Home | Overview | Browser | Access | Login | Cookbook | nonSurvey

DR2 Release - March 2007

• This release contains one UKIDSS database, UKIDSSDR2plus. As with previous plus versions of database releases, DR2plus will contain some UKIDSS survey data from areas that have incomplete filter coverage. Users only interested in areas/objects that have been observed in all nominal passbands should use the provided source table views, see below.
• During the post-processing of data included in DR2 quality error bit flags were set. The flags appear as ppErrBits attributes in detection and source tables and can be used to refine database queries. Issues flagged in this way include: possible saturation, source close to frame boundary and possible contamination by cross-talk. See the quality error bit page for further information and the SQL cookbook for examples of queries using the flags.
• This release includes a photometric recalibration of all stacks and catalogues. In regions of low extinction any changes to the DR1 zero points are very small, of order 0.01mag. The ongoing uncertainty of the overall zero point for the Z and Y bands (see Warren et al, 2007, MNRAS 375, 213) has not yet been addressed, and is deferred to DR3, when a comprehensive assessment of calibration issues, and a fundamental recalibration will be implemented. For DR2 the main change is the inclusion of corrections to the zero points in regions of high extinction (a consequence of calibrating to 2MASS). These corrections are small in JHK (because the terms are small) as well as in Y (because the extinction in the LAS is small) and likely to be reliable at the level of 0.02mag. The corrections in the ZY band (GCS) in regions of moderate to high extinction (E(B-V)>0.2mag) should be treated with caution. If accurate calibration of the ZY bands at the 0.01mag level is required the user will need to make their own assessment of the zero point. Two new attributes in MultiframeDetector, are useful in this respect. The attributes nightZPCat, nightZPErrCat give the average ZP for the night, and the scatter. Further details can be found on the CASU website.
• New archive databases holding external data are: bestDR5 (SDSS DR5), MGC (Millennium Galaxy Catalogue) and NVSS. Neighbour tables joining these surveys with some of the UKIDSS surveys have been created in UKIDSSDR2plus (see the schema browser).
• Difference images (between broad-band K and narrow-band H2) are now available for a small subset of the GPS area.

The DR2 release now includes source merging for the GPS (this was not available in the initial relase).

In terms of the Ultra Deep Survey (UDS) the DR2 release is the same as DR1 (no new UDS observations in the semesters ingested since DR1).

A description of the data and how it is held in the archive is given in the data overview.

DR2 Release Summary

Survey	No. of multiframes	No. of objects in detection table	No. of sources in source table	Approximate depths	Plots
LAS	96,789	29,741,573	16,814,636	Y=20.2, J=19.6, H=18.8, K=18.2	link
GPS	113,500	604,031,396	360,215,693	J=19.8, H=19.0, K=18.1	link
GCS	54,496	38,320,483	22,775,753	Z=20.4, Y=20.1, J=19.6, H=18.8, K=18.2	link
DXS	150,602	8,680,100	862,933	J=21.3-22.4,K=19.2-21.1	link
UDS	29,783	192,203	96,101	J=22.6, K=21.5	link

Particular attention should be given to the following notes detailing limitations and bugs (see also generic issues and bugs noted under Known Issues):

KNOWN BUGS:

• Incorrect version of extractor run on DXS in DR2
  An older version of the CASU extractor was run by WFAU on the DXS deep stacks in the DR2 data release. While we did have the correct version, we inadvertently checked out an older version (May 2006 rather than July 2006). This version overestimates the background noise, so the faintest objects are not extracted. The magnitude limit is approximately 0.1 magnitudes too bright. The DXS will be correctly processed in data release 3.
• Error in a UKIDSSDR2 DXS deep stack
  A multiframe (multiframeID=451891) with incorrect astrometry in one extension (extNum=3) was identified during curation for DR3. The extension has now been deprecated and will not be a problem for UKIDSSDR3. However, this stack is a component of a deep stack in the UKIDSSDR2 release. The deep stack has multiframeID=1104153 and all objects that have dec<0.718889 on extNum=3 should be excluded from any data sets. These objects are stars and galaxies that have been shifted by ~74". This can be seen in the image of this extension.
  

• Due to time constraints, at the time of initial release, the GPS source table (gpsSource) in DR2plus is empty (i.e. contains zero sources). It is anticpated that gpsSource will be loaded and released in DR2plus in late March.
• For the Ultra Deep Survey (UDS), the DR2 release is the same as DR1 (no new UDS observations in the semesters ingested since DR1).

• April 12th 2007: UKIDSS GALACTIC PLANE SURVEY NOW AVAILABLE IN DR2.
  As noted below, GPS merged sources were excluded from the original DR2 release to expedite delivery of the LAS, GCS and DXS. GPS merged sources are now available.
  
• The LAS YJ-HK offset frame association bug (see below) has been fixed.

• Mar 2007: Small number of LAS frames incorrectly merged
  A small number of frame sets in the LAS are affected by a frame association bug where YJ and HK frames offset by much more than 1 arcmin have been merged into single frame sets. This has resulted in YJ- and HK-only merged sources in regions within those frame sets where the frames do not overlap well (owing to retiling of the areas affected during the observations), but those sources appear nonetheless in the lasYJHK views for example. Thanks are due to R.G. McMahon for producing a comprehensive and clear analysis of this problem comparing DR2 to DR1. FIXED: 12 APR 2007

• July 3rd 2007: Seaming errors in LAS and GPS in DR2
  A major bug has been identified in the seaming attribute (priOrSec) for the LAS in the post-April 12th re-release (see below) and the GPS in DR2. If you have made seamless LAS or GPS selections since April 12th using the WHERE predicate "... AND (priOrSec=0 OR priOrSec=frameSetID) AND ...", or if you have made selections from any of the LAS/GPS source table views, then your results may be affected by incompleteness in the overlapping regions between adjacent frame sets. For the LAS, this bug will be fixed asap (watch this space); for the GPS, a fix may have to wait until DR3. WFAU would like to apologise for any inconvenience caused by this error.
  FIXED for LAS: 13-15 July 2007 - lasSource has been reseamed and the priOrSec flags re-determined in DR2.
  Due to the size of the GPS a decision was made to initially reset the priOrSec flag to zero in gpsSource as reseaming takes a considerable time. The consequence of this is that in the interim period when querying gpsSource duplicates will be returned in overlap regions.
  RESET for GPS on our secondary server (used by Astrogrid): 17 July 2007 - the priOrSec flag for all sources in gpsSource on our seondary server have been set to zero. The main server (accessed through the WSA website) will be updated at a later date and a note added here when done.
  FIXED for GPS on the main server: 17th August 2007. Seamless selections can now be made using (priOrSec=0 OR priOrSec=frameSetID) in the usual manner. The secondary server copy (for crossmatch queries) has been taken offline for the next day or so in order that the update can be applied there also.
  

• June 26th 2007: lasPointSource and gcsPointSource views bug fix
  Redefinition of the views for the UKIDSS LAS and GCS Source tables for DR2 resulted in bugs being introduced. For the LAS in particular, any selection through the view was seriously incomplete. FIXED: 26 Jun 2007 16:00 BST