Introduction

IRAF tasks are grouped into packages which contain a number of subpackages. Each task has its own parameter file which can be be edited according to your requirements. The edited parameter files are stored in the parameter file directory, defined in the command file 'login.cl'. For every task the exhaustive on-line help facility is available, i.e. if you encounter any problems, you can type 'help taskname' to see what is going wrong.

The first step at the beginning of work with an IRAF is to create your IRAF environment. For this purpose serves the file 'login.cl'. In this file you define the names of your working directory, data directory and parameters directory, set the graphic device and image display. Now, you can start the IRAF session by typing 'sir', which is already present in your directories. You will see the image display window arriving on your monitor and the inviting message <cl.

1.

Converting FITS data to the IRAF images. Use the IRAF task 'rfit'. Create a file with the names of all necessery FITS files. This can be done by the command

<cl files /home/norfa99/ser/algis/BI*.fts $>$ biaslist

for the files that names begin from the letters 'BI'. Then, by typing

<cl rfits @biaslist "" BIAS

you will get a full set of IRAF images BIAS0001, BIAS0002,... with the extensions .imh in your working directory and extensions .pix in your data directory.

2.

Examination of images. You can use the tasks 'implot' or 'display' and 'imexamine' image. The help is available from inside of these tasks: just type "?". The useful commands at this step are ':c column No.' and ':l line No' (in both tasks).

3.

Overscanning and trimming the BIAS images. The reductions of CCD data takes place within the 'ccdred' package. The tasks in this package rely on the image headers, so we need first to set the translation table between image headers and the internal IRAF language. The command for setting up the translation is 'setinstrument' in noao.imred.ccdred package. So type
'noao', 'imred', 'ccdred' consequently and <cd setinstrument

Now, typing '?' you will have a list of image header's translators available. If you choose 'direct', i.e. direct CCD photometry, you will go to the editor of 'ccdred'. You can change the translator's table to your own if you have it (there is an example file thx.dat in the image directory). Then typing 'Cntrl D' you will finish this editor and go to the editor of task 'ccdproc', which is the main task in this package. Now, if you want to overscan and trim the BIAS images, you should set the parameters as follows:

   input=      BI*.imh

(or list of your BIAS files - you know, how to do it)

   imagetype=  zero 
   ....... 
   overscan=   yes 
   trim=       yes 
   zerocor=     no
   darkcor=     no 
   flatcor=     no
      ...... 
   interactive= yes

(or no if you are not interested in the overscan fit) quit editing by typing 'Cntrl D' and run ccdproc.
If interactive=yes, you will see the averages of lines of overscan region in the graphic window plotted against number of columns. There you can change the fitting function and order. If you are satisfied with what you see, type 'q' (in the graphic window) and go to the next image.

4.

Combination of BIAS images The simpliest way to do that is to use a task 'zerocombine'. So, set parameters of this task by typing

<cd epar=      zerocombine 
   input=      @biaslist (or B*.imh) 
   output =    ZERO 
   combine=    average  (or median, whatever you prefer) 
   reject=     minmax 
   ...... 
   scale =     none 
   ....

(see the help imcombine for rejecting algorithms) and you will get the averaged BIAS image. Display it by typing

  <cd disp BIAS 1

and see how it looks.

5.

Combination of DARK images. This procedure is about identical to the previous one, but settings are slightly different. Remember, dark frames were not processed before, so you can do that at this step: set parameters in 'ccdproc':

      zerocor=     yes 
      zero=       ZERO

and set parameters in 'darkcombine':

      input=     darks 
      output=     DARK 
      combine=  average 
      reject=    minmax 
      process=      yes 
      scale=   exposure
      .....

Now, run

 <cd darkcor

and you will get a combined image DARK.

6.

Processing of flat images with zero and dark corrections. Set the parameters of 'ccdproc'

     input =  list of flat frames  (say, @flatlist) 
     darkcor= yes 
     dark=    DARK

and run
<cd ccdproc

7.

Combining of DOME flats. At his step you may want to include subsets. Then IRAF will know how to recognize which expositions could be combined together. For this purpose you should check, are there the filter names or numbers present in the image keywords. If not, you must add them by typing
<cd hedit DOMu* field='FILTER', value='u',add+
The headers of DOMu flats will be updated, adding the keyword FILTER=u Then edit the parameters of 'flatcombine', setting

     input=   list of DOME images 
     output=  Flat 
     ccdtype= flat 
     combine= average 
     reject=  crreject 
     subsets= yes 
     scale=   mode 
     ....

As a result you will have the combined DOME flats, with the subset value added to the name of your output, i.e. Flatu, Flatv, etc. Be careful with naming: subsets want, that Flat fields were named as Flat.

8.

Combination of SKY flats. There is useful to check how different are the SKY flats. The best way to do it is to divide one image by another and imexamine the result. This is done by using the simple image arythmetic, i.e.

 <cd imarit sky01 / sky02 ts12 
 <cd display ts12 1 
 <cd imexam

If you will see serious gradients of severel percents in resulting image, you should decide, which images to use in the combining. The procedure is identical to the previous one, if you have the bright twilight flats, but usually you don't have a lot of SKY flats, so maybe you will change the parameter reject= avsigclip (consult help imcombine for the details).
If you have blank SKY flats with low signal level, better use the IRAF procedure 'imcombine', which has more parameters and set

     input=   list of SKY images 
     output=  SKYf 
     ccdtype= "" 
     combine= average 
     reject=  avsigclip 
     scale=   mode 
     weight=  mode

In the output you will have the averaged flats SKYf. There are no subsets in this task.

9.

Creation of superflat from averaged DOME and SKY flats. At this step we use the IRAF procedure 'mkskyflat'. Before running this task we should flatfield the averaged SKY flats itself. We can temporarily change the imagetype of SKY from 'flat' to 'object' and run 'ccdproc' with

     input= SKY* 
     ... 
     flatcor=yes 
     flat= Flat 
     ...

(If you want to check what goes, make a copy of SKY before processing). Now set the parameters of 'mkskyflat'

     input = SKY 
     output= Sflat 
     .....

This task creates an 'improved' SKY flat by extensively smoothing it and correcting for large scale gradients. Divide the resulting image by the uncorrected one to see for improvement. By the way, you can do it in another way, by using the procedure 'mkskycor'. Consult help for this task. Change the parameters of 'ccdproc':

     input= list of scientific images 
     .... 
     zerocor=  yes 
     darkcor=  yes 
     flatcor=  yes 
     ... 
     zero=    Zero 
     dark=    Dark
     flat=    Sflat

and run 'ccdproc' once again. You are done. Congratulations!