No description available.

The primary concern of this thesis is to develop inversion methods which may be applied to observations of seismic anisotropy. I achieve this through the implementation of a non-linear global optimization scheme known as a Genetic Algorithm. I apply this technique to the inversion of quasi-compressional velocities obtained from laboratory measurements and shear-wave splitting observations measured from an azimuthal VSP experiment and two near-offset VSPs. These VSP experiments were conducted at the Conoco Borehold Test Facility, Oklahoma which presents an ideal setting due to its geological and structural simplicity and the availability of a-priori information, for example from borehole televiewer images, core samples and logging information. I compare the information derived from these sources with the results obtained from the inversion of shear-wave birefringence measurements. Comparative studies of this type are important in establishing the relation between the observed seismic anisotropy and equivalent media systems due to such phenomena as aligned fractures or fine layering. The inversion results for the azimuthal VSP and near-offset VSPs are in good agreement with each other and also with the a-priori data. The azimuthal VSP inversion allowed the identification of a significant feature associated with shear-wave propagation in anisotropic media known as a shear-wave singularity. Another significant inversion result is the suggestion that the observed seismic anisotropy is due to a sub-vertical fracture system. This result is supported through the analysis of shear-wave observations obtained from the near-offset CSP experiments. I suggest that fracture dip may be determined using appropriate acquisition geometries such as an opposite azimuth VSP for which two VSPs are recorded with sources located diametrically opposite each other about the well. A significant correlation is observed to exist between the lithology at the Conoco Borehole Test Facility and the degree of shear-wave birefringence. Specifically it appears that the shear-wave birefringence is most pronounced for wave propagation through the sandstone formations which are known to be heavily fractured. Correlations of this type are important if any confidence is to be placed in the interpretation of shear-wave birefringence in terms of fracture systems.

No description available.

Initially, microearthquakes associated with an active continental region in Central Greece are studied. The distribution of epicentres is associated with the main E-W trending normal fault in the locality, and their scatter confirms a complicated velocity structure. Although shear-wave splitting is identified, it is difficult to interpret because the shear waves are frequently obscured by extensive <I>P</I>-wave coda. The incorporation of shear waves in the focal mechanism determination augments the available observations in the case of small local networks, but, it imposes difficulties because shear waves are sensitive to the structure along the raypath and in particular anisotropy. A procedure is developed in order to identify shear-wave splitting by testing the compatibility of the observed shear wave polarisations with those generated by a double-couple source and, if splitting exists, to remove its effect in order to reconstruct the polarisation of the first split shear wave at the source, and hence to determine valid focal mechanisms for single events. The procedure is applied to two datasets from intraplate regions, Arkansas-USA and NE Brazil, in order to exploit the simplicity of seismograms typical of such regions. The simple waveforms from the Arkansas sequence reveal that two well-distributed shear waves together with <I>P</I> waves can be sufficient to determine well-constrained focal mechanisms, but more may be required in order to discriminate between their interpretation as unsplit and split shear waves. However, the proximity of raypaths in this case of swarm-type activity offers the opportunity to extend the interpretation derived from one event to other events which may have too few observations to confirm anisotropy from the focal mechanism procedure. The results reveal strike-slip focal mechanisms with the P axis oriented NE-SW which is consistent with the regional stress field. The NE Brazil dataset, reveals shallow seismicity concentrated in three tight clusters. Shear-wave splitting is observed with the fast shear wave aligned approximately N-S. This agrees with previous observations at a single three-component station, but is inconsistent with conventional interpretation in terms of fluid filled cracks aligned with the known axis of maximum compressive stress; this would require an E-W alignment. It may be that the anisotropy is instead intrinsic to the rock fabric, possibly associated with known north-south Precambrian ductile shear zones. One well-constrained focal mechanism was obtained assuming the validity of the shear-wave splitting interpretation. The mechanism reveals an almost horizontal strike-slip fault plane, trending NW-SE, with a N-S oriented P axis, which contradicts other studies which reveal an E-W oriented P axis.

The central Indian Ocean exhibits a higher level of earthquake activity than is normally associated with a region distant from conventional plate boundaries. This anomalous seismicity has been the subject of previous studies which assigned large focal depths (as deep as 40km below the seafloor), despite crustal faulting and folding observed on seismic reflection profiles. I re-examine the distribution and source parameters of teleseismically recorded earthquakes and relate them to the tectonic setting of the Central Indian Ocean Basin. Previous studies have used long-period waveform modelling, which gives unreliable focal mechanisms for small earthquakes (<I>m<SUB>b</SUB></I><SUB> </SUB><5.5), and poor resolution of depth for shallow earthquakes (<100km). Zero-phase broad-band conversions from short period and broadband waveforms are used with a detailed velocity structure derived from seismic reflection data obtained by the <I>Marion Dufresne</I> in the Central Indian Basin, to assign the more accurate focal depths (±3km). Measurements of the relative amplitudes of <I>P</I> arrivals, and surface reflections from the seafloor and sea surface, are used to constrain fault plane solutions even when <I>P</I>-wave first motion polarities cannot be determined. Unlike in other oceanic intraplate settings, the seafloor reflection is small and unclear, due to high sediment deposition in the Bengal Fan. The earthquakes have either thrust or strike-slip mechanisms and their depths are in agreement with previous studies, with no depths shallower than 22km below the seafloor in the Central Indian Ocean Basin or Wharton Basin. Earthquakes in the Bay of Bengal are shallower, possibly due to loading by Fan sediments raising lithospheric stresses beyond failure. The directions of the compressional axes of the earthquakes are in agreement with those predicted by plate models. The seismicity shows a concentration between the Afanasy Nikitin Seamount and the Ninety East Ridge. Other earthquakes are evenly distributed between 10°N and 10°S, and 80°E and 95°E. The existence of an aseismic region west of 80°E suggests that there are low strain rates in the west Central Indian Ocean Basin, associated with the proximity of the proposed Euler pole of rotation between separate Indian and Australian plates.

No description available.

No description available.

The introduction contains a review of published work on the oceanic induction problem up to the end of November 1979 and concentrates on induction in the ocean by magneto variations, although a brief account is also given of dynamo processes. The aim of the research was to model the effects of mutual induction between oceanic electric currents and a finitely conducting mantle, since previous authors had only considered the simpler case of a perfectly conducting mantle. A method for calculating these effects is derived in Chapter 2 by considering induction by a vertical magnetic dipole at the earth's surface leading to a generalised form of the Hobbs-Price mutual induction kernel. A simple uniform earth and a more realistic seven-layered model represent the mantle while considering induction in a hemispherical ocean, in Chapters 3 and 4 respectively, where details of the calculation of the electromagnetic responses are given. Alterations and improvements were made to an existing computer program so that Bunks' profile could be used to model induction in a thin sheet in the shape of the oceans and contour plots of results obtained with Banks' model and with the perfect conductor are compared in Chapter 5. Using a spherical harmonic representation of the Sq inducing field, a qualitative comparison is made between the equivalent current systems of the modelled results with those computed from observatory measurements at different instants of Universal Time. The use of Banks' profile did not give a better comparison with the observations than the perfectly conducting mantle. The application of spherical harmonic analysis to the complete solutions showed that the electromagnetic responses to each of the principal Sq harmonica were greater with the finitely conducting mantle than with the perfect conductor, while both were greater than the observed responses. When the analysis was restricted to using only the values at the sites of geomagnetic observatories the coefficients of the vertical magnetic field were found to be different from before, while the ratios of internal to external parts were found to be smaller. These discrepancies are due to aliaising by harmonics absent from the analysis and they suggest that the determination of the responses based on spherical harmonic analysis of observatory measurements must be suspect, which will influence any global geoelectric profile determined from these ratios.

This study investigates the processes involved in the formation and evolution of faulting in high porosity sandstone using laboratory triaxial compression testing. Faults in highly porous sandstone significantly affect the porosity and permeability of the rock, and typically occur as anastamosing compound bands of damage. Previously only the individual unit of these deformation band structures had been re-produced in the laboratory, possibly due to limitations on sample size. Now by deforming large specimens, I have not only produced zones of deformation bands, but also observed their hierarchical development as a function of strain - for the first time. A series of dry tests were carried out on initially intact 100mm diameter cores of Locharbriggs sandstone, at a constant confining pressure of 34.5MPa, a constant axial strain rate of 5x10E-6/s and increasing amounts of axial strain. Samples were driven over their failure curves and then subjected to differing amounts of post failure sliding. A second series of tests were carried out at increasing confining pressure (in the range of 13.8MPa to 55.2MPa) and constant amounts of axial strain. The response of the rock to the loading conditions was continuously monitored throughout the tests by recording the axial load, the axial strain and volumetric strain. Deformation structures and gouge material produced were studied by hand specimen, thin section and laser particle size analyses. The samples exhibited essentially brittle behaviour with small amounts of strain hardening and softening occurring near peak stress. Dynamic failure was accompanied by a measurable stress drop, which systematically decreased in magnitude as a function of confining pressure. Both compaction and dilatancy are observed in the volumetric strain curves. The amount of dilatancy decreased systematically with increasing confining pressure, until at high confining pressures, compactive behaviour dominated.

No description available.