Anders Scherstén

We present new geochemical data for three Mesoarchaean supracrustal belts (Ravns Storo, Bjornesund and Perserajoorsuaq) situated north of Frederikshab Isblink in southern West Greenland for which we propose the collective name 'The Ikkattup Nunaa Supracrustal Association'. They comprise mainly amphibolites of tholeiitic basalt composition and leucoamphibolites of calc-alkaline andesite composition. Both lithological units are cut by aplite sheets of tonalite-trondhjemite-granodiorite (TTG) composition with U-Pb zircon ages of c. 2900 Ma. Lu-Hf and Sm-Nd isochrons based on whole rock amphibolite and leucoamphibolite samples yield ages of 2990 +/- 41 Ma and 3020 +/- 78 Ma, respectively, which are within error of the age of the Fiskenaesset Complex situated 5-25 km to the north. Leucoamphibolites from the three supracrustal belts show apparent geochemical mixing trends between tholeiitic amphibolites and TTG gneisses, as the end-members. By assimilation-fractional-crystallisation (AFC) modelling we can show that one group of leucoamphibolites can indeed be explained by contamination of the parental melts by a TTG-like end-member and another group of high P2O5, La and Nb leucoamphibolites can be explained by contamination involving a hypothetical low-silica adakite end-member. However, the leucoamphibolites are juvenile with epsilon Nd-(2970 Ma) from +2.1 to +3.5 and epsilon Hf-(2970 (Ma)) of +3.5 to +4.3. Thus, the mafic source of the felsic contaminant melts must have been derived from a depleted mantle source more or less at the same time (<60 Ma) as the volcanism took place. Our preferred interpretation of the geochemical and isotope data is that the protoliths of the supracrustal rocks formed in an island arc setting, where early tholeiitic volcanism gave way to calc-alkaline volcanism in a maturing arc. The apparent AFC trends are thus explained by in situ partial melting of basaltic arc crust to form juvenile TTG- and adakite-melts that mixed with mafic magmas or contaminated their mantle source to produce the calc-alkaline leucoamphibolite protolith. This model has important implications for the general interpretation of other Archaean supracrustal belts, because AFC and geochemical mixing trends towards a TTG end-member are not uniquely diagnostic of crustal contamination, but may rather reflect processes operating at lower crustal or source levels in arcs, such as melting-assimilation-storage-homogenisation (MASH) or slab-melt (low-silica adakite) metasomatism of the mantle source, respectively. (C) 2012 Elsevier B.V. All rights reserved.