Jean Michel Gomes, Physical and Mathematical foundations of the pyFADO spectral synthesis code
Despite significant progress over the past decades, all state-of-the-art population spectral synthesis (PSS) codes suffer from two major conceptual deficiencies that limit their potential of gaining sharp insights into the star formation history (SFH) and Chemical Enrichment History (CEH) of star-forming (SF) galaxies which introduce significant biases in studies of their physical properties (e.g., stellar mass and sSFR - Cardoso, Gomes & Papaderos 2018): i) the neglect of nebular continuum emission in spectral fits and ii) the lack of a mechanism that ensures consistency between the best-fitting SFH and the observed nebular emission characteristics (e.g., hydrogen Balmer-line luminosities and equivalent widths-EWs, shape of the continuum in the region around the Balmer and Paschen jump).
FADO (Fitting Analysis using Differential evolution Optimization) is a conceptually novel, publicly available (http://www.spectralsynthesis.org) PSS code originally written in object-oriented Fortran 2008 standard with the distinctive capability of permitting the identification of the SFH and CEH that best reproduces the observed nebular characteristics of a SF galaxy. This so far unique self-consistency the concept allows to significantly alleviate degeneracies in spectral synthesis, thereby opening a new avenue to the detailed exploration of the assembly history of galaxies. FADO is the first PSS code employing genetic Differential Evolution Optimization and machine learning techniques. A new python-bundle pyFADO (which stands for python-FADO) contains a set of bindings to FADO Fortran library in order to facilitate the runs and creation of automated plots, which will also be presented.
pyFADO was applied to synthetic SEDs that track the spectral evolution of stars and gas in mock galaxies that form their stellar mass (M*) according to different parametric SFHs. Our analysis indicates that FADO can recover the key physical and evolutionary properties of galaxies, such as M* and mass- and light-weighted mean age and metallicity, with an accuracy significantly better (~0.2 dex) than state-of-the-art purely-stellar fits.
This, in conjunction with various other currently unique elements in its mathematical concept and numerical realization, results in powerful improvements with respect to computational efficiency and uniqueness of the best-fitting SFHs. An outline of pyFADO, tests and illustrative examples from the local volume to high-redshift universe will be presented.