Exploring the cosmological consequences of JLA supernova data with improved flux-averaging technique

Abstract: In this work, we explore the cosmological consequences of the "Joint Light-curve Analysis" (JLA) supernova (SN) data by using an improved flux-averaging (FA) technique, in which only the type Ia supernovae (SNe Ia) at high redshift are flux-averaged. Adopting the criterion of figure of Merit (FoM) and considering six dark energy (DE) parameterizations, we search the best FA recipe that gives the tightest DE constraints in the $(z_{cut}, \Delta z)$ plane, where $z_{cut}$ and $\Delta z$ are redshift cut-off and redshift interval of FA, respectively. Then, based on the best FA recipe obtained, we discuss the impacts of varying $z_{cut}$ and varying $\Delta z$, revisit the evolution of SN color luminosity parameter $\beta$, and study the effects of adopting different FA recipe on parameter estimation. We find that: (1) The best FA recipe is $(z_{cut} = 0.6, \Delta z=0.06)$, which is insensitive to a specific DE parameterization. (2) Flux-averaging JLA samples at $z_{cut} \geq 0.4$ will yield tighter DE constraints than the case without using FA. (3) Using FA can significantly reduce the redshift-evolution of $\beta$. (4) The best FA recipe favors a larger fractional matter density $\Omega_{m}$. In summary, we present an alternative method of dealing with JLA data, which can reduce the systematic uncertainties of SNe Ia and give the tighter DE constraints at the same time. Our method will be useful in the use of SNe Ia data for precision cosmology.