The difficulty of measuring the local dark matter density

Institut für Astrophysik, Georg-August-Universität, Friedrich-Hund-Platz 1, 37077 Göttingen, Germany
e-mail: hessman@astro.physik.uni-goettingen.de

Received: 4 March 2015
Accepted: 18 May 2015

Abstract

The analysis of the vertical velocity dispersion of disc stars in the local Milky Way is the most direct astronomical means of estimating the local dark matter density, ρ_DM. Current estimates for ρ_DM based on the mid-plane dynamic density use a local baryonic correction that ignores the non-local effects of spiral structure and significantly underestimates the amount of dynamically relevant gas now known to be present in the ISM; the additional gas plus the remaining uncertainties make it practically impossible to measure ρ_DM from mid-plane kinematics alone. The sampling of inhomogeneous tracer populations with different scale-heights and scale-lengths results in a systematic increase in the observed dispersion gradients and changes in the nominal density distributions that, if not properly considered, can be misinterpreted as a sign of more dark matter. If the disc gravity is modelled locally using an infinite disc, the local variation in the vertical gravity due to the globally exponential disc components results in an underestimation of the baryonic contribution by as much as ~40%. Given only the assumptions of stationarity, an axially and vertically symmetric disc, doubly exponential tracer and mass-component density profiles, a phenomenologically justified model for the cross-dispersion component σ_Rz, and a realistic model for g_z(z), it is possible to solve the full vertical Jeans equation analytically for the vertical dispersion σ_z(z) and hence test the robustness of previous attempts at measuring ρ_DM. When the model parameters for σ_Rz are estimated from SEGUE G dwarf star data, it is still not possible to explain the difference in behaviour seen in the simple thick- and thin-disc datasets reported by Buedenbender et al. (2014, MNRAS, submitted). Rather than being a fundamental problem with the kinematical model, this effect appears to be a further sign of the difficulty of defining and handling kinematically homogeneous tracer populations.