Geomagnetic reversals - the role of the mantle

It has long been postulated, and is now widely accepted, that the earth's magnetic field originates in the liquid outer part of the core. On the other hand, the periodicity of magnetic field reversals is not consistent with the physics of the outer core, and it appears that the properties and behaviour of both the solid inner core and the mantle have important influences over the reversal mechanism.

Our understanding of the geomagnetic field has been substantially enhanced in recent years, largely based on results from a very complex numerical modelling program developed in 1995 by Gary Glatzmaier and Paul Roberts of the Los Alamos Laboratory in New Mexico. The model, which takes thousands of hours to complete when run simultaneously on two supercomputers, has been used by Glatzmaier and Roberts to simulate the geomagnetic field over a period of 500,000 years. Although this is geologically short, it does provide important information on the various parameters which influence the geomagnetic field.

The liquid metal outer core has a low viscosity (its runny), a relatively even temperature distribution and a fairly rapid and consistent convection motion. (Convection velocities are in the order of tens of kilometres per year.) This motion produces the geomagnetic field, but the motion is also continually trying to reverse that field. The solid metal inner core (which is not convecting) has a much longer period of natural magnetic variability, and it has the effect of inhibiting the magnetic reversals. Attempts at reversals are only rarely successful, and hence the pattern of reversals (as shown to the left) appears to be random.

In fact the reversal pattern is not entirely random. Looking back over the past 83 m.y. there has been a general decrease in the frequency of reversals, and during the 37 m.y. period from 83 to 120 m.y. there were no reversals at all. The long period of normal magnetism from 83 to 120 is known as the Cretaceous Superchron (a "chron" is a magnetic interval), or the Cretaceous long-normal⁽¹⁾. Going back beyond the long-normal there is a general increase in the frequency of reversals.

This type of pattern cannot be explained by the physical characteristics of either the liquid or solid parts of the core, but might be related to what goes on in the mantle.

Whereas convection in the liquid part of the core is very fast, with a period in the order of hundreds of years, convection in the mantle is much slower. The rate of motion is measured in centimetres per year, and the period is in the order of millions of years. A major rearrangement of the convection system can take 100s of millions of years. The mantle convection results in an uneven distribution of heat throughout the mantle, including the area of the lower mantle near to the boundary with the core (core-mantle boundary - CMB). We can get an idea of temperature differences in the mantle using seismic tomography (detailed 3-dimensional analysis of seismic data).

It is predicted that spatial differences in the temperature of the mantle should affect the geomagnetic field and how it changes with time. Glatzmaier and Roberts modelled geomagnetic field characteristics with a spatially even flow of heat across the CMB, and also with an uneven flow - as would be expected on the basis of what we know about convection-related temperature variations in the lower mantle. They found that the observed variations in geomagnetic field lie somewhere between the even and uneven heat-flow models. In other words, it is evident that there is uneven heat-flow at the CMB, but the scale of the variability is probably smaller than the existing models suggest.

The implications of this work are that the long-term variability in the rate of geomagnetic reversals might be derived from some long-term variabilities in mantle convection - the type of variation which is associated with the break-up and coming together of super-continents, and the opening and closing of ocean basins. Such features have not been tested using the Glatzmaier and Roberts model, as that would require simulations of hundreds of millions of years of the earth's history - rather than hundreds of thousands of years as has been done to date.

References:

Glatzmaier, G. and Roberts, P., A three-dimensional self-consistent computer simulation of geomagnetic field reversal, Nature, V. 377, p. p. 203-209, 1995.

Glatzmaier, G., Coe, R., Hongre, L. and Roberts, P., The role of the earth's mantle in controlling the frequency of geomagnetic reversals, Nature, V. 401, p. 885-890, October 1999.

Buffett, B., Role reversal in geomagnetism, Nature, V. 401, p. 861-862, October 1999.

You might also be interested in the following websites which show neat pictures of what the geomagnetic field looks like:

ees5-www.lanl.gov/IGPP/EarthsMagneticField.html

es.ucsc.edu/~glatz/geodynamo.html

1. Recent work by Randy Enkin of the Pacific Geoscience Centre in Sidney shows that the end of the Cretaceous Long-normal is actually represented in the Nanaimo Group rocks in the lower part of the Malaspina Cut. The 83 m.y. old rocks at the bottom of the succession show normal magnetism, while those immediately above are reverse magnetized.

Steven Earle, 1999. Return to Earth Science News