Use of Dopcor

rohit wrote on May 13, 2009

I have been trying to combine two spectra taken at two different epochs. But before adding them (using sarith task), I need to correct for doppler shift. From the help under dopcor, I find that I need to use rvcorrect to find vhelio:

spectra1.fits has a vhelio = -13.09 km/s
spectra2.fits has a vehelio = +7.12 km/s

As I want to combine, I consider spectra 1 as reference. Thus spectra 2's vhelio needs to be changed by -20.21 km/s in order it coincides with spectra 1, right? I use the following:

rv> dopcor spectra2.fits -vhelio isvel+
Redshift or velocity (km/s): -20.21

This gives me a file called "-vhelio.fits" !!! However when I plot spectra 1 and -vhelio.fits, they do not coincide.

What is going wrong here?

Thanks!

---

Francisco Valdes wrote on May 13, 2009

You may have already figured out the output part, you were missing an argument which is the output name; i.e.

dopcor input output velocity

As to getting the shift to be what you want the first thing to check is the sign of the velocity which is always confusing to me. Other than that I'm not sure what to suggest. You could send me the two spectra if you want.

Yours,
Frank Valdes

---

rohit wrote on May 13, 2009

Hello Frank/Fitz

I am using observations in the J-band. During observations, we believe that the wavelength calibration lamp was incorrectly set so the telluric lines in the observed A-type star is slightly shifted than those in the target star. Thus a division of target star by the A-type star gives an awfully noisy spectrum. I therefore believe in shifting the telluric lines in the observed A-type star in order for them to match those in the target spectrum.

In order to obtain the shift, I used FXCOR. After looking at each order, the mean shift is ~ 5 km/s. Using this velocity I used DOPCOR to shift the A-type star. This process is not working with my spectra. I am not sure why this is so. I could send you the two spectra if this helps you. Do let me know.

Thanks a lot!

---

Francisco Valdes wrote on May 13, 2009

Note that because I considered it likely that such shifts between the calibration and target spectra might occur the TELLURIC task has a feature to optimize the matching of telluric features in both wavelength and flux. This should be exactly what you want. DOPCOR/FXCOR should not be necessary for small shifts where the difference between a zero point shift and a doppler shift is insignificant.

Frank

---

rohit wrote on May 13, 2009

Thanks Frank! Once I ran it, the only expression I had was "Wow". You can't believe how much and how long I have been struggling with these files!

Thanks a lot!

Rohit

---

Last post on May 13, 2009