Pathwise convergence of a rescaled super-Brownian catalyst reactant process (Q867080)

Let \(\rho\) be a super-Brownian motion in \(\mathbb{R}^d\) and \(X\) be the process mentioned in the title of the paper. Loosely speaking, given a realization of \(\rho\), the infinitesimal \(X\)-particles undergo Brownian motion and critical binary branching with the local branching rate proportional to the local density of the catalyst process \(\rho\). The reactant process \(X\) can be determined rigorously by its log-Laplace functionals, see [\textit{D. A. Dawson} and \textit{K. Fleischmann}, J. Theor. Probab. 10, No.~1, 213--276 (1997; Zbl 0877.60030)]. The first aim of the authors is to characterize \(X\) in terms of a martingale problem and, for \(d=1\), by an SPDE. The main goal of the paper is to study, for \(d=1\), the longtime behavior of the process \((\rho,X)\) via a mass-time-space rescaling procedure. For a fixed scaling index \(\eta\geq 0\) and a measure-valued path \(\mu=\{\mu_t, t\geq 0\}\) one passes to \(K^{-\eta}\mu_{Kt}(K^{\eta}B)\) where \(B\) is a Borel subset of \(\mathbb{R}\), \(K\geq 1\) and \(t\geq 0\). The principle point is to show that convergence happens as \(K\to \infty\) not only for finite-dimensional distributions but also in appropriate path space. While the catalyst's limiting process exhibits a phase transition at \(\eta=1\), the reactant's limit is always the same degenerate process.