gully
commented
3 years ago Here is a plot of the mean and variance spectrum of the (provisionally-telluric corrected, barycentric-corrected) Luhman 16A spectra. I compare the variance of all of these target spectra to the variance of the A0V spectra themselves. I also overplot the best-fit-grid-point-model, doctored to match the observed data's kinematic properties. We see excellent---albeit imperfect---agreement between the data and model.
The spectra are not all the same signal-to-noise ratio, so comparing the variance spectra in this way takes a bit of care, and the exercise serves as a mere stepping stone to our ultimate goal of inferring the per-visit perturbations. Here we can get a clue of where the most conspicuous perturbations can be, and at-a-glance we can understand if the telluric variations could be dominant.
Visits 1, 2, and 4 share the same A0V calibrator. The visit 3 spectrum in green exhibits conspicuous departures in the region of the strong A0V absorption lines, just as we'd expect (see link in above comment for the context).
The next step is probably to visually flag those regions where the targets exhibit high variance above-and-beyond the A0V variance.
The default IGRINS
plpleaves divides the target spectrum by an unprocessed A0V template. This step can imbue large humps in the spectrum for reasons I pointed out in this 2017 plp Issue comment.