Asymptotic properties of a continuous-space discrete-time population model in a random environment (Q2639006)

Biological populations in randomly varying environments have been modelled by use of stochastic difference equations of the form \[ (1)\quad X_{t+1}=H_{\alpha_ t}(X_ t),\quad t=0,1,2,... \] where \(X_ t\) is the population density of the t-th generation and \(\alpha_ t\) the random environmental condition. In the event that the random variable \(X_ t\) is real valued, the asymptotic properties of \(X_ t\) have been determined for a large class of biologically interesting maps \(H_{\alpha}\) [see \textit{S. Ellner}, ibid. 19, 169-200 (1984; Zbl 0536.92022), and references therein]. In many applications (e.g. in studies of dispersal strategies) \(X_ t\) is a random vector in \(({\mathbb{R}}^ k)_+\) (e.g. patch models) or more generally a random function (e.g. continuous space models). We extend some of the results of Ellner (1984) to the case where \(X_ t\) takes values in \(C_+(\Omega)\), the nonnegative continuous functions on \(\Omega\) where \(\Omega\) is some compact region in \({\mathbb{R}}^ 2\). The case where \(X_ t\) is a random vector then follows as special case.