Photometric Redshifts of Galaxies in COSMOS

We measure photometric redshifts and spectral types for galaxies in theCOSMOS survey. We use template fitting technique combined with luminosityfunction priors and with the option to simultaneously estimate dust extinction(i.e. E(B-V)) for each galaxy.Our estimated redshifts are accurate to i<25 andz~1.2. Using simulations with sampling and noise characteristics similar to those inCOSMOS, the accuracy and reliability is estimated for the photometric redshiftsas a function of the magnitude limits of the sample, S/N ratios and the numberof bands used. From the simulations we find that the ratio of derived 95%confidence interval in the redshift probability distribution to the estimatedphotometric redshift (D95) can be used to identify and exclude the catastrophicfailures in the photometric redshift estimates. We compare the derived redshifts with high-reliability spectroscopicredshifts for a sample of 868 normal galaxies with z < 1.2 from zCOSMOS.Considering different scenarios, depending on using prior, no prior and/orextinction, we compare the photometric and spectroscopic redshifts for thissample. This corresponds to an rms scatter of 0.031, with a small number ofoutliers (<2.5%). We also find good agreement (rms=0.10) between photometricand spectroscopic redshifts for Type II AGNs. We compare results from our photometric redshift procedure with three otherindependent codes and find them in excellent agreement. We show preliminaryresults, based on photometric redshifts for the entire COSMOS sample (to i < 25mag.).Comment: 38 pages; 14 Figures; 7 Tables. Accepted for Publication in ApJS. COSMOS Special Issu