[Top] [Prev] [Next] [Bottom]

Data Reduction

Introduction

A short description on how to process the MOS/ARGUS data will be described in the next pages. The intent of this chapter is only to provide a general information on the tasks which will allow a correct reduction of ARGUS data. A manual dedicated to the reduction package is under construction.

The reduction procedures have been developped at Meudon, first with the EVE software (a clone of MIDAS), then as a package of IRAS scripts. The different steps for data reduction with EVE are shortly described in the thesis of Marie-Christine Angonin (1993, Paris).

The IRAF package may be sent to potential users. Please, contact C. Vanderriest for further information. Be aware, however, that IRAF is a very unstable software (a probable consequence of its hypertrophy) and that a script working under version xxx.yyy.zzz.1 usually does not work anymore under version xxx.yyy.zzz.2... The package which is presently available has been tested and found reliable for the IRAF version V2.10.4 (May 1995).

We list now the different steps of the data reduction process and the corresponding IRAF procedures (thereafter identified as 'IRAF') or ARGUS procedures written as IRAF scripts (thereafter identified as 'ARGUS').

Five well defined steps will be presented and bring (hopefully) the observer to a complete reduction. Some of the steps presented here can be handled in a different way using another approach. The user is supposed to have a minimal knowledge of data reduction using IRAF.

The first four steps could be consider as a "pre-processing" which treats the data as it would do for long-slit data and leads to a fully calibrated bidimensional file,
The last one i the real "processing" which deals with data from individual fibres and reconstructs various images (flux, velocity, FWHM...) of the object. The idea is to treat the data as a single bidimensional file whenever possible and to reduce as much as possible the processing of hundreds of 1-D files.

What do you start with ?

You need, for a good reduction, to start with the following frames :

A set of biaises
A flat-field image
A spectrophotometric standard star
A calibration lamp(s) spectrum
An image of your object

Step 1 : Cleaning...

Combining your biaises

You use imcombine in order to get a good bias image of your detector from the sequence of biaises you made. This step could be avoided by using the combining possibility of the exposure in the MOS environment.

Removing the biais.

You use imarith on flat, lamps, cal_star and object frames for removing the biais.

Removing the cosmic rays and other artefacts

You have a wide choice of programmes for this task. IRAF provides for example cosmicrays or imcombine. ARGUS proposes spectrenet. At least object should be cleaned that way.

Step 2 : Calibration in wavelength

Identifying the lines and preparing the transformation

You use the three routines identify, reidentify, fitcoords on your cal_star frame. this is a classical procedure for spectroscopic processing.

Transforming the images...

The IRAF routine transform has now to be applied to the flat, lamps, cal_star and object frames

Step 3 : Correcting for the field distorsion (X calibration)

Identification of the dead fibers

The ARGUS procedure identfmn will use flat for finding the dead fibers. You end with a files analogue to the one used by the spectral line identifications...

X transformation computation

Using the ARGUS procedure calibrx (which in fact chaines the IRAF procedures identify, reidentify, fitcoords) on the flat frame will allow you to obtain the parameters of the X transformation needed for the field distorsion correction.

Applying the X transformation

Tha ARGUS procedure transformx will then be applied on the flat, lamps, cal_star and object frames

Step 4 : Calibration in flux

Correcting for the flat field

By the IRAF routines blkavg and blkrep, one builds a mean flat which is used for computing the flat flat_mean. The frame obtained by dividing through imarith flat by flat_mean, let us call it flat_clean, is then used for correcting the cal_star and object. imarith is used for the division of cal_star and object by flat_clean.

Using the cal_star frame

The ARGUS procedure extrastar will take care of isolating the lines where the signal of the calibration star is present on the cal_star frame and to look for the sky signal around. You will end with a single line file of the calibration star profile cal_star_line.

Calibrating the object frame

The three classical IRAF procedures standard, sensfunc and calibrate are used for this last step using cal_star_line for the final reduction of the object frame.

Step 5 : Fiber separation and image reconstruction

Extracting the individual fibers

silficut is the ARGUS routine which will cut object into individual line image files. As, along all the varions steps already made (especially for the wavelength calibration), some lines could be slightly off compared to the others, the ARGUS procedure velosky allows you to insure that all the fibers have really well aligned in wavelength.

Building an image at a given wavelength

The final (ouf!...) ARGUS procedures silfiref, modraie and lisshexa will give you the possibility to build an image at a given wavelength, or on a wavelength band, from the fiber line images. You will the be able to identify emission areas, and even to extract part of the image through the ARGUS procedure refsomme.

[Top] [Prev] [Next] [Bottom]

Send comments to: vander@gin.obspm.fr

Copyright © 1997, CFHT Corporation. All rights reserved.