doi:10.1016/j.jafrearsci.2006.01.016
Copyright © 2006 Elsevier Ltd All rights reserved.
Mesoproterozoic intraplate magmatism in the Kalahari Craton: A review
R.E. Hansona,
,
, R.E. Harmerb, 1, T.G. Blenkinsopc, D.S. Bullend, I.W.D. Dalziele, W.A. Gosee, R.P. Halld, A.B. Kampunzuf,
, R.M. Keyg, J. Mukwakwamih, H. Munyanyiwah, 
, J.A. Pancakea, 2, E.K. Seidela, 3 and S.E. Wardd
aDepartment of Geology, Texas Christian University, Fort Worth, TX 76129, USA
bCouncil for Geoscience, Pretoria 0001, South Africa
cDepartment of Earth Science, James Cook University, Townsville, QLD4811, Australia
dSchool of Earth, Environmental and Physical Sciences, University of Portsmouth, Portsmouth, UK
eDepartment of Geological Sciences and Institute for Geophysics, University of Texas, Austin, TX, USA
fDepartment of Geology, University of Botswana, Gaborone, Botswana
gBritish Geological Survey, Murchison House, West Mains Road, Edinburgh EH9 3LA, UK
hDepartment of Geology, University of Zimbabwe, Harare, Zimbabwe
Received 15 February 2003;
accepted 15 January 2006.
Available online 17 July 2006.
References and further reading may be available for this article. To view references and further reading you must
purchase this article.
Abstract
The Kalahari Craton was initially stabilized following cessation of Palaeoproterozoic orogenesis in southern Africa at ca. 1.8 Ga. Subsequent Mesoproterozoic intraplate magmatism at ca. 1.4–1.35 Ga formed a series of alkaline and carbonatitic complexes in the southern part of the craton. Original volcanic structures are partly preserved in some of the complexes, and a variety of intrusive rocks (e.g., quartz syenite, nepheline syenite, pyroxenite, ijolite, carbonatite) are present. The Premier kimberlite cluster was emplaced in the same region at ca. 1.2 Ga, but available geochronology indicates that the main alkaline magmatism preceded 1.2–1.0 Ga orogenesis in the Namaqua–Natal–Maud Belt along the southern craton margin. Another, more extensive intraplate magmatic event at ca. 1.1 Ga formed the Umkondo Igneous Province, which is recognized over an area of
2.0 × 106 km2 on the Kalahari Craton, including a detached fragment now located in Antarctica. Much of the province comprises high-level mafic intrusions, but erosional remnants of basalt lava piles and bimodal basalt/rhyolite assemblages are also present. Most of the mafic rocks are continental tholeiites, but trace-element geochemistry reveals distinct subgroups that cannot be related by crustal-level assimilation/fractional crystallization processes or by partial melting of a uniform mantle source. Geochronological and palaeomagnetic data indicate that enormous volumes of tholeiitic magma were emplaced within the province in a narrow time frame at ca. 1112–1106 Ma, which is inferred to record uprise of a mantle plume behind the Namaqua–Natal–Maud Belt.
 |
Fig. 1. Mesoproterozoic intraplate magmatic rocks in the Kalahari Craton in southern Africa and Dronning Maud Land, Antarctica (restored to pre-Mesozoic position after Reeves et al., 2002). Present-day coordinates are shown for Africa. Covered tectonic boundaries: 1: edge of Archaean cratonic basement; 2: boundary between Palaeoproterozoic terrains and Northwest Botswana Rift; 3: boundaries between Mesoproterozoic and Pan-African orogens. Alkaline and carbonatitic complexes: BR: Bull’s Run; CR: Crocodile River complexes; G: Glenover; Go: Goudini; M: Martin’s Drift kimberlite cluster; P: Premier kimberlite cluster; Pi: Pilanesberg; PR: Pienaars River complexes; Sp: Spitskop; St: Stukpan. Inferred extent of Umkondo Igneous Province shown by bold dashed lines. Known or probable members of Umkondo Province: B: dolerite at Barnardskop; De: Deweras and associated dykes in Guruve area; K: Kamativi dykes; Ka: Kartatsaus, Langberg and Skumok rhyolites and Opdam basalts; Mu: Mutare dykes; OP: Oorlogsende Porphyry; R: Rakops Dyke; Ro: rhyolite at Rostock massifs; TiG: Timbavati Gabbro; UD: dolerite dyke correlated with Umkondo dolerites by Wilson et al., 1987 and Allsopp et al., 1989b; V: dolerite intruding Vredefort Dome. Deweras and associated dykes and Kamativi and Mutare dyke suites are shown schematically; all three dyke suites may contain dykes unrelated to Umkondo Province, as discussed in text. Other labeled features: A: Amersfoort; DI: Dete Inlier; KC: Archaean Kaapvaal Craton; KH: Kgwebe Hills; Ko: Koras Group; ODS: Okavango Dyke Swarm; OI: Okwa Inlier; ORDS: Olifants River Dyke Swarm; Ph: Phoenix leucogranites; ZC: Archaean Zimbabwe Craton. Locations of boreholes CKP-8C in Tshane Complex, CKP-6 and 6A in Xade Complex, CKP-4 in Kgwebe Formation and CKP-10A in subsurface gabbro farther west are also shown. Modified from Hanson et al. (1998) and references therein. Distribution of Umkondo mafic intrusions from Mortimer, 1984, Zimbabwe Geological Survey, 1994 and Keyser, 1997.
Fig. 2. Geological map of Pilanesberg Complex, modified from Lurie, 1973 and Lurie and Cawthorn, 1984. See Fig. 1 for location. Tick marks indicate location of N–S cross section.
Fig. 3. Pilanesberg Dyke Swarm in South Africa and adjacent part of Botswana, modified from Emerman (1991). Mesoproterozoic alkaline and carbonatitic complexes are also shown. B: Buffelskraal; N: Nooitgedacht; K: Kruidfontein; T: Tweerivier; B/W: Bulhoekkop, Bulhoek South and Welgevonden; E: Elandskraal; R: Roodeplaat; L: Leeuwfontein; D: Derdepoort; F: Franspoort.
Fig. 4. Alkaline and carbonatitic complexes in and adjacent to Crocodile River Fragment. See Fig. 1 and Fig. 3 for location.
Fig. 5. Pienaars River alkaline and carbonatitic complexes, from Harmer, 1985 and Frick and Walraven, 1985. See Fig. 1 for location.
Fig. 6. Geological map of Spitskop Complex, from Harmer (1999). See Fig. 1 for location.
Fig. 7. Palaeomagnetic poles for various members of Umkondo Igneous Province. Poles are labeled in (a) (symbols as in Table 3). Circles or error ellipses of 95% confidence for poles are shown in (b). Specific information and source for each pole are given in Table 3.
Fig. 8. Geological map of the Umkondo Group and associated dolerite sills in eastern Zimbabwe and adjacent parts of Mozambique, modified from Pinna and Marteau, 1987 and Zimbabwe Geological Survey, 1994. See Fig. 1 for location.
Fig. 9. Geological map of part of southeastern Botswana showing dolerite intrusions related to Umkondo Province. Locations of sample sites with Umkondo-type palaeomagnetic poles are indicated. Modified from Mortimer (1984).
Fig. 10. (a) Th/Yb versus Nb/Yb plot for Guruve and Mutare dyke samples. Type I dykes plot close to enriched mantle, whereas type II dykes show more depleted characteristics. Type III dykes show a subduction-related influence with a component of crustal contamination. Vectors show influence of subduction components (S), crustal contamination (C), within-plate mantle enrichment (W) and fractional crystallization (F). After Pearce (1983). (b) Th/Nb versus La/Yb plot for Guruve and Mutare dyke samples relative to partial melting curves for depleted MORB mantle (DMM) and bulk silicate earth (BSE) in undepleted spinel (US), depleted spinel (DS), undepleted garnet (UG) and depleted garnet (DG) lherzolite facies. Compositions of E-MORB, upper crust and ocean-island basalt (OIB) from Sun and McDonough (1989). BSE and DMM compositions from Kostopolous and James (1992). Black arrows show possible vectors for upper crustal contamination.
Fig. 11. Geochemical data for Umkondo mafic intrusions in Botswana and South Africa. (a) Zr/TiO2 versus Nb/Y diagram of Winchester and Floyd (1977). (b) Ti versus Zr diagram of Pearce and Cann (1973). (c) Ti/Zr versus MgO. (d) SiO2 versus MgO. Note that three subgroups occur in the Botswana data set and two in the South Africa data set. Botswana subgroups are outlined by dashed boundaries, South Africa subgroups by solid lines.
Table 1.
Geochronological data for Mesoproterozoic alkaline and carbonatitic complexes

Table 2.
Geochronological data for the Premier kimberlite cluster

Table 3.
Palaeomagnetic poles for the Umkondo Province

Lat. and Long.: latitude and longitude of pole position transferred to northern hemisphere. A95: radius in degrees of 95% circle of confidence or semiaxes of error ellipse. Of the 10 sites used by McElhinny and Opdyke (1964), one pole deviates from the mean by 2.3 angular standard deviations and has been deleted from the calculations. Antarctic poles have been rotated to Africa using Euler pole of Reeves et al. (2002). See Gose et al. (2006) for complete discussion of data.
Table 4.
Geochronological data for known and probable members of the Umkondo Province
