The SCUBA Half-Degree Extragalactic Survey — I. Survey motivation, design and data processing

A. M. J. Mortier S. Serjeant J. S. Dunlop S. E. Scott P. Ade D. Alexander O. Almaini I. Aretxaga C. Baugh A. J. Benson P. N. Best A. Blain J. Bock C. Borys A. Bressan C. Carilli E. L. Chapin S. Chapman D. L. Clements K. Coppin M. Crawford M. Devlin S. Dicker L. Dunne S. A. Eales A. C. Edge D. Farrah M. Fox C. Frenk E. Gazta naga W. K. Gear E. Gonzales-Solares G. L. Granato T. R. Greve J. A. Grimes J. Gundersen M. Halpern P. Hargrave D. H. Hughes R. J. Ivison M. J. Jarvis T. Jenness R. Jimenez E. van Kampen A. King C. Lacey A. Lawrence K. Lepage R. G. Mann G. Marsden P. Mauskopf B. Netterfield S. Oliver L. Olmi M. J. Page J. A. Peacock C. P. Pearson W. J. Percival A. Pope R. S. Priddey S. Rawlings N. Roche M. Rowan-Robinson D. Scott K. Sekiguchi M. Seigar L. Silva C. Simpson I. Smail J. A. Stevens T. Takagi M. Vaccari G. Tucker,32 C. Vlahakis,3 I. Waddington,23 J. Wagg,5 M. Watson,33 C. Willott34 - A B S T R A C T The Submillimetre Common-User Bolometer Array (SCUBA) Half-Degree Extragalactic Survey (SHADES) is a major new blank-field extragalactic submillimetre (submm) survey currently underway at the James Clerk Maxwell Telescope (JCMT). Ultimately, SHADES aims to cover half a square degree at 450 and 850 m to a 4 depth of 8 mJy at 850 m. Two fields are being observed, the Subaru/XMMNewton Deep Field (SXDF) (02h18m 05) and the Lockman Hole East (10h52m + 57). The survey has three main aims: (i) to investigate the population of high-redshift submm galaxies and the cosmic history of massive dust-enshrouded star formation activity; (ii) to investigate the clustering properties of submm-selected galaxies in order to determine whether these objects could be progenitors of present-day massive ellipticals; and (iii) to investigate the fraction of submm-selected sources that harbour active galactic nuclei. To achieve these aims requires that the submm data be combined with cospatial information spanning the radio-to-X-ray frequency range. Accordingly, SHADES has been designed to benefit from ultra-deep radio imaging obtained with the Very Large Array (VLA), deep mid-infrared observations from the Spitzer Space Telescope, submm mapping by the Balloon-borne Large Aperture Submillimetre Telescope (BLAST), deep near-infrared imaging with the United Kingdom Infrared Telescope, deep optical imaging with the Subaru Telescope and deep X-ray observations with the XMMNewton observatory. It is expected that the resulting extensive multiwavelength data set will provide complete photometric redshift information accurate to z < 0.5 as well as detailed spectral energy distributions for the vast majority of the submm-selected sources. In this paper, the first of a series on SHADES, we present an overview of the motivation for the survey, describe the SHADES survey strategy, provide a detailed description of the primary data-analysis pipeline and demonstrate the superiority of our adopted matched-filter source-extraction technique over, for example, Emerson-II style methods. We also report on the progress of the survey. As of 2004 February, 720 arcmin2 had been mapped with SCUBA (about 40 per cent of the anticipated final total area) to a median 1 depth of 2.2 mJy per beam at 850 m (25 mJy per beam at 450 m), and the sourceextraction routines give a source density of 650 50 sources deg2 > 3 at 850 m. Although uncorrected for Eddington bias, this source density is more than sufficient for providing enough sources to answer the science goals of SHADES, once half a square degree is observed. A refined reanalysis of the original 8-mJy survey Lockman hole data was carried out in order to evaluate the new data-reduction pipeline. Of the 17 most secure sources in the original sample, 12 have been reconfirmed, including 10 of the 11 for which radio identifications were previously secured. Theories of galaxy formation and evolution, embedded within hierarchical structure-formation models, can describe many of the observed features of galaxies (Cole et al. 2000; Granato et al. 2000; Hatton et al. 2003). While local galaxies can in some cases still provide constraints on the high-redshift populations (Panter, Heavens & Jimenez 2003; Heavens et al. 2004), the bulk of the constraints on models of galaxy evolution come either from the integral constraint from the far-infrared background (e.g. Dwek et al. 1998; Gispert, Lagache & Puget 2000, and references therein), or directly from high-redshift galaxy surveys (e.g. Steidel et al. 1999), the most ground-breaking of which were the CanadaFrance Redshift Survey (Lilly et al. 1995) and Hubble Deep Field North (HDF-N) (Williams et al. 1996). Such optical surveys have led to a great deal of progress in understanding the assembly of stellar populations, and hierarchical galaxy-formation models are in increasingly good agreement with many (but not all) of these observations (e.g. Guiderdoni et al. 1998; Blain 1999; Kauffmann et al. 1999; van Kampen, Jimenez & Peacock 1999; Cole et al. 2000; Somerville, Primack & Faber 2001). However, the discovery of a substantial population of faint submillimetre (submm) galaxies (Smail, Ivison & Blain 1997; Barger et al. 1998; Hughes et al. 1998) has posed serious problems for the current generation of galaxy-formation models based on hierarchical structure growth. Models of the optical/ultraviolet (UV) spectral energy distributions (SEDs) of the galaxy population in the HDF-N (Thompson, Weymann & Storrie-Lombardi 2001) predict only submilliJansky/ Jy-level 850- m flux densities, but the submm point sources in the HDF-N field have 850- m flux densities of several mJy (Hughes et al. 1998; Borys et al. 2003; Serjeant et al. 2003a). This shows that there is a population of starforming galaxies that are heavily obscured by dust and have much higher infrared luminosities than would be inferred from the optical/UV observations alone. If these galaxies are at high redshifts (as current data imply), and if their emission is powered by star formation with a standard solar neighbourhood initial mass function (IMF), then their observed 850- m flux densities of several mJy imply star formation rates 1000 M yr1. Moreover, the farinfrared (FIR) luminosity density implied by the submm galaxy population suggests that these infrared-luminous galaxies contributed several tens of per cent of the volume-averaged star formation density at z 2 (e.g. Smail et al. 1997; Barger, Cowie & Sanders 1999). One attractive interpretation of the submm galaxy population is that these violently star-forming galaxies are progenitors of presentday massive ellipticals (Hughes et al. 1998; Scott et al. 2002). There are four main pieces of evidence in support of this (e.g. Dunlop 2002). First, the star formation rates inferred from the submm flux densities are sufficient to construct the stellar population of even the most-massive elliptical galaxy in 1 Gyr; secondly, the K-band morphologies of submm-selected galaxies resemble those of radio galaxies, which (...truncated)