HR: 0800h
AN: T21C-0508    [Abstracts]
TI: Models of Laterally Variable Viscous Flow in the Earth's Mantle With Constraints From Mineral Physics and Surface Observations
AU: * Niehuus, K
EM: niehuus@geophysik.uni-frankfurt.de
AF: IMGF Goethe-University, Feldbergstr. 47, Frankfurt, 60323 Germany
AU: Schmeling, H
EM: schmelin@geophysik.uni-frankfurt.de
AF: IMGF Goethe-University, Feldbergstr. 47, Frankfurt, 60323 Germany
AB: Earth's mantle viscosity is commonly assessed by mineral physics, studies of post-glacial rebound and models of global mantle flow. As has been shown in previous studies of large-scale mantle flow, details of the viscous stratification may not be deduced unambiguously under the commonly employed constraint of static geoid, Haskell constraint of PGR, observed plate motions and estimates of dynamic topography. However, the added difficulty of the inferior determination of observable fields other than the geoid and/or artificialities inherent in their implementation in flow calculations (e.g. boundary conditions for plate tectonic behaviour) serves to complicate the interpretation of flow models by introducing artifacts which are very difficult to assess, rendering a mere addition offurther observables to fit, such as core-mantle boundary topography, lithospheric stresses, radial heat flow, in an effort to reduce this ambiguity, unfeasible. B. Steinberger and A. Calderwood [1] deduced detailed radially averaged temperature-dependent viscosity variations from independent mineral physics results and, assuming density heterogeneities and seismic tomography anomalies to be of thermal origin, computed conversion profiles for these quantities. Here, their recent study is extended to laterally variable viscosities (LVV) using an analytical spherical code. For laterally variable viscosity, each spherical harmonic mode is treated separately under the assumption of a suitable laterally averaged viscous stratification, while the mode-coupling is accounted for iteratively in terms of a viscous-load term on the right-hand side of the set ofODEs describing the flow field [2]. Lateral viscosity contrasts of up to a factor of 10³ can be treated. This range of viscosity variation suffices for the modeling of tomography- and upper-mantle slab sinking model-derived temperature-dependent LVV. We optimize the viscosity profiles obtained from mineral physics by dynamic modeling of geoid, surface motion and dynamic topography while satisfying the Haskell constraint, employing a minimum of fitting variables (3-4 anchor viscosities within the upper mantle, an additional factor balancing the laterally variable viscosity to account for errors entering this quantity). We find that inclusion of LVV necessitates a modification to our previously optimized, merely radially stratified, viscosities. While the latter displayed no clear preference for the location with depth of the lowest viscosities, models with LVV seem to favor the occurence of a soft layer just below the lithosphere over one in the transition zone. The misfit between predicted and constraining surface observables varies little between the cases radially and laterally variable viscosity, while differences are more pronounced between input tomographies themselves. [1] B. Steinberger and A. Calderwood. Models of viscous flow in the Earth's mantle with constraints from mineral physics and surface observations. GJI, 2005, submitted. [2] S. Zhang and U. Christensen. Some effects of lateral variations on geoid and surface velocities induced by density anomalies in the mantle. Geophys. J. Int., 114:531-547, 1993.
DE: 8100 TECTONOPHYSICS
SC: Tectonophysics [T]
MN: Fall Meeting 2005