Abstract
Data from the ongoing Euclid survey will map out billions of galaxies in the Universe, covering more than a third of the sky. This data will provide a wealth of information about the large-scale structure (LSS) of the Universe and will have a significant impact on cosmology in the coming years. Cosmological N-body simulations are a vital ingredient in these analyses, offering a means to connect the distribution of galaxies to the underlying dark matter field. In this thesis, we develop an emulator-based halo model approach to forward model the relationship between cosmological parameters and the projected galaxy-galaxy twopoint correlation function (TPCF). We emulate the TPCF by generating mock-galaxy catalogues with the Halo Occupation Distribution (HOD) framework, using the large AbacusSummit simulation suite. The suite covers a wide 8-dimensional cosmological parameter space, whereas the HOD model we employ has 5 free parameters. The emulator constructed is efficient and can predict the TPCF accurately given a set of cosmological and HOD parameters. By integrating the TPCF, we obtain the projected correlation function, which is independent of redshift space distortions, allowing it to be directly confronted with observational measurements. We demonstrate that our emulator-based model for the projected correlation function achieves 1% precision at the scales of interest and recovers the true cosmology in a Markov Chain Monte Carlo analysis. Our model has certain limitations and is unable to properly discriminate among the effects between different parameters. The lack of information contained in the projected correlation is insufficient to robustly constrain parameters in a 13-dimensional parameter space with high degeneracy levels. A second limitation is that the simplistic HOD model implemented cannot accurately account for complex physical effects related to galaxy formation. Despite these limitations, our model shows promising results that agree with comparable studies. A major advantage of the model is its flexibility. Using our model as a starting point, we can combine with additional emulators constructed to account for other clustering statistics, thereby putting stronger constraints on the cosmological and HOD parameters. The contents of this work will be followed by a paper that will be submitted to A&A: Vikenes et al., “An Emulation-Based Model for the Full-Shape Projected Correlation Function”.
Supervisors:
Professor David F. Mota, Institute of Theoretical Astrophysics, UiO
Postdoctoral Fellow Cheng-Zong Ruan, Institute of Theoretical Astrophysics, UiO
Intern. assessor: Professor Boris V. Gudiksen, Institute of Theoretical Astrophysics, UiO
Extern. assessor: Post-DocTiago Batalha de Castro, Department of physics of UNITS / OATS