Late Archean volcanic rocks of the Belingwe Greenstone Belt have long been known to contain a small continental crustal component evident in their radiogenic isotope and trace element signatures (Chauvel et al., 1993). Recently, we discovered new geochemical evidence suggesting that all volcanic units contain members with a distinctive crustal fingerprint. It was concluded that the volcanic rocks were erupted through pre-existing continental crust, although a specific tectonic interpretation was deliberately avoided. Hofmann and Dirks' claim that the new data are also compatible with an oceanic deposition, provided that enriched mantle sources were melted or that dismembered continental crust was present in the oceanic lithosphere. We wish to draw the reader's attention to the fact that when all available evidence is considered, continental emplacement of the rocks is an unavoidable conclusion.
The formation and tectonic evolution of the Belingwe Greenstone Belt has attracted considerable attention. At the heart of the debate lies the question of whether igneous assemblages of the Upper Greenstones formed in an oceanic setting and were later thrust along a major shear zone onto volcano-sedimentary sequences of the Lower Greenstones. A major structural hiatus could suggest, but not necessarily imply, that the greenstones were tectonically emplaced onto continental crust via horizontal accretion, after their formation as oceanic plateaus, arcs, or at mid-oceanic ridges (e.g., Kusky and Kidd, 1992).
The investigation of the geochemical data by Bolhar et al. (2003) revealed a ubiquitous crustal signature in volcanic rocks from all five stratigraphic units in the Belingwe Greenstone Belt, providing direct evidence for the presence of continental crust prior to or during emplacement, in agreement with previous studies:
Chauvel et al. (1993) reported correlated model source 238U/204Pb and initial 143Nd/144Nd ratios in Reliance Formation volcanic rocks, consistent with assimilation of 3.5 Ga continental material.
Wilson et al. (1995) found xenocrystic zircons in volcanic samples from various Zimbabwean greenstone belts including the Belingwe Greenstone Belt, with U/Pb ages substantially older than those of their host rocks.
Lower crustal garnets were recovered from Reliance Formation komatiites (Shimizu et al., 2002), consistent with assimilation of lower crust.
Re-Os systematics of chromites from ultramafic complexes in the vicinity of the Belingwe Greenstone Belt provide geochronological and geochemical evidence for successive growth of the Zimbabwe craton lithosphere since 3.8 Ga (Nägler et al., 1997), suggesting that the craton was already (part of?) a continental entity by the time of Belingwe Greenstone Belt basalt eruption.
Blenkinsop et al. (1993) summarized evidence in favor of an ensialic origin for the Upper Greenstones, including the absence of a major high-temperature shear zone at the base of the Reliance Formation. They also noted the presence of quartzose sandstones intercalated with the Reliance Formation, negating an oceanic plateau origin.
Hunter et al. (1998) recognized continental-type depositional facies in the Manjeri Formation underlying Upper Greenstone volcanic rocks. Highly variable rare earth element and Nd isotope compositions with model ages of 3.5–2.9 Ga imply derivation of sediment source material from basement units.
2.7–2.6 Ga stromatolites in the Belingwe Greenstone Belt have distinctive Pb isotopes derived from a long-lived source with substantially higher 238U/204Pb than mantle (Bolhar et al., 2002). Pb of such isotopic character is only known from a few so-called high-μ cratons, of which the Zimbabwe craton is a prominent representative. It thus appears that the Pb isotopes were derived from the old gneisses of the Zimbabwe craton, strongly arguing against an oceanic origin. Sm/Nd isotope systematics are also consistent with derivation from water with a strong continental signature.
In view of this overwhelming body of evidence, we remain unconvinced by Hofmann and Dirks' suggestion that these greenstones formed in the oceanic realm. We re-emphasize that the existence of a major metamorphic and structural disconformity has not been verified by subsequent studies (Blenkinsop et al., 1993). The validity of inferring horizontal accretion on the basis of sheared contacts alone appears problematic too, since the extent of tectonic displacement cannot be quantified. Even tectonic transport involving large-scale displacement does not preclude extrusion onto pre-existing crust, because interaction of asthenospheric melts with crust could have occurred prior to horizontal accretion.
Finally, we refute Hofmann and Dirks' suggestion that lithological variation provides a more reliable discriminator of tectonic setting than trace element geochemistry. The remarkably coherent geochemical variations between different stratigraphic units require very similar processes of magma genesis. Once magmas are formed, lithological variation can be imposed by a variety of environmental factors that are common to many tectonic settings.
This reply benefited greatly from comments by Balz Kamber and Tom Blenkinsop.