Shear-wave splitting beneath Fennoscandia - evidence for dipping structures and laterally varying multilayer anisotropy

The geodynamic evolution of Fennoscandia in northern Europe (Finland, Sweden and Norway) is coined by ca. 3Ga history of tectonic processes including continental growth in its central and eastern parts and Neogene uplift processes of the Scandinavian mountains (Scandes) located along its western edge. Many details are still under debate and we contribute with new findings from studying deep-seated seismic anisotropy. Using teleseismic waveforms of more than 260 recording stations (long-running permanent networks, previous temporary experiments and newly installed temporary stations) in the framework of the ScanArray experiment, we present the most comprehensive study to date on seismic anisotropy across Fennoscandia. The results are based on single and multi-event shear-wave splitting analysis of core refracted shear waves (SKS, SKKS, PKS and sSKS). The splitting measurements indicate partly complex, laterally varying multilayer anisotropy for individual areas. Consistent measurements at permanent and temporary recording stations over several years and for seismic events of specific source regions allow us to robustly constrain dipping anisotropic structures by adding systematic forward modelling. Although the data coverage is partly limited to only few source regions, our findings support concepts of continental growth due to individual episodes of (paleo-) subduction, each affecting a plunging of the anisotropic fast axis direction due to collisional deformation. Along the northern Scandes the fast axis direction (phi) is parallel to the mountain range (NE-SW), whereas an NNW-SSE trend dominates across the southern Scandes. In the south, across the Sorgenfrei-Tornquist Zone, a NW-SE trend of phi dominates which is parallel to this suture zone. The Oslo Graben is characterized by an NNE-SSW trend of phi. In northern Norway and Sweden (mainly Paleoproterozoic lithosphere), a dipping anisotropy with phi towards NE prevails. This stands in contrast to the Archean domain in the NE of our study region where phi is consistently oriented NNE-SSW. In the Finnish part of the Svecofennian domain, a complex two-layer anisotropy pattern is found which may be due to lateral variations around the seismic stations and which requires a higher data density than ours for a unique model building. Based on these findings our study demonstrates the importance of long recording periods (in the best case > 10 yr) to obtain a sufficient data coverage at seismic stations, especially to perform meaningful structural modelling based on shear-wave splitting observations.