Bug Trails in Final Product Tiles
In the list of highest-quality tiles to be used in Final Product Generation a total of 47 scans in 23 nights have one or more instances of residual bug trail imprints in the coadds. The goal is to evaluate what impact these trails have on potential release data. The two key elements checked are:
Indications are that the impact of the bug trails on both overall reliability and data quality are minimal and that these scans should be included in the final release product. The spurious sources introduced by the bug trails have been cataloged for potential removal from the release catalog.
A list of the affected scans is presented below.
Overall the bug trails show up in the coadds as K-bright trails about 1-2 DN higher than surrounding backgrounds. Some of the strongest trails can depress the H backgrounds by just under 1 DN lower than surrounding backgrounds (but in most cases there is no obvious H imprint at all). Generally there is no detectable impact on J backgrounds. The tiny values of the "holes" at H suggest there is almost no loss of flux due to shading by the bug since the fractional drops in background level are only at the 0.1% level or smaller. The large spots in the image are thus from trivially small bugs that are strongly defocussed on the aperture window.
Datehem Affected Scans ------- ----------------------- 980421n 074 075 980422n 067 076 980503n 037 081 980505n 013 980510n 010 011 980516n 055 980529n 038 039 080 980601n 053 054 980605n 070 980618n 026 067 980710n 036 981122n 028 990521n 016 990607n 136 138 139 990608n 014 015 019 028 029 094 990609n 014 020 033 041 990610n 014 015 017 080 990611n 044 045 990612n 013 990622n 015 990813s 047 073 081 991011n 049 000423n 029* *note that this scan has been dropped from the list of v3 tiles used for final product generation in favor of 980212n 078 (also q=1)
In the following summaries, all questionable point and extended sources have been catalogued for followup and possible removal from the dataset.For these nights we find the following breakdown:
The data tables (including unique counter ID) are available below:
Very bright trails are seen in coadds 80 & 91 of scan 067, and lots of false K point sources are found here. Scan 076 barely has a short trail in coadd 232, and two false sources are noted. 70 total this night. NO false extd sources noted
Overall faint imprints in coadds, only slightly enhanced at K. No false point sources noted in the two affected scans, but one questionable extended source was noted.
Overall faint imprints in 3 coadds, only slightly enhanced at K. No false point sources noted in the two affected scans, but one questionable extended source was noted.
Bug trails are moderately bright and affect 12 coadds in scan 010, and 3 in 011. No false point sources appear to be generated by them (seems to happen only with brighter trails, often along sharp edge artifacts). However, 5 false extd sources are generated in scan 010, and a total of 3 real point sources fall along the trails and may suffer corrupted photometry.
Significant lengths of trail were successfully cleaned up by meteor blanking on this night, but 5 coadds of scan 055 still show moderate bug residuals. One real extended source does fall in a trail, but no false point sources are noted.
Two scans show moderate trails spanning 3 coadds in scan 039 and 2 coadds in 080. No false extended sources noted. A tight cluster of 12 suspicious point sources is noted in a segment of scan 080 affected by a trail.
Moderate trail starts midway through scan 053 and continues through first two coadds of 054. No false point sources noted. 9 false extended sources noted in scan 053, all K-bright.
Moderate trail appears in 5 coadds. No false point or extended sources noted.
Weak trail seen cutting across two scans. 4 false extended sources noted (2 per scan), but no false point sources seen.
Very bright trail in first two coadds of scan 036 that generates 10 false extended sources. A total of 92 false point sources have been cataloged.
Very bright trail in last coadd of scan 028 generates 62 false point sources but no false extended sources.
Very faint trail in two coadds of scan 016 generates 1 false extended source but no false point sources.
Faint trail shows up in scans 136-132 but generates no false extended or point sources.
Trails in 4 scans almost entirely cleaned up by meteor blanking. Only slight glitches in coadds 162-173 remain in scan 014, but no false sources are generated.
Trails in 4 scans, though scan 017 appears to have been completely cleaned up by meteor blanking. Faint bug artifacts remain in scans 014, 015, & 080. No false point or extended sources are noted, but 4 real extended sources lie along the residual trails and have corrupted photometry.
Faint trails seen in scans 044-045. No false point sources noted. 2 false extended sources are present, as well as 2 real extended sources that fall within the trail.
Faint trail seen in last 2 coadds of 013. No false point or extended sources noted.
Bright short trail seen in 1 coadd of 015. No false point or extended sources noted.
Very bright segments seen in 3 scans; all are partially blanked with sharp edges. No false extended sources noted. 24 false point sources extracted.
Faint trail seen in last 4 coadds of scan 049. No false point sources noted. 3 false extended sources extracted.
Faint emission features seen in both H (unusual) and K in first 7 coadds of scan 029; fortunately another instance of this tile has been substituted for final product generation, so the bad sources have not been tabulated.
A first analysis has been made of 990608n since this night has the largest number of affected scans and that the bug imprints are among the worst in the sample.
Only the final two coadds (256 & 266) are affected but have no obvious false point sources. No false extended sources detected.
A very bright trail is evident throughout coadds 009-091, and 12 questionable point sources are noted. No false extended sources detected.
A full discussion of each affected scan follows (based on visual inspection of the coadds). For point sources there is only one likely false K-only source generated by the bug trail that meets final catalog selection criteria, thought the bug trails generate multitudes of poor quality detections. There are two suspicious extended sources in scan 094 that may be triggered by the K halo.
The only affected coadd is 232, and only 4 sources are potentially in the affected area (all of which are unambiguously real).
The overall success in source selection criteria screening out the spurious detections can be seen by comparing images of source overlays of all sources and catalog quality sources (image of coadd 232).
The only affected coadds are 044 056, and only 4 sources are potentially in the affected area (all of which are unambiguously real).
In the two affected coadds (232 244), no catalog-quality sources are in the areas near the edge contaminated by bug trails.
The worst scan in this night, trails are seen in coadds 021 044 068 080 091 103 197 209 232 256 266. Out of a total of 33 point sources possibly affected by bug trails, 2 appear questionable.
The overall success in source selection criteria screening out the spurious detections can be seen by comparing images of source overlays of all sources and catalog quality sources (image of coadd 232).
Coadds affected include 009 021 032. Of 20 potenially affected point sources, only one appears to be a marginal detection:
Coadds affected include 021 033 044 056 068 (bug meanders diagnoally across scan). This scan is higher density than the others this night, and about 150 point sources are potentially affected by the bug trail. Of these sources, only 3 look marginal (but none of these is a K-only detection).
This scan also provides a good opportunity to study possible effects on photometry from the bug trails (it has higher source density than the others this night and the trail affects only about half the scan with a clear separation in Dec).
Color-color and color-mag plots show the bug trail to create no distinct population of sources (either false sources or systematically shifted photometry for real sources) in this scan. In the plots below, red points are from the portion of the scan affected by the bug trail while blue points are from the remainder of the scan. Only JK detections are plotted in the color-mag plot and the color-color plot has only 3-band detections.
Catalog quality sources: color-color plot ... color-mag plot
From the catalog quality plots it is clear that the bug trail creates no distinct population of sources (either false sources or systematically shifted photometry for real sources) in this scan. The large number of spurious detections are almost all K-only and thus do not contribute to these plots. In addition, there is no sign of systematic color shifts induced in the real sources, though admittedly the total number of potentially affected sources is quite small.
There are two likely false extended source detections in this scan. Both appear point-like in the coadds and show no elongation in the extended source extractions.
