American Scientist

Since the advent of plate tectonic theory in the 1960s, geologists have recognized that that many parts of the earth's crust and mantle move horizontally. But most of that motion is exceedingly sluggish: usually less than 10 centimeters a year, which would barely outpace a growing fingernail. Four years ago, seismologists Xiaodong Song and Paul Richards (both then at Columbia University's Lamont-Doherty Earth Observatory) discovered that one part of the solid earth—the inner core—moves quite a bit faster, rotating in such a way that much of its surface shifts more than 10 kilometers a year. Such breakneck speed is possible because the inner core floats within a immense liquid shell.

Other seismologists have confirmed the inner core's "super-rotation" (relative motion to the east). But estimates for the rate have differed widely. Wei-jia Su, Adam Dziewonski (both of Harvard University) and Raymond Jeanloz (University of California, Berkeley) reported in 1996 that it was about three times faster than Song and Richards found. Later some seismologists (including Su and Dziewonski) argued that zero relative rotation could not be ruled out. Yet others affirmed the eastward rotation but suggested rates three to five times smaller than Song and Richards's best estimate. Most recently, John Vidale, Paul Earle (both of the University of California, Los Angeles) and Doug Dodge (Lawrence Livermore National Laboratory) found strong evidence for a comparatively slow rate of rotation. Their analysis, which appeared in the May 25th issue of Nature, is especially notable because they approached the problem from a novel angle.

Other studies examined the speed of seismic waves refracted through the inner core. These signals can reveal motion because the inner core has a "grain," like a ball sculpted from a single piece of wood. Seismic waves move faster when they travel from the north pole of the inner core to the south (or vice versa) than they do when they take a more equatorial route. The grain, or fast direction, runs largely north-south but not entirely so: It appears to be tilted over by about 10 degrees.

It was this slight deviation from perfect axial symmetry that allowed Song and Richard to test for relative rotation in the first place. They examined seismic waves passing through the inner core along essentially one path and saw how the speed changed with time. What they found was a slight shift—about a third of a second difference in arrival times over three decades. The effect was subtle, because they did not have a lot of older measurements to work with: Finding sources and seismometers with just the right geometry to track waves that penetrate the inner core is quite difficult, especially when one is trying to find records from decades ago. Although their fundamental observation of changing travel times has held up, the model of the inner core that they used proved overly simple, which probably explains why their original estimate for the speed of rotation was exaggerated.

Vidale and his colleagues also compared seismic records from different years. But instead of using waves that had shot clear through the inner core, they looked at ones scattered backward from it. In a sense, they did not rely on the grain of the inner core but on its lumpy constitution. (It is as if the wooden-ball core has many small knots in it.) Such scattered seismic waves have small amplitudes and are hard to distinguish from background noise. But Vidale and his co-workers took advantage of recordings collected with one of the most sensitive instruments ever, the long-defunct Large Aperture Seismic Array (LASA)—a vast deployment of seismometers in eastern Montana, installed to monitor Soviet nuclear tests between 1965 and 1978.

In fact, the seismic waves Vidale used came from two such Soviet blasts, conducted in 1971 and 1974. These huge explosions injected a considerable amount of seismic energy into the deep earth, some of which reflected off of lumps in the inner core and reached the array of instruments in Montana. Because LASA was hundreds of kilometers across, it can, in a sense, be steered like a radar antenna (even now) by suitably combining the many individual signals that were recorded. With the array pointed toward the eastern side of the inner core, Vidale and his colleagues found that the scattered waves arrived after a slightly longer delay in 1974 than in 1971; with it pointed to the western side, the scattered signals arrived after a slightly shorter delay. This pattern coincides with the inner core rotating to the east at about 0.15 degrees per year, which shifts its equator annually by about 3 kilometers.

"It was serendipity," says Vidale of how he first conceived of this approach. He was studying the mantle using seismic waves that by happenstance also penetrated the core. "I got bored one day and tried to identify every arrival I had in this 40-minute recording," Vidale recalls. Some of the signal he saw did not fit with what was expected for the known layering within the earth: "We finally figured out that it was energy scattering around within the inner core," indicating that there must be some complicated structure therein. "Once we identified it, we realized we had a very sensitive means of testing inner-core rotation."

So Vidale's team was somewhat lucky to have stumbled on the key observation. Indeed, they were lucky to see these LASA data at all. Although LASA was a massive, well-funded project, the measurements collected for military purposes during those Cold War years had not been archived. In fact, the data tapes had been languishing in a subcontractor's warehouse in Virginia and were about to be dumped when they were rescued by members of the U.S. Geological Survey's Albuquerque Seismological Laboratory. These scientists saved the tapes from destruction but have not yet been able to recover much data from them, something that has proved a considerable challenge.

Harold Bolton, a seismologist at the Albuquerque lab who helped Vidale's colleague Paul Earle recover the pertinent records, notes that one tape in five is just barely readable now, because "the oxide just kind of sloughs off." Bolton is working with others in the USGS to find ways to bake the oxide in place, keeping it there long enough to read the data onto another medium.

Older seismic data are valuable for many studies, not just those that examine how the inner core has changed with time. Yet the general lack of enthusiasm for keeping such measurements in an orderly form makes it difficult for seismologists to gain access to vintage records. Indeed, the library of seismic signals that Song and Richards used to such advantage in 1996 (microfilmed traces from the World-Wide Standardized Seismograph Network) is precariously preserved too. The original microfilms, once stored at the USGS National Earthquake Information Center in Boulder, were discarded, leaving only three extant copies (one of which is now housed at the Albuquerque Seismological Laboratory, not far from the LASA tapes). Bob Hutt, director of that laboratory, reports that another set, at USGS headquarters in Reston, Virginia, is unorganized and "in imminent danger of being thrown out." Such worries have prompted him and his coworkers to try to save this vast library from oblivion by scanning the microfilm copy he oversees and putting the digital images on CD-ROM. But as yet, they have not obtained the necessary support from the rest of the seismological community.

Earle notes that old seismic records contain signals from large nuclear-weapons tests, making them (in all probability) unique: "Hopefully we're not going to have any more blasts of [magnitude] seven." So studies like the one he carried out with Vidale would be impossible to mount with new data. Richards also laments the continuing loss of seismological measurements from prior decades. "I've got projects all over the world trying to look at old archives." He points out that neglect of the microfilms from the World-Wide Standardized Seismograph Network and the poor treatment of the LASA tapes pales in comparison to what he has seen happen to seismological data obtained by the Soviet military during the Cold War: "They did a much better job of throwing things out than we did."—David Schneider