Shift in dispersion direction due to piont source target not centered on slit

[kfkaplan]

This is not necessarily a problem to be solved in the PLP perse, but it is an issue we have come across that is worth noting for future reference. When a point source target (ie. a star) is not centered on the IGRINS slit, it can induce a shift in dispersion on the detector akin to what we see with shifts from flexure. This shift is not related to the flexure, but creates a similar effect where telluric corrections then result in "p-cygni" type residuals for deep telluric lines. This shift is normally subpixel. It can be seen in the 2D echellograms as a shift in the telluric absorption lines relative to the sky emission lines, which should not move (barring flexure). It mostly occurs in data from McDonald Observatory where the slit is widest on the sky when guiding on the target is poor due to bad seeing, wind, or being too dim to track on. The latest PLP version (currently branch reimplement_cr_reject but will be part of a future version 3) outputs text files in the outdata directory for a night called telluric_shift_H.csv and telluric_shift_K.csv. These show the shift in pixels in telluric absorption features from the cross-correlation between individual exposures within a single set of A or B nods on a target.

There is no universal solution within the PLP to try to fix this but should probably be addressed on a case by case basis as follows:

• If the shift is from a single exposure in a target with many exposures, the single exposure could be excluded from the data reduction for that target.
• Portions of the spectrum with bad residuals where there are deep telluric absorption features could be masked out.
• A target with this shift in all its exposures could possibly be shifted in pixels to match its standard star.