Tracing subducted sediment inputs to the Ryukyu arc-Okinawa Trough system: Evidence from thallium isotopes

Abstract

Sediments are actively subducted in virtually every arc worldwide. However, quantifying their contributions to arc lavas and thereby establishing budgets of how sediments participate in slab-mantle interaction is challenging. In this contribution we use thallium (Tl) abundances and isotopic compositions of lavas from the Ryukyu arc (including south Kyushu) and its back-arc basin, Okinawa Trough, to investigate the influence of sediments from arc to back-arc. We also present extensive geochemical data for sediments and altered oceanic crust (AOC) outboard of the northern (DSDP Sites 296, 442B, 443 and 444) and central (DSDP Sites 294 and 295) part of the Ryukyu arc. The Tl isotopic compositions of sediments change systematically from lighter outboard of northern Ryukyu arc to heavier outboard of central Ryukyu arc. The feature reflects the dominance of terrigenous material and pelagic sedimentation outboard of the northern and central Ryukyu arc, respectively. Central and northern sections of Ryukyu arc and Okinawa Trough display larger range of Tl isotopic variation than southern section, which is consistent with more pelagic provenance for sediments outboard of central and northern Ryukyu arcs than that of expected sediments outboard of southern Ryukyu arc. Identical Tl, Sr, Nd and Pb isotope variations are found when comparing arc and back arc lavas, which indicates that sediments fluxes also account for the Tl isotopic variations in the Okinawa Trough lavas. Two-end-member mixing models of Tl with Pb, Sr and Nd isotopes require sediment inputs of< 1%, 0.1–1% and 0.3–2% by weight to the depleted mantle source to account for all these isotopic compositions of lavas from northern, central and southern portion of the Ryukyu arc and Okinawa Trough. Bulk mixing between mantle and sediment end members predict very similar sediment fluxes when using Tl, Sr, Nd and Pb isotopes, which indicates that fractionation of these elements must have happened after mixing between mantle and sediments. This conclusion is corroborated by model calculations of mixing between sediment melts with fractionated Sr/Nd ratios and mantle wedge, which show that no arc lava plot on such mixing lines. Thus bulk sediment mixing, rather than sediment melt, is required for the generation of the lavas from the Ryukyu arc and Okinawa Trough. The requirement of bulk sediment mixing occurring before trace element fractionation in the sub-arc mantle is consistent with models where mélange layers form at the top of the slab and are the principle source material for arc lavas. In addition, the fact that sediment components observed in the Ryukyu arc and Okinawa Trough lavas are similar, suggests that transport of mélange material to the source regions of the arc and back arc is equally efficient. This feature is most readily explained if mélange material is transported from the slab as diapirs.