fxcor on echelle data producing one average crosscorrelation

Michael Weber wrote on Jan 16, 2007

I would like to use fxcor on echelle spectra. One way is to average the result from each individual order and average afterwards, another one is to use the combined spectrum, but I really would like to make an average cross correlation function from all appropriate orders and do the radial velocity measurements there.

Looking at the fxcor source it seems feasible, but not trivial. It goes far beyond what I have ever done (or have ever wanted to do) with iraf.

And before I start (or convince somebody to start :wink:) I would like to make sure that something like an ecfxcor task does not exist yet.

---

Mike Fitzpatrick wrote on Jan 16, 2007

When you say "make an average CCF...." do you really mean you want to combine the individual CCFs from each order? This could be possible in a script task where you could use the ":wccf <filename>" command to write out the CCF from each order (e.g. run FXCOR in batch mode using a cursor command file to step through the orders and automatically write the CCF". You'd have to use a different task to "stack" the results (the 'ccftype' parameter lets you select text or image), and you would have to do the peak fitting an RV calculations yourself in the script (the equations are in the help page and RVCORRECT can give you most of the needed quantities). I'm not convinced however that this is any better than just averaging the results from the individual orders, but what I described should be possible. Hope it helps.

Cheers,
-Mike

---

Michael Weber wrote on Jan 16, 2007

Thanks for the tip. I thought about that, but didn't want to go through the trouble of doing the velocity part myself. Calculating the ccf is actually not all that difficult from a c program (the most difficult being the WCS), but there again I would have to do the velocity correction myself.
As for usefulness, there is a good paper out: S. Zucker 2003, mnras 342, 1291.
Seems like I will have to dig into RVCORRECT a bit.

Thanks,
Michael

---

Orsola De Marco wrote on Jan 16, 2007

Hi Mike and Michael, this is only a comment to your posts.

I have been working a lot on echelle data and fxcor. Mike had written a small script for me to run fxcor aperture by aperture and then average the RV results. I used that a lot, but now I have totally convinced myself that getting RV values aperture by aperture and then averaging, is NOT the same as getting RV values using the entire spectrum (combined apertures) all together. So, of course the only thing I can do is combine the apertures into one spectrum. However, the nature of my absorption lines is such that I really do not want to attempt at getting the apertures straight (ie fitting the continuum) because I do not really know where the continuum ends and the lines' wings start. So I ended up scombining all the apertures without taking out the blaze. It looks bad but, provided that the absorption lines are not on aperture overlaps, I can use it.

It would surely be great to be able to use fxcor on a non-scombined echelle spectrum selecting multiple ranges on different multiple apertures and get one Xcorrelation value...

Best, Orsola

---

Last post on Jan 16, 2007