Journal of Petrology Advance Access originally published online on March 7, 2006
Journal of Petrology 2006 47(6):1095-1118; doi:10.1093/petrology/egl004
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
Correlations between U, Th Content and Metamorphic Grade in the Western Namaqualand Belt, South Africa, with Implications for Radioactive Heating of the Crust
1 SCHOOL OF GEOSCIENCES, UNIVERSITY OF THE WITWATERSRAND, PRIVATE BAG 3, WITS 2050, JOHANNESBURG, SOUTH AFRICA
2 ITHEMBA LABS (GAUTENG), P. BAG 3, WITS 2050, JOHANNESBURG, SOUTH AFRICA
3 COUNCIL FOR GEOSCIENCES, P.O. BOX 112, PRETORIA 0001, SOUTH AFRICA
RECEIVED JULY 1, 2004; ACCEPTED JANUARY 13, 2006
 |
ABSTRACT |
|---|
 TOP ABSTRACT INTRODUCTIONGEOLOGICAL SETTING OF THE...SOURCES OF NATURAL RADIATIONDISCUSSIONCONCLUSIONSREFERENCES |
|---|
The digital image of airborne radiometric data across South Africa reveals that the largest anomaly,
100 nGy/h, is caused by the granulite-facies rocks of the Namaquan metamorphic complex, whereas most of the country is <60 nGy/h. This observation is consistent with geochemical data that show that the 
1900 ± 100 Ma greenschist-facies Richtersveld Terrane near Namibia (max. U = 3·4 ppm; Th = 20·1 ppm) and the adjacent, 1100 ± 100 Ma, amphibolite-facies Aggeneys/Steinkopf Terranes (max. U 
10 ppm; Th 52 ppm) are the least enriched in U, Th and K. In contrast, the lower-T granulite-facies Okiep Terrane near Springbok hosts more enriched granites (max. U 17 ppm; Th 66 ppm) and noritic intrusions (max. U = 14 ppm; Th = 83 ppm). The most enriched rocks are found in the 1030 Ma higher-T granulite-facies core of the Namaquan belt and include quartzo-feldspathic gneisses (max. U = 46 ppm; Th = 90 ppm) and charnockites (max. U = 52 ppm; Th = 400 ppm). Our findings contradict the notion that granulite-facies terrains are characteristically depleted in U and Th. In this study we modeled the heat production in the core of the Namaquan complex, where the granulites have had a very unusual metamorphic history, and show that ultra-high-T (1000°C, P 10 kbar) metamorphic conditions could have been achieved by radiogenic heating without invoking external heat sources. However, monazite-rich veins of charnockite and patches of granulites mark the passage of CO2-dominated melts and fluids derived from fractionated noritic intrusions. KEY WORDS: charnockite; granulite; Namaqualand; thorium; uranium; radioactive heating; metamorphism
|
INTRODUCTION |
|---|
|
TOPABSTRACTINTRODUCTIONGEOLOGICAL SETTING OF THE...SOURCES OF NATURAL RADIATIONDISCUSSIONCONCLUSIONSREFERENCES |
|---|
Geochemical and field investigations at Vaalputs, Namaqualand (Fig. 1), revealed a granitic to charnockitic basement unusually enriched in U and Th (Hart & Andreoli, 1985
; Andreoli et al., 1986; Andreoli & Hart, 1990). These works suggested that many of the U- and Th-rich rocks of Namaqualand could not be explained as the fractionated products of typical granites (Robb, 1986; Raith, 1995). Instead, these anomalous lithologies seemed to require more complex geochemical processes, including the increase of uranium, thorium and the Rare Earth Elements (REEs) during high-grade metamorphism (Andreoli et al., 1992, 1994; Read et al., 2002). These observations were confirmed by a gamma-ray exposure (total count) map of South Africa (Fig. 1), where Namaqualand defines a major radiometric anomaly with exposure rates 100 nGy/h, double the typical terrestrial values (
60 nGy/h; UNSCEAR, 2000, p. 90). This study provides an account of the petrology, geochemistry and metamorphism of the rocks causing these anomalies, focusing on a very prominent anomaly near Vaalputs (Fig. 2). We also demonstrate that the most radioactive rocks include granitic orthogneisses, granulites and charnockites that experienced high- to ultra-high-T metamorphism at 1030 Ma (Mouri et al., 2003; Raith et al., 2003). Heat produced during radioactive decay in these rocks was largely responsible for their metamorphic conditions, especially where granulite and charnockite formation was promoted by the circulation of U–Th–REE-rich H2O-deficient fluids and melts. We also report on a previously unrecognized suture represented by the Buffels River shear zone, separating older crust of the Okiep Terrane near Springbok from the younger higher-grade Vaalputs Terrane in the south.
|
|
|
GEOLOGICAL SETTING OF THE NAMAQUALAND  ANOMALY
|
|---|
|
TOPABSTRACTINTRODUCTIONGEOLOGICAL SETTING OF THE...SOURCES OF NATURAL RADIATIONDISCUSSIONCONCLUSIONSREFERENCES |
|---|
Current literature describes the western sector of the Namaquan metamorphic complex as the amalgamation of four fault/suture-bound crustal blocks yielding ages in the
2000 to 1000 Ma range (Fig. 2; Jacobs et al., 1993; Thomas et al., 1994; Praekelt et al., 1997; Colliston & Schoch, 2003). These blocks are known as the Richtersveld, Steinkopf, Aggeneys and Okiep Terranes, with the latter extending south to Nuwerus and Steenkampskraal. The precise number and definition of these terranes is still debated, in part because the Steinkopf, Aggeneys and Okiep Terranes have also been considered as members of a much larger Mesoproterozoic Bushmanland Terrane (Frimmel, 2004). From N to S, the Richtersveld Terrane represents juvenile 1900 ± 100 Ma crust comprising calc-alkaline, mainly andesitic lavas and younger intrusions. These rocks are suggestive of a mature island arc/back-arc environment above a northerly dipping subduction zone (Reid & Barton, 1983; Reid et al., 1987a; Moore, 1989). In terms of metamorphism, the Richtersveld Terrane is in greenschist- to lower amphibolite-facies (Zone A, Fig. 2) and has largely escaped the 1200 to 800 Ma Namaquan Orogeny (Thomas et al., 1994; Clifford et al., 2004). To the south, the Steinkopf and Aggeneys Terranes are both largely in upper amphibolite-facies (Zone B, Fig. 2) and have significant granitic components coeval to the Richtersveld Terrane (1900 ± 100 Ma; Reid et al., 1997a). The same terranes also comprise orthogneisses and supracrustal sequences with ages of 1700 to 1300 Ma (Reid et al., 1987b, 1997a, 1997b). In contrast, there are no rocks older than 1250 Ma in the Okiep (also spelled O'okiep) Terrane where most orthogneisses, granites and mafic intrusions yield emplacement ages of 1195 ± 15 Ma or 1040 ± 20 Ma (Thomas et al., 1994; Clifford et al., 1995, 2004; Robb et al., 1999; Raith et al., 2003). With the exception of some metasediments, quartzo-feldspathic granulites and rocks of intermediate to mafic composition, 80% of the crust exposed in the Okiep Terrane consists of granite gneisses and post-tectonic granites that are locally pyroxene-bearing (Jackson, 1979; Reid & Barton, 1983; Joubert, 1986; Reid et al., 1987b; Moore, 1989; Moore et al., 1990; Keyser, 1997). Among the latter intrusions, the most important lithologies are represented by the commonly plug-like bodies of the anorthositic to noritic Koperberg Suite, whose emplacement coincided with the so-called Klondikean Episode (1030 ± 10 Ma) of the Namaquan Orogeny, characterized by ductile D3 folding and low-T granulite-facies metamorphism (Clifford et al., 2004). The metamorphic patterns of the Okiep Terrane are complex but essentially represented by four E–W trending belts with either tectonic or gradational contacts and designated, from north to south, as Domains C, D, C' and B' in Fig. 2. Domains C and C' are in the lower-T granulite-facies, whereas the core Domain D is in the upper-T granulite-facies. Finally, Domain B' is situated at the southern edge of the metamorphic complex and is represented by gneisses and metapelites in upper amphibolite-facies (Waters & Whales, 1984; Waters, 1989).
It has been proposed that during the Neoproterozoic, between 1000 and 800 Ma, the metamorphic complex experienced isobaric cooling and by 570 Ma was eroded close to the current level of exposure, before burial beneath the sediments of the Vanrhynsdorp Group (southernmost subsidiary basin of the Nama basin; Waters, 1989; Gresse, 1995; Robb et al., 1999). Closer to the Atlantic Ocean, the Namaquan gneisses and granulites and their Neoproterozoic cover were subsequently overprinted by the Pan-African orogenic cycle, between 545 ± 2 Ma and 507 ± 6 Ma (the staurolite zone west of the line marked E, Fig. 2; Moore & Reid, 1989; Frimmel, 2000).
|
SOURCES OF NATURAL RADIATION |
|---|
|
TOPABSTRACTINTRODUCTIONGEOLOGICAL SETTING OF THE...SOURCES OF NATURAL RADIATIONDISCUSSIONCONCLUSIONSREFERENCES |
|---|
Radiometric surveys and scintillometer-assisted field traverses were conducted over a number of radiometric highs situated near Springbok, Vaalputs and Steenkampskraal (Fig. 2). These areas are underlain by granulite-facies ortho- and paragneisses, often with above-average contents of U and Th (Albat, 1984
; Robb, 1986; Andreoli et al., 1986, 1994; Moore, 1989; Thomas et al., 1996; Raith et al., 2003). For the purpose of this paper the term ‘above-average’ will be applied to lithologies whose U and Th contents exceed the average values in the upper continental crust (U = 2·5 ppm; Th = 10·3 ppm; Wedepohl, 1995). Petrographic characteristics, field relations and stratigraphic position were used to subdivide the anomalous rocks into four major groupings: (1) granite gneisses and granites, (2) charnockites, (3) the norite-anorthosite kindred and (4) granulites of supracrustal and undetermined origin.
Granite gneisses and granites
Foliated granites, granitic orthogneisses and late- to post-tectonic granites are extensively represented in all the metamorphic zones of Fig. 2, including the high-T granulite-facies Domain D. With the exception of the 1033 Ma Rietberg Granite, most of these rocks have emplacement ages ranging from 1200 Ma to 1060 Ma, therefore predating metamorphism at 1030 Ma (Thomas et al., 1996; Ashwal et al., 1997; Robb et al., 1999; Clifford et al., 2004). In view of this, most of the less deformed Namaquan age granitoids found in the Terranes of Fig. 2 represent relics whose fabrics have escaped obliteration during granulite-facies metamorphism (Andreoli et al., 1986; Raith & Harley, 1998; Raith et al., 2003; Frimmel, 2004). As presented in Table 1, the Okiep Terrane hosts granulite-facies granitic plutons with the highest concentrations of heat-producing elements (K, U and Th) when compared with other acid intrusions in the lower-grade Steinkopf, Aggeneys and Richtersveld Terranes (Figs. 2 and 3a). Foremost among these intrusions of the Okiep Terrane are the Rietberg and Kweekfontein granites, and the variably foliated Concordia, Bloukop and Jakkalshoek granite gneisses (Columns 1–3, 7 and 8 respectively, Table 1). In the northern Okiep Terrane, close to the margins of the Concordia granite (named after the village at Loc. 4, Fig. 2), there are also numerous late-stage smaller intrusions of U-rich leucogranite that may constitute potential resources of uranium (Nooitgedacht, Fig. 2; Column 2, Table 1; Jacob et al., 1986; Robb, 1986). Further intrusions of highly radioactive granite were also reported near Garies (profile X–X', Figs. 2 and 4; Jack, 1980). Detailed accounts of the petrology and mineralogy of these rocks with a view to their complex U phases are available for several areas, especially those near Springbok, Steenkampskraal and Bitterfontein (Robb & Schoch, 1985; Robb, 1986; Robb et al., 1986; Andreoli et al., 1994; Raith, 1995; Thomas et al., 1996). Less enriched, but still with concentrations above upper crustal averages are the regionally extensive, Nababeep augen gneiss (Column 4, Table 1; Robb et al., 1999) and the foliated granites, frequently megacrystic, found at Vaalputs and Steenkampskraal (Columns 5 and 6, Table 1, Fig. 3a; Andreoli et al., 1986; Ashwal et al., 1997; Knoper et al., 2001). The granitic rocks of the Okiep Terrane, with the exception of the Bloukop granite (Column 7, Table 1), are enriched in Th relative to U compared with average upper continental crust (e.g. Th/U 
4·2, Column 18, Table 1 and Fig. 3a). Our own observations indicate that the elevated contents of U and Th in the granitic rocks described above are hosted mainly by zircon and less commonly by monazite. In a detailed description of the accessory phases of the Concordia granite (Column 3, Table 1), Raith (1995) mentions primary (Ce-rich) monazite, zircon (including U- and Th-rich varieties) and secondary (Ce-) allanite. The zircons of this high-U granite are largely metamict and euhedral with high-U and low-U cores mantled by high-U rims (Clifford et al., 2004). Similarly, the Kweekfontein granite (Column 2, Table 1) includes zircons with homogeneous, euhedral, high-U and low-U cores mantled by high-U rims (Clifford et al., 2004). In the same rock type, quantitative fission track micro-mapping showed that U is hosted by both primary and secondary phases (Robb & Schoch, 1985). Primary phases include biotite, magnetite, zircon, allanite and monazite-like (Ca, LREE-) silico-phosphates such as cheralite and britholite. The secondary minerals include chlorite, epidote and sphene-like products.
|
|
|
|
As mentioned previously, the granitic rocks in the amphibolite-facies Aggeneys and Steinkopf Terranes tend to have lower U and Th contents. In particular, plutonic rocks with anomalous U and Th in the Aggeneys Terrane are represented by the
1185 ± 15 Ma Achab orthogneiss and, in lesser measure, by the 1180 ± 20 Ma Aroams and related gneisses (Columns 9–12, Table 1 and Fig. 3a; Reid et al., 1997a; Bailie & Reid, 2004). Our data show that the Achab gneiss, and to a lesser extent the Aroams gneiss, are enriched in Th (range: 26–52 ppm) and have higher Th/U ratios (range: 5–7·5) when compared with the average upper crust (Columns 9–12 and A, Table 1 and Fig. 3a).
The Steinkopf Terrane shows weaker radiometric anomalies, consistent with the low-U content of this crustal segment (Columns 13–15, Table 1 and Figs. 3a and 4). These granitic rocks are also U-depleted relative to K, as suggested by the K/U ratios (>1·7E+4), which are higher than world averages (1·15E+4; Column 18, Table 1). In contrast, the Th content of the same granites (17–47 ppm) exceeds average upper crust values (10·3 ppm) and is closer to that of the Aggeneys Terrane. In particular, the Th-rich Wyepoort intrusion (Column 14, Table 1 and Fig. 3a) may be correlated with the Concordia granite in the Okiep Terrane on the basis of geochemical affinities (Reid & Barton, 1983). Similarly, less-enriched gneisses with 23 ppm Th may be compared to the regionally extensive Nababeep augen gneisses of the Okiep Terrane (Columns 13 and 4, Table 1 and Fig. 3a; Barton, 1983; Reid & Barton, 1983; Joubert, 1986; Robb et al., 1999). Data for the 1800 Ma Gladkop orthogneisses (Column 15, Table 1; Robb et al., 1999) show that apart from U, for which there are no reliable data, these rocks are mildly enriched in Th (17 ppm) relative to upper crust values (Th = 10·3 ppm; Column 18, Table 1 and Fig. 3a). Finally, the 1700 Ma tonalite and granite in the Richtersveld Terrane (Column 16, Table 1 and Fig. 2) are also enriched in K2O (4·5%), U (3·4 ppm) and Th (20·1 ppm) and have a higher Th/U ratio (5·9) compared with the average upper continental crust (Column 18, Table 1 and Fig. 3a). In summary, the data presented in Table 1 and in Fig. 3a are consistent with the airborne spectrometric and geochemical profiles (X–X', Figs. 2 and 4; Muller & Smit, 1983) as all show a regional increase in radioactive elements from a low in Zone A to a maximum in Zone D.
Charnockites
Orthopyroxene-bearing rocks of igneous origin and granitic composition, by definition called charnockites (Streckeisen, 1976), have widespread distribution in the Okiep Terrane, especially in the Garies–Kliprand area south of the Buffels River shear zone (Fig. 2; Joubert, 1971; Jackson, 1979; Albat, 1983; Thomas et al., 1996). The following three groups of charnockitic rocks were distinguished as described below.
Charnockite veins, dykes and megabreccia
The charnockite veins are a few centimeters wide and a few meters in length. They are slightly wavy, discontinuous and tend to occur in swarms. At their type locality (no. 13; Fig. 2), the veins develop along incipient and steeply dipping zones of warped older D2 foliation (Raith & Harley, 1998) in the Nababeep augen gneiss host. As such, these warped zones are compatible with the 1030 Ma D3 deformation in the area (Raith & Harley, 1998; Robb et al., 1999). The overall appearance of these orthopyroxene-bearing veins (dark green with oily luster) is reminiscent of the charnockite veins described at Kabbaldurga, southern India (Janardhan et al., 1979; Friend, 1985; Newton, 1992). The Th- and U-enriched nature of the individual veins may be detected by using a spectrometer (e.g. Koperberg West: Loc. 5, Fig. 2; Saunders et al., 1995), and narrowly spaced clusters/swarms of veins may produce radiometric anomalies recognizable even in airborne surveys (e.g. Steenkampskraal; Fig. 2; M. Knoper, personal communication). Under the microscope, the charnockite veins seem to display igneous-looking textures as many samples contain subhedral plagioclase, perthitic K-feldspar and chloritized orthopyroxene. Also conspicuous are nodular myrmeckitic intergrowths of quartz + plagioclase. Biotite and hornblende in most cases appear to be texturally secondary as they replace orthopyroxene. Accessories comprise zoned, euhedral zircon, apatite, opaque grains and occasional crystals (0·7 mm across) of a yellow and metamict mineral phase surrounded by expansion cracks. Some allanite grains were also observed as inclusions in the center of subhedral plagioclase. When compared with their gneissic/granitic host rocks (Table 2) the charnockite veins tend to be markedly enriched in Th and the LREE, and more moderately in U, Zr, Nb, Ta, Y, Rb, K, and Cr. U, Th and the LREE show an increase even in texturally unmodified host gneiss, up to 18 cm from the veins (Columns 1–5, Table 2 and Fig. 3b), and inter-element ratios tend to display erratic variations.
The charnockite dykes differ from the veins in terms of lateral continuity, thickness and the large (up to 8 cm long) size of the euhedral orthopyroxene. The radioactivity appears to be heterogeneously distributed, and at Kliphoogte (Loc. 9, Fig. 2) it varies from three times background, where the dyke is 2 m thick, to background values where its thickness has decreased to 5 cm, over a distance of 200 m. The charnockite dykes have pegmatoid characteristics and are frequently developed in country rocks close to their contacts with the Koperberg Suite (Loc. 3, 5 and 9, also at Kliprand and Steenkampskraal, Fig. 2; Boshoff, 1951; Mostert, 1964; Andreoli et al., 1994; Saunders et al., 1995; Read et al., 2002). The matrix of the dykes contains common accessory monazite, responsible for the measured radioactivity, while the groundmass in places consists of graphic intergrowths of quartz and alkali feldspar ± plagioclase. The analytical data reveal that Th, the LREE, and in lesser measure 
[CaO + MgO + FeO + Fe2O3] and Sr may be highly enriched in the dykes when compared with their host Concordia granite. However, the latter has higher K, Na, U, Rb and Zr content than the charnockite dyke (Columns 8–10, Table 2; Fig. 3b).
The megabreccia charnockite is similar to the previous rock type and constitutes, in intimate association with leucogranite, the matrix between blocks of structurally disrupted country rocks in small stocks that are minor though characteristic constituents of the Springbok area (Kisters, 1993; Gibson et al., 1996). These bodies, also found near Steinkopf and Vaalputs (Table 4), tend to occur close to D3 cusp-shaped anticlines known as steep structures, which are another distinctive feature of the Springbok area (Gibson et al., 1996; Watkeys, 1996). Typically, the charnockite and the leucogranite grade into each other (Kisters, 1993). In rare cases, the megabreccia matrix is represented by other rock types, such as glimmerite, leucotonalite and enderbite, a dark bluish green rock with a vitreous luster that consists of quartz, plagioclase and minor orthopyroxene (Table 4; Streckeisen, 1976). The distribution of radioactive minerals in the megabreccia charnockite is poorly known.
|
Foliated charnockites
Foliated greenish-brown, orthopyroxene-bearing granulites of granitic composition are common in the domain of highest metamorphic grade and especially in the area between Paulshoek and Vaalputs, recognizable in Fig. 2 by the most prominent radiometric anomaly. A site easily accessible for the study of these rocks is located at the southwest boundary of the Vaalputs property (5 km SW of Loc. 6, Fig. 2), close to the edge of a >150 km2 sheet of the 1056 ± 10 Ma megacrystic Stofkloof granite (Andreoli et al., 1986
; Andreoli, 1996; Ashwal et al., 1997). Mapping at this locality has identified a crescent-shaped exposure of Stofkloof granite gneiss, at least 6·5 km2 in area, that is more intensely sheared and whose colour grades inward from pink (lower gamma-ray counts; Column 5, Table 1) to greenish brown (higher gamma-ray counts; Column 3, Table 3) over a distance of 500 m (Andreoli et al., 1986). The highest radiometric reading (max. K, 4·5 wt %; eU, 13 ppm; eTh, 160 ppm) was measured at the core of this anomaly, next to the historical Stofkloof farmhouse, on a 20 m2 dyke-like exposure of massive charnockite with conspicuous megacrystic orthopyroxene. A detailed, three-dimensional view of the granite gneiss–charnockitic orthogneiss relationships was also gained from stratigraphic boreholes drilled to depths of up to 1000 m on the eastern side of the Vaalputs property. In particular, alternating bands 50 cm to >1 m wide of light pinkish (less foliated) and dark green (more foliated, charnockitic) Stofkloof granite gneiss are common in boreholes HLD-1 and -3 from Vaalputs (Loc. 7 and 14, Fig. 2; Columns 5 and 6, Table 1, and Columns 1 and 3, Table 3; Fig. 3c; Andreoli, 1996). The elevated concentrations of U and Th in these charnockitic rocks are typically hosted by monazite, zircon and, in places, by a yellow and highly metamict mineral resembling allanite (sample NAM 524: Column 6, Table 3).
In general, the granite to charnockite transition in the area of the Vaalputs–Paulshoek anomaly is geochemically complex because in some places U, Th and the LREE remain unchanged, whereas in others they increase sharply from granite to charnockite (cf. Columns 5 and 6, Table 1 vs Columns 1–4, Table 3; Fig. 3c). The most spectacular increase was observed 10 km N of Paulshoek, where a granulite-facies metamorphic overprint on augen gneisses is accompanied by a 10-fold increase in U, Th (Figs 3c and 5; cf. Columns 5 vs 6, Table 3). A significant yet less striking increase in the LREE contrasts a decrease in Rb and in the K/U, Th/U and Rb/Sr ratios. The Th/U ratio of the foliated charnockites is variable, and in only a few samples does it approach the value of the average lower continental crust (Column A, Table 3). A large group of enriched samples, however, has a Th/U ratio that exceeds by a factor of 3, the value for average lower crust (Columns 2 and A, Table 3). The unusual geochemical characteristics of the high-Th charnockitic granulites are further emphasized by their inter-element ratios. In particular, their very low K/Rb and K/U, coupled with high Rb/Sr ratios (209, <0·7 E+4, and 3·3–4·96, respectively), indicate that these rocks are enriched in U (Th) and Rb relative to K when compared with average continental crust (Columns A–D, Table 3 and Fig. 3c). This type of strong enrichment of incompatible trace elements relative to incompatible major elements is reminiscent of processes of extreme magmatic differentiation.
Charnockite plutons
Charnockitic plutons of irregular shape and size (up to 400 km2) are important constituents of Zone D, and display a variable radiometric response (Fig. 2). The best known of these plutons, dated at 1063 ± 18 Ma, is exposed in a small quarry 2 km southeast of Kliprand, where it appears undeformed, almost black in colour and megacrystic (Fig. 2; Albat, 1984; Frimmel, 2004; and Frimmel, personal communication, 2004). A sample from this locality has the lowest U and Th values among all the analyzed granites and charnockites (Column 7, Table 3), even though Fig. 2 reveals several small radiometric anomalies within the same pluton. The U and Th data are close to the averages for the lower continental crust (Columns 7 and A, Table 3). Instead, the K/Rb and Rb/Sr ratios demonstrate that Rb is enriched relative to K and Sr when compared with the average of lower and upper crust rocks, as observed in most other granite gneisses and charnockites listed in Tables 1–3.
Norite–Anorthosite kindred
A distinctive feature of the Springbok area is a swarm of 1700 irregular and discontinuous dykes, plugs and sheets of generally mafic to ultramafic igneous rocks attributed to the 1030 Ma (syn-D3) Koperberg Suite. These rocks, locally mined for copper, crystallized under low-T granulite-facies conditions (Table 4; Clifford et al., 1995, 2004; Gibson et al., 1996; Robb et al., 1999). Other swarms of Koperberg Suite bodies intrude the Vaalputs, Kliprand and Steenkampskraal areas (Fig. 2 and Table 4). Anorthosite, leuconorite, diorite, norite and hypersthenite represent the more common lithologies of this suite, which displays some affinities to massif-type anorthosite (Conradie & Schoch, 1988; Clifford et al., 2005). The emplacement style of the Koperberg Suite is peculiar, as it exploited the dilatant sites caused by strain incompatibilities during the formation of D3 steep structures, during peak granulite facies conditions (Table 4; Gibson et al., 1996; Watkeys, 1996). Despite the fact that their bulk chemistry is typically mafic (biotite norite) to intermediate (diorite), the rocks of the Koperberg Suite may host significant amounts of U and Th (Table 5 and Fig. 3d). This is consistent with the observation that the 238U/204Pb (µ2) values for typical mafic to intermediate rocks of the Koperberg Suite are in the 9·98–10·20 range, higher than the corresponding values (8·7–9·8) in anorthosite and related rocks from Norway and Canada (Clifford et al., 1995). The fact that these U/Pb values are only slightly high, whereas the actual U concentrations are more than an order of magnitude above normal, indicates that the rocks are enriched in Pb as well as in U. REE, Th and U enrichments are also associated with the more differentiated/contaminated (?) members of the Koperberg suite, such as glimmerite and leucocratic tonalite (Table 4; Columns 8 and 9, Table 5; Andreoli et al., 1994; Read et al., 2002). Hyper-enrichment in U, Th, REE, Zr and F distinguishes the Koperberg Suite in the Steenkampskraal mine area, where 5 x 104 metric tons of monazite (chalcopyrite) concentrate with 0·7% F were extracted from a 400 m x 10 m dyke comprising leuconorite, enderbite, charnockite and leucocratic tonalite (Fig. 2 and Table 4; Andreoli et al., 1994; Read et al., 2002).
|
Detailed studies have indicated that in the Koperberg Suite the radioactivity is largely carried by monazite, zircon (max. U, 9729 ppm; max. Th, 1821 ppm), allanite, titanite and apatite (Andreoli et al., 1994
; Clifford et al., 1995; Robb et al., 1999; Read et al., 2002). Geochemical oddities of the Koperberg Suite include a positive correlation between Cu mineralization and radioactive element anomalies (Andreoli et al., 1994; Saunders et al., 1995), enrichment in 18O, Rb, F, Cl, negative 
Nd values (Tables 4 and 5; Conradie & Schoch, 1988; Boer et al., 1993) and, near Springbok, a notable S deficiency (Boer et al., 1994).
Supracrustal granulites and rocks of undetermined origin
Dismembered relics of highly metamorphosed sedimentary and volcanic rocks are widespread within the metamorphic complex. Although these rock suites were not investigated in the same detail as the preceding igneous rocks and granitoids, some Th and/or U values are listed for completeness in Table 6 (Fig. 3e). These data demonstrate that even the supracrustal rocks may be enriched in U and/or Th relative to the average upper crust (Fig. 3e). Some of the highest values are encountered among calc-silicate granulites and iron formations (Columns 1, 2 and 7, Table 6), but these anomalous rocks are rare and unlikely to contribute to the regional patterns given in Fig. 2 (Moore, 1989; Moore & McStay, 1990). Instead, the metapelites in the Okiep and Aggeneys Terranes and the volcanic rocks (porphyries) of the Richtersveld Terrane have wider distribution but are only moderately enriched in Th (and U) relative to shale and the average upper crust (Columns 6, 10 and 11, and A and B, Table 6; Fig. 3e). The U content and the K/U ratio of a metapelite from Vaalputs are close to the values of an average shale (Columns 4 and A, Table 6; Fig. 3e).
|
The origin of the quartzo-feldspathic pink gneisses, which are very common in the western metamorphic complex, is highly controversial, and this lithological "sack name" probably includes components of supracrustal (e.g. acid volcanics and/or arkosic sediments), intrusive and anatectic origins (Reid et al., 1997a
; Macey et al., 2006). Consequently, the status of the pink U-rich (up to 46 ppm) quartzo-feldspathic Lekkerdrink gneiss in the area of Fig. 5 is also uncertain (Column 8, Table 6; Figs. 2, 3e and 5; Macey et al., 2006). Conversely, the pink gneisses of the lower grade Aggeneys Terrane (zone B, Fig. 2) are also less enriched in U and Th than those from the area shown in Fig. 5 (Columns 8 and 9, Table 6, and Fig. 3e). Even so, the pink gneisses of zone B are still highly enriched in K, U and Th (and in Th, U relative to K) when compared with the average upper continental crust (Columns 9 and B, Table 6). The main geochemical characteristic of these rocks, which helps to distinguish them from the majority of the granites listed in Table 1, is their low Th/U ratio (3). |
DISCUSSION |
|---|
|
TOPABSTRACTINTRODUCTIONGEOLOGICAL SETTING OF THE...SOURCES OF NATURAL RADIATIONDISCUSSIONCONCLUSIONSREFERENCES |
|---|
The K, U and Th values for rocks of the metamorphic complex are sufficient to explain the elevated airborne radiometric anomalies displayed in Fig. 2, especially those in the highest metamorphic grade zone D. We also note that U and Th are largely hosted by primary igneous or high-grade metamorphic minerals, mainly zircon and monazite. This observation implies that their presence is not related to secondary, i.e. hydrothermal/low grade metamorphic processes, but rather to the emplacement history of their igneous host rocks or to metamorphic processes in the granulite-facies (Andreoli et al., 1994
). These observations are clearly inconsistent with the commonly held views that the granulites are generally depleted in heat-producing elements, especially uranium (Taylor & McLennan, 1985; Rollinson, 1993). Instead, Namaqualand displays the almost paradoxical phenomenon in which increases in metamorphism and the distribution of charnockites rich in U and Th are connected.
Terranes, crustal growth and radioactivity
Studies in the western metamorphic complex have shown that the lower grade terranes consist wholly (Richtersveld Terrane), or partly (Steinkopf Terranes), of a 1700–2000 Ma juvenile volcanic and plutonic crust that probably accumulated in a mature island arc (Reid, 1979; Reid & Barton, 1983; Moore, 1989; Thomas et al., 1994). Some of the oldest rocks could also represent amalgamated fragments of a juvenile, 2000 Ma continent (Moore, 1989). The elevated High Field Strength Element (HFSE) and Large Ion Lithophile Element (LILE) contents of these older rocks were tentatively considered by Reid et al. (1987a, 1987b) to reflect magmas derived from source regions (e.g. the mantle wedge) modified by subduction zone metasomatism. However, the Steinkopf and Aggeneys Terranes differ from the largely pristine 2000–1700 Ma Richtersveld Terrane in that they were extensively intruded by granite and migmatized at 1100 ± 100 Ma. In addition, they host several stratigraphic sequences comprising metapelites, quartzite, calc-silicates and metavolcanics, which near Aggeneys yield ages of 1700 to 1300 Ma (Reid et al., 1997a, 1997b). Reid et al. (1987a, 1987b, 1997a, 1997b) and Raith & Meisel (2001) considered the modest enrichment in U and Th of these mafic metavolcanics (amphibolites) to be primary, and compared the latter to continental (tholeiitic) basalts or arc magmas with a subduction-related fertile source component.
The Okiep Terrane north of the Buffels River shear zone (Figs 2 and 4) consists predominantly of syntectonic to post-tectonic granites intruded in a 30 Myr interval, between 1200 and 1170 Ma (Clifford et al., 1995, 2004; Robb et al., 1999). Of these, the 1192 ± 9 Ma syntectonic Nababeep augen gneiss is the least enriched in U and Th, whereas the late-syntectonic and massive to poorly foliated 1206 ± 16 Ma Concordia granite that intrudes the Nababeep gneiss is U-enriched and fractionated in the LILE (Raith, 1995; Clifford et al., 2004). These intrusions were derived from the partial melting of two different crustal protoliths with similar, 1700 to 2000 Ma, Sm–Nd model ages. The protolith of the Nababeep gneiss is chemically undetermined, whereas that of the Concordia granite was either peraluminous granite, possibly contaminated by HFSE-enriched magma from a deep source (Raith, 1995; Clifford et al., 1995, 2004), or juvenile crust of high-K (-HFSE?) andesitic composition (Duchesne et al., 1999). The undeformed and high-U 1186 ± 15 Ma Kweekfontein granite may also represent an anatectic melt of a heterogeneous and enriched lower crustal source (Robb, 1986; Clifford et al., 2004). Similarly, the moderately enriched 1035 ± 7 Ma Rietberg granite and syenite in the northwestern part of the Okiep Terrane (Figs 2 and 4; Clifford et al., 2004) display a low (87Sr/86Sr)i 0·705 that is consistent with a deep-seated source (Barton, 1983).
The origin of the large swarm of norite, anorthosite and related rocks of the Koperberg Suite remains controversial, but most authors view their enrichment in incompatible elements as the result of contamination of tholeiitic melts by 1700–2100 Ma crust (Boer et al., 1993; Brandriss & Cawthorn, 1996; van Zwieten et al., 1996). McIver et al. (1983) hypothesized instead the contamination of REE-enriched alkaline mantle-derived magma by peraluminous, crustal anatectic melts; others derived the Koperberg Suite from the direct melting of a dry, dioritic lower crust (Clifford et al., 1995, 2004; Duchesne et al., 1999). Given that the above mechanisms are not normally observed to produce extreme enrichments in REE, U and Th, Conradie & Schoch (1988) and Andreoli et al. (1994) proposed that the Koperberg Suite was derived by melting a fertile mantle source. The prevailing view is that the high levels of U and Th (and Rb, REE, etc.) in the Okiep Terrane north of the Buffels River shear zone derive from the reworking of highly enriched repositories of sedimentary or igneous origin in the lower crust. Sm–Nd model ages for this crustal component in the Koperberg Suite falls in the 1600–2000 Ma (TCHUR) or 1900–2200 Ma (TDM) age ranges (Table 4; Clifford et al., 1995, 2004). This time span is not too different from that of the 1700–2000 Ma crustal remnants in the Aggeneys and Richtersveld Terranes (see above), and may define an enlarged northern province formed in a mature island arc, in which the mantle source was enriched by subduction zone metasomatism (cf. Heaman et al., 2002).
South of the Buffels River Shear Zone, we note a significant change in the amplitude and width of the anomalies, a feature that may imply a major crustal discontinuity (Figs 2 and 4). Furthermore, detrital zircons from metasediments south of the shear zone yield crystallization ages that in nearly all cases do not exceed 1300 Ma (Raith et al., 2003). This age stands in sharp contrast with that of the metapelites and quartzites from the Springbok area, which were deposited between 1300 and 1650 Ma, and whose detrital zircons yield 1900 Ma ages (Robb et al., 1999). More specifically, detrital zircon grains in a supracrustal granulite from Bitterfontein were deposited in a basin with a maximum age of 1157 Ma (Raith et al., 2003). This sedimentary episode, previously unrecognized, received detritus from a 1250 ± 50 Ma source region (Raith et al., 2003) and was of regional extent because igneous zircons in metavolcanic granulites from Vaalputs yield an age of 1137 ± 17 Ma (Ashwal et al., 1997). Quartzite and calcic granofelses are also common in the Okiep and Aggeneys Terranes, but rare south of the shear zone, where metavolcanic rocks appear to be prevalent (Reid et al., 1987b; Moore, 1989; Robb et al., 1999; Macey et al., 2006). In view of the above, we propose to use this high-angle (60° to vertical) and up to 2 km wide shear zone as the southern limit of the Okiep Terrane (Fig. 4; Joubert, 1971, 1986; Blignault et al., 1983). The crustal block south of the shear zone is referred instead as the Vaalputs Terrane (Fig. 2). In view of these features, the Buffels River shear zone may not be comparable to major, typically shallow, extensional detachment faults (Parrish, 1995 and references therein).
|
The Vaalputs Terrane
The timing of collision of the Okiep and Vaalputs Terranes is constrained by the age of the characteristic intrusions within the two regions. The coupling occurred after the regionally widespread emplacement of granites and orthogneisses in the Vaalputs Terrane between 1109 ± 17 Ma and 1065 ± 2 Ma (Ashwal et al., 1997
; Raith et al., 2003; Frimmel, 2004), but before the common age of metamorphism and of emplacement of the Koperberg Suite at 1030 ± 10 Ma (Table 4; Clifford et al., 1995, 2004; Ashwal et al., 1997; Robb et al., 1999). The existing data suggest that the Vaalputs Terrane is extensive because its distinctive megacrystic granites, here referred to as the Southern Megacrystic Suite (Fig. 2), correlate with similar rocks in southern Natal (Thomas et al., 1996; Ashwal et al., 1997; Knoper et al., 2001; Grantham et al., 2001; Eglington et al., 2003; Frimmel, 2004).
In the Vaalputs Terrane, the temporal evolution of the crust and its U and Th inventories are less clearly defined than in the Okiep (and Aggeneys) region. However, samples from Bitterfontein and Vaalputs point to a crustal age significantly less than that near Springbok and Aggeneys, as only one detrital zircon core was found yielding a Pb–Pb age of 1700 Ma, all others having crystallized between 1300 and 1170 Ma (Ashwal et al., 1997; Raith et al., 2003). Similarly, the Sm–Nd model ages of granites and Koperberg Suite rocks from Garies, Vaalputs and Steenkampskraal (TCHUR: 1040–1200 Ma, TDM: 1570–1710 Ma) are distinctly younger than comparable rocks near Springbok (Table 4; Clifford et al., 1995, 2004; Ashwal et al., 1997; Yuhara et al., 2001). If the model proposed for the enriched crust of the Okiep Terrane is correct (i.e. subduction zone metasomatism 
enriched volcanics erosion enriched sediments deep burial and anatexis enriched granite plutons and contamination of Koperberg Suite), then the same complex cycle might have been repeated in the Vaalputs Terrane. Here the process would have been responsible for the enriched and, locally, hyper-enriched character of the Southern Megacrystic and Koperberg Suites (Columns 5–8, Table 1; Table 4; Jack, 1980; Andreoli et al., 1994; Reid et al., 2002).
Causes of metamorphism
One school of thought holds that the high-grade metamorphic event (T 750–850°C, P 4–6 kbar) peaked at 1030 Ma and was caused by magmatic underplating of the crust, with heat transfer mainly caused by magmatic convection (Waters, 1989; Robb et al., 1999). A different model suggests that the granulite-facies metamorphism was caused by continental crustal doubling during the Kibaran event at 1210 – 1180 Ma, which increased the thermal gradient to 35°C/km (T 800–850°C, P 6–7 kbar; Clifford et al., 1995, 2004; Raith & Harley, 1998). This episode was followed by slow isobaric cooling until 1020–1040 Ma (T 580–660°C, P 5·8 ± 0·5 kbar). The latter model requires a persistence of granulite-facies conditions for 170 Myr after the Kibaran event but this has been questioned (Gibson et al., 1996; Robb et al., 1999).
Studies by Kramers et al. (2001) and Chamberlain & Sonder (1990) have shown that high-grade conditions may be attained by radioactive heating if the crust is sufficiently enriched in K, U and Th. Calculations based on the data provided in our work, including the more conservative values of K, U and Th by Holland & Marais (1983), show that the heat production in the Namaquan crust at 1000 Ma considerably exceeded that of the Northern Marginal Zone of the Limpopo belt, where granulite facies conditions were reached at 2600 Ma due to radioactive heating (Table 7; Kramers et al., 2001). Additional constraints for heat production, expressed as H(
), may be obtained from published heat flow data for the metamorphic complex (Table 7). The latter suggest that the Vaalputs charnockitic granulites with H() >10 µW/m3 may not have significantly deep roots because the heat flow at Loc. 7 (Fig. 2) is 61 mW/m2, equal to the average for the metamorphic complex (Table 7). On the other hand, localities of higher than average heat flow (i.e. >70 mW/m2) were identified by Jones (1987) in the lower grade western Kakamas Terrane (81 mW/m2; Table 7) and in northern Lesotho, where they are thought to reflect the presence of more radioactive rocks of the Namaquan Belt at depth (Jones, 1992). The presence of K, Th and U anomalies at depth is also inferred by the occurrence of these elements in most of the intrusions derived from, or passing through, the lower crust between 1200 and 1020 Ma. Clearly, while these synmetamorphic (?) intrusions probably did not purge all the U and Th from their deep seated enriched sources, they effectively transported heat-producing elements upward through the crust. From the data of Table 7, and values quoted by Kramers et al. (2001), we may derive an indication of how fast the peak conditions were attained. In the Limpopo area, these authors reported that a rock unit with the density of granite (
= 2·7 g/cm3) and that is entirely insulated (a condition approachable with increasing depth in the crust) will heat up at a rate of 9·5°C/Myr, if H() is 1 µw/m3. Conditions approaching full thermal insulation and granitic bulk density may also characterize Domain D of Namaqualand (Fig. 2) where, however, the heat production is five times that of the Limpopo (i.e. H() = 5 µw/m3; Table 7). Consequently, the highest-T domain of Namaqualand could have experienced a heating rate approaching 45–50°C/Myr, capable of increasing its temperature from medium amphibolite facies conditions (600°C) to 1000°C (Mouri et al., 2003) in just 8 Myr, within the short time interval (<10 Myr at 1030 Ma) during which the Koperberg Suite was intruded throughout the western Namaquan metamorphic complex (Clifford et al., 2004).
|
Based on the available data, we propose here a model whereby the radioelement-rich
1200 Ma Okiep Terrane collided with the even more enriched Vaalputs Terrane at 1040 Ma. The resulting tectonic imbrication transported the radioactive metasediments, metavolcanics and megacrystic granite plutons of the latter to lower crustal depths, where they contributed to the anomalous heating of the newly established Namaquan belt. According to our model, the rare corundum–quartz assemblages may record a transient, early high-P, ultra-high-T episode of the Vaalputs Terrane that was followed by ductile deformation, orogenic collapse (lateral spreading) and subsequent isobaric cooling (Waters, 1988; Mouri et al., 2003).
Charnockite-forming processes
The charnockite veins of the western Namaquan metamorphic complex bear a striking similarity to the patchy and vein-like charnockites found in southern India and in Sri Lanka, which a number of authors (Janardhan et al., 1979; Friend, 1985; Harley, 1989; Newton, 1992) have related to the circulation of CO2-rich fluids and/or melts. On this basis and from our data, we propose that the Namaquan charnockite veins formed when water-deficient U-Th-REE-CO2-rich melts or fluids started to propagate along dilatation structures within gneisses close to their solidus temperatures at the time of deformation (cf. Gibson et al., 1996). In these veins, the replacement of orthopyroxene by hornblende and biotite suggests a late increase in aH2O as a result of the local dehydration of their host rocks. Geochemical data in Table 2 (Fig. 3b) also indicate that trace amounts of U and Th infiltrated the host rocks by more than 10–20 cm from the veins (Columns 1–7, Table 2 and Fig. 3b; Andreoli et al., 1994). Where more infiltrating melt was available and the fractures/shear zones linked and expanded, especially in the proximi