DI41B-4334:
Mid-mantle Seismic Anisotropy in the Southwestern Pacific Subduction Systems
Thursday, 18 December 2014
Anwar Mohiuddin, Yale University, Geology and Geophysics, New Haven, CT, United States and Maureen D Long, Yale University, New Haven, CT, United States
Abstract:
Relatively direct constraints on the pattern of mantle deformation can be obtained from observations of seismic anisotropy. An understanding of the geometry of mantle flow in the transition zone and uppermost lower mantle is particularly important for our understanding of whole mantle dynamics, as the transition zone represents an important internal boundary layer in the mantle. While much of the lower mantle is thought to be isotropic, several recent studies have found evidence for anisotropy in the transition zone and uppermost lower mantle. Here we use the source-side shear wave splitting technique to investigate anisotropy at mid-mantle depths (transition zone and uppermost lower mantle) in the Tonga-Kermadec, Sumatra, New Britain, New Hebrides, and Philippines subduction zones. We measure splitting of direct teleseismic S phases originating from deep events (>350km) that have been corrected for the effect of upper mantle anisotropy beneath the seismic stations. We find evidence for considerable anisotropy at mid-mantle depths in all subduction systems studies, with delay times ranging up to ~3 sec and average delay times of ~1-1.5 sec. Several individual measurements on phases originating at depths greater than 600 km exhibit delay times greater than 1 sec, suggesting a significant contribution from anisotropy in the uppermost lower mantle. We document significant variability in the dominant fast directions both within and among individual subduction systems, with a mix of directions that are parallel, perpendicular, or oblique to the slab strike. This variability suggests differences in the deformation geometry among different subduction systems. As further constraints on the elasticity and deformation of mid-mantle minerals become available, our data set can be used to constrain mid-mantle flow patterns associated with subduction.