Sm–Nd isotopic mapping of lithospheric growth and stabilization in the eastern Kaapvaal craton


Blair Schoene, Department of Mineralogy, University of Geneva, Rue des Maraîchers 13, Geneva CH-1205, Swtizerland. Tel.: +41 22 379 31 76; fax: +41 22 379 32 10; e-mail:


Whole-rock Sm–Nd isotope systematics of 79 Archean granitoids from the eastern Kaapvaal craton, southern Africa, are used to delineate lithospheric boundaries and to constrain the timescale of crustal growth, assembly and geochemical differentiation c. 3.66–2.70 Ga. Offsets in εNd values for 3.2–3.3 Ga granitoids across the Barberton greenstone belt (BGB) are consistent with existing models for c. 3.23 Ga accretion of newly formed lithosphere north of the BGB onto pre-existing c. 3.66 Ga lithosphere south of the BGB along a doubly verging subduction margin. The Nd isotopic signature of c. 3.3–3.2 Ga magmatic rocks show that significant crustal growth occurred during subduction–accretion. After c. 3.2 Ga, however, the Nd signature of intrusive rocks c. 3.1 and 2.7 Ga is dominated by intracrustal recycling rather than by new additions from the mantle, signalling cratonic stability.


Models of the structure and composition of continental lithosphere are constrained by the integration of geological and geophysical data and rely on the relatively rare occurrence of xenolith bearing volcanic rocks. Alternatively, the geochemistry and isotopic characteristics of exposed igneous rocks can be viewed as probes of their source regions and therefore provide insight into magmatic processes and lithospheric structure. In particular, Sm–Nd isotope systematics can be used to constrain the time-scales of early crust-mantle segregation and subsequent magmatic and geochemical differentiation of the lithosphere (DePaolo and Wasserburg, 1976a; DePaolo, 1980; Nutman et al., 1993). In addition, differences in the Nd signature of plutonic rocks can be used to identify important lithospheric boundaries that may not be recorded in the ages of rocks at the surface (Bennett and DePaolo, 1987; Milisenda et al., 1988; Davis and Hegner, 1992; Dickin, 2000).

The Kaapvaal craton in southern Africa is one of the most extensively studied fragments of Archean lithosphere and has long served as a crucible for the development of models for Archean crustal growth (Viljoen and Viljoen, 1969; Green, 1975; Anhaeusser, 1983; Richardson et al., 1984; de Wit et al., 1992; Grove and Parman, 2004; Moyen et al., 2006). However, high-resolution seismic studies and the occurrence of xenolith bearing kimberlites are restricted to the mostly Neoarchean central and western craton, whereas the best exposed and oldest basement rocks occur in the eastern craton (Carlson et al., 2000; Schmitz et al., 2004). To bridge this gap in knowledge of the 3-D petrologic and geochemical structure of the eastern Kaapvaal craton, we analysed whole-rock Sm–Nd isotopic systematics of 79 c. 3.66–2.70 Ga plutonic and orthogneiss samples from a ∼200-km long transect across potential Mesoarchean lithospheric discontinuities. We use these data to identify and refine lithospheric boundaries and to evaluate crustal growth and differentiation as a function of time. These new data, which nearly quadruple the existing database for such rocks in the region, lead to improved tectonic models for the assembly and stabilization of the eastern Kaapvaal craton.

Geological setting and sampling

The present-day crustal architecture of the Kaapvaal craton flanking the Barberton greenstone belt (BGB) is commonly explained within the context of a c. 3.2–3.3 Ga NW–SE directed subduction–accretion event, which is recorded by coeval syntectonic magmatism, syn-contractional to strike–slip basin development, and high-grade metamorphism (Kröner et al., 1991; de Wit et al., 1992; Kamo and Davis, 1994; Lowe, 1994; de Ronde and de Wit, 1994; Lowe and Byerly, 1999; de Ronde and Kamo, 2000; Stevens et al., 2002; Diener et al., 2005; Dziggel et al., 2005; Moyen et al., 2006). This period was followed by c. 100 Ma of continued transform boundary deformation, culminating in c. 3.1 Ga granitic magmatism and differential exhumation of rocks on reactivated faults (Westraat et al., 2005; Schoene and Bowring, 2007; Schoene et al., 2008).

It has been inferred that c. 3.23 Ga deformation was the result of the amalgamation of at least two microcontinental blocks, with the Saddleback-Inyoka fault system (SIFS) within the BGB representing a lithospheric suture zone (Fig. 1; de Wit et al., 1992; de Ronde and de Wit, 1994; Heubeck and Lowe, 1994; Lowe and Byerly, 1999). However, there is considerable uncertainty about the boundaries of the different terranes, their nature with depth, and how they were amalgamated. The 3.66–3.45 Ga mafic-silicic banded gneiss of the Ancient Gneiss Complex (AGC; Jackson, 1984; Compston and Kröner, 1988; Kröner et al., 1989; Kröner and Tegtmeyer, 1994) to the south of the BGB is one of the terranes (Jackson et al., 1987) and it was intruded by granodiorites to tonalites forming the Usutu Intrusive Suite from 3236 to 3220 Ma that record NW–SE compression in magmatic fabrics (Schoene and Bowring, in revision). Further south, a suite of complexly deformed orthogneisses called the Nhlangano gneiss (NGT), are c. 3240–3280 Ma (Schoene and Bowring, in revision), although its relationship to the AGC is not well-understood. The terrane to the north of the BGB has been suggested to represent a young island-arc type continental fragment that was accreted onto the AGC c. 3.23 Ga along a NE-trending boundary (de Wit et al., 1992), consistent with zircon dates from the Nelshoogte and Kaap valley plutons of c. 3236 and 3227 Ma respectively, dates ranging from c. 3240 to 3280 Ma for the Badplaas orthogneisses and a c. 3.3 Ga xenolith from within a c. 3.1 Ga granitic batholith (Kamo and Davis, 1994; Kisters et al., 2006; Schoene et al., 2008). Bordering the SW BGB, orthogneiss complexes (e.g. the c. 3.45 Ga Stolzburg and the c. 3.52 Ga Steynsdorp complexes) were also intruded by c. 3.23 Ga syntectonic tonalitic to granodioritic magmas (Stevens et al., 2002; Dziggel et al., 2005; Schoene et al., 2008), which may be coeval with high-P, low-T metamorphism in this terrane (Kisters et al., 2003; Diener et al., 2005; Moyen et al., 2006). Schoene and Bowring (in revision) propose a NE–SW trending doubly vergent subduction zone from c. 3.28 to 3.22 Ga to account for the synchronicity of magmatism and deformation north and south of the BGB and the generation of the Nhlangano gneiss. Finally, the Neoarchean saw the eruption and deposition of the Pongola Supergroup c. 2.985 Ga and the intrusion of the Usushwana complex and associated mafic to ultramafic dikes c. 2.850 Ga (Fig. 1; Hegner et al., 1984).

Figure 1.

 Geological map of the eastern Kaapvaal craton, showing sample locations and crystallization ages from this study. Also shown are the extents of lithospheric blocks: NBGB, rocks north of the Barberton Greenstone Belt; SST, the Stolzburg and Steynsdorp Terrane; AGC, Ancient Gneiss Complex; NGT, Nhlangano Gneiss Terrane. Terrane boundaries are denoted by red dashed lines. SIFS, Saddleback-Inyoka Fault System. A–A′ denotes the line onto which the samples were perpendicularly projected in Fig. 2. Map compiled from Wilson, 1982; de Wit, 1982; de Ronde et al., 1994; Lowe and Byerly, 1999; Schoene et al., 2008; Schoene and Bowring (in revision).

On the basis of geochronology of exposed (meta)igneous rocks, we divided the study area into four areas (Fig. 1): The NBGB (rocks NW of the SIFS, NW of the BGB), the SST (the Stolzburg and Steynsdorp terrane), the AGC (Ancient Gneiss Complex including and the Usutu intrusive suite) and the NGT (the Nhlangano gneiss terrane and associated rocks). From these four areas, we collected plutonic and orthogneiss samples that range in composition from granite to gabbro with U–Pb crystallization ages between c. 3.66 and 2.7 Ga (Fig. 1). Sample descriptions and Sm–Nd analyses are presented in Table 1. Analytical methods are described in the Supporting Information.

Table 1.   Sample summary and Sm–Nd data.
Sample descriptionAnalytical data
Rock unit
inline image
inline image
  1. (a) Sample name, grouped by approximate age. *Duplicate analysis.

  2. (b) Name of rock unit or locality for rocks whose associations are difficult.

  3. (c) Lithology of rock based on hand-sample description.

  4. (d) Latitude and longitdue derived by hand-held global positioning system. Terr., terrane name from Fig. 1.

  5. (e) Approximate U–Pb crystallization age in Ma, with references.

  6. No symbol = exact rock sample dated; *# = same unit dated, but different sample in this study; **# = inferred same unit dated, based on rock type or map correlation; **no number = poor age constraints.

  7. (1) Layer et al. (1989); (2) Schoene and Bowring (in revision); (3) Wilson (1982); (4) Kamo and Davis (1994); (5) Schoene et al. (2008); (6) Schoene and Bowring (2007); (7) Schoene et al. (2006); (8) Kröner et al. (1989); (9) Kröner and Tegtmeyer (1994).

  8. (f) Concentrations in p.p.m., as determined by isotope dilution.

  9. (g) Isotopic ratios corrected for mass fractionation (see text) and 100 pg Sm and Nd blank. Precision on Sm/Nd ratios ≤0.1%.

  10. (h) 2-sigma standard error of fractionation-corrected ratio.

  11. (i) Epsilon Nd value at the present day (0) and the time of crystallization (t), calculated as follows: [(143Nd/144Nd)sample/(143Nd/144Nd)CHUR) − 1] × 104 see text for CHUR values used.

c. 2.73 Ga granites
 WKC0066Mbabane plutonCo. grained granite26.304631.1551AGC2690*19.6659.520.098150.5105636−40.47−6.38
 WKC0071Hlatikulu graniteMed. grained granite26.748431.3851NGT273027.2541.760.104930.5106396−39.00−6.56
 WKC0078Hlatikulu graniteMed. grained granite26.755631.4540NGT273029.5659.280.097530.5104996−41.73−6.67
 WKC0081Ngwempisi graniteCo. grained granite26.600031.1614AGC2730**316.44105.680.094040.5104998−41.72−5.48
 EKC02-25Malendela plutonCo. grained granite26.202931.3340AGC2730**313.6394.270.087390.5102358−46.87−8.23
 EKC02-30Mbabane plutonCo. grained granite26.330231.2626AGC2690*114.7796.720.092300.5103566−44.51−8.41
 EKC02-34Ngwempisi graniteCo. grained granite26.598531.1524AGC2730**322.32136.330.098990.5104456−42.77−8.24
 EKC02-65Hlatikulu graniteMed. grained granite26.734931.4008NGT2729216.6294.460.106400.5106096−39.58−7.90
 EKC02-68Sinceni plutonCo. grained granite26.666731.5505NGT2730**310.1048.560.125770.5110326−31.32−6.26
 BS04-13Hlatikulu graniteFi. grained granite26.865131.2911NGT2730*228.39161.930.105990.5105754−40.24−8.18
 BS04-20Hlatikulu graniteMed. grained granite26.854331.2933NGT2733219.29112.780.103410.5105736−40.29−7.37
 BS04-21Hlatikulu graniteMed. grained granite26.872631.2727NGT2730222.08149.100.089540.5103096−45.42−7.61
c. 3.1 Ga rocks
 KPV99-78Nelspruit batholithMed. grained granite25.077230.1304NBGB3106*4, 515.9291.070.105680.51075812−36.67−0.23
 KPV99-79Nelspruit batholithPorphyritic granite25.359930.9937NBGB3106*4, 52.7710.850.154560.5120988−10.536.39
 KPV99-80Nelspruit batholithMed. grained granite25.359930.9937NBGB3106*4, 56.6742.100.095840.5105818−40.130.24
 KPV99-83Nelspruit batholithMed. grained granite25.473930.9640NBGB3106*4, 514.3986.600.100490.5105236−41.25−2.76
 KPV99-84Nelspruit batholithPeg. dike25.473930.9640NBGB3106*4, 513.4687.600.092920.5104788−42.14−0.59
 KPV99-86Stentor plutonHomo. granodiorite25.617431.3064NBGB3106*4, 56.3440.610.094410.5104898−41.91−0.96
 KPV99-87Stentor plutonFol. granodiorite25.617431.3064NBGB3106*4, 54.9322.020.135390.51152022−21.802.80
 KPV99-88Stentor plutonPegamtitic granite25.617431.3064NBGB3107*4, 52.486.960.215290.51304367.900.46
 KPV99-89Stentor plutonFol. granodiorite25.649731.2426NBGB310756.6942.850.094370.5105276−41.18−0.21
 KPV99-91Stentor plutonLate peg. dike25.649731.2426NBGB310452.7611.590.144060.51150428−22.12−1.10
 KPV99-91*Stentor plutonLate peg. dike25.649731.2426NBGB310452.138.890.145160.51150416−22.12−1.49
 KPV99-102Mpuluzi batholithMed. grained granite26.083430.7348SST3107*41.258.520.088950.5101908−47.75−4.64
 WKC0045Mpuluzi batholithMed. grained granite26.156530.7532SST3107*415.1581.390.112540.5107546−36.76−3.08
 WKC0055Pigg’s Peak batholithMed. grained granite26.067431.1830AGC3140*65.6335.180.096670.5104746−42.22−2.09
 EKC02-49Mpuluzi batholithFi. Grained granite26.189830.9604SST3107*45.9133.970.105100.5106256−39.27−2.61
 EKC02-55Mpuluzi batholithMed. grained granite26.464531.0050SST3107*45.5730.680.109700.5106996−37.83−3.01
 EKC03-1Pigg’s Peak batholithMed. grained granite26.033331.1667AGC3140*65.3529.420.110030.5107806−36.23−1.54
3.2–3.3 Ga rocks
 KPV 99-81Nelspruit batholithAmphibolite xenolith25.359930.9937NBGB3250**11.1762.060.108800.51112714−29.477.72
 KPV 99-82Nelspruit basementBanded gneiss25.473930.9640NBGB3258*511.2870.570.096650.5105148−41.430.60
 KPV99-85Kaap Valley plutonFol. grained tonalite25.768831.0613NBGB3227*4, 72.9615.350.116680.5109768−32.430.88
 KPV 99-90Stentor basmentBanded gneiss25.649731.2426NBGB325854.3323.840.109920.5108218−35.441.02
 KPV 99-90*Stentor basmentBanded gneiss25.649731.2426NBGB325854.3023.830.109060.5108018−35.831.00
 KPV99-92Kaap Valley plutonFol. tonalite25.757030.8452NBGB3227*4, 73.3316.730.120200.5110786−30.421.43
 KPV99-94Nelshoogte PlutonFol. granoriorite25.891230.6235NBGB323650.703.970.105840.51076310−36.571.35
 KPV99-95Stolzburg pluton – youngAplitic dike26.021830.7442SST3213*4, 53.0718.820.098590.5105506−40.73−0.09
 KPV99-97Stolzburg pluton – youngFi. grained granodiorite26.021830.7442SST3213*4, 52.7115.530.105580.5106458−38.87−1.14
 KPV99-98Stolzburg pluton – youngFi. grained granodiorite26.021830.7442SST3213*4, 52.1611.730.111330.5107676−36.50−1.14
 KPV99-101Dalmein plutonMed. grain granodiorite26.073330.9007SST3216*46.4734.800.112500.5108006−35.86−0.95
 KPV99-103Mpuluzi batholithMafic tonalite xenolith26.083430.7348SST3200**5.4525.460.129370.5112028−28.01−0.10
 WKC0068Ancient Gneiss ComplexLate pegmatitic dike26.345531.1453AGC3200**2.1311.130.115870.5108746−34.41−1.08
 WKC0079Usutu suiteMed. grain granodiorite26.521631.2529AGC3227*20.734.960.089190.5102426−46.73−1.98
 WKC0088Vlakplaats plugMed. grain granodiorite26.140330.9809SST323053.4719.970.105150.5106518−38.77−0.65
 EKC02-1Dalmein plutonMed. grain granodiorite26.088430.9373SST3216*48.0252.030.093190.5104136−43.41−0.49
 EKC02-51Kaap Valley plutonFol. tonalite25.757830.8455NBGB322772.4913.600.110810.5108258−35.360.38
 EKC03-18Usutu suiteMed. grain granodiorite26.284131.3904AGC318622.3714.130.101510.5106116−39.54−0.32
 AGC01-4Usutu suiteFol. tonalite25.895231.3069AGC322667.6241.980.109780.5107166−37.50−1.36
 EKC02-23Usutu suiteMegacryst. granodiorite26.265731.4450AGC323025.0631.240.097960.5107344−37.154.00
 EKC02-24Usutu suiteMed. grained tonalite26.265731.4450AGC322728.9744.720.121230.5109716−32.52−1.13
 EKC02-32Usutu suiteMed. grained tonalite26.284131.3904AGC323228.8760.360.088820.5102176−47.22−2.28
 EKC02-35Usutu suiteMed. grain leucotonalite26.601031.1666AGC323623.3020.120.099230.5104546−42.61−2.04
 EKC03-21Usutu suiteMed. grained tonalite26.553331.3096AGC323222.6516.330.098060.5104824−42.06−1.00
 EKC03-23Usutu suiteFol. tonalite26.054731.1472AGC323028.2741.800.119590.5110746−30.501.62
 EKC03-33Usutu suiteMegacryst. granodiorite26.054031.1529AGC3227210.6160.240.106480.5105876−40.00−2.46
 EKC02-64Nhlangano gneissPorph. banded gneiss26.667531.3977AGC/NGT326622.6820.870.077620.5099696−52.06−1.92
 EKC02-66Nhlangano gneissTonalitic banded gneiss26.800331.3371NGT326621.6810.220.099700.5105426−40.88−0.03
 EKC03-35Nhlangano gneissTonalitic banded gneiss26.876031.2805NGT330025.8937.810.094180.5103586−44.47−0.84
 EKC03-36Nhlangano gneissMegacryst grano. gneiss26.876431.2819NGT330023.9922.400.107620.5106896−38.01−0.10
 EKC03-37Nhlangano gneissAmphibolite banded gneiss26.876331.2834NGT330024.9319.610.151880.5115746−20.76−1.70
 BS04-6Usutu suiteMed. grain granodiorite26.576631.3455AGC/NGT321920.846.490.077880.5100988−49.54−0.23
 BS04-8Usutu suiteMed. grain leucotonalite26.576631.3455AGC/NGT323222.4217.710.082560.5100776−49.96−2.42
 BS04-18Nhlangano gneissAplitic banded gneiss26.854031.2987AGC/NGT3300*26.9138.560.108350.5107156−37.510.09
>3.45 Ga rocks
 KPV 99-96Stolzburg plutonTon./grano. banded gneiss26.021830.7442SST344553.1616.760.114020.5107888−36.080.67
 WKC0048Ancient Gneiss ComplexTonalitic banded gneiss25.909831.3030AGC3664*5, 819.60143.330.082680.5101926−47.716.29
 WKC0049Ancient Gneiss ComplexTonalitic banded gneiss25.909831.3030AGC3550*5, 86.1822.700.164600.5116766−18.77−4.10
 WKC0056Ancient Gneiss ComplexTonalitic banded gneiss26.064231.1831AGC3550*5, 85.3124.600.130470.5110746−30.51−1.81
 WKC0067Ancient Gneiss ComplexTonalitic banded gneiss26.345531.1453AGC3500*5, 814.0172.840.116310.5108156−35.560.75
 WKC0082Ancient Gneiss ComplexAmphibolite banded gneiss26.674731.1266AGC3550*5, 86.6024.720.161520.5117576−17.18−1.07
 WKC0083Ancient Gneiss ComplexPeg. banded gneiss26.674731.1266AGC3550*5, 83.1512.310.154660.5115856−20.53−1.28
 WKC0084Ancient Gneiss ComplexTonalitic banded gneiss26.674731.1266AGC3550*5, 82.5212.710.119690.5108794−34.310.98
 WKC0087Steynsdorp plutonFol. tonalite-granodiorite26.177930.9871SST3517*4, 53.0614.340.129050.5113168−25.794.96
 AGC01-1Tsawela gneissTonalitic banded gneiss26.755830.9794AGC3450*93.2216.380.118790.5108506−34.88−0.20
 AGC01-2Ancient Gneiss ComplexTonalitic banded gneiss26.714131.0448AGC3550*5, 81.067.570.084570.5101828−47.913.49
 AGC01-5Ancient Gneiss ComplexTonalitic banded gneiss25.896131.3104AGC366257.5937.980.120730.5107906−36.05−0.11
 EKC02-20Phophonyane graniteFi. grained fol. granite25.896131.3104AGC354552.8215.220.111880.5105336−41.06−2.27
 EKC02-40Steynsdorp plutonFol. tonalite-granodiorite26.177930.9604AGC351752.5613.990.110510.5107096−37.631.50
 EKC03-3Stolzburg plutonHomo. granodiorite25.974030.7630SST344550.784.080.116390.51081310−35.610.09
 BS04-7Tsawela gneissTonalitic banded gneiss26.576631.3455AGC341123.8622.960.101680.5104336−43.02−1.21

Sm–Nd results and discussion

Nd isotopic map of the eastern Kaapvaal craton

To evaluate the role of older crust in the generation of exposed rocks, we examine the range of εNd(t) (εNd at the time of crystallization) across the crustal transect for rocks of similar age. Magmatism occurred across the entire study area c. 3.2–3.3 Ga; c. 3.1 Ga granitic magmatism also extends from the AGC across the SST into the NBGB. Abundant c. 2.7 Ga magmatism occurred in the AGC and NGT only (Fig. 1). Figure 2 shows the sample locations projected onto a line trending 325° (approximately perpendicular to the NE–SW regional structural grain; cross-section line shown in Fig. 1) plotted against εNd(t). Both c. 3.2–3.3 and c. 3.1 Ga rocks show higher εNd(t) values in the NBGB. In the NBGB, c. 3.1 Ga samples have εNd(t) > −2 and samples to the south have εNd(t) < −2. For 3.2–3.3 Ga rocks, there are no samples with εNd(t) < 0 in the NBGB, whereas εNd(t) for samples in other areas ranges from 0 to −3 (with two exceptions). No 3.2–3.3 Ga samples in the SST have εNd(t) < −2. There is similar scatter in εNd(t) in 3.2–3.3 Ga samples on both sides of the boundary between the AGC and NGT, indicating that a similar previously enriched reservoir contributed to those rocks. Similarly, for c. 2.7 Ga granites, there is no difference in εNd(t) across the surface expression of that boundary.

Figure 2.

 Nd isotopic data and Sm/Nd plotted as a function of distance along the transect A–A′, shown in Fig. 1, for magmatic rocks 2.73, 3.1 and 3.2–3.3 Ga. Lithospheric boundaries inferred from surface geology indicated by dashed lines. See text for abbreviations and Fig. 1 for terrane locations in map view. Error bars on εNd are fixed at ±0.5 at the 2-sigma level (see Supporting Information), and errors in Sm/Nd are smaller than symbol.

The marked offsets in εNd(t) between coeval magmas in the NBGB, SST and AGC probably represent isotopically distinct crust/lithosphere in each terrane. Alternatively, the data may reflect similar crustal age, but different mixtures of enriched and depleted reservoirs in the resulting magmas. To test this, we also plotted the Sm/Nd ratios of rocks in Fig. 2 in each terrane. If elevated εNd(t) in the NBGB results from greater contribution of mantle-derived melt rather than from the absence of more evolved, low εNd material at depth, there may be a correlation between relatively high εNd(t) and elevated Sm/Nd. Although some of the c. 3.1 Ga granite samples from the NBGB do show such a correlation, 3.2–3.3 Ga rocks show nearly identical Sm/Nd across the SIFS despite the offset in εNd(t) (Fig. 2). Thus, the 3.2–3.3 Ga granodiorites to tonalites probably inherited some component of older crust south of the SIFS.

Timescales of crustal growth and differentiation

εNd(t) values for rocks >3.45 Ga vary from −3 to +5 (ignoring two outliers), similar to that observed previously, which probably reflects the generation of these rocks by mixing mantle and older crustal components (Carlson et al., 1983; Kröner and Tegtmeyer, 1994; Kröner et al., 1996; Kröner, 2007). Similarities in εNd(t) between these and younger rocks south of the SIFS permit us to model this domain as a single lithospheric block, allowing evaluation of Sm–Nd evolution for nearly a billion years of cratonic evolution (c. 3.66–2.73 Ga). For rocks south of the SIFS, we have plotted εNd(t) and Sm/Nd as a function of time, using data from this study and from the literature (a total of 104 data points; Fig. 3). In general, εNd(t) of plutonic rocks define a linear array between 2.7 and 3.3 Ga, while εNd(t) values of rocks >3.45 Ga are more variable and plot mostly below an extension of that array to the depleted mantle curve (Fig. 3B). Metavolcanics of the Neoarchean Pongola Supergroup and the predominantly gabbroic Usushwana complex (Fig. 1) plot well above this array.

Figure 3.

 Sm/Nd (A) and εNd (B) values from this study and from the literature plotted as a function of crystallization age, for samples south of the SIFS. Gray circles, this study; black cirles, Carlson et al. (1983); white squares, (Kröner et al. (1996); gray diamonds, Kröner and Tegtmeyer (1994); white diamonds, Hegner et al. (1984). Shaded areas outline the evolution curves for rocks of the age that correspond to the oldest portion of a given trajectory. Depleted mantle curve results from drawing a line between 0 and +10 at 4500 and 0 Ma, respectively.

To quantify these trends, we calculated εNd for each data point at five different dates corresponding to periods of magmatism between c. 2.73 and 3.25 Ga (Fig. 4). We then calculated the average εNd at those times for (1) all the magmatic rocks of that age and (2) all the older host rocks present at that time, as a proxy for average crustal εNd. For example, host rocks for c. 3.1 Ga magmas include both 3.2–3.3 Ga and >3.45 Ga rocks; Pongola and Usushwana rocks were not included as host rocks, because they are either volcanic or not regionally extensive respectively (Fig. 4). We find that at 3.25 Ga, εNd(t) of rocks south of the SIFS is elevated relative to the average crustal εNd and therefore require input from a high εNd source such as the mantle or older basaltic crust (Fig. 4). The Nd isotopic signatures of granitic magmas at c. 3.1 and 2.73 Ga can be entirely explained by in situ crustal melting with little or no mantle input. Pongola volcanics and the Usushwana intrusives also have elevated εNd(t) relative to the average crustal value, but still well below a depleted mantle value (which was ∼+4; Fig. 3B; DePaolo, 1981; Bowring and Housh, 1995). Sm/Nd of Pongola and Usushwana rocks is also consistent with a contribution from a non-crustal source (Fig. 3A), in that they mostly plot above the c. 3.1 and 2.73 Ga granites. The other data follow a trend of lower Sm/Nd with time that is consistent with an increased importance of crustal melting and differentiation over time. The εNd(t) values of the Pongola and Usushwana rocks resulted from mixing mantle and crust, but those magmatic episodes had little effect on the bulk crustal Sm–Nd systematics in that the εNd(t) of 2.73 Ga granites can be explained entirely by melting average crust.

Figure 4.

 crustal εNd evolution for rocks south of the SIFS. Points and 2-sigma errors correspond to weighted mean values of εNd for magmas of a given age and the basement rocks present at the time of crystallization. If the magma value is higher, it requires input from a depleted reservoir, if magma and basement rocks values are equal, then εNd(t) can be accounted for by intracrustal Nd recycling.

The above analysis assumes that calculated εNd(t) of >3.45 Ga basement rocks in the AGC are accurate, despite suggestions that some Archean rocks may have had their primary Sm/Nd altered through metamorphic processes based on expected εNd(t) when compared to Hf isotopic data (Gruau et al., 1996; Vervoort et al., 1996; Moorbath et al., 1997). We argue that this is not a problem in the above analysis because (1) the clustering of a vast majority of the rocks in the study between +2 and −3 over a large geographic area, (2) this cluster corresponds well to the range observed in low-grade metavolcanics at the base of the BGB (−2 to +2 at 3.45 Ga; Kröner and Tegtmeyer, 1994; Kröner et al., 1996; Chavagnac, 2004), and (3) the only regional metamorphic event documented is at c. 3.23 Ga (see references above), which corresponds to the starting point of the above analysis and thus the εNd(t) values at that time are unaffected. The few outliers that fall near or above the mantle curve in Fig. 3 may have had their Sm/Nd locally altered by metamorphic or fluid interaction processes, although it cannot be determined. We deem those samples worthy of future work, but regard the average of the cluster of analyses used above to be robust for the evolution of the crust as a whole.

Tectonic implications

Current models for the tectonic evolution of the BGB region are characterized by c. 3.23 Ga oblique subduction and accretion of an immature exotic terrane onto older lithosphere. These models are well-supported by metamorphic, petrological, geochronological, structural, and stratigraphic studies from the BGB and the flanking terranes (de Wit et al., 1992; Lowe, 1994; Clemens et al., 2006; Kisters et al., 2006; Moyen et al., 2006; Schoene et al., 2008; and references above). Our Sm–Nd isotopic mapping of the eastern Kaapvaal craton suggests that there were at least two distinct lithospheric blocks present c. 3.28–3.23 Ga, represented in Fig. 1 as the NBGB and the AGC + NGT ± SST. These new data, in combination with recently documented geochronology and field evidence south of the BGB, require that existing models be modified to include a doubly verging subduction zone to account for coeval magmatism and contractional deformation north and south of the BGB (Schoene and Bowring, in revision; Fig. 5). In such a model, from c. 3.28 to 3.24 Ga, the NBGB was formed and accreted onto pre-existing AGC and SST lithosphere, while the AGC hosted arc magmatism and deformation recorded in the Usutu intrusive suite and NGT. Consistently low εNd(t) across the AGC–NGT boundary requires older crust contributed to the NGT Nd-budget. Thus, the AGC is, or once was, more extensive than is presently exposed (Fig. 1). Inherited zircons ≤3.45 Ga in c. 3.24–3.28 Ga Nhlangano gneiss in the NGT support this hypothesis (Condie et al., 1996; Kleinhanns et al., 2003; Schoene and Bowring, in revision). The relationship between the SST and the AGC remains elusive, although slightly elevated εNd(t) of 3.2–3.3 Ga magmas compared to the AGC suggests the SST may represent one or more distinct crustal slivers, as has been hypothesized based on sedimentary and structural studies from within the BGB (Heubeck and Lowe, 1994; Lowe, 1994). Additionally, as pointed out by Schoene and Bowring (in press) and; Schoene et al. (2008), major structural modification of the crust c. 3.1 Ga needs to be addressed more closely in making c. 3.2 Ga paleoreconstructions of the eastern Kaapvaal craton.

Figure 5.

 Cartoon illustrating the c. 3.3–3.2 Ga tectonic evolution for basement rocks from the eastern Kaapvaal craton, after Schoene and Bowring (in revision), and references therein. Arrows in top panel show relative plate motions of the NBGB, and the AGC/SST. Continued subduction and hypothesized slab break-off led to melt generation by multiple mechanisms c. 3240–3220 Ma, resulting in a mixed crustal/mantle εNd(t) signature. See Fig. 1 for surface extent of lithospheric blocks and abbreviations.

Our analysis (Fig. 4) further supports a subduction origin for c. 3.28–3.22 Ga magmatism. εNd(t) for the c. 3.22–3.24 Ga Usutu intrusive suite in the AGC and the Nhlangano gneisses requires sources in the mantle and in older continental crust rather than melting lower continental crust exclusively (Figs 3 and 4). When combined with evidence for contractional deformation throughout the region and submarine basin closure in the BGB, a subduction model involving considerable crustal growth is the simplest explanation. Our data suggest that after c. 3.2 Ga, however, the eastern Kaapvaal craton behaved as a coherent lithospheric block dominated by crustal recycling through lower-crustal melting and little mantle Nd input. The proposed tectonic setting for the 3.1 Ga granites ranges from a transtensional to compressional to predominantly strike–slip kinematics (e.g. Kamo and Davis, 1994; de Ronde and de Wit, 1994; Westraat et al., 2005; Schoene et al., 2008). The geodynamic setting for c. 2.73 magmatism remains even more ambiguous, although we note the potential correlation in timing between the widespread Ventersdorp rifting event throughout the central craton (e.g. de Wit et al., 1992). In any case, these geochemical reorganizations of the crust were likely an important factor in the c. 3.1–2.7 Ga stabilization of the eastern Kaapvaal craton, in that it transported heat-producing elements (e.g. U, Th, K) into the upper crust, which can act to cool and strengthen the entire crustal column over time (see Schoene et al., 2008 and references therein).


The authors would like to thank B. Eglington, I. Schoenberg-Kleinhanns, A. Kröner, and an anonymous reviewer for insightful comments that improved the manuscript. Funding came in part from NSF EAR9526702 to S.B. This is AEON contribution #59.