Includes bibliographical references. / The Blesbokspruit is situated south-east of Johannesburg on South Africa's East Rand, famous for its prolific gold and uranium deposits. With the onset of mining in the region around 1930, the Blesbokspruit was altered from a non-perennial to a perennial stream and a wetland due to the construction of roads and embankments as well as the discharge of mine effluent. Various studies have highlighted the detrimental impact South African gold mines and their slimes dams have on the environment especially the release of metals into aquatic ecosystems. Unable to degrade naturally, metal pollution accumulates within aquatic ecosystems with the potential to move up the biologic chain. Water and sediment samples were taken from throughout the Blesbokspruit in an attempt to characterise the aqueous and sediment geochemistry of the stream. Filtered and acidified water samples were analysed for major cations and anions using Atomic Absorption Spectrophotometry and Ion Chroma topography respectively. The waters were found to have elevated levels of Cl⁻, SO₄²⁻, Na⁺ and Ca²⁺. These elevated salt concentrations are derived from the discharge of saline effluent into the Blesbokspruit by the Grootvlei Proprie ary Mines Ltd and are responsible for the increase in electrical conductivity downstream of the mine. The generally mildly alkaline nature of the Blesbokspruit is the result of the underlying dolomites of the Chuniespoort Group . ICP-MS was used to quantify aqueous and sediment metal enrichment relative to natural concentrations reported in the literature. The results of the ICP-MS indicate aqueous and sediment enrichment in chalcophile metals derived from the oxidation of sulphide minerals. U is also enrichment within the water and is derived from the ore minerals mined from the underlying Witwatersrand Supergroup. However, total metal concentrations are not indicative of bioavailable metal concentrations. For this reason it is important to determine how metals are partitioned within the sediment. A sequential extraction was performed on sediment samples from the Blesbokspruit as a method of determining metal partitioning and mobility under various possible environmental conditions. The sequential extraction extracted metals from the following five operationally defined fractions: the exchangeable fraction, sensitive to changes in ionic concentration; the pH- sensitive bound to carbonates fraction; the redox-sensitive bound to Fe and Mn oxides and bound to organic matter fractions and essentially inert and unavailable residual fraction. Metal extracts from the five fractions were analysed by ICP-MS. The results of the sequential extraction indicate that only minor proportions of metals are exchangeable, while a significant proportion of many of the metals analysed for were hosted in the residual phase and are therefore not bioavailable under most natural conditions. Varying amounts of the meta ls analysed were partitioned within the carbonates, Fe and Mn oxides and organic matter.

Bibliography: p. 118-132. / The examination and interpretation of surface features on pyrope, picroilmenite and chrome diopside from kimberlite gives valuable information on the history of these minerals both during and after kimberlite emplacement. Descriptions and scanning electron photo-micrographs of surface features observed on these mantle-derived minerals from thirty-four kimberlites having a global distribution are presented. Careful observations established the age relationships between the different surface features on each of the minerals studied, and enabled the compilation of a classification scheme of these surface features. This proposed classification scheme is based on the chronological sequence of formation of th.e surface morphological characteristics observed on these mantle-derived minerals and separates thoie which are formed during kimberlite emplacement (primary features), from those resulting ffom in-situ weathering and diagenetic etch processes (secondary features). The mineralogy, structure and mode of formation of kelyphite are discussed in relation to kimberlite emplacement. Pyrope encrusted by kelyphite is commonly found in most garnet peridotites, some eclogite xenoliths from kimberlite, as well as a few garnet-bearing serpentinites. Although the relative abundance of kelyphite-encrusted garnet varies from one kimberlite to another, kelyphite shells are most commonly developed on mauve garnets which are most probably derived from peridotite and are least commonly found on orange varieties from eclogites.

Generation of granitic magma predominantly occurs by melting through the breakdown of hydrous minerals. However, melting due to the influx of water has been recognised in anatectic amphibolite-facies composite grey gneisses, metagreywackes and low-pressure metapelites, and has consequently been proposed as an alternative mechanism for granite generation and crustal differentiation. Water-fluxed melting is recognised by voluminous melt production at relatively low temperature, where hydrous minerals are stable and anhydrous minerals are preferentially consumed during melting. Mineral equilibrium modelling to determine the P–T conditions, melt volumes, melting reactions and viable fluid sources reveal that water-present melting in all target lithologies is confined to the wet solidus and does not extend to temperatures higher than 700–710 ◦C. Melting at suprasolidus conditions does not involve the mechanical flow of a free water phase. Instead, the process is driven by diffusion of H2O along chemical potential gradients and is therefore more appropriately described as diffusively-fluxed melting. Diffusively-fluxed melting is not restricted to specific compositions or P–T conditions, although it is more efficient at lower pressure and in lithologies with a low hydrous mineral content. Melting reactions in all lithologies primarily consume quartz and feldspars to yield 5–6 mol.% melt for each mol.% of H2O added. aH2O remains constant at ∼0.70–0.77 during progressive melting as long as alkali feldspar is present. Once alkali feldspar is exhausted, plagioclase becomes the main reagent, producing more tonalitic melt compositions with gradually higher aH2O. Melting will initiate and proceed as long as a µH2O gradient exists between the fluid source and target lithology. Our calculations show that an ordinary magma, such as an I-type magma with typical, undersaturated H2O content, has a µH2O high enough to be a viable fluid source, allowing diffusively-fluxed melting to produce melt volumes and fertility comparable to that of dehydration melting. However, voluminous melt production requires a considerable volume of H2O, which necessitates a focussed fluid source such as a magma conduit or melt-bearing shear zone. Any other magmatic fluid source will undergo a similar amount of crystallisation as the melt fraction produced in the target rock, such that there will be no net melt production. Considering that shear-zone hosted magma conduits are relatively rare, diffusively-fluxed melting appears to only be viable in a small fraction of the anatectic orogenic crust. Therefore, whereas it may play a significant role in locally raising melt volumes and modifying magma chemistry through mingling and hybridisation, it does not appear to, of itself, be able to meaningfully contribute to crustal differentiation.

Includes abstract. / Includes bibliographical references. / A lot of work has been done on the upper Orange River but little attention has been given to the source of its suspended sediment load. The source of the suspended load is important because it can provide insight into soil erosion, a major problem in South Africa and globally. In this study the composition of the suspended sediment load of upper Orange River tributaries is related to bed rock geology, topography, soils and land use. Orange River catchment bed rock includes sedimentary rocks of the Karoo Supergroup and flood basalts of the Drakensberg Group each with distinct geochemical signatures.

Bibliography: leaves 112-120. / Metamorphism observed in the Vredefort structure increases in grade on a regional scale from lower greenschist facies in the outer perimeter of the Potchefstroom syncline, throughupper greenschist and amphibolite grades within rocks of the Witwatersrand Supergroupand Dominion Group in the collar of the Vredefort dome, to granulite grades in the core ofthe dome. In the collar, two metamorphic assemblages have been identified, one representing a peak metamorphic event pre-impact in age and constrained to temperatures >420-620°C at 3.0-4.0 kbar, and a younger retrograde event, constrained to 500-540 ± 15°C, which overprints the latter assemblage and overgrows pseudotachylite in a small number of samples.

Bibliography: pages 180-197. / The Cainozoic Bredasdorp sediments along the south coast of the Cape Province, South Africa, have come under investigation through a few intermittent studies undertaken since the early part of the century. In this presentation the literature is reviewed and a stratigraphic subdivision, based on both lithostratigraphic and biostratigraphic principles, is applied to the sediments of the Bredasdorp Group. The Bredasdorp deposits can be classified, according to origin, as shallow marine and aeolian. The marine deposits are subdivided into the Pliocene De Hoopvlei and the Middle to Late Pleistocene Klein Brak Formations. The Late Pliocene to Early Pleistocene Wankoe Formation, the Late Pleistocene Waenhuiskrans Formation and the Holocene Strandveld Formation constitute the coastal aeolian deposits. These marine and marine-related (aeolian) formations, characterised by calcareous elastics have been grouped together in a newly defined Bredasdorp Group. In order to construct a depositional model for the Bredasdorp Group, various facies have been identified on the basis of geometry, lithology, fossil contents, palaeocurrent data, biogenic and sedimentary structures. The facies are related to environments constituting a shoreline setting with offshore, transitional, shoreface, foreshore and backshore zones, followed by back-barrier lagoons, estuaries, backshore dunes and coastal dunefields associated with transgressive/regressive shorelines. The various deposits, as well as distinctive geomorphological features, are correlated with the relative sea-level movements throughout the Cainozoic, which have shaped the southern Cape coastal plain. Sea-level curves for Southern Africa, drawn by several authors are compared. A relative sea-level curve is constructed for the south coast of South Africa. Several Early Cainozoic transgression/ regression cycles are recognised at places along the South African coast. The earliest cycle started in the Palaeocene and was followed by a less pronounced transgression/regression cycle occurring in the Late Eocene to Early Oligocene. Remnants of surfaces related to these cycles are recognised in the Southern Cape Province, but these are interpreted as products of subaerial processes. The next cycle, reaching a transgressive maximum of c. 180m, started in the Miocene and terminated in the Early Pliocene. Again, no evidence of marine deposits is preserved on this marine-planed surface. The Early Pliocene transgression reached a maximum present-day elevation of c. 120m. Marine planation of the coastal platform took place during the transgression, whereas the De Hoopvlei Formation situated below 120m, was deposited during the subsequent Late Pliocene regression. The Wankoe Formation was deposited during the same regression as backshore dunes and coastal dunefields. During the Quaternary transgression/regression cycles, of which at least three are indicated, the transgression reaching a maximum of about 50m, in places eroded part of the Neogene Bredasdorp Group. The Klein Brak Formation (preserved below 20m) was deposited during Middle to Late Pleistocene regressions. The Waenhuiskrans Formation, which is extensively developed on the present-day continental shelf, was deposited during this regression with sea-level receding to about -130m below present sea-level. The aeolian Strandveld Formation, which is still being deposited, originated from the Flandrian transgressive maximum at the start of the Holocene.

Magmatism that occurs within the Central Volcanic Zone (CVZ) of the Andes is heavily influenced by processes of differentiation and assimilation that occur within the extremely thick continental crust. These processes, that result in crustal contamination of mantle derived magmas, are not uniform across the region and are determined locally. Data from Azufre is combined with that of previously studied volcanoes of the CVZ, and specifically the Altiplano-Puna Volcanic Complex (APVC) in northern Chile, to further understand such processes that have occurred in this region. Factors that need to be taken into consideration when examining volcanoes of the APVC are the aforementioned thickness of the continental crust of the Andes, which differentiates the Andes from other volcanic arcs globally, and the presence of a zone of partial melt, the Altiplano-Puna Magma Body (APMB), below the APVC as they could both have had a major effect on erupting lavas. Elemental concentration and isotope data are used here to examine the role that the APMB has had on the evolution of Azufre. This data set is also compared to those of previously studied volcanoes of the APVC to further provide understanding as to the interaction between the APMB and the APVC. Major element data reveals well defined sub-alkaline trends, with samples predominantly being andesitic in composition. 87Sr/86Sr ratios of 0.7067 - 0.7075 and δ 18O values of 7.85 - 9.72 are high in comparison to those of other Central Andean lavas, however in comparison to other erupted lavas of the APVC they are low. A transition from the edges of the APVC towards the centre show an increase in 87Sr/86Sr and δ 18O. Conversely 143Nd/144Nd ratios decrease towards the centre, with Azufre displaying ratios of 0.5123 - 0.5124. These changing ratios from the edge to centre of the APVC correspond with changing thickness of the APMB below the APVC and are therefore viewed to represent the amount of interaction between parental magmas of the APVC with the partial melt of the APMB, primarily with increasing 87Sr/86Sr indicating increasing amounts of interaction. δ 18O data indicates that there are multiple sources of O enrichment, evidenced by two opposite data arrays, with one array indicating assimilation by magmas of high δ 18O crust and the second array indicating assimilation of low δ 18O crust. Modelling of δ 18O suggests that the magmas which have assimilated low δ 18O crustal material have interacted with between -2.01 and 3.49 . Low LREE/HREE ratios and a negative Eu anomaly support the idea that assimilation of crustal material occurred at shallow crustal levels by indicating shallow plagioclase fractionation and evolution in a garnet free environment, most likely within the APMB.

The evolution of the Southern Outeniqua Basin has been analysed within a tectono-stratigraphic framework by the interpretation of 3,700km of 2D seismic data constrained by five exploration wells, drilled on the adjacent shelf. The mapping of different fault families into distinct structural orientations, NE-SW, NNE-SSW and NW-SE, was done and the basin sub-divided into four sectors (far-west, west, central and east) based on the geographical position of these faults. Sequence stratigraphic analysis of the basin lead to the identification of seven secondorder unconformities which have been further defined into five main seismo-stratigraphic units. The variation in major structural orientations suggest the evolution of the basin developed during three phases of extension, which have been related to the broader plate tectonic context during the Mesozoic break-up of Gondwana. Early Jurassic (170-150Ma) rifting between West and East Gondwana is proposed to have created weaknesses in the crust along the Southern African margin in a NNE-SSW orientation leading to the development of small rift basins and the deposition of syn-rift sediments. A series of intra-continental strike slip faults, including the proto-Agulhas Falklands Fracture Zone, are proposed to have developed along the southern margin during the incipient formation of the Weddell Sea (147Ma) between Antarctica and West Gondwana. Later mid-Jurassic (150-130Ma) extension along the North-Falkland rift zone lead to the oblique extension along the southern margin and the development of normal faults orientated NE-SW. Continued rifting (134Ma) in the proto-South Atlantic lead to translation along the southern African margin and the coalescence of the individual segments of the intra-continental strike slip faults and ultimately the formation of the Agulhas Falklands Fracture Zone. This phase of wrench faulting (134- 130Ma) is characterised by the strike-slip deformation of syn-tectonic sediments defined by second-order unconformities BR and B - late Valanginian - Hauterivian (134-130Ma). The Diaz Marginal Ridge (DMR), which bounds the Southern Outeniqua Basin and AFFZ to the south and north respectively, shows a marked variation is seismic character along strike. In the eastern and western sectors of the basin the DMR is made of syn-rift sediments. The discontinuous and non-parallel seismic character of the DMR in the central sector of the basin is interpreted to be continental basement. Significant bevelling of the top of the DMR in this sector, along with on-lapping reflector geometries of horizon B, Hauterivian (130Ma), is proposed to signify that the continental block existed prior to the formation of the Southern Outeniqua Basin and remained a basement promontory until the development of horizon G - Tertiary (66Ma). The Southern Outeniqua Basin is therefore proposed to have formed through four periods of tectonism: 1) early Jurassic (170-150Ma) orthogonal rifting of West from East Gondwana; 2) late Jurassic to early Cretaceous (150-130Ma) oblique extension and translation along intra-continental strike-slip faults; 3) early-late Cretaceous deformation associated with transverse motion along the AFFZ and development of an active transform margin; 4) mid-late Cretaceous thermal uplift of the Southern Outeniqua Basin associated with migration of the Natal Valley spreading centre along the AFFZ. Thermal subsidence and true passive transform margin conditions developed in the Southern Outeniqua Basin from early Tertiary to present day.

Includes bibliographical references. / Images of the electrical conductivity of the subsurface below the Barberton Greenstone Belt (BGB) reveal two electrically conductive zones that extend to a depth of at least 5km. These high conductivity zones correlate well with regional shear zones geologically mapped at surface. The serpentinised Stolzburg Layered Ultramatic Body (SLUB) occurs along one of these shear zones, the Saddleback-Inyoka Shear Zone (SISZ). In an attempt to identify the conductive rocks or phases responsible for the observed electrical conductivity anomalies, impedance spectroscopy measurements on samples collected from regional rock types and shear zone related lithologies were undertaken, and show that the rocks from the BGB are generally very resistive. Sulphide rich greenschists and greywackes collected at depth from the northern extension of the SISZ have elevated conductivities, but are not comparable to the anomalies observed in inversions of the magnetotelluric data. Magnetite veins from the SLUB were the only materials that had electrical resistivities low enough to be comparable with the conductivity anomalies. In order to better understand the structural evolution of the western extent of the greenstone belt, a field investigation was undertaken in the SLUB.

The primary focus of this dissertation is to develop a predictive rock physics theory that establishes relations between rock properties and the observed seismic and to present the results of different seismic characterization techniques to interpret a tight gas sand reservoir off the south coast of South Africa using as input rock physics analysis and inverted seismic outcomes. To perform the aims and goals of this study a workflow that involves the execution of three main processes was implemented: (1) rock physics modelling, (2) a simultaneous seismic inversion, and (3) seismic reservoir characterization techniques. First, a rock physics model was generated as a bridge between the seismic observables (density, Vp and Vs) and reservoir parameters such as fluid content, porosity and mineralogy. In situ and perturbational log - derived forward modelling was performed. Both in situ and perturbational forward modelling were used to generate synthetic seismic gathers, which were used to study the AVA attribute responses. Overall, the effect of fluid fill on this tight gas sand seismically is modest compared with the effect of porosity changes. Second, there follows a detailed description of a workflow implemented to simultaneously invert P and S pre - stack seismic data. The derived elastic properties (acoustic impedance, Vp/Vs and density) were then used in combination with the rock physics analysis to characterize seismically the reservoir. The predicted acoustic impedance and Vp/Vs volumes show a good tie with the log data. However, the density outcome was of limited quality compared with the two mentioned above. Finally, using outcomes from rock physic s analysis and/or inverted data, four seismic techniques to characterize the reservoir were conducted. The techniques involved are: (1) AVO cross - plotting to generate a good facies property based on AVO attributes (intercept - gradient) and rock physics in the area of study , (2) rock physics templates (RPTs) to compute discrete rock property volumes (litho - Sw, litho - porosity) using a collection of curves that cover all possible "what if" lithology - fluid content - porosity scenarios for the reservoir and the inverted data, (3) a lithological classification to calculate litho - facies probability volumes based on a litho - facies classification using petrophysical cut - off s , multivariate probability functions (PDFs) and inverted data, and (4) an extended elastic impedance (EEI) inversion to derive rock property volumes (Vclay, porosity) based on AVO attributes (intercept, gradient). Despite differences in the input and theory behind each technique, all outcomes share parallels in the distribution of good and poor facies or reservoir and non - reservoir zones.