Flux-averaging analysis of type Ia supernova data

Because of flux conservation, flux-averaging justifies the use of the distance-redshift relation for a smooth universe in the analysis of Type Ia supernova (SN Ia) data. We have combined the SN Ia data from the High-z SN Search and the Supernova Cosmology Project and binned the combined data by flux-averaging in redshift intervals of Delta z = 0.05 and Delta z = 0.1. We find that the unbinned data yield a Hubble constant of H-0 = 65 +/- 1 km s(-1) Mpc(-1) (statistical error only), a matter density fraction of Omega(m) = 0.7 +/- 0.4, and a vacuum energy density fraction of Omega(Lambda) = 1.2 +/- 0.5. The binned data for Delta z = 0.1 yield H-0 = 65 +/- 1 km s(-1) Mpc(-1) (statistical error only), Omega(m) = 0.3 +/- 0.6, and Omega(Lambda) = 0.7 +/- 0.7. Our results are not sensitive to the redshift bin size. Flux-averaging leads to less biased estimates of the cosmological parameters by reducing the bias due to systematic effects such as weak lensing. Comparison of the data of 18 SNe Ia published by both groups yields a mean SN Ia peak absolute magnitude of M-B = -19.33 +/- 0.25. The internal dispersion of each data set is about 0.20 mag in the calibrated SN Ia peak absolute magnitudes. The difference in analysis techniques introduces an additional uncertainty of about 0.15 mag. If the SNe Ia peak absolute luminosity changes with redshift due to evolution, our ability to measure the cosmological parameters from SN Ia data will be significantly diminished. Assuming power-law evolution in the peak absolute luminosity, (1 + z)(beta), we find a strong degeneracy between the evolution power-law index beta and the matter density fraction Omega(m). For Omega(m) = 0.3, we find that the unbinned data yield H-0 = 65 +/- 1 km s(-1) Mpc(-1) (statistical error only), Omega(Lambda) = 1.4 +/- 1.1, and beta = 0.5 +/- 1.6, and the binned data (with Delta z = 0.1) yield H-0 = 65 +/- 1 km s(-1) Mpc(-1) (statistical error only), Omega(Lambda) = 0.6 +/- 1.4, and beta = 0.0 +/- 1.0.