| Abstract: |
Extracting robust insights on galaxy physics from upcoming surveys requires a fast and flexible galaxy formation simulator. Current models face three major challenges. First, models capable of generating high-fidelity predictions are computationally prohibitive for the vast volumes these surveys will explore. Second, less resource-intensive models fail to capture the intricate physical processes and scales that these surveys will target. Third, the lack of a forward model for galaxy formation limits our understanding of observational uncertainties, weakening constraints on galaxy physics. Therefore, I am developing a galaxy-halo connection model that forward models populations of galaxy spectral energy distributions (SEDs) with consistent galaxy growth histories within cosmological structure formation. Building on the UniverseMachine framework (Behroozi et al. 2019), our model incorporates dark matter halo property-dependent star formation rates, dust, and metallicity to derive galaxy SEDs from dark matter-only simulations. This population-level SED analysis approach enables a fully physical, self-consistent model of galaxy stellar masses, star formation histories, dust, and metallicity for z=0-15, with significantly reduced uncertainties. A key outcome is the creation of realistic mock catalogs for current and future wide-field surveys, calibrated to match the latest observations. |
