The paper studies fluid flows governed by the two-dimensional Navier-Stokes equations, but with random noise added to the system. These equations describe how fluids like air or water move, and the random forcing models turbulent or unpredictable inputs. The authors focus on how tiny disturbances or tracer particles evolve when carried along by such a randomly driven fluid. Specifically, they look at two related mathematical objects: one describes how small perturbations to the fluid velocity itself grow or shrink over time, and the other describes how a passive substance (like a dye or a pollutant that does not affect the flow) spreads through the fluid.
The central result is that each of these "linearized" processes has a unique long-run statistical behavior, no matter where you start. In technical terms, there is a unique stationary measure for what is called the projective process, which tracks the direction of a disturbance vector as it evolves, ignoring its size. Uniqueness of this stationary measure is important because it underpins rigorous results about Lyapunov exponents, which quantify the average rate at which small disturbances grow or decay, and ultimately connect to questions about mixing and predictability in fluid systems.
The proof combines two main ingredients. The first is a coupling argument, which shows that two copies of the system started from different initial conditions can be forced to behave very similarly after a long time. The second is a condition called non-degeneracy of the Malliavin matrix, which is a way of verifying that the random noise genuinely spreads its influence throughout the relevant directions of the system. For the passive scalar case this second step is harder, and the authors handle it by showing, using tools from complex analysis, that generically the passive scalar field cannot be described in a dimensionally reduced way. Together these arguments establish the uniqueness result rigorously.