Global attractors for 3-dimensional stochastic Navier-Stokes equations

Sell''s approach 35 to the construction of attractors for the Navier-Stokes equations in 3-dimensions is extended to the 3D stochastic equations with a general multiplicative noise. The new notion of a process attractor is defined as a set A of processes, living on a single filtered probability space, that is a set of solutions and attracts all solution processes in a given class. This requires the richness of a Loeb probability space. Non-compactness results for A and a characterization in terms of two-sided solutions are given.