Abstract

Recent works have provided clues to the physical drivers of the existence of the observed strong correlation between stellar mass and star formation rate (SFR) in ‘normal’ star forming galaxies. To date, the low mass end of this sequence is best measured at the highest redshifts with Hubble+Spitzer photometry, with ground-based K-band data to bridge the gap. We revisit the measurement of this relation from the Hubble Frontier Fields using BEAGLE , a Bayesian SED fitting code (Chevallard & Charlot 2016), to compute the posterior mass, SFR and redshift for main sequence galaxies spanning cosmic time. Our Bayesian hierarchical model of the main sequence (an extension of the work of Curtis-Lake et al. 2021) self-consistently propagates uncertainties on mass, SFR and redshift estimates to uncertainties on the main sequence parameters, whilst explicitly accounting for potential main sequence outliers. In this study we measure the main sequence relation between 1