Abstract
The study of the large-scale structure of the Universe is an integral part of modern cosmology, which is entering a golden era with the arrival of new generations of ambitious cosmological surveys. This thesis examines different cross-sections of the interface where theoretical predictions are confronted with observational measurements, and proposes a number of novel methodologies that can improve the constraints of cosmological models from the large-scale clustering of galaxies tracing the underlying matter distribution:1) a hybrid-basis Fourier analysis approach, which balances the accuracy required for modelling the anisotropic galaxy distribution in wide and deep surveys against the computational
cost of processing an unprecedented amount of cosmological data;
2) forward modelling from measurements of the tracer luminosity function to constraints
on the amplitude of relativistic corrections, which can give rise to scale-dependent modifications to clustering statistics on scales close to the Hubble horizon;
3) statistical techniques within the existing survey analysis framework to mitigate the impact on cosmological inference from non-Gaussian likelihoods and the parameter dependence
of covariance matrices.
A special focus in designing these new techniques for galaxy clustering analysis is the detection of local primordial non-Gaussianity ƒNL, which parametrises initial conditions of structure formation as set by inflationary models and imprints a scale-dependent signature in the bias of tracers on very large scales. These new approaches to analysing large-scale structure observations are also broadly applicable to constraining other standard cosmological parameters and should be particularly beneficial to forthcoming galaxy surveys such as the Dark Energy Spectroscopic Instrument and the Euclid mission, the scientific objectives of which include drastically tighter bounds on primordial non-Gaussianity as well as the test of gravitational theories on cosmological scales,
| Date of Award | Jan 2021 |
|---|---|
| Original language | English |
| Awarding Institution |
|
| Supervisor | Robert Crittenden (Supervisor) & William James Percival (Supervisor) |