As earthquake waves probe the Earth at all depths, seismology is a very powerful tool for providing insights into the structure and evolution of our planet. Elastic or seismic anisotropy, the dependence of the propagation velocity of seismic (elastic) waves on the propagation or polarization direction, is a widely used phenomenon to gain crucial insights into past and present geological and geodynamic processes within the Earth’s interior. These processes cause deformations within the materials and rocks, which lead to a preferred orientation of structures and crystals, and by this to (bulk) anisotropy. Shear wave splitting (SWS; analogous to optical birefringence of light) is a unique indication for anisotropy. The two splitting parameters, the fast polarization direction of the anisotropic medium and the delay time the two split shear waves have accumulated after traveling through the anisotropic medium, are commonly used to quantify the anisotropy.
In addition to anisotropy related to large-scale processes such as absolute plate motion, subduction, continental rifting, and orogeny, anisotropy can vary on regional and local length scales. ... mehrEspecially for continental upper mantle anisotropy, small-scale complex 3-D anisotropy scenarios are suspected.
The Upper Rhine Graben (URG) area in Central Europe is characterized by a complex geological and geodynamic setting, and previous anisotropy studies propose indications for small-scale variations of anisotropy in the upper mantle. However, no 3-D anisotropy model is available for this area yet. According to this, the main goal of this doctoral thesis is to resolve the anisotropy in the upper mantle underneath the URG area in 3-D.
The Black Forest Observatory (BFO) is one of the most reliable operating and quietest recording stations worldwide and provides freely available seismological data of more than 25 yr. A quite surprising difference between the SWS measurements in the northeast and southwest is clearly indicating lateral variation: SWS with a characteristic variation of the fast polarization direction indicating a two-layer scenario in the northeast, but numerous nulls over a backazimuth range of 100° in the southwest (referred to as null anomaly).
To understand this striking observation, seismological data from five (semi-)permanent recording stations (10-25 yr) surrounding the BFO were analyzed uniformly. SWS of core-refracted phases PKS, SKS, and SKKS are manually measured with the widely used MATLAB package SplitLab (single-event analysis) with the plugin-in StackSplit (multi-event analysis).
The correct relative temporal alignment of the vertical (Z), North (N), and East (E) component traces is essential to achieve correct results (not only in terms of shear wave splitting). During the SWS analysis, an error source was found within SplitLab, leading to a wrong temporal alignment and, by this, to wrong shear wave splitting measurements. This bug was fixed and affected measurements were redone.
Complex anisotropy scenarios, i.e., vertical and lateral variations as well as a tilted symmetry axis, lead to backazimuthal variations of the splitting parameters. Many anisotropy studies suffer from poor directional or backazimuthal coverage due to the analysis of short-term data (2-3 yr) which leads to a small number of recorded earthquakes and, by this, a small number of shear wave splitting measurements. Additionally, often averaging over wide directional (backazimuthal) ranges is done. In case of less observations or / and averaging the splitting parameters, these variations are missed or ruled out, leading to oversimplified or wrong anisotropy models.
For the URG area, due to the analysis of long-term data, complicated splitting patterns could be observed. Due to missing differences between the observations on SKS and SKKS phases, there are no direct indications for a lowermost mantle contribution to the anisotropy, and based on a delay time longer than 0.5 s the anisotropy is assumed to be located in the upper mantle, including both the lithosphere and asthenosphere. To explain the splitting observations, structural anisotropy models with two layers and a tilted symmetry axis are tested via the comparison of forward calculated synthetic splitting parameters with the observed ones. Subregions of the study area are outlined, based on similarities between the splitting patterns in the stereoplot representations and overlapping piercing points calculated at different depths in the upper mantle.
For the anisotropy in the upper mantle underneath the URG area a 3-D block model is proposed: (i) two layers in the Moldanubian Zone (south) but one layer in the Saxothuringian Zone (north), (ii) different fast polarization directions on the east and west sides of the URG in the Moldanubian Zone, and (iii) indications for a tilted symmetry axis close to the Lalaye-Lubine-Baden-Baden fault. For the null anomaly in the southwest of BFO, explanations besides the trivial solution of isotropy could be (i) orthogonal symmetry axes of the two layers canceling out the splitting effect, (ii) a vertical symmetry axis due to a C-type olivine fabric in case of high temperature and water content, (iii) vertical symmetry axis due to vertical mantle flow, and (iv) wave scattering due to heterogeneities. Possible is a link to the Kaiserstuhl Volcanic Complex due to a modification of the mantle during the Middle Miocene melting processes, including fluid-rock interaction. Magmatic processes may have produced structural modifications, including large intrusions resulting in a different olivine CPO or wave scattering. A tilted symmetry axis could be related to the terrane boundary being a remnant of the deformation during the Variscan subduction.
The anisotropy studies at BFO and in the URG area demonstrate that continental upper mantle anisotropy can be quite complex, which cannot be explained by large-scale processes such as simple asthenospheric flow or vertical coherent deformation. Analyzing long-term data allows us to observe both interstation and intrastation lateral variations. Especially such complicated splitting signals from the upper mantle can contaminate splitting signals originating from the LMM due to the integral nature of the measurements.
The variations of the observed splitting parameters and the modeled anisotropy occur on a significantly smaller length scale than expected from the size of the first Fresnel zone. However, XKS phases have a dominant period of 8-10 s. Thus, ray theory is only limitedly valid, and wave effects could become relevant. In a follow-up study, simulations with AxiSEM3D can be used to test the proposed or observation-based anisotropy model and considering arbitrary anisotropy and heterogeneities.
1