Mantle Plumes

           

Principal Investigator

Alison Leitch

Research School of Earth Sciences

This project looks at the way mantle plumes rise and melt, producing vast amounts of lava, as they approach the Earth's surface.
The Earth's mantle is a rocky layer about 3000km thick between the crust and the molten iron core. Over a timescale of hundreds of millions of years, the mantle convects as a fluid. The convection is driven by "internal" heat from the decay of radioactive elements, and "bottom" heating from the core. The internal heat is lost at the surface by the formation and subduction of tectonic plates, while the bottom heat is lost from the core by mantle plumes. A mantle plume consists of a large, spherical head of hot material with an even hotter, narrow "tail" connecting the head to the source region just above the core. Plume heads and tails melt when they get close to the surface, the heads producing millions of cubic kilometers of lava (e.g. Deccan Traps in India), and the tails making strings of volcanic islands (e.g. Hawaiian Islands). We aim to understand the influences of mantle composition and viscosity structure, and the core temperature on the rate and total volume of melting.
 
         
     
         

 

Projects

r01 - VPP

     
         

 

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

Just off the east coast of North America, next to the continental shelf, is a long narrow region of thickened oceanic crust called the East Coast Margin Igneous Province (ECMIP). Oceanic crust is produced where the lithosphere (the top 100-200km of the mantle, which is cold and stiff) is cracked along a ragged line thousands of kilometers long (a rift). The two sides of the rift pull apart and underlying mantle rises up and partially melts, producing a basaltic crust. Unusually thick crust is often attributed to higher than normal temperatures in the mantle. We proposed that when the Atlantic Ocean opened the rift was underlain by the flattened head of a hot mantle plume. To test this, we carried out numerical models of mantle convection where we rifted the lithosphere over various model plume heads. To match the observations, we need a thin, flat plume head where the temperature tapers sharply from the top boundary. This structure results if, on its way from the core-mantle boundary to the surface, a plume head passes through a step decrease in the background viscosity of the mantle at 660 km depth. Such a step decrease is independently indicated from studies of gravity signals around subduction zones and post-glacial rebound.

Appendix A -
 
     

Flood basalt eruptions are massive outpourings of lava (105 - 107 cubic km) which take place within very short periods of time (1-3 million years). They occur sporadically over the globe about once every 10 million years. It has been suggested that they are due to decompression melting of a mantle plume head. If there is a step reduction in the mantle viscosity at 660 km depth, then the plume head necks down as it passes through the step, and the small, hot central part of the plume rises quickly and impinges directly on the base of the lithosphere. This allows about 105 cubic km of melting in 3 million years, but to explain the larger, faster eruptions we must look further. Geochemical studies suggest plumes contain a component of ancient oceanic crust, which has a much lower melting point than the "normal" mantle. When we add 15% by volume of this component to a plume passing through a step viscosity decrease, we can match the melt-rates of the larger eruptions.

An important realization that came from the work above was that while the rate of melting in a plume tail depends on the rheology and temperature in the lowermost mantle, melting in a plume head also depends greatly on the initial conditions. Continuing work aims to reconcile observations of melt-rates in plume heads and tails with plausible mantle rheologies and compostions.

What computational techniques are used?

The code used in this project, CONMG is a multigrid finite-difference code written by Dr Geoff Davies of the Research School of Earth Sciences and developed for use in this project by the author. The code is specialized to study mantle convection problems in two dimensions in a cartesian or cylindrical domain. It rapidly and efficiently solves the equations of conservation of mass, momentum and energy in an incompressible, highly viscous medium with boundary conditions appropriate to the Earth. To study melting problems (the aim of this project) very high resolution is needed, involving 40-240MB of core memory.

Publications

A. M. Leitch, M. J. Cordery, G. F. Davies, and I. H. Campbell, Flood basalts from eclogite-bearing mantle plumes, South African Journal of Geology, 104, (in press).

A. M. Leitch, G. F. Davies, and M. Wells, A plume head melting under a rifting margin, Earth and Planetary Science Letters (in press).

A. M. Leitch, V. Steinbach, and D. A. Yuen, Centerline temperature of mantle plumes in various geometries: Incompressible Flow, Journal of Geophysical Research, 101, 1996, 21829-21846.

     
- Appendix A