The Impact of Peculiar Velocities on Supernova Cosmology

Martin, Bryony

The Impact of Peculiar Velocities on Supernova Cosmology

Thesis / Dissertation

2011

Abstract

Thesis We investigate the impact of correlated peculiar velocities on the accuracy of the estimated cosmological parameters calculated using type Ia supernovae. We use mock galaxy catalogues constructed from the Millennium Simulation, the largest dark matter N-body simulation run to date. Supernovae are assigned to the galaxies in the mock survey data according to supernova rates that define the probability of a supernova occurring in a given time. The rates are based on the physical properties of the galaxies, in particular their colour, which is linked to the rate of star formation, and their mass. The peculiar velocities of each galaxy are provided and we assume that the supernovae have the same peculiar velocities as their host galaxies. Therefore, each supernova has a true cosmological redshift, and an observed redshift which is the redshift measured by an observer and is subject to contamination by radial peculiar velocities. We find that the correlations between the peculiar velocities can be reduced by combining observational data from different survey fields that have large angular separations. Using CosmoMC, the cosmological parameters are estimated, with the Markov-Chain Monte Carlo technique, from the simulated supernova data, firstly using the true cosmological redshifts, and subsequently, using the observed redshifts. We compare the estimated values of the parameters Ωm and H0 that are obtained with the two different redshifts to get a measure of the inaccuracies caused by correlated peculiar velocities. Our results provide upper bounds on the effect. The results suggest that a minimum redshift cut at z = 0.05 is the most effective at reducing the effects of the correlated peculiar velocities without compromising the accuracy of the parameter estimation. In general, the shifts in parameter estimates due to correlations in the peculiar velocities are around 1σ or less, with H0 slightly more biased than Ωm. In specific cases, where only low redshift data is used (and therefore, which are not very relevant to reality) the shift can be up to 3σ. Since H0 is determined by low redshift data, it is not surprising that it is more affected by the correlations which occur primarily at low redshift. Combining the 1σ shift with standard sample variance means that, in general, peculiar velocity correlations can cause a shift in the recovered cosmological parameters of up to 2σ. This suggests that future surveys will have to take these correlations into account to ensure unbiased parameter estimates.

Keywords

Astronomy

Reference:

Collections

Masters

Full item page