Frozen Earth: The Once and Future Story of Ice Ages

by Doug Macdougall

  The work of Shakun and his colleagues examined only the most recent deglaciation, spanning approximately the past twenty thousand years. But the Pleistocene Ice Age is characterized by multiple cycles of warming and cooling, of ice retreats and advances, stretching back two and a half million years or more. Detailed, high-resolution records through all of these cycles are rare. For example, Greenland ice cores, a primary source of information about past Northern Hemisphere climate changes, extend to only 130,000 years ago, covering little more than one complete cycle. However, during the winter of 2008–9, a group of scientists and engineers operating under the aegis of the International Continental Scientific Drilling Program retrieved sediment cores that record the local climate in northeastern Siberia through nearly all of the Pleistocene Ice Age cycles. The cores were drilled from a lake (with a tongue-twisting name: Lake El’gygytgyn) that occupies a 3.6-million-year-old meteorite crater about one hundred kilometers (sixty-seven miles) north of the Arctic Circle. The availability of a continuous record of local Arctic environmental change through the Pleistocene Ice Age is tremendously important because it permits climatologists to compare the real climate variability, as recorded in the sediment cores, with that predicted by climate simulations run with different forcing factors. This is especially valuable for the Arctic because both climate models and observations (including temperature records from the past few decades) indicate that northern polar regions are considerably more sensitive to global warming than other parts of the Earth.

  The scientists who examined the Lake El’gygytgyn sediment cores recently summarized their work in the journal Science (“2.8 Million Years of Arctic Climate Change from Lake El’gygytgyn, NE Russia,” by Martin Melles and colleagues, Science 337, 20 July 2012). What did they learn? Two observations stand out. The first is that in northern Siberia, many “super interglacials,” short intervals when local summer temperatures reached levels considerably higher than those of today, punctuated the long Pleistocene Ice Age. The second is that these periods of high temperatures in Siberia correspond closely in time with episodes of ice sheet meltback in the Antarctic that are known from ocean sediment cores.

  Melles and his colleagues looked in detail at several especially warm super interglacials, with summer temperatures 4°C to 5°C (7°F to 9°F) higher than those of today, and investigated possible forcing factors that could have produced such temperatures. What they discovered is surprising. Climate simulations that included the effects of both local summer insolation and greenhouse gas forcing (the latter probably more important) could not reproduce the observed high temperatures and instead predicted temperatures that were no higher than those of non-super interglacials. And because the super interglacials at Lake El’gygytgyn correspond to periods of sharp deglaciation in Antarctica, it is clear that these high-temperature intervals were not simply the result of localized extreme warmth. The super interglacials were global.

  Why didn’t the climate models reproduce the high super interglacial temperatures experienced at the Siberian lake? Clearly, still-unrecognized processes or forcing factors must have been involved. Melles and his colleagues speculate that ocean circulation—that great mover of heat around the globe—might be part of the answer, but they can’t be sure exactly how. These results are another reminder of just how complex the climate system is, and how difficult it is to construct simulations or models to predict accurately how temperatures, rainfall, and the like will change in the future. More often than not questions answered spawn new questions, and climatologists—indeed, all scientists—always seem to face more work to get to the bottom of things.

  Studies such as those described in the past few pages are remarkable achievements; they have detailed how surface temperatures, precipitation, vegetation, ocean circulation, and other aspects of the environment changed during the Pleistocene Ice Age. Even though questions remain, they have gone a long way toward elucidating the mechanisms behind glacial-interglacial cycles. But what about the ultimate question: what initiated the Pleistocene Ice Age in the first place?

  In chapter 12, I describe one possible answer, an idea that was suggested not long before the initial publication of this book in 2004: that chemical weathering of the evolving Himalayan Mountains “drew down” carbon dioxide in the atmosphere, reduced the greenhouse effect, and cooled the planet. This may seem a bit confusing because the Pleistocene Ice Age began only about two and a half million years ago, when large-scale glaciers began to form in northern polar regions, yet the Himalayas are much older (they began to form about fifty million years ago when plate tectonic forces caused India to crash into Asia). However, temperature proxies in deep-sea sediment cores show that global temperatures declined steadily from approximately the time of the India-Asia collision (when they were much higher than they are today) until the start of the Pleistocene Ice Age. By about thirty-five million years ago, global temperatures were low enough for ice to begin to cover the Antarctic (which had previously been unglaciated), and climate feedbacks related to this ice cover further cooled the Earth until, eventually, Northern Hemisphere glaciation began. So the question of what initiated the Pleistocene Ice Age rests on what caused the long-term cooling that began around fifty million years ago.

  It is well known that carbon dioxide from ordinary air, when dissolved in rainwater, is the primary agent of rock weathering and that extensive weathering depletes its abundance in the atmosphere. That young, rising mountain ranges are sites of intense chemical weathering is also well known. The coincidence in timing between the rise of the Himalayas and a global temperature decrease suggests that weathering of this young mountain range could have been responsible for the lower temperatures, through its effect on atmospheric carbon dioxide. But recently a new candidate has joined carbon dioxide drawdown as a possible cause of the global cooling: sulfur. What, you may ask, does sulfur have to do with climate? Potentially quite a lot. Sulfur is plentiful; in the form of sulfate (SO2-4), it is the fourth-most-abundant ion in seawater. Because of this, the oceans are a major source of sulfur-bearing aerosols in the atmosphere—suspended microscopic droplets that reflect incoming solar radiation. When their concentration increases, they reflect more solar radiation and the Earth cools. This effect was illustrated clearly in 1991, when a large eruption of Mt. Pinatubo in the Philippines injected sulfur-bearing aerosols into the atmosphere, lowering global average temperatures by about 1°F for more than a year.

  In a recent paper in the journal Science (“Rapid Variability of Seawater Chemistry over the Past 130 Million Years,” Science 337, 20 July 2012), Ulrich Wortmann and Adina Paytan note that the record of past seawater sulfur content shows large and quite rapid changes, and they conclude that deposition and dissolution of vast quantities of the sulfurrich mineral gypsum almost certainly caused at least some of this variability. Gypsum is abundant in so-called evaporite deposits, which are assemblages of minerals that form in hot, arid regions when salty seawater trapped in restricted basins evaporates. Large-scale evaporite deposits have formed many times during our planet’s long history, as evidenced by the numerous salt mines found around the globe (in addition to being important sources of sulfur, evaporites provide us with table salt and potassium for fertilizer). But evaporite minerals are not very stable at the Earth’s surface; when exposed to ordinary precipitation, they dissolve readily.

  Wortmann and Paytan’s analysis indicates that the sulfur content of the oceans started to increase rapidly (geologically speaking) approximately fifty million years ago—near the time when uplift associated with Himalayan mountain building began. The authors conclude that this uplift exposed large-scale evaporite deposits to erosion. (Undissolved remnants of these deposits still exist, stretching from Oman to Afghanistan, Pakistan, and India.) Large amounts of gypsum in the uplifted deposits dissolved, substantially raising seawater sulfur content and thereby increasing the concentration of sulfur-bearing aerosols in the atmosphere, which ultimately resulted in global cooling. Plate tecton
ics—in this case the collision of India and Eurasia—thus played a major role in the cooling that led to the Pleistocene Ice Age, through both the drawdown of carbon dioxide and the supply of sulfur to the oceans. These observations illustrate how deeply interconnected even seemingly disparate Earth processes are.

  Did plate tectonics and the movement of continents play a role in the Earth’s earlier ice ages? We don’t know for sure, because as scientists probe further and further back into our planet’s history the evidence becomes increasingly fragmentary. The question of timing is crucial: for example, was the onset of an ancient ice age coincident with a continent-to-continent collision like the one that raised up the Himalayas, or not? Dating events accurately enough to answer such questions is more easily said than done. But one thing is clear from recent research: greenhouse gases, particularly carbon dioxide, played a major part in the initiation and the cessation of past ice ages, just as they have for the Pleistocene Ice Age.

  Take, for example, the “Snowball Earth” theory, described in chapter 8. Over the past several years evidence has continued to accumulate that severe glaciation, with permanent glaciers on the continents and ice covering even tropical seas, occurred during several discrete ice ages between about 600 and 750 million years ago. One of the problems many scientists initially had with the concept of a completely frozen Earth was that it would have been very difficult to melt: an ice-covered planet would reflect so much of the sun’s energy that it would stay frozen. However, under such conditions it is likely that enough carbon dioxide (from volcanic eruptions) would eventually accumulate in the atmosphere to produce a “super greenhouse” world, leading to collapse of the ice sheets. Ending Snowball Earth–like glaciations may not have been as difficult as once thought. But what initiated these extreme events?

  Since the first publication of this book, computer models of global climate have become ever more sophisticated, capable of incorporating more, and more varied, factors that influence climate. Several groups of scientists have used these models to investigate the probable forcing factors most important for initiating Snowball Earth–like conditions.

  At the time of Snowball Earth glaciation, the planet was a very different place than it is today. For starters, the surface received about 6 percent less solar radiation (this is well known from studies of how stars like our sun evolve). Furthermore, all the evidence points to low-latitude locations for most of the existing continents, with none at the poles. Both of these boundary conditions are important for understanding the Snowball Earth glaciations.

  The computer models don’t tell us exactly what happened, and different versions give slightly different results. But all of the simulations point to the importance of two primary climate forcings: the reflectivity (albedo) of sea ice, and the amount of carbon dioxide in the atmosphere. Even with solar radiation only 94 percent as strong as it is today, very low greenhouse gas concentrations are crucial for initiating Snowball Earth episodes in all climate models because—with no continents in polar regions—extremely low temperatures are necessary to initiate freezing of the high-latitude seas and maintain year-round ice cover. As cooling proceeds under low greenhouse gas conditions and ice cover expands, however, albedo becomes the dominant factor and eventually results in runaway cooling. Exactly how much of the planet must be covered with ice and snow for this to happen varies depending on the model used. But the point at which runaway cooling begins can’t be reached at all without very low greenhouse gas concentrations.

  What lessons do the climate models have for the Anthropocene (an informal but very useful label for the time in our planet’s history when human activity has overtaken natural processes as a primary driver of atmospheric chemistry and other aspects of our environment)? One startling conclusion from the best and most recent models is that even after anthropogenic carbon dioxide emissions slow down or stop, their effects will persist for much longer than is generally realized: tens of thousands of years. As the science journalist Mason Inman put it, “carbon is forever” (and he wasn’t referring to diamonds, which are pure carbon).

  Why do the effects of greenhouse gas emissions last so long? Won’t the Earth start to cool down when humans stop putting greenhouse gases into the atmosphere? The simple answer to the first of these questions is that the climate system is complex and takes a long time to approach a new equilibrium state; the answer to the second is yes, but slowly and only (for thousands or perhaps even tens of thousands of years) to temperatures well above those of the period before the emissions began.

  Throughout the glacial-interglacial cycles of the Pleistocene Ice Age, carbon dioxide in the atmosphere has fluctuated between a low near 170 parts per million during the coldest intervals to about 300 ppm during the warmest. Today it stands near 395 ppm, the high value mainly due to the burning of fossil fuels. Even taking into account pledged emission reductions, the concentration is expected to continue rising and will likely exceed 850 ppm by the end of the twenty-first century. If carbon emissions were to miraculously fall to zero then, which appears less and less likely with each passing year, climate models indicate that atmospheric carbon dioxide would still be close to 500 ppm a thousand years later. Global average temperatures would still be several degrees Celsius (more than 5°F) higher than those of today. If we end up burning all of the Earth’s fossil fuel reserves, atmospheric carbon dioxide will rise even higher over the next few centuries, to levels approaching 2,000 ppm, and recovery to conditions resembling those of today will take correspondingly longer—hundreds of thousands of years. Although about half of the anthropogenic carbon dioxide will eventually dissolve in the ocean, and chemical weathering of surface rocks will gradually consume most of the rest, these are slow processes. The atmospheric content—and the Earth’s surface temperatures—will remain high for a very long time.

  In the absence of human activity, the cycles of glacial and interglacial periods that characterize the Pleistocene Ice Age would continue, paced by Northern Hemisphere insolation changes. The next severe glaciation would occur some fifty thousand years from now, when the Earth’s orbital parameters will result in low summer insolation at high northern latitudes. Once again ice would advance over large swaths of North America, northern Europe, and Asia. But if human activity releases so much carbon dioxide into the atmosphere that greenhouse warming overwhelms the cooling effect of decreased insolation, there will be no Northern Hemisphere glacial advance in fifty thousand years. The next glacial period will not occur for at least another half a million years, by which time most anthropogenic carbon dioxide will be gone. It is astonishing to realize that human activity over just a few centuries could have such a profound effect on our planet, stretching tens to hundreds of thousands of years into the future.

  To put things in perspective, I should point out that the Earth has experienced periods in the past—even very long periods—with atmospheric carbon dioxide of several thousand ppm, high global average temperatures, and no permanent glaciers except perhaps for a few small high-altitude ice fields. However, that was long before humans arrived on the scene and existing life had adapted to conditions we would consider extreme. The greenhouse gas content of the atmosphere is now rising at a rate unprecedented in the Earth’s long history, entirely because of human activity. Most of the consequent environmental changes will occur over the next few centuries. Unless geoengineering solutions can be found—large-scale projects designed to slow or stop global warming by a variety of methods, including extracting carbon dioxide from the atmosphere and storing it permanently—humankind will have to adapt very nimbly in order to avoid the partial or wholesale collapse of nations and societies. The environmental changes, including higher global temperatures, higher sea level, and potentially drastic changes in biological diversity and species distribution, will affect agriculture, human health, and all populations living close to sea level. Who would have thought that studies of ice ages could give us such insight?

  Doug Macdougall


  October 2012

  CHAPTER ONE

  Ice, Ice Ages, and Our Planet’s Climate History

  The American author and historical popularizer Will Durant once wrote, “Civilization exists by geological consent, subject to change without notice.” That is not a new idea, even if Durant phrased it especially well, but nowadays many historians scoff at the notion of environmental determinism, the possibility that climate or geology may have seriously affected the course of human history. And yet there are still many places on this planet where Durant’s observation rings true, especially places with extremes of climate. One such is the arctic regions, particularly Greenland. Ninety-five percent of that island country is covered by ice. Towns and villages cling to the coastline; at their backs loom glaciers a thousand meters thick: gleaming, white, blue, clear, transparent ice. The icecap weighs on the land like a lead brick on a floating plank, pressing it down below the level of the surrounding sea. If the ice were suddenly removed, the waters of the ocean would rush in to take its place. The glaciers seem fixed and static, but in reality they are dynamic, in constant slow movement outward from their thick centers. New snowfall adds to their mass every year, but at the margins they calve off apartment-block-sized chunks of themselves and send flotillas of weirdly shaped icebergs sizzling and crackling and sometimes eerily and silently floating down the fjords to the sea. The icebergs carry pieces of Greenland with them too, sand, pebbles, and boulders gouged and scraped from the land, later to be dropped far out at sea as the ice melts. The Inuit of Greenland have lived with the ice of glaciers for thousands of years. They are truly people of the ice age. Most of the rest of us have been affected by the ice age too, but in less obvious ways.

  Permanent icefields—that is, large glaciers—are not common in mainland North America. In the mountainous west, in Alaska and in the Yukon, there are small high-altitude glaciers, but in the overall scheme of things, they are fairly minor features of the landscape. However, as a boy, like many others both in North America and northern Europe, I grew up surrounded by the work of ice. Like most others, I was, at the time, completely unaware of that fact. I am not referring to the ice of a skating rink or of a January puddle. Rather, this was ice just like that of Greenland today, or of Antarctica, ice of vast extent and kilometers thick that blanketed huge swathes of the Northern Hemisphere thousands of years ago. It reached down from centers in Canada and Scandinavia and covered the sites of cities such as Boston, Detroit, and Hamburg. Its legacy is everywhere even today, from the geography of our waterways to the distribution of native peoples in the New World. It ground up solid rock to make the sand of countless beaches and the soil of midwestern farms in the United States. It sculpted rolling hills and long valleys across the landscape. It scraped up soil and rocks as it flowed, and dumped the debris as terminal moraines in places like Cape Cod and Long Island, New York, far from its original home. It even picked up diamonds from still-undiscovered deposits in Canada and transported them to the United States, twenty thousand years before NAFTA was conceived.