Introduction

Photometric redshifts

With ~1,000,000 galaxies in the SXDS Field, photometric redshifts will be especially important to quantify the evolution of galaxies, detect distant clusters, and select samples for detailed spectroscopic study. The SXDS field will be the target of deep near-infrared imaging with the UKIRT Wide Field Camera (WFCAM), to limits of K~23, H~24, and J~25. This survey will take several months to complete and is unlikely to be surpassed until NGST. When combined with similarly deep (say, R-K~5, etc.) four-color optical images, photometric redshift determination will be possible for the vast majority of sources. Although photometric redshifts can be obtained from optical photometry alone, there is a substantial improvement when it is combined with near-infrared data (Connolly et al.\\ 1997), in particular in the redshift range 1 < z < 2, where the star formation rate is likely to peak (Madau et al. 1996), and we stress the importance of the near-infrared observations (both existing and planned) to this project. This procedure can be applied not only to galaxies but also to quasars and our quasar sample will be 5-6 magnitudes deeper than that of the SDSS, and therefore a very powerful tool for investigating quasar evolution in luminosity and spatial density.

We further note that at faint magnitudes where spectroscopy is impossible, the accuracy of photometric redshifts is limited by the photometric accuracy, which will still be small for our ultra-deep images. We will therefore be able to construct the most accurate photo-z catalogue yet produced, to fainter magntiude limits than previously attempted over such a wide area.

Although the deep WFCAM survey will not start before 2004, we already started pilot observations with SIRIUS on the University of Hawaii 2.2-m telescope to image approximately 175 square arcmin of the SXDS Field to J~22, H~21, and K~20. We also intend to obtain coverage of the entire field to K~19 (and similar depths in J and H). Although obviously much shallower than the proposed WFCAM observations, K=20 corresponds to an L* galaxy at z~2, and will therefore be able to constrain photometric redshifts throughout much of the surveyed volume.

From the optical data alone, we will be able to construct samples of `B-drops' and `R-drops', i.e., Lyman-break galaxies at z~4 and z~6, respectively. While more than 1,000 B-drops are expected to be found, the number of R-drops is highly uncertain and will be important in determining at what look-back time the star formation rate begins to decline. By simple extrapolation of the HDF-N result with photometric redshifts, we expect about 10 to 100 galaxies with z~6 in this field.