Importance of lateral viscosity variations in the whole mantle for modelling of the dynamic geoid and surface velocities

We investigate the effect of lateral viscosity variations (LVV) in the mantle on dynamic geoid and near-surface mantle velocities. Contrary to the previous studies, we analyse the effect not only of the lithospheric keels and asthenosphere but also of the whole mantle. A 3D global viscosity model is constructed using (a) the S20 seismic tomography model converted to temperature and (b) assumptions about homologous temperature in the mantle Paulson et al. [Paulson, A., Zhong, S., Wahr, J., 2005. Modeling post-glacial rebound with lateral viscosity variations. Geophys. J. Int. 163, 357–371]. The conversion parameters provide a depth-dependent viscosity profile, which is generally consistent with the existing studies. Therefore, the horizontal variations are produced self-consistently within this approach. We estimate them as ‘conservative’, thus providing a minimum limit for possible horizontal changes; consequently the obtained results show a low limit for dynamic topography, geoid and surface velocities’ disturbances. To estimate the effect of LVV we have employed the perturbation method of Zhang and Christensen [Zhang, S., Christensen U., 1993. Some effects of lateral viscosity variations on geoid and surface velocities induced by density anomalies in the mantle. Geophys. J. Int. 114, 547–551], which is modified to implement mantle compressibility. It has been found that the impacts of the upper and lower mantle are comparable in amplitude. The difference between the initial (only radial viscosity) dynamic geoid and the geoid with implemented LVV reaches −27 to +19 m for the whole mantle LVV, while the effects of the upper and lower mantle are equal to −24 to +16 and −20 to +16 m correspondingly. In general these effects are not correlated. The difference in the geoid response leads to substantially altered velocity-to-density scaling factors (up to 30% compared to the initial ones) obtained in least squares inversion. The effect of the lower mantle LVV on surface velocities is also substantial, in particular with respect to the toroidal component, which does not exist in the initial model.