Frozen Earth: The Once and Future Story of Ice Ages (29 page)

The ice cores have yielded a massive amount of information about the Earth’s climate and environmental conditions through the few most recent glacial-interglacial cycles of the Pleistocene Ice Age, including data about the local temperatures at the drilling sites, the rates of snow accumulation, a record of distant volcanic eruptions (which are indicated by thin layers of very fine volcanic ash that is distributed globally through the atmosphere), and even the intensity of winds.
The latter
comes from analysis of the amount of ordinary dust and sea salt in the ice cores: the greater the wind intensity, the higher the content of these materials.
And, of course, the ice also traps samples of the atmosphere as small bubbles, as we have already seen.
The most recent part of the ice cores, especially for the time since the Industrial Revolution, also provides a very good record of how man’s activities have affected the global environment.
Anything that forms gaseous molecules, or attaches itself to the very tiny particles that are transported around the globe by winds, can end up being deposited in polar ice.
Mercury, lead, and freon are just a few examples of substances that have been measured in glaciers and can be traced directly to industrial processes.

Some of the properties measured in the long Vostok core are shown in figure 21.
A remarkable feature of such graphs is that almost all of the properties that have been measured in the ice show patterns similar to those of oxygen isotopes in deep-sea sediments over the same time period.
The Vostok core is long enough to show four complete glacial-interglacial cycles, and the approximately 100,000-year periodicity is readily apparent.
As with the deep-sea sediments, setting up a reliable timescale for the variations observed in the ice cores is critical for their interpretation.
Here glaciers have one great advantage over ocean sediments: snow accumulates relatively quickly, and annual layers are usually easily discernible.
Even deep in a glacier, where high pressure and the flow of ice have thinned them, the layers can often be distinguished.
If annual layers can be counted—and if you can be certain that none are missing—then the cores can be dated very precisely and all of the interesting environmental proxy indicators can be placed accurately on a timescale.
Also, because of the very high resolution provided by the annual layers, even events of quite short duration can be identified accurately.
But counting layers one by one is a tedious process—can you imagine counting tens of thousands of layers without making an error?
Automation has helped.
It turns out that the electrical conductivity of the ice changes subtly with the seasons, because different amounts of various trace compounds from the atmosphere are incorporated into
snow in summer and winter.
Long sections of core can be scanned quickly for changes in conductivity, and the wiggles of the output interpreted in terms of yearly cycles.
But the human eye and brain have remained primary tools for constructing ice-core timescales.
Where
cross-checks are possible, layer counting has proved to be very accurate.
For example, the dates of quite a few large volcanic eruptions are well known from historical records that stretch back several thousand years.
Fine-grained volcanic ash and chemicals such as sulfur dioxide that are spewed out in these eruptions are quickly distributed through the atmosphere and deposited on the icecaps as discrete layers.
Dating by layer counting generally places these marker horizons in the ice cores within a few years of their actual occurrences.
Uncertainties get larger for older parts of the ice cores, but in the Greenland cores, where layer counting has been very successful, dating appears to be accurate to a few percent over the past 100,000 years—a remarkable accomplishment.
In Antarctic cores, too, layer counting gives correct ages for events that can be dated independently in other localities—for example, a particularly rapid change in global temperature that shows up in proxy records of different types in different places.
Such agreement provides a high degree of confidence in the method.

Figure 21.
Data from the Antarctic ice cores at Vostok Station show that temperature changes (relative to the present) calculated from isotopic data and atmospheric CO
₂
content (in parts per million) from bubbles in the ice track each other very closely.
Cold glacial periods had low CO
₂
, whereas the warm interglacials had much higher values.
The amount of dust in the ice, on the other hand, is highest during the cold periods, signifying generally windier conditions.
These graphs are based on data from a paper by J.R.
Petit
et al.
in the journal
Nature,
June 3, 1999.

The properties of the Vostok ice core shown in figure 21 illustrate some of the insights that have been gained into Pleistocene Ice Age climate from examination of polar ice.
One, the close correspondence between temperature and the concentration of CO
₂
in the atmosphere, we have already discussed.
Both seawater temperatures, deduced from proxy measurements in sediments, and the local temperature at the Vostok drilling site—also based on a proxy, the isotopic composition of hydrogen in the ice—are closely correlated with carbon dioxide.
The amount of dust in the Vostok core, a proxy for windiness, shows a strong anti-correlation with other parameters.
It is obvious from the graph that the windiest times of the past 400,000 years, when most dust was deposited on the ice, occurred at the height of the glacial periods, when the temperature in the Antarctic—and presumably also globally—was at a minimum.
This is consistent with other evidence that at the times of maximum ice cover, the global climate was cold, arid, and windy.
In such an environment, the extent of desert areas increased, providing more dust, and grasslands expanded at the expense of forests.
Extensive loess deposits formed in some parts of the world.
Yet another
feature that is apparent in figure 21, especially in the temperature and CO
₂
plots, is the very sharp transition from cold to warm periods.
This had already been noted in sediment cores, but it is even more striking in the Vostok data.
Wally Broecker, an early investigator of the glaciation record in deep-sea cores, termed these quick shifts in temperature “rapid terminations,” implying a swift end to glacial episodes after long periods of cold.
The best estimate of the actual temperature change at the Vostok site during these transitions is about 12°C.
Note, too, that the warm periods do not last very long—in fact, the present warm period is considerably longer than any other in the ice-core record—and that during each glacial period, the temperature gradually cools to a minimum value just before the next “rapid termination,” giving the whole graph a saw-tooth appearance.

For completeness, there is one aspect of figure 21 that should be explained in more detail.
It has to do with how measurements of the components of air bubbles correspond with other properties.
Although both CO
₂
and the temperature proxy were measured in the same ice core, at a given depth in the ice, these parameters do not record conditions at the same time.
It turns out that an adjustment has to be made to the air bubble data in order to bring it into correspondence with the other properties.
The reason is quite simple.
As anyone who has walked in a snowfall knows, fresh snow is very light, because it is mostly air.
But as more and more snow accumulates, pressure increases on the underlying layers, and air is squeezed out.
On a glacier, as long as there are still spaces around the snow crystals, air in the snow will continue to exchange with the atmosphere, moving in and out with the winds and with changes in atmospheric pressure.
But eventually the pressure of the overlying layers causes the snow crystals to grow into a continuous mass of ice, trapping whatever air remains as sealed bubbles and preventing it from further exchange with the atmosphere.
That instant, when there is no longer communication with surface air, is the critical one for scientists attempting to decipher the gas bubble data.
Depending on how quickly snow accumulates, this may happen at a
depth in the glacier where the surrounding ice is anywhere from a few hundred to almost a thousand years old.
In the latter case, a gas bubble analysis would refer to conditions that prevailed a thousand years after the temperature was recorded in ice at the same depth.
Obviously, to make an accurate comparison of how various properties track one another, the timing must be known fairly accurately.
Usually, it’s possible to estimate the rate of snow accumulation from the thickness of annual layers in the ice cores, so that the uncertainty in the age of the air that’s being analyzed can be minimized.
Still, it should be realized that there may be small offsets in the records.

Up to this point, I have focused on the results from Greenland and Antarctic ice cores.
That is where the remaining large ice caps are, and where the major drilling efforts have been mounted.
But a few people have recognized that mountain glaciers, virtually all of them small remnants of much larger ice fields that existed at the height of the most recent glacial period, may also have interesting stories to tell.
Because these smaller glaciers are widespread across the Earth, they also may hold clues about the global reach of the ice age climate.
In particular, those that still exist at tropical or subtropical latitudes are often the only available storehouses of information about ice age climate away from the harsh environment of the polar regions.

Many scientists credit one man, a glaciologist at Ohio State University named Lonnie Thompson, with almost single-handedly inventing and shaping the study of ice cores from small mountain glaciers.
In recent years, Thompson’s decades-long contributions to climate science have been thrust into the limelight as he won a series of awards.
These include the Heineken Prize of the Royal Netherlands Academy of Arts and Sciences, and, together with his wife Ellen Mosley-Thompson (also a glaciologist), the prestigious Common Wealth Award for Science and Invention.
Commenting on his work, Thompson noted that the glaciers he studies are like canaries in a coal mine—sensitive indicators of change.
Many are fast disappearing.
It was the Thompsons who predicted that the Snows of Kilimanjaro—mentioned
earlier in this book, and one of a few equatorial glaciers—will completely disappear by 2020.
Less well-known glaciers in South America, Asia, and North America are vanishing just as quickly.

Lonnie Thompson’s early ambition was to be a coal geologist—he came from West Virginia, where coal mining is a major industry.
But his horizons broadened as he worked toward a Ph.D.
at Ohio State University.
The university had been a center for data from polar regions collected during the IGY, and soon Thompson and his wife were both working on ice-coring projects at the university’s Institute of Polar Studies (now the Byrd Polar Research Center).
But research on Antarctic and Greenland ice was a new and very hot topic, and Thompson began to look for a niche not already filled by senior scientists and ambitious younger colleagues.
While still a graduate student, he made an expedition to explore the five-kilometer-high Quelccaya glacier in the Peruvian Andes.
That trip, physically demanding though it was, convinced him that drilling on the glacier was possible and would provide information that couldn’t be had from the cores in Greenland or the Antarctic.
Thompson wrote a proposal to the U.S.
National Science Foundation (NSF) to mount a full-scale drilling program along the lines of those that were already operating in the polar regions.
But his request was turned down—the consensus was that it just wasn’t feasible to drill at such high altitudes.
Thompson was not deterred, and managed to carry out annual summer research on the Peruvian glacier from 1976 onwards.
By 1979, he had won over the NSF and secured money for drilling.
Although his first attempt was a failure—flying the drilling rig to the glacier proved too dangerous—he eventually succeeded, in 1983, using people and pack animals to bring in the equipment and supplies.
Thompson and his crew drilled through the glacier to the underlying rock—a short distance by the standards of polar ice cores, only some 160 meters—and obtained a well-preserved sequence of annual layers that extended back 1,500 years.
It was the first core through a tropical glacier.
Even so, some in the science community were not impressed, because the timespan was short.
But for those interested in tropical climate, especially
South American climate, it was a bonanza.
Here, in great detail, was a record of environmental change.
Variations in average temperature, wet periods, dry periods, even volcanic eruptions were all recorded in the cores.
Because all of these would have affected the peoples of the region, archeologists could begin to make connections between climate and what they knew about the history of various societies in the area.
Thompson’s persistence had paid off.