Keck Deep Fields. III. Luminosity‐dependent Evolution of the Ultraviolet Luminosity and Star Formation Rate Densities atz∼4, 3, and 2

We use the Keck Deep Fields UGRI catalog of z~4, 3, and 2 UV-selectedgalaxies to study the evolution of the rest-frame 1700A luminosity density athigh redshift. The ability to reliably constrain the contribution of faintgalaxies is critical and our data do so as they reach to M*+2 even at z~4 anddeeper still at lower redshifts. We find that the luminosity density at highredshift is dominated by the hitherto poorly studied galaxies fainter than L*,and, indeed, the the bulk of the UV light in the high-z Universe comes fromgalaxies in the luminosity range L=0.1-1L*. It is these faint galaxies thatgovern the behavior of the total UV luminosity density. Overall, there is agradual rise in luminosity density starting at z~4 or earlier, followed by ashallow peak or a plateau within z~3--1, and then followed by the well-knowplunge at lower redshifts. Within this total picture, luminosity density insub-L* galaxies evolves more rapidly at high redshift, z>~2, than that in moreluminous objects. However, this is reversed at lower redshifts, z<~1, areversal that is reminiscent of galaxy downsizing. Within the context of themodels commonly used in the observational literature, there seemingly aren'tenough faint or bright LBGs to maintain ionization of intergalactic gas even aslate as z~4. This is particularly true at earlier epochs and even more so ifthe faint-end evolutionary trends we observe at z~3 and 4 continue to higherredshifts. Apparently the Universe must be easier to reionize than some recentstudies have assumed. Nevertheless, sub-L* galaxies do dominate the total UVluminosity density at z>~2 and this dominance further highlights the need forfollow-up studies that will teach us more about these very numerous but thusfar largely unexplored systems.Comment: Accepted for publication in the Astrophysical Journal. Abstract abridge