Cosmology with Photometric Surveys of Type Ia Supernovae

TL;DR

This paper evaluates the effectiveness of using solely photometric data from large surveys like LSST to identify Type Ia supernovae, determine their redshifts, and reduce contamination from other supernova types for cosmological research.

Contribution

It introduces a Markov Chain Monte Carlo method for fitting supernova parameters from photometric data and a Bayesian statistical approach to reduce contamination from non-Ia supernovae.

Findings

Photometric redshift estimation is accurate for z < 0.5.

Photometric data at z < 0.2 is nearly as precise as spectroscopic measurements.

The Bayesian method effectively reduces contamination from SNIb/c supernovae.

Abstract

We discuss the extent to which photometric measurements alone can be used to identify Type Ia supernovae (SNIa) and to determine redshift and other parameters of interest for cosmological studies. We fit the light curve data of the type expected from a survey such as the one planned with Large Synoptic Survey Telescope (LSST) and also to remove the contamination from the core-collapse supernovae to SNIa samples. We generate 1000 SNIa mock flux data for each of the LSST filters based on existing design parameters, then use a Markov Chain Monte-Carlo (MCMC) analysis to fit for the redshift, apparent magnitude, stretch factor and the phase of the SNIa. We find that the model fitting works adequately well when the true SNe redshift is below 0.5, while at $z < 0.2$ the accuracy of the photometric data is almost comparable with spectroscopic measurements of the same sample. We discuss the contamination of Type Ib/c (SNIb/c) and Type II supernova (SNII) on the SNIa data set. We find it is easy to distinguish the SNII through the large $\chi^2$ mismatch when fitting to photometric data with Ia light curves. This is not the case for SNIb/c. We implement a statistical method based on the Bayesian estimation in order to statistically reduce the contamination from SNIb/c for cosmological parameter measurements from the whole SNe sample. The proposed statistical method also evaluate the fraction of the SNIa in the total SNe data set, which provides a valuable guide to establish the degree of contamination.