PETROLOGICAL MODELS FOR THE FORMATION OF TTG GNEISSES – DO WE NEED SUBDUCTION?
However, a number of lines of evidence may be taken to imply that subduction in the modern sense first began during or in the late Archaean, and so did not operate at the time of formation of the earliest crust. If this is true, then what do the geochemical subduction signals from early Archaean rocks have to say? Here, it is important to recognise that tectonic settings can only be indirectly implied from geochemical patterns: the primary control on the geochemistry of igneous rocks is exerted by the mineral assemblages involved in melting and remaining in the residue. The requirement for low Nb/Ta in TTG gneisses can be realised either in a one-stage process by melting of basaltic crust that melts in the form of garnet amphibolite (Foley et al., 2002), or in a two-stage process in which an intermediate low-Nb/Ta source for the TTG gneisses is produced. Two-stage options include arc-related stacking of basalts giving rise to eclogite sources for the TTG gneisses (Rapp et al., 2003), and melting of amphibole pyroxenites to produce low Nb/Ta basalts that melt to give TTG gneisses in a second event (Foley et al., 2003). Both these two-stage hypotheses invoked subduction, but need not do so - the latter petrological mechanism is compatible with Mg-rich melts and higher water contents in the early Earth and, perhaps most importantly, offers a mechanism to cause low Nb/Ta reservoirs and lead to TTG gneisses without subduction in the modern sense.
S. Foley, M. Tiepolo & R. Vannucci (2002) Nature 417, 837-840; S. Foley, S.Buhre & D. Jacob (2003) Nature 421, 249-252; R. Rapp, N. Shimizu & M. Norman (2003) Nature 425, 605-609