1. Introduction
2. Questions to Ask before You Start
3. Starting IA
4. Data Formats
5. Reading Your Data
6. Displaying Your Data
7. Creating and AAR from an SPD
8. Further Processing of an AAR
9. Saving Your Results
10. Printing A Spectrum
11. Other Processing Steps

---

Introduction

This manual is for researchers who are not familiar with the Interactive Analysis (IA) software packages but who are familiar with the objectives of their SWS observations. This manual gives a brief introduction to using IA and provides links for more information on each command. After reading this manual you should be able to:

1. Start IA.
2. List the different Data Formats.
3. Read your data.
4. Display your data in SPD and AAR formats.
5. Process, by hand, an SPD to an AAR.
6. Save your results on disk.
7. Make a hard copy of your spectrum.
8. List useful IA routines.
9. Know who to contact for further questions.

There are quite a few very useful Online resources concerning SWS spectra. Many of these resources may be found on the ISO SWS homepage. The SWS Instrument Data Users Manual (IDUM) is one of the very important resources that is helpful in addressing specific issues about the data.

Back to Table of Contents

---

Questions to Ask before Your Start

1. Who do I talk to if I have any problems with the account or processing?
2. Which machine should I use to run IA?
3. What is the path to my DATA directory?
4. Is there a SCRATCH disk? If so, what is the correct path?
5. How do I print Postscript files? And where is the printer located?

Back to Table of Contents

---

Starting IA

In your working directory type ia3. This command will start an IDL (Interactive Data Language) session with all the keywords and paths set correctly for the Interactive Analysis. You will see the following prompt:
IA3 TEST>
If you have any troubles at this stage contact the system manager.

The prompt indicates that you've started the IA session. A quick word on the IA session. IA runs within the Interactive Data Language (IDL) environment. Within IDL, data which is found on disk in FITS format must be read into memory in order to be manipulated in IDL. You read the data into memory using the routines listed below. You will specify a variable name such as spd, spd0, new_spd, aar, aar10, my_aar, etc. that means something to you. IA does not care what the variable name is as long as the data tagged by the name has the correct format. Believe me, you will hear from IDL if the format is incorrect.

One final note. At times, all your variable may seem to disappear! Don't worry, they haven't. For some reason or another the routine that was running has crashed and left you within it. To get back to your data and the main level of IA3 just type retall.

Back to Table of Contents

Data Formats

There are three different data formats provided for you and one that you create:

• ERD: Edited Raw Data. Time oriented raw detector readouts.
• SPD: Standard Processed Data. Time oriented uncalibrated data. In the current form of the pipeline, this is the most useful data structure to begin with.
• AAR: Automated Analysis Results. Wavelength oriented calibrated data.
• Final Spectrum: Wavelength oriented, cleaned and possibly rebinned data. This is usually in the form of an AAR.

See the SWS IDUM for further descriptions of the data formats.

Back to Table of Contents

---

Reading Your Data

Below are listed two IA commands to read data on disk into IDL structures (memory).

spd = read_fspd(<spd-filename>)

aar = read_faar(<aar-filename>)

spd, spd-filename, aar, and aar-filename, are variables that you give. You have the choice of any name to give the spd and aar variables. The spd-filename and aar-filename are the names, in quotes, of your SPDfile and AAR file respectively including the PATH to the file.

Back to Table of Contents

---

Displaying Your DATA

In IA you can display the SPD or the AAR. The SPD gives you a time stamp of what occurred during the observation.

• Displaying an SPD
• plotspd, spd, /status, /flag, band=band : This routine will create a window in IDL and display the SPD. I find the /status and /flag keywords to be very useful in following the time sequence of an SPD. The band keyword allows you to specify which SWS band (1,2,3,4, 5 or 6) you display. Note that none of the keywords are necessary for the routine to work. Therefore the simplest command is: plotspd,spd. Below is an example of plotspd (plotspd,spd,band=1,/status,/flag). This example plots all the detectors of band 1 separately and overlays the /status and /flag keywords. This is an excellent way to review the progress of the observation.
• plotspd example
• Displaying an AAR
• plotaar, aar, xr=[low,high], yr=[low,high], band=band : This routine will create a window in IDL and display the AAR. Often, I will want to concentrate on a specific wavelength range, xr, or limit the flux range of the plot, yr. Again, the band keyword allows you to specify which SWS band (1,2,3,4, 5 or 6) you display.
The example below displays an AAR using the command: plotaar,aar,yr=[-100,100.],/flag.
• plotaar example

Back to Table of Contents

Creating an AAR from an SPD

We now come to what is a bit of an art. I call this section the SWS Soup. First of all, you can create an aar from an spd with all the defaults that are used in the pipeline with the following command:

aar=daar(spd)

It is also possible to create an AAR following a series of steps, this is the SWS Soup. The defaults of all these routines equates to the daar command.

1. spd=antimem(spd): This routine applies a first cut correction to memory effects by marking the first slope in a dark current as NODATA.
2. spd=dark(spd): This routine subtracts the dark current from the scan. The dark current is estimated from the median value in the dark measure. Whenever possible the dark current is linearly interpolated between two dark measures. In many cases, the slope of this interpolation is zero but when memory effects are present in the data the slope is non-zero. This is then another correction for the memory effects.
3. spd=respcal(spd): This routine applies the responsivity calibration to the spd.
4. spd=fluxcon(spd): This routine converts the slope values of the spd to flux values (in Jy).
5. spd=velcor(spd): This routine corrects for the relative velocity of the satellite and object.
6. aar=extract_aar(spd): This routine rearranges the spd into wavelength and AOT band or line order.

---

Further Processing of an AAR

Although this is a prefectly fine AAR, there is still more processing that can occur to clean up the spectra. In the following, I give a brief description of the commands and keywords and some idea of how I use them in my work.

1. aar=sws_flatfield(aar)

The flatfielding routine applies and offset or scaling factor to the flux values per detector. It is a method of placing all detectors on the same calibration. This routine is a recognition of the imperfections in the calibration since upto this point all detectors are treated completely separately. Note: this routine will also work on an spd.

This routine comes with a number of useful keywords:

• /offset or /scaling: You may choose whether the correction to the flux values is a constant addition (offset) or a multiplication (scaling). At the aar stage, I assume that the correction is a constant scaling factor. When operating on an spd, sws_flatfield can be used to bring all the detectors to the same zero point offset.
• band=band: this alows you to flatfield only the aot bands specified.
2. aar=sig_clip(aar)

This routine removes points that are outside a level decided by the standard deviation within a calculated data bin. The bin size is either taken as a default value or input by you through a keyword.

3. aar=sws_rebin(aar)

---

Saving Your Results

There are a number of different methods for saving your results.

• Writing a FITS Table

The most straightforward and easily understood method is to write fits tables of you data structures. This is the inverse of the read_fspd and read_faar commands: write_fitstable. The use is quite simple:

write_fitstable,aar,aar-filename

write_fitstable,spd,spd-filename

Note that there is only ONE command to write any IA structure in the proper fits table format.

• Writing an IDL Save file

If you want to save specific variables, and you know you will remember them in the future, you may use the IDL command save. A complete description of this command can be found in the IDL Reference Guide.

Printing A Spectrum

Below, I've listed a series of commands that will write a Post-Script file and send the file to a printer. Since there are many different computer systems, this 'recipe' is not at all guaranteed to work. In IDL, this recipe will produce a Post Script file, but it may be the wrong syntax to print the file. Contact your system manager if there is any problem.

set_plot,'ps'

device,file=<ps-filename>,/landscape

plotaar,aar,<parameters> or plotspd,spd,<parameters>

device,/close

$lpr -P<printer> ps-filename

Back to Table of Contents

---

---

Last Update: 29 December 1997 by Russell F. Shipman russ@sron.rug.nl

Revised: 29 January 1998 by Alberto Noriega-Crespo alberto@ipac.caltech.edu