Testing Mantle Circulation Models

Over the past decade, a new family of mantle convection models have been developed, which are conditioned by recent plate motion history (e.g. Bunge et al., 1997). They are commonly known as 'mantle circulation models' and allow for comparisons between present-day model predictions and ever improving seismic tomography images (e.g. Li et al. 2008). In this work, we present results from systematic investigations into the influence of various model parameters upon final model prediction/tomography correlations, to obtain a better understanding of their relative importance. These include a range of material properties, such as the radial viscosity structure, the Clapeyron slope of mineral phase transitions and compressibility; in addition to other aspects, such as the initial condition for the simulation. For our comparisons, we focus in particular on two large robust mid-mantle seismic anomalies, which others have related to the subduction of the Farallon and Tethys plates (e.g. Romanowicz, 1980). While these features are recovered with some fidelity in most simulations, the match can vary greatly. We find that there is a great deal of information in this mismatch, which includes information on the plate motion history.