A K-band-selected photometric redshift catalog in the Hubble Deep Field South: Sampling the rest-frame V band to z=3

We present the first results from the Faint Infra-Red Extragalactic Survey of the Hubble Deep Field South (HDF-S). Using a combination of deep near-in frared (NIR) data obtained with the Infrared Spectrograph and Array Camera at the VLT and the WFPC2 Hubble Space Telescope data, we construct a K-band -selected sample which is 50% and 90% complete for and K-s,K-AB less than o r equal to 23.5 K-s,K-AB less than or equal to 22.0, respectively, where th e magnitudes are measured over a diameter aperture. For z less than or equa l to 3, our selection by the K-band flux chooses galaxies based on waveleng ths redder than the rest-frame V band, and so selects them in a way that is less dependent on their current star formation rates than selection in the rest-frame UV. We developed a new photometric redshift technique that mode ls the observed spectral energy distribution (SED) with a linear combinatio n of empirical galaxy templates. We tested this technique using 150 spectro scopic redshifts in the Hubble Deep Field North (HDF-N) from the Cohen et a l. sample (published in 2000) and found Deltaz/(1 + z) approximate to 0.07 for z < 6. We show that we can derive realistic error estimates in z(phot) by combining the systematic uncertainties derived from the HDF-N with error s in z(phot), which depend on the observed flux errors. We estimate photome tric redshifts for 136 galaxies in the HDF-S from the full seven-band, 0.3- 2.2 <mu>m SED. In finding the correct z(phot), our deep NIR data is importa nt for breaking the redshift degeneracy between templates of identical obse rved optical colors. The redshift histogram of galaxies in the HDF-S shows distinct structure with a sharp peak at z approximate to 0.5 and a broad en hancement at z similar to 1-1.4. We find that 12% of our galaxies with K-s, K-vega < 21 lie at z <greater than or equal to> 2. While this is higher tha n the fraction predicted Omega (M) = 1 hierarchical models of galaxy format ion, we find that published predictions using pure luminosity evolution mod els produce too many bright galaxies at redshifts greater than unity. Final ly, we use our broad wavelength coverage to measure the rest-frame UBV lumi nosities, L-rest, for z less than or equal to 3. There is a paucity of gala xies brighter than L-V(rest) greater than or equal to 1.4 x 10(10) h(-2) L. at z similar to 1.5-2, similar to what Dickinson found for the HDF-N (publ ished in 2001). However, z(phot) is particularly uncertain in this regime, and spectroscopic confirmation is required. We also note that at z > 2 we f ind very luminous galaxies with L-V(rest) h(-2) L. (for Omega (M) = 0.3, Om ega (Lambda) = 0.7, H-0 = 100 h km s(-1) Mpc(-1)). Local B-band luminosity functions predict 0.1 galaxies in the redshift range 2 less than or equal t o less than or equal to 3.5 and with L-B(rest) greater than or equal to 5 x 10(10) h(-2) L.(,B), but we find nine. The discrepancy can be explained if L-B(*) increases by a factor of 2.4-3.2 with respect to locally determined values. Random errors in the photometric redshift can also play a role, an d spectroscopic confirmation of the redshifts of these bright galaxies is r equired.