The Earth's magnetic field is a subtle clue to a turbulent past, when sheer gravitational pressure melted rock, separating out the heavier iron and nickel which sank inward, forming a metallic spinning core. Measurements of the resulting magnetic field are almost like X-rays, allowing a glimpse of the position and mass of the underlying metallic "bones" of Earth.

          The other planets of the Solar System have magnetic fields, whose strength and variation hint at buried secrets that may be quite different from what has been found on Earth. The planet Mars for instance, may not have melted in its early stages. Its magnetic field is generally weak, does not seem to emanate from the core, and varies in direction and strength all over the surface. There are places on the surface of Mars where the field is 10 times as strong as Earth's. These anomalies seem to arise from local concentrations of solid metal in the crust. The fact that the metal has been magnetized suggests that at some point there was a much more powerful planetary magnetic field which has since disappeared. Another set of anomalies shows up as "holes" in the field, occurring just where asteroids are known to have punched holes into the surface. In other places, disruptions in the magnetic field lie along a line, just as they do at points on Earth where the sea floor is spreading.

         These tantalizing facts come from the massive amounts of data returned by the Mars Global Surveyor satellite as it weaved an orbit around Mars, making hundreds of thousands of measurements. Such a huge amount of data is actually useless until someone can assemble it into a comprehensive model of the planetary magnetic field. With a good model, scientist can hunt for patterns, make inferences about the interior of the planet, decide where to send future satellite missions towards areas of interest, make comparisons with Earth, and possibly shed light on some poorly understood features of Earth's own field.

          Such a model has been constructed by Joseph Cain and his collaborators. The FSU researchers had to clean up the data, which had been taken at various altitudes, some during the Martian day and some at night. A grid was laid over the Martian surface to organize and group the data, which represented hundreds of thousands of individual observations. Then a computer program, originally developed for Earth data, analyzed the data on the SCS IBM SP3 supercomputer.

          To understand the modeling process, imagine hearing a particular musical chord played on a piano, and being asked to reconstruct that chord by naming the keys to be played, and how hard each is to be struck. For the Martian magnetic field model, the "keys" come from a generic set of functions commonly used to describe patterns centered on a sphere. The computer program is told the maximum number of such patterns it is allowed to use, and must determine how to combine these patterns so as to come as close as possible to the given measurements.

          The researchers made two calculations, first using pattern functions up to degree n = 50, and then a much harder calculation up to degree n = 90. To test the models, they predicted the magnetic field along a line where actual measurements were available. In the figure below, only the X component of the magnetic field is displayed, as predicted by the two models, and as recorded in the data. The shortcomings of the n = 50 model (in turquoise) are clear, but the n = 90 model (dark blue) seems to do a very good job of approximating the data (black). This gives scientists confidence in Cain's model.

          Now the researchers are preparing even more difficult calculations to be run on SCS's new IBM SP4 supercomputer. This new phase of the work will be more difficult; not only will the number of functions be increased to get a more accurate model, but the researchers also hope to include the outer reaches of Mar's ionosphere, where strong magnetic currents interact with plasma streaming from the Sun.

          Dr Cain is working closely with a group of graduate students, including Bruce Ferguson, who made the original runs on the IBM SP3, and David Mozzoni, who worked hard to modify and and extend the original program, and to create graphic images of the results.

         Professor Joseph Cain,
         Geophysical Fluid Dynamics Institute and
         School of Computational Science,
         Florida State University

To find out more...

The research group maintains a "Planetary Magnetic Fields" web site at http://geomag.gfdi.fsu.edu where you can find out more about this project.

You may also refer to the paper:

An n = 90 internal potential function of the Martian crustal magnetic field,
Joseph Cain, Bruce Fergusion, David Mozzoni,
Journal of Geophysical Research,
Volume 108, Number E2, 25 February 2003.