julienguy
commented
2 years ago They are due to cosmic rays aligned with the fiber traces, which are difficult to flag for bright stars because of the Poisson noise in the spectral traces.
ds9 /global/cfs/cdirs/desi/spectro/redux/fuji/preproc/20210105/00070933/preproc-r5-00070933.fits /global/cfs/cdirs/desi/spectro/redux/fuji/preproc/20210105/00070934/preproc-r5-00070934.fits(count 152 fibers starting from the left)
Same images with the masked pixels set to zero. One can see some unmasked pixels.
desi_preproc --zero-masked -i ~/data/20210105/00070933/desi-00070933.fits.fz --cam r5
desi_preproc --zero-masked -i ~/data/20210105/00070934/desi-00070934.fits.fz --cam r5
ds9 preproc-r5-00070933.fits preproc-r5-00070934.fits
Target 39633178008228311 was observed several nights and usually does not have strong emission lines. For example in fuji per night a typical night
plot_spectra -i 20210107/spectra-5-80675-20210107.fits -t 39633178008228311
But on one night there appears an emission line. In the following plot there are two exposures for one night. In each exposure there is a sharp emission line, whose position is slightly different between exposures.
plot_spectra -i 20210105/spectra-5-80675-20210105.fits -t 39633178008228311