Modeling multiple melt loss events in the evolution of an active continental margin

The Fosdick migmatite-granite complex in West Antarctica records evidence for crustal melting during two periods of tectonism along the East Gondwana margin. Initial high-temperature metamorphism in the Devonian-Carboniferous (M-1) was broadly contemporaneous with emplacement of calc-alkaline arc magmas during Pacific-style accretionary margin convergence. This event involved metamorphism of arc plutonic rocks soon after their emplacement and partial melting and migmatization of host metasedimentary rocks. Preservation of M-1 garnet-bearing assemblages and mineral equilibria modeling of the metasedimentary rock and calc-alkaline plutonic rock protolith compositions regionally exposed constrain conditions of M-1 metamorphism to 820-870 degrees C and 7.5-11.5 kbar. A second anatectic event during the Cretaceous (M-2) resulted in metamorphism of plutonic rocks and partial melting of fertile metasedimentary rocks that had remained at a high enough structural level to have been subsolidus during the first anatectic event, and a metamorphic overprint on now residual paragneisses characterized by the growth of M-2 cordierite after garnet, and after biotite + sillimanite. Mineral equilibria modeling of para- and orthogneiss compositions in the Fosdick migmatite-granite complex constrain conditions of M-2 metamorphism to 830-870 degrees C and 6-7.5 kbar. We use the results of mineral equilibria modeling to assess the constituents of the Fosdick migmatite-granite complex as melt sources and as domains of melt transfer and melt accumulation during the two anatectic events. Modeling the range of metasedimentary rock protolith compositions shows that similar to 4-25 vol.% melt was produced at the conditions of M-1 metamorphism, although most compositions would have been fertile enough to reach the melt connectivity transition (similar to 7 vol.%) leading to the development of a melt extraction pathway and subsequent melt loss. The preservation of peak-M-1 assemblages in the paragneiss is consistent with melt loss, and modeling based on a representative protolith composition indicates that a minimum of 70% of the total melt produced must have been extracted from the metasedimentary rock source. The intrusive plutonic rocks produce similar to 2-3 vol.% melt during the M-1 event. Although the plutonic rocks were not a significant melt source at the level exposed, granites derived from these rocks but sourced at a deeper crustal level accumulated within the Fosdick migmatite-granite complex during the M-1 event. The elevated geotherms in a magmatic arc environment allow the possibility that higher crustal levels in the Fosdick migmatite-granite complex remained subsolid us during the M-1 event, and could be fertile sources during the M-2 event. At peak M-2 conditions, these fertile metasedimentary rocks would produce similar to 5-30 vol.% melt, whereas the plutonic rocks were not likely a significant source of melt at this crustal depth. The residual paragneisses that underwent melting and melt loss during the M-1 event are estimated to produce similar to 12 vol.% additional melt during M-2. The mechanical anisotropy created by the residual gneisses likely produced a gradient in melt pressure. This gradient, together with shallow fabrics in the hosting gneisses, could have acted to focus M-2 melts derived from the fertile metasedimentary rocks into a horizontally-sheeted leucogranite complex. The accumulation of these melts would have lead to pronounced weakening of the crust, facilitating the exhumation of the Fosdick migmatite-granite complex during the transition from regional shortening during M-2 to regional extension at ca. 100 Ma. (C) 2009 Elsevier B.V. All rights reserved.