Equivalent width for double peak spectra
anggarendra wrote on Feb 22, 2010
hello everyone,
I have a double peak Be star Spectra (shell) in H alpha region and want to determine the equivalent width value using splot. which method should I use for the best result?
direct Integration (e), single profile fitting (k), or multiple profile deblending (d)
I've tried using all of them and personally like the (d) method better.
after determining the short and long wavelength with d, I tick the mark for the violet and read peaks with gaussian (g) and absorption center with the voigt (v)
(The supervisor in my university taught me that) the result I got was the EW value for each of the peak.
How can I determine the value of EW for the whole H alpha line?
Thx,
anggarendra
I have a double peak Be star Spectra (shell) in H alpha region and want to determine the equivalent width value using splot. which method should I use for the best result?
direct Integration (e), single profile fitting (k), or multiple profile deblending (d)
I've tried using all of them and personally like the (d) method better.
after determining the short and long wavelength with d, I tick the mark for the violet and read peaks with gaussian (g) and absorption center with the voigt (v)
(The supervisor in my university taught me that) the result I got was the EW value for each of the peak.
How can I determine the value of EW for the whole H alpha line?
Thx,
anggarendra
Francisco Valdes wrote on Feb 22, 2010
I'm not sure what you mean by the whole Halpha line. If you mean the two peaks taken together than the direct integration method would be used. In dealing with double peaked data, if they are far enough apart and you expect symmetric profiles then another good method is to measure only one side (from the peak to the uncontaminated side). There are options for that. However, deblending is often the method of choice.
Frank Valdes
Frank Valdes
Julien Marocco wrote on Feb 22, 2010
Hi there,
I have a quite similar problem with equivalent width measurements in absorption lines, and as I'm totally newbie with IRAF I may need your help if you don't mind. ;)
I have to measure EW in hundreds of galaxy spectras (and 4 EW per spectra), that I corrected from doppler effects given their redshift and using the dopcor tool.
So as there are many spectras I figured I should do this using a script rather than doing it by hand.
However I encounter issues when it comes to running my script, because I use the deblending tool with a cursor file (I drew inspiration from the IRAF scripts pdf on this site, but noticed a few mistakes with cursor files and tried to correct them), and the cursor position is fixed, but absorption lines may be slightly at higher or lower wavelength depending on the accuracy of my redshift list.
My problem is that in some spectras there are emission lines very close to the absorption line that I wish to measure, and when my script is supposed to fit (gaussian fit) them, sometimes it fits the closer emission line, or sometimes it doesn't fit anything, and it returns INDEF values in my .log file or EW as large as hundreds of angströms, or negative EW when emission lines are measured...
So I was wondering if there is a way, using the deblending tool in a script, to constraint only closer absorption lines EW measurements, and not to use a fixed value for the gaussian fit (because of the possible precision errors on the redshift that may imply inaccurate positions for my absorption lines) ?
if it may help here are excerpts my script files :
Cursor file
1360 1.2E-18 1 d
1380 1.2E-18 1 d
1370 1.2E-18 1 g
0 0 1 q
0 0 1 f
0 0 1 f
0 0 1 y
0 0 1 q
0 0 1 q
Cl file
lis="atmcor1370.lis"
while (fscan (list,s1) != EOF) {
cl.ukey="ukey.dat"
splot (s1, save_fi="atmeqw1370.log",cursor="eqw1370.cur")
cl.ukey=""
}
unlearn splot
- my .lis file in the .cl file is the list of .fits images for which I need to measure specific EW (I have 4 cl files; 4 .lis files and 4 .cur files because there are spectras in which I can't measure certain lines because of their redshift)
- concerning the cursor file, I tried to adapt what I read in the pdf about scripts, I'm not sure of using the right options of the deblend tool for my fit. Actually I tried several cases, when it asks for fit positions and gaussian widths I didn't find documentation on what the options do, that's to say "fixed, single or all". My supervisor first told me to do all and all, but it returned too many INDEF values and weird values, so I tried various combinations, single-fixed, fixed-fixed, etc... and the fixed-fixed one was the one which worked best regarding the amount of INDEF values, but I'm not sure if in term of accuracy it is the best.
- I haven't normalized my spectras (as indicated in the IRAF scripts pdf) but I don't think this is the cause of my problem, my supervisor told me this wasn't necesary, are we wrong ?
anyway thank you in advance for taking the time to read my post, and hope you can help me ! :)
Julien
PS : and don't hesitate to tell me if I wasn't clear, i'm not sure of my english :lol:
I have a quite similar problem with equivalent width measurements in absorption lines, and as I'm totally newbie with IRAF I may need your help if you don't mind. ;)
I have to measure EW in hundreds of galaxy spectras (and 4 EW per spectra), that I corrected from doppler effects given their redshift and using the dopcor tool.
So as there are many spectras I figured I should do this using a script rather than doing it by hand.
However I encounter issues when it comes to running my script, because I use the deblending tool with a cursor file (I drew inspiration from the IRAF scripts pdf on this site, but noticed a few mistakes with cursor files and tried to correct them), and the cursor position is fixed, but absorption lines may be slightly at higher or lower wavelength depending on the accuracy of my redshift list.
My problem is that in some spectras there are emission lines very close to the absorption line that I wish to measure, and when my script is supposed to fit (gaussian fit) them, sometimes it fits the closer emission line, or sometimes it doesn't fit anything, and it returns INDEF values in my .log file or EW as large as hundreds of angströms, or negative EW when emission lines are measured...
So I was wondering if there is a way, using the deblending tool in a script, to constraint only closer absorption lines EW measurements, and not to use a fixed value for the gaussian fit (because of the possible precision errors on the redshift that may imply inaccurate positions for my absorption lines) ?
if it may help here are excerpts my script files :
Cursor file
1360 1.2E-18 1 d
1380 1.2E-18 1 d
1370 1.2E-18 1 g
0 0 1 q
0 0 1 f
0 0 1 f
0 0 1 y
0 0 1 q
0 0 1 q
Cl file
lis="atmcor1370.lis"
while (fscan (list,s1) != EOF) {
cl.ukey="ukey.dat"
splot (s1, save_fi="atmeqw1370.log",cursor="eqw1370.cur")
cl.ukey=""
}
unlearn splot
- my .lis file in the .cl file is the list of .fits images for which I need to measure specific EW (I have 4 cl files; 4 .lis files and 4 .cur files because there are spectras in which I can't measure certain lines because of their redshift)
- concerning the cursor file, I tried to adapt what I read in the pdf about scripts, I'm not sure of using the right options of the deblend tool for my fit. Actually I tried several cases, when it asks for fit positions and gaussian widths I didn't find documentation on what the options do, that's to say "fixed, single or all". My supervisor first told me to do all and all, but it returned too many INDEF values and weird values, so I tried various combinations, single-fixed, fixed-fixed, etc... and the fixed-fixed one was the one which worked best regarding the amount of INDEF values, but I'm not sure if in term of accuracy it is the best.
- I haven't normalized my spectras (as indicated in the IRAF scripts pdf) but I don't think this is the cause of my problem, my supervisor told me this wasn't necesary, are we wrong ?
anyway thank you in advance for taking the time to read my post, and hope you can help me ! :)
Julien
PS : and don't hesitate to tell me if I wasn't clear, i'm not sure of my english :lol:
Francisco Valdes wrote on Feb 22, 2010
If you haven't found it already you might want to check the task BPLOT. This is basically what you are doing, calling SPLOT in a batch way. BPLOt is a script and you could look at the code for hints. However, if you learn how to do the various redirections of input and output then using SPLOT directly as you are doing is fine.
The problem with using cursor files in an automated way is, obviously, if the data require special care to account for small wavelength scale differences then it may not work correctly. I am not sure what is the specific question. You are asking about constraints on the solution? I think if you've looked at the help page and done some fits by hand you will probably understand the options for constraints. But if not try and ask with a specific example of what you would like the fitting to be able to do and we can try and help.
Frank Valdes
The problem with using cursor files in an automated way is, obviously, if the data require special care to account for small wavelength scale differences then it may not work correctly. I am not sure what is the specific question. You are asking about constraints on the solution? I think if you've looked at the help page and done some fits by hand you will probably understand the options for constraints. But if not try and ask with a specific example of what you would like the fitting to be able to do and we can try and help.
Frank Valdes
Julien Marocco wrote on Feb 22, 2010
Sorry I may have been a little confusing : basically what I'd like to do is to fit and get the Equivalent Widths of my spectras.
But my problem is that fitting thanks to Splot requires a specific wavelength in the cursor file, what I would like to do is to find a way to fit not on a specific wavelength, but for example on a few angstroms around a specific wavelength, and for a specific kind of line (in my case I don't want to fit emission lines, only absorption).
But I don't know if it is possible with IRAF scripts. What I would like to do is to be able, in my script, to tell IRAF not to measure emission lines when it encounters one.
For example, if I take the CIV absorption line at 1549 A, sometimes it is slightly shifted (because of the uncertainties of my redshift corrections), then my script fits the closest line to 1549 A and sometimes it is an emission line very close to it, sometimes it just doesn't fit anything and returns INDEF values...
In brief I'd like to know if there is a way, using splot, to select an area and to set constraints on the type of line so that it measures the closest absorption line around the position I defined.
But I'll take a look at the Bplot task tomorrow at the lab, maybe it is able do what I'm looking for, anyway thanks for your reply :wink:
But my problem is that fitting thanks to Splot requires a specific wavelength in the cursor file, what I would like to do is to find a way to fit not on a specific wavelength, but for example on a few angstroms around a specific wavelength, and for a specific kind of line (in my case I don't want to fit emission lines, only absorption).
But I don't know if it is possible with IRAF scripts. What I would like to do is to be able, in my script, to tell IRAF not to measure emission lines when it encounters one.
For example, if I take the CIV absorption line at 1549 A, sometimes it is slightly shifted (because of the uncertainties of my redshift corrections), then my script fits the closest line to 1549 A and sometimes it is an emission line very close to it, sometimes it just doesn't fit anything and returns INDEF values...
In brief I'd like to know if there is a way, using splot, to select an area and to set constraints on the type of line so that it measures the closest absorption line around the position I defined.
But I'll take a look at the Bplot task tomorrow at the lab, maybe it is able do what I'm looking for, anyway thanks for your reply :wink:
Last post on Feb 22, 2010