Contact: Tim Stephens, (831) 459-2495,

WASHINGTON, D.C. — Deep in Earth’s interior is a dynamo that creates the planet’s magnetic field —
a kind of generator driven not by spinning turbines but by swirling flows of liquid iron. The workings
of this dynamo cannot be observed, but scientists have used computer simulations to gain powerful
new insights into the operation of the “geodynamo” and the behavior of Earth’s core.

The first self-consistent, three-dimensional computer simulation of the geodynamo was achieved in
1995 by Gary Glatzmaier, now a professor of earth sciences at the University of California, Santa
Cruz, and Paul Roberts, professor of mathematics at UCLA. Glatzmaier, Roberts, and their
coworkers have since refined and extended their simulations, shedding new light on the planet’s inner

Glatzmaier presented the group’s latest findings on Sunday, February 20, at the annual meeting of
the American Association for the Advancement of Science in Washington, D.C.

The Glatzmaier-Roberts model of the geodynamo is essentially a complex set of equations describing
the physics of Earth’s core. Scientists had long speculated that the mechanism behind the
geomagnetic field involved the motion of the Earth’s fluid outer core, which surrounds a solid inner
core. Both are composed mainly of iron. The solid inner core is about the size of the moon and as
hot as the surface of the sun.

The flow of heat from the core ultimately drives the geodynamo. “Basically, the whole thing works
because the Earth is cooling off,” Glatzmaier said. The cooling process results in fluid motions in the
outer core that produce an electric current, which, like any electric current, generates a magnetic

One of the initial achievements of the Glatzmaier-Roberts model of the geodynamo was the
simulation of a reversal of Earth’s magnetic field, when the north and south magnetic poles trade
places. This phenomenon has occured many times in the history of the planet, according to
paleomagnetic records preserved in rocks that show the direction and strength of Earth’s
magnetism at the time the rocks formed.

“We were able to get a magnetic field generated by the model that looks a lot like the Earth’s and
undergoes reversals,” Glatzmaier said.

The model also predicted that the solid inner core should rotate slightly faster than the surface of
the Earth. This prediction was later supported by other researchers using evidence from seismic
waves that pass through the core.

Over the past five years, Glatzmaier and his coworkers have improved the precision and resolution
of their model, taking advantage of advances in computer capacity. They have now run simulations
spanning as much as 300,000 years and showing a pattern of magnetic-field reversals very similar
to that seen in the paleomagnetic record.

“We can run the simulation for 200,000 years and the magnetic field will be stable for a very long
time — millions of time steps for which we solve these equations. Then within a thousand years it
reverses polarity, and then it remains stable again for another long period. We were very happy to
see that, because that’s also what we see in the Earth’s record,” Glatzmaier said.

He noted that the reversals are not triggered by an external influence on the geodynamo. “It is
simply due to the very nonlinear, chaotic nature of the dynamo system,” he said.

The group’s most recent efforts have focused on the role of the mantle in controlling the frequency
of geomagnetic reversals. Temperature variations in the mantle, causing an uneven pattern of heat
flow from the outer core into the mantle, may affect the fluid dynamics of the outer core. So
Glatzmaier’s group ran their simulation using eight different patterns of heat flow across the
core-mantle boundary.

The results, published in the October 28, 1999, issue of the journal Nature, showed that the pattern
of heat flow determined by the mantle does have a big influence on the behavior of the geodynamo.
The most Earthlike pattern of magnetic-field reversals occurred with a relatively uniform heat-flow
pattern. This suggests that scientists may have overestimated the extent of thermal variation in the
mantle, or that variations in mantle composition may compensate for thermal variations.

“We’re still far from satisfied that we have all the answers,” Glatzmaier said. “The model is a way of
exploring the unknown, and it looks very promising because the results are so much like the real
magnetic field. But we have less confidence in the details, and that’s where more powerful computers
will help.”


Editor’s note: You may contact Glatzmaier after the AAAS meeting at (831) 459-5504 or

Additional information about the geodynamo, including color images derived from computer
simulations, can be found on the Web at