A charge-coupled device study of high-latitude Galactic structure: testing the model parameters

We interpret published charge-coupled device (CCD) UBVI data to deduce the stellar density distribution and metallicity distribution function in the region of 2-8 kpc from the Galactic plane, and compare our results to several star-count models. A feature of extant star-count models is degeneracy between the adopted scaleheights of the thin and thick discs, and their local normalization. We illustrate the utility of this small data set, and future larger sets (e.g. Sloan Digital Sky Survey, SDSS), by explicitly considering consistency between the derived density laws, and the implied solar neighbourhood luminosity function. Our data set, from Hall et al.'s 1996 paper (l = 52degrees, b =-39degrees), contains 566 stars, selected to be consistent with stellar loci in colour-colour diagrams. The effective apparent V -magnitude interval is 15.5 less than or equal toV (o) less than or equal to 20.5<. Our analysis supports the parametrization of the recent (SDSS) galaxy model of Chen et al., except in preferring the stellar halo axial ratio to be η= 0.84. Photometric metal abundances have been derived for 329 stars with (B -V )(o) ≤ 1.0 using a new calibration. This shows a multimodal distribution with peaks at [Fe/H]=-0.10, -0.70 and -1.50 and a tail down to -2.75 dex. The vertical distance-dependent metallicity distribution function, if parametrized by a single mean value, can be described by a metallicity gradient d[Fe/H]/dz ∼-0.2 dex kpc(-1) for the thin disc and thick disc, and d[Fe/H]/dz ∼-0.1 dex kpc(-1) for the inner halo, to z = 8 kpc. However, the data are better described as the sum of three discrete distribution functions, each of which has a small or zero internal gradient. The changing mix of thin disc, thick disc and halo populations with distance from the plane generates an illusion of a smooth gradient.