THE CHEMICAL COMPOSITION OF SUBDUCTING SEDIMENTS

Sediments on the seafloor are records of erosion of the continents and productivity of marine life. They are also input to the world's subduction zones, supplying water and chemical components to arc volcanoes, and contaminants to the deeper mantle. Here I provide a new estimation of global subducting sediment, GLOSS-II, motivated by an expanding inductively-coupled plasma mass spectrometry (ICP-MS) dataset for trace elements. Specific focus is on key elements in solid earth recycling that are now better constrained by new data: Li, Be, B, Nb, Ta, Pb, Th and U. The abundances of these elements reflect, to first order, dilution of continental detritus by biogenic opal and carbonate, and secondly, various continental and marine processes, such as the extent of continental weathering (Li/K2O), hydrogenous vs. hydrothermal oxides (Pb/Fe), biophosphate accumulation and exposure age on the seafloor (Th and U), and organic carbon burial (U). The Be/K2O ratio (0.86 +/- 0.04) and Nb/Ta ratio (14 +/- 1) of terrigenous marine sediments reflects average upper continental crust, and indeed refine those estimates. The mass flux of Li subducted into different trenches correlates strongly with the Li/Y ratio in adjacent volcanic arcs, pointing to a direct control on Li enrichment in arcs by subducted sediment. Such a control may be consistent with Li isotope variations in some arcs, with little fractionation required in the slab or mantle. Sediments vary significantly in their Nb anomaly (i.e., Nb depletion with respect to La and Th), and some arcs inherit the sedimentary Nb anomaly (e.g., South Sandwich) with no new fractionation in the subduction zone, while others (e.g., Marianas, Aleutians) likely require residual rutile in the slab. The Nb anomaly is thus a nearly ubiquitous feature in arcs that has a complex origin, dependent on at least three variables: the sedimentary Nb anomaly, the presence of rutile in the slab, and Nb/La variation in the mantle. Trench sediments and GLOSS-II should thus continue to be useful in elucidating processes occurring in subduction zones, the continents and mantle.