We use FIRE-2 zoom cosmological simulations of Milky Way size galaxy halos to calculate astrophysical J-factors for dark matter annihilation and indirect detection studies. In addition to velocity-independent (s-wave) annihilation cross sections <σv>, we also calculate effective J-factors for velocity-dependent models, where the annihilation cross section is either either p-wave (∝ v^2/c^2) or d-wave (∝ v^4/c^4). We use 12 pairs of simulations, each run with dark-matter-only (DMO) physics and FIRE-2 physics. We observe FIRE runs produce central dark matter velocity dispersions that are systematically larger than in DMO runs by factors of ∼ 2.5 − 4. They also have a larger range of central (∼ 400 pc) dark matter densities than the DMO runs (ρFIRE/ρDMO ∼ 0.5 − 3) owing to the competing effects of baryonic contraction and feedback. At 3 degrees from the Galactic Center, FIRE J -factors are 3 − 60 (p-wave) and 10 − 500 (d-wave) times higher than in the DMO runs. The change in s-wave signal at 3 degrees is more modest and can be higher or lower (∼ 0.3−7). We find these results for s-wave are broadly consistent with the range of assumptions in most indirect detection studies, though our p-wave and d-wave values are significantly enhanced compared to what is commonly adopted. Contrary to past estimates, we suggest that thermal models with p-wave annihilation may be within range of detection in the near future. We then look at the shapes of the emission signal and find that the shape of the predicted J-factor-scaled emission is significantly different in FIRE compared to DMO. Contours of constant J on the sky are well-fit by ellipses. At a fixed fraction of peak J-factor on the sky, DMO runs have short-to-long axis ratios that are typically elliptical at ∼ 0.6, though demonstrate a broad range from 0.4 − 0.95. The FIRE runs are usually rounder, with axis ratios ∼ 0.8, and, importantly have narrower range of expected shapes (∼ 0.7 − 0.85). The long axisis always aligned with the Galactic plane in the FIRE simulations, to within ∼ 5◦. These predictions should be useful as priors in dark matter indirect detection studies, providing new constraints for a detection signal that our high resolution simulations have allowed.

This thesis presents a computational study of the formation and evolution of Milky-Way-mass galaxies using FIRE-2 cosmological zoom-in simulations. All simulated galaxies experience an early period of bursty star formation that transitions to a late-time steady phase of near-constant star formation. Stars form at different time have very different kinematic properties, morphologies, and chemical abundances. We classify stars in the central galaxy as bulge, thick-disc, and thin-disc according to their orbital properties and discover that these components are assembled in a time-sequence from early to late times, respectively. Stars born during the early bursty phase have more radial orbits and they evolve into a central spheroidal structure at z=0, while stars form during the late-time steady phase have more circular orbits and thin-disc-like morphology. Thick-disc stars form during a transitional period, mostly before the star formation settles down. Thus, we find a strong correlation between the age of thick-disc population at z=0 and the transition time. We also find that galaxies with earlier transition from bursty to steady star formation have higher thin-disc fractions at z=0. Bursty star formation at early time also triggers a significant amount of strong outflow events. During starbursts, the instantaneous fraction of stars born in outflows can be as high as ~ 20 - 50%. These outflow stars contribute substantially to the in-situ stellar halo. Statistically, stellar populations form at different times tend to maintain their disc/bulge-like kinematics, with only moderate evolution over time. Mergers do alter kinematics, but play a secondary role in ''disc-settling'' process. They do not destroy the thin disc nor change broad trend between star formation transition time and thin/thick-disc/bulge properties. If our simulations are representative of the Universe, then stellar archaeological studies of the Milky Way can provide a window into past star-formation modes in the Galaxy. Current age estimates of the Galactic thick disc would suggest that Milky Way transitioned from bursty to steady phase > 6.5 Gyr ago; prior to that time the Milky Way likely lacked a stable thin disc and underwent intense outflow activities.

The Universe’s mass is predominately comprised of dark matter, yet its exact particle nature remain elusive. The dark energy (Λ) plus Cold Dark Matter (ΛCDM) model of cosmology reigns as the prevalent model of our Universe, as it matches the observed large-scale structure. The formation mechanism of these so-called dark matter halos introduces many possible particle candidates for CDM models and further manifests attributes in the global abundance and the underlying mass distribution. Precise detection of these objects in the low-mass regime could further pave way to revealing the particle nature of dark matter. An exciting route to reveal the existence of dark matter structures within our Universe that have mass scales smaller than the faintest galaxies (i.e., Ultra-faint Dwarfs) is based on their lingering presence on strong gravitational lensed systems. The contents within this thesis provides, for the first time, in depth analysis of crucial lensing quantities derived from numerical simulations of dark matter halos and galaxies within the ΛCDM paradigm. Importantly, we explore models pertaining to the “line-of-sight” (LOS; i.e., field) perturbers, as these are one of most contributing population of dark matter halos to significantly impact the lensing signal, as compared to the subhalo population. We demonstrate that clustering in the vicinity of the lens host halo is an important component in any lensing configuration, as it produces a clear enhancement to the lensing signal relative to an assumption of unclustered halos with substructure. We also introduce an analytical surface density profile for low-mass dark matter halos that are parameterized by the so-called “projected concentration”, making it useful for modeling LOS perturbers in strong gravitational lensing models. Finally, we then extend this surface density model using a statistical sample of simulated dark matter halos to infer the median evolution of the projected concentration and the scatter as a function of mass.

The framework developed here inevitably broadens the scope of strong-lensing analyses for probing dark matter in a ΛCDM universe, thus demonstrating the power of strong gravitational lensing as a unique probe of dark matter physics.

We introduce an extension of the ELVIS project to account for the effects of the Milky Way galaxy on its subhalo population. Our simulation suite, Phat ELVIS, consists of twelve high-resolution cosmological dark matter-only (DMO) zoom simulations of Milky Way-size $\Lambda$CDM~ haloes ($M_{\rm v} = 0.7-2 \times 10^{12}$ M$_\odot$) along with twelve re-runs with embedded galaxy potentials grown to match the observed Milky Way disc and bulge today. The central galaxy potential destroys subhalos on orbits with small pericentres in every halo, regardless of the ratio of galaxy mass to halo mass. This has several important implications. 1) Most of the {\tt Disc} runs have no subhaloes larger than $V_{\rm max} = 4.5$ km s$^{-1}$ within $20$ kpc and a significant lack of substructure going back $\sim 8$ Gyr, suggesting that local stream-heating signals from dark substructure will be rare. 2) The pericentre distributions of Milky Way satellites derived from \textit{Gaia} data are remarkably similar to the pericentre distributions of subhaloes in the {\tt Disc} runs, while the DMO runs drastically over-predict galaxies with pericentres smaller than 20 kpc. 3) The enhanced destruction produces a tension opposite to that of the classic `missing satellites' problem: in order to account for ultra-faint galaxies known within $30$ kpc of the Galaxy, we must populate haloes with $\vpeak \simeq 7$ km s$^{-1}$ ($M \simeq 3 \times 10^{7} \Msun$ at infall), well below the atomic cooling limit of $\vpeak \simeq 16 ~\kms$ ($M \simeq 5 \times 10^{8} \Msun$ at infall). 4) If such tiny haloes do host ultra-faint dwarfs, this implies the existence of $\sim 1000$ satellite galaxies within 300 kpc of the Milky Way.

While astronomers believe that we understand how the universe works on large scales, the processes that shape the interiors of galaxies are still very much an area of ongoing research. Today, we are in a unique position to utilize incredibly powerful supercomputers to make predictions about our universe that can, in turn, be confirmed by groundbreaking observations made by modern telescopes such as the James Webb Space Telescope (JWST) or the upcoming Vera C. Rubin Observatory. Using high-resolution, cosmological “zoom” simulations of galaxies run by the Feedback In Realistic Environments (FIRE) collaboration, I advance our understanding of the internal galactic processes responsible for two galaxy scaling relations: (1) A new predicted scaling relation between a dwarf galaxy’s age and strength of its metallicity gradient (the Gradient-Strength Galaxy Age relation, or GSGA). (2) The Radial Acceleration Relation (RAR), which I show is a natural consequence of ΛCDM, contrary to some predictions. I also present a test of two self interacting dark matter implementations in two different simulation codes. Finally, I describe how understanding the effects of these internal processes allows us to use scaling relations as tools to test galaxy formation and dark matter theory.

We study dwarf satellite galaxy quenching using observations from the Geha et al. (2012) NSA/SDSS catalog together with CDM cosmological simulations to facilitate selection and interpretation. We show that fewer than 30% of dwarfs (M* ~ 10^8.5-10^9.5 Msun) identified as satellites within massive host halos (Mhost ~ 10^12.5-10^14 Msun) are quenched. We conclude that whatever the action triggering environmental quenching of dwarf satellites, the process must be highly inefficient. We investigate a series of simple, one-parameter quenching models in order to understand what is required to explain the low quenched fraction and conclude that either the quenching timescale is very long (> 9.5 Gyr, a “slow starvation” scenario) or that the environmental trigger is not well matched to accretion within the virial volume.

We further present FIRE/Gizmo hydrodynamic zoom-in simulations of isolated dark matter halos, two each at the mass of classical dwarf galaxies (Mvir ~ 10^10 Msun) and ultra-faint galaxies (Mvir ~ 10^9 Msun). The resulting central galaxies lie on an extrapolated abundance matching relation from M* ~ 10^6 to 10^4 Msun without a break. Our dwarfs with M* ~ 10^6 Msun each have 1-2 well-resolved satellites with M* = 3 - 200 x 10^3 Msun. Even our isolated ultra-faint galaxies have star-forming subhalos. We combine our results with the ELVIS simulations to show that targeting the ~ 50 kpc regions around nearby isolated dwarfs could increase the chances of discovering ultra-faint galaxies by ~35% compared to random pointings.

The well-resolved ultra-faint galaxies in our simulations (M* ~ 3 - 30 x 10^3 Msun) form within

Mpeak ~ 0.5 - 3 x 10^9 Msun halos. Each has a uniformly ancient stellar population (> 10 Gyr) owing to reionization-related quenching. More massive systems, in contrast, all have late-time star formation. Our results suggest that Mhalo ~ 5 x 10^9 Msun is a probable dividing line between halos hosting reionization “fossils” and those hosting dwarfs that can continue to form stars in isolation after reionization.

Finally, we perform a systematic Bayesian analysis of rotation vs. dispersion support (vrot/sigma) in 40 dwarf galaxies throughout the Local Volume (LV) over a stellar mass range 10^3.5 Msun < M* < 10^8 Msun. We find that the stars in 80% of the LV dwarf galaxies studied – both satellites and isolated systems – are dispersion-supported. These results challenge the traditional view that the stars in gas-rich dwarf irregulars (dIrrs) are distributed in cold, rotationally-supported stellar disks, while gas-poor dwarf spheroidals (dSphs) are kinematically distinct in having dispersion supported stars. We apply the same Bayesian analysis to four of the FIRE/Gizmo hydrodynamic zoom-in simulations of isolated dwarf galaxies (10^9 Msun < Mvir < 10^10 Msun) and show that the simulated isolated dIrr galaxies have stellar ellipticities and stellar vrot/sigma ratios that are consistent with the observed population of dIrrs and dSphs without the need to subject these dwarfs to any external perturbations or tidal forces. We posit that most dwarf galaxies form as puffy, dispersion-dominated systems, rather than cold, angular momentum-supported disks. If this is the case, then transforming a dIrr into a dSph may require little more than removing its gas.

Dwarf galaxies are among the most numerous objects in the Universe, and also appear to be the most dark matter (DM) dominated; consequently, they provide strong tests on the standard paradigm of hierarchical galaxy formation: cold dark matter with a cosmological constant (ΛCDM). Due to their low luminosities, however, observational studies of dwarfs have remained limited to the nearby Universe, with a primary focus on the satellites of the Milky Way (MW). Upcoming surveys will relax the observational constraints, allowing for studies of dwarf galaxies well beyond the virial radius (Rv) of the MW, where the presence of the Andromeda (M31) galaxy may have a measurable impact. In this thesis, I introduce the ELVIS (Exploring the Local Volume in Simulations) suite, which resolves the formation and evolution of Local Group (LG)-like pairs of galaxies, chosen to resemble the MW and M31 host halos, along with thousands of dwarf halos that surround them. Using ELVIS, I demonstrate that simulations of isolated MW-size hosts do not correctly predict dwarf counts and kinematics beyond Rv of the MW, and also explore the faint-end relationship between stellar mass (Mstar) and halo mass (Mhalo). I then use ELVIS to explore the “too-big- to-fail” (TBTF) problem, a challenge for ΛCDM that points out the overabundance of large (maximum circular velocity Vmax 􏰆~ 30 km/s) dwarf halos relative to observations, finding TBTF to be ubiquitous to MW-size hosts and the fields around them. Finally, I explore two possible explanations for TBTF in depth: fluctuations in the baryonic mass in the center of

a dwarf halo, driven by supernovae feedback, which I find lacks the energy to solve TBTF, and a “running” initial power spectrum, as motivated by the BICEP-2 measurement of primordial gravitational waves, which significantly alleviates the problem and may enhance the impact of other processes, including baryonic feedback.

Dark Matter (DM) accounts for the vast majority of mass in the universe, but the particle identity of dark matter remains a mystery. Uncovering the fundamental nature of DM remains one of the greatest challenges facing modern physics. Because the only information about DM comes from astrophysical observations, these are the best sources to constrain models. Dwarf galaxies present an especially tantalizing regime to investigate dark physics, as they have the highest ratio of dark to luminous matter and therefore will be most affected by differences between DM models. Additionally, this is precisely the scale where generic dark matter theories have the most difficulty reproducing astronomical observations, leading to the missing satellites, core-cusp and too-big-to-fail (TBTF) problems.

A particular class of models with nuclear scale self-interactions (called SIDM) has emerged as a promising candidate. SIDM naturally forms cored halos, which may alleviate both the core-cusp and TBTF problems. However, at larger scales the the interplay between SIDM halos and the galaxies residing in them is poorly understood, complicating this picture greatly. In this thesis I present numerical simulations of SIDM and CDM halos investigating these issues. I show that at dwarf scales SIDM cross sections as small as $0.5~\cmg$ solve the TBTF and core-cusp problems, and that cross sections 2 orders of magnitude larger are not ruled out. I have also embedded gravitational potentials that approximate realistic galaxies in simulations of larger haloes in order to test the impact of galaxy formation on SIDM halos. These simulations show that SIDM is indistinguishable from CDM in systems where the galaxy dominates the central region, but in galaxies with higher mass-to-light ratios or less centrally concentrated baryons it is possible to constrain SIDM cross sections. In the galaxy cluster regime I show that an SIDM cross section of $0.1-0.2~\cmg$ is preferred to CDM or other SIDM cross sections.

One of the goals of theoretical galaxy formation is to use simulations combined with observations in order to constrain the fundamental physics that govern the formation of galaxies. Of particular interest are physical processes that can impact the formation of small-scale structures. Among the things we can potentially constrain with a combination of observations and simulations of small-scale structures are stellar feedback, reionization, and dark matter particle physics. We investigate several promising avenues to constrain fundamental physics with simulations of galaxy formation. First, we use the star formation histories (SFH) for Local Group dwarf galaxies in order to constrain high-redshift stellar mass functions (SMFs). Second, we look at the destruction of substructure around large galaxies, and how it may impact the detection of substructure via subhalo lensing. Finally, we use full hydrodynamics simulations of Local Volume dwarf galaxies in order to investigate the impact of stellar feedback on the age distributions of stars within these galaxies.