spectrum fron an echelle spectrograph

Guest wrote on Apr 14, 2006

To whom it may concern,

I am faced for the first time with a spectrum taken with an echelle spectrograph. As far as I can understand, each line of the CCD image contains a small portion of the whole spectrum in say roughly 50-100 Angstrom chunks. What I'd like to be able to do is to take the individual pieces, line them all up to produce the entire spectrum. How would I go about doing that??

Thanks,

Alex

---

Mike Fitzpatrick wrote on Apr 14, 2006

Alex,

See the "User's Guide to Reducing Echelle Spectra" (http://iraf.net/irafdocs/ech/) for the initial reduction and Example 1 of the SCOMBINE help page for how to combine the orders into a single spectrum. Followup if you still have questions.

-Mike

---

Guest wrote on Apr 14, 2006

Hi Mike,

Ok, so I checked out the example in the SCOMBINE help page... and it worked... almost. I'll try to explain... using SPLOT to view the combined image, I get the entire spectrum as I wanted... except, it seems almost like alternate "chunks" are shifted up higher with respect to the rest of the spectrum. I am compeltely unfamiliar with echelle spectra, I usually deal with longslit spectra, so perhaps there's a nuance here that I'm missing. If it helps, I posted my combine images here:

http://www.astro.umontreal.ca/~gianninas/iraf/

If you should like the originals, I can post those as well.

Thanks for your help so far, and I hope you can help me just a bit more!

Alex

---

Mike Fitzpatrick wrote on Apr 14, 2006

Hi Alex,

It looks to me like the image was combined along the wrong axis, i.e. the plot looks like a cut across the orders rather than a combination along the orders. The dispersion axis is controlled by either a DISPAXIS header keyword, the 'dispaxis' package param (e.g. "cl> epar onedspec" to reset it) or in some cases with an interactive command. If you could post the url to the original image I might be able to be more specific but try resetting dispaxis and combining again. Also note that SCOMBINE alone won't extract the orders for combination, that is covered in the full reduction in the User's Guide I mentioned.

-Mike

---

Guest wrote on Apr 14, 2006

Hello again,

OK, first off, I added the original images to the same site as above. Second, I tried changing the DISPAXIS but I got the same result. Finally, with regards to your last comment: here's the thing, these are not my spectra. They are the spectra of another researcher who made them available to the public... and since this object is now of interest to me, I wanted to have a look at this data. So as far as what has been done to the raw images to get them to the form which I downloaded... well the article where these are published doesn't go into too much detail.

All that being said, the badly combined images I produced do actually show the proper spectral features, it's just these misplaced "chunks" that mess it up. Correct me if I'm wrong, but that wouldn't be the case, seeing the proper spectral lines, if I was cutting across the orders would it? (... my echelle ignorance is a factor here :? )

Well, hope that can help you to help me... any other information you need I would be glad to provide.

Thanks,
Alex

---

Mike Fitzpatrick wrote on Apr 14, 2006

Now it makes sense, I was confused because the original plot show column number. In any case, you apparently took the example literally and set the 'combine' param to be "sum". What you're seeing is simply the sum of the overlapped pixels in each order, try something like 'combine-average' and you'll see a more reasonable looking spectrum (still noisy but no obvious jumps).

Cheers,
-Mike

---

Guest wrote on Apr 14, 2006

Hi Mike,

It worked! Thanks a million. :D

If you gimme your last name, I'll make sure to acknowledge your contribution when we publish our article. ;)

Regards,
Alex

---

Last post on Apr 14, 2006