Toroidal mode observations and their interpretation in terms of anisotropy and density

Knowledge on variations in seismic anisotropy and density in the Earth's mantle will provide constraints on Earth’s dynamic properties, which helps to understand Earth’s mantle flow and deformation. Whole Earth oscillations, which are standing waves along Earth’s surface and radius, are an ideal method to investigate 3D variations in seismic anisotropy and density in the deep mantle. These oscillations, also called normal modes, are excited after strong earthquakes and provide important constraints on Earth’s large scale structures and their sensitivity reaches from crust to core. Normal modes can be divided into (i) toroidal modes, dominated by horizontal surface motion (SH), and (ii) spheroidal modes, involving a combination of horizontal and vertical surface motion (P-SV). Toroidal modes, which are similar to Love surface waves, are our main interest here. In combination with spheroidal modes, they provide important large scale constraints on the anisotropy and density structure of Earth's mantle. In this thesis we measure normal mode splitting functions of mainly toroidal (and some spheroidal) modes with the aim to interpret these in terms of 3D variations in mantle anisotropy and density. Studying toroidal and spheroidal modes and their cross-coupling will enable us to recover radial and azimuthal anisotropy for P- and S-velocity simultaneously. We fill a 20 year gap of toroidal mode splitting functions measurements and expand more recent spheroidal mode studies by focussing specifically on toroidal mode observations. Toroidal modes are mainly visible on the horizontal component of the seismogram, so we added horizontal component data for all large earthquakes from the last 35 years. Using these new horizontal data recordings, we were able to refine and extend isolated self-coupling splitting function measurements for toroidal modes compared to earlier studies. Toroidal mode energy may also become visible on the vertical component instead of only the horizontal components due to cross-coupling or resonance (i.e. exchange of energy) between fundamental toroidal and spheroidal modes. The effect of rotation of the Earth on toroidal-spheroidal cross-coupling is well known. Here, we also investigated the occurrence of additional cross-coupling, which may be due to radial and azimuthal anisotropy and thus provide important information on anisotropic structure of Earth’s mantle. We confirm previous anomalously strong cross-coupling observations, extending the earlier observations from degree s=2 to s=4. Furthermore, we investigated the influence of toroidal-spheroidal mode cross-coupling on the measurement of core-mantle boundary sensitive Stoneley modes and their interpretation in terms of lowermost mantle density models. Our new Stoneley mode measurements confirm previous interpretations of lighter LLSVP's on top of an elevated CMB.