spectro-aras.com

Hi Andy, I'm glad if that helps in some way!

Yesterday I have observed nu Gem with Lhires III 150. I also observed 26 Gem and two Miles star (HD65900, HD072660) about which we have library spectra, all targets around 41 degree altitude above horizon. A very bright star like Castor should be avoided for an intensity calibration purpose because the atmospheric scintillation could have a great impact on continuum with relatively short exposures.

I "cross-checked" the response curve between the two Miles stars (checking the result using the other star response), finding a good agreement. It is the same as to say that the response curves obtained with the two stars are very similar.

I compare nu Gem with François low-res spectrum in BeSS:



Nu Gem is a complex system that undergone Be and Be-shell phases and the circumstellar disk changes with time, so I think all this may have impact on continuum. More than a year has gone between observations.

Comparison with a nu Gem profile calculated using 26 Gem reference star (response with Pickles A2V library spectrum):



Profiles should have been coincident but, because of the problem described by Robin that said:

Is the catalogue spectrum of your reference star reliable ? If it is an actual spectrum of the star eg like the MILES stars then it will be but if you are using a generic eg Pickles spectrum then it might not be identical to the actual spectrum of the star used.

they are different. The profile using 26 Gem as reference star and Pickles A2V is not ok. Note I observed the stars at the same altitude, so atmosphere extinction is not responsible for that deviation.

Finally a spectrum extended to far red using a Pickles spectrum similar to Miles HD65900 for response curve calculation. Miles library spectra ending at about 7400A while Pickles go beyond (I created a "pseudo-Miles" joining a Miles spectrum and the "tail" of a similar continuum Pickles spectrum)



I see some raising of intensities in the far red part of your nu Gem spectrum. As you can see, my profile tends always to decrease toward red wavelengths (I obtain the same behaviour with any usual Pickles library spectrum). The raising of red part of such hot star profiles seems to me quite unusual (but I have not a great experience!). This is why I told you to check for any non-zero value offset in the 2d spectrum frames used for extracting profiles or check if red part of calculated response curve "enters" into telluric absorption bands.

For spectra comparisons I suggest you to scale all low-res profiles in the range 6650-6750A (intensity equal to one). It will be easier to evaluate hotter/cooler continua (but maybe it's a matter of habit...)

Clear sky!
Paolo

PS: you can download my nu Gem profile in FITS format here:
http://quasar.teoth.it/html/spectra/_nu ... tended.fit

Hi Paolo,

Thanks for all the trouble you've gone to in investigating this. I've downloaded a copy of your fits file so I can use it in comparisons. That is some very useful information and tips you have given me.

Nu Gem is now getting too low for me to get decent data from my latitude (latitude 51 degrees with Nu Gem now a poor 25 degrees when it is fully dark), so I'll pick a higher altitude star for some more experimentation and learning.

I notice the Miles stars that you used are in quite different parts of the sky though at similar altitude. When taking the spectra of standard stars, is there any rule of thumb on how close in altitude they should be? It still takes me a little while to relocate to a new star and get it onto the slit before I start acquiring the spectrum, by which time the stars altitude will have changed a little. I'm not sure if a change of say 5 degrees in altitude would be a problem?

Thanks for all your help,

Andy

Your welcome Andy!

I'm not sure if a change of say 5 degrees in altitude would be a problem?

The tolerance depends on the target altitude. The effect of atmospheric extinction is related to air mass passed through by the light (sorry I don't know if this is proper English). Therefore, it vary (non-linearly) with the angular distance of the target from the zenith.

ISIS allows to simulate the effect by means of atmospheric transmission curve calculation. So I create some profiles comparison to give you an idea of the +/- 5 degree altitude difference impact on continuum. I considered average atmospheric conditions for the current season with the observer at sea level (no elevation). Results will differ in extreme different climates (like dry deserts or very humid weather places).

Target at 65 degree above horizon and reference star used to calculate the general response curve at 60 (-5), same altitude (correct profile) and 70 (+5) degree altitude:



Target at 35 degree above horizon:



Target at 15 degree above horizon:



As you can see, when the target is very high above horizon, 5 degree altitude "error" in the position of the reference star has a very small impact on continuum. At low altitudes the observation of reference star becomes critical. Five degrees really distort a profile when we observe at 15 degree above horizon. Further down (i.e. 5 degree altitude), less than one degree error is enough to produce a similar effect. This is why, for example, Nova Sgr 2015 #2 continuum is very difficult to calibrate, especially in early observations.

Note that the impact with +5 and -5 altitude error of reference star is asymmetric. This is due to different airmass change between the two situations (see values on graphs). This means that an altitude error above the target is less severe than the same amount below.

Here you find a spreadsheet using the Miles spectral library that helps to find the optimal reference star. It provides several useful parameters like differential airmass and altitude values, angular distances and more:

http://www.spectro-aras.com/forum/viewt ... ?f=8&t=941

Miles is not an archive designed for intensity calibration of spectra but it is very useful for this purpose, being careful to avoid peculiar stars that could change the spectrum over time. Normally I check the selected one on Simbad, preferring main sequence hot stars. If there's a choice between reference stars with similar altitude, it is preferable to choose the closest (in angular distance) to the target.

Ciao!
Paolo

Hi Andy,

Paolo has covered the limits to using a reference star at approximately the same altitude and that is the technique most commonly used but there are a couple of other techniques which can be used to correct for the atmosphere. They are:-

Measuring reference star(s) at two different altitudes so a better estimate of the contribution of the atmosphere at the altitude of the target can be made.
Modelling the effect of the atmosphere directly.

All three are described by Christian Buil here.
http://www.astrosurf.com/buil/isis/guid ... method.htm

Cheers
Robin

Hi Paolo & Robin,

Thanks very much for all your help. This has been immensely useful to me and will enable me to improve my response calibrations.

Paolo, those are great graphs from ISIS. They have given me a much better understanding of the effect of changing altitude and how to be careful in my choice of reference star.

Robin, thanks for link. It is useful to know there is a way to correct using 2 stars if I cannot obtain a spectrum of a single star at the right altitude.

Now what I need is more practice

Cheers,

Andy