Probing bulk flow with nearby SNe Ia data

TL;DR

This study tests the isotropy of the local universe using low redshift supernova data, finding moderate evidence for anisotropy that aligns with expectations from large-scale bulk flows in the standard cosmological model.

Contribution

It introduces a non-parametric method to analyze supernova data for anisotropy and assesses the impact of data distribution biases and velocity perturbations on the results.

Findings

Rejection of isotropy hypothesis at moderate significance (p ~ 0.07)

Estimated maximum anisotropy direction at (b, l) = (20°, 276°) with uncertainties

Including velocity perturbations reduces the anisotropy significance (p ~ 0.29)

Abstract

We test the isotropy of the local Universe using low redshift supernova data from various catalogs and the non-parametric method of smoothed residuals. Using a recently developed catalog which combines supernova data from various surveys, we show that the isotropic hypothesis of a Universe with zero velocity perturbation can be rejected with moderate significance, with $p$-value $\sim 0.07$ out to redshift $z < 0.045$. We estimate the direction of maximal anisotropy on the sky for various pre-existing catalogs and show that it remains relatively unaffected by the light curve fitting procedure. However the recovered direction is biased by the underlying distribution of data points on the sky. We estimate both the uncertainty and bias in the direction by creating mock data containing a randomly oriented bulk flow and using our method to reconstruct its direction. We conclude that the inhomogeneous nature of the data introduces a directional bias in galactic latitude of approximately $|\Delta b_{\rm max}| \sim 18^{\circ}$ for the supernova catalog considered in this work, and after correcting for this effect we infer the direction of maximum anisotropy as $(b,\ell) = (20^{\circ}, 276^{\circ}) \pm (12^{\circ},29^{\circ})$ in galactic coordinates. Finally we compare the anisotropic signal in the data to mock realisations in which large scale velocity perturbations are consistently accounted for at the level of linear perturbation theory. We show that including the effect of the velocity perturbation in our mock catalogs degrades the significance of the anisotropy considerably, with $p$-value increasing to $\sim 0.29$. One can conclude from our analysis that there is a moderate deviation from isotropy in the supernova data, but the signal is consistent with a large scale bulk velocity expected within $\Lambda$CDM.