This CD-ROM contains a small sample of the first imaging data from the Sloan Digital Sky Survey. We emphasize that these are engineering data; they were observed and reduced before the SDSS instrumentation and software had reached their final level of performance. Nevertheless, the data are good enough to support some preliminary science. This sampler both shows you a sample of the images and tables and presents the data in a form close to that in which the SDSS data products will be released.
This writeup describes:
Source code is provided in the src directory on this CD-ROM
to aid in reading the atlas
and mask data files. Instructions for
compiling and running the code are given in the file
readme in that directory.
The data directory hierarchy is organized by run number and
camcol number. All outputs on this CD-ROM
are in the data/77/5 subdirectory.
The SDSS imaging camera is a mosaic of 54 CCD detectors in the focal plane, 30 large devices arranged in 6 columns of 5 each and 24 smaller devices around the periphery. The SDSS imaging data are taken in time-delay-and-integrate (TDI) mode at the sidereal rate almost simultaneously in five bands: the sky tracks through 5 CCD detectors in succession, each located behind a different filter, as shown below:
The filters and their effective wavelengths are: u' (3540 Å), g' (4760 Å), r' (6280 Å), i' (7690 Å) and z' (9130 Å). The imaging CCDs are 2048 x 2048 pixel devices which image an area of about 13' x 13' on the sky with 0.4'' pixels. The filter responses and SDSS photometric calibration are described by Fukugita et al. (1996) and in the SDSS Project "Black Book" (1996). The camera is described in detail by Gunn et al. (1998) and the entire SDSS is described in several places, including the Black Book, Gunn and Weinberg (1995) and Knapp (1997).
Because it operates in TDI mode, the camera must scan great circles. The first data, including those on this CD-ROM, were taken by drift-scanning at the celestial equator. These data were taken when the telescope collimation was not yet finalized, and, due to residual astigmatism, the images sizes range from 1.2'' to 1.5'' full width at half maximum (FWHM). The data also contain some horizontal streaking that is an artifact of a problem in the data acquisition readout electronics, which has subsequently been fixed. During more recent observing runs, image sizes < 1'' FWHM have been achieved.
Some SDSS nomenclature will be useful:
A scanline is the data from a single set of CCDs which sweep the same area of sky. Each set of 5 CCDs is housed in a single dewar: there are 6 sets of CCDs separated by about 80% of the CCD width. The area of sky swept by the 6 CCD columns is called a strip. A given area of sky is imaged by performing two successive scans, offset by almost a CCD width, to fill in a stripe. The scanlines in a stripe overlap about 10% on each side with the adjacent scanlines.
The data stream from a single CCD in a scanline is cut into a series of frames which measure 2048 x 1489 pixels and overlap 10% with the adjacent frame. The frames in the 5 filters for the same part of the sky are called a field.
A run is the set of data collected from one continuous pass of the 2.5 m telescope across the sky, and thus covers one strip. Typically, a run lasts for a few hours. Each scanline is labelled by the camcol number (1-6, increasing from S to N) of the camera column of CCDs covering that scanline.
Astrometric calibrations are performed by the astrometric pipeline (Astrom). The astrometric solutions for the frames on this CD-ROM were obtained by comparing the pixel centroids of stars on the frames with the catalog Right Ascension and Declination of matching stars found in Stone's (1997) equatorial astrometric catalog. The matching stars' centroids and coordinates were then fit by a low-order least squares polynomial. The residuals from this fit were used to generate spline functions which were then applied to the polynomial fit to reduce the residuals. The astrometric accuracy of these fits is about 0.1 arcsec (rms, per coordinate).
The photometric calibrations are produced by the monitor telescope pipeline (MT). The calibration procedures, absolute standards and color conversions to the SDSS filters are described by Fukugita et al. (1996). A network of about 100 primary standard stars is being set up using a 20" Monitor Telescope located at APO and a 40" telescope at the US Naval Observatory station in Flagstaff, AZ. The zero points are tied to stars with good absolute spectrophotometric calibrations. Because the primary stars are too bright to be observed by the 2.5 m telescope, secondary sky patches that overlap the 2.5 m scans are observed with the Monitor Telescope. These secondary patches are used to tie the MT photometric system to the 2.5 m observations.
The imaging data reduction is performed by the photometric pipeline (Photo). Photo consists of three successive pipelines. The last of these, the Frames pipeline, operates on an individual set of five frames covering a field. It corrects the frames for artifacts (bad columns, cosmic rays, bleed trails); bias-subtracts and flattens them; finds objects (as linked pixels above the sky) in two stages (bright and faint), as well searching for lower surface brightness objects in a binned image; combines the objects in the five filters; measures the objects (centroid, size, shape, counts); measures the radial profiles; extracts the PSF filtered counts; fits to three simple models (PSF, de Vaucouleurs profile, exponential disk); classifies the objects; deblends compound objects; cuts atlas images (including both parent and child images for blended objects) and writes out a series of outputs: tables of measured parameters; corrected frames; 4x4 pixel binned sky frames (with objects subtracted); atlas images; mask frames (location of saturated pixels, interpolated pixels etc); and summary statistics for each frame.
To perform these calculations, Frames needs to know the flat field, bias, sky, and PSF. These are calculated for the entire run by the Postage Stamp Pipeline, (PSP), which calculates these quantities for the whole run and interpolates them to the center of each frame. The PSP uses cutout images of bright (but unsaturated) stars, at present those used by the ASTROM pipeline; rejects bad stars (double etc) and calculates the parameters of a simple PSF fit. The cutouts are made by the Serial Stamp Collecting Pipeline (SSC) which also aligns the frames in a field.
Several caveats need to be kept in mind when interpreting the data:
More detailed information on data processing is given in chapter DATASYS of the Black Book. A major technical paper describing the photometric pipelines and their outputs is in preparation by Lupton et al. (1998).
These data were collected during the night of June 27, 1998. In our nomenclature, these data are for run 77, camcol 5, fields 311-320.
The data directory hierarchy is organized by run number and camcol number. All outputs on this CD-ROM are thus in the data/77/5 subdirectory.
The imaging outputs are of the following types:
c_r_0311.fit
etc, where c means corrected, r means r' filter and 0311 is the field number from the observing run. Each frame is padded with an additional 128 rows of data from the next frame in the scanning sequence; this is done in order to allow the processing of frames independently of one another.
The astrometric calibrations are contained in the header using the WCS (World Coordinate System) convention. The header also contains a GAIN keyword that gives the conversion from ADU to flux in units of 0th magnitude stars. The zero point is chosen so that this magnitude is on the AB system (Oke and Gunn 1983).
The frames are oriented so that increasing column number (NAXIS1) is the north direction and increasing row number (NAXIS2) is the east direction.
b_r_0311.fit
Much image analysis and examination of the SDSS data will be done by recreating the frames from the atlas images and binned sky images.
a_0311.fit
The cutouts contain pixels deemed to belong to the object plus extra pixels around the outside. The "object" pixels are the union of those defined in each of the five bands. An atlas image is cut even if the object was not detected in that band. SDSS also plans to cut atlas images at the positions of objects identified in catalogues at other wavelengths (IRAS, ROSAT, FIRST etc.) whether or not an optical object is detected there. This is not implemented in the present release.
Part of the operation of the photometric pipeline is to deblend the images of overlapping objects. Atlas images are cut both for the composite parent object and for its deblended children.
The atlas image format is opaque outside the SDSS computing environment. Stand-alone code to convert the atlas images to FITS image files, plus a description of how to compile and use it, is available in the src directory.
m_r_0311.fit
The mask image format is opaque outside the SDSS computing environment. Stand-alone code to convert the mask files to FITS image files, plus a description of how to compile and use it, is in the src directory.
o_0311.fit
These files are in binary FITS table format with three tables per file. The first table is the object list itself. Object quantities are expressed in instrumental units. A full description of the outputs is given in outputs.htm and details of the flags are given in flags.htm and t.htm. The second table defines the numerical values for flags attached to each object, and the third table define the numerical values for the object type attached to each object.
t_0311.fit
These contain the identical quantities one-to-one to the Photo output files except that the instrumental quantities are converted to magnitude and sky position. The format is binary FITS.
t_0311.txt
These files also contain only a subset of the objects in the binary files. The excluded objects are bright objects, objects near the frames edges, and objects in the "padded" areas of each frame.
f_0311.fit
Code to read FITS images and FITS binary tables can be obtained in C or FORTRAN from NASA:
http://legacy.gsfc.nasa.gov/fitsio
See also the FITS primer at this server.
Code for FV, a FITS binary table browser, and other FITS tools can be found at:
http://legacy.gsfc.nasa.gov/ftools
Gunn, J.E., Carr, M.A., Rockosi, C.M., Sekiguchi, M., et al. 1998, AJ (in press, December 1998).
Gunn, J.E., & Weinberg, D.H. 1995, in ``Wide Field Spectroscopy and the Distant Universe'', ed. S.J. Maddox and A. Aragon-Salamanca (Singapore: World Scientific), p 3.
Knapp, G.R. 1997, Sky and Telescope, August, p 40.
Lupton, R.H., Ivezic, Z., Gunn, J.E., et al. 1998, in preparation.
Oke, J.B., & Gunn, J.E. 1983, ApJ, 266, 713.
SDSS Collaboration, 1996 ``The SDSS Black Book'' Black Book. See also http://www.sdss.org.
Stone, R.C. 1997, AJ, 114, 2811.