|
Dietmar Muller, The University of Sydney (Australia)
Lydia Di Caprio, The University of Sydney (Australia)
Michael Gurnis, California Institute of Technology (United States)
Trond Torsvik, Geological Survey of Norway (Norway)
|
|
The past latitudes and longitudes of the tectonic plates are reconstructed via absolute plate motion models. These models describe how the fragments of Earth's outer shell have moved relative to a fixed reference system, such as the spin axis. Palaeomagnetic data provide palaeo-latitudes alone, not longitudes, due to the radial symmetry of the Earth's magnetic field. In contrast, hotspot tracks - volcanic chains with age progression reflecting the passage of a plate over a hotspot - could theoretically provide both palaeo-latitudes and -longitudes. However, none of these tracks are preserved before 130 million years ago, and thus we have no useful information for those earlier times. In addition, we know that currently available moving hotspot reference frames, which take into account the deflection of plume conduits in the convecting mantle, are not reliable before 100 Ma. The contrasts between alternative absolute plate-motion models have triggered major controversies, which are born out in geodynamic interpretations of seismic shear wave anomalies in the lower mantle. It has been suggested that the lower mantle underneath Antarctica is "too blue", referring to the intense positive seismic shear-velocity anomalies found underneath Antarctica in the lowermost mantle. These anomalies reflect subducted slabs that have descended from the surface deep into the mantle, eventually piling up on the core-mantle boundary. The lower-mantle anomaly below Antarctica appears to far exceed what would be expected from the subduction history along West Antarctica, where subduction ceased about 105 Ma. However, this conclusion is based on outdated plate-motion models, which place the eastern Gondwanaland margin in an incorrect longitudinal position in the Cretaceous. The Cretaceous position of this margin is constrained by the well-mapped extent of subducted slab material underneath the mid-ocean ridge south of Australia at the Australian-Antarctic Discordance. We propose that Earth's subduction history can be used to create a subduction reference system. In this system, the positions of subducted plates are used as reference markers for modelling the time-dependent effect of mantle convection on the Earth's surface. Our system also includes a workflow to attempt to correct for True polar Wander. This involves computing the global average of continental motion and rotation through time in a palaeomagnetic reference frame. In this way, we are able to separate motions with the characteristics of True Polar Wander, which are reflected in coherent rotations of all continents around a point close to their common centre of mass, from those of continents over the underlying mantle. We show preliminary CitcomS geodynamic models from eastern Gondwanaland that illustrate the workflow we are developing to iterate towards a subduction reference frame.
|
