Spectropolarimetry Reductions Primer
For all images with a common format:
For each grating and/or grating setting:
Convert from FITS to .imh format
SECTION the FITS images to an input
list (e.g. "SECTIONS f* > inlist")
COPY the input list to an output list
EDIT the output list of filenames to
desired (e.g. fa.001 > a001)
RFITS the images from the "fits_file"
= @inlist to the "iraf_file" = @outlist
Remove bias and fix bad columns
BIASFIX does this in one step.
The parameter file should appear something like:
(biassec = "[1201:1220,*]")
(trimsec = "[2:1196,*]")
(biasorder = 9)
(mask = "badpix")
The bad pixel mask file should be checked against a test image before running.
Prepare flat-field ratio image for each aperture used
- Combine multiple flat-field images to a single image, using median filtering to remove cosmic rays and reduce noise.
e.g., IMSUM a001,a002,a003,a004 q600_2.sum option=median
Determine x and y boundaries of the two polarized spectra using SPLOT
and edit these into POLFLAT, for example:
POLFLAT is run interactively, adjusting the fitting orders until
an acceptable fit is achieved.
Compute a 2-D dispersion solution for each polarized spectrum
POLID begins line identification from
input centerlines of the two polarized spectra, a coordinate database file
(one of those in home$linelists), and calls REIDENTIFY as well as
FITCOORDS to prepare dispersion solutions. Representative
input parameters are:
Remove cosmic rays
This can be done in one of several ways, including my script IMCLEAN
Reduce flux standards, followed by polarization standards
image1 = "a065"
image2 = "a066"
stoke = "v"
imflat = "q600_2.rat"
outroot = "aruma_9"
imfit = "lamp600"
lam1 = 4000.
lam2 = 8000.
dlam = 4.
invgain = 0.79
noise = 6.
subtsky = yes
okvalues = yes
delab = yes
(nwide = 30)
(torder = 7)
(review = yes)
Shift dispersion zero-point of standards to correct for flexure
This is done using SPECSHIFT and
typically amounts to sub-pixel adjustments, using the output of POLRED.
Note that since SPECSHIFT only adjusts the dispersion solution zero-point
(e.g. CRVAL1), it is 100% reversible.
Compute spectral response functions
As in normal spectrophotometric reductions, these are done using STANDARD
and SENSFUNC. Note that the latter
often requires a rather high spline order (13 or more) to adequately fit
the bumps and wiggles in the response function.
Reduce all other observations
This uses POLRED as per the standard
Shift dispersion zero-point to correct for flexure
Using SPECSHIFT as before.
Compute corrected total flux spectra
The task POLFLUX is self-explanatory,
using the response functions computed previously.
Average multiple flux spectra into one mean
Use AVGFLUX. Note that each
input spectrum is entered explicitly or via wildcards, not as a root name.