Fuzzy Supernova Templates II: Parameter Estimation

TL;DR

This paper extends the SOFT method to estimate redshift and luminosity distance of Type Ia supernovae from photometric data alone, enabling efficient analysis of large survey datasets for cosmological studies.

Contribution

It introduces an extension of the SOFT method for photometric redshift and distance estimation, validated with SDSS and SNLS data, achieving accuracy comparable to spectroscopic methods.

Findings

Redshift residuals RMS_z=0.051 with light curves alone.

Hubble residuals of 0.18 mags with spectroscopic priors.

Reliable redshift and distance measurements without spectroscopic data.

Abstract

Wide field surveys will soon be discovering Type Ia supernovae (SNe) at rates of several thousand per year. Spectroscopic follow-up can only scratch the surface for such enormous samples, so these extensive data sets will only be useful to the extent that they can be characterized by the survey photometry alone. In a companion paper (Rodney and Tonry, 2009) we introduced the SOFT method for analyzing SNe using direct comparison to template light curves, and demonstrated its application for photometric SN classification. In this work we extend the SOFT method to derive estimates of redshift and luminosity distance for Type Ia SNe, using light curves from the SDSS and SNLS surveys as a validation set. Redshifts determined by SOFT using light curves alone are consistent with spectroscopic redshifts, showing a root-mean-square scatter in the residuals of RMS_z=0.051. SOFT can also derive simultaneous redshift and distance estimates, yielding results that are consistent with the currently favored Lambda-CDM cosmological model. When SOFT is given spectroscopic information for SN classification and redshift priors, the RMS scatter in Hubble diagram residuals is 0.18 mags for the SDSS data and 0.28 mags for the SNLS objects. Without access to any spectroscopic information, and even without any redshift priors from host galaxy photometry, SOFT can still measure reliable redshifts and distances, with an increase in the Hubble residuals to 0.37 mags for the combined SDSS and SNLS data set. Using Monte Carlo simulations we predict that SOFT will be able to improve constraints on time-variable dark energy models by a factor of 2-3 with each new generation of large-scale SN surveys.