An outstanding problem in the field of nuclear fusion research is the precise mechanism by which a hot, magnetically inhomogeneous plasma is heated when illuminated by a constant beam of small amplitude radio waves matched in frequency to harmonics of the ion Larmor frequency. An accurate model must include microscopic dynamics and inevitably a kinetic theory is required. Highly energetic ions (> 1MeV) born from fusion reactions or powered by gyroresonance have large Larmor radii (> 10cm) which are comparable in size to the wavelength of the incident radiation. In particular we will focus on fast magnetosonic waves. Exact full wave equations describing a thermal plasma in a weakly inhomogeneous field are presently at least fourth order integro-differential equations (Sauter, 1992). These are computationally taxing. Recently a method was proposed to reduce the problem to a second order integro-differential equation at the expense of information related to the propagation of mode-converted waves (Holt, 1992). We present here a generalisation of the theory to allow for arbitrary velocity-dependent equilibria while at the same time retaining a general functional form for the field profile. We consider the specific case of a bi-Maxwellian plasma immersed in a linearly inhomogenous magnetic field. We find that thermal anisotropy produces resonance localisation when the perpendicular ion temperature is greater than that parallel to the ambient field. A study of the symmetry properties of the conductivity tensor reveals that the Onsager reciprocal relations are obeyed only for an isotropic plasma in an inhomogeneous field. This is a generalisation of the result obtained by Nambu (1995). We present a generalisation of the reduction method to include effects due to changes in wave amplitude. We find that we are able to include the odd-order field derivatives responsible for energy conservation. Our numerical study of fundamental Helium-3 gyroresonance in a majority Deuterium plasma reveals that we have > 99.9% energy conservation in all cases. We show that locally-uniform theory can be very inaccurate (≃ 70% in one case presented in our recent paper, Cairns et al., 1995) particularly for higher energy ions whose non-locality is more extreme. We present a representative sample of results for minority heating and mode conversion heating schemes. We report the appearance of an unexpected cut-off on the low field side of the minority gyroresonance which may have important consequences for antennae presently placed on the outside of Tokamaks.

This thesis studies the time dependent evolution of MHD waves in cold, fully compressible non-uniform plasmas. We used a 1-D box model (e.g., Southwood (1974)) to study wave mode coupling, and concentrate upon developing an understanding of the underlying physics that governs waves in the Earth's magnetosphere. We begin by discussing the form of the (often singular) governing eigenmodes of the system, and subsequently use these eigenmodes as a basis with which to construct the solution to a variety of initial value problems. We consider a detailed analysis of both the widths and the internal length scales developed by cavity mode driven held line resonances (FLRs), and compare our results to observations presented in the literature.

Astrophysical outflows are common in a large variety of objects with very different length-scales. They can be almost spherical, as in the case of the solar wind, or show a high degree of anisotropy as in pre-main sequence stars, star-forming regions or even extragalactic objects. This work is aimed at finding exact solutions of the axisymmetric wind equations in which all variables depend not only on the distance to the central object but on latitude as well. The geometry of the stream/field-lines is taken as helicoidal and this seems to be a good approximation in some examples of collimated flows. From a simple hydrodynamic approach, a straightforward technique based on separation of the variables yields the most general solution of the wind equations under the above assumptions. The way the different variables depend on latitude is controlled by three anisotropy parameters which are related to typical ratios at the base of the atmosphere. The density needs to be higher at the equator than at the pole for the outflow to be able to accelerate. In these circumstances, the radial velocity always increases from equator to pole. Contrary to Parker's model of the solar wind, the solution does not pass through any critical point, since no polytropic law is assumed. However, the general behaviour is similar, with a high acceleration at the base and the velocity rapidly attaining an almost constant asymptotic value. The heating rate that sustains this rapid increase is mostly concentrated near the surface of the central object. The inclusion of the magnetic field in the analysis introduces two critical points: the Alfvenic point and an extra X -type point filtering the solution that gives a vanishing pressure at infinity. If the density anisotropy is too low the wind is unable to accelerate to large asymptotic values. The dependence of the angular velocity of the roots of the fieldlines with latitude reproduces well the observed rotation profile of photospheric magnetic features. The mass loss rate can be substantially increased if the structure of the outflow is highly anisotropic. Some applications to the solar wind are also discussed. In particular, recent results from ULYSSES (pointing out that solar speed increases with latitude while the density decreases from equator to the pole) are in good agreement with the general behaviour of the solutions presented in this work.

It has been supposed since the middle of this century that it is the global magnetic field surrounding a quiescent prominence that provides the force to prevent its collapse due to the sun's gravitational field. Many theoretical models, assuming that the prominence plasma is supported in a dip in the magnetic field lines associated by the magnetic tension force, have since been put forward. The aim of this thesis is to propose further models of quiescent prominences to widen our understanding and knowledge of these remarkable features. A short overview over the magnetohydrodynamic equations used to describe solar prominences, or most of the solar phenomena for that matter, are discussed in chapter 2, and a short summary of prominence observations and attempts to model them is given in chapter 3. A brief description of the numerical code used in chapters 5 and 7 is given in chapter 4. Observations of Kim (1990) and Leroy (1985) have found that most large quiescent prominences are of inverse polarity type for which the magnetic field passes through the prominence in the opposite direction to that expected from the photospheric magnetic field. Many theoretical models have been proposed, but failed. Hence, in chapter 5 we investigate first – without the inclusion of a prominence sheet – when an inverse polarity magnetic field must have the correct topology for an inverse polarity configuration before the formation of the prominence itself. Only very recently, the first basic successful model of an I-type polarity prominence was proposed by Low (1993). In chapter 6 we examine this model and investigate current sheets more complicated and realistic than the one used by Low. These analytical models deal with the force-free solution, which is matched onto an external, unsheared, potential coronal magnetic field. These solutions are mathematically interesting and allow an investigation of different profiles of the current intensity of the magnetic field vector and of the mass density in the sheet. The prominence properties predicted by these models have been examined and have been found to match the observational values. The mathematics of current sheets in general is also briefly discussed. Chapter 7 deals with numerical solutions of inverse polarity prominences embedded in a force-free magnetic flux tube, matched onto an unsheared potential coronal field. Unfortunately the solutions gained are quite sensitive to the boundary conditions imposed on them through the numerical box, showing a loss of convergence and a tendency for the solution to blow up. Finally, a short summary as well as possible future work is given in chapter 8.

Oscillatory fluid flows arise naturally in many systems. Whether or not these systems are stable is an important question and external periodic forcing of the flow may result in rich and complicated behaviours. Here three distinct oscillatory fluid flows are examined in detail, with the stability of each being established using a range of analytical and computational methods. The first system comprises standing surface capillary-gravity waves in second-harmonic resonance subject to Faraday excitation. Using the perturbation technique of multiple scales, the amplitude equations for the system are derived. At exact resonance, and with the absence of damping, the only fixed point of the equations is found to be the origin. A computational approach reveals that the amplitudes of the two waves remain either bounded or grow to infinity depending on initial data. With the introduction of detuning and damping into the system families of fixed points now exist and some special cases are considered. The second class of flows are unbounded time-periodic flows with fixed ellipsoidal stream surfaces, and having spatially uniform but time-periodic strain rates. Using a recently developed method based on theoretical study of the Schrodinger equation with quasi-periodic potential, a computational approach is adopted which determines the stability of the flow to three-dimensional plane wave disturbances. Results for the growth rate and winding number of the disturbance clearly reveal the regions of instability. It is found that almost all these flows are highly unstable. The third class is another set of three-dimensional time-periodic flows with spatially uniform strain rates. These flows are non-axisymmetric and have sinusoidally-fluctuating rates of strain directed along the fixed coordinate axes. The same computational method is employed and it is found that instability increases along with the non-axisymmetric nature of the flow.

Small brightenings called x-ray bright points (Golub et al, 1974) occur in the solar corona. They are observed with the soft x-ray telescope on Skylab to be approximately 22 Mm in diameter with a brighter inner core of width 4-7 Mm although with the Normal Incidence X-ray Telescope their dimensions are observed to be typically 6 Mm x 9 Mm. By comparison with magnetograms of the photosphere it has been noticed recently that there is a high correlation between the occurrence of x-ray bright points and the mutual reduction of flux between two opposite polarity magnetic fragments. These fragments are originally unconnected magnetically, but move towards each other and simultaneously lose equal amounts of flux (cancel): they are called cancelling magnetic features (Martin et al, 1984). The observations relating to these features were reviewed by Priest et al. (1994) who suggested that they naturally evolve through three phases: the pre-interaction, interaction and cancellation phases. From this evidence qualitative pictures of the magnetic field structure for an x-ray bright point and associated cancelling magnetic feature were established. The aim of this thesis has been to build on the ideas of Priest et al. (1994) to produce a detailed theoretical model of an x-ray bright point and a cancelling magnetic feature. The magnetic field structures are estimated, and the position and lifetime of the bright point are calculated, as is the total amount of energy released during the bright point. This work is also extended to study more complex cancelling configurations representing the main basic types of cancelling magnetic feature. The results of these models determine the factors that affect the lifetime and position of a bright point and indicate which types of cancelling magnetic features are most likely to produce bright points that are long-lived, lie directly above the cancellation site and occur simultaneously with the cancellation phase. The complex structure of a bright point cannot be explained from the above two-dimensional models: thus two recently observed bright points were studied to see if the above model could be extended into three dimensions to explain the structure seen in soft x-ray images. The available observational data was used and leads to reasonable explanations for the complex shapes of both bright points. Finally, a more realistic model for the overlying field was set up involving a model of the field above a supergranule cell field with fragments of finite width. The interaction of an ephemeral region within this field was then studied and led to five different scenarios. The results obtained reaffirmed those found in the previous simpler models and suggest where bright points may appear in a cell relative to the cancelling magnetic feature and for how long the bright points might last. Predictions for the lifetimes of cancelling magnetic features are also made, indicating when the cancelling magnetic feature occurs relative to the bright point.

It has become clear that the closed-field regions of the solar atmosphere are not static (as was once thought) but that many types of steady and unsteady flows and other dynamical, processes such as flares are continually occurring, in them. This thesis investigates some theoretical aspects of these dynamical phenomena. Steady, one-dimensional flow along a coronal loop is investigated first of all. Such a flow may be driven by a pressure difference between the foot points, and a wide range of shocked and unshocked flows are found. The presence of steady flows removes the symmetry present in most static loop models, and these models are shown to form only one class of a much wider range of dynamic solutions to the equations of motion. Thermal non-equilibrium in hot coronal loops occurs if the pressure in a loop becomes too big. The non-linear evolution of this non-equilibrium state is followed, and the loop is found to cool from of order 10[super]6 K to below 10[super]5 K in a few hours. An upflow is driven, and non-equilibrium is suggested as a means of formation of either cool loop cores or prominences. Thermal non-equilibrium is also discussed as a possible mechanism for the simple-loop flare. It is suggested that a cool equilibrium at a temperature of a few times 10[super]4 K can flare to over. 10[super]7 K if the mechanical heating in the cool loop becomes too large. The evolution is followed and the loop is found to flare to over 10[super]7 K in approximately 5 minutes. Magnetohydrodynamic shock waves have long been regarded as a potentially efficient heating mechanism. Their behaviour is re-examined here, and it is found that certain types of shock can release very large amounts of energy. These results are then applied to the heating of "post"-flare loops, for which temperatures of 10[super]7 K at the loop summit may be obtained. Finally, some solutions of the magnetostatic equation are discussed, and it is pointed out that if the gas pressure is too big then magnetostatic equilibrium will break down. It is suggested that the subsequent evolution may give rise to a surge or other mass ejection.

Ever since the first recorded observation of a solar flare in September 1859, it has been a key question - for physics as a whole and for astrophsics in particular - to ask what mechanism lies behind the sudden, violent release of energy from the sun. It has become increasingly apparent that the complex structure of the solar magnetic field lies at the heart of the answer. The process of magnetic reconnection has, over the years, become the accepted explanation by which magnetic energy can be released on both large and small scales in astrophysical and laboratory plasmas. The results of reconnection can be seen, for instance, in star formation, solar flares and the earth's aurorae; indeed the 1859 flare was followed by exceptional auroral activity. The mechanism of magnetic reconnection was first postulated by Giovanelli (1947) as a way of releasing the magnetic energy stored in the Sun. He, and later Dungey (1953), realised that the behaviour of the plasma in the vicinity of a magnetic neutral or null point, where the field disappears, is quite different from other regions of space. In this thesis the nature of magnetic neutral points and their role in the process of reconnection is investigated. Firstly, a general classification of magnetic neutral points is presented. The chapter includes equilibrium and steady-state solutions for two-dimensional magnetic neutral points. The differences in the field behaviour close to each type of neutral point are explained and criteria for the existence of steady-state solutions and equilibria involving pressure balance are presented. In the last section, a self-similar solution for a collapsed X-point is explored. The X-point necessarily becomes cusp-like in nature if shearing is applied in the ignorable direction. Two reconnection models are considered. The first is an extension of the Priest-Lee model (1990). It incorporates large pressure gradients in the inflow corresponding to the Forbes-Priest Almost-Uniform Model. The investigation includes both analytical and numerical solutions and a study of the separatrix jet. In the numerical study, current spikes are found at the end of the current sheets and a much increased reconnection rate is found analytically in the extreme flux file-up limit. The second reconnection model presented is also based on the Priest-Lee configuration. A uniform field is imposed on the basic structure producing a cusp-point with a non-zero field strength as the neutral point is approached from above. This results in the removal of the singularity in the flow above the separatrix. A non-singular solution is found analytically for a double-cusp. A much larger reconnection rate is found and a numerical solution is presented.

A technique has been developed for studying surface waves on liquids. The measuring device employed Fourier analyzes the surface wave being studied. This property of the technique has been verified by three independent tests. The method developed has been applied to the study of the Rayleigh-Taylor instability of fluid surfaces. The results of this study are in good agreement with theory. / Science, Faculty of / Physics and Astronomy, Department of / Graduate

This study investigates the propagation of a surge wave in a power canal following load rejection or reduction. Dimensionless relationships are derived to predict (a) the initial wave height, (b) the variation of the wave height along the canal and (c) the maximum stage of water depth at the downstream end for straight prismatic canals of rectangular, triangular and trapezoidal cross-sections. The effects of various parameters, such as velocity and depth of initial flow, frictional coefficients, bed slope, cross-section of the canal, distance of wave propagation and initial wave height of the surge are studied. A computer program is developed for the calculations required. It is found that, as a positive surge propagates along the canal, the wave height decreases linearly with distance for a short canal, according to an exponential function for a long canal. An approximate logarithmic relationship is also found between the variation of wave height of a positive surge and canal cross-sectional parameters. The variation of water depth at the downstream end of the canal is not linear with respect to time. An almost linear relationship between the maximum water depth at the downstream end of the canal and the length of the canal is noted. The dimensionless relationships derived in this study may be used to establish design criteria for crest elevations of the banks and walls of power canals to avoid overtopping. / Applied Science, Faculty of / Civil Engineering, Department of / Graduate