The nature of use_src=0 dup_src=0 sources with high SNR
Below are examples of a dozen bright (typically Ks<13.5)
rd_flg="222" cc_flg="000" sources for which use_src=0
and dup_src=0.
Observations:
- Most all of the source examined are real point sources.
- The most common cause for rejection from the catalog is that they were influenced by persistence on one overlapping scan but not on the other (due to the alternating scan direction).
Conclusions:
- These sources should be rejected from the catalog.
At scan center certain sources have a 50/50 chance of being rejected due to persistence of nearby bright sources depending on scan direction. At scan edges such sources have two chances at detection -- usually with alternating scan direction because of the survey strategy -- and thus close to 100% chance that they will be included in the catalog. If the goal is to have the overlap regions statistically indistinguishable from the scan center regions then there needs to be a mechanism to reject 50% of the good detections of such sources. The duplicate resolution acts exactly in this manner.
Example images
Note: If a pair of images appears below the circle on the left image is centered on the source in question. The requested source position had to be shifted in order to get the adjacent scan's coadd to appear in the image server so for subsequent images the location of the circle is arbitrary. The starfield must be used to locate the source.
Example 1 -- counterpart lost due to persistence
Example 2 -- counterpart lost for unknown reasons
Example 3 -- counterpart lost due to persistence
Example 4 -- counterpart lost for unknown reasons
Example 5
-- counterpart(s) lost for persistence (obviously...)
Example 6 - persistence
Example 7 - persistence
Example 8 - persistence
Example 9 - persistence
Example 10 - persistence
Example 11 -- faint source, unknown cause for rejection
Example 12 - this source is rd_flg="200" cc_flg="000" just to show an example. It is likely an artifact of a bright star. It has J-snr just above 10, but is not detected in the other bands. Although it lies in the overlap region the WSDB has only one entry.
Mike Skrutskie
April 16, 2002