Imagine standing on a tropical island in the Pacific, one of those coral atolls that barely rises high enough above sea level to avoid being inundated by passing typhoons. Now imagine standing in that exact same spot 20,000 years ago. What would be different? The waves would still be crashing against the edges of that same island, but now some 375 feet below the place you stand. The difference would be the result of water taken from the ocean and stored in ice sheets.
After almost 40 years' study of past climates, nothing amazes me more than the fact that enough water was stored in the great ice sheets of North America and Europe during the last glacial maximum 20,000 years ago to lower the level of the entire world ocean by that amount. The next time you stand at the coast looking out to sea, knowing you are looking at just a tiny part of that one ocean, try to imagine draining the world ocean by 375 feet. Or the next time you fly across the Atlantic or Pacific, taking hours to get to the other side at 400 or 500 miles per hour, try to imagine.
With all that water withdrawn from the ocean, you could walk from Ireland to Britain and then on to the mainland of Europe, or from Tasmania to Australia to New Guinea, or from Siberia to Alaska, or from the mainland of Southeast Asia through Indonesia all the way to Borneo. The modern-day Sea of Japan was then just a coastal lake. The Persian Gulf was dry land. In flatter coastal regions, the coastline was tens or even hundred of miles seaward of its present location.
Or consider how much ocean water is trapped in the ice on land. Viewed from space today, the largest feature in North America is the broad bulge of bedrock extending from the high plains of Colorado and Wyoming westward across the Rockies and the Colorado Plateau to the Sierra, Cascade, and Olympic Mountains of the far West and Pacific coast. On average, this terrain bulges about 2,500 meters (8,000 feet) above sea level. Yet, 20,000 years ago, the great ice sheet of North America formed a bulge this same size. At its crest, the ice sheet was at least 2,000 meters (6,500 feet) high, and broad, rounded domes of ice covered all of Canada and parts of the northern United States. The weight of all this ice caused the bedrock underneath to sag at least 600 meters (2,000 feet) below its present-day level, forming a bowl-shaped depression. The mass of the ice was so large that Earth's axis of rotation began to shift slowly toward the ice center.
This great feat of natural engineering, dwarfing anything humans have done, has astounded generations of scientists, and the resulting sense of awe has inspired some of them to life-long pursuit of explaining how this could have happened. For those who sensed that orbital changes might be the explanation, the first problem was to figure out which season was critical and why.
In the 1860s and 1870s, James Croll proposed that the amount of solar radiation received in winter is the critical control on the size of ice sheets. At first, this idea sounds entirely sensible: snow must fall in order for ice to accumulate, and winter is certainly the most glacial-like of the seasons. Lower amounts of solar radiation would make winters colder and longer, allowing more snow to fall, and favoring the growth of ice sheets. But on further reflection, this idea makes little sense. Croll had overlooked the power of ablation, a word that summarizes all of the processes that melt snow and ice. Had Croll lived through the warm summers here in the southern Shenandoah Valley of Virginia, he might have seen things differently. Try to imagine even the thickest snow bank in some deep mountain cove surviving the blazing Sun, midday heat, and warm rains of summer.
Even the winters around here don't let snows lie on the ground for long. In early January 1996, we had an unusual 30-inch blizzard of powdery driven snow that drifted to depths of 5 and 6 feet. This storm hit the entire east coast from Georgia to Maine and was one of the biggest storms of the 1900s. After the snow stopped, the temperature stayed cold for about a week, as the powder slowly settled to 17 inches of more compacted snow.
One evening the local forecast called for rising temperatures and rain overnight. Not long before midnight, and before the rain started, I turned on our floodlights, looked out at the deep snow, and went to bed. That night, strong rains pounded on the roof of our house, bringing me half-awake several times. Around 6 in the morning, I got out of bed and turned on the floodlights. To my disbelief, the meadow was almost completely free of snow, except for a few patches where the drifts had been deepest. Somehow, almost a foot and a half of compacted snow had disappeared in less than 6 hours. I later learned that the temperatures overnight had risen into the 50s and 60s (°F). The rain that fell was so warm that it worked on the snow like warm faucet water on an ice cube. That night, the local hillsides shed floodwaters normally seen only during heavy summer cloudbursts. Here was an example of rapid ablation even during the night and in the middle of winter! At this latitude (37°N), snow rarely covers the ground for more than a month per year each winter because of the effect of the Sun and above-freezing daytime temperatures.
To our north, of course, ablation is not so dominant. Across the populated southern tier of Canada (near 50°N), snow may lie on the ground for as much as half the year. Farther north, at latitudes above 65°N, the snow-free season shrinks to just two months. Yet in today's climate, no matter how snowy a winter may be, the summers are still warm enough to melt all the snow that falls, except for scattered patches during unusually cool summers. Even that far north, it is impossible to leave any significant amount of snow in place at summer's end so that the next winter's accumulation can add to it and start the process of building an ice sheet. This key insight—the power of summer melting—was missing from James Croll's hypothesis.
In the early 1900s Milutin Milankovitch, following up on guidance from meteorologist Wladimir Koppen, proposed that the amount of solar radiation received in summer is the critical factor that determines the growth and melting of ice sheets. He hypothesized that ice sheets would have grown during times when summer radiation was low and melted when it was high. In those far northern regions of the Arctic where snow barely melts in modern summers, a small decrease in solar radiation would allow some snow to persist through the summer in favorable locations. With more snow added the next winter, the snowfields would expand, and their bright surfaces would reflect more radiation and chill the climate further, allowing still more snow and ice to accumulate.
Milankovitch's idea that summer is the critical season initially made some headway, but most climate scientists were not convinced that he was correct. Still lacking was persuasive supporting evidence based on Earth's actual climatic history. For half a century, Milankovitch's idea remained in limbo, neither completely supported nor rejected by the evidence. He died in 1958, just as the first critical evidence testing his hypothesis began to emerge.
The first important advance was the development of radiocarbon dating by Willard Libby and colleagues in the late 1940s. Climate scientists finally had a means of dating some of the deposits left by the ice sheets. Radiocarbon dating is normally useful for ages back to about 30,000 years ago, so the piles of glacial debris (moraines) strewn across the landscapes of northern North America and Europe could be dated if they contained organic carbon, or their age could be constrained by dating the carbon-rich soil layers lying immediately above or below the rubble. Most of the glacial deposits yielded ages of 20,000 years or younger.
Other kinds of deposits confirmed this finding. Scientists working on cores taken from lakes in areas south of the former ice sheets found that pollen grains preserved in the deeper lake sediments had come from trees that were far more cold-adapted than the forests growing in those regions today (for example, spruce pollen grains in regions where oak forests grow now). Radiocarbon dating confirmed that the layers of pollen from the cold-adapted trees were the same age as the glacial debris.
Radiocarbon dating also showed that the large ice sheets of the last glacial maximum had gradually melted between about 16,000 and 6,000 years ago, consistent
with Milankovitch's theory. During that interval, solar radiation in Northern Hemisphere summers was as much as 9% higher than it is today because Earth's poles were tilted more toward the Sun, and because Earth was closer to the Sun during the northern summer. Over those 10,000 years, the ice sheets slowly retreated, ending up as small remnants in northeast Canada (fig. 4.1).
Radiocarbon dating also showed that the ice sheets had earlier reached glacial-maximum size a few thousand years after an interval when solar radiation had been considerably lower during Northern Hemisphere summers. This lag of icesheet size behind solar radiation had also been predicted by Milankovitch, who had noted that ice sheets do not grow or melt immediately but take thousands of years to respond to this solar "driver."
But a single glaciation could hardly be taken as a definitive test of the Mi-lankovitch hypothesis. Scientists needed a long record spanning many glaciations to compare with the solar radiation changes calculated from astronomy. Unfortunately, this record was not to be found in the regions where the ice sheets had existed. Ice sheets are agents of erosion that bulldoze loose material lying in their paths. Each successive glaciation destroys most of the rubble deposited by previous ice advances. The only record left largely intact is the retreat of the ice sheet from its most recent maximum 20,000 years ago (fig. 4.1). Think of a classroom blackboard used for a day of lectures. With each erasing, the record of earlier lectures is lost, except for scattered words around the edges. In a similar way, only a few fragments of previous glaciations remain untouched by later erosion, and those are usually too old to be dated by the radiocarbon method. So the place where the ice sheets existed is not the place to look for a continuous record of their history.
In contrast, ocean basins are places where deposition dominates. Shallow coastal regions are subject to wave erosion, but most deep ocean basins lie beyond the reach of even the largest storms. For this reason, in the middle and late 1900s, attention turned to ocean sediments as a possible archive of the complete history of the ice ages. Techniques were developed for dropping large, hollow pipes into the soft, sandy mud of the sea floor during oceanographic expeditions to retrieve sediment cores. In many places, these cores provided continuous records extending hundreds of thousands of years into the past.
Cores retrieved from ocean basins adjacent to the ice sheets were found to contain layers of debris delivered directly by the ice—fragments of continental rocks like granite and smaller mineral grains like quartz and feldspar. Initially, this debris had been incorporated into the ice sheets as they eroded the land. Then, as ice flowed out from the continental interiors and reached the sea, icebergs broke off, floated away, melted, and dropped the debris into the sediments on the sea floor far from its source regions. In contrast, during intervals when the ice sheets disappeared, ocean sediments were free of continental debris. These layered sediment sequences from the ocean held long and complete records of the presence and absence of ice sheets.
Ocean sediments also contain an even more important measure of the size of the ice sheets. Like other elements found on Earth, oxygen comes in several forms called isotopes, which differ slightly in mass and weight. Both ocean water and ice (liquid and solid H2O) contain oxygen-18 and oxygen-16, in amounts that vary from region to region. When ice sheets form, they take more of the lighter oxygen-16 isotope from the ocean and store it as ice, leaving more of the heavier oxygen-18 isotope behind in the ocean. Living in the ocean are hard-shelled plankton that form CaCO3 shells, and the oxygen in their shells comes from sea-water. When the plankton die, their shells fall to the sea floor and become part of the permanent record piling up over time.
The layered history of deposited plankton shells tells us about the ice sheets. When the ice sheets grow, and more oxygen-16 is taken from the ocean, the shells deposited in the ocean sediments are enriched in oxygen-18. When the ice sheets melt, the extra oxygen-16 is returned to the ocean, and the plankton living at that time incorporate it in their shells. By analyzing layer after layer of fossil shells in sediment cores, scientists can reconstruct the history of growing and melting ice sheets even in regions far from the ice.
Pioneering investigations of sediment cores by geochemist Cesare Emiliani during the 1950s and early 1960s had shown that a sequence of at least 5 glacial-interglacial alternations had occurred over an interval thought to span the last several hundred thousand years. The fact that these glacial-interglacial repetitions occurred in rather regular cycles gave a major boost to the Milankovitch hypothesis because Earth's orbit also changes in regular cycles (chapter 3). But the age of these oscillations was still uncertain because these records extended far beyond the range of radiocarbon dating.
The problem of dating these younger cycles was soon overcome by a clever but independent approach that made use of fossil coral reefs as sea-level "dipsticks." Coral reefs form at or just below sea level and leave hard skeletal structures in the fossil record. As ice sheets grow, they take water from the ocean, and when they melt they return it. As the level of the ocean moves up and down the sides of ocean islands, the living coral reefs follow, since they can grow only near sea level. Over time, the fossilized corals leave behind a history of past sea level along the island slopes, and changes in sea level can be used to calculate the volume of sea-water trapped in the ice sheets. Also coming into use during the 1960s was a method for dating coral reefs by measuring small amounts of radioactive uranium that exists in seawater and is incorporated in the skeletons of corals. Uranium allowed the corals to be dated.
Three coral reefs that formed between 80,000 and 125,000 years ago proved to be invaluable in dating the last interglaciation, a time, like today, of minimum ice volume and high sea level. The dates of these reefs matched three levels in the ocean cores where the oxygen isotopes and the scarcity of glacial debris indicated minimum amounts of glacial ice. These reefs helped to anchor the time scale of the ice-age cycles and set up a major discovery.
In 1976 marine geologists Jim Hays and John Imbrie and geophysicist Nick Shackleton published the first strong confirmation that ice-age cycles are tied to variations in Earth's orbit. Using a marine sediment record extending back more than 300,000 years, they found three major cycles of variation in the oxygen-isotope (ice-volume) record: cycles of 100,000, 41,000, and 22,000 years, precisely equivalent to the orbital cycles of eccentricity, tilt, and precession (chapter 3). In addition, the timing of the ice-volume cycles at 41,000 and 22,000 years matched the prediction made by Milankovitch: it lagged several thousand years behind the changes in Northern Hemisphere solar radiation. The story of this exciting discovery, and related efforts, is told in a book written by John Imbrie and his daughter Katherine, Ice Ages: Solving the Mystery.
41,000-year cycles 100,000-year cycles
Was this article helpful?