Seismic Velocity Anisotropy In D" Layer Of Mantle Beneath The Western Pacific

D" layer, the lowermost few hundred kilometers of the mantle, serves as a dynamical, thermal and chemical boundary layer between the solid mantle and the liquid outer core. Heat, angular momentum, and possible some materials are exchanged across the core-mantle boundary (CMB) and this layer is postulated to influence mantle convection, the plate motion and the earth's magnetic field. Seismological studies indicate the presence of anisotropy in D" layer. Seismic anisotropy is an important tool in understanding dynamic processes in the Earth and carries potential information on the evolution and dynamics of the solid Earth. In the favorable conditions it can provide a seismic signature to mantle flow invisible to other methods such as tomography. Therefore the location, orientation, and magnitude of seismic anisotropy in D" layer is helpful to constrain how chemical and melt heterogeneity or anisotropic minerals are oriented by patterns of flow near CMB. The western Pacific is one of the active and complex tectonic regions presently in the Earth and may have extremely complicated structure in the deep mantle. Therefore study of the D" layer beneath the western Pacific is critical to understand the deep structure and dynamic process of this region.At first, Using seismic shear phases from 146 Tonga-Fiji and its adjacent region events during 1994 and 2007 recorded by 36 stations of the Incorporated Research Institute for Seismology (IRIS) broadband arrays, and from 26 northeast Asia and north Pacific events during 2000 and 2005 recorded by 18 stations of IRIS, we studied the shear wave anisotropy in D" layer beneath the western Pacific utilizing the ScS-S differential travel time method. We obtained 512 splitting time values between the radial and transverse components of ScS wave and calculated the anisotropy strength. The pattern of the shear wave splitting is different between the seismic waves propagating in the direction of N-S and that in the direction of NW-SE: The former mainly involve major V_ScSH> V_ScSV (V_ScSH is velocity of horizontally polarized ScS wave, V_ScSV is velocity of vertically polarized ScS wave) splitting value, the latter mainly contain the V_ScSH < V_ScSV value, and the anisotropy strength of the latter is obviously larger than that of the former. On the whole, the large majority of shear waves involve the pattern of V_ScSH < V_ScSV value. The shear wave splitting with V_ScSH < V_ScSV is focused on the central and eastern part of the study area, while the shear wave splitting with V_ScSH < V_ScSV is mainly distributed in the southwestern part of the study area. The splitting time values of ScS wave range from -4.08s to 4.53s with an average absolute value of 1.26s. The strength of anisotropy varies from -1.96% to 2.55% with an average absolute value of 0.61%. The distribution trend of the splitting time values and the anisotropy strength are consistent.Using the ScS-S differential time method can remove the effect of upper mantle anisotropy conveniently and reveal the anisotropy in D" layer preferably. But the splitting observation measure is limited on the radial and the transverse components, thus attempting to examine transverse isotropy with a vertical axis of symmetry (VTI). A more complex method must be used to resolve more general forms of anisotropy. Therefore, using seismic shear phases from 15 Tonga-Fiji and its adjacent region events recorded by 8 stations of IRIS, we obtained the splitting parameters (i.e. the polarization direction of the fast wave and the time delay between the separated fast and slow waves) of the ScS phase in D" layer beneath the western Pacific utilizing the rotating-correlation method and compared the result with that of ScS-S differential time method. The polarization direction (given by azimuth from north) of the fast wave vary greatly from 4°to 170°(note: it has n times 180 degree ambiguity). The angles between the fast direction and the radial direction range from 7.48°to 73.86°, and most of the angles are less than 45°, suggesting that there is a pattern of V_ScSV>V_ScSH in most of the study area. The delay time is 0.2s-3.9s with an average value of 2.2s and the strength of anisotropy is 0.11%-2.01% with an average value of 1.06%. The estimation about the relative size of V_ScSV and V_ScSH from the two methods has 75% similarity, but their delay time and anisotropy strength has some difference. Due to the uncertain of the upper mantle anisotropy correction and the computing errors of the correlation value from the selection of the waveform windows and the shift of time, there may be larger errors in the anisotropy result using the rotating-correlation method. But it can examine the polarization direction of fast wave at a wider variety of azimuths.Seismic anisotropy may reveal the dynamical process of the Earth interior, by which the dominant deforming mechanism and rheology property of the different layer of the mantle can be inferred. But it is very complex to interpret the seismic anisotropy results and there are often contrary explanations. Due to geographical limitations in the distribution of earthquake sources and seismic sensors, at present, none of the deep mantle anisotropy studies has significant azimuthal raypath sampling. Limited azimuth coverage makes it difficult to distinguish one anisotropy pattern from another. Therefore assumptions are necessary to proceed from measurements to interpretations. Based on the observation and analyse of the shear wave splitting, we inferred the possible rheological field mechanism for the D" layer beneath the western Pacific. In this area, the vertical upwelling flow is expected to be dominant. The horizontal flow structures may exist but the magnitude may be relatively small and mainly located at the southwestern part of the study area. There may be vertical fabrics formed by the aligned heterogeneous materials resulting from the ascending flow. Lattice preferred orientation (LPO) of the lower mantle minerals in this region is a possible mechanism for the observed anisotropy too. Additionally, flow out of horizontal plane may also exist and align anisotropic crystals or heterogeneous materials to form azimuthal anisotropy which can be explained as transverse isotropy with a tilted axis of symmetry.(Mg, Fe) SiO₃ post-perovskite may be the main mineral phase in D" layer and is remarkably anisotropic. It can provide a new approach for explain the seismic observations such as the seismic anisotropy in D" layer. Therefore, the slip systems of those phases and resultant LPO are important for understanding the observed seismic anisotropy. But the nature of the dominant slip system for post-perovskite phase has yet to be clarified. Considering the complex stress environment of D" layer and properties of minerals associated with the stress field and the orientation of crystallographic axes, single-crystal energy in triaxial stress field was calculated to study elastic constants, single crystal and aggregated acoustic velocities. The conditions were found under which energy is a minimum. The calculations show that the orientation of crystallographic axes and the differential stress significantly affect on the properties of the post-perovskite mineral. The a-axis tend to align paralleling to the maximal compression direction, and transverse anisotropy in shear wave velocity with the b-axis as vertical symmetry axis is larger than that under the condition in hydrostatic stress field. The crystal orientation type with minimum energy becomes more stable than the other orientation type with the differential stress increasing. The results also support (010) plane as the dominant slip plane. There is anisotropy with V_SH< V_SV in the upwelling region of the D" layer when [010] axis orients horizontally. While in the horizontal flow region, vertically oriented [001] and [010] axes both result in anisotropy with V_SH> V_SV.