SHRIMP zircon dating and Nd isotopic systematics of Palaeoproterozoic migmatitic orthogneisses in the Epupa Metamorphic Complex of northwestern Namibia

Precambrian Research 183 (2010) 50–69

Contents lists available at ScienceDirect

Precambrian Research journal homepage: www.elsevier.com/locate/precamres

SHRIMP zircon dating and Nd isotopic systematics of Palaeoproterozoic migmatitic orthogneisses in the Epupa Metamorphic Complex of northwestern Namibia A. Kröner a,b,∗ , Y. Rojas-Agramonte a,b,c , E. Hegner d , K.-H. Hoffmann e , M.T.D. Wingate f,g a

Institut für Geowissenschaften, Universität Mainz, 55099 Mainz, Germany SHRIMP Centre, Chinese Academy of Geological Sciences, Beijing 100037, China c Instituto Superior Politécnico José Antonio Echeverria, Avenida 114 No 11901 entre 119 y 127, Marianao, CP 19390, Habana, Cuba d Department für Geo- und Umweltwissenschaften, Universität München, Theresienstraße 41, 803233 München, Germany e Geological Survey of Namibia, Private Bag 13297, Windhoek, Namibia f Tectonics Special Research Centre, University of Western Australia, Crawley, WA 6009, Australia g Geological Survey of Western Australia, 100 Plain Street, East Perth, WA 6004, Australia b

a r t i c l e

i n f o

Article history: Received 3 September 2009 Received in revised form 10 April 2010 Accepted 22 June 2010

Keywords: Congo Craton Epupa Complex Namibia Palaeoproterozoic Zircon dating

a b s t r a c t The Epupa Metamorphic Complex constitutes the southwestern margin of the Congo Craton and is exposed in a hilly to mountainous terrain of northwestern Namibia, bordering the Kunene River and extending into southern Angola. It consists predominantly of granitoid gneisses which are migmatized over large areas. This migmatization locally led to anatexis and produced crustal-melt granites such as the Otjitanda Granite. We have undertaken reconnaissance geochemical studies and single zircon U–Pb SHRIMP and Pb–Pb evaporation dating of rocks of the Epupa Complex. The granitoid gneisses, migmatites and anatectic melts are similar in composition and constitute a suite of metaluminous to peraluminous, calc-alkaline granitoids, predominantly with volcanic arc geochemical signatures. The zircon protolith ages for the orthogneisses range from 1861 ± 3 to 1758 ± 3 Ma. Anatexis in the migmatitic Epupa gneisses was dated from a melt patch at 1762 ± 4 Ma, and the anatectic Otjitanda Granite has a zircon age of 1757 ± 4 Ma. Migmatization and anatexis therefore occurred almost immediately after granitoid emplacement and date a widespread high-temperature Palaeoproterozoic event at ∼1760 Ma which has not been recorded elswhere in northern Namibia. The Nd isotopic systematics of all dated samples are surprisingly similar and suggest formation of the protolith from a source region that probably separated from the depleted mantle about 2.4–2.0 Ga ago. A major Archaean component in the source area is unlikely. Structural reworking of the Epupa gneisses during the transpressional Neoproterozoic to early Palaeozoic Kaoko orogeny led to partial or complete obliteration of the older structures and resulted in spectacular low- to high-grade shear and mylonite zones. This reworking did not affect the U–Pb isotopic system in the zircons but documents partial destruction of the Congo cratonic margin. The Epupa granitoid rocks formed during an event generally referred to in Africa as the Eburnian orogeny, but the nature and tectonic setting of the Congo craton of southwestern Africa during this time remain largely unknown. © 2010 Elsevier B.V. All rights reserved.

1. Introduction The Epupa Metamorphc Complex, located in Kaokoland, NW Namibia (Miller, 2008), is a general term for a widespread assemblage of mostly amphibolite-grade granitoid othogneisses with rare interlayered paragneisses and minor gabbroic intru-

∗ Corresponding author at: Institut für Geowissenschaften, Universität Mainz, 55099 Mainz, Germany. Tel.: +49 6131 3922163; fax: +49 6131 3924769. E-mail address: [email protected] (A. Kröner). 0301-9268/$ – see front matter © 2010 Elsevier B.V. All rights reserved. doi:10.1016/j.precamres.2010.06.018

sions. It constitutes the southwestern margin of the Congo craton and is exposed in a hilly to mountainous terrain extending from the Hoanib River near Sesfontein in the south to the Kunene River and beyond into Angola in the north (Fig. 1). Martin (1965) originally proposed the name Epupa Formation for these rocks and correlated them with other pre-Neoproterozoic assemblages farther south in Namibia. In southern Angola the unit is known as the Gneiss–Migmatite–Granite Complex (Carvalho, 1982; Carvalho and Alves, 1993). The complex is poorly mapped and constitutes the geologically least known terrain in southern Africa, due to its remoteness and poor

A. Kröner et al. / Precambrian Research 183 (2010) 50–69

51

Fig. 1. Overview map of the Epupa Complex and environs in NW Namibia (after Geol. Map of Namibia 1:1,000,000, 1980). Abbreviations in legend: Ka = Karoo; Da = Damara sequence; No = Nosib sequence; KC = Kunene Gabbro–Anorthosite Complex; OF = Okapuka Formation; EC = Epupa Complex.

accessibility. On the provisional Geological Map of Namibia, 1:250,000 (Sheet 1712-Swartbooisdrif, Schreiber, 2002) it is shown as consisting of a variety of gneisses or just “mixed gneiss”.

In the northeast the Epupa gneisses are intruded by the ca. 1385 Ma Kunene gabbro–anorthosite complex (KGAC, Drüppel et al., 2007) and are in tectonic contact with the high-grade Epembe Unit dated between 1640 and 1500 Ma and with granulite-facies

52

A. Kröner et al. / Precambrian Research 183 (2010) 50–69

Fig. 2. Simplified geological map of part of northern Kaokoland with localities of samples dated in this study. Geography based on “The Shell Map of Kaokoland, Kunene Region”, Namibian Ministry of Environment and Tourism, 2001. Geology simplified from Schreiber (2002).

rocks dated between 1490 and 1450 Ma (Seth et al., 2003; Brandt et al., 2007, see Fig. 2). In the Baynes and eastern Otjihipa Mountains south of the Kunene River and in the Steilrand Mountains in the southeast, the Epupa gneisses are unconformably overlain by a sequence of quartzites and shales, locally with a basal conglomerate (Fig. 2), and these are part of the Neoproterozoic Damara sequence of the Pan-African Kaoko belt. Farther west, in the western Otjihipa Mountains, a sequence of quartzite, quartz-sericite schist and chlorite or amphibolite schist, intruded by gabbroic sills and dykes, represents an older, pre-Damara succession known as Okapuka Formation (Geol. Map of Namibia 1980; Schalk, 1985, see Figs. 1 and 2). This succession has been folded together with some of the Epupa gneisses, presumably during the PanAfricasn event, and our unpublished zircon dating indicates an age >900 Ma. A granitic augen-gneiss that was severely deformed and partly mylonitized during the late Neoproterozoic to early Palaeozoic Kaoko orogenic event is exposed in a small tectonic window below Neoproterozoic metasedimentary rocks of the Damara Supergroup near Sesfontein in southern Kaokoland (located south of Fig. 1), and zircons from these gneisses yielded ages between 2645 and ca. 1500 Ma (Seth et al., 1999). However, it is uncertain whether these rocks are part of the Epupa Complex or constitute a separate pre-Kaoko terrane that was accreted to the Kaoko belt in the Neoproterozoic (Seth et al., 1999). One of the purposes of the present study was to search for additional Archaean rocks in the main Epupa Complex farther north, and because no other Archaean rocks were

discovered, we still consider that the Sesfontein gneisses may be part of an exotic terrane.

2. Main rock types and previous geochronology The main body of the Epupa Metamorphic Complex in northern Kaokoloand consists of upper greenschist- to amphibolite-facies granitoid orthogneisses that are variably migmatized and locally become anatectic. This process locally leads to the development of granite, the largest body of which is known as Otjitanda Granite (Geol. Map of Namibia, 1980; see Fig. 1). The progressive development of migmatites through agmatites into granite is described below. The only detailed mapping of the various pre-KGAC metamorphic rocks was undertaken in a broad east-west zone around the village of Epembe in the east (Fig. 1) where Brandt et al. (2003, 2007) distinguished a variety of amphibolite-facies gneisses (Orue Unit) and a fault-bounded block of granulites (Epembe Unit). The relationship between these rocks and the migmatitic Epupa gneisses farther west is not known because the contact zone is not exposed, but is probably faulted. Seth et al. (2003) dated metamorphic zircons from granulitefacies metamorphic rocks of the fault-bounded Epembe Unit south of the KGAC (Fig. 1) at 1510–1520 Ma. These rocks were previously included in the Epupa Complex (Miller and Schalk, 1980; Miller, 2008) but may constitute a separate crustal unit because granulite-

A. Kröner et al. / Precambrian Research 183 (2010) 50–69

53

Fig. 3. Diagrams for chemical characterization of dated granitoid samples from the Epupa Complex. (a) Feldspar triangle after O’Connor (1965; (b) FeOtot/FeOtot + MgO after Frost et al. (2001). (c) Na2 O + K2 O + CaO after Frost et al. (2001). (d) Data plotted in Shand’s diagram (1943).

facies rocks of the above age were not found in the main Epupa Complex, as shown below. However, zircon cores (detrital grains?) in several granulite-facies paragneisses of the Epembe Unit yielded 207 Pb/206 Pb SHRIMP ages of 1810 ± 11 to 1635 ± 13 Ma (Seth et al., 2003), and such ages do indeed occur in the Epupa Complex proper. It is therefore possible that the Epembe Unit contains metasedimentary rocks that were derived from erosion of the older Epupa gneisses. There are only few zircon ages previously reported from the main Epupa Complex. Tegtmeyer and Kröner (1985) dated a coarsegrained granitic augen-gneiss collected at the viewpoint above the Ruacana Falls on the Kunene River and obtained five fairly discordant multigrain size-fraction analyses that were fitted to a regression line intersecting Concordia at 1795 + 33/−29 Ma. Seth et al. (1999) dated two granitic to dioritic orthogneisses from the Ganamub antiform west of Sesfontein (just S of Fig. 1) and obtained zircon ages of 1985 ± 23 and 1961 ± 4 Ma. However, as in the case of the Hoanib River Archaean gneisses, it is uncertain whether this basement terrane is really part of the Epupa Complex. The area along the lower Hoarusib River southwest of Puros in western Kaokoland (Fig. 1) was investigated in detail by Kröner et al. (2004) who described intimate tectonic interlayering of migmatitic gneisses of Palaeoproterozoic and Neoproterozoic age on the west-

ern margin of the Kaoko orogenic belt. They dated zircons from migmatitic orthogneisses and granitoid gneisses at 1448–2028 Ma and ascribed these to discrete magmatic events in the Palaeoand Mesoproterozoic. The original intrusive contacts are not preserved due to pervasive reworking during the Pan-African event. Goscombe et al. (2005) separated zircons from melt segregations in a coarse-grained meta-pelitic paragneiss from north of Puros and of presumed Palaeoproterozoic age and obtained numerous nearconcordant and strongly discordant analyses for inherited grains with the least discordant zircons defining a mean 207 Pb/206 Pb age of 1768 ± 8 Ma. Zircons representing the melt phase were dated at 562 ± 5 Ma, thus confirming the strong Pan-African overprint. Carvalho et al. (2000) interpreted Rb–Sr whole-rock isochron ages of 1884 ± 39 and 1826 ± 48 Ma for granitoid gneisses and migmatites, respectively, in the Gneiss–Migmatite Complex of southwest Angola as reflecting high-grade metamorphism (Namib event of Torquato, 1974, later equated with Eburnian event in Central and West Africa), whereas an undeformed granite yielded an emplacement age of 1686 ± 69 Ma. However, the interpretation of these Rb–Sr ages is ambiguous because no metamorphic event of this age range has so far been recognized in northwestern Namibia. McCourt et al. (2004) recently reported SHRIMP zircon ages of 1959 ± 6.4 and 2038 ± 28 Ma for little deformed granites directly

54

A. Kröner et al. / Precambrian Research 183 (2010) 50–69

Fig. 4. Field photographs showing progressive deformation of coarse augen-gneiss and migmatite in the area around Okongwati. (a) Very coarse-grained porphyritic granite–gneiss with well rounded Kfsp-porphyroclasts. (b) Same rock showing higher strain with porphyroclasts drawn out into large “augen” or thin lenses. (c) Same rock with strong strain increase in shear zone where porphyroclasts are drawn out into thin layers. (d) Same rock deformed into mylonite along E-W striking shear zone. (e) Migmatitic gneiss with dark grey-green chlorite- and epidote-rich palaeosome and criss-crossing network of granite veins around Okongwati. (f) Same rock a little farther west where higher strain has brought granitic veins into parallelism and melt begin to develop in small shear zones.

underlying the undeformed Chela Group sediments in southwestern Angola. They also found detrital zircon ages between 1765 and 2977 Ma in sediments of the Chela Group which they consider to be derived from the underlying basement. We undertook reconnaissance dating of granites, granitoid gneisses and migmatites in the north-central part of the Epupa Complex, complemented by whole-rock geochemistry and Nd isotopic systematics for selected samples. The analytical procedures are summarized in the Appendix. Our traverse mainly followed the road from Otjiveze to Okongwati and Ondova, and then the track westwards to Etengwa, Otjitanda and south to Etanga, including a

very rough track to Otjihaa (see Fig. 2). This area consists almost exclusively of granitoid rocks which are largely migmatized and display variable stages of preservation and overprinting. 3. Geochemistry Major and trace element compositions together with the normative An, Ab and Or values for the dated samples are given in Table 2. Rock names are based on the normative ternary plot of O’Connor (1965, see Fig. 3a) and/or the silica-alkali plot of Middlemost (1985). Except for one sample of hornblende-rich monzodioritic gneiss, the

A. Kröner et al. / Precambrian Research 183 (2010) 50–69

55

Table 1 GPS locations for dated samples and brief characterization. Sample No.

Location ◦



Rock type and characterization 

Na 01/43

S17 35 31.3 E12◦ 25 31.8

Strongly foliated granodioritic gneiss, Hartmann Mts.

Na 01/48

S17◦ 39 56.1 E12◦ 40 53.0

Mylonitic migmatitic gneiss W of Van Zyl’s Pass

Na 01/49

S17◦ 28 00.4 E13◦ 19 46.1

Very coarse-grained porphyritic granitic augen-gneiss

Na 01/50

S17◦ 26 01.8 E13◦ 16 13.9

Granite–gneiss with anastomosing veins of granite

Na 01/51

S17◦ 28 17.5 E13◦ 03 51.6

Aplitic granite vein cutting banded migmatite

Na 01/54

S17◦ 29 55.2 E13◦ 00 08.8

Melt patch in migmatitic gneiss

Na 01/56

S17◦ 36 57.8 E12◦ 51 34.7

Homogeneous porphyritic Otjitanda granite

Na 01/57

S17◦ 52 17.9 E13◦ 04 22.5

Streaky hbl-gneiss derived from porphyritic tonalite

Na 03/80

S17◦ 15 53.1 E13◦ 12 29.0

Coarse porphyritic granodioritic gneiss

Na 03/85

S17◦ 35 37.2 E12◦ 54 21.4

Palaeosome of partly migmatitic, foliated granite–gneiss

Na 03/86

S17◦ 30 53.2 E12◦ 58 36.1

Agmatitic migmatitic granite–gneiss

Na 03/88

S17◦ 54 48.1 E12◦ 50 18.7

Coarse porphyritic granite–gneiss

Na 03/89

S17◦ 57 35.3 E12◦ 48 39.4

Porphyritic granite–gneiss, partly migmatitic

Na 03/90

S17◦ 54 27.4 E13◦ 00 28.7

Palaeosome of well foliated migmatitic granite–gneiss

Na 03/93

S17◦ 14 41.0 E12◦ 20 17.6

Well foliated red granite–gneiss within migmatite

Na 03/97

S17◦ 54 05.1 E12◦ 42 28.8

Muscovite-rich polydeformed migmatitic gneiss

granitoids show a wide range in SiO2 concentrations from ca. 62 to 77 wt. % and variable Na2 O/K2 O ratios of ca. 0.5–2.3, and mobilization of Na and K during migmatization may have contributed to this wide range. The composition of the granitoids corresponds to granodiorite, granite, and quartz monzonite in the above diagrams (Fig. 3a). Following the classification of Frost et al. (2001), most Epupa samples are characterized by low Fe-numbers, but a few are high and resemble A-type granitoids (Fig. 3b). However, Ga/Al ratios and Nd, Zr and Y are generally low and La/Y ratios are high and untypical of A-type granites (Whalen et al., 1987), except for sample Na 03/93 which is high in Y, Zr and Nb and has a high 1000× Ga/Al ratio of 3.5. A calc-alkaline, Cordilleran-type crustal protolith is supported for most samples by high La/Nb ratos and high primitive mantle-normalized (La/Y)n ratios of 6–39 (Table 2). Low abundances in the high-field strength elements Nb, Ti and Y relative to similarly behaving REE further support the calc-alkaline character (Table 2 and Taylor and McLennan, 1985). In Fig. 3c, a predominantly calc-alkalic and S-type granite affinity for samples with high SiO2 is suggested. Mixed crustal protoliths comprising igneous as well as sedimentary material are supported in Shand’s diagram (Fig. 3d). The dated samples predominantly plot in the field of Volcanic Arc Granite (Pearce et al., 1984), but some samples plot in the Within-Plate Granite field (diagram not shown), consistent with the classification in Fig. 3b. Thus, the chemical data suggest that there is no remarkable and distinct chemical difference between the dated rocks. Yet, primitive

mantle-normalized (La/Yb)n ratios indicate highly variable degrees of fractionation of light and heavy REE in the samples that may have been caused by melting of garnet-bearing as well as garnet-free crustal protoliths. The Sm–Nd isotopes listed in Table 5 show an overall uniform crustal source with respect to the average mean crustal residence age. Initial εNd values of 3.2 to −2.5 indicate involvement of juvenile as well as older crust, corresponding to DM model ages of 1.8–2.4 Ga. 4. Field relationships, zircon geochronology, and whole-rock Nd isotopic systematics Some 6 km southeast of Okongwati an extremely coarse-grained granitic augen-gneiss is exposed where individual K-feldspar phenocrysts reach diameters up to 10 cm (Fig. 4a). The rock displays a weak to very strong foliation dipping south at 55–70◦ , and at many localities the K-feldspar porphyroclasts or aggregates are strongly elongated (Fig. 4b). Locally, the gneiss shows remarkable strain gradients over short distances, and narrow shear zones have transformed the rock into a banded gneiss (Fig. 4c). At several localities within the Epupa gneisses covered in this paper the east-west striking shear zones are up to tens of metres wide and have transformed the augen-gneisses and/or migmatites into a finely banded mylonitic gneiss (Fig. 4d). These shear zones are not related to the Pan-African structural overprint which is broadly north-south and occurs near the western margin of the Epupa Complex (Fig. 2) but

56

A. Kröner et al. / Precambrian Research 183 (2010) 50–69

Fig. 5. Concordia diagrams showing analytical data for SHRIMP II analyses of granite–gneiss and migmatite samples of the Epupa Complex, Kaokoland, NW Namibia. Data boxes for each analysis are defined by standard errors in 207 Pb/235 U, 206 Pb/238 U and 207 Pb/206 Pb. Inset in Fig. 4a shows histogram with distribution of radiogenic lead isotope ratios derived from evaporation of 5 single zircons from sample Na 01/48, integrated from 769 ratios. Sample number and short characterization is given in each diagram. For sample locations see Table 1 and Fig. 2.

may be related to the same event that gave rise to the east-west trending high-grade Orue and Epembe Units farther east (Fig. 2). We dated granitic augen-gneiss sample Na 01/49 from the locality 6 km southeast of Okongwati (Table 1 and Figs. 2 and 5a). The zircons are stubby to long-prismatic with little to well-rounded terminations, due to “metamorphic corrosion”, i.e. partial zircon dissolution and recrystallization during upper amphibolite-facies metamorphism (Kröner et al., 1994). Most grains display excellent oscillatory zoning, typical of magmatic growth (see inset in Fig. 5a). Six grains were analyzed on SHRIMP II (Table 3) of which three are concordant, and the others are variably discordant, but all analyses have similar 207 Pb/206 Pb ratios and define a best-fit line (MSWD = 0.19, pof = 0.94) with a Concordia intercept age of 1758 ± 6 Ma (Fig. 5a). The mean 207 Pb/206 Pb age of all six grains is 1758 ± 3 Ma. A further six grains were evaporated individually and yielded identical 207 Pb/206 Pb ratios which provide a mean age of 1756.7 ± 0.4 Ma (Table 3 and Fig. 5a, inset). This age is very precise in view of the large number of isotopic ratios measured (Table 4) and is interpreted to most closely reflect the time of protolith emplacement. Because the SHRIMP data suggest that the discordance was caused by Pb-loss in recent times (alignment of analyses along a discordia line through the origin in Fig. 5a), the excellent agreement between the SHRIMP and evaporation data show the latter technique to yield meaningful ages.

The εNd(t) value for sample Na 01/49 is −1.3, and the corresponding mean crustal residence age is 2.15 Ga (Table 5). This can either be interpreted as reflecting generation from older crust caused by a depleted mantle melting event at about 2.15 Ga or mixing of unspecified older crust with a juvenile reservoir. The significance of the Nd model ages will be discussed more fully below when all data have been presented. Very similar rocks are exposed along the road from Okongwati to Ondova, and we dated granodioritic gneiss sample Na 03/80 from an exposure near the road just northeast of Ondova (Table 1 and Fig. 2). The zircons are long-prismatic to oval, and most have well rounded terminations. Striped and oscillatory zoning are common (see inset, Fig. 5b). Six grains were analyzed (Table 3) of which the isotopic ratios of four are concordant or near-concordant, and two are strongly discordant (Fig. 5b). The four least disturbed grains have a mean 207 Pb/206 Pb age of 1767 ± 6 Ma, whereas strongly discordant grain 3 has a similar 207 Pb/206 Pb ratio but was not considered for mean age calculation. However, if regressed together with the other four grains (MSWD = 0.13, pof = 0.88), the Concordia intercept age is 1766 ± 8 Ma (Fig. 5b). Grain 1 is indistinguishable in morphology from the other grains but is much younger with a minimum 207 Pb/206 Pb age of 1347 ± 12 Ma (Table 3). This age is difficult to explain in view of the fact that the host rock does not show any strong metamorphic overprint other than the duc-

A. Kröner et al. / Precambrian Research 183 (2010) 50–69

57

Fig. 6. Field photographs showing progressive development of anatexis, agmatite formation and granite intrusion in the Epupa Complex of northern Kaokoland, NW Namibia. (a) Partial melt in small-scale shear zones following developing axial-plane foliation in well layered migmatitic granite–gneiss west of Okongwati. (b) Advanced stage of melt formation and beginning of flattening of melt veins during progressive deformation. (c) Very advanced stage of melt formation accompanied by ductile deformation. Note alignment of surviving rock fragments and melt in shear zone cutting older foliation parallel to hammer handle. (d) Agmatitic granite-migmatite “mush” cutting older banded migmatitic gneiss. (e) Undeformed, homogeneous Otjitanda Granite with xenoliths of migmatitic granite–gneiss. (f) Undeformed porphyritic Otjitanda Granite cutting fabric in older migmatitic granite–gneiss N of Otjitanda.

tile deformation giving rise to a pervasive foliation. We consider laboratory contamination highly unlikely because this sample was processed together with other ca. 1700–1800 Ma old rocks. However, we found granitoid rocks of a similar age in the western part of the Epupa Complex near Marienfluss (Fig. 2, see Kröner et al., 2006), and details on these rocks will be reported elsewhere. At the village of Okongwati the coarse augen-gneiss changes into a migmatitic granite–gneiss, but the contact between these rock

types is not exposed. The migmatitic gneiss has a massive, homogeneous palaeosome with strong retrogression defined by chlorite and epidote and a network of intrusive granite (Fig. 4e). A sample of this granitic palaeosome (Na 01/50, Table 1 and Fig. 2) was dated, and the zircons are predominantly long-prismatic with little to well rounded terminations and well-developed oscillatory zoning (inset in Fig. 5c). Seven idiomorphic grains with oscillatory zoning were analyzed on SHRIMP II of which the results for three grains are concordant or near-concordant, whereas four results are

58

A. Kröner et al. / Precambrian Research 183 (2010) 50–69

Fig. 7. Concordia diagrams showing analytical data for SHRIMP II analyses of migmatitic gneisses and granite of the Epupa Complex, Kaokoland, NW Namibia. Data boxes as in Fig. 4. Inset shows histogram with distribution of radiogenic lead isotope ratios derived from evaporation of five single zircons from same sample, integrated from 382 ratios. Sample number and short characterization is given in each diagram. For sample locations see Table 1 and Fig. 2.

variably discordant (Fig. 5c), but all yielded identical 207 Pb/206 Pb ratios (Table 3) with a mean age of 1789 ± 2 Ma, suggesting Pb-loss in recent times. A best-fit line (MSWD = 0.07, pof = 0.99) defines a Concordia intercept age of 1790 ± 8 Ma (Fig. 5c). We interpret this age as reflecting the time of protolith emplacement, and the slight difference in age between Na 01/49 and Na 01/50 suggests the precursor of the coarse augen-gneiss to be intrusive into the Okongwati gneiss. The εNd(t) value for sample Na 01/50 is −0.7, and the corresponding mean crustal residence age is 2.14 Ga (Table 5). This is identical to the model age for the coarse augen-gneiss south of Okongwati and suggests a similar souce fort his palaeosome. A little farther west on the track to the Otjijandjasemo hot spring the relatively low-strain migmatite changes gradually into a more deformed variety where the neosome granite veins become aligned and folded (Fig. 4f). Still farther west on the track from Okongwati to Etengwa (Fig. 2), one can observe an advanced stage of migmatization and beginning of anatectic melting. The migmatitic granite–gneiss is strongly layered, and small, homogeneous domains representing former melt patches begin to develop in small-scale shear zones parallel to a well-developed axial-plane foliation (Fig. 6a). The next stage is that these homogeneous veins become more abundant, evidence of melting can also be seen in the granitic portions of the migmatitic veins, and the entire rock begins to lose its coherency and deforms plastically (Fig. 6b). The rock gradually creates the impression of having been a “mush”, and

where shearing is prevalent the melt portions align themselves in the direction of shear motion (Fig. 6c). A further increase in the degree of melting led to agmatitic portions (Mehnert, 1968) where the once melted material becomes predominant but where the dissolving migmatite can still be recognized (Fig. 6d). Finally, a relatively homogeneous igneous rock is developed, representing a new granite known as Otjitanda Granite Our sample Na 01/56 contains xenoliths of the not yet melted migmatite in various stages of dissolution (Fig. 6e). This granite becomes widespread around the village of Otjitanda and is shown on the 1:1,000,000 Geological Map of Namibia (1980). Its boundaries with the surrounding Epupa migmatites are largely gradational although sharp intrusive contacts are also found, where the unfoliated granite cuts the older migmatitic veining (Fig. 6f). We dated an agmatitic, migmatitic, granodioritic gneiss similar to that shown in Fig. 6d, collected on the track between Etengwa and Okauwa (sample Na 03/86, Table 1 and Fig. 2). The zircons are stubby to long-prismatic, and mostly show some rounding at their terminations (inset in Fig. 5d). Some grains have very narrow lowU (bright under CL) rims, perhaps related to the melting event, but these could not be analyzed on SHRIMP. Three grains, two concordant and one discordant (Fig. 5d), have identical 207 Pb/206 Pb isotopic ratios (Table 3) with a mean age of 1773 ± 3 Ma. The corresponding regression line (MSWD = 0.04, pof = 0.85) defines an identical Concordia intercept age of 1774 ± 12 Ma. We interpret this age as reflecting inheritance of the original granodioritic gneiss,

Table 2 Major and trace element data for dated granitoid samples of the Epupa Complex. Na01/43 Granod. gneiss

Na01/48 Migm. gneiss

SiO2 TiO2 Al2 O3 Fe2 O3 MnO MgO CaO Na2 O K2 O P2 O5 LOI

66.70 0.54 15.34 3.99 0.08 1.32 3.40 3.77 3.15 0.14 0.78

Total Rb Sr Ba Zr Y Nb U Th Pb Ga Cr Ni Co Cu Zn Sc La Ce Nd Sm (La/Y)n Na2 O/K2 O Or Ab An

Na01/49 Augengneiss

Na01/50 Granitic palaeosome

Na01/51 Aplitic dike

Na01/54 Monz. melt patch

Na01/56 Otjitanda granite

74.45 0.28 13.40 2.00 0.04 0.63 1.42 2.66 4.76 0.07 0.83

69.21 0.56 14.04 3.49 0.06 0.84 2.31 2.40 5.10 0.12 1.06

71.62 0.43 13.75 2.61 0.07 0.63 1.90 2.52 4.64 0.10 1.16

99.20

100.54

99.19

100 419 867 155 21 12 2 9 16 19 19 10 99 8 58 9 39 70 30 5 12 1.20 28.1 48.2 23.6

161 173 573 126 30 7 2 12 29 15 10 8 135 4 22 7 31 60 28 8 6.6 0.56 49.2 39.3 11.5

217 182 1157 215 43 14 4 26 24 18 23 10 55 15 42 10 71 136 53 13 11 0.47 49.3 33.2 17.5

Na01/57 Monz. gneiss

Na03/80 Granod. gneiss

Na03/85 Migm. granite gneiss

77.72 0.14 11.36 0.77 0.01 0.05 0.73 2.20 5.57 0.01 0.32

62.91 0.47 17.84 4.05 0.08 1.28 2.71 3.58 5.10 0.08 1.52

99.43

98.89

171 231 1295 198 45 12 3 19 29 15 16 5 84 4 29 11 65 106 48 11 9.2 0.53 47.7 37.1 15.2

132 171 1182 96 9 5 <1 9 22 12 8 4 145 3 4 3 36 54 21 6 26 0.39 59.7 33.8 6.46

Na03/86 Migm. granod. gneiss

Na03/88 Porph. granite gneiss

70.01 0.51 13.96 3.44 0.06 0.51 1.75 3.20 5.17 0.14 0.73

49.25 0.48 20.53 6.8 0.12 6.28 7.97 2.92 2.68 0.14 2.75

67.93 0.42 14.56 3.59 0.08 1.97 2.61 3.81 2.58 0.18 1.80

70.00 0.32 14.78 1.90 0.04 0.37 1.56 3.00 5.95 0.11 0.88

72.36 0.29 13.35 2.63 0.06 0.36 2.83 3.62 3.03 0.07 0.79

70.83 0.49 13.66 3.58 0.08 1.05 1.52 3.30 3.96 0.15 1.14

99.61

99.47

99.91

99.60

98.90

99.40

262 239 819 185 49 14 10 46 28 22 14 8 63 6 63 9 38 75 34 7 4.9 0.70 41.1 41.3 17.6

253 168 1479 478 56 33 8 60 45 22 8 4 69 3 72 6 172 325 128 20 20 0.62 46.7 41.4 11.9

70 204 1242 56 13 4 <1 4 14 17 586 130 54 27 89 16 13 21 12 4 6.4 1.09 21.0 32.7 46.3

116 514 1419 114 19 6 1 5 31 18 41 24 57 18 50 9 26 48 24 7 8.8 1.48 25.7 54.4 19.9

138 192 1789 255 13 10 7 39 30 16 7 3 65 5 28 4 64 144 37 7 32 0.50 52.0 37.6 10.4

247 157 846 261 25 8 21 117 45 15 7 4 68 5 26 3 150 255 102 16 38 1.19 30.0 51.2 18.8

Na01/89 Migm. granite gneiss

Na03/90 Granod. gneiss

Na03/93 Augengneiss

Na03/97 Musc. granite gneiss

66.84 0.37 17.10 1.12 0.04 0.21 3.70 3.28 4.97 0.02 0.97

63.54 0.73 15.28 5.86 0.11 1.78 4.15 2.89 3.08 0.19 1.52

75.03 0.32 11.91 2.98 0.03 0.99 3.20 4.14 1.79 0.13 1.06

71.58 0.57 13.35 4.23 0.05 1.19 1.88 2.93 3.07 0.07 0.73

99.70

100.40

99.10

99.96

99.70

177 184 783 183 58 12 5 23 24 15 14 10 82 4 44 6 67 114 53 9 7.4 0.83 40.4 48.2 11.3

132 378 1129 305 31 12 2 11 19 17 12 11 54 12 64 11 40 73 37 7 8.3 0.66 28.1 42.6 29.2

134 287 828 226 34 11 1 14 17 17 27 16 39 25 71 11 35 100 32 5 6.6 0.94 29.4 39.4 31.2

172 47 745 508 100 28 5 28 15 22 1 5 70 1 24 2 101 202 92 19 6.3 2.31 48.8 46.1 5.14

138 178 603 188 22 9 3 13 22 15 43 14 57 4 48 11 36 68 27 4 11 0.95 35.0 47.9 17.1

A. Kröner et al. / Precambrian Research 183 (2010) 50–69

Sa. #

n = normalized to primitve mantle values.

59

60

A. Kröner et al. / Precambrian Research 183 (2010) 50–69

Fig. 8. Field photographs showing rock relationships and progressive Pan-African structural overprinting in gneisses of the Epupa Complex, Kaokoland, NW Namibia. (a) Undeformed aplitic granite dyke (sample Na 01/51) cutting migmatitic granite–gneiss just W of Okongwati.(b) Refoliation of Epupa augen-gneiss sample Na 01/48 in PanAfrican Puros shear zone along margin of the Kaoko transpressional orogenic belt W of Van Zyl’s Pass. (c) Intense mylonitic reworking of Epupa augen-gneiss in Puros shear zone W of Van Zyl’s Pass. (d) Reddish-grey augen-gneiss sample Na 03/93 with strong Pan-African reworking near western margin of Epupa Complex at Kunene River.

and this is in the same range as other protolith ages for the Epupa gneisses. An early quartz monzonitic former melt patch in a migmatite which intruded along the axial plane of tight folds deforming the migmatitic banding was collected between Etengwa and Okauwa (sample Na 01/54, Table 1 and Fig. 2). This rock is itself ductilely deformed, suggesting that it formed early in the anatectic process when deformation was still intense. This sample contains many long-prismatic, mostly euhedral zircons with good oscillatory zoning (inset in Fig. 7a). Five grains were dated of which four are concordant and one is about 10% discordant (Table 3 and Fig. 7a). All five grains yielded a mean 207 Pb/206 Pb age of 1762 ± 4 Ma, whereas the Concordia intercept age is 1763 ± 10 Ma (MSWD = 0.09, pof = 0.96). We interpret this age as reflecting the peak of highgrade metamorphism, leading to anatexis in the Epupa migmatite. The εNd(t) value for sample Na 01/54 is −2.5, the lowest value of all samples analyzed in this study. The DM model age of this sample is 2.42 Ga and must be interpreted as a maximum value because the Sm/Nd ratio of this sample of an arrested melt patch is higher than a typical crustal value of 0.12 (Table 5). Nevertheless, the Nd isotopic data preclude this melt to be derived from an Archaean source. The porphyritic Otjitanda Granite was sampled just north of Otjitanda (sample Na 01/56, Table 1 and Fig. 2) where it is homogeneous and can be seen to intrude the banded migmatite (Fig. 6f). Our sample contains abundant long-prismatic, mostly idiomorphic zircons with well-developed oscillatory zoning (inset in Fig. 7b), although there are also numerous metamict high-U grains. Five

grains with bright CL-images were analyzed (Table 3) of which three are least disturbed and define a mean 207 Pb/206 Pb age of 1757 ± 4 Ma (Fig. 7b), whereas grain 4 is grossly discordant but has a similar 207 Pb/206 Pb ratio, thus suggesting significant recent Pb-loss (Table 3, not shown in Fig. 7b). Grain 1 is also very discordant (Table 3) but seems to have lost Pb at unspecified times in the past and is therefore not further considered here and is not shown in Fig. 7b. The best-fit line for the three least disturbed zircons (MSWD = 0.14, pof = 0.71) defines a Concordia intercept age of 1757 ± 10 Ma, identical to the mean 207 Pb/206 Pb age. If the strongly discordant analysis is added to the regression, the intercept age becomes 1757 ± 8 Ma. The 207 Pb/206 Pb mean age seems slightly younger than the melt-derived material of Na 01/54 but is, in fact, identical within analytical error. The agreement between these two ages suggests that the peak of amphibolite-facies metamorphism and anatectic granite-formation in the Epupa Complex of northwestern Namibia was reached at about 1760 Ma. The εNd(t) value for granite sample Na 01/56 is +3.2, the highest value of all samples analyzed in this study, and the corresponding Nd mean crustal residence age is 1.81 Ga (Table 5). This is only slightly older than the intrusive age, and the Nd isotopic systematics are therefore compatible with the anatectic nature of this granite and also rule out an Archaean age for the precursor material from which the granite was derived. Note than sample Na01/56 is relatively high in Nb and Y concentrations (Table 2) and plots in the within-plate-granite field of Pearce et al. (1984) (not shown graphically).

A. Kröner et al. / Precambrian Research 183 (2010) 50–69

Fig. 9. Histogram showing distribution of radiogenic lead isotope ratios derived from evaporation of 5 zircon grains from aplitic granite sample Na 01/51, Kaokoland, NW Namibia. The spectrum plotted has been integrated from 530 ratios. Mean age is given with 2-sigma (mean) error.

An undeformed and unmetamorphosed aplitic granite dyke was found cross-cutting the banded migmatite (Fig. 8a) near the track west of the Otjijandjasemo hot spring (sample Na 01/51, Table 1 and Fig. 2). The zircons of this sample are clear to medium brown, mostly long-prismatic and euhedral, and five grains were analyzed individually by the evaporation method. All analyses yielded identical 207 Pb/206 Pb ratios for which a mean age of 1670.1 ± 0.5 Ma was calculated (Table 4 and Fig. 9) and which we interpret to reflect the time of aplite dyke emplacement. This result unambiguously shows that the event of migmatite formation and strong ductile deformation in this part of the Epupa Complex was over by 1670 Ma ago and that the granulite-facies event as recorded in the Epembe Unit farther east did not affect the Epupa Complex in the region covered by this study. The εNd(t) value for aplitic dyke sample Na 01/51 is −1.9, and the corresponding Nd mean crustal residence age is 2.07 Ga (Table 5). Here again these Nd isotopic data are similar to those in samples Na 01/49 and Na 01/50 and suggest a similar source. A significantly older gneiss was found in a small pit near the track east of the township of Etanga (Fig. 2). This is a streaky, banded, non-migmatitic, hornblende-rich monzodioritic gneiss (Na 01/57, Table 1) whose field relationships with the migmatitic rocks could not be established due to poor outcrop conditions in this generally flat region. The zircons are mostly equant to stubby with well rounded terminations and display broad, striped zoning or sector zoning (see inset in Fig. 7c). These morphologies and

61

internal structures are quite different from the zircons in all the other gneisses analyzed here. Four grains were dated on SHRIMP II (Table 3 ) and show various degrees of discordance but have similar 207 Pb/206 Pb ratios with a mean age of 1861 ± 3 Ma, whereas the discordia line (MSWD = 0.10, pof = 0.90) defines an identical upper Concordia intercept age of 1861 ± 7 Ma (Fig. 7c). We interpret this to reflect the time of protolith emplacement. The εNd(t) value for hornblende-gneiss sample Na 01/57 is −1.0, and the DM model age of 2.26 Ga (Table 5) suggests a similar source to that of the previous samples. The entire area west of Etanga is made up of migmatitic gneisses interspersed with more porphyritic granite–gneiss. We sampled some of these gneiss varieties on a track from south of Otjitanda to the west towards Otjihaa (Fig. 2). Our sample Na 03/90 of granodioritic composition (Table 2, Fig. 3a) is from an isolated hill southwest of Etanga (Table 1 and Fig. 2) and represents the dark, biotite-rich palaeosome part of a fine-grained well layered migmatitic gneiss. The zircons are long-prismatic with well rounded terminations and show excellent oscillatory zoning (inset in Fig. 7d). Five grains were analyzed on SHRIMP II (Table 3) of which two are concordant, one is slightly reversely discordant and one is discordant (Fig. 7d). These four grains have identical 207 Pb/206 Pb ratios corresponding to a mean age of 1759.7 ± 3.5 Ma (Fig. 7d) which is the same as the other migmatitic gneisses farther north and northeast. The upper Concordia intercept age is 1759 ± 6 Ma (MSWD = 0.18, pof = 0.84). One further concordant result represents a grain morphologically identical to the others but much older at 1854 ± 8 Ma, and this is interpreted as a xenocryst. This age is similar to that for monzodiorite gneiss Na 01/57 and may suggest that at least some of the migmatitic gneiss precursors may be derived from melting of older crust. The εNd(t) value for granodioritic gneiss sample Na 03/90 is −0.6, and the corresponding mean crustal residence age is 2.12 Ga (Table 5). A similar source material is indicated as in the previous samples. The medium-grained palaeosome portion of a well foliated, slightly migmatitic granite–gneiss (Na 03/85) was sampled northeast of Otjitanda near the Ovireva Dam (Table 1 and Fig. 2). The zircons are mostly long-prismatic with slightly to well rounded terminations and oscillatory zoning (inset in Fig. 10a). Four grains were analyzed on SHRIMP II and are concordant or near-concordant with similar 207 Pb/206 Pb ratios (Table 3 and Fig. 10a) that provide a mean age of 1786 ± 7 Ma, whereas the Concordia intercept age (MSWD = 0.39, pof = 0.68) is 1785 ± 14 Ma. This is a little older than most other migmatitic gneisses in this region and testifies to the slight age variation in the Epupa Complex. A further sample is a well foliated, coarse, porphyritic granite–gneiss (Na 03/88) collected on a hill near the poor track to Otjihaa and about 2 km N of the village of Okazorwe (Table 1 and Fig. 2). The zircons are mostly stubby to rarely long-prismatic and show well-developed oscillatory zoning (inset in Fig. 10b). Five grains were analyzed (Table 3) and define a linear array with one concordant and four variably discordant data points (Fig. 10b). We ignore the most discordant result, and the remaining analyses have identical 207 Pb/206 Pb ratios defining a mean age of 1778 ± 3 Ma, whereas the Concordia intercept age is 1780 ± 14 Ma (MSWD = 0.03, pof = 0.97). We consider this to reflect the time of emplacement of the gneiss precursor. Another sample along the same track but farther southwest in hilly country is a well foliated and partly migmatitic granite–gneiss (Na 03/89, Table 1 and Fig. 2) that is interlayered with migmatite and unmigmatized granite–gneiss. The zircons are large and mostly long-prismatic with well rounded terminations and excellent oscillatory zoning. Three grains were analyzed (Table 3), all with identical 207 Pb/206 Pb ratios, of which two are concordant and one

62

A. Kröner et al. / Precambrian Research 183 (2010) 50–69

Table 3 Analytical data for SHRIMP analyses of zircons from the Epupa Complex, Kaokoland, northwest Namibia. Samle No.

U ppm

Th ppm

206Pb 204Pb

208Pb 206Pb

207Pb 206Pb

Na01/43-1 Na01/43-2 Na01/43-3 Na01/43-4 Na01/43-5 Na01/43-6

402 980 156 457 700 523

188 369 83 198 297 318

31374 125628 60423 38852 31917 32336

0.1306 0.1089 0.1549 0.1287 0.1232 0.1746

± ± ± ± ± ±

9 6 20 9 9 9

0.1079 0.0694 0.1075 0.1076 0.1020 0.1077

± ± ± ± ± ±

5 3 9 5 4 4

0.2962 0.1989 0.2815 0.2798 0.2272 0.3049

± ± ± ± ± ±

28 18 27 26 21 28

4.405 2.644 4.174 4.152 3.195 4.526

± ± ± ± ± ±

47 27 58 44 34 47

1672 1169 1599 1590 1320 1715

± ± ± ± ± ±

14 10 14 13 11 14

1713 1313 1669 1665 1456 1736

± ± ± ± ± ±

9 7 11 9 11 9

1764 1556 1758 1760 1660 1760

± ± ± ± ± ±

8 6 16 8 8 7

Na01/48-1 Na01/48-2 Na01/48-3 Na01/48-4 Na01/48-5

157 244 208 141 307

126 30 248 136 257

29732 8979 70003 921 4546

0.2356 0.0393 0.3514 0.2967 0.2481

± ± ± ± ±

22 10 17 38 16

0.1084 0.1085 0.1082 0.1081 0.1083

± ± ± ± ±

9 6 6 16 7

0.2522 0.2847 0.3154 0.2923 0.2228

± ± ± ± ±

19 21 23 22 16

3.769 4.258 4.704 4.359 3.328

± ± ± ± ±

46 42 46 75 33

1450 1615 1767 1653 1297

± ± ± ± ±

10 10 11 11 8

1586 1685 1768 1705 1488

± ± ± ± ±

10 8 8 14 8

1772 1774 1769 1768 1771

± ± ± ± ±

16 11 10 27 11

Na01/49-1 Na01/49-2 Na01/49-3 Na01/49-4 Na01/49-5 Na01/49-6

353 379 304 120 267 306

257 293 220 97 236 250

1667 62500 30453 53694 22457 7315

0.2388 0.2210 0.2141 0.2417 0.2576 0.2381

± ± ± ± ± ±

19 8 1 24 13 13

0.1079 0.1075 0.1074 0.1073 0.0770 0.1077

± ± ± ± ± ±

8 3 5 10 5 5

0.2701 0.3146 0.2921 0.2912 0.3103 0.3142

± ± ± ± ± ±

19 23 21 22 23 23

4.020 4.662 4.325 4.308 4.607 4.664

± ± ± ± ± ±

4 38 40 55 42 43

1542 1763 1652 1647 1742 1761

± ± ± ± ± ±

10 11 11 11 11 11

1638 1761 1698 1695 1751 1761

± ± ± ± ± ±

9 7 8 10 8 8

1765 1757 1756 1754 1760 1760

± ± ± ± ± ±

14 6 9 17 9 9

Na01/50-1 Na01/50-2 Na01/50-3 Na01/50-4 Na01/50-5 Na01/50-6 Na01/50-7

219 259 198 151 106 260 135

169 364 226 114 106 194 120

39551 25597 83126 38745 13176 13445 17950

0.2189 0.4116 0.3452 0.2403 0.2836 0.2076 0.2603

± ± ± ± ± ± ±

14 16 18 18 24 14 20

0.1094 0.1093 0.1096 0.1093 0.1095 0.1092 0.1093

± ± ± ± ± ± ±

6 5 6 8 10 6 8

0.3214 0.3163 0.3087 0.3086 0.3165 0.2950 0.3074

± ± ± ± ± ± ±

24 23 23 24 25 21 23

4.850 4.768 4.665 4.649 4.779 4.441 4.633

± ± ± ± ± ± ±

48 44 47 51 59 43 53

1797 1772 1735 1734 1773 1666 1728

± ± ± ± ± ± ±

12 11 11 12 12 11 11

1794 1779 1761 1758 1781 1720 1755

± ± ± ± ± ± ±

8 8 8 9 10 8 10

1790 1788 1793 1787 1791 1786 1788

± ± ± ± ± ± ±

10 9 11 13 16 10 14

Na01/54-1 Na01/54-2 Na01/54-3 Na01/54-4 Na01/54-5

120 146 161 242 118

104 180 125 269 130

14133 16326 10180 45190 22341

0.2567 0.3511 0.2469 0.3235 0.3196

± ± ± ± ±

19 19 19 15 22

0.1079 0.1079 0.1075 0.1077 0.1078

± ± ± ± ±

8 7 8 5 8

0.3094 0.3146 0.2952 0.3172 0.3097

± ± ± ± ±

45 46 43 46 45

4.605 4.681 4.373 4.708 4.603

± ± ± ± ±

78 77 74 75 79

1738 1763 1667 1776 1739

± ± ± ± ±

22 22 21 22 22

1750 1764 1707 1769 1750

± ± ± ± ±

14 14 14 13 15

1765 1765 1757 1760 1763

± ± ± ± ±

13 11 13 9 14

Na01/56-1 Na01/56-2 Na01/56-3 Na01/56-4 Na01/56-5

2773 306 150 544 527

905 379 146 129 765

1690 41459 18767 100 16233

0.1313 0.3715 0.2864 0.4010 0.4237

± ± ± ± ±

9 14 22 15 12

0.0663 0.1075 0.1073 0.1073 0.1074

± ± ± ± ±

4 5 9 65 4

0.1009 0.3120 0.3149 0.0561 0.2986

± ± ± ± ±

7 2 24 5 21

0.922 4.627 4.659 0.830 4.425

± ± ± ± ±

9 41 55 5 36

620 1751 1765 352 1684

± ± ± ± ±

4 11 12 3 11

664 1754 1760 614 1717

± ± ± ± ±

5 7 10 29 7

816 1758 1754 1754 1757

± ± ± ± ±

12 8 15 110 7

Na01/57-1 Na01/57-2 Na01/57-3 Na01/57-4

659 371 292 372

900 408 366 461

28956 44799 5232 48045

0.3956 0.3206 0.3778 0.3577

± ± ± ±

9 11 17 11

0.1139 0.1137 0.1139 0.1138

± ± ± ±

3 4 6 9

0.3284 0.3234 0.2920 0.3354

± ± ± ±

23 23 21 24

5.157 5.068 4.586 5.630

± ± ± ±

40 42 44 44

1831 1806 1652 1865

± ± ± ±

11 11 10 12

1846 1831 1747 1863

± ± ± ±

7 7 8 7

1863 1859 1863 1861

± ± ± ±

5 6 10 6

Na02/72-1 Na02/72-2 Na02/72-3 Na02/72-4 Na02/72-5

466 71 114 227 660

145 50 35 111 86

968992 11835 34031 26010 180865

0.0907 0.2085 0.0880 0.1388 0.0365

± ± ± ± ±

5 36 15 12 4

0.0821 0.0015 0.0825 0.0821 0.0823

± ± ± ± ±

3 15 8 6 3

0.2136 0.2141 0.2142 0.2143 0.2139

± ± ± ± ±

12 15 13 12 11

2.420 2.425 2.432 2.426 2.426

± ± ± ± ±

17 48 29 23 16

1248 1251 1251 1252 1250

± ± ± ± ±

6 8 7 6 6

1249 1250 1252 1250 1251

± ± ± ± ±

5 14 9 7 5

1249 1249 1254 1249 1252

± ± ± ± ±

7 35 18 13 7

Na03/80-1 Na03/80-2 Na03/80-3 Na03/80-4 N03/a80-5

341 279 600 477 400

6.8 66 681 521 402

26542 97475 201 21631 978

0.0056 0.0680 0.3642 0.3196 0.3048

± ± ± ± ±

8 7 52 12 23

0.0863 0.1082 0.1082 0.1079 0.1080

± ± ± ± ±

5 5 22 4 9

0.2205 0.3097 0.2033 0.3146 0.3104

± ± ± ± ±

20 29 19 29 29

2.625 4.619 3.032 4.682 4.623

± ± ± ± ±

31 50 71 49 62

1284 1739 1193 1763 1743

± ± ± ± ±

11 14 10 14 14

1308 1753 1416 1764 1753

± ± ± ± ±

9 9 18 9 11

1346 1769 1769 1765 1766

± ± ± ± ±

12 8 37 7 16

Na03/85-1 Na03/85-2 Na03/85-3 Na03/85-4

154 372 297 171

141 872 193 130

838926 6310 10251 126167

0.2656 0.6783 0.0962 0.2250

± ± ± ±

17 21 10 18

0.1091 0.1095 0.1090 0.1090

± ± ± ±

8 6 6 9

0.3082 0.3157 0.3152 0.3178

± ± ± ±

33 32 33 34

4.635 4.766 4.737 4.777

± ± ± ±

63 58 59 67

1732 1769 1766 1779

± ± ± ±

16 16 16 17

1756 1779 1774 1781

± ± ± ±

11 10 10 12

1784 1791 1783 1783

± ± ± ±

13 10 10 14

Na03/88.1 Na03/88.2 Na03/88.3 Na03/88.4 Na03/88.4

204 238 961 494 363

176 179 635 960 304

173130 21699 13571 30590 110400

0.2531 0.2288 0.1988 0.2639 0.2343

± ± ± ± ±

16 16 8 10 10

0.1088 0.1089 0.1038 0.1086 0.1087

± ± ± ± ±

7 7 4 4 5

0.3194 0.2999 0.2379 0.2927 0.3085

± ± ± ± ±

34 31 24 30 32

4.792 4.501 3.406 4.383 4.625

± ± ± ± ±

63 59 38 50 54

1787 1691 1376 1655 1733

± ± ± ± ±

12 15 12 15 16

1783 1731 1506 1709 1754

± ± ± ± ±

11 11 9 9 10

1780 1780 1693 1776 1778

± ± ± ± ±

12 12 7 7 8

Na03/89-1 Na03/89-2 Na03/89-3

206 191 293

166 141 409

611247 30797 125188

0.2388 ± 13 0.2263 ± 16 0.4286 ± 18

Na03/90-1 Na03/90-2 Na03/90-3 Na03/90-4 Na03/90-5

192 122 259 433 180

178 155 400 603 128

270270 36732 69037 84431 55072

0.2570 0.3497 0.3156 0.4045 0.1995

± ± ± ± ±

13 20 13 11 12

206Pb 238U

0.1083 ± 6 0.1080 ± 7 0.1082 ± 6 0.1078 0.1076 0.1077 0.1075 0.1133

± ± ± ± ±

5 7 5 3 5

207Pb 235U

0.3159 ± 33 0.3217 ± 34 0.2870 ± 30 0.3129 0.3193 0.3001 0.3126 0.3336

± ± ± ± ±

18 20 17 17 19

206/238 Age ± 1s

4.718 ± 59 4.792 ± 64 4.281 ± 53 4.650 4.736 4.460 4.635 5.212

± ± ± ± ±

37 46 33 30 41

1770 ± 16 1798 ± 17 1626 ± 15 1755 1786 1692 1753 1856

± ± ± ± ±

9 10 8 8 9

207/235 Age ± 1s

1770 ± 10 1784 ± 11 1690 ± 10 1758 1774 1724 1756 1855

± ± ± ± ±

7 8 6 5 7

207/206 Age ± 1s

1771 ± 10 1766 ± 12 1769 ± 10 1762 1759 1762 1758 1854

± ± ± ± ±

9 12 8 6 8

A. Kröner et al. / Precambrian Research 183 (2010) 50–69

63

Table 3 (Continued ) Samle No.

U ppm

Th ppm

206Pb 204Pb

208Pb 206Pb

207Pb 206Pb

Na03/93-1 Na03/93-2 Na03/93-3 Na03/93-4

612 458 644 647

421 311 489 434

170358 33628 136388 131389

0.1986 0.2091 0.2450 0.2029

± ± ± ±

7 10 9 7

0.1079 0.1077 0.1077 0.1078

± ± ± ±

3 4 3 3

0.3149 0.2922 0.2887 0.2880

± ± ± ±

29 27 26 26

4.684 4.337 4.288 4.279

± ± ± ±

47 46 43 43

1765 1652 1635 1631

± ± ± ±

14 13 13 13

1764 1700 1691 1689

± ± ± ±

8 9 8 8

1764 1760 1761 1762

± ± ± ±

5 7 6 6

Na03/97-1 Na03/97-2 Na03/97-3 Na03/97-4

195 137 809 178

115 124 508 146

4941 571755 2391 401284

0.1744 0.2639 0.1767 0.2426

± ± ± ±

22 16 12 15

0.1100 0.1103 0.1101 0.1100

± ± ± ±

10 7 6 7

0.3233 0.3441 0.2477 0.3131

± ± ± ±

34 37 25 33

4.905 5.233 3.759 4.751

± ± ± ±

74 70 45 62

1806 1906 1427 1756

± ± ± ±

17 18 13 16

1803 1858 1584 1776

± ± ± ±

13 11 10 11

1800 1804 1891 1800

± ± ± ±

17 12 10 11

is discordant (Fig. 10c). The mean age of 1769 ± 6 Ma and the Concordia intercept age of 1769 ± 10 Ma (MSWD = 0.18, pof = 0.67) are again virtually identical to the previous gneisses and migmatites. A muscovite-rich migmatitic granite–gneiss with particularly complex folding is exposed on a small dam N of Otjihaa (Table 1 and Fig. 2). Our sample Na 03/97 contains many stubby to rarely long-prismatic oscillatory-zoned zircons that show well rounded terminations (inset in Fig. 10d). Four grains were analyzed (Table 3) and are spread along a well-defined discordia line (MSWD = 0.056) with one grain concordant, one reversely discordant, and the other two discordant (Fig. 10d). However, as in all previous cases, recent Pb-loss probably caused the discordance, and the uniform 207 Pb/206 Pb ratios of all four analyses correspond to a mean age of 1801 ± 3 Ma and an identical but less precise Concordia inter-

206Pb 238U

207Pb 235U

206/238 Age ± 1s

207/235 Age ± 1s

207/206 Age ± 1s

cept age of 1802 ± 10 Ma (MSWD = 0.06, pof = 0.94, Fig. 10d). The reversely discordant grain plotting above the Concordia curve signifies U loss or Pb gain that is not commonly observed but has been previously reported on a 30 ␮m scale in SHRIMP analyses of zircons from metamorphic terrains (Williams et al., 1984; Compston and Kröner, 1988). Four additional zircons were evaporated individually and yielded identical 207 Pb/206 Pb ratios which provide a mean age of 1800.3 ± 0.4 Ma (Table 4, Fig. 10d, inset). This precise age, in view of the large number of isotopic ratios measured (Table 4), is interpreted to most closely reflect the time of protolith emplacement Our last few samples come from the western part of the Epupa Complex where the effect of Pan-African overprinting during the Kaoko orogeny is variably strong. This overprinting is due to oblique

Fig. 10. Concordia diagrams showing analytical data for SHRIMP II analyses of Epupa migmatitic granite–gneisses in western Kaokoland, NW Namibia. Data boxes as in Fig. 5. Sample number and short characterization is given in each diagram. For sample locations see Table 1 and Fig. 2.

64

A. Kröner et al. / Precambrian Research 183 (2010) 50–69

Table 4 Isotopic data from single grain zircon evaporation of samples from the Epupa Complex, Kaokoland, Namibia. Evaporation temp. in ◦ C

Mean 207 /Pb/206 Pb ratiob and 2-␴m error

207 Pb/206 Pb age and 2-␴m error

82 98

1598 1597

0.108254 ± 68 0.108229 ± 97

1770.4 ± 1.1 1769.8 ± 1.6

3 1–3

80 260

1600

0.108293 ± 74 0.108260 ± 48

1770.9 ± 1.2 1770.3 ± 0.8

Dark redbrown, Stubby to longPrismatic, ends Slightly rounded

1 2 3 4 5 6 1–6

138 147 96 175 111 102 769

1596 1598 1593 1604 1600 1599

0.107452 0.107481 0.107427 0.107408 0.107513 0.107474 0.107456

± ± ± ± ± ± ±

78 36 74 33 31 48 21

1756.6 1757.1 1756.2 1755.9 1757.7 1757.0 *1756.7

± ± ± ± ± ± ±

1.3 0.6 1.3 0.6 0.5 0.8 0.4

Dark redbrown, Stubby, ends little Rounded

1 2 3 4 5 1–5

88 85 106 170 81 530

1600 1598 1598 1596 1599

0.102516 0.102486 0.102506 0.102525 0.102535 0.102515

± ± ± ± ± ±

36 75 54 21 79 22

1670.2 1669.6 1670.0 1670.3 1670.5 *1670.1

± ± ± ± ± ±

0.6 1.4 1.0 0.4 1.4 0.5

Clear to yellowBrown, short- and Long-prismatic, ends Well rounded

1 2 3

84 95 174

1599 1597 1598

0.110057 ± 49 0.110032 ± 46 0.110046 ± 31

1800.3 ± 0.8 1799.9 ± 0.8 1800.2 ± 0.5

4 1–4

96 449

1597

0.110085 ± 44 0.110054 ± 27

1800.8 ± 0.7 1800.3 ± 0.4

Sample Number

Zircon colour and morphology

Na 01/48

Clear to light pink, Long-prismatic, ends Slightly rounded

Mean of 3 grains Na 01/49

Mean of 6 grains Na 01/51

Mean of 5 grains Na 03/97

Mean of 4 grains a b

Grain #

1 2

Mass scansa

Number of 207 Pb/206 Pb ratios evaluated for age assessment. Observed mean ratio corrected for non-radiogenic Pb where necessary. Errors based on uncertainties in counting statistics.

transpression during collision of terranes of the Kaoko orogen with the southwestern margin of the Congo craton (Goscombe et al., 2003; Goscombe and Gray, 2008) and has given rise to the spectacular Puros shear belt (Fig. 11) in which Neoproterozoic rocks of the Kaoko belt are tectonically juxtaposed against reworked Epupa gneisses. The Kaoko rocks in low-strain domains of the Puros shear belt are metamorphosed to upper greenschist- to amphibolitefacies, and these rocks as well as the Epupa gneisses show all transitions from well foliated gneisses to ductile ultramylonites with accompanying retrogression. A particularly well-exposed example of this overprinting can be seen in the cliffs just west of the bottom of Van Zyl’s Pass where our sample Na 01/48 was collected (Table 1 and Fig. 2). The migmatitic Epupa granitic augen-gneiss here is completely refoliated in a northnorthwest-south-southeast direction with well-preserved shear-sense indicators confirming sinistral shear sense (Fig. 8b). None of the Palaeoproterozoic structures are preserved. A few metres away from this locality shearing is even more severe and has transformed the augen-gneiss into a north-south striking mylonite (Fig. 8c). Sample Na 01/48 contains mostly long-prismatic zircons with slightly to well-rounded terminations and variably preserved oscillatory zoning (inset in Fig. 12a). The late Neoproterozoic overprinting does not seem to have disturbed the isotopic system in these grains, because five zircons analyzed all have similar 207 Pb/206 Pb ratios (Table 3) although four analyses are discordant and suggest recent Pb-loss (Fig. 12a). The mean 207 Pb/206 Pb age of 1771 ± 3 Ma is within the same range as those for the other Epupa gneisses, and a discordia line (MSWD = 0.05, pof = 0.99) defines an upper Concordia intercept at 1771 ± 11 Ma. Three additional zircons were evaporated and yielded identical 207 Pb/206 Pb isotopic ratios for which the mean age is 1770.3 ± 0.8 Ma (Table 4 and Fig. 12a, inset). We adopt a protolith crystallization age of 1770 Ma for this reworked gneiss which is clearly part of the Epupa Complex. The εNd(t) value for granitic augen-gneiss sample Na 01/48 is −1.7, and the corresponding mean crustal residence age is 2.26 Ga

(Table 5). These values confirm that the gneiss was derived from a similar source as the migmatitic gneisses farther east. A further example of a refoliated augen-gneiss occurs at the confluence of the Otjindjani River with the Kunene River (Fig. 2, Table 1) where a reddish, grey-brown, strongly N-S foliated gneiss derived from a porphyritic granite (Fig. 8d, sample Na 03/93) is interlayered with grey-green migmatite. The relatively rare zircons with vague oscillatory zoning are stubby and frequently fragmented and most have bright, narrow rims that could not be analyzed (inset in Fig. 12b). Three grain cores were measured on SHRIMP II, and the pattern is similar to the other gneisses. One grain is concordant whereas two grains are discordant, but all have similar 207 Pb/206 Pb isotopic ratios (Table 3) suggesting recent Pb-loss, and the mean age is 1762 ± 2 Ma (Fig. 12b), again virtually identical with many of the ages reported above. The best-fit line for four grains (MSWD = 0.02, pof = 0.99) defines a Concordia intercept age of 1764 ± 11 Ma. The εNd(t) value for this red augen-gneiss sample Na 01/93 is +0.7, and the corresponding mean crustal residence age is 2.05 Ga (Table 5). Again, these values confirm that the gneiss was derived from a similar source as the migmatitic gneisses farther east although its appearance in the field is quite different from the grey gneisses farther east. Finally, we dated a granodioritic gneiss from the hilly terrain west of the broad Marienfluss valley and known as the Hartmann Mountains. This is a strongly deformed zone mapped as the Hartmann Domain of the Kaoko belt by Goscombe et al. (2005) and Goscombe and Gray (2008) and situated west of the Puros Mylonite Zone (Fig. 11b). It consists predominantly of Neoproterozoic turbiditic metasedimentary rocks with amphibolitic bodies. Goscombe et al. (2005) reported ages between 572 and 507 Ma for late metamorphic granites and metamorphic minerals, reflecting the strong Pan-African imprint. Goscombe and Gray (2008) mapped one narrow body of Palaeoproterozoic granite–gneiss close to the Kunene River (Fig. 11b). We dated a strongly deformed and north-northwest-south-southeast elongated narrow sliver of granodioritic gneiss (Na 01/43) from another such body farther

A. Kröner et al. / Precambrian Research 183 (2010) 50–69

65

Fig. 11. (a) Satellite image showing spectacular Puros shear belt in northwestern Kaokoland and southern Angola as a result of transpressive deformation on the margin of the Congo Craton. Dark rocks on upper right are part of the Palaeoproterozoic Epupa Complex which becomes reworked to the west in the Pan-African Kaoko belt. Broad valley on right is Marienfluss, isolated hills on left are Hartmann Mts. Kunene River is dark, thin band in upper part of image. (b) Geological map of part of same image showing tectonic boundary between Pan-African Kaoko belt and Palaeoproterozoic Epupa Complex in northwerstern Namibia (from Fig. 4 of Goscombe and Gray, 2008). TPMZ is Three Palms Mylonite Zone; HMZ is Hartmann Mylonite Zone; PMZ is Puros Mylonite Zone. For details of geology see Goscombe and Gray (2008). Location of samples Na 01/43 and 3/93 is indicated.

southeast (Figs. 2, 11b), which is tectonically interlayered with amphibolite-facies Damara meta-turbidites and which, in the field, we interpreted as an early syntectonic granitoid of the Kaoko Belt. All original contacts with the enclosing meta-turbidite are overprinted by strong shearing due to the Kaoko transpressional event. The zircons are predominantly long-prismatic with well rounded terminations and excellent oscillatory zoning (inset in Fig. 12c). Seven grains were dated on SHRIMP II, and all results are discordant (Table 3 and Fig. 12c). Five grains have identical 207 Pb/206 Pb isotopic ratios and are therefore well aligned in the Concordia diagram (MSWD = 0.101) and, as in all previous cases, indicate recent Pb-loss. The mean 207 Pb/206 Pb age is 1763 ± 4 Ma, whereas the Concordia intercept age is less precise at 1762 ± 17 Ma (MSWD = 0.20,

pof = 0.90). The two additional analyses marked in black in Fig. 12c suggest Pb-loss at around 545 Ma, in line with the late Neoproterozoic Kaoko event dated by Goscombe et al. (2005), but we do not consider these analyses for age assessment. The above age, interpreted as reflecting the time of protolith emplacement, is indistinguishable from the other gneiss/migmatite ages reported above and clearly identify the original granodiorite as part of the Epupa Complex. The εNd(t) value for granodioritic gneiss sample Na 01/43 is +0.6, and the corresponding mean crustal residence age is 2.00 Ga (Table 5). As in all previous samples, these values confirm that the gneiss was derived from a similar source as the migmatitic gneisses farther east and belongs to the Epupa Complex.

66

A. Kröner et al. / Precambrian Research 183 (2010) 50–69

Fig. 12. Concordia diagrams showing analytical data for SHRIMP II analyses of Epupa granite–gneisses reworked during the Pan-African Kaoko event in western Kaokoland, NW Namibia. Data boxes as in Fig. 5. Inset in (a) shows histogram with distribution of radiogenic lead isotope ratios derived from evaporation of 3 single zircons from sample Na 01/48, integrated from 260 ratios. Sample number and short characterization is given in each diagram. For sample locations see Table 1 and Fig. 2.

Fig. 13. Nd isotopic evolution diagram for dated samples of the Epupa Complex, Namibia.

A. Kröner et al. / Precambrian Research 183 (2010) 50–69

67

Table 5 Whole-rock Sm–Nd isotopic data for granitoid samples of the Epupa complex. Sample

Rock type

Age [Ma]

Sm [␮g/g]

Nd [␮g/g]

147

Sm/144 Nd

Na 01/43 Na 01/48 Na 01/49 Na 01/50 Na 51 Na 54 Na 55 Na 56 Na 57 Na 03/90 Na 03/93

Granodiorite gneiss Migmatitic gneiss Augen-gneiss Granitic palaeosome Aplitic dyke Monzonitic melt patch Otjitanda granite Otjitanda granite Monzodiorite gneiss Granod. gneiss Augen-gneiss

1763 1770 1758 1790 1670 1763 1760 1760 1865 1760 1762

5.873 5.331 9.484 8.909 3.027 7.456 16.83 19.54 2.100 7.560 17.38

34.14 26.12 52.54 47.97 19.27 33.03 104.0 124.1 10.55 39.77 85.93

0.1040 0.1234 0.1091 0.1123 0.09496 0.1365 0.09785 0.09520 0.1204 0.1149 0.1223

143

Nd/144 Nd [m]

0.511593 0.511699 0.511559 0.511610 0.511424 0.511813 0.511568 0.511628 0.511653 0.511663 0.511809

± ± ± ± ± ± ± ± ± ± ±

11 12 10 11 11 12 12 12 12 12 12

143

Nd/144 Nd [t]

0.510387 0.510262 0.510297 0.510288 0.510381 0.510231 0.510435 0.510526 0.510176 0.510332 0.510392

εNd[t]

tDM

0.6 −1.7 −1.3 −0.7 −1.9 −2.5 1.5 3.2 −1.0 −0.6 0.7

2.00 2.26 2.15 2.14 2.07 2.42 1.93 1.81 2.26 2.12 2.05

143 Nd/144 Nd ratios are normalized to 146 Nd/144 Nd = 0.7219. 143 Nd/144 Nd ratios are relative to 143 Nd/144 Nd = 0.511847 ± 8 (2 SD, N = 10) in the La Jolla Nd standard. m = measured ratio; within-in run precision (2 SE) referring to the last digits of the ratio; t = initial value. Parameters for chondrite-uniform reservoir: 143 Nd/144 Nd = 0.512638 and 147 Sm/144 Nd = 0.1967 (Jacobsen and Wasserburg, 1980). tDM = depleted mantle Nd model age (DePaolo, 1988); n.c. = sample composition not suited for model age calculation.

It seems obvious, therefore, that the Neoproterozoic turbiditic sequence of the Hartmann Domain was originally underlain by basement rocks of the Epupa Complex. It is likely that basement slivers were tectonically interdigitated with the tubidites during the Kaoko transpressional event. 5. Discussion and conclusions Our field observations and age data from a hitherto undated and largely unmapped basement terrain in northwestern Namibia have led to recognition of a widespread late Palaeoproterozoic structural/metamorphic event that affected a large area of predominantly granitoid gneisses. Remarkably, this terrain is almost completely devoid of supracrustal rocks and only contains isolated bodies of gabbro, suggesting that the gneisses and migmatites were derived from a very large granitoid batholith that extended into southern Angola. The crust into which the original granitoid rocks were intruded is nowhere exposed, and we failed to find remnants of Archaean crust or Archaean zircon xenocrysts. The Nd isotopic systematics (Table 5) also rule out derivation of the original Epupa granitoids from Archaean crust. The available age data for the gneiss–migmatite complex in southern Angola also do not suggest any Archaean inheritance. We therefore reiterate our suggestion that the Archaean rocks previously identified farther south in the Hoanib River near Sesfontein (Seth et al., 1999) either define a crustal segment of the southwestern Congo Craton that is different from the Epupa Complex discussed here or represents a separate Archaean to Palaeoproterozoic exotic terrane incorporated into the Kaoko belt during transpressional convergence (Seth et al., 1999). The Epupa granitoid rocks show transitional meta to peraluminous, calc-alkaline compositions (Taylor and McLennan, 1985). High to moderate La/Y ratios suggest melting of crustal protoliths in the garnet-stability field as well as at shallower depths. It is surprising that the zircons analyzed in this study largely preserved their original isotopic composition, despite amphibolite-facies metamorphism and ductile deformation, and only experienced variable Pb-loss in recent times. Pb-loss related to the late Neoproterozoic Kaoko orogenic event is only obvious in some rocks in the western part of the Epupa Complex where Pan-African overprinting was severe. The Epupa granitoids in the present study area were emplaced between ca. 1860 and 1760 Ma ago and were tectonized and metamorphosed almost immediately after their formation, as documented by the ages for a former melt patch in migmatite and for the Otjitanda Granite of ∼1760 Ma. This also implies that granitoid generation, subsequent deformation into gneisses, and prograde metamorphism leading to migmatite formation, partial anatexis and, finally, complete anatexis to form the Otjitanda Granite, all occurred within the same late Palaeoproterozoic oro-

genic event. The relatively short duration of this event, coupled with the geochemical and Nd-isotopic signature of the dated samples, suggests that the Epupa gneisses were derived from a large (subduction-related?) granitoid batholith that may have been part of a Palaeoproterozoic magmatic arc. The Nd isotopic systematics of all samples are surprisingly similar (Fig. 13) and suggest formation of the protolith from a source region that either separated from the depleted mantle about 2.2–2.0 Ga ago or represents a mixture of juvenile magma with minor older continental crust. A major Archaean component in the source area seems unlikely, as also shown by a lack of Archaean zircon xenocrysts. The highest initial εNd values of 1.5 and 3.2 found in the Otjitanda Granite contrast with the mostly negative values in the melted surrounding migmatitic gneisses. This relationship suggests a major juvenile crustal source for the Otjitanda Granite protolith in addition to melting of the surrounding gneisses. In terms of age, this event is generally referred to in Africa as the Eburnian orogeny (Roques, 1948) which has affected large parts of the continent (Kröner, 1977; Cahen et al., 1984) but whose nature and tectonic setting in the Congo craton of southwestern Africa is largely unknown. Tectono-magmatic-metamorphic events in the same age bracket are also known from many other parts of the world (Goodwin, 1991) and have led to the now popular speculation that they may be related to formation of the hypothetical Palaeoproterozoic supercontinent Nuna (Hoffman, 1987; Reddy and Evans, 2009) or Columbia (Rogers and Santosh, 2002). However, the position of the Epupa Complex in this puzzle is completely uncertain because its tectonic setting and evolution cannot be reconstructed with confidence without associated supracrustal assemblages which are largely developed in southern Angola but are poorly studied there (Cahen et al., 1984; Carvalho et al., 2000). Condie et al. (2009) compiled global zircon age spectra for Precambrian granitoid rocks but did not have any age data for southwestern Africa. In their compilation there is a marked peak in granite generation in Central Africa at about 1830 Ma and, looking at the continent as a whole, there is a minor peak at ∼1900 Ma (see Condie et al., 2009, their Figs. 6 and 10). Our present study documents a peak in granitoid activity between ∼1760 and ∼1800 Ma, and this seems most compatible with similar peaks in Australia and Laurentia (Condie et al., 2009, their Fig. 6). Whether this means a possible past continental connection between the southwestern Congo craton and either of these late Palaeoproterozoic continents remains to be established. Acknowledgements We thank Bruce Eglington, Clark Friend and an anonymous reviewer for improvements of the manuscript. This project was funded by Deutsche Forschungsgemeinschaft (KR 590/89-1) and

68

A. Kröner et al. / Precambrian Research 183 (2010) 50–69

Mainz University. A.K. acknowledges a Gledden Senior Fellowship of the University of Western Australia during which part of the SHRIMP analyses were performed in 2005. Y.R.-A. acknowledges a Humboldt-Foundation Georg Forster Fellowship. Analytical facilities for zircon evaporation were kindly provided by the Max-Planck Institut für Chemie in Mainz. Linda Iaccheri carried out the Sm–Nd isotope analyses at the University of Munich, and Petra Sigl prepared Figs. 1 and 2. We thank the Geological Survey of Namibia for providing field support. This is a contribution to IGCP Project 509 and publication no. 688 of the Mainz Research Centre “Earth System Science”. Appendix A. Appendix: Sample preparation and analytical procedures Heavy minerals were separated from whole-rock samples by standard procedures using jaw crusher, steel rolling mill, Wilfleytable, Franz magnetic separator and heavy liquids. Zircons were then hand picked and either mounted in epoxy resin together with chips of the Perth Consortium standard CZ3 for SHRIMP analyses or selected for evaporation. A.1. SHRIMP II procedure The mount was ground down and polished so that the zircon cores were exposed, and zircons were photographed in reflected and transmitted light and under cathodoluminescence (CL) to enable easy and best location on the mount during SHRIMP analyses. The mount was then cleaned and gold-coated. Isotopic analyses were performed on the Perth Consortium SHRIMP II ion microprobe whose instrumental characteristics were described by De Laeter and Kennedy (1998). The analytical procedures are outlined in Compston et al. (1992), Claoué-Long et al. (1995), Nelson (1997) and Williams (1998). Prior to each analysis, the surface of the analysis site was precleaned by rastering of the primary beam for 2–3 min, to reduce or eliminate surface common Pb. The reduced 206 Pb/238 U ratios were normalized to 0.09432, which is equivalent to the adopted age of 564 Ma for CZ3. Pb/U ratios in the unknown samples were corrected using the ln(Pb/U)/ln(UO/U) relationship as measured in standard CZ3 and as outlined in Compston et al. (1984) and Nelson (1997). The 1 error in the ratio 206 Pb/238 U during analysis of all standard zircons during this study was between 1.03 and 1.43%. Primary beam intensity was between 2.2 and 3.8 nA, and a Köhler aperture of 100 ␮m diameter was used, giving a slightly elliptical spot size of about 30 ␮m. Peak resolution was about 5000, enabling clear separation of the 208 Pb-peak from the HfO peak. Sensitivity varied between 19 and 36 cps/ppm Pb. Analyses of samples and standards were alternated to allow assessment of Pb+/U+ discrimination. Raw data reduction and error assessment followed the method described by Nelson (1997). Common Pb corrections have been applied using the 204 Pb-correction method and assuming the isotopic composition of Broken Hill, because common Pb is thought to be surface-related (Kinny, 1986). The analytical data are presented in Table 1. Errors given on individual analyses are based on counting statistics and are at the 1␴ level and include the uncertainty of the standard added in quadrature. Errors for pooled analyses are at 2␴. The ages and errors of intercepts of the best-fit line with concordia were calculated using the Isoplot program (Ludwig, 2003). MSWD- and probability of fit (pof)-values are given in the text and on the concordia diagrams. A.2. Single zircon evaporation We used the method developed by Kober (1986) involving repeated evaporation and deposition of Pb from chemically

untreated single zircons in a double-filament arrangement (Kober, 1987). Our laboratory procedures as well as comparisons with conventional and ion-microprobe zircon dating are detailed in Kröner and Todt (1988) and Kröner and Hegner (1998). Isotopic measurements were carried out on a Finnigan-MAT 261 mass spectrometer at the Max-Planck-Institut für Chemie in Mainz. The calculated ages and uncertainties are based on the means of all ratios evaluated and their 2 mean errors. Mean ages and errors for several zircons from the same sample are presented as weighted means of the entire population. During the course of this study we repeatedly analyzed fragments of large zircon grains from the Palaborwa Carbonatite, South Africa. These zircons, used as an internal standard, are euhedral, colourless to slightly pink, and completely homogeneous when examined under cathodoluminescence. Conventional U–Pb analyses of six separate grain fragments from this sample yielded a 207 Pb/206 Pb age of 2052.2 ± 0.8 Ma (2, W. Todt, unpublished data), whereas the mean 207 Pb/206 Pb ratio for 18 grains, evaporated individually over a period of 12 months, is 0.126634 ± 0.000027 (2 error of the population), corresponding to an age of 2051.8 ± 0.4 Ma and identical to the TIMS U–Pb age. The above error is considered the best estimate for the reproducibility of our evaporation data and corresponds approximately to the 2 (mean) error reported for individual analyses in this study (Table 3). In the case of large combined data sets the 2 (mean) error may become very low, and whenever this error was less than the reproducibility of the internal standard, we have used the latter value (that is, an assumed 2 error of 0.000027 for the 207 Pb/206 Pb isotopic ratio). The analytical data are presented in Table 3, and the 207 Pb/206 Pb spectra are shown in histograms that permit visual assessment of the data distribution from which the ages are derived. The evaporation technique provides only Pb isotopic ratios, and there is no a priori way to determine whether a measured 207 Pb/206 Pb ratio reflects a concordant age. Thus, principally, all 207 Pb/206 Pb ages determined by this method are necessarily minimum ages. However, many studies have demonstrated that there is a very strong likelihood that these data represent true zircon crystallization ages when (1) the 207 Pb/206 Pb ratio does not change with increasing temperature of evaporation and/or (2) repeated analyses of grains from the same sample at high evaporation temperatures yield the same isotopic ratios within error. Compara-tive studies by single grain evaporation, conventional U–Pb dating, and ion-microprobe analysis have shown this to be correct (see Kröner and Hegner, 1998, for discussion and references). A.3. Sm and Nd isotope analyses Whole-rock powders were and spiked with a 149 Sm-150 Nd tracer solution and decomposed in a mixture of 2 ml concentrated HF and ∼50 ␮l HClO4 in steel-lined PTFE-bombs at ∼170 ◦ C over a period of one week. After drying the sample solution, the fluorides/perchlorates were converted in the same bomb at ∼170 ◦ C for one day. This was followed by repeated treatment of the fluoride/perchlorate residue with 6 N HCl in PFA vessels until the sample cake was completely dissolved. Chromatographic element separation followed the procedure described in Hegner et al. (1995). Nd and Sm were loaded on Re-filaments using diluted phosphoric acid and measured as metals in a double-filament configuration. Total procedural blanks of <200 pg for Nd and Sm are not significant for the analyzed concentration levels. Sm and Nd isotope abundance ratios were determined on a MAT 261 mass spectrometer at Munich University, upgraded with new high voltage- and filament current supply units, low-noise amplifier resistors, and a new magnetic field probe by Spectromat® , Bremen. The Nd isotopic abundances were measured employing a triple-collector, and the monitoring of

A. Kröner et al. / Precambrian Research 183 (2010) 50–69

interfering 144 Sm. The 143 Nd/144 Nd ratios were normalized to 146 Nd/144 Nd = 0.7219. The La Jolla Nd reference material yielded 143 Nd/144 Nd = 0.511847 ± 0.000008 (2 SD of population, N = 10) during this study. The Nd reference material JNdi was measured at 143 Nd/144 Nd = 0.712109 ± 0.000003 (2 SD, N = 4). The long-term external precision for 143 Nd/144 Nd is ∼10−5 (2 SD). The external error of the 147 Sm/144 Nd ratio was determined as ∼0.02% (2 SD) using the CIT Sm/Nd standard solution. The analyzed 147 Sm/144 Nd ratio in BCR-1 was 0.1381 ± 0.0001 (2 SD; N = 10). All uncertainties reported in this study are given at the 95% confidence interval. We interpret the Nd model ages in terms of mean crustal residence ages (Arndt and Goldstein, 1987). References Arndt, N.T., Goldstein, S.L., 1987. Use and abuse of crust-formation ages. Geology 15, 893–895. Brandt, S., Klemd, R., Okrusch, M., 2003. Ultrahigh-temperature metamorphism and multistage evolution of garnet–orthopyroxene granulites from the Proterozoic Epupa Complex, NW Namibia. J. Petrol. 44, 1121–1144. Brandt, S., Will, T.M., Klemd, R., 2007. Ultrahigh-temperature metamorphism and anticlockwise PT paths of sapphirine-bearing orthopyroxene-sillimanite gneisses from the Proterozoic Epupa Complex, NW Namibia. Precambrian Res. 153, 143–178. Cahen, L., Snelling, N.J., Delhal, J., Vail, J.R., 1984. The Geochronology and Evolution of Africa. Clarendon Press, Oxford, 512 pp. Carvalho, H. de, 1982. Geologia de Angola (Geological map of Angola) 1:1,000,000. Instituto Investigacao Cientifica Tropical (Centro Geologia), Lisbon, Portugal. Carvalho, H., Alves, P., 1993. The Precambrian of SW Angola and NW Namibia. Inst. Inv. Cientifica Tropical, Série de Ciencias da Terras. Communicacoes 4, 1–38. de Carvalho, H., Tassinari, C.J., Alves, P.H., Guimarães, F., Simões, M.C., 2000. Geochrono-logical review of the Precambrian in western Angola: links with Brazil. J. Afr. Earth Sci. 31, 383–402. Claoué-Long, J.C., Compston, W., Roberts, J., Fanning, C.M., 1995. Two Carboniferous ages: a comparison of SHRIMP zircon dating with conventional zircon ages and 40 Ar/39 Ar analyses. Soc. Sediment. Geol., Spec. Publ. 54, 3–21. Compston, W., Williams, I.S., Myer, C., 1984. U–Pb geochronology of zircons from Lunar Breccia 73217 using a sensitive high mass-resolution ion-microprobe. J. Geophys. Res. 89 (Suppl.), B525–B534. Compston, W., Kröner, A., 1988. Multiple zircon growth within early Archaean tonalitic gneiss from the Ancient Gneiss Complex, Swaziland. Earth Planet. Sci. Lett. 87, 13–28. Compston, W., Williams, I.S., Kirschvink, J.L., Zhang, Z., Ma, G., 1992. Zircon U–Pb ages for the Early Cambrian time scale. J. Geol. Soc. London 149, 171–184. Condie, K.C., Belousova, E., Griffin, W.L., Sircombe, N., 2009. Granitoid events in space and time: constraints from igneous and detrital zircon age spectra. Gondwana Res. 15, 228–242. De Laeter, J.R., Kennedy, A.K., 1998. A double focusing mass spectrometer for geochronology. Int. J. Mass Spectrom. 178, 43–50. Drüppel, K., Littmann, S., Romer, R.L., Okrusch, M., 2007. Petrology and isotope geochemistry of the Mesoproterozoic anorthosite and related rocks of the Kunene Intrusive Complex, NW Namibia. Precambrian Res. 156, 1–31. Frost, B.R., Barnes, C.G., Collins, W.J., Arculus, R.J., Ellis, D.J., Frost, C.D., 2001. A geochemical classification for granitic rocks. J. Petrol. 42, 2033–2048. Miller, R.McG., Schalk, K.E.L. (Eds.), 1980. Geological Map of Namibia 1:1 000 000. Geological Survey, Namibia, Windhoek. Geological Map of Namibia, Sheet 1712 – Swartbooisdrif, 2002. Geol. Survey, Namibia, Windhoek. Goodwin, A.M., 1991. Precambrian Geology. The Dynamic Evolution of the Continental Crust. Academic Press, London, UK, 666 pp. Goscombe, B.D., Hand, M., Gray, D., 2003. Structure of the Kaoko Belt, Namibia: progressive evolution of a classic transpressional orogen. J. Struct. Geol. 25, 1049–1081. Goscombe, B., Gray, D., Armstrong, R., Foster, D.A., Vogl, J., 2005. Event geochronology of the Pan-African Kaoko Belt, Namibia. Precambrian Res. 140, 103.e1–103.e41. Goscombe, B., Gray, D., 2008. Structure and strain variation at mid-crustal levels in a trans-pressional orogen: a review of Kaoko Belt structure and the character of West Gondwana amalgamation and dispersal. Gondwana Res. 13, 45–85. Hegner, E., Walter, H.J., Satir, M., 1995. Pb–Sr–Nd isotopic compositions and trace element geochemistry of megacrysts and melilitites from the Tertiary Urach volcanic field: source composition of small volume melts under SW Germany. Contrib. Mineral. Petrol. 122, 322–335. Hoffman, P.F., 1987. Tectonic genealogy of North America. In: Van der Pluijm, B.A., Marshak, S. (Eds.), Earth Structure: An Introduction to Structural Geology and Tectonics. McGraw-Hill, New York, pp. 459–464. Jacobsen, S.B., Wasserburg, G.J., 1980. Sm–Nd isotopic evolution in chondrites. Earth Planet. Sci. Lett. 90, 315–329. Kinny, P.D., 1986. 3820 Ma zircons from a tonalitic Amitsoq gneiss in the Godthab district of southern West Greenland. Earth Planet. Sci. Lett. 79, 337–347. Kober, B., 1986. Whole-grain evaporation for 207 Pb/206 Pb-age-investigations on single zircons using a double-filament thermal ion source. Contrib Mineral. Petrol. 93, 482–490.

69

Kober, B., 1987. Single-zircon evaporation combined with Pb+ emitter-bedding for 207 Pb/206 Pb-age investigations using thermal ion mass spectrometry, and implications to zirconology. Contrib. Mineral. Petrol. 96, 63–71. Kröner, A., 1977. The Precambrian geotectonic evolution of Africa: plate accretion versus plate destruction. Precambrian Res. 4, 163–213. Kröner, A., Todt, W., 1988. Single zircon dating constraining the maximum age of the Barberton greenstone belt, Southern Africa. J. Geophys. Res. 93, 15329–15337. Kröner, A., Jaeckel, P., Williams, I.S., 1994. Pb-loss patterns in zircons from a highgrade metamorphic terrain as revealed by different dating methods: U–Pb and Pb–Pb ages for igneous and metamorphic zircons from northern Sri Lanka. Precambrian Res. 66, 151–181. Kröner, A., Hegner, E., 1998. Geochemistry, single zircon ages and Sm–Nd systematics of granitoid rocks from the Góry Sowie (Owl) Mts., Polish West Sudetes: evidence for early Palaeozoic arc-related plutonism. J. Geol. Soc. London 155, 711–724. Kröner, S., Konopasek, J., Kröner, A., Passchier, C.W., Poller, U., Wingate, M.T.D., Hofmann, K.H., 2004. U–Pb and Pb–Pb zircon ages for metamorphic rocks in the Kaoko Belt of northwestern Namibia: a Palaeo- to Mesoproterozoic basement reworked during the Pan-African orogeny. S. Afr. J. Geol. 107, 455–476. Kröner, A., Rojas-Agramonte, Y., Wingate, M.T.D., Liu, D.Y., 2006. The Epupa Complex of northwestern Namibia: Late Palaeoproterozoic to Mesoproterozoic granitoid magmatism along the southwestern margin of the Congo craton and Pan-African reworking of a craton margin. Abstr.-vol. In: 3rd Int. Symposium on Gondwana to Asia, University of Hong Kong, pp. 21–22. Ludwig, K.R., 2003. User’s Manual for ISOPLOT/Ex 3.0. A Geochronological Toolkit for Microsoft Excel. Berkeley Geochronology Center Special Publ. 4, 70 p. Martin, H., 1965. The Precambrian Geology of South West Africa and Namaqualand. Precambrian Research Unit, Univ. Cape Town, 159 pp. McCourt, S., Armstrong, R.A., Kampunzu, A.B., Mapeo, R.B.M., Morais, E., 2004. New U–Pb SHRIMP ages on zircons from the Lubango region, southwest Angola: insights into the Proterozoic evolution of south-western Africa. Geoscience Africa 2004, Abstr.-Vol., Univ. Witwatersrand, Johannesburg, South Africa, 438–439. Mehnert, K.R., 1968. Migmatites and the Origin of Granitic Rocks. Elsevier, Amsterdam, 393 pp. Middlemost, E.A.K., 1985. Naming materials in the magma/igneous rock system. Earth Sci. Rev. 37, 215–224. Miller, R.McG., 2008. The Geology of NAMIBIA, vol. 1 (Archaean to Mesoproterozoic). Ministry of Mines and Energy, Geological Survey, Windhoek, Namibia, 631 pp. Miller, R.McG., Schalk, K.E.L., 1980. Geological Map of Namibia, Scale 1:1 Million. Geological Survey of Nambia, Windhoek, Namibia. Nelson, D.R., 1997. Compilation of SHRIMP U–Pb zircon Geochronology Data, 1996. Geol Surv Western Australia, Record 1997/2, 189 pp. O’Connor, J. T., 1965. A Classification for quartz-rich igneous rocks based on feldspar ratios. U.S. Geol. Survey Prof. Paper B525, pp. 79–84. Pearce, J.A., Harris, N.W., Tindle, A.G., 1984. Trace element discrimination diagrams for the tectonic interpretation of granitic rocks. J. Petrol. 25, 956–983. Reddy, S.M., Evans, D.A.D., 2009. Palaeoproterozoic supercontinents and global evolution: correlations from core to atmosphere. In: Reddy, S.M., Mazumder, R., Evans, D.A.D., Collins, A.S. (Eds.), Palaeoproterozoic Supercontinents and Global Evolution, vol. 323. Geol. Soc, London, Spec. Publ., pp. 1–26. Rogers, J.J.W., Santosh, M., 2002. Configuration of Columbia, a Mesoproterozoic supercontinent. Gondwana Res. 5, 5–22. Roques, M., 1948. Le Précambrian de l’Afrique occidentale francaise. Bull. Soc. Géol. France 18, 869–875. Schalk, K.E.L., 1985. Preliminary note on the Okapuka Formation of northwestern Kaokoland. Unpubl. Report, Geological Survey of Namibia, 8 p. Schreiber, U.M., 2002. Provisional Geological Map 1712 – Swartbooisdrif, scale 1:250,000. Geological Survey of Namibia, Windhoek, Namibia (unpublished). Seth, B., Kröner, A., Mezger, K., Nemchin, A.A., Pidgeon, R.T., Okrusch, M., 1999. Archaean to Neoproterozoic magmatic events in the Kaoko belt of NW Namibia and their geodynamic significance. Precambrian Res. 92, 341–363. Seth, B., Armstrong, R.A., Brandt, S., Villa, I.M., Kramers, J.D., 2003. Mesoproterozoic U–Pb and Pb–Pb ages of granulites in NW Namibia: reconstructing a complete orogenic cycle. Precambrian Res. 126, 147–168. Shand, S.J., 1943. Eruptive Rocks. Their Genesis, Composition, Classification, and their Relation to Ore-Deposits with a Chapter on Meteorite. John Wiley and Sons, New York. Taylor, S.R., McLennan, S.M., 1985. The Continental Crust: Its Composition and Evolution. Blackwell, Oxford. Tegtmeyer, A., Kröner, A., 1985. U–Pb zircon ages for granitoid gneisses in northern Namibia and their significance for Proterozoic crustal evolution of southwestern Africa. Precambrian Res. 28, 311–326. Torquato, J.R., 1974. Geologia do Sudoeste de Mocamedes e Suas Relacoes com a Evolucao Tectonica de Angola. Unpubl. doctoral dissertation, Universidade Sao Paulo, Brazil, 243 pp. Whalen, J.B., Currie, K.L., Chappell, B.W., 1987. A-type granites: geochemical characteristics, discrimination and petrogenesis. Contrib. Mineral. Petrol. 95, 407–419. Williams, I.S., 1998. U–Th–Pb geochronology by ion microprobe. In: McKibben, M.A., Shanks III, W.C., Ridley, W.I. (Eds.), Applications of Microanalytical Techniques to Understanding Mineralizing Processes. Rev. Econ. Geol., 7, 1–35. Williams, I.S., Compston, W., Black, L.P., Ireland, T.R., Foster, J.J., 1984. Unsupported radiogenic Pb in zircon: a cause for anomalously high Pb–Pb, U–Pb and Th–Pb ages. Contrib. Mineral. Petrol. 88, 322–327.