Christopher Finlay talked about the gyre in the heart of the Earth as if it were alive. It was eccentric, he said. It had limbs. It twisted and stretched. The gyre’s actions were coercing the Earth’s magnetic dipole to decay. To me, it sounded as if this mysterious gyre were an organism secretly sucking energy from the dipole, feeding enemy factions within the core, destabilizing the regime. Far from stable and orderly, this was anarchy. It was a breathtaking peek at a covert drama taking place within the invisible magnetic force.
If anybody could say what the gyre looked like, it was Finlay. Tall and lanky, with curly brown hair and a quick smile, he was one of a small number of people on the planet with the prowess to work it out. He grew up near Belfast, Ireland, ambling around the countryside with a compass in hand, fascinated with the heroic tales of Edmond Halley and the magnetic crusade. By the time I met him, he was a geophysicist at the National Space Institute at the Technical University of Denmark in Copenhagen, the university Hans Christian Ørsted set up in 1829. Along with institutes in Potsdam and Paris, Finlay’s was one of the three European scientific centers monitoring data from Swarm, the triplet of satellites now tracking the magnetic field from space. His boss in Copenhagen, Nils Olsen, was known as the dean of Swarm data. The bottom line: If you’re a scientist working on the Earth’s evolving magnetic field, you know the work that Finlay and Olsen are producing.
It was through Finlay that I had come to be at the meeting in Nantes. When I saw him in Copenhagen, he had told me about it, under questioning. It would, he said carefully, be highly technical. Nevertheless, I had written to the organizers and received enthusiastic permission to attend as a journalist guest. Once there, Finlay was one of my guides to the goings-on, good-naturedly pointing out people to consult and explaining basic concepts. As the conference progressed, I ended up with a couple of pages of questions at the back of my notebook under the heading: Ask Chris!
In Copenhagen, I had caught him in the midst of preparing teaching materials and getting ready for another meeting. He was on the run. Every now and then, he would jump up and dash over to his computer to show me what he was talking about. He would click through PowerPoint presentations filled with notes for the classes he taught, and pull up colored maps showing the Earth’s magnetic field. As for Gubbins before him, and then Jackson, who supervised Finlay’s PhD at the University of Leeds, those maps represented an important way of describing his understanding of the field and how it is changing. Above his desk, nine printed-out maps were carefully thumbtacked to a large bulletin board. Each map save one had the same color tacks on each of the four corners. A great whiteboard dominated the wall beside it, covered with mathematical formulae and calculations in tidy blue, green, and black markings.
All this work is part of a quest to understand the dynamo inside the Earth. To do that, Finlay and his colleagues have created computerized numerical simulations to see if they can replicate the one that generates our magnetic field. The idea is to understand today’s field but also predict its future movements. It turns out that a key component in the model seems to be the gyre.
The best matches in the models show the gyre to be a thick swath of molten metal in the outer core, clenched between the grip of the solid inner core and that of the mantle. Mantle and inner core are locked together by gravitational forces, but convection deep in the core tends to drag the inner core to the east. To stay in balance, the gyre close to the top of the outer core is forced to move westward. That explains the westward drift of the Earth’s magnetic field that Halley first noticed. Because the inner core is growing lopsidedly—cooling faster underneath Indonesia—that too puts pressure on the gyre, distorting it into eccentricity. That explains why the long-term variation in the magnetic field happens mainly on the Atlantic side of the planet, and less on the Pacific. Bad luck for all those navigators who were tracking declination and dip on well-worn crossings from Europe to North America over the centuries. Had the greater variation in the field happened on the North America–to–Asia side instead, sailors would have had an easier time of it. Perhaps the magnetic crusade and the race to understand the magnetic field would not have happened at all.
One of Finlay’s papers contains a visualization of what the models say the gyre might have looked like in 2015. Colored bloodred and dark blue, the gyre pulsates in fleshy ropes, looking for all the world like viscera. One of the many revelations here is that the Earth’s inner anatomy is not symmetrical. It is fundamentally askew. This is not the image of the plodding bar-magnet within the Earth that researchers held until the last century. Or the hard-boiled-egg vision of the planet that was common just over a hundred years ago. Oldham, Jeffreys, and Lehmann, whose seismographic analysis produced the first glimpses of the Earth’s inner structure, would hardly recognize this stroppy and complex creature.
And in Finlay’s models, it was the combination of oddities in this limb-stretching gyre that was driving several critical phenomena affecting the dipole, and therefore the magnetic field. In the models, the gyre’s limbs reached up and down through the outer core in muscular columns. In the northern hemisphere, they transported magnetic flux in a balanced fashion: What went up to the pole also came down to the equator. In the southern hemisphere, things were badly off balance. A strong magnetic flow ran from southwest of Australia toward the equator, but it was not offset by a similar flow toward the south pole under South America. On the contrary, the reversed-flux patch that Gubbins and Jackson had been tracking in the south was growing both larger and stronger. It was this lack of symmetry in the south that was draining the dipole.
That was important, Finlay explained, because the magnetic anatomy within the core is already highly diverse, another revelation that would have stunned early magnetic investigators like Gilbert and Halley. In 2017, observed from about 64,000 kilometers from the Earth’s center, the dipole component of the field accounted for 99.9 percent of the field’s energy. At the Earth’s surface, it was 93.2 percent. Where the mantle meets the outer core, it was only 38.6 percent. The rest is smaller-scale, more complex magnetic components. The numbers at the core–mantle boundary most accurately reflect what’s happening inside the Earth because it is closest to where the field is produced. It means even small changes in the dipole show that the field as a whole is changing from within the core.
In addition, the reversed-flux patch on the core–mantle boundary was linked to an odd bruise of decay in the field on the Earth’s surface, like a soft spot on an apple. Again, Finlay turned to his computer. He brought up a colored map to show me. It was a two-dimensional globe in vivid greens, blues, reds, and yellows, with a spate of white dots scattered across it. This was an image of the strength of the Earth’s magnetic field at the surface of the planet, a descendent of the maps made since the times of the magnetic crusade, cousins of the ones Gubbins and his team had made of the boundary between the mantle and the core in the 1980s.
Most of it was green, which corresponded to a field strength in the range of 40,000 nT (nanoTeslas). Patches near the two poles, north and south, were dark red, indicating a stronger field closer to the range of 60,000 nT. But the curious feature was the vast expanse of blue. It stretched from the eastern edge of southern Africa across the Atlantic Ocean far to the west of South America, and from the equator almost to Antarctica. It was a region of low field strength, down in the range of the 20,000s or so. On top of the blue, and in a few other places, were splotches of white dots representing episodes when satellites flying overtop the weak field had had memory failure.
This was the South Atlantic Anomaly, named for the fact that it had a weirdly low field strength and was centered in the Atlantic Ocean below the equator. In fact, the field was so depleted that solar radiation reached close enough to the Earth’s surface to disrupt satellite technology.
Its existence came as a surprise to geophysicists. They were unaware of its impact before satellites began to transmit information, partly because there had been relatively little measuring of the magnetic field in the southern hemisphere over time. They caught the first glimpses of how it was changing with the onset of the Ørsted and CHAMP satellite data in the early 2000s. But it was when the Swarm trio started transmitting far more precise information beginning in 2015 that its influence became clear. Not only is the anomaly big, it is heading to the west, growing rapidly, and causing the field to weaken fast.
And fast not just in geological terms, but in human terms too. A paper published in 2016 found that if you define the anomaly as having a field strength below 32,000 nT, its area grew by more than half—53 percent—from 1955 to 2015. By then, it covered a little more than one-fifth of the Earth’s surface—20.3 percent—up from 13.3 percent sixty years earlier. In that same period, the very weakest part of the anomaly degraded fast too, dropping from 24,000 nT to 22,500 nT, a decline of 6.7 percent.
So then the question emerged: Was this twisting gyre, linked to an off-balance dipole that in turn produced the South Atlantic Anomaly, the long-sought mechanism for the reversal of the poles? Was it evidence that the inner Earth’s warring magnetic tribes were in the throes of toppling the dominant dipole even now? Are the poles poised to switch places?
Finlay was constitutionally unable to tell anything but the whole truth. That was the alpha and omega of his scientific training. He would love to be able to say for sure that the gyre is the key to a reversal—or that it isn’t. He would love to be able to say whether a reversal is on its way right now. But he can’t. The most definitive thing he can say is that the poles will certainly reverse at some point, as they have so many other times, and that we cannot rule out that this could be the early stages of a reversal.
The truth is that neither he nor anyone else knows what the start of a reversal looks like. There are theories. There is no consensus. There’s little detailed evidence in the rocks of precisely what was happening in the core when the earlier reversals took place. In geological terms, the reversals happen too quickly to lay down a good record of the transition from one field direction to the next. In fact, it’s not clear that rocks can always capture a magnetic signal in a highly disturbed, reversing field. Often they are only able to record the fact of the reversal itself.
There’s not even a consensus on what causes a reversal. Maybe it’s the gyre. Maybe it’s something else. When Finlay and others make their computer simulations produce a reversal in the field, sometimes it resembles what’s happening today. A patch of inverted polarity can grow and move from equator to pole and then—a flip. But in some of the simulations, the reversed-flux patches grow powerful and then the dipole beats them back and regains control. The reversal is averted.
And if this is a reversal, how advanced might it be? That depends where we are in the process and how long the process takes. There’s no consensus on any of it. A reversal has three separate phases. There’s the period when the dipole is weakening, followed by the relatively swift movement of the poles to opposite sides of the planet, followed by the regrowing of the dipole. No one knows exactly how long each of these phases lasts, or even whether they are consistent during each reversal. The general thinking is that each phase is at least centuries long, perhaps longer, and that anything quicker is implausible. Even that is still under dispute. A recent paper by the Italian researcher Leonardo Sagnotti that looked at a continuous layer of sediments in the Apennine Mountains calculated that the last time the poles reversed, 780,000 years ago, it happened in less than a century. Most other evidence suggests it took about 10,000 years from beginning to end. This is a crucial point because the concern is not primarily the reversal itself. It is the weakening of the field while the reversal is in process and how much additional radiation will strafe how close to the Earth for how long.
And while the issue of how long the dipole has already been weakening is contentious, geophysicists can say that today, it is about twice as strong as it was just before the five previous reversals. It’s been decaying by an average of 16 nT a year since the first Gaussian measurements of the field’s intensity around 1840, which is less than two hundred years. That’s a total of about 10 percent since 1840. And that’s on the surface, not at the core–mantle boundary. If the whole dipole were to continue to decay at that rate, that part of the field would be gone in less than two thousand years. But it’s unclear just how far the dipole would have to decay in order for non-dipole factions to force a pole reversal. Would it have to go all the way to zero? Or something short of that?
And, as Finlay pointed out repeatedly, the magnetic field does not function according to a linear model. It is profoundly, intrinsically nonlinear. To physicists and mathematicians, nonlinearity has a precise meaning. Linear means you can add up all the components of something and get a correct answer. So, if I’m making a cake and I double the ingredients, I get twice as much cake. Nonlinear means the answer isn’t directly proportional to the sum of the components that go into the problem. If each component doubles, the endpoint is not necessarily double. So you can solve individual parts of a problem, but when you put them together, you might not get the answer you expect. Not only that, but when components of this nonlinear system are changing, it gets even harder to figure out what the answer will be.
And then there’s chaos. Some nonlinear systems are also chaotic, which means that a tiny change in initial conditions can have a profound, unpredictable, and counterintuitive change in outcome. Just because something acted a certain way before doesn’t mean it will act that way in the future. It doesn’t mean random; these systems still follow well-defined laws. Also, a chaotic system doesn’t show a discernable pattern over time. The most famous way of explaining the idea of chaos comes from the world of meteorology. Trying to work out a way to predict weather in 1961, the American mathematician and meteorologist Edward Lorenz ran a computer simulation and then reran it, starting it in the middle of a time sequence. But he inadvertently truncated one of the numbers he input by a few decimal places. The computer program, which didn’t change, produced a dramatically different forecast. Lorenz eventually described it this way: If a butterfly flaps its wings in Brazil, does that set off a tornado in Texas? His explanation became known as the “butterfly effect.” Small changes can lead to big differences.
I asked Finlay if the core is chaotic. Maybe, he said. He paused and thought for a moment. The core is a turbulent place with strongly nonlinear dynamics. The reversals don’t show a simple pattern in time. The models show it is sensitive to initial conditions. Does that make it chaotic? It is certainly a very plausible hypothesis, he said.
The idea of nonlinear chaos has much deeper historical roots than the 1960s, though. For more than 250 years, ever since Isaac Newton presented his theory of gravity, mathematicians have been trying to solve something called the three-body problem. It goes like this: You have three particles (or heavenly bodies, originally) moving in space, connected by their gravitational pulls. You know where they are now. Tell me exactly where they will be in the future. It turns out to be unsolvable except in a few weird scenarios. You can solve it for one body and for two, but not for three. It means the movement of the bodies is nonlinear and therefore impossible to predict over time.
It’s similar in the core. You can know roughly what the core is doing now. You can know more or less what it did in the past. You can know the rules of physics the core must abide by and that they say the field’s direction must change at some point. But you can’t say for sure when it will happen. The nut of the problem is precisely that difficulty in knowing the initial conditions. Any tiny changes make for enormous differences. Added to that, the reversals are aperiodic, meaning, unlike in the sun, they don’t happen in a time pattern that anyone can discern.
This business of trying to predict the field’s future has been going on in some form or another since Henry Gellibrand’s measurements in John Welles’s garden in 1634. It has been primarily theoretical and mainly carried on in the absence of enough knowledge or data. But the efforts based on more precise information are far more recent—only a few decades old. The Swarm data, which are the most precise of the lot and which allow for more detailed calculations, only began after the satellites were launched in late 2013. This is the cutting edge.
Here’s a measure of how much of a frontier these scientists are on. One of the first things I noticed in Finlay’s office in Copenhagen was a full-color replica of the map Halley made of declination lines across the Atlantic Ocean, based on his observations on the Paramore in 1700. Halley thought that once he had mapped those lines, they would be a long-term, invaluable record of the angle of deviation at sea between geographical north and magnetic north. In fact, they were out of date almost as soon as they were published because the field is so changeable.
But today, even with the three centuries of understanding and information that have followed Halley’s work, geophysicists can only predict the movement of the field lines at the surface for as much as five years. Not more. Anything past that is soothsaying. So, every five years, geophysicists gather as a community to work out their mathematical forecast for the next half decade and publish it, making it freely available to all. Called the International Geomagnetic Reference Field, each version is out of date by the time the new one is made. The models are extraordinarily detailed. Finlay was lead author of the forecast in 2010. Knowledge of the magnetic field’s direction is an essential part of many modern navigation and orientation applications, and for industries working underground. It works for aviation during those times when GPS satellite systems don’t work well enough or fail altogether. Even smartphones rely on the magnetic models Finlay and his colleagues produce.
And this five-year forecast is for the field as a whole on the surface of the Earth. Predicting what will happen in the dynamo or to the gyre within the core is a much more remote frontier. Trying to predict what the poles will do, even tougher. And here’s the sobering truth: While geophysicists are avidly examining past reversals for clues about the next one, there’s also the possibility that every reversal acts differently, or that the dynamo itself is changing and that the clues from the past won’t solve the riddle.