Relativistic gravity in the inhomogeneous Universe

TL;DR

This thesis investigates how inhomogeneities and anisotropies in the Universe affect cosmological observations and tests of gravity, extending formalism and analyzing light propagation to challenge standard assumptions.

Contribution

It develops a new framework for studying anisotropy and inhomogeneity effects on cosmological scales, including averaging methods and constraints on post-Newtonian parameters.

Findings

A theory-independent approach to cosmic expansion and perturbations.

Constraints on post-Newtonian parameters from cosmic microwave background data.

Averaged models accurately reproduce the Hubble diagram in inhomogeneous universes.

Abstract

Cosmology is built on a relativistic understanding of gravity, where the geometry of the Universe is dynamically determined by matter and energy. In the cosmological concordance model, gravity is described by General Relativity, and it is assumed that on large scales the Universe is homogeneous and isotropic. These fundamental principles should be tested. In this thesis, we explore the implications of breaking them. In order to understand possible modifications to gravity on cosmological scales, we extend the formalism of parameterised post-Newtonian cosmology, an approach for building cosmological tests of gravity that are consistent with tests on astrophysical scales. We demonstrate how this approach can be used to construct theory-independent equations for the cosmic expansion and its first-order perturbations. Then, we apply the framework to observations of the anisotropies in the cosmic microwave background. We use these to place novel cosmological constraints on the evolution of the post-Newtonian parameters. We investigate the consequences of inhomogeneity and isotropy by developing a new approach to studying anisotropy in the Universe, wherein we consider how an anisotropic cosmology might emerge on large scales as a result of averaging over inhomogeneous structures, and demonstrate how the emergent model is affected by backreaction. We perform a detailed study of light propagation in a wide class of inhomogeneous and anisotropic spacetimes, exploring the conditions under which the Hubble diagram can be accurately predicted by an anisotropic model constructed using explicit averaging, even in the presence of large inhomogeneities. We show that observables calculated in a suitable averaged description closely reproduce the true Hubble diagram on large scales, as long as the spacetime possesses a well-defined homogeneity scale.