Question on order merging and standard in flux calibration
christip wrote on Sep 07, 2006
Hi!
I've been trying a lot of time now to achieve a uniform merging of orders in echelle spectra. I have been using scombine to do this. The output merged spectrum has peaks or dips at the points of merging of the orders. How can I improve this?
Could the problem be caused earlier on, maybe from the flux calibration? or is it scombine I don't use right?
One more question:
When using standard for flux calibration and at the point of editing the bandpasses:
if the standard has an absorption line the bandpasses should also fit the absorption region or only the continuum leaving out any lines?
Thanks a lot
Looking forward to your reply!
I've been trying a lot of time now to achieve a uniform merging of orders in echelle spectra. I have been using scombine to do this. The output merged spectrum has peaks or dips at the points of merging of the orders. How can I improve this?
Could the problem be caused earlier on, maybe from the flux calibration? or is it scombine I don't use right?
One more question:
When using standard for flux calibration and at the point of editing the bandpasses:
if the standard has an absorption line the bandpasses should also fit the absorption region or only the continuum leaving out any lines?
Thanks a lot
Looking forward to your reply!
Francisco Valdes wrote on Sep 07, 2006
Hi,
Pasting together spectral orders is difficult and there is currently no special software in IRAF for doing this. Physically adding the orders should get all the flux. However there are other effects like the sensitivity of the detector that interfer. Another "in principle" kind of method is to model the blaze function. But there is nothing in IRAF for this.
The best way I know is to normalize each order using CONTINUUM and then average the results with SCOMBINE. Actually, if I understand, you have flux calibrated each order independently using standard stars. In this case the overlap should, again in principle, have produced the same values. In this case SCOMBINE is used. However, because you are dealing with the ends of the sensitivity function there can be enough uncertainties that this might not work to well.
Concerning the standard stars a flux calibration point is simply the flux over some wavelength region whether or not there are lines. But typically the standard stars are hot, often rapidly rotating, stars with as few lines as possible. Your observation of the standard star should have exactly the same features and use the same bandpasses so that you don't need to ignore the regions with lines. If the spectrum is very low resolution or the calibration information has very small band passes then there can have edge effects. But the simple answer to your question is that the flux is based on the observed data and not fitting to the continuum.
Yours,
Frank Valdes
Pasting together spectral orders is difficult and there is currently no special software in IRAF for doing this. Physically adding the orders should get all the flux. However there are other effects like the sensitivity of the detector that interfer. Another "in principle" kind of method is to model the blaze function. But there is nothing in IRAF for this.
The best way I know is to normalize each order using CONTINUUM and then average the results with SCOMBINE. Actually, if I understand, you have flux calibrated each order independently using standard stars. In this case the overlap should, again in principle, have produced the same values. In this case SCOMBINE is used. However, because you are dealing with the ends of the sensitivity function there can be enough uncertainties that this might not work to well.
Concerning the standard stars a flux calibration point is simply the flux over some wavelength region whether or not there are lines. But typically the standard stars are hot, often rapidly rotating, stars with as few lines as possible. Your observation of the standard star should have exactly the same features and use the same bandpasses so that you don't need to ignore the regions with lines. If the spectrum is very low resolution or the calibration information has very small band passes then there can have edge effects. But the simple answer to your question is that the flux is based on the observed data and not fitting to the continuum.
Yours,
Frank Valdes
Last post on Sep 07, 2006