Understanding patterns of both large (plate) scale mantle flow and smaller scale buoyant flow in subduction zones is key to models of subduction zone thermal structure, dehydration reactions and volatile distribution, and magma generation and transport. Different patterns of flow in the mantle wedge can generate distinct signatures in seismological and geochemical observables, and ample evidence indicates that a simple two-dimensional, plate-driven corner flow model is inadequate in a significant number of regions. For example, observed shear-wave fast polarization directions in several subduction zones are inconsistent with predictions based on two-dimensional wedge corner flow even when taking into account possible deviations from the standard A-type olivine slip system (e.g. the Marianas, Central America, Tonga, South America, Kamchatka, Alaska). In addition, some arcs and back-arcs contain lavas that exhibit characteristics very similar to nearby hot spot volcanics (e.g. northern Tonga, Costa Rica- Nicaragua), suggesting mantle is entrained into the wedge along highly three-dimensional (3-D) (non-corner flow) trajectories. The combination of recent observations and 3-D modeling suggests subduction-induced wedge circulation has more spatial and temporal complexity than predicted in previous two-dimensional (2-D) modeling. We propose two sets of laboratory experiments to model 3-D subduction zone flow and anisotropy which build upon the PI's previous efforts and which are strongly constrained by seismic data. One set of experiments will investigate how mantle flow is driven by a variety of physical subduction zone parameters related to plate motions and the subducting slab (alongstrike slab dip variations, trench roll-back, slab edges and tears, and upper plate morphology and deformation). In a second set of experiments, The second will consider how wedge material with anomalous viscosity and/or buoyancy interacts with flow and alters predicted anisotropy; sources of such material include hydrated or partially melted mantle from the slab-wedge interface, volatile depleted mantle produced by decompression melting or enriched mantle entrained into the wedge. In both sets of experiments, parameters will be systemically varied to quantify the relative impact of different factors. Flow models will be tested through comparison with seismic anisotropy observed in subduction zones around the globe; this involves transformation of flow fields to crystallographic orientations, and calculation of shearwave splitting. Model development and data-model comparisons will focus on two regions: Nicaragua-Costa Rica and the Marianas. These systems have key differences in plate motions, slab geometry and age, and upper plate deformation, and both are well-sampled by shear-wave splitting, as well as other seismic and geochemical studies. Intellectual merit. This proposal will contribute to an understanding of the physical processes that drive 3D mantle wedge flow and will provide a better context for interpreting the growing number of data sets in subduction systems. Specific questions to be addressed include: 1) What effect does spatial complexity in along-arc slab morphology and sinking mode have on flow in the wedge? 2) Can combinations of these parameters and upper plate shape and deformation produce 3D flows that are consistent with observed anisotropy (e.g., arc-parallel fast directions)? 3) How do altered or chemically distinct regions of mantle wedge interact with 3- D, subduction-induced flow? 4) Does the long-term deformation and entrainment of these features produce LPO patterns that are consistent with observations? 5) How do these patterns change with density/viscosity contrasts between ambient and altered mantle reservoirs and what are the implications for geochemical models of arc magmagenesis? Broader impacts. The proposed work would help to constrain mantle flow and its implications for melting processes in the Nicaragua-Costa Rica and Izu-Bonin-Mariana subduction zones, the two MARGINS Subduction Factory focus areas. This project would provide a Brown graduate student and a URI graduate student with training in laboratory fluids experiments, their integration with seismic observables, and their interpretation. The URI fluids lab will be used as a teaching tool in URI and Brown courses.
