Principal Investigator Geoffrey F Davies Projects k43, r20

Research School of Earth Sciences Machine VP, CM

Mantle Convection

Convection in the Earth's mantle is the fundamental process underlying plate tectonics and continental drift. There are in fact two active components driving mantle convection: plates and plumes. The plates sink from the top and the plumes rise from the bottom of the mantle. There have been considerable technical difficulties involved with the large viscosity gradients associated with plates and plumes, and only in the past year have these been overcome. A number of projects in this area are in various stages.

What are the basic questions addressed?

How can observations at the Earth's surface be used to constrain the form of mantle convection, so that it can be used to understand the evolution of the Earth?

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

The interaction of descending plates and rising plumes with the mantle transition zone, (where pressure-induced phase transformations occur) is a topic of great current interest because of questions about possible layering in the mantle and the long-term chemical and thermal evolution of the mantle. New models have shown that both plates and plumes tend to break through the transition zone more readily than the upwellings and downwellings of constant viscosity models published to date. This means that the mantle is less likely to be layered, and this is consistent with several lines of observational evidence to that effect.

The generation and rise of plumes through the mantle is being studied systematically. This work quantifies the conditions under which plumes will form by instability of the thermal boundary layer at the bottom of the mantle. This is required to better understand the role of plumes in early Earth history and their importance in controlling the cooling of the Earth's core. It has also been demonstrated that the form of plume heads arriving at the top of the mantle is strongly affected by a viscosity step believed to exist in the transition zone, but not much affected by phase transformations, except that the latter may completely block the rise of plumes in some circumstances. This is important for understanding better the nature of flood basalt eruptions.

The amount of melting in plumes due to the release of pressure as they approach the surface, is being calculated from these models, but the work is at an early stage. This will complement much more detailed melting calculations being done by Dr M Cordery, in project r08.

The work on plume generation and ascent is near completion. The work on melting will continue. There are important questions to be explored concerning the effect of mantle composition and rheology on the amount of melt produced. This is one of the outstanding questions concerning the plume model of flood basalt eruptions.

What computational techniques are used and why is a supercomputer required?

A multigrid finite difference method is used to solve for the flow at each timestep in the convection calculations. Fine numerical grids are required to resolve the details of the model, with up to 512 grid intervals per dimension. This is the most difficult and computationally intensive part of the calculation. The temperature field is then stepped in time using a standard alternating direction implicit method.


Penetration of plates and plumes through the mantle transition zone, G F Davies, accepted for publication.