Global geodynamic models constrained by tectonic reconstructions including plate deformation

In order to investigate the effect of mantle flow on the Earth's surface, imposing the kinematics predicted by plate reconstructions in global convection models has become common practice. Such models are valuable to investigate the effect of the mantle flow beneath the lithosphere on surface topography. Changes in surface topography due to lithospheric deformation are so far not part of top-down tectonic models in which plates are treated as rigid in traditional tectonic reconstructions. We introduce a new generation of geodynamic models that are based on tectonic reconstructions with deforming plates at both passive and convergent margins. These models allow us to investigate the relationships between lithospheric deformation and mantle flow, and their combined effects on surface topography. In traditional tectonic reconstructions, continents are represented as rigid blocks that either overlap or are separated by gaps in full-fit reconstructions. Reconstructions that include a global network of topological plate polygons avoid continental overlaps and gaps, but velocities are still derived on the basis of the Euler poles for rigid blocks. To resolve these issues, we developed a series of deforming plate models using the open source plate modeling software GPlates. For a given area, our methodology requires the relative motions between major rigid continental blocks, and a definition of the regions in which continental lithosphere deformed between these blocks. We use geophysical and geological data to define the limit between rigid and deforming areas, and the deformation history of non-rigid blocks. The velocity field predicted by these reconstructions is then used as a time-dependent surface boundary condition in global 3-D geodynamic models. To incorporate the continental lithosphere in our global models, we embed compositionally distinct crust and continental lithosphere within the thermal lithosphere. We define three isostatic columns of different thickness and buoyancy based on the tectonothermal age of the continents: Archean, Proterozoic and Phanerozoic. In the fourth isostatic column, the oceans, the thickness of the thermal lithosphere is assimilated using the half-space cooling model. We also use this capacity to define the thickness of the thermal lithosphere for different continental types, with the exception of the deforming areas that are fully dynamic. Finally, we introduce a new slab assimilation method in which the thermal structure of the slab, derived analytically, is progressively assimilated in the upper mantle into the dynamic models. This method not only improves the continuity of slabs in our models, but it also allows us to model flat slab segments that are particularly relevant for dynamic topography. This new generation of models allows us to analyse the contributions of continental deformation and of mantle flow to surface topography. We compare our results to geological and geophysical data, including stratigraphy, paleo-altimetry, paleo-environment and mantle tomography. This allows us to place constraints on key model parameters and to refine our knowledge of plate-mantle interactions during continental deformation.