Scans Needing Special Handling During Version 3 Processing

Scans Requiring Special Handling in Version 3

Quality checking of the v2 output has uncovered a few problems not always detected during nightly QA analysis.

When problem scans were found, the observatories were alerted and generally the bad tiles were reacquired with acceptable scans. For the scans discussed below, however, higher quality rescans were not obtained due to unfavorable sky placement for late-mission operations, due to bad luck, or due to the fact that the problem was uncovered only after observatory operations had ended.

Each of these scans will therefore require special handling during v3 processing to assure that they meet the standards of the rest of the survey. It should be noted that the v3 pipeline will not (in most cases) handle these cases automatically. Rather, each will involve its own hand manipulation:

1) H glitches -- This is a problem thought to be unique to the northern H array. In a handful of scans identified to date, a single frame has greatly increased noise, and this imprints into the coadds and creates hundreds of false sources. After a few such cases were detected serendipitously by looking at coadd images, a tool was developed by Gene Kopan to help flag this problem in new data. The tool was also run retroactively on older data to identify other scans not previously diagnosed with the problem.

2) Clouds -- For scans which had no overlapping scans on the night of observation, the presence of a cloud might be detected after comparing that scan to adjacent scans taken on other nights. For such cases the cloud might not be obvious in the background plots, used during QA as a surrogate photometric check when no overlaps are available. For all such singleton scans, we have run a tool which pairs that scan with its overlapping neighbors in the database. Sherry Wheelock has written code to do the comparison and to flag scan pairs with large deltas or sigmas. This tool has now been run on the entire working database (both hemispheres) and the worst cases have been identified.

3) Zero-point problems -- The above mentioned overlap tool also showed that there was an occasional block of scans on a night for which the overlaps indicated a bad zero-point determination for that block. These are photometrically stable sci scans for which the photometric solution obviously went astray.

4) Astrometric problems -- Occasionally, pattern mismatches or large frame steps were found during post-processing analysis of the v2 data. See Howard's memo for more information. In these cases, we need to pay special attention to the astrometric reconstruction in v3 to see if hand processing may be warranted to recover a better astrometric solution.

Below is the list of problem scans, ordered by observation date.

980319s, most scans on night

These are scans which need to be monitored for their astrometric reconstruction as they come through the pipe because they've either been rerun or for some other reason there remains some question about their dtp (date and time of processing) tags.

Have now been run through v3 processing. The astrometric reconstruction for these scans is found to be within tolerances.

980801s, Scans 054-057 (Tiles 329730-329733)

   TILE   DATE  SCAN   Ra      Dec   Qc Qo Qn Qs  S 
 329730  980801 054 250.1062 -86.962 10 10 10 10  1
 329731  980801 055 251.2670 -86.962 10 10 10 10  1 
 329732  980801 056 252.4248 -86.962 10 10 10 10  1
 329733  980801 057 253.5809 -86.962 10 10 10 10  1

Have now been run through v3 processing. The astrometric reconstruction for these scans is found to be within tolerances.

980921n, Scan 007 (Tile 29791)

   TILE   DATE  SCAN   Ra      Dec   Qc Qo Qn Qs  S 
  29791  980921 007 320.6633  86.963 10  3  3  5  0

This is a scan which needs to be monitored for its astrometric reconstruction as it comes through the pipe because it's either been rerun or for some other reason there remains some question about its dtp (date and time of processing) tag.

980924s, Scans 053-054 (Tiles 302923-302924)

   TILE   DATE  SCAN   Ra      Dec   Qc Qo Qn Qs  S 
 302923  980924 053 352.0942 -2.9276 10 10 10 10  1
 302924  980924 054 352.2150 -2.9270 10 10 10 10  1

Have now been run through v3 processing. The astrometric reconstruction for these scans is found to be within tolerances.

980926n, Scans 069-075 and 082 (Tiles 29552-29559)

   TILE   DATE  SCAN   Ra      Dec   Qc Qo Qn Qs  S 
  29552  980926 069  44.0040 86.9688 10 10 10 10  1
  29553  980926 070  45.1561 86.9680 10 10 10 10  1
  29554  980926 071  46.3116 86.9682 10 10 10 10  1 
  29555  980926 072  47.4693 86.9683 10 10 10 10  1 
  29556  980926 073  48.6266 86.9686 10 10 10 10  1
  29557  980926 074  49.7864 86.9687 10 10 10 10  1
  29558  980926 075  50.9477 86.9687 10 10 10 10  1
  29559  980926 082  52.1147 86.9695 10 10 10 10  1

980927n, Scans 070-071, 074, and 076 (Tiles 29545-29546, 29549, 29551)

   TILE   DATE  SCAN   Ra      Dec   Qc Qo Qn Qs  S 
  29545  980927 070  35.9082 86.9674 10 10 10 10  1
  29546  980927 071  37.0626 86.9671 10 10 10 10  1
  29549  980927 074  40.5401 86.9683 10 10 10 10  1
  29551  980927 076  42.8578 86.9683 10 10 10 10  1

Have now been run through v3 processing, after a redelivery of the Martin files. The astrometric reconstruction for these scans is found to be within tolerances.

981028s, Scan 122 (Tile 325574)

   TILE   DATE  SCAN   Ra      Dec   Qc Qo Qn Qs  S 
 325574  981028 122 69.2051 -62.9269 10 10 -9 -9 -9

This scan has a large frame step that appears to be caused by a real astrometric slip further compromised by poor astrometric reconstuction near the step. The health of this scan's astrometry needs to be checked carefully after v3 processing to see if special handling is required. Additional info can be found on Howard's pages.

Has now been run through v3 processing. The astrometric reconstruction for this scan is found to be barely within tolerances, so the scan has been flagged for possible uncertainty adjustments later. (Note that the cause of the problem is the bright star R Doradus.)

981122s, Scan 017 (Tile 321410)

   TILE   DATE  SCAN   Ra      Dec   Qc Qo Qn Qs  S 
 321410  981122 017 335.7432 -44.927 10 10 -9 -9 -9

This is a scan with a reconstruction step problem. Unlike similar problem cases 981130n s041 and 000113n s143, the sharp astrometric step seen here occurs in the middle of the scan. We need to assess the quality of the v3 astrometric reconstruction here to see if special handling will be needed. Additional info can be found on Howard's pages.

Have now been run through v3 processing. The astrometric reconstruction for this scan is found to be barely within tolerances, so the scan has been flagged for possible uncertainty adjustments later. (Note that the cause of the problem is the bright star Pi-1 Gruis.)

981130n, Scan 041 (Tile 88017)

   TILE   DATE  SCAN   Ra      Dec   Qc Qo Qn Qs  S 
  88017  981130 041 103.7074 44.4894 10 10 -9 -9 -9

This scan has a large frame step that appears in the first couple of frames. The health of this scan's astrometry needs to be checked carefully after v3 processing to see if special handling is required. We could presumably drop the first couple of frames entirely, for example, and not create a hole on the sky. Additional info can be found on Howard's pages.

990312n, Scan 083 (Tile 20793)

   TILE   DATE  SCAN   Ra      Dec   Qc Qo Qn Qs  S  D   AF5  Hpsp  Kpsp Bla0 Bla1 Bla2
  20793  990312  83 235.010  44.930   4  4  4  4  1  1  -9.9  10.2   7.9   1    0    0

This is an H-glitch scan, and the bad frame is evident in coadd 173 (bottom) and coadd 185 (top). Gene has found a way to blank out a single frame in a scan, though an end-to-end test through the pipeline will be needed to verify that the fix works without further ramifications.

990419s, Scan 135 (Tile 305155)

   TILE   DATE  SCAN   Ra      Dec   Qc Qo Qn Qs  S  D   AF5  Hpsp  Kpsp Bla0 Bla1 Bla2
 305155  990419 135 263.680  -8.928  10 10 -9 -9 -9  -  -9.9   7.0   8.6   0    0    0

This is a scan with a pattern mismatch. (Refer to Howard's memo for more details.) For v3 processing we need only to insure that the proper fpos file is used for this scan when the night is processed as the scan itself had to be run special through v2 software to produce the correct position reconstruction.

990715s, Scan 056 (Tile 320944)

   TILE   DATE  SCAN   Ra      Dec   Qc Qo Qn Qs  S
 320944  990715 056 259.6622 -44.930 10 10 -9 -9 -9

This is a scan with a pattern mismatch. For v3 processing we need only to insure that the proper fpos file is used for this scan when the night is processed as the scan itself had to be run special through v2 software to produce the correct position reconstruction.

991129n, Scans 015-016 (Tiles 24980-1)

   TILE   DATE  SCAN   Ra      Dec   Qc Qo Qn Qs  S  D   AF5  Hpsp  Kpsp Bla0 Bla1 Bla2
  24980  991129  15 296.291  56.933   5  5  5  5  1  1  -9.9   8.2   8.4   0    0    1
  24981  991129  16 296.497  56.933   5  5  5  5  1  1  -9.9   8.4   8.6   0    0    1

Both of these suffer from a slightly incorrect zero-point determination, as judged from overlaps in the database. We will run these tiles normally through 2MAPPS v3.0 and apply a zero-point offset fudge factor to the scan's .cal file after we have the surrounding tiles for comparison. As can be seen from the plots the size of the correction for scan 15's overlap to the west and for scan 16's to the north and to the south is about 0.05 mag in all bands based on v2 processing.

991129n, Scan 030 (Tile 28773)

   TILE   DATE  SCAN   Ra      Dec   Qc Qo Qn Qs  S  D   AF5  Hpsp  Kpsp Bla0 Bla1 Bla2
  28773  991129  30 311.407  74.935   3  3  3  3  1  1  -9.9   8.9   9.1   0    1    1

This scan has a brief cloud as shown in its complimentary overlaps to the east and to the west. The correction will be similar to the fudge factor described above, but the size of the photometric correction will depend upon the declination of the source. (That is, the cloud was most noticeable at one point in the scan, which is where the correction is at a maximum of about 0.2 mag based on v2 processing.) Again, the correction itself will be computed after the surrounding, comparison scans have also been run through 2MAPPS v3, and this correction will be applied to the scan's .cal file.

000113n, Scan 143 (Tile 10547)

   TILE   DATE  SCAN   Ra      Dec   Qc Qo Qn Qs  S 
  10547  000113 143 216.5555 20.9307 10 10 -9 -9 -9

000116n, Scan 058 (Tile 18261)

   TILE   DATE  SCAN   Ra      Dec   Qc Qo Qn Qs  S 
  18261  000116  58 196.1847 38.9303  9  9 -9 -9 -9

000127n, Scan 091 (Tile 7338)

   TILE   DATE  SCAN   Ra      Dec   Qc Qo Qn Qs  S 
   7338  000127 091 172.5648 14.9309  4  4 -9 -9 -9

000218n, Scan 014 (Tile 29270)

   TILE   DATE  SCAN   Ra      Dec   Qc Qo Qn Qs  S  D   AF5  Hpsp  Kpsp Bla0 Bla1 Bla2
  29270     218  14 217.669  80.934   3  3  3  3  1  1  -9.9   9.3   8.8   0    1    0

This scan also has a brief cloud, as can be seen from its complimentary overlaps to the east and to the west. Here, the maximum deviation caused by the cloud is about 0.15 mag based on v2 processing. Its correction will be handled the same way as that for 991129n, scan 030 above.

000928s, Scan 114 (Tile 203271)

   TILE   DATE  SCAN   Ra      Dec   Qc Qo Qn Qs  S 
 203271  000928 114  34.3910  9.0725 10 10 -9 -9 -9

D. Kirkpatrick - IPAC
Last updated 24 October 2001