Abstract: This thesis presents a comprehensive investigation into the accuracy of spectral energy distribution (SED) modeling inferences of galaxy star-formation histories. SED modeling is an essential tool for inferring star-formation histories from nearby galaxy observations, but it is fraught with difficulty due to our incomplete understanding of stellar populations, chemical enrichment processes, and the nonlinear, geometry-dependent effects of dust on our observations. The dust attenuation curve of a galaxy, defined as the ratio of the total observed to emitted flux as a function of wavelength, depends on the chemical composition and grain size distribution of the dust as well as the geometry of the stars and dust relative to the observer. Utilizing hydrodynamic simulations and radiative transfer calculations, we generate simulated galaxy SEDs spanning far-ultraviolet (FUV) through far-infrared (FIR) wavelengths that account for the geometry-dependent effects of dust. These simulations enable the validation of SED modeling techniques against galaxies with known ground truth. We find that subgrid post-processing recipes that mitigate limitations in the temporal and spatial resolution of the simulations are required for producing FUV to FIR photometry that statistically reproduce the colors of galaxies in the nearby Universe. These simulations demonstrate a large variation in attenuation laws among galaxies, and that energy balance between dust attenuation and re-emission can be violated by up to a factor of 3. Inspired by the diversity among these simulated attenuation laws and the inability of commonly used existing models to reproduce them, we propose a novel dust attenuation model with three free parameters that can accurately recover the simulated attenuation curves as well as the best-fitting curves from the commonly used models. This new model is fully analytic and treats all starlight equally, in contrast to two-component dust attenuation models. Finally, we use our simulated observations to test the accuracy of SED modeling techniques. We find that the combined effect of model mismatches for high mass galaxies leads to inferred star-formation rates (SFRs) that are on average underestimated by a factor of 2 when fit to UV through IR photometry, and a factor of 3 when fit to UV through optical photometry. These biases lead to significant inaccuracies in the resulting sSFR-mass relations, with UV through optical fits showing particularly strong deviations from the true relation of the simulated galaxies. In the context of massive existing and upcoming photometric surveys, these results highlight that star-formation history inference from broadband photometry remains imprecise and inaccurate, and that there is a pressing need for more realistic testing of existing techniques.
Professor Michael Blanton (Thesis Advisor)
Professor Kyle Cranmer
Professor David Hogg
Professor Anthony Pullen
Professor Jeremy Tinker |