Submillimetre-wave surveys: first results and prospects

The population of distant dusty submm-luminous galaxies was first detected last year [19]. Forms of evolution required to account for both this population and the intensity of background radiation have now been determined [6], and are used to investigate the most efficient observing strategies for future surveys. Submm-wave galaxy surveys detect the redshifted thermal far-infrared radiation emitted by dust grains that reprocess optical/ultraviolet light from young stars and active galactic nuclei (AGN). The selection function in redshift in such surveys is very broad [3], and extends out to redshifts of order 10. Hence they provide an efficient and direct technique for selecting distant dust obscured galaxies and AGN [14]. High-redshift biased selection makes submm-wave surveys an ideal way to search for gravitational lenses [2]. Consistent models of the evolution of dusty galaxies can be constructed to account for submm counts [1, 8, 12, 13, 19], mid-infrared counts [7, 15, 16] and the submm/far-infrared background radiation intensity [9, 11, 17, 18]. While not necessarily correct, these models are well constrained and can be used to predict source counts on a range of different angular scales and at a range of different wavelengths [6]. Here the resulting counts are used to investigate the optimal strategies for future mm/submm-wave surveys. The source confusion noise expected in these surveys is discussed elsewhere [4, 5]. Observed submm-wave counts are presented in Fig. 1, alongside the counts predicted by taking an ensemble average of seven different models of galaxy evolution that are consistent with both the observed counts and the intensity of background radiation [6]. The associated redshift distributions are shown in Fig. 2. Excluding [14], no redshifts have yet been determined for submm-selected sources. A spectroscopic redshift distribution of submmselected sources will provide a powerful test of plausible galaxy evolution models [6]. The broad-band colours of the optical identifications in general suggest agreement with the predictions [20]. The predicted counts of mm/submm-wave sources can be used to determine the most efficient strategy for detecting distant dusty galaxies, in both general galaxy surveys (Fig. 3a) and in surveys for lensed objects (Fig. 3b). The detection rate achieved in the crucial first detections using SCUBA will be greatly exceeded in future surveys; for references to suitable instruments see [4]. The fraction of lensed sources expected in a survey is shown in Fig. 3(c). The rate of confirmation of lenses, including the time required to image lensed structures at sub-arcsec resolution using the MMA, is shown in Fig. 3(d). The detection and confirmation rates differ most at faint flux densities, at which Figure 1: Observed source counts and the ensemble mean (thick lines) and 1σ uncertainty (thin lines) derived from well-fitting count and background models [6]. References to the data are B[1], H[12], Hu[13], K[15], L[7, 16], S[19] and WW[21]. Points without a number correspond to 850-μm data. the follow-up MMA observations require more time. The detection rates predicted for future instruments [10] and telescopes – large ground-based interferometer arrays like the MMA, the 50-m LMT, a 10-m telescope at the South Pole and the 3.5-m space-borne FIRST – exceed those of existing instruments by about two orders of magnitude. In concentrated efforts over a number of years, catalogues of order 10 distant galaxies and AGN will be compiled. The fine angular resolution of large interferometer arrays will allow the detected sources to be resolved directly in the mm/submm waveband. In some cases, their redshifted mid-infrared fine-structure line emission could be used to determine redshifts directly in the submm waveband. In addition, the cosmic microwave background imaging space mission Planck Surveyor will provide an all-sky arcmin-resolution 100-mJy survey, which will be very useful for selecting both extremely luminous submm-wave sources and gravitational lenses [2]. • The first detections of distant dusty galaxies indicate that there is an abundant population of such sources. These can be exploited to investigate galaxy formation and evolution, large-scale structure at high redshifts and the values of cosmological parameters. • It is most important to determine the redshift distribution of submmwave sources selected in blank-field surveys, and thus test the first observationally constrained models of the evolution of dust obscured galaxies at high redshifts [6]. • Huge samples of distant dusty galaxies and gravitational lenses can be compiled using future instruments – especially the LMT and MMA. Figure 2: The ensemble mean (thick lines) and 1σ uncertainty (thin flanking lines) of predicted submm-selected redshift distributions [6]. Acknowledgements. This work has benefited greatly from SCUBA/JCMT observations in collaboration with Ian Smail, Rob Ivison and Jean-Paul Kneib. I thank Malcolm Longair for helpful comments on the manuscript, and Gislaine Lagache and Dave Clements for discussing the results of the ISO FIRBACK programme.