HR: 16:30h
AN: V34A-03    [Abstracts]
TI: Mass transfer in subduction zones: an elemental and isotopic perspective
AU: * Turner, S
EM: sturner@els.mq.edu.au
AF: Department of Earth and Planetary Sciences, Macquarie University, Sydney, 2109 Australia
AU: George, R
EM: rgeorge@els.mq.edu.au
AF: Department of Earth and Planetary Sciences, Macquarie University, Sydney, 2109 Australia
AB: Little doubt remains that subduction zone lavas contain elements recycled from the subducting slab. However, whether the key agents of this mass transfer are fluids, supercritical fluids or melts has major implications for the thermal structure of the mantle wedge. The evidence for contributions from both subducted sediment and altered oceanic crust are compelling and in most arcs their relative proportions vary inversely. Thus, so-called "fluid-rich" lavas with high Ba/Th and Sr.Th ratios have low Sr and Be isotopes etc and the converse is true for the so-called "sediment-rich"lavas with elevated La/Sm. A complicating factor is that many individual arcs tend to be dominated by one end-member. Nevertheless, experimental partition coefficient data are consistent with the differences between the fluid and sediment components being formed in the presence of different residual mineralogies. Sediment fluids appear to be poor in incompatible elements, relative to those derived from altered oceanic crust and cannot easily replicate the sediment end-member. We suggest that subducted sediments dehydrate at relatively shallow levels and that these fluids are not strongly sampled by arc lavas. Altered oceanic crust may dehydrate more extensively and to greater depths and may be buffered against melting. Model melts of dehydrated sediment residues provide a much better simulation of the inferred sediment end-member but may require ~800 C at ~ 2GPa, consistent with recent temperature-dependant viscosity models. These general inferences are strongly supported by Be and U-series isotope data which suggest that the sediment (melt) end-member is added 100's kyr to several Myr prior to eruption whereas addition of fluid components continues until a few 1000 yrs prior to eruption. Thus, the fluid and sediment end-member contributions are separate in composition, space and time. These data argue strongly against the involvement of any single supercritical fluid.
DE: 8410 Geochemical modeling (1009, 3610)
DE: 8413 Subduction zone processes (1031, 3060, 3613, 8170)
SC: Volcanology, Geochemistry, Petrology [V]
MN: Fall Meeting 2005