Glaciers That Speak in Tongues and other tales of global warming

October 2001

New Zealand mistletoes that bear strange, sealed flowers
depend on savvy native pollinators to thrive.



A century and a half ago, the world’s mountainous regions were somewhat colder than they are today. We know this because historical records—writings, paintings, early photographs—show that glaciers were larger then. In the Swiss Alps, the area for which the most detailed documentation is available, glaciers expanded during the Little Ice Age, a cold episode that ran from about 1300 to 1860. (This ice age is called “little” because, even at its worst, the cooling required to produce it was only one-tenth the cooling that induced the “big” ice age that peaked 20,000 years ago.)

Though records for glaciers in other parts of the world are less detailed, we know that in the high peaks of the tropical Andes, in the temperate-zone Andes of Chile and Argentina, and in the Southern Alps of New Zealand’s South Island, glaciers were substantially larger in 1850 than they are now. So we suspect that the Little Ice Age cooled not just Europe but the world.

The Little Ice Age ended abruptly. Starting in 1860, the world’s glaciers began a retreat that has continued right up to the present. Without a doubt, therefore, planet Earth has gotten warmer over the past century and a half. But humanity’s exact contribution to the warming is still under debate. Along with most atmospheric scientists, I take very seriously the results of computer simulations showing that human-produced forces are very likely driving the rise in temperature that we have seen over the past quarter century. Yet roughly half the overall warming since 1860 occurred before carbon dioxide (CO2) emissions from human activities had reached significant levels. Some take this as evidence that most of the current upswing in temperature is merely a continuation of the natural events that brought the Little Ice Age to a close.

To truly understand the scenario of global warming, we need to know how much Earth’s temperatures would have fluctuated in the absence of the Industrial Revolution and whether we are now exacerbating or counteracting these fluctuations. And we can know these things only if it can be shown that Earth has had predictable temperature cycles during the past several thousand years. Many scientists doubt that such large-scale regularities exist, but assuming for the moment that they do, how would we detect them? Even year-to-year changes in average global temperature are difficult to chart, despite the help we get from the worldwide network of thermometers and from the scores of satellites orbiting high above Earth. Getting a sense of relevant natural fluctuations is especially difficult because during the period that geologists call the Holocene Epoch (the 11,000 years since the end of the last ice age), Earth’s temperatures have been remarkably stable. Unlike the previous 100,000 years, when the climate underwent numerous large jumps and drifts, measured in many degrees Fahrenheit, the changes during the entire Holocene have been only 1° or 2° F—too small to ascertain with the natural climate indicators we have been using until now (such as tree rings and fossil pollen), whose accuracy is no better than 2° F.

This is why climatologists have turned to mountain glaciers. The record created by these glaciers is an excellent proxy for climate, standing in for hundreds of years of thermometer readings. Not only does this proxy tell us about past temperatures, its margin of error is less than 0.4° F.

Everywhere on Earth, the higher you climb, the colder it gets. The reason is that as air rises, it expands and therefore cools. On average, air temperature changes at a rate of 1° F for every 300 feet of elevation. At some point on the way up, air temperature reaches the freezing point. Glaciers can’t form or endure, of course, unless the air temperature remains low, and how far down a mountain they reach depends on what glaciologists call the equilibrium snowline—the boundary between an upper zone where accumulation of snow outpaces melting and a lower zone where the reverse is true. This boundary corresponds fairly closely to the altitude at which air temperature reaches the freezing point. In the Alps,
Records of glacial advances and retreats provide an accurate gauge of temperature changes.
large glaciers have tongues of ice streaming down past the equilibrium snowline and on into valleys in the lower zone. The melting in such tongues below the snowline is balanced by the downward flow of ice from above the snowline. When the air temperature shifts, such tongues can shrink or enlarge dramatically. For example, since 1860 the Rhône Glacier, in the Alps of southwestern Switzerland, has retreated about three miles in response to a rise in the equilibrium snowline of about 300 feet. This rise corresponds to an atmospheric warming of slightly more than 1° F.

Such astounding sensitivity suggests that a record of glacial advances and retreats during the Holocene Epoch would yield the information we seek about recent climate cycles. But this is easier said than done. The problem is that glaciers act like giant erasers. Each advance eradicates almost all traces of what’s come before. At their greatest size during the Little Ice Age, Europe’s glaciers covered an area at least as large as was covered in any previous Holocene advance, so the record of earlier Holocene advances is mostly obscured.

We do, however, have evidence that well before the Little Ice Age—several times during the past 11,000 years, in fact—Alpine glaciers pushed out to roughly the same position they occupied in 1850. One such indication is the size of the moraines—the looping walls of debris that mark the edge of each major Alpine ice tongue. These piles of rocky rubble are huge, some standing 300 feet above the valley floor. It is difficult to imagine that they could have been formed during a single forward push of ice; instead they appear to be the result of a great many such pushes in the course of thousands of years. Within these huge debris piles are layers representing ancient soils, indicating that many earlier advances occurred. But carbon 14 dates derived from soil materials are often misleading and so cannot be relied on to provide the precise chronology we seek.

During the last decade, however, a major breakthrough occurred, owing to the appearance of a new climate proxy: wood and peat that have been washing out from beneath the retreating Alpine glaciers. For this kind of material, carbon dating is quite reliable and allows us to determine precisely the warm periods when trees and other plants were able to grow in places that are now covered by glaciers.

In September 2000, I had the opportunity to witness a harvest of this ancient wood and peat when Christian Schluechter, a University of Bern geologist who pioneered such studies, led a small group of interested scientists to the terminus of the Unteraare Glacier in the northern Swiss Alps. He explained to us that once each year, toward the end of summer, the meltwater that has accumulated beneath the ice suddenly breaks out of its confinement and sweeps over an apron of large cobbles lying at the foot of the glacier. To our amazement, we saw pieces of wood and peat wedged here and there among the cobbles. During the single hour our helicopter taxi service allotted us for exploring the apron, our group found fifty separate pieces. Many of them showed evidence of compression, shearing, and twisting caused by the weight and motion of the overlying ice.

Our finds were not the first “warm artifacts” to be harvested from the forelands of retreating Swiss glaciers. Schluechter and his colleagues had been collecting wood and peat fragments for several years and have obtained carbon dates for nearly a hundred of them. Some of the wood dates all the way back to the earliest Holocene. Even more important, rather than being spread evenly through time, the dates for these pieces of wood and peat fall into distinct groups, with each group presumably representing a warm episode when Alpine glaciers were even smaller than they are today.

Geologists are now investigating whether these groupings correspond to another new source of evidence of cyclic patterns in Earth’s recent history. This evidence comes from studies of sediment in the deep waters of the North Atlantic. The rock fragments in these sediments are much too large to have been transported there by ocean currents; they could have reached their present location only by having been frozen into large icebergs that floated long distances from their point of origin before melting. During the past decade, Gerard Bond, my colleague at Columbia University’s Lamont-Doherty Earth Observatory, has studied the makeup of such ice-rafted debris. Noticing that some of the sediment grains were stained with iron oxide, he reasoned that they must have come from locales where glaciers had overrun outcrops of red sandstone. Bond concluded that a detailed analysis of deep sediment cores would reveal changes in the mix of sediment sources over time. This proved to be an excellent strategy, for Bond found something so unexpected that it stunned all of us who study climate history. The proportion of these red-stained grains fluctuated back and forth over time from lows of 5 percent to highs of about 17 percent, and these fluctuations had a pattern: a nearly regular, 1,500-year cycle. Even more amazing, he found that the cycles ran virtually unchanged, in both amplitude and duration, through both ice-age and non-ice-age periods during the last 100,000 years.

Bond puzzled over what might be pacing this cycle. As a geologist, he knew that the sources of the red-stained grains were generally closer to the North Pole than were the places yielding a high proportion of “clean” grains. At certain times, apparently, more icebergs from the far north were making their way well to the south before finally melting and shedding their sediment. Bond hypothesized that the alternating cycles might be evidence of changes in ocean-water circulation.

Ocean waters are constantly on the move, and water temperature is both a cause and an effect. As water cools, it gets denser and sinks to the bottom. In one part of what I like to call the “bipolar seesaw,” the bottom layer of the world’s oceans comes from cold, dense water sinking in the far North Atlantic. This causes the warm surface waters of the Gulf Stream to be pulled northward, as they are today. Bond realized that during this part of the ocean cycle, a large proportion of the icebergs that bear red grains would melt while still fairly far north. But sometimes the ocean reorganizes itself, and the Southern Hemisphere holds sway in driving ocean circulation. At such times, surface waters in the North Atlantic would generally be colder, permitting icebergs bearing red-stained grains to travel farther south before melting and depositing their sediment.

The onset of the Little Ice Age in about 1300, which followed the so-called Medieval Warm Period of the eighth through tenth centuries, may represent the most recent time that such a switchover occurred. The contrast in the North Atlantic is apparent if we consider that in the tenth century, Erik the Red and his band of Vikings colonized the lands surrounding the fjords in southwestern Greenland. Not only did the Vikings navigate their wooden vessels back and forth between Scandinavia and Greenland without being thwarted by sea ice, but they were also able to grow enough grass to support sizable flocks of sheep. As time went on, however, conditions deteriorated. The last recorded communication from the colonists occurred in the early fourteenth century—just at the onset of the Little Ice Age in the Alps—and eventually the colony died out (see “The Vikings’ Silent Saga,” Natural History, November 2000).

Further evidence of the impact of the Little Ice Age comes from records kept by Icelanders, whose writings indicate that between 1650 and 1850, their island was icebound for several months each year—a great hardship, since fishing was a main source of sustenance. They reported with pleasure that the ice began to wane in 1880, permitting them to extend the fishing season. Readings from their thermometers (which they began to use in about 1870) also suggest that the mean annual temperature was rising.

Could it be that these ocean oscillations, because of their effect on air temperature, also explain the snowline fluctuations seen in the Swiss Alps? So far, none of the wood or peat fragments sluiced from beneath the ice have yielded carbon 14 dates from the eighth through the thirteenth centuries, which would correspond to the Medieval Warm Period and the interval leading up to the Little Ice Age. But another source of evidence demonstrates that Alpine glaciers were smaller during this time. Medieval farmers living below the huge Aletsch Glacier, in what is now south-central Switzerland, constructed a crude aqueduct of hollowed-out larch tree trunks to carry water from a small mountain lake down to a village. We know from written records that parts of this aqueduct had to be rebuilt after being overrun by the 1350 advance of the Aletsch Glacier.

If future dating of wood and peat expelled from beneath retreating glaciers in the Alps and other mountainous regions worldwide supports Bond’s 1,500-year cycles, and if the Medieval Warm Period/Little Ice Age oscillation can be shown to be part of the most recent of these cycles, we will have taken an important step toward establishing a major natural rhythm in Holocene Epoch temperatures. This rhythm could then be extrapolated into the future. Because the midpoint of the Medieval Warm Period was about A.D. 850, an extension of Bond’s cycles would place the midpoint of the next warm interval in the twenty-fourth century.

While offering a useful basic framework, this pattern alone does not account for all aspects of past fluctuations and thus is not a sufficient predictor for the future. For example, we usually think of cycles as having a regular, bell-like shape. But during the millennia corresponding to the last ice age, Bond’s 1,500-year cycles were closer to rectilinear, indicating sudden starts and stops. As is clearly recorded within the deep layers of Greenland’s ice, the transitions often took just a few decades. This abruptness was especially pronounced as the climate warmed. Has this also been the case during our own epoch, the Holocene? If the Holocene’s “Bondian” cycles, too, have been rectilinear, one would expect the post-Little Ice Age warming to have been completed within a few decades. One would also expect that in the absence of the Industrial Revolution, global temperature would have stabilized for a warm plateau of several hundred years. But in fact, global temperatures in the decades immediately after the Little Ice Age did not simply jump to a new plateau.

Also, studies of Alpine glaciers show that the Little Ice Age had three cold peaks, in about 1350, 1650, and 1850. Do we have any clues about the causes of these additional, smaller fluctuations? I pondered this problem in the early 1970s. At the time, Holocene climate records were few and far between. In fact, only one—a 70,000-year record obtained from a one-mile-long, four-inch-diameter ice core drilled in northern Greenland—had enough length and detail to provide any clues. Danish paleoclimatologist Willy Dansgaard and his colleagues had managed to obtain paleotemperature results from samples of this core and had concluded that much of the variation in climate could be accounted for by a combination of 80-year and 180-year cycles, which they thought reflected periodic fluctuations in the Sun’s energy output. When I merged Dansgaard’s pattern with that of the warming expected from the steady increase in man-made greenhouse gases, I obtained a composite that matched the major features of the actual global thermometric record—namely, a warming phase extending from 1860 to the time of World War II and followed by a thirty-year pause. I predicted that when, in the near future, Dansgaard’s natural cycle turned from a cooling into a warming phase, the natural and the man-made factors would join forces and produce a prominent renewal of the warming trend. My warning was published in the journal Science in 1975. In 1976 the thirty-year plateau came to an end, and the warming that then began has continued right up to the present.

My prediction was correct, but was it soundly based? Ten new ice-core records—from Greenland, from Antarctica, and from high-mountain sites elsewhere on the planet—are now available. None show Dansgaard’s combined 80-year and 180-year cycles. I thus have been inclined to write off the success of my prediction as just a happy accident. Still, the changes we are attempting to document have a magnitude of only a few tenths of a degree, and perhaps in most records they are masked by regional climate change.

What drives Bond’s 1,500-year cycle? High on the list of possibilities must be the Sun. But sufficiently accurate, satellite-based measurements of solar activity cover only the past twenty years—not long enough to warrant our drawing firm conclusions. We do, however, have a longer-term record of the number of dark spots resulting from magnetic storms on the Sun’s surface. These spots usually wax and wane in an eleven-year cycle. During an interval known as the Maunder Minimum (A.D. 1650-1710), however, no sunspots were observed. Minze Stuiver, an isotope geologist at the University of Washington, has shown that more carbon 14 atoms were created in our atmosphere during the Maunder Minimum than either before or after it. The reason, Stuiver postulated, is that the electrically charged particles streaming out of sunspots generate a magnetic field that deflects incoming cosmic rays from our solar system. During periods of low sunspot activity, such as the Maunder Minimum, however, the magnetic shield is turned off and more cosmic rays bombard Earth’s atmosphere, manufacturing extra carbon 14 atoms. Having demonstrated the link between sunspots and carbon 14 production, Stuiver was then able to use carbon 14 measurements on tree-ring-dated wood to deduce sunspot minimums prior to the invention of the telescope. His data indicate that these occurred at roughly two-century intervals and might be the cause of Dansgaard’s 180-year cycle. Stuiver’s record contains no hint of Dansgaard’s 80-year cycle or of Bond’s 1,500-year cycle, however.

If not the Sun, then what might be the driver of Bond’s cycle? For me, the top candidate remains the ocean’s bipolar seesaw. In the global ocean circulation pattern we are accustomed to seeing, sinking cold surface water in the North Atlantic is replaced by warm Gulf Stream waters, which are drawn northward. What may underlie this pattern is actually salt, since an extra gram of salt per liter makes seawater denser by an amount equivalent to a cooling of 8° F. The North Atlantic is unusually salty, because the location of mountains and the direction of prevailing winds lead to the export of water vapor from the Atlantic to the Pacific Ocean. The salt left behind makes that surface water not only saltier but denser, and the water sinks in the familiar pattern.

Meanwhile, acting like a giant conveyor belt, the normal circulation of the world’s oceans transports excess salt out of the Atlantic Ocean. This process may not occur at a steady rate. Rather, it may oscillate, leading to Bond’s
Are underlying natural cycles retarding or reinforcing human-induced global warming?
1,500-year cycle. As I see it, at times the conveyor mechanism carries salt away faster than it is being replaced. This may lower the water density in the North Atlantic enough to make the conveyor shut down. At this point, the warm Gulf Stream weakens, and the circulation pattern reorganizes itself under the domination of the Southern (Antarctic) Ocean. With the new regime, salt ceases its rapid departure from the North Atlantic, and salt levels there begin to rise again. When the northern waters are again dense enough to rejuvenate deep-water formation in the North Atlantic, the cycle is complete.

The details of this mechanism are still not clear, but what is clear is why such a cycle might lead to warmings and coolings of the lands surrounding the North Atlantic. When the north rules, an enormous amount of heat—as much as would be generated by a million large power plants—is carried north of the Strait of Gibraltar by the Gulf Stream. This heat is released into the atmosphere in winter and is carried to northern Europe by the prevailing westerly winds. When the south rules instead, this source of heat is lost and the Alps cool.

But how does any of this explain the cooling in the Southern Hemisphere—the Little Ice Age in New Zealand’s Southern Alps and in South America’s Andes? Instead of being limited to the land area adjacent to the North Atlantic, this cold episode appears to have affected much of the planet. There must have been some link between patterns of ocean circulation and conditions in the atmosphere that affected both hemispheres. Although the nature of this link has not been discovered, it probably involves Earth’s “tropical heat engine”—the way air rising from the equator fuels the atmosphere with heat and water vapor.

So where do we stand? For a start, certain regular fluctuations in the Holocene climate seem to occur, but scientists are still left with many uncertainties about them. This prevents us from making a meaningful prediction concerning how the climate would have changed in the absence of the Industrial Revolution. We cannot prove the existence of either the 1,500-year cycle—which the available Greenland ice cores fail to record—or the combined 80- and 180-year cycles. However, we can state with some confidence that natural Holocene temperature fluctuations have been on the same scale as the human-caused effects estimated to result from greenhouse gases. Hence, we cannot assume that in the absence of human intervention, Earth’s temperatures would have remained stable.

Unfortunately, we cannot even say whether natural changes are at this point retarding or reinforcing human-induced greenhouse warming. The situation will be much clearer two decades from now, however, as computer simulations predict an additional 1.5° F warming by the year 2020. If such an increase in global temperatures occurs, there will not be any doubt: natural causes alone would not have been sufficient to account for it.

Does this mean we can all sit back, do nothing, and wait for the results to roll in? Certainly not. In twenty years, we may well conclude that we must stem the rise of CO2, and if so, we’ve got a lot of preparation to do. Very likely, fossil fuels will remain our primary source of energy. With more people and a higher standard of living in the less developed
We may have to remove CO2 from power-plant exhausts and even from the atmosphere itself.
countries—and with even a sustained per capita level of demand in the United States—energy use will at least double by the year 2050. We must learn how to remove CO2 from power-plant exhausts and probably from the atmosphere itself. (Estimates of the cost of CO2 removal suggest that it would add a few cents per kilowatt-hour to the cost of electricity or about forty cents per gallon to the cost of gasoline.) Plus, of course, the CO2 we remove must be put somewhere: stored in the deep ocean or in deep saline aquifers or, if we want to be sure it never comes back to haunt us, converted into magnesite (MgCO3), a geologically inert mineral. More difficult than the technical aspects are the political ones. Ready or not, we will have to face them all.

To strengthen their case, corporate spokespersons, avid consumers, and plenty of other people and institutions inclined to dismiss the ongoing rise in atmospheric CO2 as inconsequential may be happy to latch on to the paleoclimatic reconstruction presented here. This would be unfortunate. Unless all the work done on climate simulations and fossil-fuel-use projections is seriously flawed, one thing is certain: our planet will indeed experience a major human-induced warming during this century.

We have learned that Holocene temperatures have undergone natural fluctuations, but the causes of these changes are so subtle that we have yet to figure them out. Apparently, our climate system responds to even tiny nudges. This being the case, the potential effects of human activities should not be underestimated. If we continue along a business-as-usual energy course, we’ll be giving the climate a large shove.

A recent visit to the Alps provided Wallace S. Broecker with a firsthand glimpse of new evidence of past climate changes. Guided by Swiss glaciologists, he and the others in the visiting field team found fifty samples of ancient peat and wood that had washed out from beneath the glacial ice—signs that vegetation once grew in areas now frozen over. The Newberry Professor of Earth and Environmental Sciences at Columbia University's Lamont-Doherty Earth Observatory, Broecker is the author of a text on the evolution of Earth, How to Build a Habitable Planet (Columbia University Press, 1998). Among his previous articles for Natural History is “Global Warming on Trial” (April 1992).

Copyright © Natural History Magazine, Inc.

Return to Web Site Archive