On dynamical invariants of coadjoint orbits of 3D volume-preserving diffeomorphisms
Robert Cardona, Julian Chaidez, Francisco Torres de Lizaur
TL;DR
This work analyzes the invariants of coadjoint orbits for 3D volume-preserving diffeomorphisms, focusing on helicity as a central but not exclusive invariant. It constructs a framework showing that helicity’s uniqueness can fail ubiquitously in $C^{1}$ on integral homology spheres: either the Ruelle invariant or the topological entropy must vary within fixed-helicity sets, depending on whether non-Anosov dynamics are present or all flows are Anosov, respectively. On the 3-sphere, the Ruelle invariant is actually independent of helicity, yielding a strong negative answer to Arnold–Khesin’s question about somewhere density of coadjoint orbits in helicity level sets; the authors extend a similar negative conclusion to arbitrary 3-manifolds using local invariants. The paper develops a toolkit of fixed-helicity perturbations (including a Franks-type lemma for orbit pieces), toric flow tubes for explicit invariant computations, and local invariants based on zeroes and short periodic orbits to establish nowhere-dense coadjoint orbits and to illuminate the subtle interplay between helicity and dynamical invariants.
Abstract
The helicity, or asymptotic linking number, is a functional of exact volume-preserving vector fields on 3-manifolds, invariant under volume-preserving transformations. It is known to exhibit remarkable uniqueness properties: many invariant functionals reduce to functions of helicity. We examine how severely this uniqueness can fail. On integral homology spheres with the $C^{1}$-topology, the failure is extreme: for every $C^{1}$-open set of nonvanishing exact fields of fixed helicity, some other global dynamical invariant is continuous and non-constant in that set; the Ruelle invariant if some field is non-Anosov, and topological entropy otherwise. In particular, on the three-sphere, the Ruelle invariant is everywhere independent of helicity. This implies, in a very strong sense, a negative answer to a question of Arnold and Khesin on the somewhere density of coadjoint orbits of 3D volume-preserving diffeomorphisms, when considered for the $C^1$-topology. On arbitrary three-manifolds, we also answer the question in the negative using local rather than global invariants.
