The Navier-Stokes equations describe how fluids like water and air move, but mathematicians still struggle with basic questions about whether these equations always have unique, well-behaved solutions. One approach to finding potentially problematic solutions involves looking for "self-similar" ones, where the flow pattern looks the same at different scales, like a fractal. This paper focuses on a special subclass called "homothetic" self-similar solutions, where the solution can be rescaled in a particularly rigid way. These are especially important because they represent the only known route for a specific mathematical strategy, called the Jia-Sverak program, that might prove non-uniqueness of physically reasonable (Leray-Hopf) solutions to the equations.
In three dimensions, the authors prove a Liouville-type theorem, which is essentially a result that says "the only solution with certain good properties is the trivial one." Specifically, they show that no non-trivial homothetic solutions exist when the initial fluid configuration is sufficiently regular. This is significant because it closes off a potential avenue for constructing counterexamples to uniqueness in 3D using this particular homothetic approach. In two dimensions, a parallel result holds: the only well-behaved homothetic solution that decays at large distances is the Oseen vortex, a well-known swirling flow pattern.
Beyond just ruling things out, the paper also analyzes the stability of the Oseen vortex by studying a linearized version of the equations around it. They prove that this linearized operator is stable, meaning small perturbations do not grow uncontrollably. However, they also uncover a subtlety: there exists a homothetic solution near which an "approximate eigenvalue" of the linearized operator is unstable. This tension between the stability result and this approximate instability suggests that the Oseen vortex sits in a delicate mathematical landscape, and the full story of uniqueness for 2D fluid flows remains nuanced.