Constraining mantle flow with seismic and geodynamic data: A joint approach

Understanding the style of convective flow occurring in the mantle is essential to understand the thermal and chemical evolution of Earth's interior as well as the forces driving plate tectonics. Models of mantle convection based on three-dimensional (3-D) seismic tomographic reconstructions have the potential to provide the most direct constraints on mantle flow. Seismic imaging of deep Earth structure has made great advances in recent years; however, it has not been possible to reach a consensus on the nature of convection in the mantle. Models of mantle flow based on tomography results have yielded variable conclusions largely because of the inherent non-uniqueness and differing degrees of resolution of seismic tomography models as well as the difficulty in determining flow directly from seismic images. Here we address this difficulty by simultaneously inverting global seismic and convection-related data sets. The seismic data consist of globally distributed shear body wave travel times including multi-bounce S-waves, shallow-turning triplicated phases, as well as core reflections and phases traversing the core (SKS and SKKS). Convection-related data sets include global free air gravity, tectonic plate divergence, and excess ellipticity of the core-mantle boundary. In addition, the convection-related constraint on dynamic surface topography is estimated on the basis of a recent global model of crustal heterogeneity. These convection-related observables are related to mantle density anomalies through instantaneous mantle flow calculations and linked to the seismic data via optimized density-velocity scaling relationships. Simultaneous inversion allows us to test various mantle flow hypotheses directly against the combined seismic and convection data sets, rather than considering flow predictions based solely on a seismically derived 3-D mantle model. In this study, we test four different mantle flow hypotheses, including whole-mantle flow and models with impenetrable flow boundaries at depths of 670 km, 1200 km, and 1800 km. This hypothesis testing shows that the combined global seismic and geodynamic data sets are best reconciled when a whole-mantle flow scenario is considered. Convection models with restrictive flow boundaries within the lower mantle provide distinctly poorer fits to these combined data sets providing evidence that the mantle flows without permanent hindrance at the boundaries considered. (c) 2006 Elsevier B.V. All rights reserved.