Strange single broad-line spectrum, an extraction artifact?

[AlexGKim]

There is a spectrum that doesn't match my limited knowledge of astronomical objects. It may be real but I suspect it could be an extraction artifact.

plot_spectra -i /global/project/projectdirs/desi/spectro/redux/denali/tiles/pernight/80635/*/coadd-4-80635-*.fits -t 39627628532992432 --rebin 4

[AlexGKim]

Another example

plot_spectra -i /global/project/projectdirs/desi/spectro/redux/denali/tiles/pernight/80635/*/coadd-4-80635-*.fits -t 39627628537186687 --rebin 4

[sbailey]

This looks like another case of fidicials being on; see #1168.

Action items:

• track down which exposures have the bump but weren't already flagged as bad
• check neighboring fibers on the focal plane (e.g. with flux in 4400-4800 vs. x,y)
• flag exposures as bad

[moustakas]

For the record, both of these examples are present in Everest as well. Not sure how we're handling these in Fuji, if at all.

[sbailey]

There is contamination on night 20210102 expids 70543 - 70545, but not on 20210105 exps 70931-32 for the same tile 80635.

Action: flag 20210102 expids 70543 - 70545 as bad

Note that petal 3 on 20210105 had bad positioning, and is already flagged as bad in the exposure tables.

@akremin

Not fiberflatted at all, but median of frame counts 4400 - 4800 A on two different nights for same tile:

[akremin]

Expids 70543 - 70545 on 20210102 are now marked as bad