Local SDSS galaxies in the Herschel Stripe 82 survey: A critical assessment of optically derived star formation rates

We study a set of 3319 galaxies in the redshift interval 0.04 < z < 0.15 with far-infrared (FIR) coverage from the Herschel Stripe 82 survey (HerS), and emission-line measurements, redshifts, stellar masses and star formation rates (SFRs) from the Sloan Digital Sky Survey (SDSS) (DR7) MPA/JHU data base. About 40 per cent of the sample are detected in the Herschel/SPIRE 250 μm band. Total infrared (TIR) luminosities derived from HerS andWidefield Infrared Survey Explorer (WISE) photometry allow us to compare infrared and optical estimates of SFR with unprecedented statistics for diverse classes of galaxies.We find excellent agreement between TIR-derived and emission line-based SFRs for HII galaxies. Other classes, such as active galaxies and evolved galaxies, exhibit systematic discrepancies between optical and TIR SFRs.We demonstrate that these offsets are attributable primarily to survey biases and the large intrinsic uncertainties of the D_n4000- and colour-based optical calibrations used to estimate the SDSS SFRs of these galaxies. Using a classification scheme which expands upon popular emission-line methods, we demonstrate that emission-line galaxies with uncertain classifications include a population of massive, dusty, metal-rich star-forming systems that are frequently neglected in existing studies. We also study the capabilities of infrared selection of star-forming galaxies. FIR selection reveals a substantial population of galaxies dominated by cold dust which are missed by the long-wavelength WISE bands. Our results demonstrate that Herschel large-area surveys offer the means to construct large, relatively complete samples of local star-forming galaxies with accurate estimates of SFR that can be used to study the interplay between nuclear activity and star formation.