Sub-slab mantle anisotropy beneath south-central Chile

Complex mantle flow exists in subduction zones; one way to reveal these is through the analysis of shear-wave splitting. In particular, asthenospheric flow beneath the subducting slab remains poorly understood. There have been numerous studies of shear-wave splitting along the Andean margin, but a gap in our knowledge remains in south-central Chile. The TIPTEQ network was deployed along the south central Chilean forearc during 2004-2005; the density of this network and its proximity to the forearc makes it useful for studying anisotropic sources other than the supra-slab mantle wedge. Here, we present results from the analysis of splitting in both teleseismic SKS phases and local S waves emanating from the subducting Nazca lithosphere, at 120 short-period stations. The magnitude of local splitting is small (~0.21 s) and is shown to be consistent with a fault-induced source in the upper crust. This is consistent with the inferred lack of supra-slab asthenospheric wedge, based on local seismic tomography models. The average teleseismic splitting delay time is ~1.2s . Fast directions are variable and show a strong dependency on the ray's back-azimuth. Forward modelling shows that this back-azimuthal dependency can be caused by a 100km thick anisotropic layer beneath the slab, with a dipping symmetry axis; this is consistent with a model of sub-slab entrained asthenosphere. We discuss the preservation potential of seismic anisotropy beneath the slab, whether these observations are unique across global subduction zones and whether there is a relationship with the slab's age.