Initial Evaluation of SNEMO2 and SNEMO7 Standardization Derived From Current Light Curves of Type Ia Supernovae

TL;DR

This study evaluates the SNEMO models for Type Ia supernova standardization, comparing their performance with SALT2 using various data sets, and finds that data quality significantly impacts the constraining power of the models.

Contribution

It introduces the use of SNEMO models for supernova standardization and assesses their effectiveness relative to SALT2 across different photometric data sets.

Findings

SNEMO2 performs comparably to SALT2 in parameter constraints.

SNEMO7 reduces the Hubble diagram RMS scatter from 0.148 to 0.141 mag.

Data quality and model complexity influence the constraining power of SNEMO models.

Abstract

To determine if the SuperNova Empirical Model (SNEMO) can improve Type Ia supernova (SN Ia) standardization of several currently available photometric data sets, we perform an initial test, comparing results with the much-used SALT2 approach. We fit the SNEMO light-curve parameters and pass them to the Bayesian hierarchical model UNITY1.2 to estimate the Tripp-like standardization coefficients, including a host mass term as a proxy for redshift dependent astrophysical systematics. We find that, among the existing large data sets, only the Carnegie Supernova Project data set consistently provides the signal-to-noise and time sampling necessary to constrain the additional five parameters that SNEMO7 incorporates beyond SALT2. This is an important consideration for future SN Ia surveys like LSST and WFIRST. Although the SNEMO7 parameters are poorly constrained by most of the other available data sets of light curves, we find that the SNEMO2 parameters are just as well-constrained as the SALT2 parameters. In addition, SNEMO2 and SALT2 have comparable unexplained intrinsic scatter when fitting the same data. When looking at the total scatter, SNEMO7 reduces the Hubble-Lemaitre diagram RMS from 0.148~mag to 0.141~mag. It is not then, the SNEMO methodology, but the interplay of data quality and the increased number of degrees of freedom that is behind these reduced constraints. With this in mind, we recommend further investigation into the data required to use SNEMO7 and the possibility of fitting the poorer photometry data with intermediate SNEMO-like models with three to six components.