Constraining galaxy-halo connection and cosmology using DESI data

Abstract

The Dark Energy Spectroscopic Instrument (DESI) survey presents a unique opportunity to investigate the nature of dark energy and the evolution of the universe, offering an abundance of data that can enhance our comprehension of both the galaxy-halo connection and cosmological parameters. This thesis delves into a comprehensive exploration of the galaxy-halo connection within the context of the DESI survey, emphasizing its pivotal role in interpreting large galaxy surveys and extracting cosmological insights. We underscore the importance of the Halo Occupation Distribution (HOD) model, a statistical tool designed to populate galaxies from dark matter halos. A profound understanding of this model is essential for accurately interpreting the galaxy clustering data from DESI and for inferring the underlying cosmological model. In this thesis, we emphasize the significance of higher-order statistics, such as the projected three-point correlation function, for constraining the galaxy-halo connection. While two-point statistics have been extensively utilized, we demonstrate that the additional information provided by higher-order statistics enables us to glean more information on 1-halo scales and constrain the HOD parameters with greater precision. We adopt and generalize a swift HOD fitting pipeline to significantly reduce the required computation time for the three-point analysis. We present the HOD analysis of Luminous Red Galaxy (LRG) and Quasi-Stellar Object (QSO) samples derived from the One-Percent survey conducted by DESI. We scrutinize the HOD parameters as well as physical parameters such as the satellite fraction and mean halo mass of LRGs within a redshift range of 0.4 < z < 1.1, and QSOs within a redshift range of 0.8 < z < 2.1. Notably, for LRGs, we detect clear trends of redshift evolution for these physical parameters. We demonstrate how a well-constrained galaxy-halo connection model can be employed to extract cosmological information from DESI data, particularly on small scales. Finally, we present a systematic study of the Jackknife resampling technique, which is widely used in HOD fitting, in the estimation of the covariance matrix.