Title: Probing the Large Scale Structure of the Universe Through Luminous Red Galaxies

 

Abstract: The study of Baryon Acoustic Oscillations (BAO) has become central to understand the nature of Dark energy. Dark energy is presumed to be responsible for driving the accelerated expansion of the universe. BAO refers to periodic fluctuations in the density of baryonic matter in the universe that were caused by acoustic oscillations created by counteracting forces of gravity and radiation pressure (analogous to the sound waves created in air by pressure differences). 

A precise measurement of the BAO feature requires distance information for a large number of galaxies. Massive galaxies are typically found in massive dark matter halos and cluster very strongly (Postman & Geller, 1984). This clustering enhances the BAO signal and also makes these galaxies ideal probes for the large-scale structure of the universe. Luminous Red Galaxies (LRGs) have played a vital role in the detection of BAO. These are relatively old massive elliptical systems dominated by old stars. These are the most massive galaxies in the z∼1 universe, showing characteristic 4000 ̊A break in their spectral energy distributions (SEDs) (Eisenstein et al., 2005). This thesis establishes a new algorithm for selecting high redshift LRGs for ongoing and upcoming spectroscopic surveys like SDSS-IV/eBOSS and DESI, which aim to precisely measure the BAO signal at high redshifts. I further adapt these methods for assembling the SDSS-IV/eBOSS LRG sample and explain these methods in detail. I present the results from SEQUELS, an ancillary program of SDSS-III/BOSS, to validate these methods. 

Large-scale clustering measurements are prone to systematic uncertainties associated with imaging surveys. This thesis employs modern statistical techniques to overcome these challenges. A multivariate regression analysis, primarily applied on the eBOSS LRG and quasar samples, is presented for a better understanding of the systematic uncertainties and their effects on the clustering measurements. Another key aspect of this work is the development of a machine-learning algorithm for estimating photometric redshifts. Applying this machinery yields accurate estimates of the distances of galaxies using their photometric properties alone, i.e., their colors and magnitudes. This work culminates with new clustering measurements that use these photometric redshifts in order to detect BAO in the eBOSS LRG sample. I have also presented initial results from an ongoing effort on clustering measurements of LRGs around quasars using the first two years of spectroscopic data from eBOSS.