Frozen Earth: The Once and Future Story of Ice Ages (28 page)

Figure 20.
This qualitative graph, based on oxygen isotope analyses of fossils from deep-sea sediment cores, shows how ocean water temperatures—and presumably also average Earth surface temperatures—have varied over the past 60 million years.
It is obvious that there has been a gradual decrease since about 55 million years ago, with especially sharp drops between 40 and 35 million years ago, and again during the past few million years.
These are the times when glaciation began in the Antarctic and the Northern Hemisphere respectively.

Parameters such as the oxygen isotope composition of shells are generally referred to by geochemists as proxy indicators, or simply proxies, because they don’t measure temperature or climate change directly.
Rather, a proxy has to be translated into some environmental variable of interest through an understanding of how the proxy itself responds
More and more proxies, many of them chemical and most of them measured in ocean sediments, are being added to the geochemists’ arsenals.
Nevertheless, oxygen isotopes remain one of the most valuable and informative proxies available to us.
Even to a skeptic, the pattern of 100,000-year cycles in oxygen isotope compositions that characterizes the past million years or so, and that shows up again and again in cores from throughout the world’s oceans (figure 16), is striking.
It is hard to believe that the correspondence of these variations with the 100,000-year cycles of eccentricity of the Earth’s orbit around the sun is purely coincidental.
Nevertheless, as implied in chapter 7, interpretation of these changes is not entirely straightforward, because at least two separate factors influence the oxygen isotope composition of shells.
One is the temperature at which the shells grew, something that is reasonably well understood, because it has been calibrated by both laboratory studies and measurements of samples from nature.
The second is the oxygen isotopic composition of the seawater, which depends on the amount of glacial ice that existed on the continents when the shells grew.
Why this should be so may not seem obvious at first, but the reason is quite straightforward.
Evaporation from the oceans preferentially causes one of the oxygen isotopes to be enriched in the water vapor, leaving behind liquid water that is depleted in that same isotope.
The oxygen isotopic composition of the ocean is therefore changed by this process.
If the evaporated water is precipitated at high latitudes and ends up as glacial ice, the ocean’s isotopic composition remains changed until the ice melts and the water again returns to the sea.
The larger the volume of the ice on the continents, the bigger the shift in the ocean.

Fortunately, both colder temperatures and larger ice volumes change the oxygen isotopes in the same way, so the overall effect is to amplify the glacial-interglacial variability in the oxygen isotope record.
This is fine for getting a general picture of the climate variations, but it would also be useful to disentangle the two effects.
Recently, some exciting progress has been made on this problem.
The approach has been to use a newly developed proxy that is independent of the oxygen isotope
variations to determine past seawater temperature.
With this knowledge, the expected temperature component can be subtracted from the oxygen data, leaving a residual record that should be due only to changes in the volume of continental ice sheets.
Two very interesting insights have emerged from this procedure.
First, it appears that ocean surface water in the tropical Pacific warmed up by 3–4°C during the most recent deglaciation, a much larger increase than earlier data had suggested (this also means that the tropical ocean had been much cooler during the glacial interval than earlier suspected).
Secondly, the data suggest that there was a lag of two to three thousand years between the temperature increase and the decrease in ice volume.
This is not very surprising—think about how long it takes for a bag of ice cubes to melt, or an old-fashioned refrigerator to defrost, even when the ambient temperature is far above freezing.
But intuitive or not, this information could not have been gleaned from the oxygen isotope data alone—the simultaneous use of more than one proxy was the critical step.
The same approach applied to the previous deglaciation, about 120,000 years ago, gives similar results.
This new knowledge paints a detailed picture of
how
the glacial periods ended, not just when, and it suggests that warming in the tropics plays a crucial role in ushering in an interglacial interval.

And the story gets even better.
Although ocean sediment cores have provided an array of proxies that can be used to track ice age climatic changes, cores of the ice itself have added an entirely new dimension.
Just as the sediment cores do, they hold a number of proxies for ice age climate, but in addition they contain direct information about environments in the past.
Entombed in the ice are air bubbles, samples of the ancient atmosphere that, in spite of their very small size, can be analyzed for many different constituents, even those present in trace amounts.
Of special interest are the greenhouse gases that can trap the sun’s energy and raise the planet’s temperature.
Measurements of these tiny time-capsule air bubbles have revealed that greenhouse gas concentrations in the atmosphere varied approximately in step with the
cycles of glaciation and deglaciation.
But deciding whether greenhouse gases are implicated as a cause of glacial-interglacial temperature variations or are simply a result requires very accurate timescales for both ice and ocean sediment cores, so that their respective records can be compared.
Such accuracy is difficult to attain with conventional dating methods.
But in recent years, an ingenious approach to this problem has been devised, based on the fact that the timing of the Earth’s orbital cycles is very precisely known.
If they really are the root cause of climate cycles, they can be used as a kind of template to examine the changes observed in deep-sea sediments or glacial ice.
All that’s required is to date one or more levels in the cores accurately and fix them relative to the orbital cycles—a procedure referred to as “tuning.”
Analyses of this type show that the CO
₂
content of the atmosphere (based on ice-core measurements) and the Earth’s surface temperature (based on oxygen isotopes in ocean sediments) both changed in sync with the 100,000-year eccentricity cycle of the Earth’s orbit around the sun.
These same analyses also show, as we saw earlier, that the volume of ice on the continents lags behind the temperature change—apparently by a few thousand years.
The very close correspondence between temperature and CO
₂
variations suggests that somehow carbon dioxide in the atmosphere is regulated by changes in the eccentricity of the Earth’s orbit, and that it in turn regulates temperature.
How this occurs is currently unknown.
But these results once again confirm the Croll-Milankovitch theory—the regular orbital cycles act like a metronome, ticking out the rhythm of the planet’s climate cycles.

As should be obvious by now, examination of ice cores is an important part of research into the Pleistocene Ice Age.
But it is a fairly recent endeavor.
The story of ice coring, especially in the Greenland ice sheet, has been nicely told in a recent book titled
The Two-Mile Time Machine
by Richard Alley, a scientist at Pennsylvania State University who has been deeply involved in that effort.
In the following, I trace the impact of ice-core science on our understanding of ice ages, drawing on both Alley’s book and other sources.

Coring a glacier, especially in the extreme climates of Greenland or the Antarctic, is no simple matter.
Equipment has to be brought in, workers have to be housed and fed, and the cores, once collected, must be stored at temperatures well below freezing.
If you hadn’t thought about it very seriously—or even if you had—you might question why anyone would want to go to such effort just to collect a bit of ice.
Part of the answer is the pure curiosity of mankind; it’s like climbing a hill to see what’s on the other side.
A feature of glaciers that must surely have piqued the curiosity of many who observed them over the years, and that certainly played a part in the desire to core into them, is their visible layering.
Like the layers of sedimentary rocks, the layers of ice in a glacier record the passing of time, and like the pages of a diary, each layer contains clues about what happened in the past.
The information is cryptic, but with the right tools it can often be deciphered.

Early attempts to recover stratigraphic (i.e., layer-by-layer) information from ice date back to 1957, when the International Council of Scientific Unions, a body that coordinates international activities in science, launched the International Geophysical Year (IGY) to promote geophysical research on a global scale.
By its nature, geophysics is international, and under the aegis of the IGY, scientific projects were conducted quite literally from pole to pole.
Sixty-seven different countries had official roles, and, in spite of its name, IGY lasted for eighteen months.
An important focus was the polar regions, which at that time were still poorly known scientifically.
Both the Arctic and Antarctic held much interest for geophysicists, because they are the home of the north and south magnetic poles, the aurora borealis, and, of course, the largest remaining ice sheets of the Pleistocene glaciation.
But the early ice-sampling attempts were quite crude.
Hand-dug pits were made in the layered ice at places like Byrd Station, a U.S.
encampment on the Antarctic ice sheet at 80° south latitude.
These surface holes, however, only penetrated through very young ice.
They provided samples and data for the topmost part of the Antarctic glaciers, but could not capitalize on the third dimension, the great thickness of the ice.
The U.S.
Army Corps of Engineers already had a major laboratory devoted to “cold regions” research at the time of the IGY, and its chief scientist, a man named Henri Bader, pushed hard for scientific drilling of the ice.
Although it didn’t happen immediately, within a decade after the end of the IGY, his organization had cored deep into both the Greenland and the Antarctic ice sheets.

The earliest drilling was done on a ships-of-opportunity basis, in places where camps or scientific stations already existed.
It produced a lot of new information about the polar ice caps, but it also soon became apparent that a more coherent strategy was required, and that the greatest rewards would come from drilling where the longest possible undisturbed cores could be obtained, or where the accumulation rate of snow was especially favorable for obtaining high-resolution records.
Several countries, both individually and in partnership, made ambitious plans for polar drilling, and by the mid 1990s, several deep-drilling projects had been completed on the Greenland and Antarctic ice sheets, and others were active.
Particularly important for research on the Pleistocene Ice Age have been cores from central Greenland and Vostok Station in the Antarctic.
In Greenland, under formidable weather conditions, teams from Europe and the United States drilled a pair of holes just thirty kilometers apart, almost dead center in the continent and at the very summit of the ice cap.
The rationale for this seemingly double effort was that independent analysis of the cores would provide a cross-check on the reliability of the data—and also, two cores would provide twice as much ice for critical analyses.
Both projects reached a depth of about three kilometers, where the ice is more than 110,000 years old.
Except for the very oldest sections of the cores near the bottom of the holes, where flow over the underlying rocks of the Earth’s crust coupled with the great pressure of the overlying ice has distorted the layering, the agreement between the two drilling sites is amazingly good.

In January 1998, at the other end of the world in the Antarctic, ice that until very recently was the oldest yet drilled (more than 420,000
years old) was retrieved from a hole that reached 3,623 meters depth.
Although ice cores have been collected at many sites in the Antarctic, including at the South Pole, this very long core from the Russian Vostok Antarctic Station has special importance, and has been heavily studied, because it reaches so far back into the Pleistocene Ice Age.
It does so because annual snowfall is much less in the Antarctic than in Greenland, and a given thickness of ice therefore represents a far greater stretch of time than it does in Greenland.
(In September 2003, a European consortium announced that ice they had recovered at another Antarctic site dates back to at least 750,000 years.
Few data are yet available for this core at the time of writing.)

Ice drilling at Vostok (where the mean temperature is a chilly
^–
55°C) actually began in the 1970s and was for a long time purely a Russian endeavor, but later became a joint Russia-France-U.S.
project.
Based on remote sensing measurements from the surface, the ice continues for more than 100 meters beyond the depth reached by drilling.
However, a decision was made to stop, because below the ice lies a large lake—Lake Vostok—that has been isolated from contact with the atmosphere for hundreds of thousands of years.
Insulated from the frigid polar air above by three and a half kilometers of ice, and heated from below by the slow but steady escape of heat from the Earth, the lake remains unfrozen.
It is thought that Lake Vostok may contain unique bacteria or other life that has long been isolated from the rest of the world’s biosphere, and biologists and chemists are anxious to devise a contamination-free sampling plan.
The last thing they wanted was to plunge a dirty drilling string through the ice into its undisturbed waters.