next previous
Up: The ATESP radio survey

1 Introduction

Sky surveys play an important role in astronomy. Large collections of objects allow reliable studies of the average properties of different constituents of the universe; new populations can be found and well defined sub-samples can be extracted for further detailed analysis.

Surveys in the radio domain have yielded many important results in the past, since the discovery of radio galaxies and quasars. The first samples of radio sources ( $S\mathrel{\mathchoice {\vcenter{\offinterlineskip\halign{\hfil
$\displaystyle ... Jy) have demonstrated that classical radio galaxies are rare in the local universe and strongly evolve with cosmic time both in density and luminosity (e.g. Longair 1966). More recently, deep radio surveys ( $S \mathrel{\mathchoice {\vcenter{\offinterlineskip\halign{\hfil
$\displaystyle ... mJy) have shown that normalized radio counts show a flattening below a few mJy, corresponding to a steepening in the actual observed counts (see e.g. Windhorst et al. 1990 for counts at 1.4 GHz). This change of slope is generally interpreted as being due to the presence of a new population of radio sources (the so-called sub-mJy population) which does not show up at higher flux densities (see e.g. Condon 1989). To explain the new faint radio population several scenarios have been invoked: strongly-evolving normal spirals (Condon 1984, 1989); actively star forming galaxies (e.g. Rowan-Robinson et al. 1993); or a non-evolving population of local (z < 0.1) low-luminosity galaxies (e.g. Wall et al. 1986). The true nature of the population is not well established. The same is true for the relative contributions from the above mentioned objects. Furthermore the source space density inferred from the faint end of the local bivariate luminosity function for both spirals and ellipticals is not well known (see e.g. Condon 1996). Therefore it is not possible to estimate the local contribution to the counts (even if expected to be small) nor is there a clear local reference frame for understanding evolutionary phenomena.

Unfortunately, due to the long observing times required to reach faint fluxes, the existing samples in the sub-mJy region are generally small. Table 1 shows a compilation of the largest 1.4 GHz surveys available in the mJy and sub-mJy regime with the surveyed areas and limiting fluxes (note, however, that the quoted limiting fluxes are often not uniform over the entire areas).


Table 1: 1.4 GHz mJy and sub-mJy radio surveys
Survey References Area $S_{\rm lim}$
    sq. deg. mJy
NVSS Condon et al. 1998 $3 \;10^4$ 2.5
ELAIS N b Ciliegi et al. 1999 4.22 1.15
FIRST White et al. 1997 1550 1.0
ELAIS S Gruppioni et al. 1999b 4.0 0.4
VLA-NEP Kollgaard et al. 1994 29.3 0.3
PDF Hopkins et al. 1998 3.0 0.2
Marano Field Gruppioni et al. 1997 0.36 0.2
ELAIS N a Ciliegi et al. 1999 0.12 0.135
LBDS Windhorst et al. 1984 5.5 0.1 - 0.2
Lockman Hole de Ruiter et al. 1997 0.35 0.12
Lynx 3A Oort 1987 0.8 0.1
0852+17 Condon & Mitchell 1984 0.32 0.08
1300+30 Mitchell & Condon 1985 0.25 0.08
HDF Richards 1999 0.3 0.04
ATESP this paper 25.9 0.47

The identification work and subsequent spectroscopy are very demanding in terms of telescope time. Typically, no more than $\sim 50$ - 60% of the radio sources in sub-mJy samples have been identified on optical images, even though for the $\mu$Jy survey in the Hubble Deep Field an identification rate of about 80$\%$ has been reached (Richards et al. 1999). On the other hand, the typical fraction of spectra available is only $\sim 20\%$. The best studied sample is the Marano Field, where $\sim 45\%$ of the sources have spectral information (Gruppioni et al. 1999a).

To establish a firm point in the radio properties of galaxies in the local (z < 0.2) universe it is necessary to survey a large area in the sky down to faint flux limits. Furthermore it is necessary to have in the same region a statistically significant sample of galaxies with well studied optical properties (radial velocities, magnitudes etc.). To alleviate the identification work, regions with deep photometry (possibly multicolor) already available provide a significant advantage. The region we have selected fulfills these requirements at least partially.

Vettolani et al. (1997) made a deep redshift survey in two strips of $22^{\circ}\times 1^{\circ}$ and $5^{\circ} \times 1^{\circ}$ near the SGP by studying photometrically and spectroscopically nearly all galaxies down to $b_{\rm J}
\sim 19.4$. The survey, yielding 3342 redshifts (Vettolani et al. 1998), has a typical depth of z=0.1 with 10$\%$ of the objects at z>0.2 and is $90\%$ complete. In the same region lies the ESO Imaging Survey (EIS, Nonino et al. 1999) Patch A (3.2 sq. degr.), consisting of deep images in the I band out of which a galaxy catalogue 95% complete to I=22.5 has been extracted. Further V band images are available over $\sim$ 1.5 sq. degr.

We used the 6 km configuration of the ATCA to make a 20 cm radio continuum mosaic of the region covered by the ESP galaxy redshift survey. The ATESP radio survey has uniform sensitivity ($1 \sigma$ noise level $\sim$79 $\mu$Jy).

The present paper essentially deals with a description of the survey, the observations, the mosaic technique and the data reduction. It is organized as follows. In Sect. 2 the survey design, in particular with respect to the mosaic technique is explained. In Sects. 3 and 4 we present in detail the calibration of our 20 cm observations and the data reduction. We discuss the problems encountered and the solutions adopted. Section 5 is dedicated to the analysis of the mosaics. A summary is given in Sect. 6.

next previous
Up: The ATESP radio survey

Copyright The European Southern Observatory (ESO)