Chen, H.W., Marzke, R.O., McCarthy, P.J., Martini, P., Carlberg, R.G., Persson, S.E., Bunker, A. Bridge, C.R., and Abraham, R.G. 2002 ApJ, in press
(Abridged) We present rest-frame R-band galaxy luminosity function measurements for three different redshift ranges: 0.5<=z<=0.75, 0.75<=z<=1.0, and 1.0<=z<=1.5. Our measurements are based on photometric redshifts for ~3000 H-band selected galaxies with apparent magnitudes 17<=H<=20 from the Las Campanas Infrared Survey. We show that our photometric redshifts are accurate with an RMS dispersion between the photometric and spectroscopic redshifts of \sigma_z/(1+z)~0.08. Using galaxies identified in the Hubble Deep Field South and Chandra Deep Field South regions, we find, respectively, that (7.3\pm 0.2) % and (16.7\pm 0.4)% of the H<=20 galaxies are at z>=1. We first demonstrate that the systematic uncertainty inherent in the luminosity function measurements due to uncertainties in photometric redshifts is non-negligible and therefore must be accounted for. We then develop a technique to correct for this systematic error by incorporating the redshift error functions of individual galaxies in the luminosity function analysis. The redshift error functions account for the non-gaussian characteristics of photometric redshift uncertainties. They are the products of a convolution between the corresponding redshift likelihood functions of individual galaxies and a Gaussian distribution function that characterizes template-mismatch variance. We demonstrate, based on a Monte Carlo simulation, that we are able to completely recover the bright end of the intrinsic galaxy luminosity function using this technique. Finally, we calculate the luminosity function separately for the total H-band selected sample and for a sub-sample of early-type galaxies that have a best-fit spectral type of E/S0 or Sab from the photometric redshift analysis.
[ Publication List | Home Page ]