At present, all physical models of diffuse Galactic γ-ray emission assume that the distribution of cosmic-ray sources traces the observed populations of either OB stars, pulsars, or supernova remnants. However, since H_{2}-rich regions host significant star formation and numerous supernova remnants, the morphology of observed H_{2} gas (as traced by CO line surveys) should also provide a physically motivated, high-resolution tracer for cosmic-ray injection. We assess the impact of utilizing H_{2} as a tracer for cosmic-ray injection on models of diffuse Galactic γ-ray emission. We employ state-of-the-art 3D particle diffusion and gas density models, along with a physical model for the star-formation rate based on global Schmidt laws. Allowing a fraction, f_{H_{2}}, of cosmic-ray sources to trace the observed H_{2} density, we find that a theoretically well-motivated value f_{H_{2}}∼0.20-0.25 (i) provides a significantly better global fit to the diffuse Galactic γ-ray sky and (ii) highly suppresses the intensity of the residual γ-ray emission from the Galactic center region. Specifically, in models utilizing our best global fit values of f_{H_{2}}∼0.20-0.25, the spectrum of the galactic center γ-ray excess is drastically affected, and the morphology of the excess becomes inconsistent with predictions for dark matter annihilation.

Observations by the Fermi Large-Area Telescope have uncovered a significant γ-ray excess directed toward the Milky Way Galactic Center. There has been no detection of a similar signal in the stacked population of Milky Way dwarf spheroidal galaxies. Additionally, astronomical observations indicate that dwarf galaxies and other faint galaxies are less dense than predicted by the simplest cold dark matter models. We show that a self-interacting dark matter model with a particle mass of roughly 50 GeV annihilating to the mediator responsible for the strong self-interaction can simultaneously explain all three observations. The mediator is necessarily unstable, and its mass must be below about 100 MeV in order to decrease the dark matter density of faint galaxies. If the mediator decays to electron-positron pairs with a cross section on the order of the thermal relic value, then we find that these pairs can up-scatter the interstellar radiation field in the Galactic center and produce the observed γ-ray excess.

We model and constrain the unique indirect detection signature produced by dark matter particles that annihilate through a $U(1)$ gauge symmetry into dark photons that subsequently decay into three-photon final states. We focus on scenarios where the dark photon is long-lived, and show that $\gamma$-ray probes of celestial objects can set strong constraints on the dark matter/baryon scattering cross section that in many cases surpass the power of current direct detection constraints, and in some cases even peer into the neutrino fog.

Self-interacting dark matter (SIDM) models have been proposed to solve the small-scale issues with the collisionless cold dark matter paradigm. We derive equilibrium solutions in these SIDM models for the dark matter halo density profile including the gravitational potential of both baryons and dark matter. Self-interactions drive dark matter to be isothermal and this ties the core sizes and shapes of dark matter halos to the spatial distribution of the stars, a radical departure from previous expectations and from cold dark matter predictions. Compared to predictions of SIDM-only simulations, the core sizes are smaller and the core densities are higher, with the largest effects in baryon-dominated galaxies. As an example, we find a core size around 0.3 kpc for dark matter in the Milky Way, more than an order of magnitude smaller than the core size from SIDM-only simulations, which has important implications for indirect searches of SIDM candidates.