UC Davis

The Lambda-cold dark matter (ΛCDM) model for the cosmos has been a huge empirical success, demonstrating robust consistency with generations of cosmic microwave background (CMB) observations and galaxy surveys. However, fundamental questions persist regarding the physics of dark matter, dark energy and inflation. Moreover, emergent cosmological tensions and anomalies between CMB and other cosmological observations suggest beyond-ΛCDM physics. In this dissertation, I present work to resolve the cosmological tensions and anomalies, as well as tests of the ΛCDM model with the latest CMB data, in conjunction with complementary cosmological probes. I first show that dimensionless cosmological observables are invariant under a uniform scaling of the gravitational free-fall rates, photon scattering rate and the recombination rate, and that a “mirror world” dark sector model can exploit this scaling direction to resolve the Hubble tension. Using the “mirror world” dark sector model as a reference model, I then study the constraints of the amount of light relics by comparing departures of model predictions to the scaling invariant direction in the parameter space. I also investigate cosmological constraints on the atomic dark matter model from current CMB and baryon acoustic oscillation (BAO) data, and understand the physical impacts on the cosmological observables. I present the joint Bayesian inference of reconstructed CMB lensing potential bandpowers and unlensed E mode bandpowers from South Pole Telescope (SPT)-3G 2019-2020 polarization data. The cosmological analysis of the measured bandpowers confirms the existing Hubble and S8 tensions from the CMB polarization signal only and also yields a good simultaneous fit to the ΛCDM model with Planck and Atacama Cosmology Telescope (ACT) data. The SPT polarization data also confirm the weak evidence of excess lensing power over ΛCDM predictions seen in Planck and ACT, with signals dominated by CMB temperature data. The non-linear structure evolution is also detected for the first time at > 3σ significance in CMB lensing. To achieve concordance, further cosmologicalobservations with improved precision will be able to guide the model building effort.