2023MNRAS.519.6065B

Galaxies, dark matter haloes, and star clusters have a finite extent, yet most simple dynamical models have an infinite extent. The default method to generate dynamical models with a finite extent is to apply an energy truncation to the distribution function, but this approach is not suited to construct models with a preset density profile and it imposes unphysical constraints on the orbit population. We investigate whether it is possible to construct simple dynamical models for spherical systems with a preset density profile with a finite extent, and ideally with a different range of orbital structures. We systematically investigate the consistency of radially truncated dynamical models, and demonstrate that no spherical models with a discontinuous density truncation can be supported by an ergodic orbital structure. On the other hand, we argue that many radially truncated models can be supported by a tangential Osipkov-Merritt orbital structure that becomes completely tangential at the truncation radius. We formulate a consistency hypothesis for radially truncated models with such an orbital structure, and test it using an analytical example and the numerical exploration of a large model parameter space using the SPHECOW code. We physically interpret our results in terms of the occupancy of bound orbits, and we discuss possible extensions of the tangential Osipkov-Merritt orbital structure that can support radially truncated models.