paper

Deriving star formation histories from photometry using energy balance spectral energy distribution modelling

arXiv:1507.07554 · doi:10.1093/mnras/stv1727

Abstract

Panchromatic spectral energy distribution (SED) fitting is a critical tool for determining the physical properties of distant galaxies, such as their stellar mass and star formation rate. One widely used method is the publicly available MAGPHYS code. We build on our previous analysis (Hayward & Smith 2015) by presenting some modifications which enable MAGPHYS to automatically estimate galaxy star formation histories (SFHs), including uncertainties, based on ultra-violet to far-infrared photometry. We use state-of-the art synthetic photometry derived by performing three-dimensional dust radiative transfer on hydrodynamic simulations of isolated disc and merging galaxies to test how well the modified MAGPHYS is able to recover SFHs under idealised conditions, where the true SFH is known. We find that while the SFH of the model with the best fit to the synthetic photometry is a poor representation of the true SFH (showing large variations with the line-of-sight to the galaxy and spurious bursts of star formation), median-likelihood SFHs generated by marginalising over the default MAGPHYS libraries produce robust estimates of the smoothly-varying isolated disk simulation SFHs. This preference for the median-likelihood SFH is quantitatively underlined by our estimates of (analogous to the goodness-of-fit estimator) and (the integrated absolute mass discrepancy between the model and true SFH) that strongly prefer the median-likelihood SFHs over those that best fit the UV-to-far-IR photometry. In contrast, we are unable to derive a good estimate of the SFH for the merger simulations (either best-fit or median-likelihood) despite being able to obtain a reasonable fit to the simulated photometry, likely because the analytic SFHs with bursts superposed in the standard MAGPHYS library are insufficiently general/realistic.

11 pages, 8 figures, accepted by MNRAS

References in corpus (28)

Cited by in corpus (38)