Add to Calendar 2026-02-27T14:00:00 2026-02-27T16:00:00 The Cosmic Imprint of Time: Decoding the Evolving Cosmos through Weak Gravitational Lensing Event Information: Abstract: A persistent discrepancy, known as the S_8 tension, challenges the modern cosmological paradigm: the large-scale structure clustering strength inferred from weak gravitational lensing consistently yields lower values than that derived from the cosmic microwave background.  This tension suggests potential limitations in our understanding of the structure formation, possibly arising from observational systematics, astrophysical processes, or new physics in the dark sector. In this thesis, I first investigate one potential origin of the tension using cosmic shear data from the KiDS-1000 and DES-Y3 surveys. I demonstrate that the data prefer a halo mass function with approximately ~ 30% fewer massive halos of M > 10^{14} solar mass, corresponding to a ~ 25% suppression in the matter power spectrum at k ~ 1 h Mpc^{-1} Adopting this suppression resolves the $S_8$ tension, but at the risk of contradicting independent cluster abundance and baryonic feedback measurements. To address the fundamental projection degeneracies in weak lensing, I develop a cleaner cosmological inference framework. I use the Bernardeau-Nishimichi-Taruya (BNT) transform, which elegantly reorganizes lensing kernels to isolate contributions from different redshifts and physical scales. I extend this into a practical pipeline that maps angular (ℓ-space) observables to well-defined three-dimensional (k-space) modes. This method delivers unbiased S_8 constraints that remain stable against changes in nonlinear modeling while retaining strong statistical power. Any technique separating redshift-dependent information must robustly account for intrinsic alignments (IA) -- the correlated shapes of galaxies due to local tidal fields, which have a different redshift dependence than cosmic shear. I test the pipeline's robustness against non-zero IA, showing it remains unbiased when the IA model is reasonably accurate. Furthermore, I propose a model-independent diagnostic using k- or z-slices to directly test the scale and redshift dependence of IA, demonstrating the practical use and resilience of the BNT approach against this major systematic. Finally, I apply this first physical-scale-based cosmic shear analysis to the KiDS-Legacy data. After using the BNT transform to remove contributions from nonlinear scales (k >= 0.33 Mpc^{-1}), I obtain a constraint of S_8 = 0.717^{+0.047}_{-0.046}. This result is 1.14σ lower than the case including nonlinear scales and 1.80σ lower than using only nonlinear scales (remove k < 0.33 Mpc^{-1}). This indicates that, at the sensitivity of the current stage of weak lensing survey, nonlinear astrophysical feedback does not significantly bias S_8. However, while the current data still does not allow us to distinguish whether this behavior arises from observational systematics or cosmology, it is notable that this trend is counterintuitive relative to most of the existing explanations proposed for solving the S_8 tension  Event Location: BUCH B312 (1866 Main Mall, UBC-V campus)