Weak gravitational lensing is now well established as a powerful cosmological probe, particularly for studying large scale structure growth in the Universe. The vast majority of weak lensing experiments to date use optical and near infrared observations which are well suited to the requirements in source densities and shape analysis. In this thesis we outline the prospects associated with weak lensing surveys from radio observations. This can offer key advantages to optical counterpart studies such as the well defined observing beam pattern of a radio telescope and a window into a much broader observed redshift distribution. In addition to the prospect of radio weak lensing surveys alone, combining with optical counterparts in a cross-correlation study has been shown to mitigate uncorrelated systematics, further motivating the case for radio based weak lensing studies. The correlation of galaxy shapes through multi-wavelength observations will affect the noise on the cosmological power spectrum in cross-correlation analysis. We use radio and optical observations of the COSMOS field with the VLA and HST respectively, accompanied with simulations for calibration in order to measure shape correlations between wavelength regimes. Although we do not detect a correlation between optical and radio shapes, a lower limit on the intrinsic astrophysical scatter was placed at >0.212pi (or 38.2 degrees), through a Monte Carlo simulation of source catalogues with the measured uncertainties. The SuperCLASS experiment aims to measure a weak lensing signal with radio observations from a super-cluster field. We introduce the radio data, collected with the e-MERLIN and JVLA, and the reduction steps taken. Assisted by simulations, we have designed a shape measurement pipeline (SuperTRAP) which performs additional phase rotation and averaging steps to extract visibility sets on a source by source basis followed by image plane shape analysis. A series of staged tests of increasing complexity are outlined here and evaluated by the shape recovery bias and efficiency. Finally we present the optical counterpart observations and shape analysis for the SuperCLASS field, with data collected by the Subaru Suprime-Cam. Observational systematics are measured to form representative PSF models in each CCD exposure and the subsequent shape analysis from the I band photometry is presented. Shear analysis from the measured power spectrum shows good agreement with theoretical predictions. From the measured shear power spectrum we detect a strong signal in the E-mode band powers, equivalent to a 9.31sigma detection. Our measurements from the B-mode and E-B cross band powers suggest negligible contamination from systematics. The optical analysis presented here will provide the counterpart analysis to the radio for future cross-correlation studies.

Observations of gravitationally microlensed quasars offer a unique opportunity to probe quasar structure on extremely small scales. In this thesis, we conduct extensive microlensing simulations and compare with observational data to constrain quasar accretion discs, and conduct preliminary probes of broad emission line region structure. This analysis is done using a new single-epoch imaging technique that requires very little telescope time, and yet produces results that are comparable to those obtained from long-term monitoring campaigns. / We begin by exploring the impact of variable smooth matter percentage and source size on microlensing simulations. Adding a smooth matter component affects minimum and saddle point images differently, broadening the magnification distribution for the saddle point image significantly. However, increasing the radius of the background source washes out this difference. The observation of suppressed saddle point images can therefore only be explained by microlensing with a smooth matter component if the background source is sufficiently small. We use these simulations, in combination with I-band imaging of the lensed quasar MG 0414+0534 to constrain the radius of the quasar source. This demonstrates the viability of a single-epoch imaging method for constraining quasar structure. / This technique is then expanded to single-epoch multi-band observations, in order to constrain the radial profile of quasar accretion discs as a function of observed wavelength. We present new Magellan observations of two gravitationally lensed quasars: MG 0414+0534 and SDSS J0924+0219. We also analyse two epochs of Q2237+0305 data obtained from the literature. Our results are compared with four fidicial accretion disc models. At the 95 per cent level, only SDSS J0924+0219 is inconsistent with any of the accretion disc models. When we combine the results from all three quasars -- a first step towards assembling a statistical sample -- we find that the two steepest accretion disc models are ruled out with 68 per cent confidence. / In addition, we are also able to use our microlensing simulations to constrain the smooth matter percentages in the lenses at the image positions. In both MG 0414+0534 and SDSS J0924+0219 we find smooth matter percentages that are inconsistent with zero. A smooth matter percentage of approximately 50 per cent is preferred in MG 0414+0534, and approximately 80 per cent in SDSS J0924+0219. Q2237+0305 is usually assumed to have a smooth matter percentage of zero at the image positions, as they lie in the bulge of the lensing galaxy. Though our measurement is consistent with a zero smooth matter percentage, there is a peak in the probability distribution at a value 20 per cent. This is perhaps a hint of additional intervening structures along the line of sight to the background quasar. / We test the sensitivity of our accretion disc constraints to a range of modelling parameters. These include errors in lens modelling, Bayesian prior probability selection, errors in observational data, and the microlens mass function. Constraints on the power-law index relating source radius to observed wavelength are found to be relatively unaffected by changes in the modelling parameters. Constraints on source radii are found to be more strongly affected. / Finally, the broad emission line region of Q2237+0305 is examined. Gemini IFU observations are presented clearly showing differential microlensing across the velocity profile of the CIII] emission line. To analyse this signature, we present three simple broad emission line region models: a biconical outflow, a Keplerian disc, and spherical infall. A method is developed to compare the shapes of simulated flux ratio spectra with the observed spectrum. We are unable to discriminate between the biconical outflow and Keplerian disc models when we average over all viewing angles and orientations. The spherical infall model, however, does not fit the observed data. We also find that for the non-spherically symmetric geometries, low inclination angles are essentially incompatible with the observations. This analysis offers hope that with sufficiently high signal-to-noise observations, differential microlensing signatures may allow us to constrain the geometry and kinematics of this poorly understood portion of quasar structure.

Gravitational lensing, initially a phenomenon used as a solid confirmation of General Relativity, has defined itself in the past decade as a standard astrophysical tool. The ability of a lensing system to produce multiple images of a luminous source is one of the aspects of gravitational lensing that is exploited both theoretically and observationally to improve our understanding of the Universe. In this thesis, within the field of multiple imaging we explore the case of maximal lensing, that is, the configurations and conditions under which a set of deflecting masses can produce the maximum number of images of a distant luminous source, as well as a study of the value for this maximum number itself. We study the case of a symmetric distribution of n-1 point-mass lenses at the vertices of a regular polygon of n-1 sides. By the addition of a perturbation in the form of an n-th mass at the center of the polygon it is proven that, as long as the mass is small enough, the system is a maximal lensing configuration that produces 5(n-1) images. Using the explicit value for the upper bound on the central mass that leads to maximal lensing, we illustrate how this result can be used to find and constrain the mass of planets or brown dwarfs in multiple star systems. For the case of more realistic mass distributions, we prove that when a point-mass is replaced with a distributed lens that does not overlap with existing images or lensing objects, an additional image is formed within the distributed mass while positions and numbers of existing images are left unchanged. This is then used to conclude that the maximum number of images that n isolated distributed lenses can produce is 6(n-1)+1. In order to explore the likelihood of observational verification, we analyze the stability properties of the symmetric maximal lensing configurations. Finally, for the cases of n=4, 5, and 6 point-mass lenses, we study asymmetric maximal lensing configurations and compare their stability properties against the symmetric case.

Gravitational lensing, initially a phenomenon used as a solid confirmation of General Relativity, has defined itself in the past decade as a standard astrophysical tool. The ability of a lensing system to produce multiple images of a luminous source is one of the aspects of gravitational lensing that is exploited both theoretically and observationally to improve our understanding of the Universe. In this thesis, within the field of multiple imaging we explore the case of maximal lensing, that is, the configurations and conditions under which a set of deflecting masses can produce the maximum number of images of a distant luminous source, as well as a study of the value for this maximum number itself. We study the case of a symmetric distribution of n-1 point-mass lenses at the vertices of a regular polygon of n-1 sides. By the addition of a perturbation in the form of an n-th mass at the center of the polygon it is proven that, as long as the mass is small enough, the system is a maximal lensing configuration that produces 5(n-1) images. Using the explicit value for the upper bound on the central mass that leads to maximal lensing, we illustrate how this result can be used to find and constrain the mass of planets or brown dwarfs in multiple star systems. For the case of more realistic mass distributions, we prove that when a point-mass is replaced with a distributed lens that does not overlap with existing images or lensing objects, an additional image is formed within the distributed mass while positions and numbers of existing images are left unchanged. This is then used to conclude that the maximum number of images that n isolated distributed lenses can produce is 6(n-1)+1. In order to explore the likelihood of observational verification, we analyze the stability properties of the symmetric maximal lensing configurations. Finally, for the cases of n=4, 5, and 6 point-mass lenses, we study asymmetric maximal lensing configurations and compare their stability properties against the symmetric case.

In this thesis, I explore how the environments of both galaxy and cluster-scale strong gravitational lenses affect studies of cosmology and the properties of the earliest galaxies. Galaxy-scale lenses with measured time delays can be used to determine the Hubble constant, given an accurate lens model. However, perturbations from structures along the line of sight can introduce errors into the measurement. I use data from a survey towards known lenses in group environments to calculate the external shear in these systems, which is typically marginalized over in standard lens analyses. In three of six systems where I compare the independently-calculated environment shear to lens model shears, the quantities disagree at greater than 95% confidence. We explore possible sources of this disagreement. Using these data, I generate fiducial lines of sight and insert mock lenses with assumed input physical and cosmological parameters and find that those parameters can be recovered with ∼ 5-10% scatter when uncertainties in my characterization of the environment are applied. The lenses in groups have larger bias and scatter. I predict how well new time delay lenses from LSST will constrain H₀ and find that an ensemble of 500 quad lenses will recover H₀ with ∼ 2% bias with ∼ 0.3% precision. On larger scales, galaxy cluster lenses can magnify the earliest galaxies into detectability. While past studies have focused on single massive clusters, I investigate the properties of lines of sight, or "beams", containing multiple cluster-scale halos in projection. Even for beams of similar total mass, those with multiple halos have higher lensing cross sections on average. The optimal configurations for maximizing the cross section are also those that maximize faint z ∼ 10 detections. I present a new selection technique to identify beams in wide-area photometric surveys that contain high total masses and often multiple clusters in projection as traced by luminous red galaxies. I apply this technique to the Sloan Digital Sky Survey and present the 200 most promising beams. Several are confirmed spectroscopically to be among the highest mass beams known with some containing multiple clusters. These are among the best fields to search for faint high-redshift galaxies.

This thesis describes an investigation into weak gravitational lensing, a unique and powerful astronomical tool for the study of dark matter on large scales. Lensing distorts background images, inducing correlations in the observed ellipticities of galaxies, and these correlations can be used to estimate many characteristics of the Universe. Key to all weak lensing studies is a reliable and unbiased method to detect weak lensing distortions from observed galaxy images that are contaminated by Earth and telescope-based shearing and smearing distortions. A new galaxy model-fitting technique is presented that has been developed in order to satisfy this requirement, which will also permit future signal-to-noise optimised measurements of weak lensing shear. Model-fitting provides a good alternative to the standard scite{KSB} method (KSB), and comparisons between the two techniques are drawn from an analysis of deep {it R} band imaging from the COMBO-17 survey, revealing strong evidence for the presence of bias in KSB galaxy shape measurement. With the galaxy model-fitting technique, an investigation into the effectiveness of the Oxford Dartmouth Thirty degree survey (ODT) for gravitational lensing studies is presented, resulting in the detection of weak gravitational lensing by large scale structure, or `cosmic shear', in 0.7 square degrees of the best seeing ODT images. One concern for all cosmic shear studies is that the weak lensing signal, manifest in the weakly correlated ellipticities of distant galaxies, is contaminated by the intrinsic alignment of close galaxy pairs, potentially induced during galaxy formation by physical interactions such as tidal forces. This contamination is investigated theoretically, through numerical simulations, and observationally, with an analysis of the COMBO-17 survey and the study of published results from the Red-sequence Cluster survey and the VIRMOS-DESCART survey, concluding that the intrinsic alignment effect is at the lower end of the range of theoretical predictions. The impact of intrinsic galaxy alignments on cosmological parameter estimation is investigated, with an analysis of the weak lensing results from the COMBO-17 survey. When marginalising over the observationally constrained intrinsic alignment signal, the amplitude of the matter power spectrum sigma_8 is reduced by ~0.03 to sigma_8(Omega_m / 0.27)^{0.6} = 0.71 pm 0.11, where Omega_m is the matter density parameter. With distance information from either spectroscopy or photometric redshifts, the down-weighting of nearby galaxy pairs in weak lensing analysis can be optimised to virtually eliminate the systematic errors in the shear signal arising from intrinsic galaxy alignments, leaving a much smaller, largely statistical error. This method is applied to the photometric redshift sample of the COMBO-17 survey. Weak lensing measurements from the forthcoming SuperNova/Acceleration Probe weak lensing survey (SNAP), and the Canada-France-Hawaii Telescope Legacy survey, are expected to be contaminated on scales >1 arcminute by intrinsic alignments at the level of ~ 1% and ~2% respectively. Division of the SNAP survey for lensing tomography significantly increases the contamination in the lowest redshift bin to ~7% and possibly higher. Removal of the intrinsic alignment effect by the downweighting of nearby galaxy pairs will therefore be vital for the lensing tomography studies of SNAP.

The origin, nature and distribution of dark matter in the universe form some of the biggest questions in modern astrophysics. Dark matter is distributed on a wide range of scales in the universe. This thesis concentrates on galactic scales, attempting to lower the veil and probe the structure of dark matter in galaxies. (For complete abstract open document)

In the formation of galaxy groups and clusters, the dark matter haloes containing satellite galaxies are expected to be tidally stripped in gravitational interactions with the host. We use galaxy-galaxy weak lensing to measure the average mass of dark matter haloes of satellite galaxies as a function of projected distance to the centre of the host, since stripping is expected to be greater for satellites closer to the centre of the cluster. We further classify the satellites according to their stellar mass: Assuming that the stellar component of the galaxy is less disrupted by tidal stripping, stellar mass can be used as a proxy of the infall mass. We study the stellar-to-halo mass relation of satellites as a function of the cluster-centric distance to measure tidal stripping. We use the shear catalogues of the Dark Energy Survey (DES) science verification archive, the Canada-France-Hawaii Lensing Survey (CFHTLenS) and the CFHT Stripe 82 surveys, and we select satellites from the redMaPPer catalogue of clusters. For galaxies located in the outskirts of clusters, we find a stellar-to-halo mass relation in good agreement with the theoretical expectations from Moster et al. for central galaxies. In the centre of the cluster, we find that this relation is shifted to smaller halo mass for a given stellar mass. We interpret this finding as further evidence for tidal stripping of dark matter haloes in high-density environments.

Strong gravitational lensing provides an independent measurement of the Hubble parameter (H-0). One remaining systematic is a bias from the additional mass due to a galaxy group at the lens redshift or along the sightline. We quantify this bias for more than 20 strong lenses that have well-sampled sightline mass distributions, focusing on the convergence kappa and shear gamma. In 23% of these fields, a lens group contributes >= 1% convergence bias; in 57%, there is a similarly significant line-of-sight group. For the nine time-delay lens systems, H-0 is overestimated by 11(-2)(+3)% on average when groups are ignored. In 67% of fields with total kappa >= 0.01, line-of-sight groups contribute greater than or similar to 2x more convergence than do lens groups, indicating that the lens group is not the only important mass. Lens environment affects the ratio of four (quad) to two (double) image systems; all seven quads have lens groups while only 3 of 10 doubles do, and the highest convergences due to lens groups are in quads. We calibrate the gamma-kappa relation: log(kappa(tot)) = (1.94 +/- 0.34)log(gamma(tot)) + (1.31 +/- 0.49) with an rms scatter of 0.34 dex. Although shear can be measured directly from lensed images, unlike convergence, it can be a poor predictor of convergence; for 19% of our fields, kappa is greater than or similar to 2 gamma. Thus, accurate cosmology using strong gravitational lenses requires precise measurement and correction for all significant structures in each lens field.

One possibility for detecting low-amplitude pulsational variations is through gravitational microlensing. During a microlensing event, the temporary brightness increase leads to improvement in the signal-to-noise ratio, and thereby better detectability of pulsational signatures in light curves. We explore this possibility under two primary considerations. The first is when the standard point-source and point-lens approximation applies. In this scenario, dividing the observed light curve by the best-fitted microlensing model leads to residuals that result in pulsational features with improved uncertainties. The second is for transit events (single lens) or caustic crossing (binary lens). The point-source approximation breaks down, and residuals relative to a simple best-fitted microlensing model display more complex behaviour. We employ a Monte Carlo simulation of microlensing of pulsating variables toward the Galactic bulge for the surveys of OGLE and of KMTNet. We demonstrate that the efficiency for detecting pulsational signatures with intrinsic amplitudes of <0.25 mag during single and binary microlensing events, at differences in χ2 of Δχ2 > 350, is $\sim \!50\!-\!60{{\ \rm per\ cent}}$. The maximum efficiency occurs for pulsational periods P ≃ 0.1-0.3 d. We also study the possibility that high-magnification microlensing events of non-radially pulsating stars could be misinterpreted as planetary or binary microlensing events. We conclude that small asymmetric features around light curve peaks due to stellar pulsations could be misdiagnosed with crossing (or passing close to) small caustic curves.