Testing Dynamic Topographic Predictions of Mantle Convection Models Using Global Palaeobiological Datasets

Over geological timescales, aside from isostatic processes arising from crustal thickness variations, flow within the mantle has long been recognised to generate a significant component of Earth's topography, i.e. "dynamic topography". Therefore, geological and geophysical evidence of Earth's surface deflection can provide spatio-temporal evidence of deep Earth processes, if tectonic/crustal processes are accounted for. Mantle convection models can be used to calculate past and present dynamic topography in a number of ways, with the aim of matching surface observations to improve our understanding of mantle properties and flow characteristics. We analyse the global spatio-temporal patterns of dynamic topography predicted by a suite of models run using the TERRA code, which solves the Stokes and energy equations for mantle flow within a spherical shell. Both compressible/incompressible models are analysed, for a range of mantle viscosity structures. We calculate dynamic topography using two widely-used methods, focussing on the present-day where the pattern of dynamic topography is constrained in greatest detail. First, we examine dynamic topography using instantaneous surface stress calculated from full-resolution 3-D TERRA output. Secondly, model output is transformed into the spherical harmonic domain, and density anomalies at depth are propagated to surface stress variations, and therefore topographic deflections, using analytic sensitivity kernels i.e. the propagator matrix method. Each method makes subtly different assumptions about boundary conditions and mantle structure and properties. We demonstrate that uplift predictions calculated using each method can be compared with observational estimates derived from palaeobiological data, oceanic residual depth measurements, and continental gravity anomalies. We highlight key similarities and differences between dynamic topographic predictions from each method across a suite of mantle convection models, and identify correlation/misfit with observational constraints.