Upper mantle stratification by P and S receiver functions

Seismic stratification of the upper mantle is investigated by applying two complementary techniques to the records of the Graefenberg array in southern Germany. The anisotropic P receiver function technique (Kosarev et al. 1984; Vinnik & Montagner 1996) is modified by using summary seismic events instead of individual events and different weighting functions instead of the same function for the harmonic angular analysis of the SV and T components of the Pds phases. The summary events provide better separation of the second azimuthal harmonic than the individual events. The parameters of the second harmonics of SV and T thus evaluated should be similar if they reflect the effects of azimuthal anisotropy. This can be used as a criterion to identify the anisotropy. To detect the Sdp phases and their azimuthal variations caused by azimuthal anisotropy we have developed a stacking technique, which can be termed the S receiver function technique. It includes axis rotation to separate interfering P and S arrivals, determination of the principal (M) component of the S-wave motion, deconvolution of the P components of many recordings by their respective M components and stacking of the deconvolved P components with weights depending on the level of noise and the angle between the M direction and the backazimuth of the event. Both techniques yield consistent results for the Graefenberg array. As indicated by the P receiver functions, the upper layer of the mantle between the Moho and 80 km depth is anisotropic with dV(s)/V-s around 0.03 and the fast direction close to 20 degrees clockwise from north. The fast direction of anisotropy below this layer is around 110 degrees. The boundary between the upper and the lower anisotropic layers is manifested by the detectable Pds and Sdp converted phases. Shear wave splitting in SKS is strongly dominated by azimuthal anisotropy in the lower layer (asthenosphere).