Best source of reference spectra?

[jnp]

I have noticed a difference in m0iii spectra originating from different sources.

I have been getting spectra from this link

http://www.ifa.hawai.../AJP/hilib.html

This alternative uses a multi column format but the data is the same

https://www.eso.org/.../lib/index.html

The attached chart shows a comparison of Pickles data from https://www.eso.org/.../lib/index.html) and the file downloaded from the ISIS site?

The ISIS file (top/blue) is 'smoother' (or less resolved - take your pick). Which source of reference spectra should we be using?

I don't know if m0iii is the only inconsistent spectrum? (I haven't compared all 113 spectra, but a few random selections seem to be the same).

Thanks for looking

John

[Merlin66]

There has been some interesting discussion on spectral references....

http://www.spectro-aras.com/forum/viewtopic.php?f=6&t=207

The Miles database (as used by Christian in ISIS etc) seems to be the prefered "weapon of choice"

If you just want "real life" low resolution comparison spectra, then Richard Walker's "Spectral Atlas" is a great starting point.

In the astronomical spectroscopy group files area, we have a copy of the original MKK spectral atlas for comparison.

[Ben Ritchie]

Are you after MK standards, or specific (northern sky) objects?

If it's the former, I use the UVES POP catalogue here

http://www.eso.org/sci/observing/tools/uvespop/interface.html

UVES is a high-resolution echelle spectograph on VLT Kueyen at Paranal, so it's much higher resolution than you need, but you can easily rebin to match your resolution. Just pick the type you want from the drop-down, it will show you the whole spectrum then pick the 'ascii file' radio button, the wavelength range you want (in Angstroms), hit plot again and it gives you a text file to download. Then just rebin as required using your favourite tool.

Paranal is in the southern hemisphere and can't observe many of the stars we UK observers take for granted, but unless you are taking spectra of a real oddball a star of the same spectral type from the UVES POP catalogue will have a very similar spectrum to your target.

[robin_astro]

Hi John,

I have noticed a difference in m0iii spectra originating from different sources.

In this case it looks like a resolution effect to me.

In general it depends what you are looking for when using reference spectra. For calibration purposes (eg instrument response) where the continuum is important I now use reliable flux calibrated actual spectra of the measured star eg from the Miles database which is proving to be reliable. Even the UVES spectra are not that good when it comes to accurate continuum as they have a claimed 10% uncertainty at best (and checks against the same stars in Miles confirm this) owing to the fact that they used historic instrument response and extinction corrections rather than taking actual calibration stars (The stars were recorded in "free time" at dawn and dusk) they are ok for high resolution over a limited wavelength range eg with the LHIRES though.

Generic spectra (such as Pickles) for each type can be different from actual stars for a number of reasons:

1. The published spectral classification of a star can be in doubt/dispute or just plain wrong. (Putting most stars into Skiff's database

http://vizier.u-strasbg.fr/viz-bin/VizieR?-source=B/mk

almost always brings up a range of published classifications for any given star)

2 The star may be spectroscopically variable,( It is particularly hard to find a cool stat that is not), pulsating stars for example can change their classification dramatically with phase

3. The star may be affected by interstellar extinction

4. metallicity, as well as affecting the strength of lines, also affects the continuum shape (or more accurately the effective temperature for a given spectral clasification)

5. Any other peculiarity not included above.

Cheers

Robin

[robin_astro]

For calibration purposes (eg instrument response) where the continuum is important I now use reliable flux calibrated actual spectra of the measured star eg from the Miles database which is proving to be reliable.

If using the Miles database stars for calibration, use the original non de-reddened spectra not the calculated de-reddened version. There is a very useful spreadsheet version of these spectra here from the thread that Ken mentioned. It is also best to stick to main sequence (ie luminosity class v) stars as they are less likely to be variable

http://quasar.teoth.it/html/varie/MILES_OBS.zip

Cheers

Robin

[jnp]

Thanks for the all the informative responses.

There has been some interesting discussion on spectral references....

http://www.spectro-a...c.php?f=6&t=207

That is an epic thread. Looked like it was all in French to start with. I have just a quick scan and this raises some questions. I modified the Excel macro to export spectral class + name as that is more user friendly than just the Miles number.

The Miles are higher resolution but the range is just 350 to 749. I know there is a lot of telluric absorption below 750nm, but isn't there anything of interest either?

Another question. The SpT column doesn't follow the stellar classification designation, e.g "O" or "A0"?

John