Surface-wave studies of the Australian region
The deployment of portable stations over the Australian continent since the beginning of the 1990s has allowed the collection of a unique dataset for surface waves at regional distances. Surface wave tomographic inversions now exploit the excellent azimuthal ray coverage available for central and eastern Australia, so that S wave tomographic models for the upper mantle can be built with a lateral resolution of few hundred kilometres and a vertical resolution of a few tens of kilometres. Our tomographic models include an anisotropic component in addition to the distribution of S-wave heterogeneities. When only Rayleigh waves are considered in the inversion, this anisotropic component is represented at each depth in the mantle by the direction of fast horizontally propagating SV waves. When Love and Rayleigh waves are inverted simultaneously, the anisotropic component reflects the difference in wave speed between S waves polarised in horizontal and vertical directions. Our results for the simultaneous inversion of Love and Rayleigh waves agree with previous studies to locate the anisotropy in the uppermost 200–250 km of the mantle. However, there are significant differences between the different models deduced from the Rayleigh wave inversions. We show that the most likely cause of these differences is the frequency band used in the analysis of the seismograms. The application of path averages directly to shear-wave slowness is a reasonable assumption for the recovered models. The shear-wave speed models are found to be robust, especially when removing paths likely to have experienced a complex propagation. In the light of these new inquiries, we attempt to extract the well-resolved part of the surface-wave inversion and to see how far it can be reconciled with other results obtained from body-waves studies. It appears that, due to the horizontal smoothing imposed by surface waves and the difficulty of estimating the best choice for the frequency band used in analysis, the details of the anisotropic directions in the upper layer should be interpreted with caution. However, the existence of at least two layers of anisotropy is well constrained. In the upper layer, the complex anisotropy would reflect ‘frozen’ deformation in the lithosphere while in the lower layer the smoother pattern is more likely to reflect present-day deformation due to the northward motion of the Australian Plate. The observation that anisotropy is not vertically coherent with depth is more easy to reconcile with other anisotropic measurements inferred from body waves.