DESI Early Data Release CV Spectra Not Displaying

Question

Hi,

DESI Early Data Release Spectra displays for star subtypes: B, A, F, G, K, M fine.

https://www.legacysurvey.org/viewer/desi-spectrum/edr/targetid39633422540342641

https://www.legacysurvey.org/viewer/desi-spectrum/edr/targetid39633402688703635

https://www.legacysurvey.org/viewer/desi-spectrum/edr/targetid39633399773663251

Subtype CV seems to have issues:

https://www.legacysurvey.org/viewer/desi-spectrum/edr/targetid39633411349941774

https://www.legacysurvey.org/viewer/desi-spectrum/edr/targetid39633411349940312

https://www.legacysurvey.org/viewer/desi-spectrum/edr/targetid39633405607936637

Joseph Karpinski

Answer 1

I think you have found some "bad luck" spectra. Looking in to the actual data files, these have inverse variance set to zero for the entire spectrum, which means the entire spectrum should be regarded as "bad". The classification of CV should also be regarded as doubtful in this case.

Comments

Well, not sure.

There are over 117,000 DESI-EDR CV stars.
It looks like none of them have a spectra.

A  count: 2288
B  count: 235
CV  count: 117877
F  count: 40012
G  count: 125959
K  count: 257266
M  count: 92047
WD  count: 4963

Please review.

Thanks

---

The vast majority of the CV objects you have identified have a non-zero value for ZWARN (see, e.g. https://data.desi.lbl.gov/doc/releases/edr/known-issues ), which indicates a problematic spectrum that should not be trusted.  Furthermore, of those CV objects with non-zero value for ZWARN, most have the bit flag 0x200 set, indicating, "No data for this fiber, e.g. because spectrograph was broken during this exposure (ivar=0 for all pixels)". The classification and redshift are automated, and in this case it appears that CV is the classification that is attempted even if there is no data. We have also created https://github.com/desihub/redrock/issues/262 to address this in the future.