Isotropy or weak vertical transverse isotropy in D" beneath the Atlantic Ocean

Shear velocity properties of D″ beneath the central Atlantic Ocean are explored using predominantly European seismic recordings of intermediate and deep focus (>100 km) South American earthquakes. Broadband data are analyzed and, when possible, corrected for upper mantle models of receiver-side anisotropic structure. Regional shear velocity heterogeneity in D″ is mapped by analysis of 306 S-SKS differential times that have been corrected for three-dimensional seismic velocity structure above D″ using a whole mantle tomographic model. This correction yields modest (less than ±2%) estimates of seismic velocity heterogeneity in D″, with a transition from high to low seismic velocities traversing from west to east beneath the central Atlantic, in agreement with global tomographic models. Additionally, shear wave splitting of S and S_diff for the same recordings was analyzed to assess seismic anisotropy in D″ . The highest-quality data provide 105 spliting times between SH and SV onsets tha are mostly within the ±1 s uncertainty level. The few larger values generally exhibit SV delayed relative to SH. Assuming an anisotropy geometry involving vertical transverse isotropy (VTI), as preferred in most regions of D″ that have been studied to date, <0.5% anisotropy strength within a 100 km thick layer, or <0.25% anisotropy within a 300 km thick layer are compatible with the data. These values are low in comparison to those found in high-velocity regions beneath the circum-Pacific Ocean or in the low-velocity region beneath the central Pacific, and many observations are, in fact, consistent with isotropic structure. The lack of strong VTI relative to other regions may be due to (1) the absence of stress from overlying midmantle downwelling, (2) relatively weaker shear flow in the D″ boundary layer, and/or (3) lack of chemical heterogeneity that could develop either lattice-preffered orientation or shape-preffered orientation. The azimuthal sampling of this region of D″ is quite limited; thus the precise geometry and meachanism of any anisotropy are difficult to constrain. It remains possible that this region may contain subtle azimuthal anisotropy that could couple the SV and SH signals; however, amplitude observations suggest that any such coupling is minor.