Does SEGUE/SDSS indicate a dual Galactic halo?

Abstract

We re-examine recent claims of observational evidence for a dual Galactic halo in SEGUE/SDSS data, and trace them back to improper error treatment and neglect of selection effects. In particular, the detection of a vertical abundance gradient in the halo can be explained as a metallicity bias in distance. A similar bias, and the impact of disk contamination, affect the sample of blue horizontal branch stars. These examples highlight why non-volume complete samples require forward-modelling from theoretical models or extensive bias-corrections. We also show how observational uncertainties produce the specific non-Gaussianity in the observed azimuthal velocity distribution of halo stars, which can be erroneously identified as two Gaussian components. A single kinematic component yields an excellent fit to the observed data, when we model the measurement process including distance uncertainties. Furthermore, we show that sample differences in proper motion space are the direct consequence of kinematic cuts, and are enhanced when distance estimates are less accurate. Thus, their presence is neither a proof for a separate population, nor a measure of reliability for the applied distances. We conclude that currently there is no evidence from SEGUE/SDSS that would favour a dual Galactic halo over a single halo full of substructure.