Session 6 - Evolution, Survey and Clusters of Galaxies.
Display session, Monday, June 10
Tripp Commons,

[6.06] Photometric Redshifts Using Galaxy Image Structure

D. R. Andersen, M. A. Bershady, A. Jangren (Penn State Univ.)

Spectroscopic redshifts of faint galaxies require large telescopes and considerable observing time. However, Connolly et al (1995) have shown that broadband photometry in UB_JR_FI_N bands can be used to estimate redshifts to an accuracy of \Delta z \sim 0.05 to B_J\sim22.5 and z\sim0.5. The accuracy is expected to improve by at least a factor of two with more precise photometry, and the technique can be extended to higher redshift with deeper spectroscopic training sets. Application of this photometric redshift technique represents a 30--fold increase in observing efficiency over multi--fiber spectrographs to this depth. The efficiency increases steeply at fainter magnitudes because of the increasing surface-density of galaxies. We show that the photometric redshift technique can be improved even further by including image structure characteristics as well. After binning a sample by a redshift--independent concentration index, photometric redshifts derived from just B,R and the log of the apparent size \theta have dispersions (\Delta z) comparable to the four-band fit. This success is due to the strong correlations of apparent surface-brightness with redshift, and rest-frame color to image-concentration. U and I band CCD images require 80% of the observing time to reach the same S/N as B and R for typical galaxy colors. By eliminating the need for these bands, the photometric redshift technique improves in efficiency by another factor of \sim5, or a net \sim150--fold improvement over spectroscopic redshifts to B_J=22.5. We discuss one way of exploiting photometric redshift technique to select volume--limited samples of field galaxies. Such samples are particularly well suited for studying the faint end of the luminosity function.