What's New in DR8?

SDSS Data Release 8 differs from its predecessors, both because there is additional data, but also in the processing of the data and in some details of the data distribution. Here we summarize the main differences between DR8 and DR7: new data, new reductions and new formats.

New data

As part of the first two seasons of SDSS-III observations, we obtained new imaging and spectroscopic data.

From the SDSS imaging camera, there is 3000 square degrees more of unique imaging coverage, mostly in the Southern Galactic Cap. There are a small number of runs that were taken as part of SDSS-I and -II that were not previously released that we include here.

As previously, we include matches to USNO-B catalog, the ROSAT All-Sky Survey Faint Source Catalogue and Bright Source Catalogue, and the Third Reference Catalog (but with object coordinates updated to those listed in the NASA Extragalactic Database as of Spring 2009). For DR8 we have added matches of the photometry to the 2MASS PSC and XSC All Sky catalogs. Furthermore, we have updated the matches to the FIRST radio survey catalog to its 2008 version.

Using the SDSS spectrograph, SEGUE-2 took 211 new plates of data, or 135,040 more spectra. These new plates use somewhat different targeting schemes than the SEGUE-1 observations did. In addition, several new calibration observations of clusters have been taken.

We include in DR8 a small number of plate observations (108) which were taken as part of SDSS-I and SDSS-II but not included in DR7. These include 12 plates with no observation at all in DR7; in the other cases DR7 did not include all observations (MJDs) of the plate. Some of these plates are excellent quality, some are of lower signal-to-noise. Quality flags are available to check each plate for its individual quality.

Finally, as part of this release, we include detailed target selection information for all the SEGUE-1 and SEGUE-2 plates. In addition, we include the results of target selection as run on the entire photometric data set.

New reductions

In addition, there have been new reductions of all of the data. The imaging data has been entirely reprocessed, and the stellar spectroscopy has been re-analyzed with a new stellar parameters pipeline.

The imaging data has been reprocessed through the latest version the photometric pipeline (v5_6_3) and re-calibrated and re-resolved. In more detail:

The new photo v5_6_3 improves the sky subtraction around bright objects.
All of the data has been recalibration with the "ubercalibration" procedure of Padmanabhan et al. (2008).
A new version of resolve, which determines the unique set of objects across the survey area, has been run. This version includes "non-standard" SDSS imaging runs more naturally, and retains more information about multiply-observed objects.

Not all of the old information is pertinent to the new reductions. For example, there are no "imaging QA pages"; instead each field is given a "score" to define its quality (as explained in the resolve documenation).

Unfortunately, the new reductions introduced some errors in astrometry, which are quite substantial at declinations greater than about 41 deg in declination. See the astrometry algorithms description for details.

As in DR7 and earlier, we provide "corrected frames" for each filter and each field. However, we have altered the format of these files somewhat, applying sky-subtraction and calibration to the pixels, as well as correcting the WCS headers to be much closer to the true astrometric solution. Thus, these files are immediately usable for science without any calibration. The new files contain enough information in them to back out both the sky-subtraction and calibration and thus recover the original corrected frame.

The spectroscopic results have been partially reprocessed. There are two basic steps to the processing:

2D Extraction: This step produces the calibrated spectra. For SDSS-I and -II data re-released here, we use the same reductions as released in DR7. For most of the new SEGUE-2 data, the same version of the reduction pipeline was used. An exception is the cluster plates, for which some adjustments to the pipeline were made.
1D Analysis: This step produces redshifts and classifications for each spectrum. For all data, we rely on the idlspec2d pipeline for redshifts (also known as "specBS"). This pipeline differs in detail wth the previous pipeline used in DR7, though for virtually all objects (> 99%) they agree.

For stars, a new version of the SEGUE Stellar Parameters Pipeline (SSPP) has been run, with improved determinations of stellar properties. This procedure has been run on both the SEGUE data, and on all stars in the full DR8 spectroscopy.

For galaxies, for the purpose of line index measurements, we include the results of the "Garching" group measurements for DR7 galaxies (e.g. Brinchmann et al. (2004), Kauffmann et al. (2003) and Tremonti et al. (2004)).

New Data Distribution

The data distribution system has changed somewhat, due to the different focus of SDSS-III and with respect to changes in the calibration and resolve of the photometric data. The Catalog Archive Server (CAS) is largely the same, with some additional data, and some data not included. The Science Archive Server (SAS) uses a significantly different front-end than did the Data Archive Server (DAS) of DR7 and previous.

A major difference in the CAS and SkyServer is that for DR7 and previous, there was a set of full photometric catalogs distributed corresponding to the version of the data reductions used for targeting the SDSS Legacy survey. In DR8 we do not distribute that catalog, only the latest version of the photometric reductions.