Kibaran Belt sensu stricto (s.s.)

The southern tip of the NE-trending Kibaran Belt extends into the map area (Fig. 1). It is composed of deformed supracrustal rocks and orogenic granitoids (Rumvegeri 1991; Tack et al. 1994, 2002), and separates the Congo Craton to the NW from the Tanzania–Bangweulu Craton to the SE. A basement rise (2.2–1.9 Ga Rusizian Rise) subdivides the belt into an Eastern External Domain, underlain by the Archaean Tanzanian Craton, and a Western Internal Domain underlain by the Palaeoproterozoic Rusizian basement. The transition zone between the Western Internal Domain and Eastern External Domain is the site of numerous layered mafic complexes and associated A-type granitoids, dated at 1383–1371 Ma^7–13 in the northern part of the belt and at 1385–1370 Ma^1,3–6 farther south (superscripts refer to the geochronological reference numbers in Tables 1 and 2). Younger magmatism includes small-volume, A-type granitoid and mafic bodies that were emplaced along shear zones at c. 1205 Ma¹⁴ (Tack et al. 1999, 2002) and intrusion of regionally widespread c. 980 Ma¹⁵ Nb–Ta-bearing pegmatites and Sn-bearing granites (e.g. Brinckmann et al. 2001; Tack et al. 2002). Recent U/Pb dating of a low Th/U metamorphic zircon from the Kisele Monzogranite has yielded an upper intercept age of 1079 Ma ± 14 Ma², interpreted to date the time of sillimanite-grade D₂ deformation and metamorphism (Kokonyangi et al. 2001, 2004). The explanation for the c. 400 Ma magmatic and metamorphic history of the Kibaran Belt remains enigmatic. The c. 1370 Ma magmas have been interpreted as either (1) mantle and crustal melts resulting from lithospheric delamination following D₁ orogenic collapse (Tack et al. 1994) or (2) suprasubduction-zone mafic magmas associated with granites, formed by mixing of the mafic magmas with the products of dehydration melting of metasediments (Kokonyangi et al. 2004). Analyses of detrital zircon populations suggest that sediments forming the Eastern External Domain were deposited at c. 1.78 Ga and sourced from the Tanzanian Craton and Ubendian Belt, whereas parts of the Western Internal Domain sediments were deposited post-1.4 Ga by reworking of Eastern External Domain supracrustal rocks (Cutten et al. 2004).

Irumide Belt s.s.

The NE-trending Irumide Belt is predominantly underlain by granitoid rocks, but a Palaeoproterozoic platformal quartzite–pelite succession, the Muva Supergroup, is widespread in the NW (Daly & Unrug 1982). Deposition of the Muva Supergroup has been dated at c. 1880 Ma (De Waele & Fitzsimons 2004) and has recently been correlated with the Itremo Group in central Madagascar (Cox et al. 2004). In the NW of the belt, near the Irumide Front (Fig. 1), foliated granitoids (inferred to be part of the basement to the Muva Supergroup) have yielded ages in the range 2050–1930 Ma (De Waele et al. 2003b, 2004a, b). In contrast, granitoid magmatism elsewhere in the Irumide Belt occurred between 1660–1550 Ma and 1119–950 Ma^16–44. These late Palaeoproterozoic to early Mesoproterozoic and late Mesoproterozoic to early Neoproterozoic granitoids have bulk-rock geochemical signatures and highly negative ε_Nd(t) values indicating that they formed entirely by recycling of continental crust (De Waele et al. 2003a). Polyphase deformation in the Irumide Belt was coeval with formation of the subassemblages cordierite–andalusite and sillimanite–K-feldspar in pelite (Mapani 1999), indicating relatively low-pressure–high-temperature (LP–HT) metamorphic conditions. Direct dating of the metamorphism in two areas has yielded U–Pb ages of 1046 ± 3 Ma on metamorphic monazite (Schenk & Appel 2001) and 1020 ± 7 Ma and 1004 ± 20 Ma on low Th/U zircon overgrowths (De Waele et al. 2003b), compatible with the ages of syn- to post-tectonic granitoids.

Tectonic windows in the Zambian Zambezi and Lufilian Belts

In the Zambezi Belt of southern Zambia, two tectonic windows through the Neoproterozoic cover expose the Irumide basement consisting of megacrystic K-feldspar augen granite (locally known as the Mpande Gneiss and Munali Hills Granite) that have been dated at 1106 ± 19 Ma⁶³ and 1092 ± 4 Ma⁶⁸, respectively. Whole-rock major, trace and rare earth element analyses of the Mpande Gneiss and Munali Hills Granite show aspects of both S-type and arc-type magmatic environments (Katongo et al. 2004, 2005). The granitoids may be juvenile products formed in an active continental-margin arc, or they may have inherited their arc-like signatures from older crustal material from which they were recycled.

In the adjacent SE Lufilian Belt, inherited zircon xenocrysts (mostly 1.1–1.6 Ga^69–72, but including grains with ages up to 3.0 Ga) are present in Neoproterozoic lapilli tuffs (Rainaud et al. 2003). These xenocrysts display a range of ages essentially indistinguishable from those of the Palaeoproterozoic basement in the central Irumide Belt (De Waele et al. 2004a), suggesting that this basement may underlie part of the Lufilian Belt.

Southern Irumide Belt

The Southern Irumide Belt (Fig. 1) in SE Zambia and Mozambique represents part of the Irumide belt s.l. that has been extensively reworked under upper-amphibolite-facies conditions during Pan-African orogenesis. Although there are few areas where reliable data are available, it appears that this part of the Irumide Belt is composed of a series of distinct terranes of various ages and tectonic setting (Fig. 1) (Mapani et al. 2004). The Rufunsa Terrane, including the Chewore Inliers and Chongwe River area, consists of a collage of accreted oceanic, island-arc and continental-margin-arc lithologies and associated sedimentary successions (Oliver et al. 1998; Johnson & Oliver 2000, 2004). The oldest dated lithology, at c. 1393 ± 22 Ma⁴⁵, is a plagiogranite from a marginal-basin ophiolite suite that that has been linked to an oceanic realm to the south (present-day coordinates) of the Tanzania–Bangweulu Craton (Oliver et al. 1998; Johnson & Oliver 2000, 2004). Suprasubduction-zone island-arc magmatism in this region occurred between 1105 and 1037 Ma^46–60 (Table 1, Fig. 3). On its NE margin, the Rufunsa Terrane is in contact with the Luangwa Terrane, which consists of polydeformed metasediments and dioritic gneisses with isoclinally folded amphibolite layers (former dykes). One of these dioritic gneisses has yielded an Archaean age of c. 2.6 Ga (Cox et al. 2002), suggesting that the Luangwa Terrane may be an accreted microcontinental fragment. The structurally overlying Nyimba Terrane is predominantly composed of marble and calc-silicate, and is itself overlain by the Petauke–Sinda Terrane, which consists of a suite of calc-alkaline plutonic rocks interpreted as an accreted island arc (Mapani et al. 2004). One of these lithologies has been dated at 1125 ± 15 Ma⁶¹. The easternmost terrane of the Southern Irumide Belt in Zambia, the Chipata Terrane, consists of retrogressed mafic and felsic granulites of supracrustal origin. Its tectonic setting is at present unknown. Mapani et al. (2004) suggested that terrane accretion was contemporaneous with deformation and metamorphism in the Irumide Belt.

Southern Malawi and Lurio–Nampula Belts (Mozambique)

Calc-alkaline granitoids with gneissic fabrics and late Mesoproterozoic emplacement ages predominate in the Lurio and Nampula Belts (previously supergroups) of Mozambique (Pinna et al. 1993; Kröner et al. 1997; Jamal et al. 1999; Jamal & De Wit 2004) and southern Malawi (Kröner et al. 2001). In Mozambique, these granitoids were emplaced between 1148 and 1009 Ma^{130–133,135–140} (Table 1, Fig. 3). The granitoids lack inherited Archaean or Palaeoproterozoic zircons, and show both positive and mildly negative εNd(t) values indicating derivation in a mature island-arc-type tectonic setting (Jamal & De Wit 2004). These rocks lack any evidence for Mesoproterozoic high-grade metamorphism that might be associated with continental collision, but are strongly overprinted by Pan-African tectonometamorphism. In southern Malawi, calc-alkaline granitic orthogneisses with crystallization ages of 1040–999 Ma^119,120,122 (Table 1, Fig. 3) have Nd isotopic signatures compatible with their formation as juvenile melts in a suprasubduction-zone, continental-margin-arc setting (Kröner et al. 2001). Neoproterozoic arc magmatism (775–555 Ma^{112–118,121,123–129}) (Table 1, Fig. 3) has also been recognized in this region (Kröner et al. 2001).

Choma–Kalomo Block

The Choma–Kalomo Block (CKB in Fig. 1) on the NW margin of the Kalahari Craton is underlain by an unnamed supracrustal sequence consisting of pelitic schist with amphibolite layers and boudins cut by a batholithic granitoid complex. The oldest dated units in the batholith were emplaced between 1352 and 1343 Ma^96,97 (Table 1, Fig. 3), providing a minimum age for deposition of the supracrustal rocks. Younger intrusions have yielded ages of 1285 ± 64 Ma⁹⁸ and 1198 ± 6 Ma⁹⁹ (Table 1, Fig. 3), indicating a prolonged magmatic history. Ages of the two main magmatic pulses have recently been confirmed by SHRIMP dating of zircon, which has yielded a crystallization age of 1368 ± 10 Ma¹⁰⁴ for the main batholith, and ages between 1188 ± 11 Ma and 1174 ± 27 Ma^100–103 for cross-cutting intrusions. The batholith is inferred to have been emplaced as a series of ‘syntectonic to late syntectonic plutons’ (Hanson et al. 1988a), thereby also providing an approximate age for the main fabric-forming event in the Choma–Kalomo Block. Whole-rock Sm/Nd isotope data for the c. 1350 Ma granites suggest that, although the magmas interacted with sialic crust on ascent, they were not derived from an Archaean source (Hanson et al. 1988a).

Ghanzi–Chobe Belt

The present northwestern margin of the Kalahari Craton is the site of a late Mesoproterozoic intracontinental rift containing mafic and felsic volcanic rocks and small granitic intrusions, known as the Ghanzi–Chobe Belt (Fig. 1). The rift stratigraphy includes tholeiitic Kgwebe Volcanic series and rhyolite and granitic stocks (Kampunzu et al. 1998). The presence of Palaeoproterozoic zircon xenocrysts in volcanic units (Singletary et al. 2003) suggests that they were intruded through the Palaeoproterozoic Magondi Group. All dated felsic units have tightly constrained crystallization ages of 1107–1104 Ma^{105–108,110,111} (Table 1, Fig. 3), and have been correlated with the volcanic rocks in the Upper Sinclair sequence (Hoal 1993) that occur along strike to the SW in Namibia and have a U/Pb age of 1094 ± 20 Ma (Hegenberger & Burger 1995). Greenschist-facies metamorphism of the Kgwebe Volcanic series and the unconformably overlying Neoproterozoic sediments in the Ghanzi–Chobe Belt took place during the late Neoproterozoic to Palaeozoic Damara Orogeny (Kampunzu et al. 2000). However, meta-sediments exposed in a borehole (CKP11), and in the Kwando Complex farther west have amphibolite-facies fabrics correlated with those that formed at 1.2–1.15 Ga in the Choma–Kalomo Block (Singletary et al. 2003).

Northern Kalahari margin

The northern Kalahari margin is a complex subregion of multiply tectonized Archaean basement gneisses interleaved with Neoproterozoic metasediments of the Rushinga Group, structurally overlain by the high-grade Zambezi Allochthonous Terrane. This terrane consists of granitoid gneisses, known as the Ocellar Gneiss and Masaso Metamorphic Suite, and the mafic Mavuradonha Complex (Barton et al. 1993; Vinyu et al. 1999; Hargrove et al. 2003). Components of the Ocellar Gneiss have been dated at c. 1050 Ma^86,87, but the majority of the unit has crystallization ages of c. 870 Ma⁸⁸. The geochemical character of the Mesoproterozoic granitoids in the Zambezi Allochthonous Terrane is unknown. The Mavuradonha Complex (formerly Metamorphic Suite) is a Palaeoproterozoic metamorphosed layered gabbro–anorthosite intrusion (Müller et al. 2000, 2001; Hargrove et al. 2003), but other components of the Zambezi Allochthonous Terrane have yielded Neoproterozoic ages in the range 870–840 Ma^80–83, which are interpreted to represent the time of A-type magmatism and granulite-facies metamorphism (Hargrove et al. 2003).

The Katangan Supergroup, Lufilian Belt–Congo–Tanzania–Bangweulu Craton

The Neoproterozoic Katangan Supergroup was deposited on the Congo–Tanzania–Bangweulu Craton margin as a passive margin sequence (Mendelsohn 1961; Binda 1994; Cailteux et al. 1994; Porada & Berhorst 2000). In the Lufilian Belt, the Katangan Supergroup is deformed into a north-verging, thin-skinned, low- to medium-grade, fold–thrust belt. The review by Porada & Berhorst (2000) presented the most comprehensive stratigraphic framework for the supergroup, which we have integrated with recent geochronological investigations (Armstrong et al. 1999; Key et al. 2001) to produce a revised stratigraphic column (Fig. 5). The supergroup, which has been defined for the Copperbelt region, has been divided into two lithotectonic sequences, an older sequence comprising the copper-bearing Roan Group, which is in tectonic contact with a younger sequence comprising the Mwashia and Lower and Upper Kundelungu Groups. The Roan Group rests unconformably on, and contains boulders of the 877 ± 11 Ma⁶⁷ (Master et al. 2002; Rainaud et al. 2002) Nchanga Granite. The Roan Group, interpreted as a rift-to-drift sequence, was divided into three subgroups (Binda 1994) that are inferred to be approximately contemporaneous, laterally equivalent units (Porada & Berhorst 2000; Wendorff 2000). To the south of the Mwembeshi Shear Zone (Fig. 1), rocks interpreted to be equivalents of the Roan Group (Porada & Berhorst 2000) are known as the Cheta Formation, and the base of the sequence is marked by a thin succession of intensely sheared meta-pelite (the Nazingwe Formation), which passes up into the voluminous and laterally extensive Kafue Rhyolite that has been dated at 879 ± 19 Ma⁶². Overlying the Kafue Rhyolite is a sequence of marble and pelite that contains thousands of exotic mafic and ultramafic blocks. The Cheta Formation is in tectonic contact with underlying basement rocks and it has yet to be conclusively proven that it is equivalent to the Roan Group and/or that it formed on the Congo–Tanzania–Bangweulu Craton. It is possible that it is an allochthonous sequence that was tectonically emplaced on the Congo–Tanzania–Bangweulu Craton.

• Download figure
• Open in new tab
• Download powerpoint

Fig. 5.

 Reconstructed lithostratigraphic section of the Lufilian Belt in Zambia, modified from Porada & Berhorst (2000) with additional age constraint information. Black ellipses within the Roan and Cheta Formation represent numerous mafic and ultramafic blocks (refer to main text). Superscript numbers following ages relate to Table 1; locations are illustrated in Figure 2. Other references (superscript letters): A, Rainaud et al. (2002); B, Cahen et al. (1984); C, Kampunzu et al. (1998); D, Vavrdova & Utting (1972). D/C, disconformity; M.S.Z., Mwembeshi Shear Zone; N, north; S, south; T, thrust; U/C, unconformity

In the Copperbelt, the Mwashia Group, overlying the Roan, is composed of platformal, shallow-water argillite and carbonate, with many horizons of rift-related bimodal volcanic rocks (Porada & Berhorst 2000), one of which has been dated at 765 ± 5 Ma⁷³. The top of the Mwashia Group passes into glacial deposits (known as the Grand Conglomérat in the Democratic Republic of Congo (DRC)) of the Lower Kundelungu Group and its associated cap carbonate, which, in light of the c. 765 Ma age determination for the volcanic rocks in the Mwashia Group, probably correlates with the global Sturtian–Rapitan glacial event (Key et al. 2001). The remaining part of the Lower Kundelungu Group comprises a thick sequence (up to 5000 m) of shale. The overlying Upper Kundelungu Group consists of a second diamictite deposit at the base (equivalent to the Petit Conglomérat in the DRC and possibly recording the global 650–600 Ma Marinoan glacial event) overlain by a sequence of sandstone and siltstone, including the Cambro-Ordovician Luapula Beds (Vavrdova & Utting 1972). Deposition of the Mwashia and the overlying Kundelungu Groups is interpreted to have occurred in a rapidly subsiding extensional basin that opened some 100 Ma after initiation of the Roan rift basin (Porada & Berhorst 2000, and references therein).

Northern Kalahari margin

Sedimentary protoliths of the Neoproterozoic Rushinga Group were presumably deposited on the northeastern margin of the Kalahari Craton. In northeastern Zimbabwe, where metamorphic grade is in the amphibolite facies, the suite consists of basal quartzite or psammite overlain by a structural package of pelite, quartzite and marble (Barton et al. 1991; Hargrove et al. 2003), with volcanic and plutonic rocks dated between 808 and 795 Ma^84,85,89. Farther west, abundant peralkaline sheets were emplaced along with mafic magmas, in a continental rift setting, presumably through the Palaeoproterozoic Magondi Belt (Munyanyiwa et al. 1997; Dirks et al. 1999). Igneous zircons from granitic gneisses have yielded a TIMS age of 831 ± 6 Ma⁹⁰ and Pb/Pb evaporation ages that range from 854 ± 0.8 to 737 ± 0.9 Ma^91–95.

The oldest Neoproterozoic magmatic rocks (870–840 Ma) on the Kalahari margin are the Ocellar Gneiss and the Mavuradohna Metamorphic Suite in the Zambezi Allochthonous Terrane; they have been interpreted to represent a period of mid-crustal A-type magmatic activity accompanied by granulite-facies metamorphism (Hargrove et al. 2003). However, these rocks occur in thrust slices, and it is not known whether they formed on the Kalahari margin or are allochthonous to it.

Contemporaneous rifting events on the Congo–Tanzania–Bangweulu and Kalahari margins?

It is apparent that extension-related magmatism and sediment formation occurred on the northern margin of the Kalahari Craton and the southern margin of the Congo–Tanzania–Bangweulu Craton in the interval c. 880–750 Ma (Figs 3 and 6). The question therefore arises as to whether this provides credence to a model involving the attempted break-up of a previously assembled Congo–Tanzania–Bangweulu–Kalahari Craton and the formation of a ‘Zambezi Aulacogen’ (Daly 1986a, b; Hanson et al. 1988b, 1993, 1994, 1998; Unrug et al. 1998; Kampunzu et al. 2001; Hanson 2003; Hargrove et al. 2003). We argue that, although there is a broad similarity in the age of magmatism on both cratons, in detail the timing does not overlap. Rifting and associated extensional-related magmatism in the Congo–Tanzania–Bangweulu Craton occurred principally in two discrete pulses, some time after c. 890 Ma forming the Roan rift basin and at c. 765–735 Ma forming the Mwashia–Kundelungu basin (Porada & Berhorst 2000) (Fig. 6), although the Ngoma Gneiss (and possibly also the Lusaka Granite, although its crystallization age has a very large error of c. 68 Ma), were intruded between these two episodes. In contrast, rift-related magmatism on the Kalahari Craton occurred intermittently from 870 Ma (if the Zambezi Allochthonous Terrane formed on the Kalahari margin) to 750 Ma, with a peak at about 800 Ma associated with the deposition of the protoliths to the Rushinga Group (Fig. 6; Hargrove et al. 2003). However, as the Zambezi Allochthonous Terrane cannot be unambiguously linked to the Kalahari Craton on which it rests, extensional magmatic activity associated with the Kalahari margin may not have started until c. 830 Ma.

• Download figure
• Open in new tab
• Download powerpoint

Fig. 6.

 Histogram and probability density curves comparing Neoproterozoic magmatic rifting episodes in the Congo–Tanzania–Bangweulu Craton with those for the Kalahari, South China and Australian Cratons that formed part of Rodinia. Data for the Congo–Tanzania–Bangweulu Craton are summarized in this paper; data for other cratons were summarized by Li et al. (2003). RMS, Rushinga Metamorphic Suite; ZAT, Zambezi Allochthonous Terrane.

If the Kalahari Craton was a component of the Rodinia supercontinent, associated with Australia, Antarctica and South China (e.g. Dalziel 1997; Li et al. 2003; Loewy et al. 2003; Pisarevsky et al. 2003), then the A-type magmatism and associated rifting could have been related to supercontinent break-up. The timing of Neoproterozoic rifting events in this part of Rodinia was reviewed by Li et al. (2003, 2004) and a comparison of the timing of Neoproterozoic magmatism is shown in Figure 6. In this figure, the Kafue Rhyolite is shown as being associated with the Congo–Tanzania–Bangweulu Craton on which it sits at present, but this may be part of the Kalahari margin or allochthonous to both cratons. Li et al. (2003, 2004) proposed that there were two main rifting episodes, at 830–795 Ma and 780–745 Ma, which Pisarevsky et al. (2003) related to sequential break-up of the various cratons from the supercontinent. Apart from the Zambezi Allochthonous Terrane, which, as noted, may be allochthonous to the Kalahari Craton, the timing of rift magmatism in the Kalahari Craton is comparable with that in the South China and Australia Cratons; that is, the main magmatic activity switched off at 795 Ma.

Ophiolitic lithologies

The mafic and ultramafic blocks embrace a range of lithologies including basalt, amphibolite, gabbro, norite, picrite, peridotite, lherzolite, troctolite, eclogite and serpentinite (Vrána et al. 1975). Some of the mafic blocks in the Lufilian Belt have incompatible trace element ratios that are similar to those of magmas formed in an ocean-island setting or during the initial stages of continental rifting (Tembo et al. 1999). Gabbro, metagabbro and eclogite blocks from the Southern Irumide Belt have incompatible trace element and isotopic ratios comparable with those of mid-ocean ridge basalts (MORB) and have been interpreted as fragments of former ocean crust (John 2001; John et al. 2003, 2004a).

In addition to mafic and ultramafic blocks, the Bancroft Subgroup and Cheta Formation to the south of the Mwembeshi Shear Zone are composed of a shelf-like succession of dolomite and subordinate argillite and quartzite (Binda 1994; Porada & Berhorst 2000) and contain blocks and fragments derived from the adjacent Mindola and Kitwe Subgroups (Porada & Berhorst 2000). There appear to be two possible explanations for the abundant mafic blocks in these sediments: (1) the unit is an ophiolitic mélange (Gansser 1974; Williams 1977), i.e. an olistostrome with blocks emplaced by sedimentary mass flow processes following the emergence above sea level of ophiolitic thrust slices; (2) the unit is a tectonic mélange and the mixing of ophiolitic and continental-margin lithologies occurred during thrusting in the subsequent Lufilian–Zambezi Orogeny. Discrimination between these two alternatives is not straightforward, as much of the Bancroft Subgroup and Cheta Formation has undergone polyphase deformation and amphibolite-facies metamorphism during the Lufilian–Zambezi Orogeny. An olistostrome origin is tentatively preferred because: (1) the blocks are apparently limited stratigraphically to the Bancroft Subgroup and Cheta Formation (Porada & Berhorst 2000); (2) there is no evidence of contact aureoles or skarns developed at the margins of any of the blocks (Mendelsohn 1961), suggesting they were not intrusive into the sediments; (3) many blocks of mafic and ultramafic igneous lithologies are undeformed; (4) the great variation in size, composition and texture of the blocks over small distances is more compatible with an olistostromal origin; (5) blocks of igneous lithologies are accompanied by blocks of the adjacent Mindola and Kitwe formations, implying that these units were also involved in thrusting associated with ophiolite emplacement. This interpretation implies that ophiolite obduction onto the margin of the Congo–Tanzania–Bangweulu Craton occurred after the deposition of the Roan Group; that is, some time after c. 880 Ma. Alternatively, if the mafic–ultramafic blocks developed in a tectonic mélange, the upper age limit of the ophiolite would be the time of the Zambezi Orogeny (i.e. c. 550 Ma; see below).

Neoproterozoic continental-margin-arc granitoid magmatism

In southern Malawi, juvenile calc-alkaline granitoids (Kröner et al. 2001) have crystallization ages of 710–555 Ma^{112–118,121,123–129} (Table 1, Fig. 3). These were emplaced into c. 1040 and 999 Ma continental-margin-arc granitoids (Kröner et al. 2001) and suggest active-margin magmatism along part of the southern margin of the Congo–Tanzania–Bangweulu Craton from at least c. 710 Ma until terminal collision at c. 550–530 Ma.

Subduction-zone metamorphism?

Some of the mafic blocks within the Roan Group and Cheta Formation of the Lufilian and Zambezi Belts display evidence for very high-pressure, moderate-temperature (HP–MT) eclogite-facies metamorphism. These blocks occur in a discrete HP thrust slice or slices some 200 km long and 40 km wide (Fig. 7), in which eclogites and associated mafic blocks typically crop out as isolated hills and contact relations with the surrounding metasedimentary rocks are rarely exposed. However, wherever present, the surrounding country rock metapelite contains kyanite (Vrána & Barr 1972; Vrána et al. 1975), suggesting that the metasedimentary matrix also underwent an HP metamorphic evolution. Peak P–T conditions have been estimated at 630–690 °C and 26–28 kbar, and are compatible with a subduction-zone setting (John & Schenk 2003; John et al. 2003). Sm/Nd whole-rock–garnet and Lu–Hf whole-rock–garnet–omphacite isochrons (John & Schenk 2003; John et al. 2004a) have yielded a range of ages between 659 ± 14 Ma and 595 ± 10 Ma, which are interpreted as the time of eclogite-facies metamorphism (John et al. 2004a).

• Download figure
• Open in new tab
• Download powerpoint

Fig. 7.

 Location of the numerous high-pressure mafic, ultramafic and whiteschist blocks in the Zambezi and Lufilian Belts (from Vrána & Barr 1972; Vrána et al. 1975; Cosi et al. 1992; John et al. 2003, 2004b). Abbreviations as in previous figures, and: B.B., Bangweulu Block; S.I.B., Southern Irumide Belt; Z.B., Zambezi Belt.

If the blocks in the metasediments constitute a tectonic mélange, following eclogite-facies metamorphism of ocean crust in a subduction-zone setting, the HP rocks would have been incorporated into the metasediments of the Roan Group and Cheta Formation during collisional overthrusting (Porada & Berhorst 2000; John & Schenk 2003; John et al. 2003). Alternatively, if the blocks were originally part of an ophiolitic mélange, they would have been situated on continental crust prior to HP metamorphism, and their tectonometamorphic evolution would have involved subduction of the outer part of the Congo–Tanzania–Bangweulu Craton margin followed by its uplift and foreland-directed transport to mid-crustal levels. Attempted subduction of continental crust is known in several orogens (e.g. Alps, Appalachians, Scandinavian Caledonides, Variscides) and the presence of whiteschists in rocks of the Congo–Tanzania–Bangweulu margin (see below) provides additional evidence in favour of this model.

Collisional tectonometamorphism: Lufilian–Zambezi Orogeny

The Lufilian and Zambezi Belts are characterized by thin- and thick-skinned thrusting, respectively, and the polarity of thrusting in the Lufilian–Zambezi Orogen is illustrated in Figure 7. In the Zambezi Belt north of the Zambezi Valley, thrust transport was toward the NNE–NE (Wilson et al. 1993; Hanson et al. 1994; Johnson & Oliver 2004), whereas structures along the northern Kalahari margin are consistently SSE-verging (Barton et al. 1991; Dirks et al. 1999; Vinyu et al. 1999; Müller et al. 2001), indicating that the Lufilian–Zambezi Orogen was doubly vergent. Considering their spatial distribution, the orogen appears to have the architecture of an asymmetric, orogenic-scale structural fan. Orientations of the transport directions and the sinistral displacement on the Mwembeshi Shear Zone suggest that convergence between the Kalahari and Congo–Tanzania–Bangweulu Cratons was oblique.

In the Lufilian and Zambezi Belts, regional metamorphism reached amphibolite facies, except in discrete HP thrust slices, which are characterized by whiteschist-facies (kyanite–talc) assemblages (Johnson & Oliver 1998, 2002). P–T conditions in whiteschists peaked at c. 750 ± 25 °C at 13 ± 1 kbar and 600–900 °C at 8–15 kbar in the Lufilian and Zambezi belts, respectively, along a clockwise P–T path (Johnson & Oliver 1998, 2002, 2004; John 2001; John et al. 2004b). The development of these assemblages in both the reworked basement and cover sequences of the Congo–Tanzania–Bangweulu Craton implies that they originated by attempted subduction of the Congo–Tanzania–Bangweulu margin during the Zambezi Orogeny.

Figures 1 and 7 shows that the HP assemblages occur where the Damara–Lufilian–Zambezi Orogen is narrowest and the Congo–Tanzania–Bangweulu and Kalahari Cratons come into closest proximity at the current erosion level. This suggests that the HP metamorphism developed as a result of subduction of a promontory on the Congo–Tanzania–Bangweulu Craton and that the Congo–Tanzania–Bangweulu Craton formed the lower plate in the collisional orogen.

Age of collisional Pan-African tectonometamorphism

A total of 32 ages for the Pan-African metamorphism are available for this part of southern Africa (Table 2; Fig. 2) which has been divided into four subregions. Metamorphic ages for all four subregions overlap and are constrained between 615 and 517 Ma^141–172. The oldest ages, between 615 and 586 Ma^{163–165,168,171}, are from the Lurio and Nampula Belts. However, metamorphic zircons from this subregion also display a younger peak at 544–523 Ma^{166,167,170,172}, which overlaps with ages in some of the other subregions (Fig. 8). In southern Malawi the ages range between 575 and 547 Ma^157–162 with the older ages overlapping with suprasubduction-zone, continental-margin-arc magmatism (Kröner et al. 2001). In the Lufilian and Zambezi Belts in northern Zimbabwe and southern Zambia, the metamorphic ages are in the range 557–520 Ma^{141–145,149–156}, except for those from the Chowe River area (CR in Fig. 1), where HP metamorphism is dated at c. 570 Ma^146–148.

• Download figure
• Open in new tab
• Download powerpoint

Fig. 8.

 Histograms and probability density curves for the 32 metamorphic ages presented in Table 2 and Figure 7.

There is thus good evidence that Pan-African metamorphism in this region took place over a period of about 100 Ma from c. 615 to 520 Ma. The distribution of ages in Figure 8 shows that the interval c. 550–520 Ma is common to all four subregions, suggesting that this c. 30 Ma period was the time of the main collision between the Congo–Tanzania–Bangweulu and Kalahari Cratons. The presence of older metamorphic ages may indicate that collision was diachronous, or they may represent accretionary events between smaller terranes and blocks prior to the main continent–continent collision. Rb–Sr mineral isochrons and ⁴⁰Ar/³⁹Ar hornblende spectra from the Zambezi Belt (summarized by Vinyu et al. (1999) and Goscombe et al. (2000)) yield ages in the range 510–480 Ma and are interpreted to record unroofing of the orogen and cooling below c. 350 °C.

The c. 550–520 Ma timing of collisional metamorphism in the Zambezi Belt is similar to the 540–510 Ma range determined for the contiguous Damara Belt (Jung et al. 2001, 2003; Jung & Mezger 2003), suggesting the closure of a single ocean basin. The timing of peak metamorphism in the East African Orogen in Tanzania appears to have occurred significantly earlier, between 640 and 630 Ma (Muhongo et al. 2001; Kröner et al. 2003; Sommer et al. 2003), although the tectonic significance of these ages is disputed (e.g. Appel et al. 1998; Sommer et al. 2003). High-grade metamorphism in southern and central Madagascar occurred at c. 550 Ma (Kröner et al. 1996; De Wit et al. 2001; Kröner 2002), essentially coeval with that in the Damara–Lufilian–Zambezi Orogen. The spatial arrangement of the metamorphic ages and the differences in timing of peak metamorphism indicate that closure of the Mozambique and Zambezi–Adamastor oceans and amalgamation of Gondwana was not a simple collision between two large blocks or mega-cratons, i.e. East and West Gondwana, but rather a series of collisional events among several cratonic fragments, which took place over a period of about 100 Ma (Maas et al. 1992; Meert & Torsvik 1995; Fitzsimons 2000, 2003; Kröner 2002; Kröner & Cordani 2003; Meert 2003). Shackleton (1996) speculated on the location of suture zones that were integral to the formation and amalgamation of the Gondwana supercontinent. The data presented here show that the Damara–Lufilian–Zambezi Orogen has all the hallmarks of a suture zone that marks the site of a former Neoproterozoic ocean basin, and thus was integral to the amalgamation of Gondwana.