It is believed that supermassive black holes (\mbh\ $\sim 10^6M_{\odot}- 10^{9}M_{\odot}$) in the center of galaxies are a vital component of galaxy evolution. This idea is driven by two observations: (i) black hole (BH) masses are tightly correlated to host galaxy properties in local quiescent galaxies, and (ii) BH accretion and star formation histories track one another closely. These two observations suggest that the growth of accreting BHs is intimately connected to that of their host galaxies. However, the origin and physical mechanism remains unknown and is an active area of astrophysical research.

The goal of this dissertation is twofold: (i) characterize the structure and kinematics of the broad emission line region (BLR) to better understand the central regions of accreting BHs and (ii) improve the way BH mass measurements are calibrated across cosmic time. While both goals aim to provide insight into the role BHs play in galaxy evolution, the latter is most relevant for studies that probe the origin of the BH scaling relations via characterization across redshift.

In this dissertation, I model the H$\beta$ BLR of nine Active Galactic Nuclei (AGN) observed during the Lick AGN Monitoring Project (LAMP) 2016 reverberation mapping campaign and investigate whether there are any luminosity-dependent trends in the structure and kinematics of the \H$\beta$ BLR. Then, I combine the LAMP 2016 modeling results with previous LAMP reverberation mapping campaigns and search for a way to improve how BH masses are calibrated across cosmic time. Finally, I expand on the initial development and testing of \textsc{caramel-gas}, which aims to model the gas density field of the BLR to model multiple emission lines simultaneously, e.g., H$\alpha$ and H$\beta$, and learn about the surroundings of active BHs. However, further testing and development of \textsc{caramel-gas} beyond the work presented in this thesis is required before achieving this goal. Overall, future reverberation mapping campaigns with sufficient data quality and variability are needed to confirm the results presented in this thesis.

Three open problems in our understanding of early-type galaxies are 1) identifying the

process(es) responsible for their rapid size evolution, 2) accurately constraining the stellar

IMF and its variations in the population, 3) measuring the density profile of their dark

matter halo. We use strong lensing as the main diagnostic tool to address these issues.

We first dissected a massive elliptical galaxy in its stellar and dark matter components,

measuring both its IMF and the inner slope of the dark matter halo. We then collected

a sample of 45 strong lenses in the redshift interval 0.2 < z < 0.8 and used them, in

combination with lenses from other surveys, to measure the slope of the total density

profile, the stellar IMF and the dark matter mass in the population of massive early-type

galaxies, and their time evolution. Finally, we used our measurements of the evolution of

the density slope to test a galaxy growth scenario based on purely dissipationless mergers.

Our main results are: the stellar IMF of massive early-type galaxies is significantly heavier

than that of the Milky Way and correlates with galaxy mass; the dark matter halo has

a steep slope in at least one system; more compact galaxies have less dark matter than

their extended counterparts at fixed redshift and stellar mass; early-type galaxies evolve

while keeping the slope of their total density profile approximately constant. This last result cannot be reproduced with purely dissipationless mergers, therefore a little amount

of dissipation is required.

This thesis presents progress in two subfields of astrophysics, first in galactic astronomy, describing how a large survey of ground-state silicon monoxide emission can probe asymptotic giant branch stellar envelope expansion, and second in observational cosmology, presenting novel constraints on primordial black holes (PBH) as a fraction of dark matter by leveraging strong gravitational lensing. The Galactic project searched the Very Large Array Bulge Asymmetries and Dynamical Evolution (BAaDE) survey for maser and thermal emission from the $J=1-0, v=0$ ground vibrational state transition. I present results characterizing these sources as a nearby disk population and derive the expansion velocity of the emission region by fitting thermal components of lines to an emission model. For the small number of sources with corresponding carbon monoxide emission, I find the expansion velocities to be similar for both the thermal SiO and CO, evidence that both emissions are produced in the same region above the star where the gas has reached terminal velocity. In the cosmology portion of the thesis, the objective is to use a flux-ratio analysis to put constraints on intermediate-mass PBH as a possible fraction of dark matter. A population of PBH within the dark matter halo of a lensing galaxy would affect the magnification, and therefore the flux ratios, of the images in a quadruply-imaged quasar (quad). By forward-modeling the flux ratios using an Approximate Bayesian Computing technique, I obtain a constraint on PBH within the mass range of $10^4 $M$_{\odot}

To explore the chemo-structural properties of galaxies and understand quantitatively the cycling of baryons at the peak epoch of cosmic star formation, I developed a highly effective method for sub-kiloparsec scale spatially resolved spectroscopy of strongly lensed galaxies using space-based wide-field slitless grism data. Applying this method to the deep Hubble Space Telescope (HST) near-infrared grism observations, I obtained precise gas-phase metallicity maps for a large sample of star-forming galaxies in the redshift range of 1.2 ≲ z ≲ 2.3. Over half of the galaxies in my sample reside in the dwarf mass regime (Mstar ≲ 10^9 Msun), making my sample the first statistically representative sample of high-redshift dwarf galaxies with their metallicity spatial distribution measured with sufficient resolution. The metallicity maps obtained in my work reveal a variety of baryonic physics, such as efficient radial mixing from tidal torques, rapid accretion of low-metallicity gas, and various feedback processes which can significantly influence the chemo-structural properties of star-forming galaxies. For the first time, I discovered two dwarf galaxies at z~2 displaying strongly inverted metallicity radial gradients, suggesting that powerful galactic winds triggered by central starbursts carry the bulk of stellar nucleosynthesis yields to the outskirts. I also observe an intriguing correlation between stellar mass and metallicity gradient, consistent with the "downsizing" galaxy formation picture that more massive galaxies are more evolved into a later phase of disk growth, where they experience more coherent mass assembly at all radii and thus show shallower metallicity gradients. Furthermore, 10% of the metallicity gradients measured in my sample are inverted, which are hard to explain by currently existing hydrodynamical simulations and analytical chemical evolution models. My method can be readily applied to data from future space missions employing grism instruments, e.g., JWST, Euclid, WFIRST. Combined with the continuous input of HST resources, these data will revolutionize our understanding of the chemo-structural evolution of galaxies throughout vast cosmic time.

The reionization of intergalactic hydrogen in the universe’s first billion years was likely driven by the first stars and galaxies, so its history encodes information about their prop- erties. But the timeline of reionization is not well-measured and it is still unclear whether galaxies alone can produce the required ionizing photons. In my thesis I have focused on two ways to use galaxies at our current observational frontiers to constrain reionization and high redshift galaxy evolution.

One tool is the UV luminosity function (LF), which traces the evolution of star-forming galaxies and their ionizing photons. Accurately measuring LFs and understanding their evolution are important for understanding the connections between galaxies and their dark matter halos. I developed a simple, but powerful, semi-analytic model for LF evolution assuming star formation is driven by dark matter accretion. This has proved remarkably consistent with observations, and implies the majority of star formation at high redshifts occurs in low mass galaxies. I also developed a technique to improve the accuracy of LF by accounting for gravitational lensing magnification bias.

Secondly, Lyman alpha (Lyα) emission from galaxies can probe the intergalactic medium (IGM) ionization state as Lyα photons are strongly attenuated by neutral hydrogen, but this requires disentangling physics on pc to Gpc scales. I developed a new forward-modeling Bayesian framework combining cosmological IGM simulations with interstellar medium models to infer the IGM neutral hydrogen fraction from observations of Lyα emission. My thesis presents new measurements of the neutral fraction at z ∼ 7 and z ∼ 8, which, along with other independent constraints, provide increasing evidence for the bulk of reionization occurring at z ∼ 6 − 8. This is consistent with reionization being driven by ultra-faint galaxies with a low average ionizing photon escape fraction. We also show that reionization impacts Lyα emission from different galaxy populations in different ways: UV bright galaxies living in overdense regions that reionize early have potentially visible Lyα even in a highly neutral IGM.

The upcoming James Webb Space Telescope was designed to observe galaxies at Cosmic Dawn. Throughout this thesis I have made predictions for what JWST may observe in the context of high redshift galaxy evolution.

Recently, a significant tension has been reported between two measurements of the Hubble constant (H_0) from early-Universe (e.g., cosmic microwave background) and late-Universe probes (e.g., cosmic distance ladder). If systematic errors are ruled out in these measurements, then new physics extending the ΛCDM model will be required to resolve the tension. Therefore, different independent probes of H_0 -- such as the strong-lensing time-delay -- are essential to confirm or resolve the tension. The measured time-delays between the lensed images of a background quasar constrain H_0, as they depend on the absolute physical distances in the lens configuration. I led a team from the STRong-lensing Insights into the Dark Energy Survey (STRIDES) collaboration to analyze the lens DES J0408-5354. I modeled the mass distribution of this lens using Hubble Space Telescope imaging, and combined it with analyses from my collaborators to infer H_0 = 74.2-3.0+2.7 km s^-1 Mpc^-1 with the highest precision (3.9 per cent) from a single lens to date. This measurement agrees well with both the previous sample of six lenses from the H0LiCOW collaboration and other late-Universe probes, thus it increases the aforementioned tension. To confirm or resolve this tension at the 5σ level -- the gold standard of detecting new physics -- we need to increase the sample size and improve precision per system while keeping the systematics under control. The large amount of required investigator time (~1 year per lens) is currently the main bottleneck to increase the sample size. I present an automated lens-modeling framework that will enable rapid increment of the sample size in the near future. I also show, through simulation, that incorporating the spatially resolved kinematics of the lensing galaxy improves the precision of H_0 per system. Additionally, I develop the first general method to efficiently compute the lensing properties of any given elliptical mass distribution. By allowing any radial shape of mass profile, this method helps to avoid any systematic that may potentially arise from adopting only a few specific parameterizations. I forecast that a sample of ~40 lenses with spatially resolved kinematics will provide sub-per-cent precision in H_0 within the next decade.

Measuring the masses of supermassive black holes in active galactic nuclei (AGN) allows us to trace their evolution over cosmic time and understand how black holes coevolve with their host galaxies. We present a new technique to measure black hole masses and constrain the structure of the broad line region in AGN using reverberation mapping data. We begin by developing a simply parameterized phenomenological model of the broad line region geometry and dynamics and apply this model to high-quality reverberation mapping data for six AGN from the Lick AGN Monitoring Project 2008 and 2011 datasets. The results of this analysis provide the most precise AGN black hole masses from reverberation mapping to date and the first detailed constraints on the geometry and dynamics of the broad line region emission. Specifically, we find that the shape of the broad line region is generally a close to face-on thick disk with preferential emission from the far side, and that the dynamics range from inflow to near-circular orbits. In addition, we present photometric AGN light curves using image subtraction for the Lick AGN Monitoring Project 2011 dataset as a first step towards modeling the broad line region in a larger sample of AGN.

There is overwhelming evidence for a link between supermassive black holes (BHs) and their host galaxies, yet the physical mechanisms behind the connection are poorly understood. Discerning between possible co-evolutionary scenarios requires precise measurements of BH masses and host properties across cosmic time. BH mass measurements outside the local universe require bright active galactic nuclei (AGNs), but the AGN light overwhelms the light of the host, making host property measurements difficult. Further, BH mass estimates rely on the unresolvable broad emission line region (BLR), and the unknown nature of the BLR limits their precision.

In this dissertation, I use a unique combination of gravitational lensing and reverberation mapping techniques to tackle both issues. First, I present a semi-supervised classification approach for identifying rare lensed quasars based solely on photometry in wide-field surveys. Lenses are powerful tools for measuring host properties, but studies are limited by the small sample size, so discovery efforts are crucial. Using this technique, I successfully discovered three new lenses in the Sloan Digital Sky Survey, with several other candidates awaiting further follow-up. This approach is flexible and can easily be modified to target specific types of lenses as well as lenses in surveys covering different portions of the sky.

Next, I use a modeling framework to measure precise BH masses for seven AGNs and infer the structural and kinematic properties of the Hβ BLR. This nearly doubles the sample size of modeled BLRs, allowing for a correlation analysis between BLR properties and AGN observables that can be used to improve BH mass measurements for all AGNs. I then model Hβ, CIV, and Lyα in NGC 5548 using data from 85 Hubble Space Telescope orbits and multiple ground-based telescopes to understand the similarities and differences between the optical and UV BLRs. High-z BH mass measurements rely on UV lines, but our current understanding of the UV BLR is severely lacking, so understanding their connection is critical for decreasing high-z BH mass uncertainties. Finally, I tie the previous projects together with the most precise high-z BH mass measurement to date, using data from a 4.5 year monitoring campaign of a strongly lensed quasar at z=2.805. Accounting for gravitational time delays, the campaign exceeds 6 years in duration. Understanding the co-evolution of BHs and their hosts will require measurements of this precision at high-z, and this research demonstrates that gravitational lensing makes this level of precision obtainable.

The abundance of low mass halos is one of the key predictions of LCDM, which remains at apparent odds with observations of luminous structure. We present new measurements of the spatial distribution and the cumulative luminosity function of satellite galaxies up to a thousand times fainter than their hosts, as a function of host stellar mass and morphology between redshifts 0.1 and 0.8, using imaging from the GOODS and COSMOS fields in conjunction with a rigorous statistical analysis. We demonstrate how these measurements provide powerful new constraints for abundance matching and cosmological simulations in the context of both warm and cold dark matter, and how future measurements of faint satellite colors using CANDELS, will provide important distinguishing power between warm and cold dark matter models.

In addition, we present results from a complementary gravitational lens modeling project in which we use strongly lensed AGN narrow-line emission in order to detect dark matter subhalos, demonstrating a promising new method for measuring the subhalo mass function in thousands of lensed systems which will be discovered in ongoing and future optical surveys.