Ergodic Poisson splittings (Q784168)

By definition, a thinning of a point process consists in removing points according to some rule, whereas the related notion of splitting operation is a decomposition of a point process into a sum of a number of other point processes. It is shown in [\textit{K. Ball}, Electron. Commun. Probab. 10, 60--69 (2005; Zbl 1110.60050)] that it is possible to deterministically choose points from a homogeneous Poisson point process on \(\mathbb{R}\) to get another homogeneous Poisson point process of lower intensity. Moreover, it is possible to proceed in a translation equivariant way. This result has been further refined in [\textit{A. E. Holroyd} et al., Ann. Probab. 39, No. 5, 1938--1982 (2011; Zbl 1277.60087)] by extending it to \(\mathbb{R}^d\) and replacing translation equivariance by isometry equivariance. Moreover, the remaining points were also shown to form a homogeneous Poisson point process. The isometry or translation equivariance plays a key role here as Meyerovitch showed in [\textit{T. Meyerovitch}, Ann. Probab. 41, No. 5, 3181--3200 (2013; Zbl 1279.60061)] that it is not possible to deterministically thin a Poisson point process in an equivariant way with respect to a transformation which is conservative ergodic on the base space. In the present paper, the authors are also interested in thinnings and splittings of Poisson point processes which are equivariant under some dynamics. The difference with [Ball, loc. cit.] and [Holroyd, loc. cit.] is that the present paper considers thinnings and splittings which are equivariant with respect to a conservative transformation. Moreover, contrary to the above mention result of [Meyerovitch, loc. cit.], the authors allow additional randomness in the splitting procedure. The main result of the paper is the following: If \(T\) is an infinite measure preserving map with infinite ergodic index (that is, all its Cartesian products are ergodic), then any ergodic \(T\)-splitting of a Poisson \(T\)-point process yields independent Poisson \(T\)-point processes (see Section 2). Section 3 proves that for \(T\) with infinite ergodic index, the only way to get an ergodic marked \(T\)-point process out of a Poisson \(T\)-point process is to take i.i.d. marks, independent of the underlying process. The authors of the present paper in [the first author et al., Ann. Sci. Éc. Norm. Supér. (4) 50, No. 6, 1301--1334 (2017; Zbl 1382.37006)] gave conditions on \(T\) under which an ergodic \(T\)-point process with moments of all orders is necessarily a cluster-Poisson process, best described as an independent superposition of shifted Poisson processes (a so-called SuShi). Section 4 extends result in [the first author, loc. cit.] to general random measures. This more general framework allows to simplify and improve some disjointness results from [the first author, loc. cit.] in the last section.