On the timescale at which statistical stability breaks down (Q2664465)

This paper focuses primarily on the timescale at which statistical stability of dynamical systems breaks down. The first few pages introduce all the necessary notation and terminology to understand the paper: some definitions of structural stability, statistical stability, stochastic stability, and the breakdown of statistical stability. The main results are presented in Section 2. In Section 3, the authors consider some preliminary results through known lemmas and definitions. Results in Section 4 are based on the assumption that \(\left\lbrace f_t\right\rbrace\) is a Misiurewicz-rooted unimodal family. Section 5 is about breakdown of statistical stability. Persistence of statistical stability is presented in Section 6. Roughly speaking, the strategy is as follows: (1) A special inducing scheme is constructed to approximate the one-parameter family of (discrete-time) maps \(f_t\) with a nonuniformly expanding map \(\hat{f}_{t}\) which admits an absolutely continuous invariant probability measure \(\hat{\mu}_{t}\); (2) \(f_{t}\) agrees with \(\hat{f}_{t}\) on time horizons smaller than \(t^{-1}\), namely if \(n(t) = o(t^{-1})\). One of the main difficulties in the proof is the construction of the approximating map \(\hat{f}_{t}\), which allows a suitable inducing scheme and which coincides with \(f_{t}\) everywhere except on a set of Lebesgue measure of order \(t\).