PALEOCLIMATOLOGY

Sultry Last Interglacial Gets a Sudden Chill

Contradictory evidence from ice cores raised questions about climatic events in Northern Europe in the late Pleistocene. Examining the evidence more closely, scientists stumbled upon a brief but intense cooling event.

by Mark Maslin, Environmental Change Research Centre, Department of Geography, University College, London; and Chronis Tzedakis, Godwin Institute of Quaternary Research, Department of Geography, University of Cambridge, United Kingdom

Over 100,000 years ago, during an interglacial period known as the Eemian, a climate event lasting only several hundred years was chronicled in climate records from the deep-sea sediments of the Atlantic Ocean and from lake sediments of Northern Europe. This event, which may provide a clue to future climate change, would not have been discovered had it not been for the "Eemian controversy" stirred by the Greenland ice core records.

    The controversy began when the oxygen isotope record from the core collected by the European Greenland Ice core Project (GRIP) was published in 1993. GRIP members found that air temperatures over Greenland —represented by oxygen isotopes in the ice—varied greatly during the last interglacial, the Eemian. Until then, the climate of the last interglacial was thought to be stable and similar to that of the Holocene. The GRIP ice cores, however, contained evidence of two extreme cold events, each lasting over 1,000 years. This was cause for concern, as the Eemian was thought to be an excellent analog for modeling future climate change, especially because the Eemian atmospheric CO₂ content may have been slightly higher than preindustrial levels.

    Initially, the finding led to controversy because oxygen isotopes in the adjacent American GISP2 (Greenland Ice Sheet Project) ice core did not show the same events in the Eemian section of the sequence. In the GISP2 core, inclined layers, which suggest that the ice has been disturbed, were detected in the ice that was older than about 110 ka (110,000 years). Some researchers thought that these layers might also be present in the GRIP core.

    In September 1995, participants at the GISP2-GRIP Joint Workshop concluded that both GRIP and GISP2 ice cores had suffered stratigraphic disturbances in ice older than 110 ka. Their determination was based on atmospheric CH₄ and ¹⁸O trapped in ice as well as detailed work on the structural properties of the cores.

     While the debate over the ice cores was heating up, oscillations observed in a number of European pollen sequences were used to illustrate the instability of the Eemian compared to the Holocene and by extension, to support the GRIP variations (see figure). However, correlating the pollen data with the GRIP record was difficult. Upon close examination, the terrestrial sites suggest a significant difference in the timing and length of cold events.

Comparison of a) lake records from Northern Europe (the % nonarboreal pollen from Central France indicating shifts in tree cover), b) the marine records from the Norwegian Sea (summer sea surface temperature), and c) the subtropical Atlantic (benthic foraminifera ¹⁸O) with d) the GRIP ice core. Because of the difficulties of absolute dating within the last interglacial, the records are aligned at the upper and lower boundaries of the Eemian. The mid-Eemian cold event is clearly shown in all of the records. Note that none of the terrestrial or marine records fit with the observed changes in the GRIP ice core record.

    At the same time, deep-sea records, including a high-resolution (70- to 200-year) record of Ocean Drilling Project site 658 (see figure), showed no evidence for the cold events observed in the GRIP record. These incompatibilities cast doubt over the robustness of the GRIP Eemian ice core section and appeared to support the view that the record was disturbed by ice tectonics.

Controversy Leads to Further Study

The Eemian debate may have provided another source of valuable climate information. The detailed work on climate records of the Eemian, primarily aimed at testing the validity of the GRIP record, may have accidentally lead to the discovery of a relatively short but climatically significant cooling event within the Eemian. A detailed study of Ocean Drilling Project site 658 in the subtropical Eastern Atlantic found that the duration of the terminations of Eemian and Holocene interglacial periods were similar. In contrast to the Holocene, the Eemian includes a short cold event shown in the benthic oxygen isotope record at about 122 ka. This cooling is accompanied by a short but significant reduction in benthic ¹³C that we attribute to a reduction in the amount of North Atlantic Deep Water being formed in the Nordic Seas.

    Comparisons between the terrestrial and marine records are difficult because we lack sufficient chronological control for the terrestrial sites. However, by drawing together a number of independent lines of evidence—for example, by correlating terrestrial and marine pollen sequences and counting annual layers in lake sediments—we can construct a preliminary terrestrial chronology for European vegetation events of the last interglacial.

Records Confirm Sudden Chill

Peak interglacial climatic conditions allow certain plant species to proliferate: the `yew phase' north of the Alps, which coincided with an `olive phase' and maximum expansion of Mediterranean vegetation in southern Europe. The olive phase, can be traced in marine pollen records from the eastern Mediterranean and coincides with the deposition of sapropel S5, which is associated with increased monsoonal activity during maximum Northern Hemisphere insolation (125–126 ka).

    Following this period, temperate conditions persisted for about 3500 years--as shown by the expansion of the hornbeam deciduous tree (Carpinus betula) across Europe. Only after this period do pollen sequences suggest climate instability. The evidence suggests that the cold events recorded in various pollen sequences occurred much later than the disputed GRIP events, and they appear to have persisted for only hundreds of years rather than thousands. The timing (about 122 ka) and duration of the cold oscillation recorded after the hornbeam expansion observed in lake sediments from Northern Europe appears to correlate more convincingly with the intra-Eemian cold event seen in the marine record from the Ocean Drilling Program.

    Further support for the Eemian cold event comes indirectly from coral reef records. It is possible to determine the age of a piece of coral by analyzing the amount of radioactive uranium it contains. This is because radioactive uranium breaks down or "decays" over time. These dated corals indicate that the last interglacial period lasted at least between 130 and 117 ka. Moreover, evidence suggests that globally, the main episode of coral reef building was confined to a narrow range of dates between 127–122 ka. Interestingly, the period of reef building seems to end at the same time the Eemian cold event begins, that is, at about 122 ka. The mid-Eemian isotope event near 122 ka may also be concurrent with the intense cooling and freshening observed west of Ireland and in the northern Norwegian Sea.

Gauging Future Climate Change

Today, a similar freshening of the North Atlantic Drift and the Norwegian Current would reduce deep water formation in the Nordic seas. This would slow the great global deep-water conveyor belt and would lead to colder conditions in Europe. We suggest that a similar chain of events may have occurred in the mid-Eemian. Based on the absence of ice, rafted debris, and other evidence of melting icebergs during the Eemian, the observed freshening of the surface waters and the corresponding reduction of deep water formation cannot be ascribed to an ice surge event but rather to overlapping processes: the increased incursion of "fresh" North Pacific water via the Bering Strait during times of raised sea level, and enhanced precipitation over the North Atlantic and Nordic Seas following the Northern Hemisphere insolation maximum.

    There is no cold event of similar magnitude in the Holocene record, but minor late Holocene ¹⁸O and ¹³C events are found in marine records, about 5500 calendar years ago during the early Subboreal chronozone. The Holocene and Intra-Eemian cold events (refer to figure) are similar in that they both occur after the interglacial peak, and they signal the beginning of a trend towards colder conditions, which, thousands of years later, resulted in the beginning of a glacial period. However, no such clear oscillations have been reported in European Holocene pollen records.

A comparison of oxygen and carbon isotope records of the benthic foraminifera C. wuellerstorfi from a) the Last Glacial Maximum (LGM)-Holocene transition and b) the Stage 6-Eemian transition. The records were collected at Ocean Drilling Project site 658. Note that the carbon isotope records of the Stage 6-Eemian and the (LGM)-Holocene are offset by approximately 0.4% due to the long-term 400-ka cyclicity observed at this site.

    In conclusion, the intra-Eemian cold event most likely lasted no longer than 400 years, and it may have had a profound, short-term effect. The Eemian cold event raises two major questions about future climate change: Will global warming increase precipitation over the North Atlantic, and could this lead to an Eemian-type shutdown of deep water formation that would bring with it cold conditions for Europe?

    Source: Eos, September 10, 1996, p. 353.

GLOSSARY