Problem: These galaxy pieces are, for the most part, not astrophysically useful, nor is the photometry reliable (due to background removal limitations with the pipeline).
Options: (1) leave as is; (2) delete from XSC (similar to incremental releases); (3) eliminate proximal sources, and insert astrophysically meaningful "sources" representing the entire galaxy -- use the 2MASS Large Galaxy Atlas (LGA).
The 2MASS extended source group (Chester, Huchra, Schneider, Jarrett) recommend option 3.
Huchra Compromise: As per the discussion at today's telecon (Feb 19), we propose to
ammend option 3 as follows. Instead of eliminating sources in proximity to large galaxies,
we instead leave them in the final catalog but null their photometry (and everything else
except ra,dec position). In this way, all sources "detected" by GALWORKS remain in the final catalog
(basically as a placeholder), with the elimination of their photometry which is undoubtedly
compromised by the poor background subtraction. **Flags are set accordingly.**
Schneider's Stopgap: In addition, we set a size limit (say diameter = 7', TBD) at which we believe
the LGA is complete, and use only those galaxies for the nulling/addition to the final catalog.
LGA
The LGA is an atlas of large galaxies built from 2MASS coadds, forming large areal coverage
mosaics. The background is carefully removed from the mosaics and the galaxy is
measured, including astrometry, orientation, SB, photometry and other metrics. The LGA was
presented at the Pasadena AAS and is described in more detail here:
2MASS Large
Galaxy Atlas. A paper is now being prepared for the A.J. presenting the
100 largest galaxies in the sky.
Thus far the ATLAS consists of ~350 galaxies, including the 300 largest as given by the RC3 (exceptions being the LMC/SMC which are not part of the XSC). It appears to be complete in diameter down to an optical (D25) diameter of 4.6 arcmin, roughly corresponding to a K-band isophotal diameter of 3 arcmin. The atlas is ever growing, with the goal to build mosaics for each galaxy that is clipped by a 2MASS scan. Several thousand galaxies may be compromised by the scan edges, so the ATLAS will take at least two years to build. For the XSC, however, the most critical galaxies are the largest ones (e.g., M31), chopped to pieces by the 2MASS pipeline, which constitute a fundamentally unique data set. These galaxies are ready to be inserted into the XSC.
Proposal
Identify XSC sources that are effectively pieces of large (LGA) galaxies and null their
photometry. Insert
LGA sources into the XSC.
step 2. Search the extended source database for sources in close proximity to LGA galaxies. Proximity is a function of the LGA galaxy position, orientation and size. We propose to use the isophotal or Kron radius as the size metric, with an additional 20% tacked on to this size to be conservative;
step 3. Visually inspect each extended source identified in step 2. Flag sources that are identified as large galaxy "pieces", *including* the piece centered on the large galaxy itself; these sources will be nulled in the final draw (XSC). Sources that are identified as background galaxies are left alone (i.e., drawn into the XSC as usual). We will use the extended source "CC" flag to identify these sources as pieces of large galaxies. A separate table of "notes" will be compiled that gives the identity of the proximal source (e.g., "piece of spiral arm"). This table can be placed in the appendix of the Exp. Suppl.
step 4. Insert the LGA galaxies into the working extended source database (or, alternatively, directly into the XSC). These objects will be flagged as coming from the LGA and thus alerting the user to the "special" status of these sources.
Important detail: the LGA does not contain all of the information that is given in the XSC for pipeline sources; hence, the LGA sources that are inserted in the database (or XSC) will have a subset of fields that are filled (and conversely, a subset of fields that are nulled). The most important fields for extragalactic work (position, orientation, SB and photometry) are part of the LGA and will be inserted into the XSC.
Concerns: (1) is the LGA ready for prime time? (2) is mixing the pipeline and LGA outputs too much for the user to handle? (3) What about the incompleteness of the LGA sample?
(1) The LGA is a maturing product, in the works for nearly two years now. Key galaxies, such as M51 and M33, have been worked and reworked many times honing the method. Since the LGA is one of a kind (there is no uniform allsky NIR catalog of large galaxies), we are working in virgin territory. However, there exists high-quality single aperture measurements from Aaronson and Co. (including John Huchra) that can be compared with the LGA H-band photometry. The results are impressively consistent (to be presented in the AJ aperp, now in prep). The algorithms and methods used to build the LGA mosaics and perform the measurements are adapted from the 2MASSS GALWORKS -- and hence have a solid lineage.
(2) Mixing sets does have precidence in 2MASS (e.g., R1 and R2-R1 photometry). The LGA photometry comes from the same algorithms as for th3 pipeline photometry. The user will not see any difference, except for a flag that states these sources come from the LGA. The XSC is greatly enhanced by the inclusion of the LGA -- the LGA is one of a kind and greatly needed by astronomers.
(3) Yes the LGA is far from completion (Jarrett will submit an ADP proposal to ask for assistence to complete the project). But the biggest and most compromised galaxies are completed and ready for insertion -- this is clearly the most important point. The current plan is to "publish" the LGA within IRSA and in NED, so the information will be available to the user as it is completed by Jarrett in the coming years(!)