Geophysicists compare the behavior of Earths outer core to a liquid,
with about the same flow as water. As the big hot ball of melted metal spins
and roils about, it creates the planets magnetic field, which changes
with the cores own chaotic behavior. Modeling that behavior and
the incipient flip-flops of the magnetic fields orientation that sometimes
accompany it has proven difficult. But a team of researchers has made
a small step toward doing so, using one of the largest supercomputers on the
planet to run the most realistic model yet.
The magnetic field, which protects the planet from bombardment by cosmic radiation, is constantly changing both in strength and configuration. Although researchers consider it unlikely to be imminent, some suggest that the magnetic field may shift polarity sometime in the next several millennia, so that the positive and negative poles flip between north and south.
Geophysicists have tried to predict the cores, and therefore the magnetic fields, behavior using models that function much like weather models do, but without the enormous benefit of being inside the fluid being modeled, says David Stevenson of Caltech in Pasadena. We actually know whats going on for the atmosphere and oceans, Stevenson says, but not in Earths core. Atmospheric and climate modelers can use direct observations so they can omit viscosity from their models, but models of the core must include viscosity, a flow characteristic represented by a parameter known as the Ekman number.
Using the smallest Ekman values yet for turbulent viscosity of the liquid outer core, a team of scientists at the Japan Aerospace Exploration Agency in Kanagawa and the Tokyo Institute of Technology used the Japanese supercomputer Earth Simulator to model the magnetic field. It took several thousand hours to model about 200,000 years of motion in Earths core. The Japanese team published their results in Science on July 15, noting that their model expresses a magnetic field that looks realistic, with nodes of magnetic variations that flow and merge over time until the entire field has switched its polarity. Such patches might be precursors to a reversal event, they say.
Although other models produce realistic-looking magnetic fields as well, says Gary Glatzmaier, a modeler at the University of California in Santa Cruz, no model is yet as realistic as we would like it to be for any kind of conclusions regarding magnetic field somersaults. Nevertheless, Glatzmaier says, the Japanese researchers are certainly on the right track in using the smaller values for turbulent viscosity.
The smaller Ekman number gets closer than ever to what geophysicists think is reality in the core, where the extremely low viscosity makes for very small-scale turbulence that is quite difficult to model on a computer of any size. Glatzmaier compares it to trying to simulate every small gust of wind down to the human scale while running a model of the planets atmosphere, to determine the global weather and future climate. That means that for even more realism, values for the Ekman number should be smaller, probably by several orders of magnitude, Glatzmaier says.
The new model tends to reverse its polarity too often, with reversals occurring every 5,000 years, rather than once in tens to hundreds of thousands, or even millions of years, as they do in the geologic record, Glatzmaier and Stevenson say. Plus, based on the smooth, large-scale patterns of the magnetic fields that the team reported, the models simulated fluid flows may still be too laminar compared to the strong turbulence expected in the Earths fluid core, Glatzmaier says.
Still, the teams results using more realistic viscosity values support the essential idea that rotation and the behavior of the magnetic field are dominant in theoretical calculations, Stevenson says. Although this model should not be regarded as the final word, he says, this is a significant step forward, in a pursuit that sees major improvements only every decade or so.
Back to top