sncosmo / sncosmo

Python library for supernova cosmology

Geek Repo:Geek Repo

Github PK Tool:Github PK Tool

Add new SALT2 models

benjaminrose opened this issue · comments

There are several new SALT2 models. Taylor+21 (JLA+Supercal). Pantheon+ then updated the calibration to Fragilistic (Brout+22). Finally, there is the soon-to-be published retraining of SALT2 with the K21 data set (same as used in SALT3) and using the Fragilistic cross calibration (Taylor+ 2023). This last model is already on sncosmo.github.io (sncosmo/sncosmo.github.io#30).

We should be able to add these three to this section of code.

sncosmo/sncosmo/builtins.py

Lines 694 to 706 in a18a8e5

# SALT2 models
website = 'http://supernovae.in2p3.fr/salt/doku.php?id=salt_templates'
g10ref = ('G10', 'Guy et al. 2010 '
'<http://adsabs.harvard.edu/abs/2010A%26A...523A...7G>')
b14ref = ('B14b', 'Betoule et al. 2014 '
'<http://adsabs.harvard.edu/abs/2014A%26A...568A..22B>')
for topdir, ver, ref in [('salt2-2-0', '2.0', g10ref),
('salt2-4', '2.4', b14ref)]:
meta = {'type': 'SN Ia', 'subclass': '`~sncosmo.SALT2Source`',
'url': website, 'reference': ref}
_SOURCES.register_loader('salt2', load_salt2model,
args=('models/salt2/'+topdir,),
version=ver, meta=meta)

Since they use the same training method, I think they can be simply a new version of SALT2. We should change the default to the most recent (Taylor's K21-fragillistic). For reproducibility, we should have all of them remain available.

Pantheon+ used https://github.com/PantheonPlusSH0ES/DataRelease/tree/main/Pantheon+_Data/3_SALT2/SALT2_B21trained_withsys/SALT2.MODEL000, but this has a few bugs. A point update, fixing these bugs, also exists (somewhere).

In total, I would like to add:

  • the JLA+Supercal presented in T21
  • the Pantheon+ model
  • the Pantheon+ patched model
  • the K21+Fragalisitc model in the submitted T23 paper