The primary objective of this study was to establish sedimentary models for peat formation in the southern part of the Klip River coalfield during Ecca (Permian) times and to assess palaeoenvironmental controls on coal seam behaviour and distribution. In order to achieve this approximately 2 400 borehole logs and 25 field sections were collected. The coal-bearing Vryheid Formation records early to late Permian fluvio-deltaic sedimentation within the northeastern main Karoo basin. Three informal lithostratigraphic subdivisions, based on the investigations of Blignaut and Furter (1940, 1952), are proposed: the Lower zone, Coal zone and Upper zone. An examination of the structural framework and history of the northeastern Karoo basin reveals that the southern and western boundaries of the Klip River coalfield are defined by zones of rapid basement subsidence : the Tugela and Oannhauser Troughs respectively. There is some doubt as to the locality of the source area to the rivers emptying into the Ecca sea. Ryan (1967) postulated the "Eastern Highlands" situated off the present southeast African coast, but it is contended that the Swaziland area, situated no more than 200 to 300 km to the northeast of the Klip River coalfield, constituted a more plausible source area. The Lower zone represents sedimentation along a westerly to southeasterly prograding coastline dominated by high-constructive lobate or braid deltas, but also showing significant influence by wave processes. The Coal zone, which varies in thickness from 35 to 60 m, represents a major phase of coastal progradation and braided-river deposition on extensive alluvial plains. Significant coal seams formed only during periods of fluvial inactivity, the duration of which was dependent on source-area processes. Coal seam geometry and behaviour in the Klip River coalfield were not influenced by the depositional environments of associated clastic sediments. The following factors were found to have of profound influence in determining the extent, distribution and rate of peat accumulation: 1. Platform stability and temporal and spatial variations therein. 2. The absence or presence of penecontemporaneous clastic sedimentation. 3. Duration of periods of peat formation. 4. Lithology and topographic expression of clastic sediments underlying peat-forming swamps. The peat-forming phase of the Vryheid Formation was terminated by an extensive transgression brought about by an eustatic rise in basin water-level and/or an increased rate of platform subsidence. / Thesis (Ph.D.)-University of Natal, 1988.

The northern KwaZulu-Natal coast is backed by a continuous aeolian dune cordon that rises in places, to a height of more than 100 metres and a width of 2 kilometres. This MSc thesis documents the geomorphology of the area, as well as the mineralogical, geochemical and textural variation of nine boreholes within a small part of the coastal dune cordon between Lake Nhlabane and Cape St.-Lucia. The results provide useful constraints on the identification of individual beach and aeolian dune systems, their age relationships and spatial distribution. Aeolian dunes within the coastal dune cordon were studied using aerial photographs and grouped into five dune classes that reflect their relative age. These comprise 1) a system of highly weathered dunes inland of the present coastal dune cordon, that are thought to represent older dune cordons; 2) a system of weathered and reworked dunes located on the most inland portion of the coastal dune cordon; 3) a less altered, large field of linear parallel dunes located in the northern part of the study area; 4) a system of large scale parabolic dunes; and 5) a system of coastal, relatively unweathered small parabolic dunes. Mineralogy, geochemistry, texture and SEM analysis of borehole samples revealed a complex internal structure within the present coastal dune cordon. In the most inland part of the dune cordon, a basal light grey unit (Unit K) presents similar characteristics to the Kosi Bay Formation. This is overlain by Unit A, comprising beach and dune systems, characterised by a very high heavy mineral content. Unit A also forms the basal unit of the central and coastal portions of the dune cordon. Unit B contains a mixture of reworked sediments from Unit A and younger sediments. Aeolian Units D and E form the upper part of the dune cordon. Units D and E were derived from beach - foredune systems and contain a high carbonate bioclast content. All units are interpreted to be derived from immature sediment from the Tugela River and mature sediment from the continental shelf. In the southern part of the study area, an additional unit (Unit C) with unique characteristics has been interpreted as an aeolian deposit reworked from local fluvial sediments. The units identified from their sedimentological characteristics can be directly correlated to the regional dune classes identified from the geomorphology. Luminescence dating of two calcareous dunes was undertaken, revealing that only the sediment of the small coastal parabolic dunes (Dune Class 5, Unit E2) is of Holocene age. The deposition of the large field of linear dunes (Dune Class 3, Unit D2) took place between 15 000 and 11 000 BP, during the marine transgression following the last glaciation. Luminescence dating also indicated that both dunes were subject to at least one major reworking event. A study on the weathering characteristics of the dunes can be used to attribute a relative age to the nine sedimentological units. With the help of sea level curves and the two luminescence dates, the nine units were attributed an approximate absolute age and regrouped into four sediment packages thought to broadly represent four interglacial periods. The three younger packages are attributed to the penultimate interglacial (lower part of Unit A), last interglacial (upper part of Unit A, Units B and C) and "Holocene" interglacial (Units D and E). Hence the northern KwaZulu-Natal coastal dune cordon under study represents a complex stacking of three generations of coastal dune cordons, and appears to be constituted of sediments with age ranging from at least two hundred thousand years ago to present. The oldest sediment package (Unit K), interpreted as the Kosi Bay Formation, and the older dune cordons (Dune Class I) must be older than 200 000 years, which is older than considered by previous studies. The "Holocene" dune cordon (Units D and E) is interpreted as the Sibayi Formation. / Thesis (M.Sc.)-University of Natal, Durban, 1999.

The Jurassic Rooi Rand dolerite dyke swarm was emplaced sub-parallel to the Lebombo Monocline during the initial stages of Gondwana breakup. The dykes extend northwards from the southern Lebombo region in northern KwaZulu-Natal, into central Swaziland, spanning a distance of approximately 200 kilometres with a width between 10 and 22 kilometres. Detailed mapping of a 600m-Iong section on the Pongolo River, established at least eleven phases of intrusion. Each dyke age was systematically sampled and analysed for whole-rock major, trace and rare earth element composition, as well as mineral chemistry. Selected samples were analysed for stable isotopes. In addition to notable intra-dyke chemical variations, there is also a high degree of inter-dyke mineralogical and geochemical variation, each dyke age bearing distinct geochemical characteristics. The apparent geochemical trend is not one of simple fractionation with time. Dyke chemistries are closely linked to magma genesis and magma volumes with time. Evolution of the magmas may be described in terms of varying degrees of partial melting and fractional crystallization, with a small degree of crustal contamination. Major, trace and rare earth element data indicate a lithospheric mantle source for the majority of dyke phases, and an asthenospheric source for only two of the eleven ages. Contrary to this, isotopic data (oxygen and radiogenic) indicate an enriched asthenospheric source for all the dolerites. This suggests that all ages may have originally been derived from the asthenosphere, with the majority of ages being intruded into the lithospheric mantle to later undergo partial melting and fractional crystallization, with some contamination. Previous studies assumed an asthenospheric source with depleted MORB-like rare earth element profiles to be representative for the majority of Rooi Rand dolerites. The Rooi Rand dolerites appear to display a geochemical link with the southern Sabie River Basalt Formation, as well as the Lebombo rhyolites. Magmatic evolution of the dykes was intimately linked to the initial rifting processes of lithospheric stretching and asthenospheric upwelling, which in this case concluded in a classic failed rift situation. / Thesis (M.Sc.)-University of Natal, 1996.

This study proposes that the submarine canyons of the northern Kwazulu-Natal continental margin formed contemporaneously with hinterland uplift, rapid sediment supply and shelf margin progradation during the forced regression of upper Miocene times. These forced regressive systems tract deposits volumetrically dominate the shelf sediments, and comprise part of an incompletely preserved sequence, amongst which six other partially preserved sequences occur. The oldest unit of the shelf corresponds to forced regression systems tract deposits of Late Cretaceous age (seismic unit A), into which a prominent erosive surface, recognized as a sequence boundary, has incised. Fossil submarine canyons are formed within this surface, and underlie at least one large shelf-indenting canyon in the upper continental slope. Smaller shelf indenting canyons exhibit similar morphological arrangements. Late Pliocene deposits are separated from Late Cretaceous lowstand deposits (seismic unit B) by thin veneers of Late Palaeocene (seismic unit C) and mid to early Miocene (seismic unit D) transgressive systems tract deposits. These are often removed by erosive hiatuses of early Oligocene and early to mid Pliocene age. These typically form a combined hiatus surface, except in isolated pockets ofthe upper slope where late Miocene forced regressive systems tract units are preserved (anomalous progradational seismic unit). These sediments correspond to the regional outbuilding of the bordering Tukhela and Limpopo cones during relative sea level fall. Either dominant late Pliocene sediments (seismic unit E), or transgressive systems tract sediments which formed prior to the mid Pliocene hiatus, overlie these sediments. Widespread growth faulting, slump structures and prograding clinoforms towards canyon axes indicate that these sediments initiated upper slope failure which served to create proto-canyon rills from which these canyons could evolve. The association of buried fossil canyons with modern day canyons suggests that the rilling and canyon inception process were influenced by palaeotopographic inheritance, where partially infilled fossil canyons captured downslope eroding flow from an unstable upper slope. Where no underlying canyons occur, modern canyons evolved from a downslope to upslope eroding system as they widened and steepened relative to the surrounding slope. Statistical quantification of canyon forms shows a dominance of upslope erosion. Landslide geomorphology and morphometric analysis indicate that this occurred after downslope erosion, where the canyon axis was catastrophically cleared and incised, leading to headward retreat and lateral excavation of the canyon form. Trigger mechanisms for canyon growth and inception point to an overburdening ofthe upper slope causing failure, though processes such as freshwater sapping may emulate this pattern of erosion. It appears that in one instance, Leven Canyon, freshwater exchange with the neighbouring coastal waterbodies has caused canyon growth. The canyons evolved rapidly to their present day forms, and have been subject to increasingly sediment starved conditions, thus limiting their evolution to true shelf breaching canyon systems. Sedimentological and geomorphological studies show that the shelf has had minor fluvial influences, with only limited shelf-drainage interaction having occurred. This is shown by isolated incised valleys of both Late Cretaceous and Late PleistocenelHolocene age. These show classic transgressive valley fills of wave dominated estuaries, indicating that the wave climate was similar to that of today. The narrowness of the shelf and the inheritance of antecedent topography may have been a factor in increasing the preservation potential of these fills. Canyons thus appear to have been "headless" since their inception, apart from Leven Canyon, which had a connection to the Last Glacial Maximum (LGM) St Lucia estuary, and Wright Canyon, which had an ephemeral, shallow LGM channel linking it to the Lake Sibaya estuarine complex. Coastline morphology has been dominated by zeta bays since at least 84 000 BP, thus littoral drift has been limited in the study area since these times. The formation of beachrock and aeolianite sinks during regression from the last interstadial has further reduced sediment supply to the shelf. The prevalence of sea-level notching in canyon heads, associated with sea levels of the LGM indicates that canyon growth via slumping has been limited since that time. Where these are obscured by slumping in the canyon heads (Diepgat Canyon), these slumps have been caused by recent seismic activity. The quiescence of these canyons has resulted in the preservation of the steep upper continental slope as canyon erosion has been insufficient to plane the upper slope to a uniform linear gradient such as that of the heavily incised New Jersey continental margin. Progressive sediment starvation of the area during the Flandrian transgression has resulted in a small shore attached wedge of unconsolidated sediment (seismic unit H) being preserved. This is underlain by a mid-Holocene ravinement surface. This crops out on the outer shelf as a semi-indurated, bioclastic pavement. Thinly mantling this surface are Holocene sediments which have been reworked by the Agulhas Current into bedforms corresponding to the flow regime and sediment availability to the area. Bedforms are in a state of dis-equilibrium with the contemporary hydrodynamic conditions, and are presently being re-ordered. It appears that sediment is not being entrained into the canyons to the extent that active thalweg downcutting is occurring. Off-slope sediment loss occurs only in localized areas, supported by the dominance of finer grained Early Pleistocene sediments of the outer slope. A sand ridge from the mid shelf between Wright and White Sands Canyons appears to have been a palaeo-sediment source to White Sands Canyon, but is currently being reworked southwards towards Wright Canyon. The prevalence of bedform fields south of regularly spaced canyon heads is considered a function of hydrodynamic forcing of the Agulhas Current by canyon topography. These bedforms are orientated in a northerly direction into the canyon heads, a result ofnortherly return eddying at the heads of these canyons. / Thesis (Ph.D.)-University of KwaZulu-Natal, Westville, 2009.

The mapped area, measuring about 400m2, is situated along the southern margin of the Archaean Kaapvaal Craton south of Melmoth in KwaZulu-Natal and comprises greenstones and metasediments forming a narrow, linear E-W trending and dominantly northerly inclined belt flanked to the north and south by various granitoids and granitoid gneisses which have been differentiated for the first time in this study. This belt is here referred to as the ILANGWE GREENSTONE BELT. The lIangwe Belt rocks are grouped into the Umhlathuze Subgroup (a lower metavolcanic suite) and the Nkandla Subgroup (an upper metasedimentary suite). The former consists of: (a) the Sabiza Formation: a lower amphibolite association occurring along the southern margin of the greenstone belt; (b) the Matshansundu Formation: an eastern amphibolite-BIF association; (c) the Olwenjini Formation: an upper or northern amphibolite-banded chert-BIF association. whereas the latter is sub-divided into: (a) the Entembeni Formation: a distinctive phyllite-banded chert-BIF association occurring in the central and the eastern parts of the belt; (b) the Simbagwezi Formation: a phyllite-banded chert-amphibolite association occurring in the western part of the belt, south-east of Nkandla; (c) the Nomangci Formation: a dominantly quartzite and quartz schist formation occurring in the western part of the belt, south-east of Nkandla. The contacts between the six major tectonostratigraphic formations are tectonic. In the eastern sector of the lIangwe Belt, the lowermost metasedimentary formation, the Entembeni Formation, cuts across both the Sabiza and Matshansundu Formations (the lower formations of the Umhlathuze Subgroup) in a major deformed angular unconformity referred to as the Ndloziyana angular unconformity. In the central parts of the belt, the Entembeni Formation structurally overlies the Olwenjini Formation in what seems to be a major local unconformity (disconformity). In the western sector of the belt, the Simbagwezi Formation occurs as a structural wedge between the lower and upper formations of the Umhlathuze Subgroup. That is, it structurally overlies the Sabiza Formation and structurally underlies the Olwenjini Formation. The uppermost metasedimentary unit, the Nomangci Formation occurs as a complex series of finger-like wedges cutting and extending into the Simbagwezi Formation and in each case showing that the Nomangci Formation structurally underlies the Simbagwezi Formation. This structural repetition of lithological units is suggestive of normal dip-slip duplex structures. Palimpsest volcanic features, such as pillow structures and minor ocelli, indicate that many of the amphibolitic rocks represent metavolcanics, possibly transformed oceanic crust. This is also supported by limited major element geochemistry which suggests that the original rocks were ocean tholeiites. Evidence suggests that the talc-tremolite schists and the serpentinitic talc schists represent altered komatiites. The nature of the metasediments (represented by banded metacherts, quartzites and banded iron formations) and their similarity to those of the Barberton, Pietersburg and Nondweni greenstone complexes suggests that they were formed in relatively shallow water environments. The lIangwe magmatism is represented by different types of granitoids and granitoid gneisses and basic-ultrabasic intrusive bodies. Based on similar geochemical and mineralogical characteristics and on regional distribution, mutual associations and contact relationships, these granitoids and granitoid gneisses can be divided into three broad associations, viz: (a) The Amazula Gneiss - Nkwa/ini Mylonitic Gneiss - Nkwalini Quartzofeldspathic Flaser Gneiss Association: a migmatitic paragneiss and mylonitic to flaser gneiss association of older gneisses of Nondweni age occurring in several widely separated areas and intruded by younger granitoids. (b) The early post-Nondweni Granitoids comprising the Nkwalinye Tonalitic Gneiss (a distinctive grey gneiss intrusive into the greenstones and older gneisses) and the Nsengeni Granitoid Suite (an association of three granitoid units of batholithic proportions flanking the greenstone belt and intrusive into the greenstones, older gneisses and Nkwalinye Tonalitic Gneiss). (c) The late post-Nondweni Granitoids comprising the Impisi-Umgabhi Granitoid Suite, a batholithic microcrystic to megacrystic association of five granitoid phases/units occurring to the north and south of the greenstone belt and intrusive into the greenstones, older gneisses and early post-Nondweni granitoids. Limited major element geochemistry suggests that the granitoids and granitoid gneisses are of calc-alkaline origin and are of tonalitic, granodioritic, adamellitic and granitic composition. An igneous derivation from material located possibly at the lower crust or upper mantle is suggested. At least three major episodes of deformation (01, O2 and 03) have been recognized in the greenstones. During 01, a strong penetrative S1 tectonic foliation developed parallel to the So primary layering and bedding. This period was characterized by intense transpositional layering, recumbent and isoclinal intrafolial folding with associated shearing,thrusting and structural repetition of greenstone lithologies. These processes took place in an essentially horizontal, high strain tectonic regime. The first phase of deformation (OG1) in the migmatitic and mylonitic gneisses was also characterized by recumbent and isoclinal intrafolial folding and is remarkably similar to the 01deformational phase in the lIangwe greenstones. Structural features of the first phase of deformation suggest that it was dominated by formation of fold nappes and thrusts and was accompanied by prograde M1 medium-grade middle to upper amphibolite facies metamorphism. During D2 deformation, the subhorizontal D1 structures were refolded by new structures with steeply inclined axial planes. This resulted in the formation of superimposed Type 3 interference folding in the amphibolitic rocks and large-scale, E-W trending, doublyplunging periclinal folds in the metasediments. These periclinal folds have steeply inclined and overturned limbs and are characterized by narrow, closed elliptical outcrop patterns well-defined by extensive banded ironstones and metacherts. The second phase of deformation in the granitoids (DG 2) was characterized by steeply plunging and steeply inclined small-scale tight to isoclinal similar folds. Large-scale folds are not present in the granitoids. Evidence suggests that the second phase of deformation was a major compressional event which resulted in the large-scale upright, flattened flexural folds. It was accompanied by widespread regional greenschist metamorphism and the intrusion of the early postNondweni granitoids. The third phase of deformation produced steeply plunging small-scale folds on the limbs and axial planes of the pre-existing large-scale F2 folds and upright open folds in the granitoid terrain. This episode was characterized by the emplacement of the late postNondweni granitoids (along the D2 greenstone boundary faults) and is associated with two significant events of prograde M3 upper greenschist facies metamorphism and retrograde M3 lower greenschist facies metamorphism. Post-D3 deformation is characterized by late cross-cutting faults and the emplacement of younger basic - ultrabasic bodies. / Thesis (Ph.D.)-University of Natal, 1997.

This study presents the interpretation of Edo-Western and Sparker seismic geophysical data acquired on the northern KwaZulu-Natal upper continental margin by various organisations since 1981. Five seismic sequences are recognised and these are traceable across the entire length of the study area. The oldest is interpreted as a late Cretaceous marine sequence (Sequence A), probably the offshore equivalent of the St. Lucia Formation exposed onshore. This sequence is overlain by a progradational, probable late Tertiary shelf sequence (Sequence B) onlapping in places against the underlying marine sequence. The outer portion of this sequence on the upper continental slope is characterised by complicated reflection termination patterns indicating the possible presence of discreet sequences within this shelf and slope unit. These shelf and slope sediments are overlain by a thin (less than 20m) reworked and eroded Pleistocene shelf unit (Sequence C), itself overlain by linear Pleistocene aeolianites (Sequence D) in places. The youngest sequence observed is the Holocene unconsolidated sediment wedge (Sequence E) on the inner shelf, attaining thicknesses of greater than 20m in places. The various sequences were mapped out and sediment isopach maps were produced (wherever possible) as well as an overall geological subcrop map of the study area. 150 kilometres of shallow penetration Edo Western seismic records acquired off the Sodwana Bay continental shelf were interpreted. Two sediment types are recognised, namely consolidated beach rock/aeolianite and unconsolidated Quaternary shelf sand/bioclastic reef derived sediment. In places, accumulations of bioclastic sediment in subaqueous dune troughs which have been subsequently buried by migrating bedforms manifest themselves on seismic records as dark semi-continuous reflectors beneath the migrating bedform. Close inshore, seismic records show prominent reflectors interpreted as consolidated sediment beneath varying thicknesses of unconsolidated sediment. Close to the shelf break (occurring at approximately -60m), seismic interpretation indicates that thin beach rock developments perch directly upon unconsolidated shelf sand, with the beach rock having been eroded through in places to expose unconsolidated sediment beneath. A sediment thickness map for this area was compiled from the seismic data. The limited penetration of the Pinger system necessitated "greater-than" values being used in many areas. Greatest sediment thicknesses occur in subaqueous dune fields where unconsolidated sediment thickness is at least 11 m. In inshore areas absent of subaqueous dune fields, sediment thicknesses are typically low, varying between 1 and 3m. A prominent submerged dune ridge close inshore limits substantial unconsolidated sediment build-up to landward of this feature. On the seaward side substantial build-up is limited by the action of the Agulhas Current which is actively transporting sediment into the head of submarine canyons which incise the continental shelf at Sodwana Bay. This study shows that on the northern KwaZulu-Natal continental shelf where there is a dearth of unconsolidated Quaternary sediment, the Edo Western seismic system is a useful tool for discerning thin veneers of unconsolidated sediment less than 4m thick. When considering the overall low volumes of unconsolidated sediment present on the shelf, this hitherto unconsidered volume of sediment constitutes an important part of the shelf sediment budget. Submarine landslide features observed on sparker seismic records are described and discussed. Submarine landslides are present which affect a) Sequences A and B, b) Sequence B only and c) Sequence A only, ages of these sediment failures can thus be inferred as being either post- Late Cretaceous or post- Late Tertiary. Offshore Kosi Bay, submarine landslide features affecting Sequence A are buried by unaffected Sequence B sediments, indicating a post- Late Cretaceous to pre- Late Tertiary age of occurrence. Style of failure tends towards mass flow in those submarine landslides in which Sequence B only sediments are affected, while those in which Sequence A is affected exhibit some slide features indicating a greater degree of internal coherency of these sediments compared to Sequence B. Slope stability analysis of a submarine landslide feature offshore St. Lucia Estuary Mouth indicates the failed sediment mass would have been stable under static conditions and that external dynamic forces such as storm waves or seismic activity would have been necessary to induce failure. It is demonstrated that the Zululand earthquake of 1932 would have exceeded the intensity necessary to induce sediment failure and this event should therefore be considered as a possible cause. Seismic evidence of fluvial incision/subaerial exposure at the boundaries between Sequences A and B and C and E are further evidence of lowered sea-levels probably during the Oligocene and Late Pleistocene. The position of the incision into Sequence C relative the present course of the Mkuze River indicates the possibility that this incision could represent the palaeo-outlet of this river. Seismic expression of 3 submarine canyons in the study area indicate that they are currently undergoing active headward erosion, independent of any direct modern fluvial influence. In the case of Ntabende Canyon, a nearby continental shelf incision postulated to be the palaeo-Mkuze outlet indicates that provision of terrigenous material to this portion of the continental shelf could well have accelerated mass wasting processes within the canyon itself. This submarine canyon could therefore have progressed more rapidly to a relatively mature phase of development. Subsurface structure indicates the lack of any post- Late Tertiary fault features beneath the canyons, thus excluding faults active in post- Late Tertiary times as a developmental factor. It is shown that the overall, external morphology of the KwaZulu-Natal upper continental margin is strongly influenced by seismic stratigraphic relationships, with the main influencing factors being outcrop position of the various sequences and depositional angle of sediments of which a sequence is comprised. External morphology has also been greatly modified in places by mass-wasting processes. It is demonstrated also that relating the observed seismic stratigraphy to onshore geological cross sections is problematic due to the distances involved and lack of confident offshore dates for the seismic sequences observed. Seismic relationships observed contribute to an understanding of relative sea-level movements since the Late Cretaceous and the overall geological evolution of the northern KwaZulu-Natal upper continental margin, details of which are discussed. / Thesis (M.Sc.)-University of Natal, Durban, 1998.

The morphology of Lake Sibaya is a product of an ancient fluvial system that drained a coastal landscape dominated by aeolian processes. The sedimentary processes within the lake are driven by wind generated currents. The dominant sedimentary process is one of lake segmentation, whereby prograding bedforms isolate the lake into smaller water bodies. The prograding bedforms include cuspate forelands and sand spits. The size and mobility of these bedforms is a function of sediment availability and current regime. The bathymetry of Lake Sibaya is discussed, with emphasis on geomorphic features derived from the ancient aeolian landscape as well as features related to modern sedimentary processes. The presence of underwater knickpoints and terraces indicate that lake level fluctuations have been common in Lake Sibaya. It is during lake highstands that large volumes of sand are eroded from aeolian dunes which surround the lake and made available for shoreline progradation. Ancient dune topography is preserved to depths of 20 m below water-level within the lake. Surface sediment distribution maps were compiled from 515 grab samples and thirteen core samples. Fine grained, well sorted, coarse skewed quartz sand comprises the majority of the surface area of the lake floor. Gyttja is the other dominant sediment type and accumulates in palaeovalleys and depressions on the lake floor. Sediment distribution in Lake Sibaya is discussed in terms of modern lacustrine processes as well as inherited sedimentary characteristics. The stratigraphy of the sediments underlying Lake Sibaya was investigated using a Uni-Boom seismic profiling system. Seismic profiles were compiled by identifying acoustically reflective surfaces that show regional development. Thirteen seismic overlays were prepared, and are illustrated as west - east and north - south seismic profiles. Five sequences ranging in age from late Cretaceous to Holocene were identified from the seismic profiles, and are described in terms of sequence stratigraphic principles. The seismic sequences were interpreted within a lithostratigraphic framework and are presented as a series of idealised geological sections. Thirteen sediment cores were collected from the Lake Sibaya area in order to ascertain the accuracy of the stratigraphic interpretation of the seismic records, to investigate reflective horizons identified from seismic records and to collect dateable material. Interpretation of the sediment cores reveals that a proto Lake Sibaya existed on drowned dune topography, during the period ± 43500 BP to ± 25500 BP prior to the Last Glacial Maximum. During the early to mid Holocene the Lake Sibaya site was occupied by a saline lagoon which underwent isolation from the sea ± 5030 BP. Since the mid-Holocene the lake has evolved to totally freshwater conditions and has undergone little sedimentation. The geological evolution of the Lake Sibaya area is discussed in terms of the geometry of the identified seismic sequences, the sedimentary characteristics of these sequences and the radiocarbon dates provided from the sediment cores. Palaeo-environmental conditions during the accumulation of the sedimentary sequences is discussed where fossil remains permit. / Thesis (M.Sc.)-University of Natal, Durban, 1998.

The eastern portion of the Namaqua-Natal Mobile Belt, the Natal Metamorphic Province is divided into four main tectonostratigraphic units. These units comprise two accreted island arcs: the Mzumbe and Margate Terranes; an imbricately thrust nappe zone consisting of four ophiolitic nappes in a hinterland-dipping duplex; and the highly deformed metavolcaniclastic/metagreywacke Mfongosi Group directly adjacent to the stable northern foreland of the Kaapvaal Craton. Theories of late-tectonic left-lateral movement in the southern island arcs are extrapolated northwards of the southern margin of the Kaapvaal Craton coincident with the Lilani-Matigulu Shear Zone. The relative timing and structural context of vein-hosted mineralization with respect to major recognized tectonic events is resolved in five separate areas, two in the Natal Nappe Zone and three in the Natal Thrust Front. The Madidima Nappe of the Natal Nappe Zone contains several north-northeast- to northeast-trending and northeast- to east-northeast trending quartzofeldspathic veined reefs considered to have formed in a late-tectonic left-lateral shear system (main shear and synthetic shear orientations, respectively). The northeast- to east-northeast-trending reef is duplicated due to infilling of normally-faulted steep structures in the semi-brittle, incremental normal faulting of the banded amphibolite component of the nappe. Later left-lateral movement has reactivated one of these steep structures along the southern margin of a regional F2-folded band of granite-gneiss in that a southwest extension of this structure may be responsible for sub-economic veining for a length of up to 9 km. The extensive flat-lying topography of the Mbongolwane Flats area, in which the reefs are situated, is accounted for by the accelerated weathering of rocks which underwent sustained late-tectonic metamorphism in the epidoteactinolite facies, accompanied by pervasive shearing and block rotation to the south of the southern limb of the regional F2 fold in the granite-gneiss. A large, kilometer-scale, open advective fluid system which provided fluid-mediated exchange between co-existing rocks existed at the time of vein formation. The fluid system was driven by early-tectonic intrusion of a granite gneiss and amphibole-rich granite. Two areas in the Mfongosi River valley, the northern and southern Mfongosi Valley areas, contain typical evidence of deformation at the leading edge of collision in a mobile belt. The southern Mfongosi Valley area, at the confluence of the Mfongosi and Tugela Rivers, contains veining which resulted from pressure solution of the host metavolcaniclastic/metagreywacke. Veining occupies predictable shear and tension fractures formed during the initial deformation of a foreland margin sequence, in addition to occupying those fractures formed by buckling on the layer-scale. The structural context of the northern Mfongosi Valley veining is defined by subsequent deformation and vein fragmentation such that the metavolcaniclastic/metagreywacke was reduced to a melange in which vein segments acted as competent clasts; a large-scale porphyroblast/matrix system. Formation of the Manyane Thrust to the south of the Mfongosi Group interrupted the normal retrograde metamorphism of the remainder of the Tugela Nappe and initiated a "hot iron effect" whereby a short-lived thermal pulse acted at the thrust plane, producing a reversed geothermal gradient in the underlying Mfongosi group. This reversed gradient would have been counteracted by a steepened normal geothermal gradient in the Mfongosi Group caused by overloading of the Natal Thrust Front by the Natal Nappe Zone. These geothermal gradients partly account for the concentration of veining in the areas of the Mfongosi Group which are directly adjacent to the Manyane Thrust, and directly adjacent to the Kaapvaal Craton, in the lower portions of the thrust front Stable isotope studies indicate fractionation between vein and wall rock under a short-lived, mainly rock-buffered, layer-scale fluid-movement system. Also forming part of the Mfongosi Group of the Natal Thrust Front, the Ngubevu area contains an apparently enigmatic distribution of veining accompanied by gold and base metal mineralization. The structural evolution of the Ngubevu area occurred during consistent left-lateral transpression into which has intruded early-tectonic veins, formed by pressure solution and having the same structural format as the early-tectonic veining in the southern Mfongosi Valley area. Subsequent deformation of the system was accompanied by 1900 -trending tension gashes which were continually ptygmatically-folded, sheared and offset to form occasionally mineralized quartzofeldspathic "blows" and along-strike stringers in the epidote- actinolite schist. Where veining cross-cuts narrow calcite - graphite - sericite - quartz - albite - tourmaline ± chlorite schist layers, gold mineralization occurred. The late-tectonic tension gashes, antitaxially filled by quartz and amorphous calcite, cross-cut the entire range of lithologies. The fluid system during vein deposition varied: during infilling of early-tectonic fractures a short-lived fluid-flow system dominated, with the emplacement of re crystallized wallrock occurring in a closed, non-advective regime under the influence of diffusion caused by pressure solution. The fluid system changed to a more open, advective, greater than layer-scale rock-buffered one with a decreasing contribution of material from immediate host rocks. An internal fluid source is implied for the entire period of vein emplacement, derived from structural analyses which indicates negative dilation across the Mfongosi Group in this area and by comparison of vein:wallrock δ180 values which indicate a lack of igneous-derived fluids. The Phoenix Mine, in the central portion of the Tugela Nappe, and the Ayres Reef, hosted in Manyane amphibolite adjacent to the Manyane Thrust, are grouped together on the basis of their cross-cutting nature and timing with respect to metamorphism and deformation of the host rock, and also due to their similarity in isotopic plots. Both vein sets occur in approximately east-west to east-northeast-trending zones which show evidence of late-tectonic left-lateral movement. Phoenix Mine veining occurs in weakly-metamorphosed meta-gabbro/meta-norite of the Tugela Rand Complex. The Manyane amphibolite demonstrates the amphibolite facies of metamorphism due to the short-lived thermal pulse at the Manyane Thrust. Both sets of veining display slickenlines which are indicative of their emplacement prior to the late-tectonic left-lateral movement. The unusually thick quartz veins of both deposits are the results of late- to post-Tugela Rand Complex fluids or the tapping of late-tectonic metamorphic fluid reservoirs. This caused silica metasomatism and redeposition of material in post-thrusting collapse features. A highly channelized, single-pass fluid system is proposed in the absence of intrusion-derived fluids. Whole rock geochemical data allow a distinction to be made between the Natal Thrust Front and the Natal Nappe Zone: the Foremost nappe of the nappe zone consists primarily of N-type mid-ocean ridge basalts/ocean-floor to within-plate basalts which were intruded prior to nappe emplacement by metaluminous orogenic volcanic arc granitiods. The thrust front displays a lateral variation in metabasite/metasediment ratio, with the ratio increasing from east to west in this inlier. In the east, in the Nkandlha area, melanged metagreywackes dominate and there is a marked paucity of associated metabasites. In the central portions of the thrust front, in the vicinity of the Mfongosi area, active continental margin/continental arc magmatogenic greywackes and arkoses are interlayered with calk-alkaline volcanic arc basalts (volcaniclastics). The greywacke geochemistry indicates little to no mafic/ultramafic influences in sediment contribution and the source of sediment is inferred to be the southern portions of the Kaapvaal Craton. The Nkandlha and Mfongosi area Mfongosi Group segments are considered to be in-situ or para-autochthonous. The western-most Ngubevu area predominantly hosts metabasites. The geochemistry of the metabasites indicates that they are N-type mid-ocean ridge basalts/ocean floor basalts from a destructive plate margin setting. The metabasites are interbanded with metapelitic/metacalcsilicate layers produced in a shallow water oxic environment, here inferred as a spatially-restricted shallow, marginal basin. The metabasites in the Ngubevu area are notably similar to those of the Madidima Nappe, indicating a similar provenance and pre-collisional mode of formation. It is proposed that the variation in the Natal Thrust Front was due to a north-east/south-west distribution of lithological proportions or mixing, with greywackes dominating in the northeast (in proximity to the Kaapvaal Craton) and metabasites dominating in the southwest. Left-lateral transpressional movement within the Mfongosi Group of the Natal Thrust Front, and the Natal Nappe Zone, was continuous throughout plate collision and obduction. / Thesis (Ph.D.)-University of Natal, Durban, 2000.

A regional field and geochemical study has allowed the identification of three primary units within the Proterozoic basement of the Valley of a Thousand Hills. The Nagle Dam Formation incorporates several chemically distinct orthogneiss series, characterised by limited intragroup fractionation, and derived from discrete sources. Intrusive into the gneisses are the megacrystic A-type granites of the Mgeni batholith, comprising the biotite granites of the Ximba Suite; the hornblende granites and charnockite of the Mlahlanja Suite; and the medium grained leucogranite of the Nqwadolo Suite. Petrogenetic modelling indicates that these are predominately cumulates. A general model for the A-type granites suggests that they were derived through variable MASH processes on an original within plate type basalt. Enclaves within the Mgeni batholith form a distinct series, the Valley Trust Formation, comprising a nongenetic orthogneiss association of amphibolite and crustal sourced quartzo-feldspathic gneiss and locally derived paragneisses. Interaction between the biotite granite and the pelitic enclaves generated a biotite garnet granite. Geothermobarometry suggests temperatures of metamorphism to a maximum of 770°c for the Nagle Dam Formation and c.850°C at a pressure of 6 kb for the Valley Trust Formation. Potential magmatic temperatures of c.760°C at 5 kb are derived for the Mgeni batholith. High Mn garnets within late veins indicate subsequent intrusion at higher levels. Derivation of a tectonic model for the Valley of a Thousand Hills is assisted by a revaluation of the chemical tectonic discrimination plots as source or initiator discriminators. These indicate an origin for the Nagle Dam Formation in an arc environment, while the bimodal orthogneiss association of the Valley Trust Formation and the A-type character of the Mgeni batholith suggests their evolution during extensional events. Geothermobarometry defines an isothermal decompression path, possibly generated during a collision event, superimposed on which is a potential midcrustal heating event, resultant on the intrusion of the Mgeni batholith. These data can be integrated with revised lithotectonic data from the southern portion of the Natal Province to derive a regional model. This comprises: the collision of a number of arcs with associated splitting to form backarcs, sedimentation, and failed rift systems; syn-collisional S-type magmatism, contemporaneous with isothermal decompression of the region; and a series of pulses of post-orogenic granites. / Thesis (Ph.D.)-University of Natal, Durban, 1999.

Mechanisms of sill and dyke intrusion require an understanding of fracture growth,stress distributions and intensities, dilation,intrusion rates, hydraulic pressure,host-rock effects,en echelon fracture arrays,and flow direction. The methods of previous studies have been applied to natural sill and dyke examples of the Karoo Igneous Province in northern Natal . An en echelon array of Jurassic dolerite sills occurs within Permian Ecca sediments along the Mhlatuze River, west of Empangeni. Dolerite emplacement occurred as two intrusive phases. The first phase resulted in thick, . coarse-grained dolerite sills. The second phase produced relatively thinner, fme-grained sills. The intrusion of fmegrained dolerite into older sills is demonstrated by abrupt variations in the whole-rock and mineral geochemistry profiles across the sills. Syn-crystallisation effects such as crystal settling and fractionation, and post-crystallisation hydrothermal activity is also manifested in the mineralogical and geochemical changes across the sills. The fine grained doleriteis associated with xenolithic dolerite which represents a contaminated magma propagation front of the fine-grained dolerite. The higher viscosity of the xenolithic dolerite hindered propagation, and was thus overtaken and engulfedby the mainmagmapulse. Consistent sinistral off setting of sill segments is interpreted to be the result of a fingered sill periphery intruding an en echelon fracture array. Dilation of individual segments, or fingers, occurred simultaneously. Subsequent interaction of near-tip stresses induced inwardly propagating curvature of adjacent segmentsin the array.Resultant linkage has produced a stepped-sill geometry; sill propagation and flow directions were orthogonal to the plane of linkage. The flow direction is confirmed by shape preferred-orientations of acicular mineral grains within the chilled margins of the sills, indicating the direction of flow to be perpendicular to the plane of the en echelon array, and parallel to strike directions of offset surfaces that link adjacent sill segments. Multiple dyke intrusion is examinedat an outcrop of the Rooi Rand Dyke Swarm, along the Pongola River. Individual intrusive episodes are identifiable on the basis of chill-zone relationships. The pattern along the Pongola River suggests that younger intrusive episodes frequently intrude through the centres of older dykes. A three dimensional analysis of en echelon dyke let segments allows a re-construction of the dilation history,and provides an explanation for the development of blunt-ended intrusion segments. Mineral geochemistry anomalies around dyke tips suggests possible facilitation of incipient fracture via decreases in mineral strengths manifested by geochemical changes. A statistical digital analysis of micro-phenocryst orientations within chilled dyke margins is shown to provide a viable method to ascertain magma flow directions within dykes, and may thus be a useful tool for future investigations. / Thesis (M.Sc.)-University of Natal, 1994.