2MASS V3 Production Processing

R1 n&m<3 Sources Analyzed


The incompleteness issues analyzed below were resolved.

Summary

The following analyses were done to find R1 real sources would be lost due to requiring n&m>2 in at least one band. That subset of the V3 working survey data base entries available on 020215 that were 3band, not flagged as artifacts, were not within 10" of a scan edge, were read type 1 or 3 in at least one band, and had n&m<3 in all 3 bands were investigated (SQL).

The summary:

  • There are 105 real sources (107 database entries) that might disappear completely, depending on the scans used for the final catalog draw.
  • 57 of these are confused sources, the rest are apparent POSFRM merge failures.
  • The completeness problem could be fixed by forcing these entries into the catalog.

    Analysis

    This query was run Feb 15 2002 at 18:09, with the following results:

    
    	918  entries were found matching the script criteria:
    
    	744  were halo junk, meteor or satellites,  or confusion with a catalog quality source
    	 32  were multiples with at least one catalog quality entry
    	 21  were confused objects with a catalog quality r2 entry
              4  were POSFRM merge failures with a catalog quality entry in the same scan
              7  were POSFRM merge failures with a catalog quality r2 entry in the same scan
    
    	Of the last 4 catagories, many have non-optimal r1 information, but have catalog 
    quality entries with reasonable r2 or r1 fluxes. 
    	
    	The 107 with potential completeness problems are as follows:
    
    	 15  were confused objects with a catalog quality entry in another scan
             33  were POSFRM merge failures with a catalog quality entry in another scan
    
    	 The above 2 catagories may result in incompleteness depending on the dup resolution and 
    final catalog draw. Those missing in the final draw can be saved by forcing a good representation
    from an overlapping scan into the catalog.
    
    	The last 3 catagories have no catalog quality entry in the datatbase and will result in 
    bright real sources missing in the catalog:
    
    	 42  were confused objects with NO catalog quality entry in the database
              4  were on 2 real sources in the halo of Antares with NO catalog quality entry
             13  were POSFRM merge failures with NO catalog quality entry in the database 
    
    	These can be salvaged by forcing the (best) database entry into the catalog. These 
    representations do not have all the data properly merged into the fluxes, but most have reasonable
    fluxes in spite of these problems.
    
    Note that confused objects often result in merge problems in POSFRM, which can result in low n&m counts.

    Proposed Fix

    It is proposed to search the database after the preliminary catalog draw within 2" of the 105 positions to identify those sources actually lost, and force entries as appropriate to avoid incompleteness for that subset.

    The annotated list of the 105 potential lost sources may be found here.

    The detailed list (with counters) of the 107 database entries may be found here.

    The detailed list (with counters) of the 105 database entries to be used if no catalog quality source survives at each position is here. This list contains the best representation of the source at each postion. The catalog will be searched within 2" of the positions in this list. The database records for which no catalog quality entry is found will be forced.

    Final List

    The detailed list (with counters) of the 44 database entries with no catalog quality source surviving dup resolution at each position is here. This list contains the best representation of the lost source at each position, these entries were forced into the catalog (12/15/02) so these sources would not be lost.

    021126 Addendum

    By examining bright (rd_flg 3 or mag<6) V3 database entries marked as artifacts that had no bright source in the catalog draw within 15", an additional 13 missing R1s were found. These sources are missing from the final catalog.

    Of this list,

  • 10 were lost due to picking up rd_flg 9 artifacts in a band (6 in Ks, 4 in H).
  • 1 was lost due to being flagged as a persistence artifact
  • 1 one was a double lost due to confusion problems.

    6 of these were also found in a list of Tycho stars that matched only artifacts.

    It was also discovered 775 R1 sources were inadvertently omitted from the catalog draw due to an oversight in the selection criteria omitting sources partially R1 saturated in all 3 bands. These entries were also forced into the catalog (12/15/02).


    This page last updated on Dec. 4, 2002.

    Gene Kopan - gene at ipac.caltech.edu