search

BEAST-Fitting / beast

Bayesian Extinction And Stellar Tool
http://beast.readthedocs.io
23 stars 34 forks source link

Add BOSZ stellar atmosphere models #180

Open karllark opened 6 years ago

karllark commented 6 years ago

Updated version of Kurucz models - including good wavelength coverage in the JWST range. Ideally would replace CK04. Need to pick a reasonable resolution. And decide on what to do with the extra grid parameters for carbon and alpha-enhancement abundances. Maybe these are not critical and we should just pick "standard" values (whatever these are).

https://archive.stsci.edu/prepds/bosz/

lcjohnso commented 6 years ago

I've started work on adding these data.

@karllark: Do you have an opinion on spectral resolution? A big constraint: for merged spectral grids (e.g., Tlusty + BOSZ), there needs to be a common wavelength grid. Tlusty data is at R~900 (sampled at 2x that), so propose resampling BOSZ data onto that grid. Current default was to match CK04's low 2 AA/pix (R of 100-200) and the wavelength gridding was not totally systematic either, so I consider this new plan an improvement. If that's acceptable, I will compute and add new library files for Tlusty and BOSZ libraries. Note that with higher spectral resolution and finer Z grid, the library file and spec grid creation will be more costly.

Regarding alpha abundance: would recommend fixed solar-scaled grid to start, as this matches the stellar evolution models (neither PARSEC nor MIST include alpha-enhanced at this time).

Regarding carbon abundance: perhaps this question should shift over to the AGB stellib issue (https://github.com/BEAST-Fitting/beast/issues/25). @marthaboyer should be involved in the discussion of 1) what additional stellar parameters need to be tracked (C/O ratio, C surface composition), and 2) what is the trigger for using Aringer atmospheres rather than CK04/BOSZ?

karllark commented 6 years ago

Cool! I agree - go for the Tlusty R~900. We can get higher resolution of Tlusty if needed, but I don't think that is needed at this point. Agree with your other points.