In January 1991 a suite of seven seismic refraction lines was acquired across the transform continental margin off Ghana. Six lines were shot parallel to the margin, from the continental shelf down to the deep Guinea Basin, while the seventh line was normal to, and bisected, these. Maslov travel-time and amplitude forward modelling of the refraction line normal to the margin shows a rapid transition from continental to oceanic crust over a 10 km wide high velocity zone at the foot of the continental slope. The high crustal P-wave velocities (5.8 - 7.3 km s<SUP>-1</SUP>) in this zone raise the possibility that this region may be heavily intruded by basic igneous rocks. Interpretation of the seismic refraction lines in the oceanic domain has shown a highly heterogeneous crust both parallel and normal to the margin. Oceanic layer 2 has a two layer structure and shows considerable variation in velocity, from 3.3 - 6.6 km s<SUP>-1</SUP>. The velocities are generally lower, and layer 2 thicker, to the northeast compared to the southwest ends of the lines. Layer 3 shows more homogeneity and has velocities of 6.6 - 7.3 kms s<SUP>-1</SUP>. The oceanic crust is anomalously thin to at least 70 km from the ocean-continent transition, with a thickness of 3.5 - 5.5 km, and slightly thicker crust is systematically observed to the northeast. This thin crust is caused by an abnormally thin layer 3 (2.0 - 3.5 km), while layer 2 attains a more normal thickness of 1.5 - 3.0 km. Two-dimensional gravity modelling along the line normal to the margin has confirmed the seismic refraction results. The continental crustal structure from the gravity modelling includes an abrupt increase in the depth to the Moho from 23 km below the continental slope and shelf edge to 37 km further inshore (northwest). This vertical boundary coincides with a southeast to northwest increase in the velocity of the upper crust from 5.00 km s<SUP>-1</SUP> to 5.80 km s<SUP>-1</SUP>. The variation in crustal thickness and structure is thought to represent the oceanward extension of the Akwapim fault zone which is mapped onshore Ghana with a southwest trend. This is a Pan-African suture zone which separates the West African craton to the west from Pre-Cambrian Dahomeyan units to the east.

No description available.

Transform continental margins form as a result of sea floor spreading and new ocean basin formation. Their evolution can be divided into four distinct phases (continent-continent shearing; continent-ocean shearing; ridge passing; and continent-ocean passive), the early phases involving active shearing, while in the final phase, oceanic and continental lithosphere are in passive contact across the transform boundary. Different thermal and mechanical factors influence the margin evolution in each of these phases. Transform continental margins have been much less well studied and are consequently less well understood than either rifted continental margins or oceanic fracture zones, related tectonic features, and the development of models to describe their evolution is in its infancy. The further development of evolutionary models is therefore essential to gain a more complete understanding of the margin forming processes involved. A 2D numerical thermomechanical model for transform continental margin evolution, which significantly improves upon previous models, is presented. The model can be used to calculate the evolution of the thermal structure and topography at any point along a margin of any length, with any spreading rate. The resultant uplift and subsidence across the margin, in response to thermal expansion and contractions, is calculated throughout the margin evolution assuming both local and regional isostasy for a variety of different lithospheric strengths (represented by T<SUB>e</SUB>, the effective elastic thickness of the lithosphere). The effects of sub-aerial erosion of the continental lithosphere are considered, and the continent-ocean boundary is modelled as being mechanically decoupled during the active shearing phases, and as either mechanically coupled or mechanically decoupled during the passive phase. The appropriate differential equations are solved numerically using the finite difference method. The model results have been compared to preliminary interpretations of geophysical data from a section of the Ivory Coast-Ghana margin and to previously published observations from other margins. Results are presented for five points along a margin of length L=900km for both slow (v=1cm/yr) and moderate (v=4cm/yr) spreading rates. The evolution of the topography at each of these points is investigated for a variety of different values of T<SUB>e</SUB>. Significant continental thermal uplift is calculated to occur as a result of the phases of continent-ocean shearing and ridge passing. Lateral variations in the magnitude and lateral extent of continental uplift along the margin length are predicted for each set of model parameters.

The aim of walkaway VSP experiments is to image the region beneath the receiver. The information in the data is obscured by propagation effects in the region above the receiver such as free surface multiples, internal reflections and mode conversions. This thesis presents a method of extracting the reflectivity response of the region beneath the receiver from walkaway VSP data, assuming the earth to be horizontally stratified. For marine experiments the source is an acoustic source in the water. Measurements of source signatures clearly show shot-to-shot variations. VSP processing, such as the separation of upgoing and downgoing waves, is based on the assumption of shot-to-shot repeatability. However, shot-to-shot variations are usually ignored during processing. I present a straightforward method for correcting for these variations. The source signature must be measured and the geometry of the measurement must be known. The recorded source wavelets are all shaped to a standard wavelet using filters in the frequency domain. The same filters are applied to the geophone data, thus removing the effect of the source variations. The method is demonstrated on real data. As a plane horizontally-layered earth is laterally invariant, a walkaway VSP can be viewed as an experiment with a single source and a horizontal array of geophones at depth. The data are processed in the horizontal wavenumber-frequency domain, in which plane wave components are separated. I present a method for recovering the reflectivity of the region beneath the receiver in this domain. The full wavelength at the receiver level can be computed for a plane-horizontally layered earth as a superposition of plane wave responses.

The thesis studies the propagation of guided-waves in crack-induced anisotropic media, and exploits the potential applications of guided-waves in production seismology by crosshole seismic surveys. Firstly, by extending Crampin's technique for calculating surface-waves dispersion, we present a procedure for determining the dispersion and amplitude/depth distribution of guided-waves in multilayered anisotropic media. By using the developed techniques, we study the propagation behaviour of guided-waves in crack-induced anisotropic waveguides. Our studies have shown that: (a) almost any interface, or combination of interfaces, will support the propagation of guided-waves in crosshole seismic surveys, if signals at appropriate frequencies, amplitudes, and polarizations can be generated in one well and recorded at appropriate levels in neighbouring wells; (b) in crack-induced anisotropic waveguides, the distinct families of Rayleigh and Love modes of guided-waves in isotropic waveguides combine into a single family of <i>Generalized</i> modes with 3-D elliptical particle motions; (c) guided-waves are very sensitive to the internal properties and geometry of waveguide, and such guided-wave characteristics vary substantially for different crack orientations, crack densities, crack saturations, and crack aspect ratios. Guided-waves have been clearly identified from a shallow crosshole survey at the Conoco Borehole Test Facility (CBTF) and, in many circumstances, have dominant energy in the seismograms. We have demonstrated that guided-waves in crosshole surveys have two potential applications: continuity tests of plane-layered reservoirs; and monitoring Enhanced Oil Recovery (EOR) operations for thin-layered sedimentary reservoirs. Finally, we have modelled the propagation of guided-waves in two particular waveguides: a crack-controlled channel which may be formed in hydraulic fracturing in geothermal reservoirs; and an active fault zone. The study of such guided-waves is useful in understanding the fracture structure and distribution in geothermal reservoirs, and in delimiting a fault zone or monitoring the stress and fracture development inside an active fault zone.

Palaeomagnetic studies have been carried out in three key regions of the eastern Mediterranean part of the Tethyan orogenic belt. Within the Troodos ophiolite of Cyprus, samples were collected from extrusives and sediments exposed along the Arakapas fault belt and around the periphery of the Limassol Forest block. These areas represent crust which formed in the Upper Cretaceous within a 'leaky' oceanic transform fault, and also a small fragment of crust generated at an 'Anti-Troodos' spreading axis. Significant clockwise intra-crustal rotations of small fault-bounded blocks have been identified within the inferred transform zone. These rotations are shown to be syncronous with crustal genesis and indicate a probable dextral sense of shear along the transform. In contrast, the 'Anti-Troodos' crust appears to have undergone an identical rotation to the non-transform tectonised Troodos crust to the north. Variations in declination upwards through the umber-radiolarite sequences overlying the extrusives demonstrate that 45<SUP>o</SUP> of the 90<SUP>o</SUP> palaeorotation of the Troodos microplate occurred within 15 Ma of formation of the Troodos crust. A wide-spread remagnetisation event has been identified in the Isparta angle region aof SWQ Turkey. Sites were located both within the Tauride carbonate platform massif's and in the overthrust units of the Antalya Complex. The latter consists of an assemblage of continental margin and ophiolitic rocks which originated in a strand of the Neotethys located to the south of the carbonate platforms. The secondary nature of the remanence at most sites is demonstrated by several negative fold tests. The magnetisation is carried by magnetite of presumed authigenic origin. The remagnetisation event was probably triggered by the migration of orogenic fluids ahead of the Antalya Complex during its emplacement onto the adjacent platforms in the Lower-Middle Miocene. Subsequent to remagnetisation, a large segment of the area underwent an anticlockwise rotation of 30<SUP>o</SUP>. This rotation was probably related to the Neotectonic bending of the Hellenic arc and the emplacement of the Lycian Nappes during the Middle-Upper Miocene. This research has identified block rotations in a variety of geological settings. These range from rotations active during oceanic crustal genesis to those associated with the late stages of continental collision. These rotations would have been impossible to identify by means of field structural studies alone. An awareness of such rotational deformation is essential if the geological evolution of complex areas is to be fully understood.

Attenuation of seismic energy is often seen as an unavoidable 'problem' in earthquake seismology. Observation of earthquakes at teleseismic distances is severely limited by the signal-to-noise ratio of the recording seismograph. However, the underlying mechanisms behind seismic attenuation are very different to those underlying other parameters and seismic attenuation can yield information on the structure of the Earth different to that obtained from, for example, seismic wave speed studies. In this study, 59 globally-distributed earthquakes and explosions, recorded at the four UK-type arrays in Eskdalemuir (Scotland), Gauribidanur (India), Warramunga (Australia), and Yellowknife (Canada), have been used to investigate the seismic attenuation along the <I>P</I>-wave propagation path in the frequency range 1-8 Hz. Two advantages of using array data, without which this study would have been unfeasible, are that the array seismographs record data over a wider frequency band (1-8 Hz) than many short-period instruments, and the method of 'beam-forming' seismic array data enhances the signal-to-noise ratio. The scatter of the data suggest that the upper mantle is inherently heterogeneous on scale-lengths of as little as 1-2 km. The difference between low-frequency and high-frequency models is attributed to scattering near-surface, and to strain rate effects in the lower mantle, however, the attenuation in the upper mantle at high frequencies shows an anomalous jump which is difficult to explain but may be due to a change in mantle behaviour for larger strain rates (increased frequency in cyclic loading).

No description available.

Shear-wave splitting is observed in a three-years dataset recorded at the Parkfield borehole digital seismic network, central California, and also at Cajon Pass with data recorded at a 2.5 km deep borehole, southern California. Shear-wave polarizations at five out of seven stations at Parkfield are aligned in the direction which is consistent with the direction of horizontal maximum compressive stress. In the fault zone, shear-wave polarizations are aligned approximately parallel to the fault. Temporal variation of time delays between two split shear-waves is detected before and after a M<SUB>L</SUB>=4 earthquake at Parkfield. The analysis of temporal variation based on the best quality data shows that the temporal change is significant at the 68% confidence level. The study of earthquake multiplets also shows similar variations. This means that there was probably a change in stresses, which modified the geometry of microcracks pervading the rockmass before and after this event, so that time delays between faster and slower shear-waves for the similar events travelling along nearly the same wave paths and occurred at different times display a distinct increase before the time of the M<SUB>L</SUB>=4 earthquake, and a decrease afterwards. Normalised time delays in the fault zone are found to be about twice as large as those in the intact rocks. This suggests that fluid-filled cracks and fractures within the fault zone are more extensive than those in the surrounding rocks, and that the alignment of fault-parallel shear-wave polarizations in the fault zone is attributed to the fault internal structures. Studying the causes of the temporal variation indicates that stress concentration before the M<SUB>L</SUB>=4 earthquake had probably increased the number of cracks or the radius of the cracks; after the release of strain energy after the earthquake, stress reduced to a lower level, accompanied by healing and flattening of the cracks. The change of crack density and aspect ratio associated with the M<SUB>L</SUB>=4 earthquake is interpreted as the cause of the observed temporal variation.

This study addresses two topics in environmental magnetism, i) the magnetic unmixing of lake sediments in terms of catchment samples and, ii) a connection between soil magnetic properties and annual rainfall. Soil and lake sediment cores from Mexico provided the material used. Palaeolimnology facilities the understanding of past environments by the analysis of lake sediments. Many palaeolimnological techniques are both time consuming and destructive. Analysis of pollen and diatoms require an expert to identify and count hundreds of pollen grains or diatoms, for each core horizon to be characterised. In contrast, magnetic remanence and susceptibility measurements can be obtained rapidly and without changing the physical properties of the sample that is being measured. The problem with magnetic measurements on lake sediment is the interpretation of the results. Previous workers have analysed trends in the magnetic properties of sediments. Linking high susceptibility to soil erosion, and low susceptibility to reduced sedimentation through, for example, the accumulation of organic material, low in magnetic content. In order to analyse environmental magnetic results I have developed a new magnetic unmixing technique. It uses non-negative least squares to unmix the magnetic properties of a target sample in terms of the magnetic properties of potential end-member samples. A stepwise technique is used to select appropriate end-members and estimate the errors involved. The newly developed magnetic unmixing procedure has been applied to the sediments of two lakes in Mexico, Lake Pátzcuaro, in the central Mexican highlands, and Lake Babicora in the northern Mexican highlands. The results for Lake Pátzcuaro indicate changes in the lake sediments’ magnetic properties that can be ascribed to climate changes and farming. The main climatic change is a reduction in lake level at approximately 11,000^{1 4}C yr B.P. Human arrival, at approximately 3,500^{1 4}C yr B.P. coincides with a dramatic increase in topsoil erosion. At Lake Babicora the results indicate that a reduction in rainfall led to a dramatic reduction in catchment material delivered to the core site. This change in the depositional regime occurred at approximately 17,000 ^{1 4}C yr B.P.