The nature of upper-mantle seismic anisotropy beneath central and eastern North America has been evaluated using the phase velocities of surface waves traversing the length of the Missouri-to-Massachusetts (MOMA) broad-band array. Frequency-dependent phase delays of fundamental-mode Love and Rayleigh waves, measured across the array using a cross-correlation procedure, require the presence of anisotropy in the upper mantle. 2-D radially anisotropic structures obtained via linearized inversion of these data contain shear anisotropy with a magnitude of similar to3 per cent between the Moho and at least 180-km depth. Models in which the mantle lithosphere is isotropic cannot satisfy the data. This anisotropy is approximately constant across the tectonic transition from the craton to the Atlantic margin. Forward models consisting of a shallow, lithospheric layer of azimuthal anisotropy derived from previous shear wave splitting observations fail to satisfy the surface wave delay times. The combined surface wave and splitting results suggest the presence of two layers of anisotropy: a lithospheric layer that produces the phase-delay differences observed in Love and Rayleigh waves but is transparent to vertically propagating shear body waves; and a deeper (presumably asthenospheric) layer that generates the shear wave splitting. One plausible model is that the lithosphere is characterized by vertically heterogeneous anisotropic fabric, which would produce minimal splitting in vertical shear waves. Such fabric has been hypothesized for regions of weak and complex splitting such as South Africa and Australia; the results here imply that it may be appropriate for North America as well.
848LYTimes Cited:12Cited References Count:62