ISIS, BeSS standard and observatory coordinates

[Andrew Smith]

I suggest we leave the header for input into BeSS as it is today because this will avoid discussions about whether BeSS is working in the best way.

We can then have a universal header for amateur spectra which can be recommended for spectra submitted to any general amateur database. This could be based on the BeSS output header for example but use standard keywords in place of the BSS_ words. The choice of type of header could be a selection in the data reduction software (ISIS BASS VSpec etc)

Robin

Robin, to take this forward I can think of no one better than you good self to make a specific proposal. You could then open it up for comment etc. before asking the software providers to do their magic.

Regards Andrew.

Regards Andrew

[Valerie Desnoux]

Hi

My recommendation would be to use the standard fits keyword sitelat and sitelong... and NOT beSS keywords. BSS_Lat and BSS_Long are specific to BeSS because they are related to a site descriptor in the BeSS database, that's why they are "BSS_" specific.

I could not find the SITELAT and SITELONG official description and clear specification of format, of someone can provide it.

Now, as BeSS "checker" I can tell you we reject about 2 sites coordinates over 10 even if format is explained... now we do systematic check on google map... we had several observatories location in the middle of the sea...

Valerie

[Robin Leadbeater]

Then it seems the simplest short term solution could be just to add SITELAT, SITELONG and SITEALT (or SITEELEV which seems more common) to the existing header already generated by the various software for BeSS input. As I understand it, these would be ignored by BeSS (Valerie can you confirm?) which would add the BeSS specific versions to the output data based on the registered observatory details. Again for simplicity it would be easiest to us the same data format as for the BeSS equivalents

Longer term we could move to a more universal set of keywords for databases other than BeSS but I suspect this will needed more lengthy discussions There are probably some other important parameters to be added to a universal fits header which are currently missing from the BeSS standard. For example if the spectrum is corrected for atmospheric extinction (which is information only the observer has) and the SNR, which in the absence of clear continuum regions can be estimated by the observer during reduction but not subsequently.

This is a source for dictionaries for existing keywords
http://fits.gsfc.nasa.gov/fits_dictionary.html

Cheers
Robin

[Christian Buil]

In the very latest version of ISIS (V5.5.0, just available), I have added GEO_LONG, GEO_LAT and GEO_ELEV keyword in the processed spectra header FITS file.

Christian

[Ken Harrison]

Robin,
Thanks for the fit header keywords link..

Interesting that none of the keywords listed addresses the issue of observer location.

[Robin Leadbeater]

Hi Ken,

I guess it is generally less of an issue for professional observatories which have published coordinates and don't tend to move around. Also you dont get so much mixing and sharing of data sets between professional observatories as you do with amateurs except in specific collaborations.

Mirametrics have a list of keywords their MIRA software uses and recognises which includes SITELAT, SITELONG, SITEELEV and SITESYS
http://mirametrics.com/help/mira_al_8/s ... ywords.htm

Personally I like to see spectra heliocentric corrected (and tellurics removed) by the observer prior to being made public so if I view a spectrum it actually represents the spectrum of the star, but that is not the preferred format for the BeSS database at least.

Cheers
Robin

[Ken Harrison]

AstroArtV5 has a capability of setting up a MPC Report.
This then loads the observers data into the fits header.
It uses "Longitude", and "Latitude"

[Christian Buil]

The keywords are not the only question. For example MaximDL use the keyword SITELONG but the argument is of the form:
DD:MM:SS, not in decimal form like DD.DDDDD. Another problem is the exact definition of coordinates (East, West sign, geocentric, geographic/topographic, ...). A uniform solution ("universal" FITS syntax) is impossible.

Christian