1. OVERVIEW

The data part of this release are deep images in four wavelength bands (B,V, r and i) obtained with SUSI CCD camera at the ESO New Technology Telescope in La Silla, Chile, and a 'true' colour image in the sense that it is blue, green and red representing B, V and r+i. The four images result from the coaddition of short, dithered observations for a total of 50400, 23400, 23400, 16200 seconds in the four bands, and have average FWHMs for the stellar objects of 0.90, 0.83, 0.83 and 0.70 arcsec respectively.

The limiting magnitudes at 3 sigma within a diameter aperture of 2 FWHM in AB magnitudes are 27.2, 27.0, 26.7 and 26.3 in B, V, r, and i respectively. The images cover a 2.3 x 2.3 arcmin "empty" field centered 100 arcsec south of the z=4.7 QSO BR 1202-0725.

While the field size is relatively small, these observations are comparable to the deepest obtained at ground based telescopes and unique in covering four bands with an image quality better than one arcsec. The color image shows objects as faint as 26th magnitude and it illustrates well the performance in deep imaging at good angular resolution with a relatively modest investment in exposure time at a 4m -class ground telescope. Of the approximately 500 galaxies detected in this field, the largest fraction are expected to be at redshifts smaller than z=1 and about 20% to be distributed at higher z, up to z=4 and possibly beyond.

2. THE SCIENTIFIC PROGRAM

The observing time for these observations was granted to the scientific program "Faint Galaxies in an ultra-deep multicolour SUSI field" proposed for the ESO period 58 and executed in service mode by the NTT team in February through April 1997 in photometric nights with seeing better than 1 arcsec. The scientific program aims at the study of the photometric redshift distribution of the faint galaxies and of gravitational lensing effects in the field. The P.I. of the proposal is Sandro D'Odorico and the Co-P.I. are J. Bergeron, H.M. Adorf, S. Charlot, D. Clements, S. Cristiani, L. da Costa, E. Egami, A. Fontana, B. Fort, L. Gautret, E. Giallongo, R. Gilmozzi, R.N. Hook, B. Leibundgut, Y. Mellier, P. Petitjean, A. Renzini, S. Savaglio, P. Shaver, S. Seitz and L. Yan. A presentation of the data together with a source catalogue and a detailed comparison with the HST data is given by D'Odorico, Arnouts, Cristiani, Zaggia, Fontana and Giallongo, 1998 (in preparation). The program, one of the first to be carried out at the NTT after the upgrading of the hardware and software to VLT standards, was a test case for ESO for the operation of the refurbished telescope system and for the new procedures and software packages, such as Phase II proposal preparation, service observing, data quality control and the archiving and distribution of the data.

3. ABOUT THIS WEB RELEASE

The observing proposal stated the intention to release the reduced data at the earliest possible time after the delivery of the last observations. The images are of value for scientific investigations beside those pursued by the authors and can be very useful to test data reduction and simulation packages to be applied to future surveys. The procedure to recover the images is outlined in section 5. The data can be used without restrictions provided that credit is given to the source (see section 6).

4. ON THE OBSERVATIONS AND DATA REDUCTION

4.1 The Observations

The observations were obtained with the SUSI CCD imager mounted at Nasmyth focus A of the ESO NTT. The instrument images at the f/11 focus of the telescope with just an additional 45 degree mirror and the window of the cryostat before the detector, a 1024x1024, 24 micron pixel thinned CCD (ESO #42). One pixel corresponds to 0.129+/-0.003 arcsec. The basic parameters are summarized in Table 1.

The B and i bands are comparable with the corresponding ones used for the HST Deep Field. The integration times of the HDF are a factor of 2.6 and 7.6 longer in B and i respectively. For stellar objects in the same bands, the HDF goes approximately a factor of 6 and 11 deeper due to the combination of reduced background and better PSF. The gain is reduced to a factor of 2 and 7 for diffuse objects.

4.2 Coaddition and radiation hits removal

The individual raw frames have been bias-subtracted, flatfielded and cleaned of radiation hits. Pixels affected by detector defects,radiation hits satellite trails etc. are given a zero value in the weight maps used for the subsequent combination of the frames. The final frames have been obtained by using the "drizzling" algorithm (Fruchter and Hook, 97) after determining for each sky-subtracted image the relevant parameters (flux scale, shift and rotation with respect to a reference). The output pixel size in the drizzled frames is the same as in the input ones. The final intensity of one output pixel is the sum of the flux in the overlapping area between input and output pixels weighted by a weight factor taking into account the flux-scaling, background r.m.s. and detector quality of each input pixel. The final frames are accompanied by a resulting weight map, which represents the expected inverse variance at each pixel position.

4.3 Astrometry

The astrometric solution has been derived from 15 objects in the field for which coordinates are available from the APM catalogue at the site http://www.ast.cam.ac.uk/~apmcat. These objects are uniformly distributed over the whole SUSI field and a first solution of the astrometry was computed with a polynomial fit. The agreement with the APM is better than 0.1 arcsec. To provide the astrometry in a more practical form, we have computed the astrometric parameters for a tangent-projection derived from the polynomial fit. The solution is again consistent with APM at better than 0.1 arcsec. The coordinates of 118 objects in the field were checked against the values for the same objects derived from a partly overlapping WFP2 image of the field. We find a systematic shift of 0.69+/-0.10 arcsec in alpha and -0.70+/-0.10 arcsec in delta (D'Odorico et al., 1998, in preparation).

The astrometric solutions of the TAN-projection are included into the header of fits files. In this way SkyCat, SAOimage or an STSDAS task of IRAF such as "xy2rd" can be used to derive the RA and Dec (2000) from the pixel coordinates.

4.4 Flux calibration

The photometric calibrations were obtained by observing several standard stars from Landolt (1992). The total magnitudes were computed for a fixed aperture (2.5 arcsec) corrected at 6 arcsec. The zero points of our instrumental system B'V'r'i' were adjusted to give the same BVrI magnitudes as in the standard system for stellar objects with B-V=V-r=r-I=0. The color coefficients used are:

B' - B = -0.14 x (B-V), V' - V = -0.08 x (B-V), r' - r = -0.33 x (V-R), I' - I = 0.01 x (V-I).

The zero-points are 32.60, 32.55, 31.96, 31.81 in B, V, r and i respectively with an accuracy of 0.03 mag. They include the extinction correction.

5. IMAGES AT THIS SITE

5.1 Fits files

The coadded, calibrated images are available in fits format (1110 x 1080 pixels) and identified as Bnew.fits, Vnew.fits, Rnew.fits, Inew.fits. The corresponding weight maps are NSDFwm_B.fits, NSDFwm_V.fits, NSDFwm_r.fits, NSDFwm_i.fits.

5.2 Color image

Three-colour NTT Deep Field [JPEG, 591x575 pixels, 49 kBytes] The three-colour image B_V_RI.jpg (1110 x 1080 pixel format) has been prepared by R.A.E. Fosbury and R.N. Hook at the Space Telescope European Coordinating Facility combining the B, V and (r+I) coadded frames. The fainter objects seen in the combined image have magnitudes ~26. The limiting magnitudes are already fainter than those which can be reached in spectroscopy of continuum sources at an 8-10m telescope. The image illustrates well the non-uniform distribution of galaxies and the frequent occurence of close pairing of objects of very different colours, which would lead to confusion under worse seeing conditions both in imaging and slit spectroscopy. The image is also available on-line in the following JPEG versions: Linear image, 1110x1080 pixels [1.3 MBytes] Linear image, 1110x1080 pixels [207 kBytes] Linear image, 591x575 pixels [49 kBytes]; Square root image, 591x575 pixels [44 kBytes]

5.3 How to retrieve the NDF files

The NDF images can be obtained directly from the web:

• Bnew.fits (4.7 MB)
• Vnew.fits (4.7 MB)
• Rnew.fits (4.7 MB)
• Inew.fits (4.7 MB)

Alternatively, the NDF Data Package can be requested from the ESO/ECF Science Archive Facility by sending a message to Archive Operations specifying the media preferred (DAT or Exabyte). The NDF Data Package includes the co-added images, the co-addition weight maps and the JPEG true color image.

6. NDF OBJECT CATALOGUE

The observations are described in more detailed and discussed in the paper "The NTT SUSI Deep Field" by Arnouts, D'Odorico, Cristiani, Zaggia, Fontana, and Giallongo (astro-ph/9810135). A description of the object catalogue and the actual catalogue are available.

7. PUBLIC AND SCIENTIFIC USE OF THE NTT DEEP FIELD IMAGES

The images part of this release can be reproduced if credit is given to the European Southern Observatory and to the P.I.. The field has to be identified as the "NTT SUSI Deep Field".

The four fits images can be used for scientific work without any restriction provided that the data are identified as "NTT SUSI Deep Field" and reference is given to the paper Arnouts et al. (astro-ph/9810135).