Sensitivity of the short-to-intermediate wavelength geoid to rheologic structure in subduction zones

It is well established that the long wavelength geoid and dynamic topography are responsive to the radial viscosity structure of the mantle, but recent studies indicate that lateral viscosity variations affect the geoid at shorter wavelengths. These studies, however, only consider Newtonian viscosity structures, although experimental deformation studies of mantle minerals and seismic observations of lattice preferred orientation in the upper mantle provide evidence that dislocation creep is an active deformation mechanism at mantle conditions. In addition, the lithosphere is expected to yield plastically at high stresses based on laboratory measurements of yield strength. To quantify the effects of lateral viscosity variations and realistic flow laws on the short-to-intermediate wavelength dynamic topography and geoid near subduction zones, we consider a composite viscosity that accounts both for Newtonian and stress-dependent deformation mechanisms, including plastic yielding. Regional models of instantaneous stokes flow models are computed on a variable resolution mesh using CitcomS, where the resolution ranges from 25 km away from the subduction zone to 5 km in the vicinity of the subducting slab. The slab is defined as an 80 million year old lithosphere temperature anomaly smoothed above and below by half-space cooling models, and extends 100 km into the lower mantle. The buoyancy and stress fields are expanded to spherical harmonic degree 360, corresponding to a spatial resolution of about 110 km. These fields include the effects of self-gravitation and are used to solve for the surface geoid, as well as for dynamic topography at the surface and core-mantle boundary. Results of preliminary, layered mantle viscosity models are consistent with previous geoid studies, the main conclusion being that a more positive geoid at subduction zones is the product of relative viscosity increases with depth. In layered models, increased viscous support of the down-going slab with depth decreases its velocity, resulting in lower pressure gradients above the slab, decreased dynamic topography at the surface and a more positive geoid relative to the case of a uniform viscosity mantle. In contrast, relatively stronger layers above the slab increase coupling to the surface, thereby increasing dynamic topography and leading to a more negative geoid. Future models with temperature-dependent viscosities will include lateral variations in viscosity due to the temperature anomaly of the slab. These model results will be compared with those from models that include both linear and non-linear viscosities to determine the sensitivity of the short-to-intermediate wavelength geoid to increasingly realistic subduction zone rheologies.