THE zCOSMOS 10k-BRIGHT SPECTROSCOPIC SAMPLE

We present spectroscopic redshifts of a large sample of galaxies with I_(AB) < 22.5 in the COSMOS field, measured from spectra of 10,644 objects that have been obtained in the first two years of observations in the zCOSMOS-bright redshift survey. These include a statistically complete subset of 10,109 objects. The average accuracy of individual redshifts is 110 km s^(–1), independent of redshift. The reliability of individual redshifts is described by a Confidence Class that has been empirically calibrated through repeat spectroscopic observations of over 600 galaxies. There is very good agreement between spectroscopic and photometric redshifts for the most secure Confidence Classes. For the less secure Confidence Classes, there is a good correspondence between the fraction of objects with a consistent photometric redshift and the spectroscopic repeatability, suggesting that the photometric redshifts can be used to indicate which of the less secure spectroscopic redshifts are likely right and which are probably wrong, and to give an indication of the nature of objects for which we failed to determine a redshift. Using this approach, we can construct a spectroscopic sample that is 99% reliable and which is 88% complete in the sample as a whole, and 95% complete in the redshift range 0.5 < z < 0.8. The luminosity and mass completeness levels of the zCOSMOS-bright sample of galaxies is also discussed.
\n