Evaluating the Trench Parallel Flow Hypothesis with Numerical Models of Convection

点击翻译按钮获取中文摘要

EAR-0838026: Evaluating the Trench Parallel Flow Hypothesis with Numerical Models of ConvectionScott D. KingWhen some seismic waves (shear waves) propagate through the upper 200 km of the Earth?s interior, they can be resolved into two orthogonal components that travel at different velocities. This information can be used to predict the pattern of flow in this part of the mantle and, is a key observation behind the hypothesis that mantle material flows parallel to the subducting oceanic plate in the vicinity of subduction zones. Using numerical models we will test recent laboratory fluid tank experiments developed to model this flow and the seismic observations from Earth. Our goal is to identify under what conditions trench parallel flow is observed in the calculations and thus place constraints on the strength and chemistry variations in the subducting slab and mantle. This proposal will test the hypothesis that shear wave splitting observations in subduction zones, both above and below the subducting slab, require that the dominant component of mantle flow that is parallel to the strike of the trench and slab. First we will demonstrate that we can reproduce laboratory tank experiments with a three dimensional numerical model. Next we will test the slab parallel flow hypothesis by addressing the following questions: 1) is slab parallel flow dependent on trench migration rate (i.e., whether or not slabs roll back)?; 2) to what extent does the pattern of mantle flow in the vicinity of a subduction zone depend on slab rheology?; 3) are the olivine to wadsleyite (410 km discontinuity) and the ringwoodite to perovskite plus ferropericlase (660 km discontinuity) phase transformations necessary for slab parallel flow?; and finally 4) is an increase in viscosity in the lower mantle by a factor of 10-30 essential to be able to generate slab parallel flow? We will calculate the finite strain from our numerical results to compare with the seismic observations. We anticipate that strong slabs are more likely to lead to slab parallel flow because strong slabs are more able to resist deforming and hence the mantle will be forced to flow around the slab while weaker slabs are more likely to deform in response to mantle flow and flow around a the slab will not be necessary. We also anticipate that the ringwoodite to perovskite plus ferropericlase phase transformation and an increase in viscosity at the top of the lower mantle will both increase the component of slab parallel flow.