Detailed comparison of Milky Way models based on stellar population synthesis and SDSS star counts at the north Galactic pole

S. Gao^1,2, A. Just¹ and E. K. Grebel¹

¹ Astronomisches Rechen-Institut, Zentrum für Astronomie der Universität Heidelberg, Mönchhofstr. 12–14, 69120 Heidelberg, Germany
² National Astronomical Observatories, Chinese Academy of Sciences, Datun Road 20A, Beijing 100012, PR China
e-mail: sgao@nao.cas.cn

Received: 11 October 2011
Accepted: 12 July 2012

Abstract

Context. This article investigates Sloan Digital Sky Survey (SDSS) star counts based on the Just-Jahreiß Galactic disc model by analysing the TRI-dimensional modeL of thE GALaxy (TRILEGAL) and Besançon models of the Milky Way.

Aims. We test the ability of the TRILEGAL and Besançon models to reproduce the colour–magnitude diagrams of SDSS data at the north Galactic pole (NGP). We show that a Hess diagram analysis of colour–magnitude diagrams is much more powerful than luminosity functions in determining the Milky Way structure.

Methods. We derive a best-fitting TRILEGAL model to simulate the NGP field in the (g − r,g) colour–magnitude diagram of SDSS filters via Hess diagrams. For the Besançon model, we simulate the luminosity functions (LF) and Hess diagrams in all SDSS filters. We use a χ² analysis and determine the median of the relative deviations in the Hess diagrams to quantify the quality of the fits by the TRILEGAL models and the Besançon model and compare this to that of the Just-Jahreiß model. The input isochrones in the colour–absolute magnitude diagrams of the thick disc and halo are tested via the observed fiducial isochrones of globular clusters.

Results. We find that the default parameter set lacking a thick disc component gives the best representation of the LF in TRILEGAL. The Hess diagram reveals that a metal-poor thick disc is needed. In the Hess diagram, the median relative deviation of the TRILEGAL model and the SDSS data amounts to 25 percent, whereas for the Just-Jahreiß model the deviation is only 5.6 percent. The isochrone analysis shows that the representation of the main sequences of (at least metal-poor) stellar populations in the ugriz filter system is reliable. In contrast, the red giant branches fail to match the observed fiducial sequences of globular clusters. The Besançon model shows a similar median relative deviation of 26 percent in (g − r,g). In the u band, the deviations are larger. There are significant offsets between the isochrone set used in the Besançon model and the observed fiducial isochrones.

Conclusions. In contrast to Hess diagrams, luminosity functions are insensitive to the detailed structure of the Milky Way components due to the extended spatial distribution along the line of sight. The flexibility of the input parameters in the TRILEGAL model is insufficient to successfully reproduce star count distributions in colour–magnitude diagrams. In the Besançon model, an improvement of the isochrones in the ugriz filter system is needed in order to improve the predicted star counts.