sigma:=\Phi(\mathbb{T}\times\{0\})\subset\mathcal{M} the base transversal section. Then, Φ(x,⋅)\Phi(x,\cdot) denotes the orbit under ϕR\phi_{\mathrm{R}} of Φ(x,0)∈Σ\Phi(x,0)\in\Sigma, and φ(x)\varphi(x) is the return time of Φ(x,0)\Phi(x,0) to Σ\Sigma.

Let p0p_0 be the (non-degenerate) saddle of ϕR\phi_{\mathrm{R}}. The rotation parameter α\alpha can be computed directly from the 1-form λ\lambda corresponding to the flow. Indeed, let z0z_0 and z0′z'_0 denote the first intersections of the stable and unstable separatrices of the saddle p0p_0 with Σ\Sigma, respectively. Then the rotation parameter is given by (see § 2.1)

α=(∫Σλ)−1∫z0z0′λ.(2)\alpha=\left(\int_{\Sigma}\lambda\right)^{-1}\int_{z_0}^{z'_0}\lambda. \tag*{(2)}

Roof function and passage constant

Here, we present some formulas that will be useful for the statement of the main results. We refer the reader to § 2.1 for further exposition and proofs.

Let T(c)T(c) be the passage time around a fixed neighborhood of the saddle p0p_0 at Hamiltonian energy level cc–close. Define the passage constant (see § 2.1)

c0=lim⁡c→0T(c)∣log⁡c∣=1−det⁡DW(p0).(3)c_0 = \lim_{c \to0} \frac{T(c)}{|\log c|} = \frac{1}{\sqrt{-\det DW(p_0)}}. \tag*{(3)}

Note that c0c_0 depends only on det⁡DW(p0)\det DW(p_0).

The parameter in (1.3) determines the return time to Σ\Sigma of Φ(x,0)\Phi(x,0) up to a bounded variation error. Indeed, the return time function is given by

φ(x)=ψ(x)+2c0∣log⁡x∣+c0∣log⁡(1−x)∣,(4)\varphi(x) = \psi(x) + 2c_0|\log x| + c_0|\log(1-x)|, \tag*{(4)}

where ψ\psi is of bounded variation, and the choice between c0c_0 and 2c02c_0 depends on whether the orbit passes once or twice near p0p_0 before returning to the Poincaré section Σ\Sigma, the latter corresponding to the side of the loop homologous to zero (see Figure 1).

Remark 1.1. Given that a vector field WW generates an Arnold flow on T2\mathbb{T}^2, all ϕt\phi_t, p0p_0, M\mathcal{M}, α\alpha, φ\varphi, ψ\psi, c0c_0, TtT_t, XX, Φ\Phi, and Σ\Sigma may depend on WW. We denote this dependence by ⋅W\cdot^W.

Main result

Let F\mathcal{F} be a family of vector fields on T2\mathbb{T}^2 and let η^\hat{\eta} be a probability measure on F\mathcal{F}.

Definition 1.2. We say that (F,η^)(\mathcal{F},\hat{\eta}) is good if

a) Every W∈FW \in\mathcal{F} generates an Arnold flow on T2\mathbb{T}^2.

b) Under the identification via ΦW\Phi^W of (MW,ϕtW)(\mathcal{M}^W,\phi_t^W) and (XW,TtW)(X^W,T_t^W) for each WW described in Section 1.3, let α(W)\alpha(W) be the rotation number of the base map RαR_\alpha (given by (1.2)). Then as W∼η^W \sim\hat{\eta}, A=α(W)\mathcal{A}=\alpha(W) has an absolutely continuous distribution with respect to Lebesgue(T\mathbb{T}).

Let ϵ>0\epsilon> 0 and f∈C2+ϵ(T2)f \in C^{2+\epsilon}(\mathbb{T}^2). For each W∈FW \in\mathcal{F}, define the mean of ff in MW\mathcal{M}^W as

f‾W:=1ω(MW)∫MWf(p) dp.(5)\overline{f}^W := \frac{1}{\omega(\mathcal{M}^W)} \int_{\mathcal{M}^W} f(p)\,dp. \tag*{(5)}

We want to study the scale and the limiting distribution as t→∞t \to\infty of the deviations from the mean of the Birkhoff integral

∫0tf(ϕsWp) ds\int_0^t f(\phi_s^W p)\,ds

References

  1. [1]V. I. Arnol’d. “Topological and ergodic properties of closed 1-forms with incommensurable periods”. In: Funct. Anal. Appl. 25.2 (1991), pp. 81–90.DOI
  2. [2]C. Aistleitner, N. Tech­nau, and A. Zafeiropoulos. “On the order of magnitude of Sudler products”. In: Amer. Jour. Math. 145.3 (2023), pp. 721–764.arxiv.org/abs/2002.06602
  3. [3]J. Beck. “Probabilistic Diophantine approximation. I. Kronecker sequences”. In: Ann. of Math. 140.2 (1994), pp. 449–502.
  4. [4]B. Borda, L. Frühwirth, and M. Hauke. “Limit laws for cotangent and Diophantine sums”. In: arXiv preprint arXiv:2308.12085 (2023). Revised 2025.arxiv.org/abs/2308.12085
  5. [5]D. Dolgopyat and B. Fayad. “Deviations of Ergodic Sums for Toral Translations I: Convex Bodies”. In: Geom. Funct. Anal. 24.1 (2014), pp. 85–115.arxiv.org/abs/1206.4853
  6. [6]D. Dolgopyat and B. Fayad. “Deviations of ergodic sums for toral translations II. Boxes”. In: Publ. Math. IHÉS 132 (2020), pp. 293–352.arxiv.org/abs/1211.4323
  7. [7]D. Dolgopyat and O. Sarig. “Quenched and Annealed Temporal Limit Theorems for Circle Rotations”. In: Ann. Prob. (2019).DOI
  8. [8]K. Frączek and M. Kim. “New Phenomena in Deviation of Birkhoff Integrals for Locally Hamiltonian Flows”. In: Ergod. Theory Dyn. Syst. (2024).arxiv.org/abs/2112.13030
  9. [9]G. Forni. “Deviation of Ergodic Averages for Area-Preserving Flows on Surfaces of Higher Genus”. In: Ann. Math. 155.1 (2002), pp. 1–103.
  10. [10]K. Frączek and C. Ulcigrai. “Ergodic properties of infinite extensions of area-preserving flows”. In: Proc. Lond. Math. Soc. 101.2 (2010), pp. 305–335.arxiv.org/abs/1102.5358
  11. [11]K. Fraczek and C. Ulcigrai. “On the Asymptotic Growth of Birkhoff Integrals for Locally Hamiltonian Flows and Ergodicity of Their Extensions”. In: Comment. Math. Helv. 99.2 (2024), pp. 231–354.
  12. [12]M.-R. Herman. “Sur la conjugaison différentiable des difféomorphismes du cercle à des rotations”. In: Publ. Math. IHÉS 49 (1979), pp. 5–233.DOI
  13. [13]A. B. Katok. “Invariant measures of flows on oriented surfaces”. In: Sov. Math. Dokl. 14 (1973), pp. 1104–1108.
  14. [14]Y. Katznelson. An Introduction to Harmonic Analysis. 3rd. Cambridge University Press, 2004.
  15. [15]H. Kesten. “Uniform distribution mod 1”. In: Ann. of Math. 71 (1960), pp. 445–471.DOI
  16. [16]H. Kesten. “Uniform distribution mod 1. II”. In: Acta Arith. 7 (1962), pp. 355–380.DOI
  17. [17]A. Y. Khinchin. Continued Fractions. Chicago: University of Chicago Press, 1964.
  18. [20]A. Kanigowski, A. Okunev, and R. Zelada. “Weak mixing for area preserving flows on surfaces”. In: arXiv preprint arXiv:2602.15719 (2026).arxiv.org/abs/2602.15719
  19. [21]O. Knill and F. Tangerman. “Self-similarity and growth in Birkhoff sums for the golden rotation”. In: Nonlinearity 24.11 (2011), pp. 3115–3127.DOI
  20. [22]I. Nikolaev and E. Zhuzhoma. Flows on 2-dimensional Manifolds: An Overview. Vol. 1705. Lecture Notes in Mathematics. Berlin, Heidelberg: Springer-Verlag, 1999.DOI
  21. [23]D. Ravotti. “Quantitative mixing for locally Hamiltonian flows with saddle loops on compact surfaces”. In: Ann. H. Poincaré 18 (2017), pp. 3815–3861.DOI
  22. [24]Ya. G. Sinai and K. M. Khanin. “Mixing for some classes of special flows over rotations of the circle”. In: Funct. Anal. Appl. 26.3 (1992), pp. 155–169.
  23. [25]Y. G. Sinai and C. Ulcigrai. “A limit theorem for Birkhoff sums of non-integrable functions over rotations”. In: Geometric and Probabilistic Structures in Dynamics. Vol. 469. Contemp. Math. Providence, RI: Amer. Math. Soc., 2008, pp. 317–340.

Paper details

vlnc:W_emjr54gr
Version
v1 · latest
Published
Fresh — added to the Hub within the last 30 days
Added
License
CC BY 4.0
Usage

views · 0

downloads · 0

A view counts once per browser per day, a download each time the PDF is requested.

  • Mathematics

AI

View statement
Looking for related papers…
Contents