by Curt Stager
Warm water on the surfaces of the tropical oceans loses moisture to evaporation, which makes it saltier than average seawater. When the Gulf Stream flows from the hot latitudes between West Africa and the Caribbean into the cooler North Atlantic, it doesn’t easily mix with those northern waters because its tropical heat content makes it less dense (warming makes both water and air expand). But the Gulf Stream gradually releases much of that heat into the cooler air over the North Atlantic, and when it finally does chill down its extra load of salt leaves it denser than usual.
That extra density makes some of the Gulf Stream water sink beneath the surface and continue its riverlike meanderings at greater depths. By the time it resurfaces, the deep flow has wormed its way around the southern tip of Africa and entered the Indian and Pacific oceans. Back on the surface again, the current recurves back across those oceans, rounds the tip of South Africa, and returns to the North Atlantic, picking up new loads of equatorial warmth along the way. Additional branches also operate in the Southern Ocean and Arabian Sea, adding extra loops to the tortuous path of the global conveyor.
There’s a lot more to the picture than that, however, and when illustrations of this common version of the THC concept appear in professional slide presentations, they can become what one speaker at a recent meeting of the British Royal Society called “oceanographer detectors,” because they make specialists in the audience “go visibly pale at the vast oversimplification.”
The THC model is not so much wrong as incomplete. Most scientists have now switched the focus of ocean-climate discussions to the more comprehensive MOC formulation because temperature and salinity aren’t the only drivers of ocean currents after all; winds and tides are at least as influential. THC-style flow does occur, but midlatitude westerly winds and tropical easterly trades do much of the actual pushing.
So why does marine MOC affect climate? As heat rises into the air from the Gulf Stream, it warms the westerly winds that blow toward Europe. Without those ocean-tempered winds, London might be as cold as … well, look at a map to see what lies at the same latitude on the opposite side of the Atlantic, and you’ll find snowy Labrador.
With this basic introduction to the topic, you’re already well enough equipped to take a pot shot at The Day After Tomorrow. The prevailing winds over Manhattan blow offshore toward the Atlantic, not from it, so why should a Gulf Stream shutdown freeze the city? The film also unrealistically subjects Europe to severe winter conditions year-round. Even if it really did become a climatic equivalent of Labrador, northern Europe would still warm up quite a bit in summer, just as Labrador does.
In reality, a MOC slowdown alone couldn’t turn Europe into a climatic twin of Labrador because it lies downwind of a temperature-modulating ocean rather than the interior of a continent. And because prevailing winds spin the North Atlantic surface current system clockwise regardless of what the salinity or temperature of the water is, some version of the Gulf Stream will exist as long as these winds continue to blow over it.
Although some computer models do simulate moderate conveyor slowdowns in a warmer future, a truly severe disruption would require extremely large floods of freshwater to pour into the sea, presumably from the melting of land-based ice. If, say, a major ice sheet were to slide off into the North Atlantic where some critical sinking zone is operating, then perhaps it might cap the ocean off with dilute, buoyant meltwater.
In 1999, oceanographer Wallace Broecker published a striking theoretical description of just such a total MOC collapse under perfect-storm conditions. Tundra replaces Scandinavian forests. Ireland becomes the climatic equivalent of Spitsbergen, an island in the Norwegian Arctic. When climate modelers working at Britain’s Hadley Center several years ago told their computers to “kill the MOC,” the virtual air outside their lab cooled by 8°F (5°C) within ten years, at least on the digital screen.
But Broecker maintains that such a scenario is unlikely today, because those theoretical events only played out in a world that had already been cooled by a prolonged ice age. Nowadays, however, we don’t have nearly enough readily meltable ice left in the Northern Hemisphere to do the job. To reset that stage we’d have to cover Canada, northern and central Europe, and Scandinavia with thick ice caps, and that would require colder, rather than warmer, conditions in the future.
Most computer models that have been upgraded so they more accurately represent the role of winds in ocean circulation foresee little, if any, cooling in the North Atlantic region from MOC disruptions during the Anthropocene. As the latest Intergovernmental Panel on Climate Change (IPCC) report concluded, “it is very unlikely that the MOC will undergo a large abrupt transition during the 21st century,” and most experts believe that future greenhouse warming will overwhelm any minor regional effects related to MOC. In light of such findings, Broecker has tried to tamp down some of the worst exaggerations of the ocean-climate link that have been made by nonspecialists, but it’s a tough struggle that pits scientific restraint against the lure of a good story.
One case in point is a study commissioned by the U.S. Department of Defense that presented a wildly extremist view of MOC collapse as a grave and imminent threat to national security. In their 2003 report, the authors noted that they were presenting only the most severe of all possibilities, as is commonly done in military planning circles, but that disclaimer was easily missed amid the frightening scenarios that followed. In their depictions, global average temperature shoots up faster and faster until, in 2010 AD, the MOC begins to collapse. Less than ten years later, according to their model, northern Europe cools by 5 to 6°F (3°C), devastating drought strikes the United States, and “a cold and hungry China peers jealously across the Russian and western borders at energy resources.”
In response, Broecker wrote an open letter for publication in Science that expressed his dismay over the hyperbole. “I take serious issue with both the timing and the severity of the changes proposed,” he wrote, pointing out that such extreme changes would take a long time to develop and would require glacial-type conditions, not global warming, to trigger them. Furthermore, he cautioned that computer models still can’t fully reconstruct complex MOC disturbances of the past, much less those of the future. He concluded his letter with this admonition: “Exaggerated scenarios serve only to intensify the existing polarization over global warming.”
Nonetheless, the idea of a total collapse of MOC is so emotionally gripping that it has become firmly lodged in public consciousness. In that context, oceanographers are watching closely for signs of conveyor responses to modern warming, just to be on the safe side. In 2005, for example, a team of British researchers described a 30 percent slowdown in MOC flow since 1957. The news caught fire among lay and professional audiences alike, but follow-up studies found that the slowdown alert was a “false alarm,” as Richard Kerr put it in a deflating news brief for Science. The pattern of MOC flow is extremely variable, and a more careful look at the numbers showed that the reported trend was indistinguishable from random fluctuations.
If MOC changes are unlikely to freeze Europe, will ice ages play any role at all in an Anthropocene world? The answer is a qualified yes, but if they do reappear, it won’t be the fault of ocean circulation disruptions. The main drivers of large-scale glaciation are the movements of Earth itself as it makes its annual elliptical dance around the sun.
The way many descriptions read in the popular media, you’d think that stopping our greenhouse gas emissions would prevent climatic change altogether. In fact, climates will always change whether we exist or not, just as they have even on Mars where similarly periodic frosts, thaws, and floods have left their signatures in deposits of red sand, gravel, and dust. Fortunately, because much of that change is cyclical, we can predict some of it by letting it play out on computer screens.
The fastest moves in the planetary dance are wobbles. Imagine a whirling top as it slows down and progressively loses its balance. The top begins to dip and swing around in p
rogressively wider circles as it spins more slowly on its axis. Earth does that, too, though not because it’s going to tumble over any time soon. In the wobble cycle, the North Pole draws a full loop roughly every 21,000 years (technically speaking, there are actually two modes of the cycle, 19,000 and 23,000 years long).
This has climatic effects because it changes how sunlight intensity, or insolation, affects different parts of the planet’s curved surface. Each year, winter comes to the Northern Hemisphere when it leans away from the sun, and northern summer reigns when it leans toward the sun. Every 21,000 years or so, the wobble cycle brings the Northern Hemisphere its annual dose of summer only when we’re farthest away from the sun on our egg-shaped orbit. When that happens, northern summers become slightly cooler than usual and less snow melts as a result.
Keep in mind here that these events don’t happen because the sun itself changes. Instead, it’s because sunlight affects the seasons and hemispheres differently over time. Those effects are amplified by geography because most major landmasses are crowded into the northern half of the globe and dry, solid ground accumulates ice sheets more readily than oceans do. For these reasons, ice ages typically begin in the Northern Hemisphere; the last one was born in the ice-friendly insolation target areas of northeastern Canada and northwestern Eurasia.
But that’s not the whole story. Several longer cycles also influence the comings and goings of monster ice sheets.
As Earth wobbles, it also tilts more or less steeply in ways that amplify seasonal temperature differences. The slower tilt cycle takes 41,000 years for the North Pole to rock back and forth between 22.1 and 24.5 degrees of arc. When this cycle tilts the planetary axis less steeply, it aims each pole less directly at the sun during its seasonal summertime, so it warms less than usual. For reasons as yet unknown, this was the dominant pacemaker of ice age recurrences until about 1 million years ago, when the wobble cycle and a third, even slower cycle joined forces with the tilt cycle.
That third pulse, the eccentricity oscillation, changes the shape of Earth’s orbital route around the sun. The path becomes more or less egglike over the course of about 100,000 years; it also changes in other ways every 412,000 years. Because the sun sits off center within that ring, snuggling a bit closer to one end of the oval than the other, Earth’s distance from the sun varies a great deal through the seasonal circuit, and distortions of that route caused by the eccentricity cycle accentuate those changes. When the orbital ring is most distorted, a good deal less solar heat reaches us at the farthest end of the egg.
As all these cycles operate at once, they interact in ways that are easily visualized by comparing them to water waves. I first learned about this from my friend and colleague, ice core researcher Paul Mayewski, who directs the Climate Change Institute at the University of Maine in Orono. He explained how most of the ragged jumps and wiggles in his polar climate records originate.
“It’s like waves on a lake,” he began. I imagined orderly ranks of swells rolling along under a brisk breeze. “The rising and falling of the main swells are like the slow eccentricity cycle. Now imagine that a motorboat wake joins the pattern. Those waves are smaller and closer together, so they don’t line up perfectly with the larger, wider ones.”
I envisioned an irregular, bumpy surface like the ones that I often encountered during the waterskiing days of my youth. It was easy to keep my balance on a single predictable wave pattern, but when the driver doubled back on our trail or another boat dragged its own chop across our path, it spawned a crazed tangle of leaping, plunging waves. Where two crests briefly collided they bounced skyward, and where two troughs briefly met they bounced lower. Add forward and crosswise motions to such a collision zone and, as I can personally attest, a fallen water-skier will bob around in it like a cork.
“That’s where a lot of long-term climate variability comes from,” Mayewski continued. “When different cycles occur simultaneously, they harmonize and strengthen each other sometimes and they weaken or cancel each other at other times. And the more cycles you mix into the climate system, the more erratic it becomes.” Mingling like waves through the ages, Earth’s insolation cycles account for a surprising amount of natural climatic instability, and when they produce an exceptionally long and low temperature combination they can trigger a full-scale ice age.
These cyclic patterns were worked out by James Croll, a nineteenth-century Scottish scientist, and later refined early in the twentieth century by a Serbian civil engineer, Milutin Milankovitch. The fit between theory and history isn’t perfect, and many of the past’s shorter temperature perturbations had other causes. Furthermore, we still don’t know exactly why northern glaciations have affected so much of the world at once; the slow gyrations of Earth tend to cool northern summers while warming southern ones, so you might expect ice ages to affect only half of the planet at a time. But the basic hypothesis of insolation-driven ice age pacing is still well supported in the geologic record. Take, for example, the history that was recently revealed by one superlong ice core from Antarctica. It represents 800,000 years of climatic change, and it also captures eight glacial cycles, reasonably in line with the 100,000-year rhythm of the eccentricity cycle.
Croll and Milankovitch worked with paper and pencil and devoted countless hours to hand-calculating past insolation values, a mind-numbing task that computers now repeat in seconds. But most interesting in the context of future Anthropocene climate are the models that run those cycles forward in time. By tracking Arctic insolation patterns into the future, we can predict when ice ages should come back to haunt us again—or, rather, when they would occur in the absence of our carbon pollution.
The wobble cycle is now trying to make northern summers slightly cooler than usual, which is one of the preconditions for launching a new ice age. And if we think of the last 11,700 years of postglacial conditions as just the latest in a series of similar interglacial warm spells, then common sense suggests that we’re about due for a new cold snap.
During the 1960s, a temporary global cooling episode and a few scientists who incorrectly attributed it to the return of icehouse conditions launched a brief flurry of “fear the global cooling” excitement in the media. That response, of course, was misguided because the decadal time scale under consideration was far too short to represent Milankovitch-scale cycles, and what at the time may have felt like the start of a neoglaciation was in fact merely a brief pause in the twentieth-century warming trend.
But we can do much better than that today. Well-trained specialists now devote their careers to calculating the patterns and climatic effects of orbital cycles. Among the most influential of these are André Berger and Marie-France Loutre, a pair of climatologists at the Institute of Astronomy and Geophysics in Louvain-de-Neuve, Belgium. Their charts and tables of insolation data are widely cited in the scientific literature, and the graph shown here was drawn with data that were kindly provided by Dr. Loutre.
Solar insolation at 60 degrees North, showing an anticipated cooling event ca. 50,000 AD that might have triggered the next ice age in the absence of lingering fossil fuel emissions. Data courtesy of Marie-France Loutre.
So, when is the next ice age scheduled to arrive? A Berger-Loutre chart that appeared in Science in 2002 lays it out clearly. Today, the Arctic has entered a phase of relatively weak summer insolation that, in earlier times when the eccentricity cycle was stronger, might have been deep enough to call another ice age down upon us within the next few thousand years. But this episode will be too mild to do the job, even without our greenhouse gases working against it. In other words, we’ve just dodged an icy bullet, thanks to a slight, temporary diversion of our route around the sun.
Berger and Loutre’s work also shows that the effects of the 412,000-year eccentricity cycle have been weakening in recent millennia, which helps to flatten out the depths of cooling pulses. The more symmetrical our orbit becomes as a result of changes in that longest of cycles, the narrower the range of s
easonal temperatures on Earth. This will set an even calmer tone for cyclic climate changes in the more distant future of the Anthropocene.
Roughly 25,000 years from now, our orbit will be about as close to circular as it ever gets, and the power of the other cycles to warm or cool the ice masses of the Arctic will be as weak as it ever gets. Under those conditions, we’ll still face brief regional disturbances, such as the North Atlantic Oscillation, El Niño, and the sheer orneriness of local weather that we know today, but they’ll only raise short-lived climatic ripples on the flattened surface of the longer orbitally driven patterns. With the gentle urging of slightly higher axial tilt than we experience today, the curve of northern summer insolation will simply add a very mild rise to a puny peak—more like a low hilltop—about 25,000 years from now.
Although that low blip will cause only a mild thermal boost, it nevertheless does mean that Arctic summers were already slated to warm very slightly in coming millennia even without our influence. If Berger and Loutre are correct, then this may be unwelcome news to fans of low sea levels and heavy ice cover on Greenland.
According to the calendar of orbital cycles, the next serious risk of an ice-spawning chill during Arctic summer isn’t due until about 50,000 AD, but here is where we humans enter the scene. Most climate models will only trigger an ice age at that distant point in future time if the CO2 content of the atmosphere is no higher than 250 parts per million (ppm).
With CO2 concentrations as of this writing at 387 ppm and rising, they clearly won’t be going back below that critical threshold for a very long time. In fact, as will be explained in Chapter 3, those values will still hang well above 250 ppm in 50,000 AD, and they won’t return to such preindustrial conditions for tens of thousands of years beyond that.