Incomplete retrieval of ALLWISE reference catalog - solved (kind of)

[bstecklum]

My photometry code fails when using ALLWISE as a Scamp reference catalog but works fine for, e.g., 2MASS. The reason is a lack of common objects, as illustrated by the attached figure. It shows detections (green), 2MASS stars (cyan), and ALLWISE objects (red). The retrieval of NEOWISE sources is incomplete. This is for SCAMP version 2.10.0 (2020-12-12).
I recommend to replace ALLWISE with the more recent catWISE2020 catalog.

[bstecklum]

Turns out that this was due to an incorrect condition in astrefcat.c, intended to avoid contaminated observations.
Commenting it out solves the issue.
case ASTREFCAT_ALLWISE:
/*-------- Avoid contaminated observations /
sflag = cols[cindex++];
/
if (sflag[0]!='0' || sflag[1]!='0' || sflag[2]!='0' || sflag[3]!='0')
continue;
*/

[ebertin]

Thanks for reporting this. Well, it depends on the sky region. For instance, in regions with many bright stars, I found that keeping flagged detections adds a lot of spurious sources along that impact astromatic matching and the solution. I will add support for the CatWISE2020 catalog.

[bstecklum]

For the above star forming region, scamp worked well after the code change. Apart from bright stars, extended dust emission becomes important for the longer WISE wavelengths, in particular close to the Galactic plane. This affects the ccf flags, as can be seen from the following numbers. For the 808 ALLWISE sources, the counts of zero value of the flag are 381, 260, 0, and 0 for W1, W2, W3, and W4. You might consider omitting the check for W3 and W4 to make it less strict.