Includes bibliographical references (p. 273-279).

Bibliography: pages 100-104. / The detection and measurability of microlensing anomalies including planet detection, resolved source, blending and parallax effects are investigated by way of an extensive computer simulation as a function of the observation strategy. The most efficient strategy has a planet detection probability exceeding 753 for the most favourable binary geometry, while results for typical event follow up and survey observations agree well with previous work. A complex dependence of planet detection on the resolved source radius is discovered leading to a possible increase in the probability of detecting planets with mass ratios smaller than q = 10⁻³

Includes bibliographical references (p. 129-146). / Rapid quasi-coherent oscillations were detected in the optical light curves of 24 cataclysmic variable stars (CVS). In 12 of these systems (TU Men, WW Cet, HX Peg, BP CrA, BR Lup, HP Nor, AG Hya, TW Vir, PU CMa, V426 Ooph, V1193 Ori, and CR Boo) for the first time. The results contribute to the observational record of the phenomenology of dwarf nova oscillations (DNOs), quasi-periodic oscillations (QPOs), and longer period dwarf nova oscillations (lpDNOs), strengthen the correlation, valid over nearly six orders of magnitude in frequency, between the ratio of time scales of different classes of oscillations in white dwarf, neutron star, and black hole binaries, and show that the recently recognized lpDNOs occur fairly commonly in high mass transfer rate CVs.

Neutral atomic hydrogen (Hi) is the raw fuel from which the star-forming molecular gas forms and is therefore an important tracer of galaxy evolution. Due to the intrinsic faintness of the Hi emission line (observed at rest at 21 cm), galaxies beyond a few hundred megaparsecs are difficult to observe directly with current radio telescopes. However, in the next year, MeerKAT and other SKA pathfinder telescopes will begin operating and enable deeper, large surveys (e.g. LADUMA) of neutral gas in galaxies. Hi Stacking is an observational technique that will be highly exploited to learn about the Hi content of galaxies that are not directly detected. Stacking involves combining the Hi spectra of all the galaxies in a distant sample, thereby generating a high signal-to-noise measure of their average Hi content. This work presents a new Python-based package capable of stacking Hi galaxy spectra. This package will be used to stack the Hi spectra of high-redshift galaxies observed with the MeerKAT telescope. In this work the package is applied to a sample of galaxies observed as part of the Nançay Interstellar Baryon Legacy Extragalactic Survey (NIBLES, van Driel et al. (2016)) to learn more about the gas properties of galaxies in the local universe. Using the stacking technique, we are able to recover the average Hi mass of different galaxy populations for which there was no Hi directly detected. In order to obtain the average gas properties that best represent the overall galaxy populations, we also stack both the Hi detected and non-detected spectra from the NIBLES survey. We find that our gas fraction vs. stellar mass distribution results agree well with previous stacking experiments (Brown et al., 2015; Catinella et al., 2010; Fabello et al., 2011a) and the NIBLES sample enables us to probe an order of magnitude lower in stellar mass. We find a dependence on the underlying stellar mass distribution for our gas fraction vs. NUV - r colour, especially when comparing to Brown et al. (2015) and Fabello et al. (2011a).

With the advent of powerful telescopes such as the Square Kilometre Array (SKA), its precursor MeerKAT and the Large Synoptic Survey Telescope (LSST), we are entering a golden era of multiwavelength transient astronomy. The large MeerKAT science project ThunderKAT may dramatically increase the detected number of radio transients. Currently radio transient datasets are still very small, allowing spectroscopic classification of all objects of interest. As the event rate increases, follow-up resources must be prioritised by making use of early classification of the radio data. Machine learning algorithms have proven themselves invaluable in the context of optical astronomy, however it has yet to be applied to radio transients. In the burgeoning era of multimessenger astronomy, incorporating data from different telescopes such as MeerLICHT, Fermi, LSST and the gravitational wave observatory LIGO could significantly improve classification of events. Here we present MALT (Machine Learning for Transients): a general machine learning pipeline for multiwavelength transient classification. In order to make use of most machine learning algorithms, "features" must be extracted from complex and often high dimensional datasets. In our approach, we first interpolate the data onto a uniform grid using Gaussian processes, we then perform a wavelet decomposition and finally reduce the dimensionality using principal component analysis. We then classify the light curves with the popular machine learning algorithm random forests. For the first time, we apply machine learning to the classification of radio transients. Unfortunately publicly available radio transient data is scarce and our dataset consists of just 87 light curves, with several classes only consisting of a single example. However machine learning is often applied to such small datasets by making use of data augmentation. We develop a novel data augmentation technique based on Gaussian processes, able to generate new data statistically consistent with the original. As the dataset is currently small, three studies were done on the effect of the training set. The classifier was trained on a non-representative training set, achieving an overall accuracy of 77.8% over all 11 classes with the known 87 lightcurves with just eight hours of observations. The expected increase in performance, as more training data are acquired, is shown by training the classifier on a simulated representative training set, achieving an average accuracy of 95.8% across all 11 classes. Finally, the effectiveness of including multiwavelength data for general transient classification is demonstrated. First the classifier is trained on wavelet features and a contextual feature, achieving an average accuracy of 72.9%. The classifier was then trained on wavelet features and a contextual feature, together with a single optical flux feature. This addition improves the overall accuracy to 94.7%. This work provides a general approach for multiwavelength transient classification and shows that machine learning can be highly effective at classifying the influx of radio transients anticipated with MeerKAT and other radio telescopes.

Includes bibliographical references.

Includes bibliographical references. / The field of Asteroseismology provides powerful techniques for the study of pulsating white dwarf stars. If these techniques could be applied to the white dwarf primaries of Cataclysmic Variable stars, the study of accretion in interacting binary systems would benefit enormously. Our understanding of normal modes in white dwarf pulsators would benefit likewise from the study of how they respond to accretion. This dissertation presents the analysis of preliminary observations of the Dwarf Nova GW Librae , the first Cataclysmic Variable star found to have a DAY primary. A brief introduction to Cataclysmic Variables and the pulsating white dwarfs is given, and the significance of GW Librae to these fields of study is discussed. In this dissertation I present the analysis of 7 weeks of high speed CCD observations obtained on GW Librae during 1997 and 1998. The power in the pulsation spectrum concentrates principally in regions near 376s and 650s. GW Librae has a pulsation spectrum that shows dramatic changes on a monthly basis, typical of the pulsating DA white dwarfs.

Includes bibliographical references (leaves 99-101).

The term cataclysmic variable comprises several types of variable star - novae, recurrent novae, dwarf novae; and to these we add the nova-like variables, stars which exhibit many of the characteristics of the novae but which have not been observed to erupt. U Geminorum, discovered by Hinds in 1855, is the nomotype of the dwarf novae. These stars are characterised by large, abrupt, temporary increases in brightness that occur erratically at intervals of ten days to a year or more. The dwarf novae are faint objects, few reaching 11th magnitude at maximum, whereas at minimum light most are fainter than 16th magnitude. Thus they are spectroscopically inaccessible to all but the largest telescopes. Conventional photometric techniques applied to these objects gave no information indicating their physical nature. The time-scale of the observations was so long that only the gross variations in brightness were detected. Much of the observing was undertaken by amateurs (Fig. 1), and from this work several of the basic features of this group have emerged.

Includes abstract. / Includes bibliographical references (leaves 95-99). / We use the HI velocity profiles of The HI Nearby Galaxy Survey (THINGS) galaxies to study the phase structure of the interstellar medium (ISM) and its relation to galaxy properties and morphology. To construct high signal-to-noise (S/N) profiles, we use a method analogous to the stacking method sometimes used in high redshift HI observations. We call these high S/N profiles 'super profiles'. By decomposing the super profiles into Gaussian components, we find broad and narrow components in all our analyzed galaxies.