Detailed properties of high redshift galaxies

Galaxies evolve throughout the history of the universe from the first star-forming sources, through gas-rich asymmetric structures with rapid star formation rates, to the massive symmetrical stellar systems observed at the present day. Determining the physical processes which drive galaxy formation and evolution is one of the most important questions in observational astrophysics. This thesis presents four projects aimed at improving our understanding of galaxy evolution from detailed measurements of star forming galaxies at high redshift.

We use resolved spectroscopy of gravitationally lensed z ≃ 2 - 3 star forming galaxies to measure their kinematic and star formation properties. The combination of lensing with adaptive optics yields physical resolution of ≃ 100 pc, sufficient to resolve giant Hii regions. We find that ~ 70 % of galaxies in our sample display ordered rotation with high local velocity dispersion indicating turbulent thick disks. The rotating galaxies are gravitationally unstable and are expected to fragment into giant clumps. The size and dynamical mass of giant Hii regions are in agreement with predictions for such clumps indicating that gravitational instability drives the rapid star formation. The remainder of our sample is comprised of ongoing major mergers. Merging galaxies display similar star formation rate, morphology, and local velocity dispersion as isolated sources, but their velocity fields are more chaotic with no coherent rotation.

We measure resolved metallicity in four lensed galaxies at z = 2.0 − 2.4 from optical emission line diagnostics. Three rotating galaxies display radial gradients with higher metallicity at smaller radii, while the fourth is undergoing a merger and has an inverted gradient with lower metallicity at the center. Strong gradients in the rotating galaxies indicate that they are growing inside-out with star formation fueled by accretion of metal-poor gas at large radii. By comparing measured gradients with an appropriate comparison sample at z = 0, we demonstrate that metallicity gradients in isolated galaxies must flatten at later times. The amount of size growth inferred by the gradients is in rough agreement with direct measurements of massive galaxies. We develop a chemical evolution model to interpret these data and conclude that metallicity gradients are established by a gradient in the outflow mass loading factor, combined with radial inflow of metal-enriched gas.

We present the first rest-frame optical spectroscopic survey of a large sample of low-luminosity galaxies at high redshift (L < L*, 1.5 < z < 3.5). This population dominates the star formation density of the universe at high redshifts, yet such galaxies are normally too faint to be studied spectroscopically. We take advantage of strong gravitational lensing magnification to compile observations for a sample of 29 galaxies using modest integration times with the Keck and Palomar telescopes. Balmer emission lines confirm that the sample has a median SFR ∼ 10 M_sun yr^−1 and extends to lower SFR than has been probed by other surveys at similar redshift. We derive the metallicity, dust extinction, SFR, ionization parameter, and dynamical mass from the spectroscopic data, providing the first accurate characterization of the star-forming environment in low-luminosity galaxies at high redshift. For the first time, we directly test the proposal that the relation between galaxy stellar mass, star formation rate, and gas phase metallicity does not evolve. We find lower gas phase metallicity in the high redshift galaxies than in local sources with equivalent stellar mass and star formation rate, arguing against a time-invariant relation. While our result is preliminary and may be biased by measurement errors, this represents an important first measurement that will be further constrained by ongoing analysis of the full data set and by future observations.

We present a study of composite rest-frame ultraviolet spectra of Lyman break galaxies at z = 4 and discuss implications for the distribution of neutral outflowing gas in the circumgalactic medium. In general we find similar spectroscopic trends to those found at z = 3 by earlier surveys. In particular, absorption lines which trace neutral gas are weaker in less evolved galaxies with lower stellar masses, smaller radii, lower luminosity, less dust, and stronger Lyα emission. Typical galaxies are thus expected to have stronger Lyα emission and weaker low-ionization absorption at earlier times, and we indeed find somewhat weaker low-ionization absorption at higher redshifts. In conjunction with earlier results, we argue that the reduced low-ionization absorption is likely caused by lower covering fraction and/or velocity range of outflowing neutral gas at earlier epochs. This result has important implications for the hypothesis that early galaxies were responsible for cosmic reionization. We additionally show that fine structure emission lines are sensitive to the spatial extent of neutral gas, and demonstrate that neutral gas is concentrated at smaller galactocentric radii in higher redshift galaxies.

The results of this thesis present a coherent picture of galaxy evolution at high redshifts 2 ≲ z ≲ 4. Roughly 1/3 of massive star forming galaxies at this period are undergoing major mergers, while the rest are growing inside-out with star formation occurring in gravitationally unstable thick disks. Star formation, stellar mass, and metallicity are limited by outflows which create a circumgalactic medium of metal-enriched material. We conclude by describing some remaining open questions and prospects for improving our understanding of galaxy evolution with future observations of gravitationally lensed galaxies.

Cosmic Explosions: Observations Of Infant Hydrogen-Free Supernovae Towards An Understanding Of Their Parent Systems

Radiation in the first days of supernova explosions contains rich information about physical properties of the exploding stars. In the past three years, I used the intermediate Palomar Transient Factory to conduct one-day cadence surveys, in order to systematically search for infant supernovae. I show that the one-day cadences in these surveys were strictly controlled, that the realtime image subtraction pipeline managed to deliver transient candidates within ten minutes of images being taken, and that we were able to undertake follow-up observations with a variety of telescopes within hours of transients being discovered. So far iPTF has discovered over a hundred supernovae within a few days of explosions, forty-nine of which were spectroscopically classified within twenty-four hours of discovery.

Our observations of infant Type Ia supernovae provide evidence for both the single-degenerate and double-degenerate progenitor channels. On the one hand, a low-velocity Type Ia supernova iPTF14atg revealed a strong ultraviolet pulse within four days of its explosion. I show that the pulse is consistent with the expected emission produced by collision between the supernova ejecta and a companion star, providing direct evidence for the single degenerate channel. By comparing the distinct early-phase light curves of iPTF14atg to an otherwise similar event iPTF14dpk, I show that the viewing angle dependence of the supernova-companion collision signature is probably responsible to the difference of the early light curves. I also show evidence for a dark period between the supernova explosion and the first light of the radioactively-powered light curve. On the other hand, a peculiar Type Ia supernova iPTF13asv revealed strong near-UV emission and absence of iron in the spectra within the first two weeks of explosion, suggesting a stratified ejecta structure with iron group elements confined to the slow-moving part of the ejecta. With its total ejecta mass estimated to exceed the Chandrasekhar limit, I show that the stratification and large mass of the ejecta favor the double-degenerate channel.

In a separate approach, iPTF found the first progenitor system of a Type Ib supernova iPTF13bvn in the pre-explosion HST archival mages. Independently, I used the early-phase optical observations of this supernova to constrain its progenitor radius to be no larger than several solar radii. I also used its early radio detections to derive a mass loss rate of 3e-5 solar mass per year for the progenitor right before the supernova explosion. These constraints on the physical properties of the iPTF13bvn progenitor provide a comprehensive data set to test Type Ib supernova theories. A recent HST revisit to the iPTF13bvn site two years after the supernova explosion has confirmed the progenitor system.

Moving forward, the next frontier in this area is to extend these single-object analyses to a large sample of infant supernovae. The upcoming Zwicky Transient Facility with its fast survey speed, which is expected to find one infant supernova every night, is well positioned to carry out this task.

Optimal Data Distributions in Machine Learning

In the first part of the thesis we explore three fundamental questions that arise naturally when we conceive a machine learning scenario where the training and test distributions can differ. Contrary to conventional wisdom, we show that in fact mismatched training and test distribution can yield better out-of-sample performance. This optimal performance can be obtained by training with the dual distribution. This optimal training distribution depends on the test distribution set by the problem, but not on the target function that we want to learn. We show how to obtain this distribution in both discrete and continuous input spaces, as well as how to approximate it in a practical scenario. Benefits of using this distribution are exemplified in both synthetic and real data sets.

In order to apply the dual distribution in the supervised learning scenario where the training data set is fixed, it is necessary to use weights to make the sample appear as if it came from the dual distribution. We explore the negative effect that weighting a sample can have. The theoretical decomposition of the use of weights regarding its effect on the out-of-sample error is easy to understand but not actionable in practice, as the quantities involved cannot be computed. Hence, we propose the Targeted Weighting algorithm that determines if, for a given set of weights, the out-of-sample performance will improve or not in a practical setting. This is necessary as the setting assumes there are no labeled points distributed according to the test distribution, only unlabeled samples.

Finally, we propose a new class of matching algorithms that can be used to match the training set to a desired distribution, such as the dual distribution (or the test distribution). These algorithms can be applied to very large datasets, and we show how they lead to improved performance in a large real dataset such as the Netflix dataset. Their computational complexity is the main reason for their advantage over previous algorithms proposed in the covariate shift literature.

In the second part of the thesis we apply Machine Learning to the problem of behavior recognition. We develop a specific behavior classifier to study fly aggression, and we develop a system that allows analyzing behavior in videos of animals, with minimal supervision. The system, which we call CUBA (Caltech Unsupervised Behavior Analysis), allows detecting movemes, actions, and stories from time series describing the position of animals in videos. The method summarizes the data, as well as it provides biologists with a mathematical tool to test new hypotheses. Other benefits of CUBA include finding classifiers for specific behaviors without the need for annotation, as well as providing means to discriminate groups of animals, for example, according to their genetic line.

Sunyaev-Zel'dovich observations using large-format millimeter arrays

Galaxy clusters are the largest gravitationally bound objects in the observable universe, and they are formed from the largest perturbations of the primordial matter power spectrum. During initial cluster collapse, matter is accelerated to supersonic velocities, and the baryonic component is heated as it passes through accretion shocks. This process stabilizes when the pressure of the bound matter prevents further gravitational collapse. Galaxy clusters are useful cosmological probes, because their formation progressively freezes out at the epoch when dark energy begins to dominate the expansion and energy density of the universe. A diverse set of observables, from radio through X-ray wavelengths, are sourced from galaxy clusters, and this is useful for self-calibration. The distributions of these observables trace a cluster's dark matter halo, which represents more than 80% of the cluster's gravitational potential. One such observable is the Sunyaev-Zel'dovich effect (SZE), which results when the ionized intercluster medium blueshifts the cosmic microwave background via Compton scattering. Great technical advances in the last several decades have made regular observation of the SZE possible. Resolved SZE science, such as is explored in this analysis, has benefitted from the construction of large-format camera arrays consisting of highly sensitive millimeter-wave detectors, such as Bolocam. Bolocam is a submillimeter camera, sensitive to 140 GHz and 268 GHz radiation, located at one of the best observing sites in the world: the Caltech Submillimeter Observatory on Mauna Kea in Hawaii. Bolocam fielded 144 of the original spider web NTD bolometers used in an entire generation of ground-based, balloon-borne, and satellite-borne millimeter wave instrumention. Over approximately six years, our group at Caltech has developed a mature galaxy cluster observational program with Bolocam. This thesis describes the construction of the instrument's full cluster catalog: BOXSZ. Using this catalog, I have scaled the Bolocam SZE measurements with X-ray mass approximations in an effort to characterize the SZE signal as a viable mass probe for cosmology. This work has confirmed the SZE to be a low-scatter tracer of cluster mass. The analysis has also revealed how sensitive the SZE-mass scaling is to small biases in the adopted mass approximation. Future Bolocam analysis efforts are set on resolving these discrepancies by approximating cluster mass jointly with different observational probes.

Beyond the blur: Construction and characterization of the first autonomous AO system and An AO survey of magnetar proper motions

Adaptive optics (AO) corrects distortions created by atmospheric turbulence and delivers diffraction-limited images on ground-based telescopes. The vastly improved spatial resolution and sensitivity has been utilized for studying everything from the magnetic fields of sunspots upto the internal dynamics of high-redshift galaxies. This thesis about AO science from small and large telescopes is divided into two parts: Robo-AO and magnetar kinematics.

In the first part, I discuss the construction and performance of the world’s first fully autonomous visible light AO system, Robo-AO, at the Palomar 60-inch telescope. Robo-AO operates extremely efficiently with an overhead < 50s, typically observing about 22 targets every hour. We have performed large AO programs observing a total of over 7,500 targets since May 2012. In the visible band, the images have a Strehl ratio of about 10% and achieve a contrast of upto 6 magnitudes at a separation of 1′′. The full-width at half maximum achieved is 110–130 milli-arcsecond. I describe how Robo-AO is used to constrain the evolutionary models of low-mass pre-main-sequence stars by measuring resolved spectral energy distributions of stellar multiples in the visible band, more than doubling the current sample. I conclude this part with a discussion of possible future improvements to the Robo-AO system.

In the second part, I describe a study of magnetar kinematics using high-resolution near-infrared (NIR) AO imaging from the 10-meter Keck II telescope. Measuring the proper motions of five magnetars with a precision of upto 0.7 milli-arcsecond/yr, we have more than tripled the previously known sample of magnetar proper motions and proved that magnetar kinematics are equivalent to those of radio pulsars. We conclusively showed that SGR 1900+14 and SGR 1806-20 were ejected from the stellar clusters with which they were traditionally associated. The inferred kinematic ages of these two magnetars are 6±1.8 kyr and 650±300 yr respectively. These ages are a factor of three to four times greater than their respective characteristic ages. The calculated braking index is close to unity as compared to three for the vacuum dipole model and 2.5-2.8 as measured for young pulsars. I conclude this section by describing a search for NIR counterparts of new magnetars and a future promise of polarimetric investigation of a magnetars’ NIR emission mechanism.

New Insights from Aperiodic Variability of Young Stars

Nearly all young stars are variable, with the variability traditionally divided into two classes: periodic variables and aperiodic or "irregular" variables. Periodic variables have been studied extensively, typically using periodograms, while aperiodic variables have received much less attention due to a lack of standard statistical tools. However, aperiodic variability can serve as a powerful probe of young star accretion physics and inner circumstellar disk structure. For my dissertation, I analyzed data from a large-scale, long-term survey of the nearby North America Nebula complex, using Palomar Transient Factory photometric time series collected on a nightly or every few night cadence over several years. This survey is the most thorough exploration of variability in a sample of thousands of young stars over time baselines of days to years, revealing a rich array of lightcurve shapes, amplitudes, and timescales.

I have constrained the timescale distribution of all young variables, periodic and aperiodic, on timescales from less than a day to ~100 days. I have shown that the distribution of timescales for aperiodic variables peaks at a few days, with relatively few (~15%) sources dominated by variability on tens of days or longer. My constraints on aperiodic timescale distributions are based on two new tools, magnitude- vs. time-difference (Δm-Δt) plots and peak-finding plots, for describing aperiodic lightcurves; this thesis provides simulations of their performance and presents recommendations on how to apply them to aperiodic signals in other time series data sets. In addition, I have measured the error introduced into colors or SEDs from combining photometry of variable sources taken at different epochs. These are the first quantitative results to be presented on the distributions in amplitude and time scale for young aperiodic variables, particularly those varying on timescales of weeks to months.

Stationary absolute distributions for chains of infinite order

Let {Ƶ_n}^∞_{n = -∞} be a stochastic process with state space S₁ = {0, 1, …, D – 1}. Such a process is called a chain of infinite order. The transitions of the chain are described by the functions

Q_i(i⁽⁰⁾) = Ƥ(Ƶ_n = i | Ƶ_{n - 1} = i ⁽⁰⁾₁, Ƶ_{n - 2} = i ⁽⁰⁾₂, …) (i ɛ S₁), where i⁽⁰⁾ = (i⁽⁰⁾₁, i⁽⁰⁾₂, …) ranges over infinite sequences from S₁. If i⁽ⁿ⁾ = (i⁽ⁿ⁾₁, i⁽ⁿ⁾₂, …) for n = 1, 2,…, then i⁽ⁿ⁾ → i⁽⁰⁾ means that for each k, i⁽ⁿ⁾_k = i⁽⁰⁾_k for all n sufficiently large.

Given functions Q_i(i⁽⁰⁾) such that

(i) 0 ≤ Q_i(i⁽⁰) ≤ ξ ˂ 1

(ii)D – 1/Ʃ/i = 0 Q_i(i⁽⁰⁾) Ξ 1

(iii) Q_i(i⁽ⁿ⁾) → Q_i(i⁽⁰⁾) whenever i⁽ⁿ⁾ → i⁽⁰⁾,

we prove the existence of a stationary chain of infinite order {Ƶ_n} whose transitions are given by

Ƥ (Ƶ_n = i | Ƶ_{n - 1}, Ƶ_{n - 2}, …) = Q_i(Ƶ_{n - 1}, Ƶ_{n - 2}, …)

With probability 1. The method also yields stationary chains {Ƶ_n} for which (iii) does not hold but whose transition probabilities are, in a sense, “locally Markovian.” These and similar results extend a paper by T.E. Harris [Pac. J. Math., 5 (1955), 707-724].

Included is a new proof of the existence and uniqueness of a stationary absolute distribution for an Nth order Markov chain in which all transitions are possible. This proof allows us to achieve our main results without the use of limit theorem techniques.