Saturday, December 10, 2016

Variable Sky lines in Vega hi-med 1361cm-1, flight 290

I looked into the negative residuals effecting the sky lines in Vega himed 1361cm-1. W3 IRS5 looked fine at the same setting and the neighboring 1376cm-1 setting. 

The effect in Vega is due to a gradual overall dimming of the sky lines during the observations. A=sky is observed first, and B sky+object is observed second. When subtracted, the sky lines in A systematically over subtracting from B's sky lines.
It is not a nod A-B positional effect.

Considering only nod-off frames, the sky line flux from the first nod pair to the last is:
Vega, -8% and for W3 IRS5 -1.0%. The change in flux is roughly monotonic, so all B-A nods end up with negative residuals for Vega. It also doesn't help that Vega's continuum is fainter than W3's so residuals will be more distracting for Vega.

Here's the reduced spectra of W3 and Vega

W3 IRS5, himed, 1361cm-1 with telluric abs. model overplotted.

alpha Lyr (Vega), himed, 1361cm-1 with telluric abs. model overplotted


Variable residuals in Vega sky subtracted frames.


Here's the same rectified data, when its not nod-subtracted.

I hand extracted one order (13th from the Left).

And then compared total fluxes in the region from 300:400 (the bg) and 550:650 (skylines) over the sequence of nods. You can see the A-B undulation in the bg from having Vega's contribution.

The telescope elevation (TELEL) went from 21.2362 deg to 25.2624 deg during the obs of Vega. That is the lowest of that night. 
The elevation of W3 IRS 5 went from 31.7891 - 25.3559 deg. Other nights' targets were sometimes observed at very low elevations, but for a variety of reasons did not display this issue. (mucep 384cm-1 was low but sky lines were "subsky"ed out and also it is *very* bright). 

Ways to correct this issue: there's a subtract residual sky nod-off mode in the pipeline that does a better job at removing sky lines but resulting data numbers are frequently made negative which screws up the extraction routines. It seems like it bug which could be corrected with a little bit of work. 









No comments: