A new method to measure seismic anisotropy of the upper mantle directly from the width and orientation of the particle motion of shear waves

Splitting of shear waves proves their propagation within an anisotropic medium. Frequently used methods of evaluating of upper mantle anisotropy search for two parameters - the delay time of the slow split shear wave and the polarization direction of the fast split shear wave. The parameters retrieved by the standard methods such as energy minimization on the transverse component of the shear waveforms or eigenvalue of cross-correlation matrix suffer from e.g., ubiquitous noise, errors in sensor orientation and numerous so-called null splits or unrealistically large values. However, well-resolved splitting parameters from core-mantle refracted shear SK(K)S phases are limited to relatively narrow fans of back azimuths. Such incomplete back-azimuth coverage prevents modelling anisotropic structures with symmetry axes oriented generally in 3D, i.e., with tilted axes, to be compatible with 3D anisotropic models from independent observables. Generally used averages of time delays and polarization pairs lead to simplified models of the upper mantle, which concentrate on modelling the present-day flow in the sub-lithospheric mantle. Therefore, we propose a new method directly exploiting variations in width and orientation of particle motion (PM) of split shear waves, which allows measuring anisotropic characteristics for a larger amount of waveforms and improves azimuthal coverage in a region. We characterize the PM by two parameters, the PM width and the PM orientation. At each station, we plot the normalized width of the PM as a ratio of lengths of the minor to major axes in dependence on back-azimuths. Variations of the PM width with back-azimuth exhibit oscillations with several extremes of different amplitudes. Such behaviour results from wave propagation through the anisotropic upper mantle. One of the advantages of the method is that the width of the PM is invariant of potential mis-orientation of sensors. We test the PM method on a set of SKS waveforms recorded at a subset of stations included in several recent or running passive seismic experiments (EASI, AlpArray, PACASE, AdriaArray). The stations form a band of about 200km broad running from the western Bohemian Massif through the Eastern Alps to the Adriatic Sea. Stations characterized by similar variations of the PM parameters group into sub-regions, which are compatible with the main tectonic features of the whole region. The formation of such lithospheric blocks of similar anisotropic signals is in agreement with 3D self-compatible anisotropic models of the mantle lithosphere domains derived from independent observables. We present complementary studies of the anisotropic structure of the mantle lithosphere in contributions by Zlebcikova et al. (GD7.1, EGU 2024), which shows anisotropic model of the upper mantle derived from 3D coupled anisotropic-isotropic teleseismic tomography (code anitomo), and in contribution by Kvapil et al. (GD7.1, EGU 2024), in which anisotropic structure of the lower crust is modelled from ambient noise.