2022MNRAS.516.3944M

We use GRUMPY, a simple regulator-type model for dwarf galaxy formation and evolution, to forward model the dwarf galaxy satellite population of the Milky Way (MW) using the Caterpillar zoom-in simulation suite. We show that luminosity and distance distributions of the model satellites are consistent with the distributions measured in the DES, PS1, and SDSS surveys, even without including a model for the orphan galaxies. We also show that our model for dwarf galaxy sizes can simultaneously reproduce the observed distribution of stellar half-mass radii, r_1/2, of the MW satellites and the overall r_1/2-M_⋆ relation exhibited by observed dwarf galaxies. The model predicts that some of the observed faint stellar systems with r_1/2 < 10 pc are ultra-faint dwarf galaxies. Scaling of the stellar mass M_⋆ and peak halo mass M_peak for the model satellites is not described by a power law, but has a clear flattening of M_⋆-M_peak scaling at $M_{\rm peak}\lt 10^8\, \, M_{\odot }$ imprinted by reionization. As a result, the fraction of low mass haloes ($M_{\rm peak}\lt 10^8 \, M_{\odot }$) hosting galaxies with M_V < 0 is predicted to be 50 per cent at $M_{\rm peak}\sim 3.6 \times 10^7\, \, M_{\odot }$. We find that such high fraction at that halo mass helps to reproduce the number of dwarf galaxies discovered recently in the HSC-SSP survey. Using the model we forecast that there should be the total of $440^{+201}_{-147}$ (68 per cent confidence interval) MW satellites with M_V < 0 and r_1/2 > 10 pc within 300 kpc and make specific predictions for the HSC-SSP, DELVE-WIDE, and LSST surveys.