The DR3 contains that part of the SDSS imaging and spectroscopy
taken through July 2003. Here we give coverage plots and the detailed
coverage tables from which they were
constructed. Further below, we outline the future
sky coverage of SDSS. The survey terms are explained in the
sections on imaging coverage and spectroscopic coverage below.
For data interfaces, see the data
For reference, the coverage tables describe the sky coverage in
terms of survey coordinates. To translate the sky coverage into
celestial coordinates, please refer to the coordinates
section of the EDR paper and the description below.
Note: The coverage tables, and a number of other
files we distribute, are given as ASCII parameter (.par) files.
Note also the list of list of quality
"holes" (small areas of bad seeing etc.) and missing data, the list of fields which are missing only
apparently (e.g., data from different runs were used in target and
best) and the list of differences in sky
coverage between target and best.
Imaging coverage: stripes, runs, reruns
The survey coverage is defined in survey
coordinates, as described in the coordinates
section of the EDR paper. For detailed astrometry please see the
following note about parts
of the survey which have a slightly different survey coordinate system
from the rest.
The imaging survey scans the sky along great circles, which are
circles of constant survey latitude eta. Scans are obtained
along stripes spaced 2.5° in survey latitude. Each stripe
has an integer stripe number n such that the inclination of a
stripe with respect to the celestial equator is -25° + 2.5°
× n. The boundaries of the region eventually to be
surveyed by the SDSS are defined in terms of survey longitude
lambda in atStripeDef.par.
Runs cover strips, two strips cover a stripe
An imaging run designates a continuous scan of the imaging
telescope along a stripe. The survey camera's focal plane has six
columns of CCDs, so that a single run consists of six camera
columns (camcols) (see the instrument descriptions). The
camera columns have a gap which is filled by an offset second run. The
six camcols taken together are called a strip. Thus
to completely cover the two strips making up one
stripe, at least two runs are required. The imaging
data reduction is done on a run-by-run basis, hence the imaging data in our Data Archive are accessible by run
Great circle coordinates
Every stripe has its own great circle
coordinate system. The coordinate along a great circle is called
mu. The range of data
actually obtained for a given stripe is reported in terms of
a range in great circle mu along that stripe as
startMu and endMu
in the coverage tables above. Both startMu
and endMu are given in integer
The calibrated object lists are accessible in the archive by stripe
number and startMu on that stripe.
For historical reasons, DR1 contained both photometric and
spectroscopic data which lay outside the official survey limits, i.e.,
the startMu/endMu for some runs extend beyond the
lambda limits for the stripes they cover. Since DR3,
target data in those areas are still marked
non-primary, but the
best data are primary. Please
refer to this note about DR3 data
outside the SDSS survey area for details on how to retrieve
A run is divided into images 2048 pix × 1361 pix
which are called fields. Each field then has the first 128
rows of the following field attached to it, so that all survey images
actually have a size of 2048 pix × 1489 pix.
Object detection, photometry and classification are done field by
field. In particular, the calibrated object lists and
survey images are
distributed separately for each field. To account for the overlap, object
detection is not performed on the first 64 rows or the last 64 rows of
each corrected frame.
Note that one therefore should use only objects in
the "primary" area of
each survey field, in order to avoid duplicate area and duplicate
object detections. Near the center of a stripe, the primary area will
be the full 2048 pix × 1361 pix, but frames overlap near the
ends of stripes, reducing the primary area there. Please refer to how
"primary" is set in
the "status" flag of an object and please also see the details of
how overlapping areas are
resolved. The primary area of each field is recorded as entry
stripeArea in the tsField*.fit file
Every time we (re)process our imaging data, we assign the outputs a
new rerun number. Thus, a rerun does not imply
re-observation, but is simply a re-reduction of the same run.
The re-reduction can use more recent software or calibrations than a
previous rerun, or both. Across runs, the rerun is an arbitrary
index, so the same rerun of different runs does not have to be
produced with the same software or calibrations.
Target and Best
We publish two reruns for each run
contained in DR3:
There is one coverage table each for the target and best version. The
tables contain the following information:
- A target rerun, which is the reduction used to select
targets for the spectroscopic survey. This is referred to as
skyVersion = 0. All target reruns in
DR3 have rerun < 20.
- A best rerun, which is the best available
reduction. This is usually a different rerun from the
target version, since our photometric pipelines have
been evolving. The best rerun is referred to as
skyVersion = 1. All DR3 best reruns
have rerun >=40.
- The survey area on the sky, given as a stripe number
and a range for the mu coordinate (
<= mu <=
endmu). These were used
to create the plots above.
- The runs and reruns and the range of
fields which cover the given area. This is the
information needed to find data for a given part of the sky in
the Data archive.
The spectrograph's 640 fibers are distributed on a plug plate of
roughly 3° diameter (7 square degree area) projected on
sky. Spectroscopic coverage is given in terms of the plate centers.
Plates are identified by a unique number assigned during targetting
and the MJD on which they were observed, to distinguish
re-observations of the same plate.
Future sky coverage of SDSS
Current and anticipated sky coverage of SDSS
(Aitoff projection of equatorial coordinates as above, but centered at
The plot above shows the current and anticipated sky coverage of
the spectroscopic survey (also see the SDSS data release
schedule). The solid outline shows the original survey ellipse in
the North Galactic Cap as defined above. Dark
areas show the coverage of DR3. Light areas show
spectroscopic sky coverage expected at the end of the currently funded
survey in June 2005 (to be relased in full in 2006). As DR3 includes
only data collected up through June 2003, data for part of the light areas
are alrady in hand. The imaging survey leads the spectroscopic survey
and will therefore not leave a gap in the sky coverage of the North
Galactic Cap. Filling the gap in the spectroscopic survey is
contingent upon the availability of funding for survey operations
beyond June 2005.
Last modified: Fri Oct 1 13:43:33 CDT 2004