Mantle Convection


Principal Investigator

Geoff Davies

Research School of Earth Sciences

Mantle convection is the fundamental driving process of tectonic processes in the earth's crust.
It includes the dynamics of plate tectonics and of mantle plumes, each of which represents a distinct mode of convection. This work is part of a long-term program to understand the dynamics of the mantle so that near-surface geological processes can be better understood. The subject has matured in the past few years to the point where major controversies are being resolved and dynamical modelling is beginning to contribute substantively to geology.


k43 - VPP




What are the results to date and the future of the work?

My work has in the past demonstrated that the earth's topography is an important constraint on the form of mantle convection. It has also shown that the picture of mantle convection inferred from this constraint is not necessarily inconsistent with geochemical evidence for long-term (billion-year) heterogeneity in the mantle.

In the past year I have shown that models of stiff (higher viscosity) lithosphere subducting and descending into a mantle whose viscosity increases with depth, in accordance with other evidence, yields a range of oblique or buckled and folded forms that are consistent in style with recent images from seismic tomography that seem to reveal the forms of subducted lithosphere in the mantle.

I have extended these models to investigate the factors that jointly control the topography and gravity perturbations at the surface. It has been claimed that these cannot be reconciled in the kind of single-layer model I have used. However I find that there are several factors at work and that a reconciliation is likely. This work is continuing.

I have also shown that a plume rising a under region with a thick continental lithosphere flows laterally to adjacent regions of thinner lithosphere, and that the locus of decompression melting can be shifted to the region where the plume material rises beneath the thinner lithosphere. This may explain a tendency of some flood basalt eruptions to concentrate near the edges of stable continental regions, and it is possibly a factor controlling the location of deposits of nickel and other minerals.

Appendix A -


What computational techniques are used?

We use a finite-difference multigrid method which I have developed. It vectorises to about 70% and is one of the more efficient codes in existence. It has been less robust than finite-element methods, but a new core algorithm is expected to considerably enhance its ability to handle large gradients in viscosity.




- Appendix A