Journal of Conference Abstracts

The Mantle Circulation Problem

H.-P. Bunge

Institut de Physique du Globe, Paris, France

Circulation models are a powerful tool to understand ocean and atmospheric dynamics, but they have not yet played a role in constraining the dynamics of the mantle. The recent advances in global seismic tomography to provide detailed images of subducted oceanic lithosphere suggest to model the evolution of mantle heterogeneity using high-resolution mantle convection calculations. An essential input-parameter to the circulation problem comes from well-determined reconstructions of past (150 Ma) motion, which provide not only an 80 percent complete velocity boundary condition for the mantle (the other 20 percent being the CMB), but also constrain the development of the internal mantle buoyancy forces through the history of subduction.

We have used the geodynamics code TERRA, implemented on a massively parallel CRAY T3E, to study the mantle circulation problem. 10 million finite elements provide a grid point resolution of less 50 km throughout the mantle, sufficient to model thermal convection at 10⁸ Rayleigh number. We find good correspondence between the circulation model heterogeneity structure and the S-wave study of Grand [1997], as expected for convection dominated by sinking of old slabs. Importantly, the circulation models demonstrate that plate reconstructions span sufficient time to overcome the uncertainty from unknown initial conditions. A remarkable observation is related to modeling the endothermic phase transition at 670 km depth, which results in a relatively poor fit with tomographic data, when a strongly negative clapeyron slope is assumed. These results indicate the potential of circulation modeling to test geodynamic hypotheses with seismic data.