Strong Borel--Cantelli Lemmas for Recurrence
Tomas Persson, Alejandro Rodriguez Sponheimer
TL;DR
This work advances dynamical recurrence theory by proving a strong Borel--Cantelli lemma with an explicit error term for recurrence events in systems with exponential 3-fold decay of correlations and a short return time condition. The authors develop a general RSBC framework for arbitrary target sets and then specialize it to recurrence targets $E_k=\{x: T^k x\in B(x,r_k(x))\}$, yielding a sharp asymptotic with error $O(\Phi(n)^{1/2}(\log \Phi(n))^{3/2+\varepsilon})$ where $\Phi(n)=\sum_{k=1}^n\mu(B_k(x))$. They prove a key large-$l$ bound ( Lemma \ref{['lem:largel']} ) and employ the Gal--Koksma lemma to convert variance control into almost-sure convergence. Applications include piecewise expanding interval maps with Gibbs measures and linear Anosov maps on the torus preserving Lebesgue measure, thereby enlarging the class of systems for which RSBC holds and accommodating critical decay rates up to and including $1/k$. The results connect recurrence rates to the underlying measure structure and correlations, with potential implications for hitting times and pointwise dimension in chaotic dynamics.
Abstract
Let $(X,T,μ,d)$ be a metric measure-preserving system for which $3$-fold correlations decay exponentially for Lipschitz continuous observables. Suppose that $(M_k)$ is a sequence satisfying some weak decay conditions and suppose there exist open balls $B_k(x)$ around $x$ such that $μ(B_k(x)) = M_k$. Under a short return time assumption, we prove a strong Borel--Cantelli lemma, including an error term, for recurrence, i.e., for $μ$-a.e. $x \in X$, \[ \sum_{k=1}^{n} \mathbf{1}_{B_k(x)} (T^k x) = Φ(n) + O \bigl( Φ(n)^{1/2} (\log Φ(n))^{3/2 + \varepsilon} \bigr), \] where $Φ(n) = \sum_{k=1}^{n} μ(B_k(x))$. Applications to systems include some non-linear piecewise expanding interval maps and hyperbolic automorphisms of $\mathbf{T}^2$.
