• 19:00 11 May 2006
• NewScientist.com news service
• Patrick Barry

The voyages of Captain Cook have just yielded a new discovery: the gradual weakening of Earth’s magnetic field is a relatively recent phenomenon. The discovery has led experts to question whether the Earth is on track towards a polarity reversal.

By sifting through ships�?logs recorded by Cook and other mariners dating back to 1590, researchers have greatly extended the period over which the behaviour of the magnetic field can be studied. The data show that the current decline in Earth's magnetism was virtually negligible before 1860, but has accelerated since then.

Until now, scientists had only been able to trace the magnetic field’s behaviour back to 1837, when Carl Friedrich Gauss invented the first device for measuring the field directly.

The field’s strength is now declining at a rate that suggests it could virtually disappear in about 2000 years. Researchers have speculated that this ongoing change may be the prelude to a magnetic reversal, during which the north and south magnetic pole swap places.

But the weakening trend could also be explained by a growing magnetic anomaly in the southern Atlantic Ocean, and may not be the sign of a large scale polarity reversal, the researchers suggest.

Crucial measurements

David Gubbins, an expert in geomagnetism at the University of Leeds, UK, led the study which began scouring old ships' logs in the 1980s, gathering log entries recording the direction of Earth's magnetic field.

It was common practice for captains in the 17^th and 18^th centuries to calibrate their ship's compasses relative to true north and, less often, to measure the steepness at which magnetic field lines entered the Earth's surface.

Even as far back as 1590, these measurements were typically very accurate �?to within half a degree. "Their lives depended on it," Gubbins explains.

Such ship-log records may not be adequate for reconstructing the planet's past magnetic fields in fine detail, but the data can estimate large-scale features quite well. "In that regard, I think it's a very solid result," says Catherine Constable, an expert in palaeomagnetism at the University of California, San Diego, US, who was not involved in the study.

Mineral evidence

Using the locations of the ships at the time of measurement, these records allowed Gubbins to construct a map of the relative strength of Earth's magnetic field between 1590 and 1840, which was published in 2003.

The data was combined with 315 estimates of the field's overall strength during that period, based on indirect clues, such as mineral evidence in bricks from old human settlements or volcanic rock.

Gubbins showed that the overall strength of the planet's magnetic field was virtually unchanged between 1590 and 1840. Since then, the field has declined at a rate of roughly 5% per 100 years.

Every 300,000 years on average, the north and south poles of the Earth's magnetic field swap places. The field must weaken and go to zero before it can reverse itself. The last such reversal occurred roughly 780,000 years ago, so we are long overdue for another magnetic flip. Once it begins, the process of reversing takes less than 5000 years, experts believe.

Growing anomaly

A large-scale reversal might indeed be underway, Gubbins says, but the acceleration of the magnetic decline since the mid-1800s is probably due to a local aberration of the magnetic field called the South Atlantic Anomaly. "It looks like that's responsible for most of the fall we're seeing," he says.

This patch of reversed magnetic field lines covering much of South America first appeared in about 1800, according to the ship-log data. It slowly grew in strength, and by about 1860 it was large enough to affect the overall strength of the planet's magnetic field, Gubbins says.

If the field does flip 2000 years from now, the Northern Lights will be visible all over the planet during the transition, and solar radiation at ground level will be much more intense, with no field to deflect it.

There is no need to worry, though, argues Gubbins, as our ancestors have lived through quite a few of these transitions already.

Journal reference: Science (vol 312, p 900)