SMALL-SCALE SUBLITHOSPHERIC CONTINENTAL MANTLE DEFORMATION - CONSTRAINTS FROM SKS SPLITTING OBSERVATIONS

It is generally believed that azimuthal seismic anisotropy is an important property of the subcontinental mantle, and that this is a relatively small-scale phenomenon compared with anisotropy of the oceanic mantle. During recent years, shear-wave splitting has come to be viewed as the most direct diagnostic of azimuthal anisotropy. The magnitude of the splitting delay times can often rule out a crustal source of anisotropy; however, it is not yet clear whether the source of anisotropy lies within the lithosphere or the sublithosphere. It is essential to know the scale of lateral variations and the distribution of anisotropy with depth in order to understand the origin of this phenomenon. Here we describe a way to constrain the depth of the anisotropy, by studying the lateral variation of the splitting parameters at neighbouring seismic stations. We use SKS and SKKS phases recorded at the NARS-NL array, a relatively dense network of broad-band stations separated by about 50 km. Fresnel zones at different depths are calculated for these phases. The depth of the anisotropy is constrained by the criterion that Fresnel zones corresponding to different splitting observations should not overlap. Variability in the splitting measurements for events with different directions of approach recorded at one station, and for single events recorded at various stations, provides evidence for a non-homogeneously distributed source of anisotropy, located below the array at different depths. We show that this variability in the measurements cannot be due to effects of crustal scattering. Our results indicate that the anisotropy is partly located in the sublithospheric mantle at about 400 km depth. This suggests that the deformation in this region is subject to significant lateral variations.