A Flexible Approach to Fitting Galactic Spectral Energy Distributions

Creating spectral energy distributions (SED) of galaxies with known age and metallicity parameters can be done computationally using standard python simulations, such as Conroy’s FSPS model. However, doing the inverse accurately has been a problem for current models; because of age-metallicity and mass degeneracies, galaxies with different characteristics can have SEDs that are extremely similar, making it hard to fit to a given SED and return the parameters that define that galaxy. Like Prospector, we have developed a computational model that uses Bayesian statistics and Markov Chain Monte Carlo (MCMC) framework to accurately determine the age, metallicity, and mass of a galaxy given the galaxy’s SED. Unlike Prospector, which can only fit models using one value for metallicity, we have developed a model that can predict multiple different metallicities of stellar populations within synthesized galactic SEDs accurately. Our model has been tested on spectral data taken from Conroy’s FSPS and can accurately fit randomly generated gaussian galactic SEDs with up to 30 different age-metallicity stellar population combinations, returning accurate mass-to-light-ratios as a metric of the accuracy of the fit. Along with this, the model can accurately predict ten different age bins of stellar populations within the galaxy as well as three different metallicity bins, with some degree of uncertainty. This makes it a more flexible model than Prospector in this regard, as Prospector can only fit to a single metallicity. Determining the unknown masses of galaxies in are universe is essential for accurately determining the dark matter fraction of the universe, as well as for galaxy formation and evolution. Along with this we hope to discover unknown degeneracies between age, metallicity, and mass when it comes to galactic SEDs.