Geomagnetic flip may not be random after all
21 March 2006

One of the most fascinating natural phenomena on Earth is the flipping
of its magnetic field, which has occurred hundreds of times in the last
160 million years. When the magnetic field flips, the North Pole becomes
the South Pole and vice versa. The last time this happened was some
780,000 years ago, so we could be heading for another reversal soon.
Now, physicists in Italy have found that the frequency of these polarity
reversals is not random as previously thought but occurs in clusters,
revealing some kind of "memory" of previous events (physics/0603086).

Although a full geomagnetic polarity reversal can take thousands of
years to complete, the implications could be enormous. As well as
affecting the migration trajectories of birds and other animals, the
disruption to the Earth's magnetic field could expose the Earth to
hazardous cosmic rays -- a scenario that some researchers have linked to
mass extinction events like the one that wiped out the dinosaurs around
65 million years ago. Geoscientists believe that our planet's internal
magnetic dynamo is responsible for pole reversals, but the actual
mechanism is not well understood.

Previous analyses assumed that the number of times the poles have
reversed over last 160 million years follows a Poisson distribution,
implying that the events are random. The Poisson distribution tells you
the probability of a number of events occurring in a fixed time if the
events are independent and the average rate is known. A good example of
the Poisson distribution in physics is the likelihood of unstable
radioactive nuclei decaying in a certain period.

Now, a team of physicists led by Vincenzo Carbone of the University of
Calabria have discovered that the sequence of polarity reversals can be
well described by a Lévy distribution instead. In contrast to Poisson
statistics, the Lévy distribution describes stochastic processes that
are characterised by the presence of "memory" effects -- or long-range
correlations between the events in time. Lévy distributions are widely
used to study many critical phenomena, such as earthquakes, and also
when analysing financial data. The researchers obtained their results by
careful statistical analysis of different sets of paleomagnetic data
containing estimates of when the Earth's poles reversed.

"The result means that polarity reversals are not random events that are
independent of each other," explains team member Fabio Lepreti.
"Instead, there is some degree of memory in the magnetic dynamo
processes giving rise to the reversals," he says. "We hope that our work
will serve as a useful reference point for models that aim to describe
the phenomenon of pole reversal." The Italy team now plans to build new
dynamic models to describe the field reversal sequences in a simple way,
so that the physical mechanisms that trigger pole reversals can be more
easily explained.

About the author
Belle Dumé is science writer at PhysicsWeb