Dealing with sky not aligned with lines/columns

Frank Valdes wrote on Jun 26, 1998

Q: I would like to learn how to correct my spectra for a slant, I explain.
The image containing a long slit spectrum with wavelength along the x 
axis and slit position along the y axis, shows emission lines which are 
slanted, that is, they are not prependicular to the x axis. How can 
correct for this using IRAF?


A: There are two types of spectral reductions for long slit data.  If the
spatial information does not matter then spectra are extracted by
summing across "apertures" with or without sky subtraction.  The
package used is APEXTRACT though the tasks are found in the generic
spectral reduction package SPECRED.  The summing is along lines or
columns (though the position of the object is adjusted at each line or
column along the dispersion) so if the change of the sky lines over the
region near the object being extracted is significant then the
simultaneous sky subtraction will not work well.  However, in this case
one can extract a sky region and the object region without subtraction,
do wavelength calibration for both regions, and sky subtract in
wavelength (which corrects for the wavelength shift due to the tilt).
One can also tweak the sky subtraction (before or after wavelength
calibration) with a new task called SKYTWEAK.  For a guide to slit
spectrum reductions of this type (plus a small addendum on long slit
reductions) see

ftp://iraf.noao.edu/iraf/docs/spect.ps.Z

When spatial information matters or to deal with distortions, including
the tilted slit image, the LONGSLIT package is used.  In this technique
the positions of objects in the slit and arc lines (or sky lines) are
traced to find a 2D distortion/dispersion function.  The function is
used to resample the data to another long slit image where the dispersion
and spatial axes are exactly aligned with image axes.  The sky can
then be subtracted with the BACKGROUND task or one can then start
with the aperture extraction tasks.  There is no on-line user's guide
for long slit reductions other than the task help.  A very old document
which describes the approach and some of the tasks is

ftp://iraf.noao.edu/iraf/docs/lslit.ps.Z

---

Last post on Jun 26, 1998