A self-consistent phase-space distribution function for the anisotropic Dark Matter halo of the Milky Way

Abstract: Dark Matter (DM) direct detection experiments usually assume the simplest possible 'Standard Halo Model' for the Milky Way (MW) halo in which the velocity distribution is Maxwellian. This model assumes that the MW halo is an isotropic, isothermal sphere, hypotheses that are unlikely to be valid in reality. An alternative approach is to derive a self-consistent solution for a particular mass model of the MW (i.e. obtained from its gravitational potential) using the Eddington formalism, which assumes isotropy. In this paper we extend this approach to incorporate an anisotropic phase-space distribution function. We perform Bayesian scans over the parameters defining the mass model of the MW and parameterising the phase-space density, implementing constraints from a wide range of astronomical observations. The scans allow us to estimate the precision reached in the reconstruction of the velocity distribution (for different DM halo profiles). As expected, allowing for an anisotropic velocity tensor increases the uncertainty in the reconstruction of f(v) but the distribution can still be determined with a precision of a factor of 4-5. The mean velocity distribution resembles the isotropic case, however the amplitude of the high-velocity tail is up to a factor of 2 larger. Our results agree with the phenomenological parametrization proposed in Mao et al. (2013) as a good fit to N-body simulations (with or without baryons), since their velocity distribution is contained in our 68% credible interval.