Seismic constraints on mantle hydration during subduction: Mariana versus Middle America Trench

We will present new results from active-source seismic experiments that constrain the amount of seawater entering the upper mantle along bending-induced faults at the outer rise of the Mariana and Middle America Trenches. This seawater may fill cracks in the upper mantle with free water; react strongly with olivine in upper mantle peridotite, filling cracks and fault zones with the hydrous mineral serpentinite; and/or diffuse between fault zones, pervasively serpentinizing the upper mantle. The upper mantle accounts for a large portion of the subducting lithosphere, and a hydrated upper mantle may supply the majority of water fluxing into arcs. Serpentinite is not stable at high temperatures and pressures, and, once subducted, would undergo a reverse reaction, releasing water into the mantle at depth and driving many arc- and global-scale geochemical and geodynamic processes. Hydration of the subducting upper mantle by fluid flow along bending-induced faults should depend on the density and depth of this faulting, as well as on the plate convergence rate and upper mantle temperature, which together control the rate of serpentinization reactions. Thus, a test of the outer-rise hydration hypothesis is to compare observations of the distribution of serpentinization in the upper mantle from subduction zones with different plate ages (temperatures); convergence rates; and angles between the relic abyssal-hill fabric, plate motion direction, and the trench (i.e. subduction obliquity), which control patterns of faulting. We will compare new seismic observations of faulting and serpentinization at the outer rise of the Middle America Trench offshore Nicaragua to observations from the central Mariana Trench. At the Middle America Trench, a moderately aged slab (24 Ma), and thus hot upper mantle, is subducting rapidly (85 mm/yr), both conditions that can limit serpentinization. Offshore Nicaragua, bending reactivates relic abyssal-hill fabric, which is oriented parallel to the trench. Here, measurements of seismic anisotropy and slow absolute wavespeeds suggest that these faults penetrate into the upper-most mantle and supply seawater that serpentinizes the mantle by up to ~13% (~1 wt% water). At the Mariana Trench, the slab is much older (140 Ma), and thus colder, and the convergence rate is slower (41 mm/yr), conditions expected to promote serpentinization. Here, preliminary analysis of new data suggests that upper mantle velocities are also significantly reduced under the outer rise. This velocity reduction is most extreme where bending-induced faulting is most pronounced, consistent with serpentinization via fluid flow along faults, although Cretaceous-age off-axis magmatism and the faults themselves may also affect seismic wavespeed. The goal of in-progress work on these Mariana data is to separate these wavespeed effects, enabling an estimate of serpentinization that can be compared to results from a similar analysis of the Middle America data, advancing our understanding of processes controlling the water input to subduction zones in general.