by Curt Stager
To find the best examples of how life as we know it responded to a major warming in the past, we should turn our attention to events that lie as close to us in geologic time as possible. As luck would have it, a recent interglacial fits the bill nicely. To reach it, we need travel only 117,000 years back in time. Only 117,000 years? Nobody but a geologist or paleoecologist would put an “only” in front of that number. For those who are used to dealing with vast chunks of geologic time, it’s easy to forget that 117,000 years is a very long stretch from the point of view of living things.
Just to make that point more clearly, let’s imagine that we’re traveling along a path that leads us back through climatic history. We’ll use the EPICA ice core’s bumpy temperature curve as our trail to the Eemian interglacial.
We feel only minor bumps and jostles in the early stages of that ride. We rumble gently through the wars, civilizations, and inventions of the last millennium. We pass the births of Mohammed and Jesus Christ further along the time track. After we’ve traversed another 2,000 years, some readers might expect to see Noah’s flood, followed by the creation of the world another 2,000 years down the line. But the track continues onward from the 6,000-year mark, rising slightly over four more millennia to the beginning of the warm early Holocene, back during the early days of grain agriculture in the Middle East. Another thousand years pass, and another.
EPICA ice core record of Antarctic temperatures from the Eemian interglacial to today. After EPICA community members, 2004
By now EPICA has led us through twelve millennia of cultural and climatic changes. But to reach our final destination, we must also slide down into a cold pit full of sharp-looking temperature spikes that takes nine times as long to cross as the stretch that we’ve already passed through. We’re getting ready to experience, in reverse, a full-blown ice age.
It’s a rough ride through the thermal chop of the last glaciation, which scraped much of Canada’s rock-and-soil skin off. More than a dozen saw-toothed warmings punctuated it, mostly in response to sudden changes in ice extent or ocean circulation, though none of them lasted more than a few centuries or millennia. As we rattle up the slope of each warming spike only to drop headlong down the older, steeper face, we might long for the more subtle ups and downs of Holocene climate.
Relief finally comes when, after covering 117,000 years of vigorous climate change, we reach the far side of the glacial temperature pit and climb to the edge of the Eemian interglacial. Once again, the path lies high and warm before us for another 8,000 years or so. At this point on our backward journey, most of the Eemian seems to be about as warm as the Holocene was, but there’s one last thermal crest ahead; those earliest five millennia were warmer still. Beyond that peak, just past the 130,000-year mark, yawns the toothy canyon of another, earlier ice age. Let’s not go there just yet.
To put this into more mundane terms and in normal temporal order, the Eemian interglacial began about 130,000 years ago. During the 5,000 years between full glacial conditions and the early Eemian peak, temperatures in East Antarctica rose by 22 to 25 degrees Fahrenheit (12 to 14°C). This averages out to a rate of roughly half a degree per century, less than the 1.3 degrees (0.7°C) of warming that Earth as a whole experienced during the twentieth century. However, the high, frigid mass of East Antarctica where the EPICA coring site lies is also warming more slowly than the global average today, so it wouldn’t be surprising if it lagged behind the rest of the world during the Eemian, too. But despite the relatively slow onset, the EPICA site did eventually became several degrees warmer than it is today for several millennia, and then it settled closer to modern conditions before dropping off into the last great glaciation, about 117,000 years ago.
The warm period’s name stems from the Eem River in the Netherlands about 20 miles (30 km) east of Amsterdam, which flows through clay-and sand-rich marine deposits laden with the shells of subtropical mollusks. Today many of those species, such as sharp-pointed needle whelk snails (Bittium reticulatum), generally avoid the chilly North Sea and instead favor warmer coasts such as those of the Mediterranean region. I remember collecting hundreds of Bittium shells from Turkish beaches as a kid when my father spent a year teaching math in an Istanbul high school. Since their discovery, during the late nineteenth century, the formerly submerged shelly sediments beside and beneath the Dutch riverside town of Amersfoort have represented the classic geologic reference point of the Eemian interglacial.
Technically speaking, the Eemian moniker refers only to European history; in North America, it’s been called the Sangamon, and in Russia it’s the Kazantsevo. At sea, it produced one of the largest of many oscillations in long geochemical records derived from sediment cores, and so marine geoscientists refer to it as Marine Isotope Substage 5e. But for our purposes, Eemian will do.
Ice core evidence suggests that Eemian CO2 concentrations stayed fairly close to 300 ppm. Presumably, because the greenhouse gas concentrations were lower back then than they are now, global average temperatures should also have been lower. But that’s not what the data tell us. Many records put Eemian temperatures higher than ours, typically ranging from 2 to 5°F (1 to 3°C) above today’s global average. What’s up with that?
Could the gas bubble records from ice cores be wrong? Probably not by much, because the same results appear in core after core. And other historical tools yield similar results, as well. For example, the Scandinavian research team of Mats Rundgren and Ole Bennike has coaxed temperature data from ancient willow leaves that they harvested from Eemian deposits at various European sites. Because plants breathe CO2, their leaves sometimes respond visibly to changes in the abundance of that life-sustaining gas. The telltale feature in this case was the number of microscopic breathing holes, or stomates, in a standard area of leaf surface; many plants react to rising CO2 concentrations by producing fewer stomates. From their stomatal index data, these investigators concluded that Eemian CO2 concentrations hovered between 250 and 280 ppm, which is remarkably similar to the ice core numbers.
Could the temperature estimates be wrong? We have several ways of getting at those values, too, most having something to do with stable isotopes (an isotope is a variant of an atom, as a breed is a variant of a domestic dog or cat). One of the most commonly used isotopes in this kind of work is deuterium, a heavy version of hydrogen that contaminates the water molecules in glacial ice. It serves as a paleothermometer because its abundance relative to normal hydrogen changes with temperature. The deuterium records of the EPICA and Vostok ice cores suggest that Antarctic temperatures were about 3 to 5°F (2 to 3°C) warmer than today for the first 4,000 to 5,000 years of the Eemian. But for the rest of the interglacial, temperatures more closely resembled those of modern times.
Ice core records, however, represent only single reference points on a map. To calculate global average temperatures today, we rely upon hundreds of widely dispersed weather stations, but detailed Eemian temperature reconstructions number only in the dozens. So it is misleading to simply say that EPICA data show that the world was X degrees warmer than today during the early Eemian. EPICA’s ice layers can’t tell you exactly what the global average was back then, just as a modern weather station situated at the coring site can only register the local conditions in that part of Antarctica. And the ice records only preserve information during seasons in which snow accumulated; if little or no snow fell during certain months of the year, then much of the annual record might not be represented at all.
The Mount Moulton horizontal ice core trench, which for now represents our only window on the Eemian climates of western Antarctica, displays just such geographic variability. It lies across the Ross Ice Shelf from the much larger eastern sheet where the EPICA and Vostok cores were drilled. Deep ice can ooze sideways under its own weight, and in the vicinity of Mount Moulton it presses forcefully against the base of the mountain and curls itself upward like layered taffy. If you walk outward from the rock-ice contact zone, you can traverse half a mil
lion years of polar history in parallel strips of translucent blue strata. Rather than drilling down through that archive, investigators from several American universities have simply sawed long sampling trenches through it.
Today, Antarctica is warming far more dramatically in the west than in the east, and that kind of regional variability also typified the Eemian. Although the Mount Moulton area warmed about as much as East Antarctica did 130,000 years ago, the subsequent cool-off there was more gradual than it was elsewhere, more of a smooth slide than a cliff dive. The basic storyline was the same: an abrupt temperature rise, several millennia of warm conditions, and a return to glacial cold. But the finer details vary from site to site, reminding us to be careful when using single records to study global history.
Our present warming trend is more intense in most of the polar regions than it is at lower latitudes, and the same situation existed in the past, too. Much of the Arctic was 7 to 9°F (4 to 5°C) warmer during Eemian summers than it is today, while the global average, as best we can tell, was slightly less than half of that. So if we rely only upon high-latitude records for estimates of past global temperatures, then we may bias our guesswork toward the high end of the scale.
But even if the Eemian Earth was only slightly warmer than today, greenhouse gas concentrations were definitely much lower. Does this mean that we’re wrong in linking modern global warming to carbon emissions? Not at all. We’ve simply gotten things backward nowadays. Today, CO2 and methane are driving temperatures up, but back then it was warming itself that drove the gas buildups. As the orbital cycles ran their appointed course, the loss of reflective snow and ice from northern high latitudes let darker lands and seas soak up even more incoming solar energy. That, in turn, would have accelerated microbial respiration rates and increased the activity of methane-belching wetlands, wrapping the planet in an even thicker insulating robe of heat-trapping gases. In our case, the insolation cycles that produced the Holocene interglacial of the last 11,700 years have already passed their peaks, and the present warming is mostly of our own making.
Despite its different origin, the Eemian example clearly shows that even moderate heating can demolish plenty of ice; indirect but compelling signs of interglacial melting appear in stranded fossil deposits all over the world. Among the hundreds of exposed paleo reef sites that formed during the Eemian, the most reliable indicators of past sea level are those found on coasts that are geologically stable and therefore unlikely to have lifted or lowered the ancient deposits. One such site has been described in a study led by Paul Hearty, a geoscientist from the University of Wollongong in New South Wales, Australia. Hearty and his colleagues examined a wave-cut bluff on Rottnest Island, a low-lying chunk of arid land rimmed with white sand beaches just offshore from Perth, and found fossilized corals of early Eemian age sitting at or above head height beyond the surf zone. Because such corals prefer to live in shallow water, this discovery demonstrated that sea levels of the time stood 6 to 10 feet (2 to 3 m) higher than they do today.
The longer the warm spell persisted, the higher sea levels continued to climb, even after the early thermal peak had passed. Moving farther east onto the mainland, thereby retracing the encroachment of the swelling interglacial ocean, Hearty’s group found petrified mollusks still attached to their parent rocks and to each other 23 feet (7 m) or more above modern sea level. These were creatures that lived and died several thousand years after the Rottnest corals did, and their higher, more easterly positions reflected a long-term sea-level advance. Together with similar findings elsewhere in the tropics, these results help to sketch a coherent sequence of Eemian sea-level change.
At the start of the interglacial, ocean surfaces rose 6 to 10 feet (2 to 3 m) higher than today and stayed there for several thousand years. During the second half of the warm phase, sea level continued to rise in several steps to a maximum of 23 feet (7m) or so, probably in response to partial collapses in lingering polar ice sheets. A long-term decline then followed as the next ice age withdrew water from the oceans and froze it back onto land.
Then, as now, the largest ice deposits lay in the Arctic and Antarctica, and these are the most likely sources of the mid-Eemian meltwater pulses. Something akin to the volume of Greenland’s ice sheet could do the job, but we can safely say that Greenland wasn’t the only donor to the water budget. Local ice cores there show that most of the central dome survived those long millennia of warmth, leaving a sizable remnant that was at least half as extensive as today’s pile. On the opposite end of the planet, the main mass of the East Antarctic ice sheet also survived; otherwise, we wouldn’t have those superlong records from EPICA and Vostok.
This means that the interglacial sea-level rise was probably the work of several ice sources, most likely a joint effort between Greenland and West Antarctica. It also means that even 13,000 years of heating didn’t destroy the Greenland ice sheet altogether, which might offer some reassurance as we watch temperatures and melting rates rise there during this century. If we can limit future warming to something akin to the Eemian situation, as would be expected for a moderate-emissions scenario, then perhaps we can also keep a fair bit of our terrestrial polar ice in place.
Fossil oysters near Durban, South Africa, still attached to rocks that were once submerged during the Eemian interglacial, now stranded well above sea level. Curt Stager
On the other hand, sediment cores collected from offshore sites north of Greenland show that most of the Arctic Ocean was seasonally ice-free then. Planktonic microorganisms of the sort that live under sea ice today gave way to other forms that are now common in ice-free waters farther south, and stable isotopes of oxygen preserved in the empty shells of those long-buried creatures also tell of open water near the pole. But the loss of that floating ice cap wouldn’t have changed global sea levels—it’s only land-based ice that does so.
Since there was nobody to chart coastlines precisely back then, we don’t have a complete picture of how sea-level rise reshaped maps of the Eemian world, but we do have some general insights into what was lost. Saltwater covered much of northern Europe and the west Siberian plains, isolating Sweden and Norway from the mainland as a sausage-shaped Fennoscandian island. And it’s a safe bet that many other low places, including those icons of modern sea-level rise, coastal Bangladesh and some of the smaller Pacific islands, went under for at least part of the Eemian.
Apart from warming and sea-level rise, some geological records also show what the Eemian did to rainfall. Moving farther east from Rottnest Island to what is now the hot, dusty heart of central Australia, we find an Eemian weather surprise. Increasing summer warmth invigorated monsoons throughout most of the lower latitudes, and this brought more rainfall to the arid outback. In the Kimberly region, the Gregory Lakes overflowed their banks, and the crusty flats surrounding what is now ephemeral Lake Eyre—often more of a desert mirage than a lake—were flooded continuously. Across the Indian Ocean in the highlands of East Africa, heavy rains overfed the Nile and poured river runoff into the eastern Mediterranean, where it capped the briny sea with a layer of less saline water. That buoyant lid prevented formerly dense, salty surface water from sinking or stirring deeply under the influence of the winds, and it therefore cut off the supply of dissolved oxygen to the bottom. This produced thick layers of slimy organic ooze that are still preserved in marine sediment cores collected near the coast of Egypt. And copious rains over the Sahara helped to imprison what are now vast seas of drifting sand under an emerald carpet of grass and trees.
The warmer and generally wetter interglacial world was inhabited by many animals that are familiar to us today; most of its dominant plant species are still with us as well. Although some notable evolutionary transformations have occurred since then, most of the largest biotic differences between now and then reflect environmental factors rather than genetic mutations. Mammoths are absent from our landscapes because of human hunting or the human-driven destruction of habitats by fire, not because
they’ve turned into a less hairy form of elephant. The time periods that separate us from the Eemian are too short even to have fully fossilized many of the bones, shells, and foliage left behind by its residents, which can seem more like mummies than stone replicas when discovered. Because of this similarity to modern organisms, we can use the remains of Eemian biota, not only to infer the effects of past climates on living things, but also to learn about the climatic conditions themselves.
During the warmest early phase of the Eemian, boreal treelines lay hundreds of miles north of their present locations, often pressing right up against the coastlines of the polar sea. Much of Baffin Island and southern Greenland were cloaked in white birch woods, and hazels and alders rustled in the breezes of northern Sweden and Finland well above the Arctic Circle. Pollen grains in lake and peat deposits show that great forests of oak, hornbeam, and yew covered much of Europe north of the Alps. Farther east, the pollen of spruce, fir, and pines blew into the crystal clear, mile-deep waters of Lake Baikal during most of Siberia’s Eemian (or Kazantsevo, if you prefer), which means that local boreal taiga woodlands of the time looked a lot like they do now. However, to the north of Baikal, evergreen forests eliminated coastal tundra while broad-leafed trees such as oaks and elms invaded the southern borders of the taiga belt.
In North America, too, the main theme was poleward migration. In central Alaska and the Yukon Territory, spruce-birch woods much like those of today moved in over retreating tundra as local weather became warmer and wetter in summer, eventually reaching much farther north than the Arctic treeline extends now. Cedar-hemlock-fir forests claimed the Pacific coasts of Washington and British Columbia, and the central plains sprouted a complex mosaic of deciduous forest and savanna. In Florida and Georgia, dry oak woodlands expanded during the opening phase of the Eemian before giving way to pinelands and cypress swamps. And in upstate New York and southern Ontario, trees that are now more typical of the southern Appalachians kept the local bears and squirrels fat with generous supplies of hickory nuts and acorns.