Craton nucleation and formation of thick lithospheric roots

Cratonisation is regarded as coincident with amalgamation and lithosphere thickening during collisional tectonics, leading to the birth of the cratonic entities we know today. These are, however, themselves rifted remnants of earlier plate configurations involving cratonic nuclei whose mode of formation conditioned their properties, hence stability and further evolution. Dating of inclusions in diamonds reveals that some thick cratonic roots existed well in advance of accretionary processes leading to craton amalgamation or reflected in the ages of eclogites residing in cratonic mantle. Major uncertainties relate to (1) whether these nuclei formed at ambient or excess mantle potential temperatures T-P, i.e. in mid-ocean ridge or oceanic plateau settings and (2) whether the deep, garnet-bearing cratonic mantle, a portion of which extends into the diamond stability field (> 150 km), formed by plume subcretion or stacking of oceanic lithosphere. In order to extract constraints from published data sets that include samples modified by mantle metasomatism, olivine inclusions in diamonds are used as a guide to the pre-metasomatic composition of the cratonic subcontinental lithospheric mantle (SCLM). The FeO-MgO relationships of the filtered data set show that many depleted cratonic SCLM sections nucleated by onset of partial melting at pressures > 5 GPa, possibly pointing to formation at excess T-P typical of plumes, rather than at ambient T-P in Archaean mid-ocean ridge settings. Cr2O3-Al2O3 relationships constrain pressures of equilibrium partial melt extraction. All cratonic depleted garnet peridotite suites have compositions consistent with formation at pressures >= 3 GPa and hence formed in situ and not by subduction of oceanic plates. Because most cratonic SCLM sections formed as high-pressure residues consisting of olivine and aluminous orthopyroxene, the absence of a garnet signature in residual peridotites that last equilibrated at pressures >3 GPa is not proof for original partial melting in the spinel stability field. Cratonic crust may have formed by shallow plate interactions leading to generation of TTGs and greenstone belts. These crustal nuclei may have been subsequently penetrated and subcreted by plume-derived melts, and, after delamination of dense eclogite and komatiites, were trapped by buoyant plume residues. This ensured craton longevity and implies that cratonic crust and mantle are not entirely cogenetic. (C) 2012 Elsevier B.V. All rights reserved.