Testing theories of large scale structure formation and growth using galaxy surveys

The distribution of galaxies in the Universe is not random: rather, galaxies cluster in a structured way. The formation and growth of these large-scale structures (LSS) provides powerful dynamical probes for cosmology. This thesis explores two of these probes, namely redshift-space distortion (RSD) and the imprints of LSS on the Cosmic Microwave Background (CMB). Using galaxy surveys, I test the theory of structure growth in the context of the ΛCDM cosmological model. RSD probes the velocity field of LSS, which is influenced by the growth of matter fluctuations. I use the galaxy and group catalogues in GAMA survey to test the robustness of RSD in recovering unbiased growth rate fσ8 with different tracers. Specifically, galaxies are split into red and blue subsamples, and groups are divided into three stellar mass bins. The 2D group-galaxy cross-correlation function between these subsamples are interpreted by a linear model and a small-scale Finger of God convolution. Given an appropriate minimum fitting scale, I show that the subsamples give consistent growth rate, fσ8 = 0.25 ± 0.15, also in agreement with the Planck 2018 results. The imprints of LSS on the CMB correspond to the effects of weak gravitational lensing and the Integrated Sachs-Wolfe (ISW) effect. I measure these effects using the public DESI Legacy Survey, exploiting its large sky coverage and substantial depth for tomographic studies. After careful selection of galaxies and correction for various systematic effects, I assign photometric redshifts to galaxies based on g − r, r − z, and z − W1 colours, and construct four tomographic redshift bins in 0 < z < 0.8. The photo-z errors are accounted for using the galaxy auto- and cross-correlations between these redshift bins. Having a clean galaxy sample, I measure the cross-correlation C` between the galaxy density fields and the Planck CMB temperature and lensing convergence maps. The amplitudes of these measurements relative to the ΛCDM prediction using the fiducial Planck 2018 best-fit cosmology are Aκ = 0.901 ± 0.026 and AISW = 0.98 ± 0.35. While the ISW result is consistent with the fiducial cosmology, the CMB lensing result is noticeably lower. This low amplitude is interpreted in terms of a lower Ωm in combination with the total CMB lensing constraints. Finally, to address the excess stacked ISW signal from supervoids claimed in literature, I construct a superstructure catalogue using the four tomographic bins in the DESI Legacy Survey, and measure their stacked CMB lensing and ISW signals. The results are compared to the ΛCDM prediction from a mock catalogue that is based on N-body simulations and carefully matched to the data. I find a similar discrepancy in the lensing amplitude as in the cross-correlation scenario. Here, it is mainly contributed by density peaks at the higher redshift end. I also show that the detection of ISW signal from superstructure stacking is only mild, but is consistent with the ΛCDM prediction with a 95% upper limit of AISW = 1.51 using the full sample. Testing a range of superstructure subsamples, I demonstrate that the claimed excess signal may be due to look-elsewhere effect.