Sascha Zertani, Johannes C. Vrijmoed, Frederik Tilmann, Timm John, Torgeir B. Andersen, Loic Labrousse

Seismological studies of large-scale processes at convergent plate boundaries typically probe lower crustal structures with wavelengths of several kilometers, whereas field-based studies typically sample the resulting structures at a much smaller scale. To bridge this gap between scales, we derive effective petrophysical properties on the 20-m, 100-m, and kilometer scales based on numerical modelling with the Finite Element Method. Geometries representative of eclogitization of crustal material are extracted from the partially eclogitized exposures on Holsnøy (Norway). We find that the P wave velocity is controlled by the properties of the lithologies rather than their geometric arrangement. P wave anisotropy, however, is dependent on the fabric orientation of the associated rocks, as fabric variations cause changes in the orientation of the initial anisotropy. As a result, different structural associations can result in effective anisotropies ranging from ~0-4% for eclogites not associated with ductile deformation to up to 8% for those formed during ductile deformation. For the kilometer-scale structures, a scale that in principle can be resolved by seismological studies, we obtained P wave velocities between 7.7 and 8.0 km s-1. The effective P wave anisotropy on the kilometer-scale is ~3-4% and thus may explain the backazimuthal dependence of seismological images of, for example, the Indian lower crust currently underthrusting beneath the Himalaya. These results imply that seismic anisotropy could be the key to visualize structures in active subduction and collision zones that are currently invisible to geophysical methods and thus can be used to unravel the underlying processes active at depth.