Price et al., available on arXiv.

Abstract: Accurate measurements of galaxy masses and sizes are key to tracing galaxy evolution over time. Cosmological zoom-in simulations provide an ideal test bed for assessing the recovery of galaxy properties from observations. Here, we utilize galaxies with Mstar ~ 10^10-10^11.5 Msun at z ~ 1.7-2 from the MassiveFIRE cosmological simulation suite, part of the Feedback in Realistic Environments (FIRE) project. Using mock multi-band images, we compare intrinsic galaxy masses and sizes to observational estimates. We find that observations accurately recover stellar masses, with a slight average underestimate of ~0.06 dex and a ~0.15 dex scatter. Recovered half-light radii agree well with intrinsic half-mass radii when averaged over all viewing angles, with a systematic offset of ~0.1 dex (with the half-light radii being larger) and a scatter of ~0.2 dex. When using color gradients to account for mass-to-light variations, recovered half-mass radii also exceed the intrinsic half-mass radii by ~0.1 dex. However, if not properly accounted for, aperture effects can bias size estimates by ~0.1 dex. No differences are found between the mass and size offsets for star-forming and quiescent galaxies. Variations in viewing angle are responsible for ~25% of the scatter in the recovered masses and sizes. Our results thus suggest that the intrinsic scatter in the mass–size relation may have previously been overestimated by ~25%. Moreover, orientation-driven scatter causes the number density of very massive galaxies to be overestimated by ~0.5 dex at Mstar ~ 10^11.5 Msun.