Formation of the Archean Kaapvaal Province revisited: Implications for the birth and growth of its diamondiferous root H. Helmstaedt1 and J.J. Gurney2 1: Department of Geological Sciences and Geological Engineering, Queen’s University, Kingston, Ont., Canada K7L 3N6 2: Department of Geological Science, University of Cape Town, 7700 Rondebosch, South Africa The Kaapvaal Province of the Archean Kalahari Craton differs from other African Archean terrains by its early cratonization, representing a nearly 1 mill. km2 area much of which was stabilized before 3 Ga (de Wit et al., 1992). With the exception of the northern margin, the boundaries of the province are well defined. The western boundary is hidden under the thin-skinned Paleoproterozoic Kheis belt but is well marked by a N-S gravimetric and airomagnetic linear, referred to as the Kalahari line. The southwestern boundary is defined by two northwesterly trending faults, the Doornberg and Brakbos fault zones, which delimit a ca. 40 km wide boundary zone, the 3.0-2.6 Ga Marydale domain, between the Kaapvaal Province and the Mesoproterozoic Namaqua belt. The change in mantle-root signature across this on-craton to off-craton transition has been described recently by Grutter et al. (1999). Most of the southern boundary with the Namaqua-Natal province is covered by the Carboniferous to Jurassic Karoo sequence, but it is reasonably well defined by its geophysical expression. It is exposed in Natal as a northerly-verging thrust. The northerly-trending eastern boundary runs along the Mozambique boder, where it is also obscured by a thick Karoo cover. The northern boundary is generally thought to coincide with the southern border of the Limpopo Province, however, Cheney and Winter (1995) draw it along the Zoetfontein-Palala fault. As none of the Transvaal and older cover sequences of the Kaapvaal Province are known to the north of this fault zone, they regard the crystalline basement to the north as parts of the Zimbabwe Province. Early Craton assembly As described by de Wit et al. (1992) and Poujol and Anhaeusser (2001), the Kaapvaal Province is not a homogeneous Archean province but consists of a number of subprovinces, referred to as tectonostratigraphic terrains, at least some of which appear to represent amalgamated Archean tectonic “terranes” in the modern plate-tectonic sense. The core region, in the eastern part of the province, is the Ancient Gneiss terrain with the tectonically juxtaposed Barberton greenstone belt. The Ancient Gneiss terrain, best studied in the Swaziland terrain, contains tonalitic intrusions as old as 3.644 Ga. The southern Barberton greenstone terrain (Onverwacht Group), dated as ca. 3.5 Ga, has been compared to parts of Phanerozoic ophiolites and is referred to as Jamestown ophiolite complex. Between 3.46 and 3.43 Ga it was thrust onto an active arc-like terrain with a 3.54 Ga trondjhemitic-tonalitic basement, and between 3.3 and 3.2 Ga it was juxtaposed, together with the northern Barberton greenstone terrain, against the Ancient Gneiss terrain. All three terrains were intruded at that time by geochemically similar

trondhjemitic plutons, and between 3.15 to 3.07 Ga regionally extensive sheets of highlevel granites were emplaced across the boundaries between the three terrains. The ancient core region was enlarged to the south by the 3.4 to 3.2 Ga Natal granite-greenstone terrain. Widespread tonalitic magmatism within this terrain at 3.3 to 3.2 Ga corresponds to that of the second trondhjemitic-tonalitic intrusive episode in the ancient core region to the north. Greenstone terrains to the north of the ancient core, including the Murchison, Giyani (Sutherland), and Pietersburg greenstone belts, and possibly also the high-grade metamorphic greenstone remnants in the southern marginal zone of the Limpopo Province, are thought to contain at least some rocks as old as those of the Barberton greenstone terrain (Poujol and Anhaeusser, 2001). The precise mode of their amalgamation with one another, and with the ancient core to the south, is still uncertain, but they were firmly welded to the core of the Kaapvaal Province by 3.1 Ga. Although they are covered by Late Archean and younger strata, the east-west trending terrains of the ancient core are thought to continue westwards at least as far as to the Vredefort structure, where crust-lithosphere coupling is purported to have occurred at 3.08 Ga (Moser et al., 2001). They are probably truncated along the north-south trending Colesberg lineament, west of which are the more or less north-south trending Colesberg, Amalia and Kraaipan granite-greenstone terrains that have yielded mainly younger Archean ages between about 3.01 and 2.79 Ga (Poujol et al., 2000). The Kraaipan greenstone association, occurring in highly deformed gneisses and migmatites, extends northwards into Botswana, where it is intruded by the 2830 Ma Gaborone granite. Eastwest shortening related to the amalgamation of the northerly trending terrains west of the Colesburg lineament is dated in Botswana as between 2840 and 2720 Ma (Carney et al., 1994). Of great interest, with regard to the stabilization of this western part of the Kaapval Province is the observation that even the oldest ages of the crustal domains here (ca. 3 Ga) are significantly younger than the model ages of the harzburgitic garnet inclusions of diamonds from the Kimberley and Finsch kimberlites (ca. 3.2 Ga and older; Richardson et al., 1984), both located in this part of the province. This suggests that the lithospheric root of the older, eastern part of the province may extend under the younger crust of the western part, implying tectonic accretion of younger over older. Late Archean history Following the stabilization of the granite-greenstone terrains, the Kaapvaal craton acted as basement for a series of unconformity-bounded, largely sedimentary sequences, including the Pongola, Dominion, Witwatersrand, Ventersdorp, Transvaal, Soutpansberg and Waterberg sequences. These were reviewed by Cheney and Winter (1995), applying sequence stratigraphy and tying the evolution of the sequences into the Late Archean to Mesoproterozoic tectonic history of the Kaapvaal Province. These authors quantified what had been noted by many previous workers, that the oldest sequences, including the Dominion, West and Central Rand (R1,R2), and lower Ventersdorp (V1), are confined to the southeastern part of the Kaapvaal Province. All of them are older then 2.714 Ga. The Middle Ventersdorp (V2), which is older than 2.709 Ga, covers part of the northern and western part of the province, however, only the Upper Ventersdorp (which is post 2.709

and pre 2.552 Ga) covers much of the western part and is the first “Proterozoic-like” sequence that overlaps the boundary between the younger western and older eastern parts of the Kaapvaal Province. The deformation history of these sedimentary sequences suggests that basement rocks successively overthrust, first the Dominion and Rand sequences, and then the Lower and Middle Ventersdorp sequences, with the Upper Ventersdorp sequence unconformably overlapping the earlier east-verging thrusts. The evidence from the sedimentary record is thus compatible with tectonic accretion of the western over eastern Kaapvaal basement, inferred on the basis of the older diamondiferous root under the younger western part of the province. Such a scenario implies that the Colesberg lineament is not a transcurrent fault (e.g. de Wit et al., 1992), but essentially a thrust front, and at least part of the Upper Witwatersrand basin (Central Rand sequence, R2) is a foreland compressional basin, not to the Limpopo Belt, as commonly assumed, but to the easterly directed Colesberg front. In fact, if the timing of Central Rand deposition (between 2.84-2.71 Ga) presently accepted in the literature (e.g. de Wit et al., 1992; Cheney and Winter, 1995) is correct, deposition overlaps with the 2.9-2.7 Ga, northerly-directed overthrusting of the Kaapvaal Province over the Limpopo Province, putting the Witwatersrand basin in the hinterland of this thrust belt. The westerly accretion model would also explain the westerly provenance of much of the Witwatersrand sediments, that represents a problem in the Limpopo foreland basin model (de Wit et al., 1992). Well-known for their rich detrital gold and uranium deposits, the rocks of the Central Rand sequence (R2) have consistently yielded small amounts of detrital diamonds and kimberlite indicator minerals (Raal, 1969; Raal and Robinson, 1980). This is of great geological interest not only because these diamonds represent the first well-documented Archean diamonds, but also because the occurrence of indicator minerals allows the inference of an igneous (kimberlitic?) source of Archean age. It is likely that these earliest inferred kimberlites intruded during the 3.1 to 2.9 Ga extensional basin interval represented by the Dominion (D) and Westrand (R1) sequences. They would have tapped the earliest root with ca. 3.2 Ga P-type diamonds (Richardson et al., 1984), but they may have predated the E-type diamonds accreted to this root at ca. 2.9 Ga (Richardson et al., 2001; Shirey et al., 2001). It is suggested that Witwatersrand diamonds be investigated to test this model. The first major extensional event affecting the combined eastern and western Kaapvaal Province, but only to the south of the Thabazimbi-Murchison lineament, is represented by the Upper Ventersdorp sequence (V3), consisting of clastic and volcanic rocks that were deposited between 2.7 and ca. 2.6 Ga in a system of NE-trending grabens at a high angle to the EW-trending Limpopo Province. The formation of these grabens is attributed by some authors to oblique convergence between the Kaapvaal and Zimbabwe provinces (Clendenin et al., 1988) along the Limpopo Province. Based mainly on geochemical similarities between the approx. 2.7 Ga Ventersdorp and Fortescue volcanic-sedimentary sequences, others have speculated that the Kaapvaal Province and the Pilbara Province of northwestern Australia may have been part of an early supercontinent “Vaalbara” (e.g., Nelson et al., 1992) that fragmented at about 2.7 Ga. Evidence for the location of the rifted margin is not preserved in the Kaapvaal Province, but it is possible that the

Ventersdorp graben system represents a failed arm of a triple junction. Whatever its origin, the Ventersdorp event caused only limited intracratonic rifting that did not destroy the diamondiferous root under the province. White et al. (1995) have attached great significance to the fact that the Ventersdorp basin broadly coincides with numerous kimberlites occurring in a wide NE trending belt between Pretoria and Sutherland, the implication being that the deep-seated, basin-controlling faults provided channelways for several generations of kimberlites. References: Cheney, E. S., and Winter, H. d. l. R., 1995, The late Archean to Mesoproterozoic major unconformitybounded units of the Kaapvaal Province of Southern Africa: Precambrian Research, v. 74, p. 203223. Clendenin, C. W., Charlesworth, E. G., Maske, S., and de Gasparis, A. A., 1988, Normal simple shear model for the evolution of the early Proterozoic Ventersdorp Supergroup, Southern Africa: Economic Geology Research Unit Information Circular, 201, 20p. De Wit, M. J., Roering, C., Hart, R. J., Armstrong, R. A., De Ronde, C. E. J., Green, R. W. E., Tredoux, M., Peberdy, E., and Hart, R. A., 1992, Formation of an Archean continent: Nature, v. 357, p. 553562. Grutter, H. S., Apter, D. B., and Kong, J., 1999, Crust-mantle coupling: Evidence from mantle-derived xenocrystic garnets, in Proceedings of the VIIth International Kimberlite Conference, Cape Town, Dawson Volume, p. 307-313. Moser, D. E., Flowers, R. M., and Hart, R. J., 2001, Birth of the Kaapvaal tectosphere 3.08 billion years ago: Science, v. 291, p. 465-468. Nelson, D. R., Trendal, A. F., de Laeter, J. R., Grobler, N. J., and Fletscher, I. R., 1992, A comparative study of the geochemical and isotope systematics of Late Archean flood basalts from the Pilbara and Kaapvaal cratons: Precambrian Research, v. 54, p. 231-256. Poujol, M., Anhaeusser, C.R., and Armstrong, R.A., 2000, Episodic Archean granitoid emplacement in the Amalia-Kraaipan terrane, South Africa: new evidence from single zircon chronology with implications for the age of the Western Kaapvaal Craton. Information Circular 346, Economic Geology Research Institute, Johannesburg, 21 p. Poujol, M., and Anhaeusser, C. R., 2001, The Johannesburg Dome, South Africa: new single zircon U-Pb isotopic evidence for early Archean granite-greenstone development within the central Kaapvaal Craton: Precambrian Research, v. 108, p. 139-157. Raal, F. A., 1969, A study of some gold mine diamonds: American Mineralogist, v. 54, p. 292-296. Raal, F. A., and Robinson, D. N., 1980, Green for rarity: Nuclear Active, v. 23, p. 5-8. Richardson, S. H., Gurney, J. J., Erlank, A. J., and Harris, J. W., 1984, Origin of diamond in old enriched mantle: Nature, v. 310, p. 198-202. Richardson, S.H. and 3 others, 2001, Episodic diamond formation in the Kaapvaal craton keel. AGU Spring Meeting, Boston, Abstract V52A-07. Shirey, S.B. and 7 others, 2001, Archean emplacement of eclogitic components into the lithospheric mantle during formation of the Kaapvaal-Zimbabwe cratons. AGU Spring Meeting, Abstract V52A-05.

White, S. H., De Boorder, H., and Smith, C. B., 1995, Structural controls of kimberlite and lamproite emplacement, in Griffin, W. L., ed., Diamond Exploration: Into the 21th Century, Journal of Geochemical Exploration, p. 245-264.