IMAGE: G. GLATZMAIER/LOS ALAMOS NATIONAL LABORATORY and P. ROBERTS/UCLA/SCIENCE PHOTO LIBRARY

Tangled Story: This simulation of a geomagnetic field reversal was done at Los Alamos National Lab.

Earth’s magnetic field has been monitored carefully since the 1830s, when the German polymath Karl Friedrich Gauss invented a way to measure its intensity. Since then, the field has decayed at the ­startling rate of about 5 percent per century [see photo, “Tangled Story”]. Has Earth’s field been in a spiral of decay for longer than that? Or do we happen to live in a period when the decline is particularly striking?

Now British geophysicist David Gubbins and his colleagues have an answer from the most unlikely quarter: data hidden in the logbooks of ships that navigated the planet’s oceans in the 16th, 17th, and 18th centuries. The results have allowed Gubbins to build a remarkable picture of the behavior of Earth’s magnetic field in the centuries before detailed measurements were possible.

But how can the magnetic field be measured in retrospect? One important clue is that the field can become trapped when molten rock solidifies—in a volcanic flow, for example. Measuring the trapped field gives you an indication of the field intensity when the rock solidified. These kinds of measurements have been made at 315 sites that date from the period between 1590 and 1840, creating a database of “paleointensity” during that period.

The trouble is that measuring a trapped field is hugely difficult. “The intensity measurements have typical errors of about 10 percent,” says Gubbins, a professor at the University of Leeds, in the north of England. That’s more than enough to swamp any effect from a change in Earth’s field. So, by itself, paleointensity data ­cannot answer the question.

But Gubbins wasn’t deterred. Instead, he found a century-old and somewhat obscure theory that links the intensity of a magnetic field to another property called its dipole moment, a quantity related to the strength of the poles and the distance between them. Given a measure of the field strength and many measurements of the field’s direction, the magnetic moment can be well defined. Gubbins realized that the results from the paleointensity data could be significantly improved if he could find accurate measurements of the field direction from the same period.

It turns out that measurements of the direction of the field relative to the position of the sun were common between 1590 and 1840. “Mariners made extremely accurate measurements, because their lives depended on it,” says Gubbins. Over the past 20 years or so, he and others have been mining this data from the many thousands of ships’ logs that have survived in museums and archives, an endeavor that has occupied a steady flow of graduate students. Gubbins says there are 50 000 measurements alone in the records of the British East India Company, which had a monopoly on sea trade between Britain and India for much of the period that interests him.

Gubbins has now combined these data with the paleointensity measurements to calculate that Earth’s field was probably stable prior to 1840, or at least decaying at a much slower rate than it is now.

So what caused the sudden decline after 1840? Gubbins says it is due to regions of reversed magnetic field flux appearing in the Southern Hemisphere in the late 18th century, probably as a result of small thermal changes in Earth’s core. The field’s abrupt drop is consistent with other studies, he says. Data from older rock analyses suggest that the intensity of Earth’s field has declined by as much as 40 percent over the past 2500 years, at an average rate of 1.6 percent per century. That’s much slower than the current rate, supporting the idea that there can be quiet periods as well as periods of rapid change.

“It’s just coincidence,” he says, “that today’s period of rapid change began at about the time we became able to measure it.”