Basic Geophysics Shear Wave Splitting

The earth consists of five major concentric layers: The crust, the upper and lower mantle and the outer and inner core. They differ in their elastic properties as for example the seismic velocities of P- and S-waves. Since the beginning of the 20th century, seismologists have used traveltime curves of earthquakes worldwide to refine the global model of the earth. Harold Jeffreys and his PhD student Keith Edward Bullen were two of them. Jeffreys observed in 1939 that P- and S-wave velocities decrease at the very bottom of the lower mantle. After analysing further data, Bullen found differences also in the pressure and compressibility gradients of this special layer, he called D’’. Hence, he noted in 1946, that »D’’ is significantly different in some respects from D’« - so he called the rest of the lower mantle. Today, the global availability of large amounts of seismic data allow seismologists to investigate the nature of D’’. They study it using a method called shear-wave splitting. Hello and welcome. In this video I introduce you to the concept of shear wave splitting and how the D’’ layer can be studied using those observations. I shall show you along which paths the shear waves SKS and SKKS travel through the Earth and how they can be used to analyse a special property of D’’: anisotropy. You will learn what anisotropy is and how a seismogram must be processed to observe shear-wave splitting. Seismologist Michael Grund from KIT will explain this in an interview and how he and his colleagues found anisotropic structures in the D’’ layer. He used data from a measurement campaign in Scandinavia to analyse the deep underground beneath northern Europe. Earthquakes radiate seismic waves that travel through the Earth and they are named corresponding to their way through different layers. The wave most commonly used to study shear-wave splitting is the SKS phase. It travels from the earthquake source down to the core-mantle boundary – short CMB – as a shear- or S-wave. It enters the earth's liquid outer core and is converted to a compressional P-wave. This is denoted with a K for the German “Kern”, which means “Core”. When it reaches the CMB on the receiver-side again, parts of the P-wave are converted back into an S-wave. Since P-waves only oscillate in longitudinal direction, the back converted S-waves now are polarised radially within the plane of the travel path. They are now useful to study seismic anisotropy, which is the directional dependence of seismic velocities. When the polarised S-wave enters an anisotropic medium under nearly vertical incidence, the wave splits up into two new S-waves as you see in the animation. One is polarised in the direction of the fast axis of the medium, the other perpendicular to it in the direction of the slow axis. Those waves can be identified in seismograms and analysed to study the anisotropic media they travel through. For this purpose, Michael Grund and his colleagues set up a temporal broadband seismometer network project in Finland and Sweden, that was combined with other stations distributed over the whole of Scandinavia. He will explain how these data were used to analyse shear-wave splitting and anisotropy in the D’’ layer. BARTH: Hello Michael! GRUND: Hi! BARTH: The ScanArray network consists of over 250 seismic stations. What is the goal of installing and running such a huge network? GRUND: The primary goal of ScanArray is to study the structure of the crust and mantle directly beneath Fennoscandia using different seismological methods. GRUND: On the map you see all used seismometers as coloured triangles. Due to the small distances between the individual stations it is possible to map potential structural variations with a high lateral resolution. This may help us to receive new constraints on the tectonic evolution of Fennoscandia. However, the recorded data also allows us to study much deeper structures far away from the network like the D” just atop the CMB. BARTH: The D’’ layer is meant to be anisotropic. What are the reasons for anisotropy in general? GRUND: Within the earth, anisotropy is mainly caused by two primary sources. First, the layering of media with different elastic properties and second the preferred alignment of anisotropic minerals on a larger scale. BARTH: This anisotropy can be studied using polarised shear waves. What kind of data did you analyse especially? GRUND: We used core-refracted shear waves like the SKS and SKKS phase. GRUND: Compared to SKS the SKKS phase has one more underside-reflection at the CMB halfway between the source and the receiver, as you can see in the figure. BARTH: And how did you process the observed SKS and SKKS waveform data to analyse anisotropy? GRUND: In general SKS and SKKS waves have dominant periods between 8 and 10 s. Therefore, we applied a bandpass filter between 5 and 15 s to remove frequencies of no interest. In a next step we looked at the waveforms - the radial and the transverse component. At this point, then we have to distinguish between two different cases of observations: GRUND: In the first case, we have a clear signal on the radial R-component but almost no signal on the transverse T. This is a so-called Null measurement and indicates that the recorded waveform was not split. Reasons for such observations are that the wave travelled only through isotropic material or that the polarization of the wave was parallel or perpendicular to the fast axis of the anisotropic medium. In the second case we observe clear arrivals on both, radial and transverse component. This indicates that the wave was split by an anisotropic medium. For such observations we now want to determine the parameters that best describe the anisotropic material. These are the time difference between fast and slow axes of the splitted shear-waves and their orientation in space. To find those parameters, we rotate the seismograms on the radial and transverse component and shift them in time systematically using a grid-search to minimise the energy on the transverse component. In this case this happens for the fast axis angle of phi=44° relative to North and a delay time of 1 s between the two split waves. BARTH: OK, this was just one example. Did the other data show similar parameters? GRUND: If possible we measured the splitting for the SKS and SKKS in phase in each seismogram. For several cases we observed clear discrepancies between the two phases, namely that one phase was clearly split and the other phase offered a clear null measurement. BARTH: How did you conclude that anisotropy is located at D’’ and not elsewhere on the travel path? GRUND: The most important aspect of the SKS and SKKS phases that they have very similar raypaths in volumes directly beneath the station, in contrast their raypaths in the lowermost mantle in D” differs significantly. GRUND: For a shallow source we would assume that the splitting results for the SKS and SKKS phases are nearly identical. However, for the lowermost mantle and the D” SKS and SKKS lie thousands of kilometres apart from each other. Additionally, other experiments suggest that the overlying mantle is nearly isotropic. Therefore the differences between SKS and SKKS are strong indicators that the anisotropic source responsible for the splitting is located in the D”. BARTH: And is the hypothesis of an anisotropic D’’ supported by other studies? GRUND: Yes, we compared our results with the horizontal tomographic slice computed for 2700 km depth by Simmons and colleagues. GRUND: In this depth we can identify different large-scale structures. In the West low velocity anomalies dominate the picture. They are related to the northern extensions of a large low-velocity structure beneath Africa and a hot region further North, which potentially is responsible for the mantle plume beneath Iceland. Our results are shown here as pierce points in red and orange for the SKS and SKKS phases, The black lines are the fast axes, which are located exactly in a transition area from fast to slow seismic velocities. One potential explanation is anisotropy due to mantle flow in those environments. In the eastern area a large high-velocity anomaly is the most prominent feature. Other studies inferred that remnants of subducted slab material arrive here at the CMB beneath Siberia. Our splitting measurements may be a result of strain-induced anisotropy atop the CMB. BARTH: These are exciting insights into the deep earth. What are you planning next? GRUND: An important step is to compare the measurements with geodynamic models for which our results can act as a valuable boundary condition. Furthermore, other seismic phases should be studied that sample the D" from other directions this can be helpful to improve the geometry of these anomalies. BARTH: Good luck and thanks! GRUND: Thank you! Like Keith Bullen and Harold Jeffreys gathered data to analyse the earth’s structure over 70 years ago, seismologists like Michel Grund use large amounts of seismic data to refine our picture of lower mantle deep down in the earth. In this video, you learned how shear wave splitting is used to analyse the deep earth mantle, especially D’’. You learned that SKS and SKKS phases are radially polarised when they leave the outer core. Thus, they can be used for splitting analyses. Michael Grund explained how seismic data of SKS and SKKS phases were used in the ScanArray project to study the anisotropy atop the Core-Mantle-Boundary. Their findings may be an indication for mantle flow beneath Western Europe and an old remnant lithospheric slab under Sibiria.