Seismic constraints on the hydration of subducting oceanic crust and mantle

Abstract:

It is widely accepted that large amounts of water are delivered to the mantle through subduction. Some water is carried in the hydrated minerals in the oceanic crust, and is released to the mantle through dehydration reactions as the slab becomes warmer. It is also thought that the subducting lithospheric mantle is highly hydrated, and carries large amounts of water to the crust. There are however few observational constraints on the depth to which this water is delivered by the hydrated subducting slab, and the amount of water delivered to the mantle.

Subduction zone guided waves spend longer interacting with the low velocity hydrous mineral assemblages than any other seismic phase, as so give us a unique opportunity to put new constraints on these features of the subduction zone. We use full waveform seismic modelling techniques to constrain these dispersed arrivals, and so image the velocity structure of the slab. This technique gives an observational constraint on both the hydration of the slab, and the onset of dehydration reactions.

We have shown that both low velocity subducted oceanic crust, and low velocity outer rise normal faults that penetrate the lithospheric mantle can act as effective waveguides producing characteristic body wave dispersion. Analysis of the spatial coda decay associated with intermediate depth earthquakes recorded in Northern Japan suggest that these hydrated normal faults may in fact carry the majority of the water transported by the subducting slab to the mantle. Seismic constraints on the dehydration reactions occurring in the subducted oceanic crust also show that full dehydration may occur at much greater depth than predicted by current thermo petrological models, meaning that water may be carried to greater depths than previously thought.

Together these methods suggest that much more water is delivered to the mantle by subduction than has previously been suggested, and that water may be carried to greater depths in the mantle by the hydrated oceanic crust than previously thought. Extending these studies to a range of global subduction zones may allow an estimate of global subduction zone fluid flux from seismic observations to compliment the estimates that have been made from geodynamic and petrological modelling of subduction zones.