Constraints on Seismic Models from Other Disciplines - Constraints on 3-D Seismic Models from Global Geodynamic Observables: Implications for the Global Mantle Convective Flow

Abstract

Mantle flow models based on seismic tomographic images of three-dimensional (3-D) mantle structure may be used to successfully explain a wide range of global surface data related to mantle convection. These tomography-based convection models place strong constraints on viscosity and lateral density variations in Earth's mantle. The convection-related observables that can be modeled include the global free-air gravity anomalies, the dynamic surface and core–mantle boundary topography, and the present-day tectonic plate motions. The fundamental mantle flow theory needed to interpret these geodynamic observables is reviewed for the case of a self-gravitating, compressible mantle with coupled surface plates. The most important input to the flow models, namely, the effective viscosity of the mantle, may be inferred from simultaneous inversions of glacial isostatic adjustment and mantle convection data. These viscosity inferences are used in numerical calculations of the geodynamic response or kernel functions, which provide the theoretical relationship between the surface observables and the mantle density anomalies. Understanding the present-day dynamics of the mantle is critically dependent on knowledge of these internal density perturbations, and they are derived here on the basis of a wide selection of recent global tomography models derived from seismic shear-velocity ( V S ) data. A detailed review of the geodynamic observables predicted on the basis of these V S tomography models is presented, and it is found that all 3-D models derived solely from seismic data provide very similar fits to the convection data. Although the geodynamic fits obtained using the published, purely seismic 3-D mantle models are reasonably good, there remains much room for improvement. The residual misfit, of the order of 50% of the signal in the geodynamic data, raises fundamental questions about the degree to which mantle heterogeneity is resolved by the seismic tomography models or whether the classical fluid mechanical theory employed to model flow in the mantle is adequate. Another outstanding issue in current efforts to understand mantle flow dynamics is the relative importance of the thermal and compositional contributions to mantle heterogeneity. To address these major questions, a discussion is presented of recent efforts to successfully reconcile the independent constraints on 3-D mantle structure and flow provided by global seismic and convection-related data sets. The tomographic procedure for jointly inverting these seismic and geodynamic data, including additional constraints from mineral physics data, is found to greatly improve the fits to the convection-related data, and this is accomplished in the context of a dominantly thermal origin for the mantle heterogeneity.