Geophysical models (De Beer and Stettler, 1992; Durheim et al., 1992) indicate that the crust beneath the 530 Ma (Allsopp et al., 1995) kimberlite pipes in the Central Zone (CZ) of the Limpopo Belt is composed of 8 km of metasedimentary rocks, trondhjemitic-tonalitic lithologies and granitoid rocks (classical CZ lithologies) underlain by a seismically transparent layer, to a depth of 37 km, in which P-wave velocities (Vp) abruptly jump to 8.1 km/s (the seismic Moho). An anomalous low velocity layer (LVL where Vp = 7.6 km/s) is present from 42 to 56 km, and was originally thought to be due to elevated fluid levels (Durheim et al., 1992). It has, however, been shown to be composed largely of anhydrous high pressure granulites (Pretorius and Barton, 1995). Below 56 km, Vp returns to normal mantle velocities of 8.1 km/s. Samples representing amphibolite, granulite (medium [<11 kbar] and high pressure [>11 kbar]) and eclogite facies assemblages were the subject of geochemical studies (e.g. mineral chemical, bulk rock chemical, rare earth elemental and isotopic). Results indicate that some of the amphibole rich samples represent two chemically distinct precursors, felsic-intermediate and mafic, that underwent dehydration melting and partial melt extraction, leading to development of two distinct restite assemblages: a garnet-quartz-ilmenite rock and a granulitic assemblage (cpx + plag + ga). Textural evidence suggests that (pre- to syntectonic) garnet growth was synchronous with partial melting, implying pressures around 10 to 12.5 kbar and temperatures above about 850 C (Wolf and Wyllie, 1993; Wolf and Wyllie, 1994; Patino-Douce and Beard, 1995) and that at least some partial melting could have taken place contemporaneously with peak metamorphic conditions in the CZ. Both precursor assemblages have small negative Eu anomalies and plagioclase is thought to be an important melt phase, implying that trondhjemite-tonalite melt extracted from them could have variable Eu anomalies which would be a function of degree of partial melting and source heterogeneity. Syn-tectonic plutons intruding the CZ have constant REE patterns and are incompatible with the predicted REE signature. However, some >3.0 Ga tonalitic-trondhjemitic gneisses of the CZ and Zimbabwe Craton to the north, and >2.8 Ga gneisses in the Kaapvaal Craton to the south, have such variable anomalies. This implies partial melting at lower crustal levels in the Archean, preceding high grade metamorphism in the CZ, in turn pointing to a largely amphibolitic, intermediate to mafic lower crust during this time. Calculated compressional wave velocities for the two restite assemblages are similar to those observed at the the seismic Moho (i.e. 8.1 km/sec). Some other medium pressure granulites as well as the high pressure granulite have REE signatures indicative of cumulate plagioclase. These medium pressure granulites are genetically unrelated to either the high pressure granulite or the amphibolite facies samples that underwent partial melting. Group B to C eclogite samples (Coleman et al., 1965), have bulk rock and mineral compositions and REE signatures incompatible with them being either primary mantle melts or subducted oceanic crust. They are more compatible with being high pressure cumulates, involving garnet. Preliminary and unconfirmed Sm-Nd TDM ages from mineral separates indicate Archean ages for the eclogites. Tectonic and petrological considerations imply large scale underplating events during crustal and mantle evolution below the Venetia pipes, which preceded the formation of the Limpopo Belt. Conductive heating associated with underplating may have elevated already steep Archean geotherms, leading to partial melting of amphibolitic lower crust. Based on the geochemical and geophysical data, the seismic Moho is interpreted to be intracrustal, and related to restite development associated with these lower crustal partial melting events. The petrologic Moho occurs at 56 km depth as evidenced by the high pressure anhydrous granulitic LVL (Pretorius and Barton, 1995). The high density (garnetiferous) restitic and LVL seem to have resisted delamination during formation of the Limpopo Belt, and thus the continental lithosphere below the pipes is thick. The anomalously mafic mid to lower average crustal compositions calculated for the crust below the pipes is consistent with this interpretation.
Allsopp, H.L et al., South African Journal of Geology 3, 239-244 (1995).
Coleman, R.G. et al., Bulletin of the Geological Society of America 76, 483-508 (1965).
De Beer, J.H. & Stettler, E.H., Precambrian Research 55, 173-186 (1992).
Durheim, R.I., Barker, W.H. & Green, R.W.E., Precambrian Research 55, 187-200 (1992).
Patino-Douce, A.E. & Beard, J.S., Journal of Petrology 36, 707-738 (1995).
Pretorius, W. & Barton, J.M. Jr., Centennial Geocongress, Extended abstracts, Johannesburg, South Africa 1, 335-337 (1995).
Wolf, M. & Wyllie, P.J., Geology 101, 357-373 (1993).
Wolf, M. & Wyllie, P.J., Contributions to Mineralogy and Petrology 115, 369-383 (1994).