Thesis (M. Sc.)--University of Canterbury, 2008. / Typescript (photocopy). Includes bibliographical references (p. 138-182).

The presence of massive objects is detectable in observations via the gravitational lensing effect on light from more distant sources. From this effect it is possible to reconstruct the masses of clusters, and the distribution of matter within the cluster. However, further theoretical work needs to be done to properly contextualize any proposed projects involving, for instance, SALT data sets. Observational lensing studies use one of two techniques to recover the lens mass distribution: parametric (model dependent) techniques; and, a more recent innovation, non-parametric methods. The latter deserves further study as a tool for cluster surveys. To this end, we provide a comprehensive analysis of existing non-parametric algorithms and software, as well as estimates on the likely errors to be expected when used as an astronomical tool. / Thesis (M.Sc.)-University of KwaZulu-Natal, Westville, 2009.

Thesis (Ph. D.)--Rutgers University, 2008. / "Graduate Program in Physics and Astronomy." Includes bibliographical references.

Thesis (Ph. D.)--University of Washington, 2002. / Vita. Includes bibliographical references (p. 301-311).

For 20 years now, gravitational microlensing observations towards the Galactic bulge have provided us with a wealth of information about the stellar and planetary content of our Galaxy, which is inaccessible via other current methods. This thesis summarises work on two research topics that arose in the context of exoplanetary microlensing, but we take a step back and consider ways of increasing our understanding of more fundamental phenomena: firstly, stellar microlenses in our Galaxy that were stereoscopically observed and, secondly, the morphological variety of binary-lens light curves. In autumn 2008, the ESA Rosetta spacecraft surveyed the Galactic bulge for microlensing events. With a baseline of ∼1.6 AU between the spacecraft and ground observations, significant parallax effects can be expected. We develop a photometry pipeline to deal with a severely undersampled point spread function in the crowded fields of the Galactic bulge, making use of complementary ground observations. Comparison of Rosetta and OGLE light curves provides the microlens parallax π[subscript{E}] , which constrains the mass and distance of the observed lenses. The lens mass could be fully determined if future proper motion measurements were obtained, whereas the lens distance additionally requires the determination of the source distance. In the second project, we present a detailed study of microlensing light curve morphologies. We provide a complete morphological classification for the case of the equal-mass binary lens, which makes use of the realisation that any microlensing peak can be categorised as one of only four types: cusp-grazing, cusp-crossing, fold-crossing or fold-grazing. As a means for this classification, we develop a caustic feature notation, which can be universally applied to binary lens caustics. Ultimately, this study aims to refine light curve modelling approaches by providing an optimal choice of initial parameter sets, while ensuring complete coverage of the relevant parameter space.

The last twenty-five years have seen our understanding of the formation and abundance of planets revolutionised, thanks to the first detections of debris discs, and, a decade later, of the first extrasolar planets. Hardly a week now goes by without a planet discovery, and the range of methods used to search for planets has expanded to include techniques that are efficient at detecting different types of planets. By combining the discoveries of the various methods, we therefore have the opportunity to build a picture of planet populations across the Galaxy. In this thesis, I am presenting work done as a basis towards such an effort: first I present work carried out to improve modelling methods for gravitational microlensing events. Since the first microlensing observing campaigns, the amount of data of anomalous events has been increasing ever faster, meaning that the time required to model all observed anomalous events is putting a strain on available human and computational resources. I present work to develop a method to fit anomalous microlensing events automatically and show that it is possible to conduct a thorough and unbiased search of the parameter space, illustrating this by analysing an event from the 2007 observing season. I then discuss the possible models found with this method for this event, and their implication (Kains et al. 2009), and find that this algorithm locates good-fit models in regions of parameters that would have been very unlikely to be found using standard modelling methods. Results indicate that it is necessary to use a full Bayesian approach, in order to include prior information on the parameters. I discuss the analytical priors calculated by Cassan et al. (2009) and suggest a possible form of an automatic fitting algorithm by incorporating these priors in the algorithm used by Kains et al. (2009). Another topic with which this thesis is concerned is the evolution of debris discs around solar-type stars. Late-type stars are expected to be the most numerous host stars of planets detected with the microlensing technique. Understanding how their debris discs evolve equates to understanding the earliest stages of planet formation around these stars, allowing us to truly put our Solar System in perspective. Using the analytical model of Wyatt et al. (2007a), I modelled the evolution of infrared excess flux at 24 and 70 microns using published data of debris discs around solar-type (spectral types F, G and K) stars from the Spitzer Space Telescope. By comparing the results of this study to an analogous study carried out by for A stars by Wyatt et al. (2007b), I find that although best-fit parameters are significantly different for solar-type stars, this may be due to the varying number of inefficient emitters around stars of different spectral types. I suggest that although effective properties are different by an order of magnitude or more, intrinsic properties, while still different, are so by a much smaller factor. These differences may be due to the longer timescales over which solar-type stars evolve, which allow for the formation of larger and stronger planetesimals.