1303.6099 (John Magorrian)

Bayes versus the virial theorem: inferring the potential of a galaxy from a kinematical snapshot    [PDF]

John Magorrian

I present a new framework for estimating a galaxy's gravitational potential, Phi, from its stellar kinematics. It adopts a fully non-parametric model for the galaxy's unknown phase-space distribution function, f, that takes full advantage of Jeans' theorem. Given an expression for the joint likelihood of Phi and f, the likelihood of Phi is calculated by using a Dirichlet process mixture to represent the prior on f and marginalising. I demonstrate that modelling machinery constructed using this framework is successful at recovering the potentials of some simple systems given perfect kinematical data, a situation handled effortlessly by traditional moment-based methods, such as the virial theorem, but in which the more modern extended-Schwarzschild method fails. Unlike moment-based methods, however, the models constructed using this framework can easily be generalised to take account of realistic observational errors and selection functions.

View original: http://arxiv.org/abs/1303.6099