Search

The geometry and structural history of high-grade granite-gneiss units, which host the Venetia diatremes in South Africa, have been modelled in 3D using an implicit, rules-based, conditional geometrical technique. The volume contains a pervasive, dominant foliation , which forms the main plane of strength anisotropy in ongoing design and mining of the Venetia pit, which has a planned final depth of approximately 465 m below surface. The well-developed S 2 foliation is a continuous feature that was locally re-oriented during D3 and D4 and localised ‘wrapping’ around competent amphibolite/hornblendite lenses or boudins. These local variations have important implications for pit design and operational risk management, particularly where they dip adversely out of the slope, or occur at the base of bench stacks, or in inter-ramp areas. These structures, when undercut by dips steeper than their friction angle, may result in planar sliding, presenting an operational safety risk. Best practice requires early identification of these occurrences, to make design adjustments where warranted. Lithological contacts, which have been transposed into parallelism with the S2 foliation, and a robust set of structural data collected over the last 10 years, are analysed to quantify and represent the variable orientation of S2 in 3D space. This contribution describes several methods, using leading software packages such as Leapfrog Geo™, Micromine™ and Python, to generate representative S2 form surfaces and anisotropy block models, both of which are used for downstream geotechnical engineering analysis. Outputs are also translated into apparent dip “heat” maps that show the angular interaction between S 2 foliation and a pit design surface. These methods augment, and lessen the dependence on, typical 2D wedge-shaped design sectors that are commonly employed in pit design. This approach has important safety and economic benefits. As the rock mass fabric is characterised, it supports quick identification of potentially higher-risk areas (e.g. adverse undercut fabric with greater likelihood of slope instability) in current mine plans, which justifies more detailed analysis and/or additional monitoring controls to ensure safe and efficient mining. It also allows rapid mine design optimisation.

The Bushveld Complex (BC) is the world&rsquo, s largest source of platinum group metals. Extensive studies on the complex have focused on its geochemistry, magma and platinum group mineral genesis, mineral characterization and intrusion mechanisms. However, relatively little work has been undertaken on the overall 3D geometry of the complex, which detracts from the adequate contextualization of such studies. Furthermore, the absence of a broader 3D model of the complex does not permit the identification of structural trends and mineralization patterns. This contribution details the construction of 3D implicitly-modelled Merensky and UG2 Reefs across the Rustenburg Layered Suite of the BC, using Seequent&rsquo, s Leapfrog software. Multiple open-source and public-domain data sources and modelling workflows were explored to account for disparities in data resolution, data spacing and clustering, and the resolution of model outputs. Key outcomes are (1) a representative, fully-implicit, dynamic geological model of the Merensky and UG2 Reefs over the main chamber of the BC, (2) identification of modelled features that augment the current understanding of the BC&rsquo, s kinematic history and cumulative deformation, and (3) identification and analysis of subtle geometrical trends and patterns, such as inter-reef spacing and modelled depths, as well as structural domains that may not have been apparent from numerous, more focused or isolated petrological or geochemical studies. It is anticipated that this baseline 3D model will form the foundation for future, possibly localized, dynamic updates as further information becomes available. The addition of proprietary (viz., non-open-source) data, such as 2D seismic sections and 3D seismic surveys, would enhance the overall resolution and quality of such a model and resulting interpretations.

X-ray computed tomography is a non-destructive 3D analytical technique, which in recent years has gained more widespread applications to characterise the internal structure of materials in minerals processing and metallurgical studies. Successful application of the technique relies on effective X-ray penetration. Ore samples with high average specific gravities (e.g. iron ore, massive sulfide ores) inhibit X-Ray penetration causing beam hardening artefacts that reduce data quality and limit further analysis. This contribution presents a practical way to evaluate the degree of beam hardening using a case study iron ore sample. The method quantifies the degree of beam hardening that leads to a loss of sample information, through comparing the known pore surface area of an aluminium standard sample with that of an iron ore sample. This comparison is defined as a %Error. Porosity and sample mineralogy are confidently quantified when the %Error is less than 10%. Above 10%, there is inconsistent loss of sample information (porosity and relatively low dense sample mineralogy) and the results from the volumes cannot be trusted which indicates that smaller sample sizes need to be scanned. This practical method can be routinely applied in other applications and samples with density higher than iron ore.

A structural review and re-interpretation of the Bushveld Complex, based on 254 georeferenced maps from research, technical reports and geophysical surveys, suggests that its current geometry differs markedly from its original geometry. Broadly N–S flattening, left-lateral transpressional shearing along an interconnected fault network, steepening of cumulate layering and 10 km of uplift, to the north of the northern margin of the complex's main chamber, were accompanied by an eastward-verging escape structure (Burgersfort Bulge) and syn-cooling gap areas on its Western Limb. Southward-verging deformation was transmitted from the Villa Nora Fragment, via thrusting in the Rooiberg area, generating 50–70 km of southward displacement, uplift and erosional loss of the Northern Limb's western portion. A string of anomalous occurrences of Bushveld Complex units and dense areas outlined by unconstrained gravity isoshells suggest a further 170 km southward displacement through the center of the main chamber. The Southern Limb underwent predominantly right-lateral, ductile offset from the Western Limb along the Crocodile River and Rustenburg Faults, expressed by an in-line series of NNW-trending magnetic highs to the south of their mapped terminations. Left-lateral shearing and progressive splaying of a NNE-trending dyke swarm, to the west of an inflection point, provides strain markers for this offset and 35° anticlockwise rotation of the Burgersfort Bulge and central parts of the main chamber. Although recent work suggests that the complex cooled within 1 Myr, late-intrusion/syn-cooling, SSE-wards movement of its deeper, ductile, central parts, within at least 5–10 Myr after its initial intrusion, the Transvaalide Fold-and Thrust event and the expulsion of fluids to form the high-Ti igneous suite (HITIS) and related bodies, appear to have closely post-dated the complex's intrusion. Removal of the effects of cumulative strain produces a more equant complex, shortly after its intrusion, occupying an orthogonal network of normal faults and perpendicular, strike-slip, transverse faults.

Kolomela Mine, formerly known as the Sishen South Project, is located approximately 9 km from Postmasburg, in the Northern Cape Province of South Africa. Iron ore is primarily concentrated near the top of the Asbesheuwels or Asbestos Hills Subgroup, within the Kuruman Formation. High-grade, hematite-rich, iron-ore deposits in South Africa have been addressed in the literature, although much of this work focusses on their geochemical or isotopic signatures and the role of paleosinkholes in ore preservation. Recent work on high-grade, BIF-hosted Fe deposits, in South Africa and elsewhere, emphasizes the role of deformation in iron mineralisation and upgrading of BIF over a sequence of events, either punctuated by supergene enrichment or terminating with a supergene overprint. The proximity of the Kheis orogenic front, the protracted tectonic history of the area, a recently-resolved local structural framework and an emerging set of literature on superimposed hypogene, metasomatic, hydrothermal, hypothermal and deformation-induced processes, justify a re-examination of the Kolomela deposits. Closely-spaced drillhole data, pit mapping over a period of four years and re-interpretation of the tectonic setting from high-resolution geophysical data, have been incorporated into fully-constrained 3D models of the five principle Kolomela deposits, which are presented here and analysed in terms of their geometry, tectonic setting, relationship to major structures, the geometry of the underlying dolomite contact and the possible role of gabbroic intrusions. A strong spatial correlation between thicker ore and thicker, underlying gabbroic bodies, in the forms of sills, is evident. We propose that protracted, multi-phase tectonic evolution, complete with several compressional-extensional events at this margin of the Kaapvaal Craton provided an ideal setting for “preparation” of BIF by early low-grade metamorphism, deformation and porosity creation, followed by fluid movement along interconnected contacts, unconformities, gabbro contacts and major structures.

The Namib Lead and Zinc (Pb-Zn) Mine is hosted within the Karibib Formation of the Swakop Group, Damara Sequence. The host marble contains sulphide-rich beds and is locally Pb- and Zn-rich over a strike extent of >100 km. Mining took place in the upper portions of a remobilized MVT or SEDEX orebody, which may occur adjacent to a reactivated major structure. Pb and Zn show locally anomalous copper, tin, indium and fluorine concentrations. Based on the position of gossans, individual orebodies comprise discrete elongate shoots around an antiform and its meso-scale parasitic folds, within calcitic marble. Non-coaxial flattening of the fold-cleavage geometry reflects a series of non-coaxial deformation events, recorded in the Southern Central Zone of the Damara Belt which were superimposed on a NE-SW structural grain and pre-existing fold axial planes. Non-coaxial flattening resulted in further remobilization of sulphides into dilational rhombs which formed from shearing along overlapping axial planar cleavage, in concert with the dilation of banding in the host marble.

Country rock at Jwaneng Diamond Mine provides a rare insight into the deformational history of the Transvaal Supergroup in southern Botswana. The ca. 235 Ma kimberlite diatremes intruded into late Archaean to Early Proterozoic, mixed, siliciclastic-carbonate sediments, that were subjected to at least three deformational events. The first deformational event (D 1 ), caused by NW-SE directed compression, is responsible for NE-trending, open folds (F 1 ) with associated diverging, fanning, axial planar cleavage. The second deformational event (D 2 ) is probably progressive, involving a clockwise rotation of the principal stress to NE-SW trends. Early D 2 , which was N-S directed, involved left-lateral, oblique shearing along cleavage planes that developed around F 1 folds, along with the development of antithetic structures. Progressive clockwise rotation of far-field forces saw the development of NW-trending folds (F 2 ) and its associated, weak, axial planar cleavage. D 3 is an extensional event in which normal faulting, along pre-existing cleavage planes, created a series of rhomboid-shaped, fault-bounded blocks. Normal faults, which bound these blocks, are the dominant structures at Jwaneng Mine. Combined with block rotation and NW-dipping bedding, a horst-like structure on the northwestern limb of a broad, gentle, NE-trending anticline is indicated. The early compressional and subsequent extensional events are consistent throughout the Jwaneng-Ramotswa-Lobatse-Thabazimbi area, suggesting that a large area records the same fault geometry and, consequently, deformational history. It is proposed that Jwaneng Mine is at or near the northernmost limit of the initial, northwards-directed compressional event.

The Sishen deposit is one of the largest iron ore concentrations in current production. Hematite mineralization occurs along a strike length of 14 km, with a width of 3.2 km and a maximum vertical extent of 400 m below the original surface. The 986-Mt reserve incorporates a suite of individual orebodies, beneath a locally preserved tectonized unconformity, with a wide range of geometries, depths, and orientations. Fully constrained, implicit 3D modeling of the entire mining volume (> 70 km3), was undertaken to the original, pre-mining topography. The model incorporates 5287 mapping points and > 21,000 drillholes and provides exceptional insight into the original configuration of ore and its relationship to contacts, unconformities, and structures in the enclosing country rock. The bulk of ore occurs to the west of a strike-extensive, partially inverted normal fault (Sloep Fault), within an asymmetrical synclinal structure on its western flank. This linear, N-S distribution of deep, thick ore is punctuated by palaeosinkholes, wherein base-of-ore dips of greater than 45°, are concentrically arranged. Localized ore volumes also occur along faults and in fault-bounded, downthrown blocks, to the north of NW-SE- and NE-SW-trending strike-slip faults that show relatively minor uplift to the south, probably due to the Lomanian Namaqua-Natal Orogeny. The revised model demonstrates the proximity of ore to a tectonized unconformity and highlights the structural control on ore volumes, implying that Fe mineralization at Sishen cannot be exclusively attributed to supergene enrichment and concentric palaeosinkhole formation.

A fully-constrained, implicit, 3D geological model of Sishen Mine reveals the original, pre-mining geometry of ore bodies, host rocks to mineralization and major structures. There are several overlapping controls, at a variety of scales, on the position, depth and geometry of laminated and conglomeratic ore. Most of these controls are structural or may be reconciled with the kinematic history of this part of the Maremane Dome. A series of near-horizontal sections, through the entire 3D model, demonstrates the manner in which these controls overlap and interact. First-order or large-scale controls comprise broad domes, which show preservation of laminated ore around their rims, outside of which conglomeratic ore occurs. Second-order controls comprise grabens and half-grabens, which are often bounded by strike-persistent normal faults, which show fault drag on their western flanks due to inversion, along with preservation of BIF-related supergene ore and conglomeratic ore. A type example is the thick, deep, linear ore to the west of the Sloep Fault. Third-order controls on the preservation of mineralization comprise downthrown blocks to the north of reactivated E-W, SE/ESE- or NE/ENE-trending conjugate faults. Upthrow to the south could be attributed to the 1.15–1.0 Ga NNW-directed Lomanian (Namaqua-Natal) Orogeny. Palaeosinkholes comprise fourth-order controls, which are superimposed on higher-order controls. Palaeosinkholes, which form the bulk of current mining, comprise deep, conical depressions with anomalous thicknesses of chert, chert breccia and haematite. Due to their limited size, the steepness of all units and the often chaotic nature of detached and slumped blocks in their centres, these volumes reflect longstanding models on palaeosinkhole development and very local ore control.

The c. 2060 Ma Phalaborwa Igneous Complex forms an elongate intrusion into Archean granitic gneiss. The carbonatite within the central pyroxenite core of the complex (Loolekop) is well-mineralized in copper. Open pit mining operations started in 1965, followed by underground block caving in 2003. Although little attention has been paid to large-scale structures associated with intrusive phases and mineralization, ongoing infrastructure development and block caving, as part of the new Lift II Project, require far greater resolution of structural discontinuities. 3D modelling of these structures, from over 50 years of data, reveals that Loolekop occurs at the confluence of several major shears or fault zones. Of these, five major structures were pivotal in the emplacement of banded carbonatite, transgressive carbonatite and very late-stage, narrow, E-W trending, sulphide veinlets with short down-dip and along-strike extensions, which form the bulk of mineralization. Modelled structures typically have two or more segments, which are rotated with respect to one another, in turn suggesting repeated rotation or torsion of the entire intrusive volume, aided by cross-cutting structures. The oldest structure is the N-S trending Mica Fault Zone, which shows the same trend as the entire carbonatite complex and the nearby eastern edge of the Kaapvaal Craton and the Lebombo Lineament. The youngest structure is the Central Fault, which shows an E-W inflection that is co-incident with the carbonatite and the E-W, vein-hosted Cu mineralization trend. Based on cross-cutting relationships, sinistral movement along the Central Fault Zone and its localized E-W dilational jog is invoked as a mechanism for transgressive carbonatite emplacement and the introduction of late-stage Cu-rich fluids into numerous tensional veinlets. This shearing would have been caused by an E-W trending maximum principal stress orientation. In turn, this corresponds with the orientation of near-field, eastward-directed stress along the eastern lobe of the Bushveld Complex during its emplacement and subsequent deformation.

Chemical-sedimentary rocks of the Transvaal Supergroup, along the northern margin of the Transvaal Basin, were imbricated during formation of the Mohlapitsi Fold-and-Thrust Belt between ca. 2.55 and 1.93 Ga. High-grade (> 63% Fe), stratiform Fe-ore bodies occur at or close to the base of the banded ironstone-dominated Penge Formation, which forms a distinct E-W trending mountain range in the area. The Penge Formation is separated from underlying Malmani dolostone by a thin, laterally continuous, rheologically-weak, carbonaceous shale layer that has been intensely sheared. This horizon denotes the position of a regional, top-to-N, low-angle detachment that has been variably deformed during progressive development of the Mohlapitsi Belt. Several of Kumba's Thabazimbi iron ore deposits, including Kwaggashoek East, Donkerpoort West, Kumba-Donkerpoort West and Buffelshoek West, are confined to a zone directly above this detachment. Recent studies of ore petrography, fluid alteration, fluid composition and stable isotope geochemistry link economic iron ore mineralization to hydrothermal or hypothermal processes. Along with localized high-grade mineralization, a structural control on fluid flow has been suggested, although a lack of structural data has hampered more concise mineralization models thus far. In this study we present and analyse data from detailed structural mapping over a protracted period of mining, across four opencast pits. This mapping is combined with drillhole data to produce accurate, intergrated 3D models of these deposits, which reveal the true geometry of these ore bodies. Data and resultant models indicate that enhanced hydrothermal fluid flow associated with significant iron mineralization is linked to thrusting/shearing in response to: 1) fold amplification and flexural flow along the E-W trending detachment between the Malmani dolostone and banded iron formation; 2) subsidiary fold-accommodation structures such as conjugate shear joints, normal and reverse faults and thrusts, associated with inner- and outer-arc shortening and extension, respectively; 3) localization of Fe at the intersections of major reverse faults, the detachment and stratigraphically-higher, low-angle thrusts, which are preferentially developed at the contacts between diabase and banded iron formation. Collectively, these features both aided and compartmentalized regional, pervasive fluid flow along the detachment and in stratigraphically-higher sites. Subsequent ore preservation was enhanced by reactivation of NW- to NNW-trending, Mesozoic normal faults and concomitant downthrow of fault-bounded blocks. Structural relationships and timing of high-grade iron ore mineralization further suggest that associated top-to-N kinematic fold-and-thrust formation of the Mohlapitsi Belt and, subsequently, the main detachment between Malmani dolostone and banded iron formation of the Penge Formation, could have been active during and immediately after intrusion of the Bushveld Complex at ca. 2.05 Ga.

The Steenkampskraal Monazite Mine was first established in 1952, to extract monazite ore for the production of thorium and rare earth element (REE) concentrate. Refurbishment of the mine in recent years has required the re-inspection and re-evaluation of the mineralized monazite zone (MMZ). This contribution presents a structural review of the MMZ and its emplacement, based on recent data and its setting at the southern extent of the Bushmanland Sub-province of the Namaqua-Natal Metamorphic Belt. New surface and underground mapping confirm that the MMZ is a moderately-dipping body within gneissic host rocks on the southern limb of a broad F 3 antiform. Thickness variations, both down-dip and along-strike, are the result of D 2 and D 3 deformation. The MMZ has been locally transected and steepened by subsequent late-D 3 , “steep-structures”, which are typical of the Okiep copper district, ∼150 km north of Steenkampskraal. Geochronological data suggest that the MMZ was intruded, emplaced or formed at 1046 ± 7.5 Ma, at the start of the D 3 Klondikean Episode (1040–1020 Ma). Unlike the analogous copper-bearing Koperberg Suite in the Okiep Copper district, the MMZ was not intruded into Klondikean-aged steep structures, but was rather transected and steepened by these. Local steepening of the otherwise moderately-dipping to flat-lying MMZ makes it locally amenable to detection by soil sampling and radiometric surveys.

The Beit Bridge Complex of the Central Zone (CZ) of the Limpopo Belt hosts the 519 ± 6 Ma Venetia kimberlite diatremes. Deformed shelf- or platform-type supracrustal sequences include the Mount Dowe, Malala Drift and Gumbu Groups, comprising quartzofeldspathic units, biotite-bearing gneiss, quartzite, metapelite, metacalcsilicate and ortho- and para-amphibolite. Previous studies define tectonometamorphic events at 3.3–3.1 Ga, 2.7–2.5 Ga and 2.04 Ga. Detailed structural mapping over 10 years highlights four deformation events at Venetia. Rules-based implicit 3D modelling in Leapfrog Geo™ provides an unprecedented insight into CZ ductile deformation and sheath folding. D 1 juxtaposed gneisses against metasediments. D 2 produced a pervasive axial planar foliation (S 2 ) to isoclinal F 2 folds. Sheared lithological contacts and S 2 were refolded into regional, open, predominantly southward-verging, E–W trending F 3 folds. Intrusion of a hornblendite protolith occurred at high angles to incipient S 2 . Constrictional-prolate D 4 shows moderately NE-plunging azimuths defined by elongated hornblendite lenses, andalusite crystals in metapelite, crenulations in fuchsitic quartzite and sheath folding. D 4 overlaps with a: 1) 2.03–2.01 Ga regional M 3 metamorphic overprint; b) transpressional deformation at 2.2–1.9 Ga and c) 2.03 Ga transpressional, dextral shearing and thrusting around the CZ and d) formation of the Avoca, Bellavue and Baklykraal sheath folds and parallel lineations.

In the Domes Region of the Lufillian Belt, world-class stratiform Cu (±Co, ±Ni, ±U, ±Pb, ±Zn) deposits are closely associated with tectonized pre-existing basement unconformities, viz. the pre-Katanga-basement decollement (KD). In the KD footwall on the Mwombezhi Dome, the Lumwana Mine exploits mineralization within

This paper presents the results of a detailed and fully integrated pressure–temperature–time-deformation (P–T–t-D) study of the country rocks around Venetia Mine in the Central Zone of the Limpopo Belt, South Africa. Detailed structural mapping around Venetia Mine delineates four deformation events (D 1 –D 4 ). Relict S 1 comprises quartzofeldspathic bands in biotite gneiss and amphibolite. S 2 , defined by biotite in biotite schist, gneissic banding in biotite gneiss, long axes of quartzofeldspathic augen, and the long axes of amphibolite lenses/boudins, is axial planar to F 2 . F 3 occurs predominantly at the contacts between biotite gneiss and biotite schist, forming open to closed, upright to inclined, E-W- to ENE-trending, shallowly plunging folds. D 4 , which was constrictional-prolate in nature, refolded S 1 , S 2 , F 2 and F 3 . F 4 folds and an L 4 mineral lineation, defined predominantly by sillimanite in metapelitic schist, plunge moderately NE to NNE and overprint all previous fabrics. Poles to refolded S 2 foliations show a characteristic great circle distribution. In turn, the pole to this great circle coincides with the orientation of L 4 . Such patterns are also found in the Avoca, Bellevue and Ha-Tshansi sheath folds in the Central Zone, albeit that the plunge of the lineations at Venetia are predominantly NE-wards, rather than SW-wards. The moderately NE- to NNE-plunging F 4 sheath folds and L 4 lineations accord with the NE-SW trend of sheath folds and associated mineral elongation lineations observed elsewhere in the Central Zone. However, PbSL dating of syntectonic garnets constrain the minimum and maximum age of D 2 and D 4 structures respectively to c. 2037–2040 +/− 22 Ma. The metamorphism developed in metapelitic lithologies is characterized by peak amphibolite facies conditions of ∼6.5 kbar and 680 °C. Pseudosection modelling of growth zoning in garnet provides evidence of a prograde pressure and temperature increase from ∼5 kbar at 600 °C to ∼6–6.5 kbar at 650–680 °C, which is interpreted to be a consequence of tectonic thickening. Collectively, the integrated P–T–t-D data unequivocally demonstrates that the country rocks around Venetia experienced a structural–metamorphic event in the palaeoproterozoic. Coupled with published data we suggest a simple two-fold tectonic model, which involves the Central Zone as separate terrane docking with the Kaapvaal during the Neoarchean and later, during the Palaeoproterozoic this Central Zone-Kaapvaal amalgam collides with the Zimbabwe Craton.

The Merensky Reef of the Bushveld Complex occurs in its highest stratigraphic position as a heterogeneous, pegmatitic, feldspathic melanorite bounded by two narrow chromitite stringers at the base of the Merensky Cyclic Unit (MCU). In the Swartklip Facies of the Rustenburg Layered Suite, the occurrence of widespread thermal and mechanical erosion termed “potholing” has led to the subdivision of the Merensky Reef into Normal Reef and Regional Pothole Reef sub-facies. The transition between the two sub-facies occurs where the MCU transgresses the lower chromitite stringer of the Normal Merensky Reef and cuts down into the underlying cumulate lithologies. In the Regional Pothole Reef at the Northam Platinum Mine, several economic reef types are identified, where the Merensky Reef becomes conformable to cumulate layering, in particular, to the footwall marker (NP2 reef type) and the upper pseudoReef (P2 reef type). The Normal Merensky Reef, as well as the P2 and NP2 Reefs, contains economic platinum group element (PGE) grades and includes the lower portion of the MCU melanorite and the Merensky Chromitite. Whole rock geochemistry indicates that this package is compositionally identical in Normal, P2, and NP2 Reefs, suggesting that the base of the MCU is a relatively homogeneous drape over both Normal and Regional Pothole Reef regions. However, the lower sections of the three Reefs are variables depending on the depth of transgression of the MCU. In the Normal and P2 reef types, transgression by the MCU was arrested within harzburgites, melanorites, and norites, resulting in coarse, pegmatitic textures in the immediate footwall units. For the NP2 Reef, transgression by the MCU was arrested within leucocratic rocks and resulted in the formation of troctolites below the Merensky Chromitite. These troctolites are characterised by a coupled relationship between olivine and sulphides and by changes in major element chemistry and PGE contents relative to equivalent units in the footwall of the Normal Reef. Along with micro-textural relationships, these features suggest that troctolization of leucocratic cumulates in the NP2 Reef beneath the Merensky chromitite was a result of a reactive infiltration of a chromite-saturated melt and an immiscible sulphide liquid from the overlying MCU, rather than a significant fluid flux from below. In all reef types, the concentration of S defines symmetrical peaks centred on the Merensky Chromitite (and chromitites from pre-existing cyclic units in Normal and P2 Reefs), whereas PGE concentrations define asymmetrical peaks with higher PGE contents in reconstituted footwall rocks relative to the MCU melanorite. This signature is attributable to a magmatic model of PGE collection followed by deposition towards the base of the MCU and within reconstituted footwall rocks. The continuity of the asymmetrical magmatic PGE signature between the Normal Reef and Regional Pothole Reef sub-facies indicates that PGE mineralization inherent to the Merensky magma occurred as a drape over a variably eroded and subsequent texturally and geochemically reworked or reconstituted footwall.

Syn-magmatic removal of the cumulate pile during the formation of the Bushveld Complex resulted in “potholes”. Erosion progressed downward in the cumulate pile, resulting in a series of steep, transgressive contacts between locally conformable potholed reefs in the regional pothole sub-facies of the Swartklip Facies in the western limb of the Bushveld Complex. The deepest of these potholes, “third-order” or “FWP2” potholing, occurs where the base of the Merensky Cyclic Unit transgresses the Upper Pseudo-Reef Chromitite marker horizon. The base of a FWP2 pothole on Northam Platinum Mine consists of an unconformable stringer Merensky Chromitite overlain by a medium-grained, poikilitic orthopyroxenite and underlain by either a pegmatitic harzburgite or the medium-grained Lower Pseudo-Reef Anorthosite. Detailed shape and distribution analysis of FWP2 potholes reveals underlying patterns in their shape and distribution which, in turn, suggest a structural control. The ratio between pothole short vs long axes is 0.624 (N=1,385), although the ratio increases from 0.48 to 0.61 in the long axis range 10 to 60 m, then decreases from 0.61 to 0.57 from 61 to 100 m, increasing again from 0.57 to 0.61 from 101 to 400 m, suggesting that there is not a simple relationship between pothole shape and size. Shape (circularity, eccentricity, and dendricity) analysis of a subset of 638 potholes indicates that potholes with long axes 100 m have an average normalized shape that is bilobate and elongate on a 120° orientation. The average aspect ratio (short axis length divided by long axis length) of potholes is highest for potholes with long axis lengths >100 m and lowest for potholes with long axis lengths between 35 and 60 m. The most common long axis orientation for potholes with long axis lengths 100 m. Fractal analysis indicates that the distribution of pothole centers is controlled neither by a single nor several interacting fractal dimensions. Autocorrelation (Fry) analysis of the distribution of pothole centers shows recurring pothole distribution trends at 038, 070, and 110° for potholes over the full range of long axis lengths, while the trends of 008 and 152° occur in potholes with long axes lengths between 60 and 100 m. Chi-squared (X2) analysis of the locations of pothole centers suggests that the distribution of small potholes is highly non-uniform but becomes exponentially more uniform with increasing pothole size. The model which best fits the observed shape and distribution analysis is a combination of protracted independent growth and “nearest neighbor” merging along specific orientations. For instance, the clustered distribution of original pothole centers resulted in merged potholes with long axes lengths of up to 60 m, exhibiting short vs long axes ratios of 0.61, preferred orientations of 150°, and alignment along 010 and 150° trends. Further independent growth allowed for merging of similar-sized (and smaller) neighboring potholes, generating potholes with long axes of up to 100 m in length, a preferred long axis orientation of 150°, and alignment along 010, 040, 075, and 150°. Subsequent preferential merging occurred along a 120° trend, thereby preserving a bilobate form. This implies that while pothole initiation and enlargement may be driven by a “top-down” (i.e., possibly thermomechanical) process, an underlying linear or structural catalyst/control is revealed in changes in pothole shape during enlargement and, furthermore, in the preferred trends along which potholes merged over a considerable period, possibly concomitant with adjustment of major structures in the footwall to the Bushveld Complex and pulses into the magma chamber.

The Mesoproterozoic Natal Metamorphic Province of Kwazulu–Natal in South Africa is an assemblage of several tectonic units, including accreted oceanic island arcs, obducted oceanic crust and deformed basin material. The highly deformed Mfongosi Group occurs at the leading edge of collision (the Natal thrust front), against and directly overlying the southern margin of the Kaapvaal Craton. Structures within the Mfongosi Group record “local” D 1 and D 2 events, the first of which was “oblique obduction”, with predominantly N- to NNE-verging thrusting ( D 1 ). This was followed by sinistral transpression combined with vertical constriction, forming SW-plunging kink folds and SW-plunging prolate pillow basalts ( D 2 ). The third and final event ( D 3 ) was E–W to ESE–WNW extension in a post-thrusting phase, defined by fibrous antitaxial quartz–calcite veining. The westernmost portion of the Mfongosi Group, the Ngubevu area, shows significantly higher finite strains (up to R f = 12) compared to central Mfongosi and eastern Nkandlha areas ( R f = 1.5 and less), suggesting highly oblique, largely NE-directed initial collision. Deformation of the NTF in the context of nappe emplacement is constrained by 40 Ar/ 39 Ar dating of post-cataclastic nematoblastic/porphyroblastic hornblende of the Manyane amphibolite close to the thrust between the Tugela nappe and the Mfongosi Group in the Mfongosi area. Hornblende overgrew the products of low-temperature deformation during the “local” D 1 and D 2 . A minimum age of 1171 ± 16 Ma (95% conf., including J -error; weighted by √MSWD; MSWD = 4.3) is obtained for the tectonic juxtaposition of the Tugela nappe against the southern portions of the “Mfongosi Basin”. This “local” D 1 and D 2 of the Mfongosi Group pre-dates the regional “oblique D 1 ” and “left-lateral D 2 ” previously determined for the central and southern terranes of the Natal Metamorphic Province by other researchers. Comparison of the 1171 ± 16 Ma age, with ages for shearing and intrusion, suggests that thrusting and/or mylonite-forming events migrated southwards throughout the Natal Metamorphic Province, being separated by approximately 25 million years. Thrusting and/or mylonite-forming events occurred in the nappe zone from ca. 1135 Ma to 1077 Ma, followed by a period of “quiescence” during which granites intruded, in turn followed by late-tectonic deformation of the southern Mzumbe and Margate terranes from ca. 1004 Ma to 970 Ma. Such a scenario supports previously-proposed indentation models with their implications of oblique convergence and late-tectonic escape of island arcs to the E/ENE (African azimuths).

Regional and local structural controls on the emplacement of 1326 Southern African kimberlites and related rocks (kimberlites sensu lato, 11% of which are dated) are analysed using a framework of lineaments defined by combining geology, aeromagnetics, gravity and geomorphological data. Spatial analysis of occurrences within clusters of kimberlites less than 100km across resolves variable trends, depending on the age and position of the cluster; but on a regional scale the distribution of these clusters is statistically controlled by four lineament trends: 040°, 096°, 134° and 165°. Similar regional trends are observed as aspect lineaments that can be followed over large distances from modelling the variation in dip direction of the Southern African topography. These observations suggest that different geological parameters exert a control on the distribution of kimberlites. Local structures may include en-echelon fault arrays, Riedel, R’-, P- or T-structures within trans-continental lithosphere structures (cryptic continental corridors). Many cryptic continental corridors are collinear with fracture zones along the Atlantic and Indian continental margins of Southern Africa, and may have found their origin in events resulting from plate reorganization during the break-up of the supercontinent Gondwana. Fault resistance may have rapidly changed the stress state of the African continent causing the deep lithospheric faults to be the loci of episodic extension, allowing kimberlite fluids to ascend through the faults and cluster within near-surface structures. A progressive age variation of kimberlite magmatism in Southern Africa may be attributed to stress propagation along deep lithospheric fractures.

The Rossing granite-hosted uranium deposit in the Central Zone of the Pan-African Damara Orogen, Namibia, is situated in the “SJ area” to the south of the Rossing Dome. The coincidence of a number of features in this area suggests that mineralization is closely linked to late-kinematic evolution of the Rossing Dome. These features include: (1) the rotation of the dome's long axis (trend of 017°), relative to the regional F 3 trend of 042°; (2) southward dome impingement, concomitant with dome rotation, producing a wedge-shaped zone of alkali-leucogranites, within which uranium mineralization is transgressive with respect to granites and their host lithologies; uranium mineralization and a high fluid flux are also confined to this arcuate zone to the south and south-east of the dome core and (3) fault modeling that indicates that the SJ area underwent late-D 3 to D 4 brittle–ductile deformation, producing a dense fault network that was exploited by leucogranites. Dome rotation and southward impingement occurred after a protracted period of transtensional tectonism in the Central Zone, from ca. 542 to 526 Ma, during which I- and S-type granites were initiated in a metamorphic core complex. Late-kinematic deformation involved a rejuvenation of the stresses that acted from ca. 600 to 550 Ma. This deformation overlapped with uranium-enriched granite intrusion in the Central Zone at 510 ± 3 Ma. Such late-kinematic, north–south transpression, which persisted into the post-kinematic cooling phase until at least 478 ± 4 Ma, was synchronous with left-lateral displacement along NNE-trending (“Welwitschia Trend”) shears in the vicinity of Rossing. Late-kinematic deformation, causing block rotation, overlying dome rotation and interaction of the more competent units of the Khan Formation with the Rossing Formation in the dome rim was pivotal in the localization of uranium-enriched granites within a highly fractured, high-strain zone that was also the site of prolonged/high fluid flux.

The Normal Reef Subfacies of the Swartklip facies of the Merensky Reef along the western limb of the Rustenburg Layered Suite, exposed on Northam Platinum mine, displays lateral and vertical variations in its inter-chromitite lithologies, content in platinum-group elements and grade, and the mode of sulfide occurrence. Reef greater than 300 cm in thickness consists of a footwall of mottled anorthosite, overlain by a granular orthocumulate chromitite layer (the lower chromitite of the reef), followed by a pegmatitic feldspathic dunite, a 5- to 10-cm-thick massive, plagioclase-poor dunite, a heterogeneous pegmatitic harzburgite, succeeded by a coarse-grained pegmatitic pyroxenite, collectively comprising a pre-Merensky cyclic unit. The upper pegmatitic pyroxenite is in turn overlain by the granular (upper) chromitite of the reef, occurring at the base of the Merensky Cyclic Unit, which is overlain by a homogeneous poikilitic orthopyroxenite. Disseminated and clustered pentlandite, pyrrhotite and chalcopyrite occur throughout the inter-chromitite pegmatite, but are not observed in the massive, plagioclase-poor dunite, which typically occurs approximately 100 cm above the lower chromitite. Thermal recrystallization and downward migration of a reconstitution front and mechanical erosion of the medium-grained heterogeneous pegmatitic harzburgite accompanied progressive thinning of the reef from an inter-chromitite reef greater than 300 cm in thickness. Breaching of the massive, relatively impermeable layer of massive dunite resulted in pronounced or accelerated thermomechanical erosion of the underlying pegmatitic dunite. Peaks in platinum-group-element content, situated just below the upper chromitite and just above the lower chromitite, merge with reef thinning, becoming indistinct in reef that is less than 160 cm thick. With further thinning, until the point where the two chromitites are juxtaposed or merged, the single peak in PGE content “migrates” into the poikilitic orthopyroxenite hanging-wall. A sulfide lag, consisting of up to 20% sulfide by volume, occurs in reef less than 160 cm in thickness. Its presence is especially evident near the point where the Regional Pothole subfacies becomes manifest, and the upper chromitite or base of the Merensky Cyclic Unit transgresses the lower chromitite. Spatial distributions and graphical correlations of sampling data, related to reef thickness, indicate that there is a strong positive correlation between ∑PGE grams in the total reef sampled and the inter-reef pegmatite. A strong inverse relationship, relative to thinning, is observed between ∑PGE reef grade and ∑PGE grams. The ∑PGE grams content of the upper or Merensky chromitite of the reef is remarkably consistent over the extent of Normal Reef of the Merensky cyclic unit exposed at Northam, indicating that it is a drape or post-erosional crystal cumulate over an eroded and recrystallized pre-Merensky cyclic unit.

The whole-rock geochemistry of metamorphosed greywackes, arenites and arkoses within the Mesoproterozoic Namaqua-Natal-Maudheim Province is interpreted with the aim of establishing geochemical correlations and defining common sediment source terrains. Metasediments of the Mfongosi Group of the Natal Sector of the Namaqua-Natal Metamorphic Province were sampled from their type area in the Mfongosi Valley. Metagreywackes from the northern limits of the Mfongosi Valley, directly adjacent to the Kaapvaal Craton, show ocean island arc signatures while metagreywackes from the southern limits of the Mfongosi Valley, near the contact with the Madidima Thrust of the Natal nappe zone, show mainly active continental margin signatures. Interleaved, geochemically distinct low-Ca+Na, high-K metamorphosed arkoses to lithic arkoses indicate a minor passive margin sediment component. Geochemical classification of low-grade Ahlmannryggen Group greywackes, arenites and arkoses of the Grunehogna Province, Antarctica, indicates both active and passive continental margin sediment sources. An oceanic island arc signature is not evident in Ahlmannryggen Group data. The active continental margin signature in both Natal Sector and Grunehogna Province metasediments potentially provides for a common link between these terranes. Discriminant Function Analysis, using three pre-defined provenance sub-sets within the Mfongosi Group and two pre-defined provenance sub-sets within the Ahlmannryggen Group, indicate that metasediments with active continental margin signatures from both groups are geochemically identical, implying that the active continental margin of the Grunehogna Province shed immature sediments westwards (African azimuths) into the developing, narrow or restricted Mesoproterozoic ‘Mfongosi Basin.’ This was accompanied by minor sediment influx from a stable continental platform, potentially the Kaapvaal Craton. Oblique and diachronous collision, initiated in the southwestern portions of the combined Natal Sector/Grunehogna Province system produced a laterally variable Mfongosi Group, which formed in the ‘Mfongosi Basin’. Coarse-grained sediments dominated in its eastern portions while basalts with thin sapropelite units dominated in its western portions.

Discordant veins, pipes and occasionally subconcordant sheets of iron-rich ultramafic pegmatite disrupt the layered cumulate sequence of the Upper Critical Zone, Rustenburg Layered Suite, Bushveld Complex. These pegmatite bodies have been studied where they replace the Merensky Reef footwall at Northam Platinum Mine, situated in the Swartklip Facies of the western lobe of the Rustenburg Layered Suite. Composed chiefly of ferroaugite and fayalitic olivine, the pegmatites appear to be formed by the preferential replacement of plagioclase-rich cumulates within the layered sequence. Fe-Ti oxides, sulphide (pyrrhotite and chalcopyrite) and plagioclase also occur in variable quantities. Differentiation within the pegmatite is observed where it has spread laterally beneath the impervious Merensky chromitite layer, with the development of subparallel cm-scale layers of massive magnetitite, massive sulphide and sulphide pegmatite. While some Fe-rich mobile phase must have been responsible for the pegmatites, it is concluded that the pegmatite bulk composition does not represent the original liquid. Furthermore the mode of occurrence precludes the injection of a crystal mush. Rather it is argued, mainlyon geochemical and isotopic grounds, that Fe-rich residual melts derived from the Upper Zone in the downward crosscutting gap areas migrated laterallyand upwards into the adjacent Upper Critical Zone. Variable reaction with the layered cumulates produced the anastomosing pegmatite bodies.

Calcite veins are invariably associated with en-echelon kimberlite dyke-fracture arrays. A detailed microstructural study of veining indicates four vein types. Type I stretched or ataxial veins are defined by high aspect ratio calcite fibers that are crystallographically continuous with calcite of the kimberlite matrix wall rock, by elongated phenocrystic phlogopite with sharp crystal terminations centered on contacts between adjacent calcite fibers and by phenocrystic phlogopite that grows or extends across these veins. Type I vein mineralogy indicates syn-dilational crystallization of vein minerals in local tensional areas within the kimberlite. Vein Types II (stretched to syntaxial elongate-blocky) and III (antitaxial) indicate late crystallization vein mineral growth during subsequent or repeated dilation. Calcite fibers in Type I to Type III veins are orthogonal to the contacts of their host dykes regardless of the orientation of vein margins. Type IV calcite veins, with blocky or mosaic/polycrystalline textures, are attributed to minor post-intrusion extension, which was potentially accompanied by repeated kimberlite intrusion within a given dyke array. Syn-crystallization/syn-intrusion Type I veins and an ubiquitous dyke-parallel fracture cleavage, in a zone up to 4 m on either side of dyke contacts, suggest that en-echelon kimberlite dyke-fracture arrays occupied the approximate center of zones of active dilation within the brittle carapace of the upper crust. Type II and III veins indicate that extension or dilation continued, independently of an occupying kimberlite fluid phase, after initial intrusion. Arrested mobile hydrofracturing, under low differential stress within the upper brittle or seismic carapace of the continental crust, followed by repeated dilation of the dyke-fracture system, is proposed as a mechanism for producing the features observed in this study. The conditions constrained in this study indicate passive dyke intrusion into dilating fracture arrays during crustal extension. © 2004 Elsevier B.V. All rights reserved.

Group II kimberlite dykes occur in small, dominantly en-echelon dyke-fracture arrays, with individual dyke-fractures showing small angular variations from their array trends (5° to 15°). The analysed dyke systems are characterized by closely matching opposing dyke contacts, “ in-situ ” breccia, multiple kimberlite stringers within a dilated dyke-parallel fracture cleavage, wedge-shaped apophyses in bent bridges at dyke-fracture offsets/overlaps, kimberlite-free offset/overlap areas and calcite vein fibres orthogonal to dyke contacts. Commonly found microscopic structures include synemplacement/syncrystallization calcite veinlets, containing high aspect ratio stretched fibrous calcite, and elongate phlogopite phenocrysts and serpentinized olivine phenocrysts growing across the width of these veins. Both macro- and microscopic structures support a model of orthogonal host rock dilation during kimberlite emplacement. Terminations of dyke-fracture segments show minimal curvature or overlap, suggesting that remote horizontal stresses dominated during their emplacement (“passive” intrusion), as opposed to magma overpressured systems wherein dyke or dyke-fracture overlaps curve strongly towards each other (“active” intrusion). The application of Mohr diagrams suggests that low differential stresses, with no or only a very minor shear component, prevailed at the time of emplacement. The dominance of remote horizontal forces, imparting small differential stresses to the brittle portions of the crust, a closely-spaced, dilating dyke-parallel fracture cleavage ahead of the dyke tip (imparting a local suction) and the low-volume, low-viscosity, highly volatile nature of kimberlitic magmas may explain their empirically-constrained high emplacement velocities. This, in turn, explains the means by which such magmas may entrain significant volumes of high specific-gravity mantle material. Mobile hydrofracturing in the fringe zones around dilated craton-scale jointing is proposed to be a viable mechanism for kimberlite emplacement.

At Northam Platinum Mine, in the Western Limb of the Rustenburg Layered Suite of the Bushveld Complex, South Africa, a regional-scale (km-scale) excursion, or “pothole” structure, commences at the level of the Merensky Reef (Normal Reef sub-facies) and transgresses underlying cumulate lithologies as part of the Regional Pothole Reef sub-facies. At several discrete levels within this sub-facies, the Merensky Cyclic Unit (MCU) becomes conformable to cumulate layering and results in four distinctive Regional Pothole reef types, namely the NP2, P2, FWP2 and FWP1 Reefs. The immediate footwall (