Mixing and Small-Scale Formation in a Passive Divergence-Free Vector Field

Abstract

We study mixing for a divergence-free passive vector field $u$ transported by another divergence-free vector field $U$, where $u$ evolves according to $ \partial_t u + (U \cdot \nabla) u + \nabla p = 0.$ In recent years, a lot of attention has been given to the question of optimal mixing in the scalar case, where there is a Sobolev constraint on the advecting velocity. In the vector setting considered here, however, the pressure term introduces substantial difficulties, since the simple Lagrangian perspective available in the scalar case is no longer applicable. In this paper, we investigate mixing on a torus $\mathbb{T}^d$ under the assumption that the field $U$ satisfies $ \|U(t)\|_{W^{1,q}} \leq C $ and we quantify mixing through the decay of the homogeneous $ H^{-\alpha}$ norm of $u$. We start with establishing conditions on $U$ that guarantee existence and uniqueness of solutions. We then derive lower bounds on the mixing rate for various ranges of $q$ and $\alpha$. In addition, we carry out numerical simulations of mixing by choosing, at each time instant, a field $U$ that maximizes the instantaneous decay of the $ H^{-\alpha}$ norm. These simulations provide evidence that the optimal mixing rate is at least exponential in time. More broadly, we view the present model and its diffusive analogue, as a useful framework for probing mechanisms of small-scale formation in divergence-free vector fields and for formulating simplified versions of open questions related to the incompressible Euler and Navier--Stokes equations.