The evolution of massive galaxies across cosmic time as traced by their stellar populations

Understanding how galaxies form and evolve is a central goal of modern astrophysics. This thesis addresses this question through the study of stellar populations, which act as fossil records of galaxies’ past star-formation and chemical enrichment histories. By combining methodological innovation with a multi-scale observational approach, the work develops and applies new tools to extract physical information from spectroscopic data of varying quality, aiming to achieve a consistent and physically meaningful description of how stellar populations encode the interplay between internal and environmental processes that drive galaxy evolution. The first part of the thesis establishes the methodological groundwork by developing and validating a strategy to measure stellar ages and metallicities from realistic, survey-quality spectra. Through an extensive set of simulations, it demonstrates that the joint use of optical and ultraviolet absorption-line indices enables accurate stellar population measurements even at moderate signal-to-noise ratios. This approach anticipates the capabilities of forthcoming intermediate-redshift surveys such as WEAVE-StePS, ensuring robust and homogeneous parameter estimates across large samples of galaxies. Building on this foundation, the Full-Index Fitting technique is introduced as an advanced Bayesian framework that exploits the full diagnostic potential of absorption features, combining their strength and shape in a unified model. Applied to the Cartwheel galaxy, a prototypical collisional ring system, the method reconstructs the spatially resolved distribution of stellar ages and metallicities using MUSE integral-field spectroscopy and multi-wavelength data from GALEX to JWST. The analysis reveals a predominantly old stellar disc with a younger population confined to the expanding outer ring, providing new insights into the dynamical and star-formation history of the system and demonstrating the capability of FIF to connect stellar population properties with interaction-driven processes. The technique is then extended to the study of massive quiescent galaxies in the COSMOS Wall at z ≈ 0.73, combining spectroscopic and photometric information within a hierarchical Bayesian analysis. The results reveal a clear environmental sequence: galaxies in X-ray-detected groups host the oldest stellar populations with the shortest formation timescales, while field galaxies are younger and formed over more extended periods. These findings provide one of the most detailed empirical characterisations of environmental quenching at intermediate redshift, showing that both internal and external mechanisms shape the evolution of massive galaxies. Finally, the methodology is applied to a large, homogeneous sample of BOSS and DESI luminous red galaxies spanning 0.8<z<0.15. The analysis finds no significant evolution in metallicity or [Mg/Fe] abundance ratios, while stellar ages evolve consistently with passive ageing. This demonstrates that the chemical enrichment and star-formation histories of the most massive galaxies were essentially completed by z ≈ 0.8, and that their subsequent evolution has been dominated by dry mergers and structural growth without renewed star formation. The thesis concludes by outlining future prospects. The developed Bayesian spectro-photometric tools can be applied to next-generation surveys such as WEAVE-StePS, 4MOST-StePS, 4MOST-WAVES, and MOONRISE, enabling comprehensive analyses of stellar population diversity across cosmic time. In parallel, the same techniques can be employed in spatially resolved studies with JWST and VLT/KMOS (e.g. EMPOWER) to map stellar population gradients and quenching mechanisms at high redshift. Together, these complementary approaches will link the internal structure of galaxies to their global evolution, paving the way toward a unified, data-driven understanding of how galaxies assembled, enriched, and quenched across cosmic time.