Organized melt, seismic anisotropy, and plate boundary lubrication

Based on observations in both the laboratory and the Earth, we develop the hypothesis that plate boundaries are lubricated by networks of melt‐rich shear zones. Such lubrication would serve to reduce effective strength and focus deformation at plates boundaries. This idea emerges from two sets of observations: (1) stress‐driven melt segregation and organization in experimentally deformed mantle rocks and (2) seismic anisotropy patterns as observed at three divergent plate boundaries (the Ethiopian Rift, the Reykjanes Ridge, and the East Pacific Rise). In all three tectonic settings, the magnitude of anisotropy is greatest at the probable locations of the lithosphere‐asthenosphere boundary within the plate boundary (“marginal LAB”). Seismic anisotropy in the upper mantle is controlled by the lattice preferred orientation (LPO) of predominant olivine and the alignment of melt structures. The observed patterns of anisotropy are controlled by the dip angle of the marginal LAB. When steeply dipping, shear wave splitting in vertically traveling waves (e.g., SKS phases) is most sensitive to the alignment of melt, and surface waves should reveal faster Rayleigh wave velocities than Love wave velocities ( V SV > V SH ). When shallowly dipping, shear wave splitting in vertically traveling body waves is controlled by olivine LPO, and surface waves show faster Love wave velocities than Rayleigh wave velocities ( V SV < V SH ). The formation of melt‐rich networks by stress‐driven segregation should be most effective where strain rates are highest. These melt‐lubricated shear zones will reduce effective viscosity relative to the direct extrapolation of viscosity values derived from laboratory creep experiments on homogenous samples. A composite model of anisotropic seismic properties is developed to test the hypothesis that melt segregates along the LAB, incorporating olivine fabrics with oriented and segregated melt over a range of length scales. This model is applied to observations from the three example plate boundaries, leaving the reader to speculate on the implications for interpretation of anisotropy patterns at other geodynamic settings.