International Workshop Global Ice Sheets and Sea Level during the Last Glacial Maximum

(EPILOG project; partly sponsored
by the Commission),
Timberline Lodge, Mt. Hood, Oregon, USA,
October 1-5, 2000.

What was the extent and volume of ice during the Last Glacial Maximum (LGM) around 21,000 years ago? These ice sheets influenced global climate by affecting the planetary albedo, atmospheric and ocean circulation, and the hydrological cycle. Growth and decay of ice sheets and concomitant changes of sea-level caused global isostatic adjustments that continue today, thousands of years following the termination of the last ice age. Most of the excess ice of the LGM was locked up in large Northern Hemisphere ice sheets where today only the Greenland Ice Sheet remains. Resolving the three-dimensional distribution of this ice, as well as the volume of excess ice in the Antarctic Ice Sheet and smaller glacial centers, remains a critical but unrealized objective in understanding the dynamics of glacial cycles.

Sixty participants from nine nations debated these issues at an international workshop on "Ice Sheets and Sea Level of the Last Glacial Maximum" held at Timberline Lodge, Oregon, on October 1-5, 2000. This was the second in a series of planned workshops on the Last Glacial Maximum under the auspices of EPILOG (Environmental Processes of the Ice Age: Land, Oceans, Glaciers) program. EPILOG?s goal is a comprehensive reconstruction of the state of the earth during (and the transitions into and out of) a full glacial state, and understanding the processes involved in those changes (Mix et al., 2001). EPILOG is revisiting the landmark CLIMAP study of the LGM earth, which created maps of sea-surface temperatures, glacial ice, and albedo (CLIMAP Project Members, 1981), as a test of our understanding of climate processes and models that predict future climate changes. This attention, and particularly the mismatches between an evolving array of data and model predictions, yields insight into the sensitivity of Earth systems to change, the role of external forcing relative to internal feedback in driving natural change, and linkages between various climate subsystems.

Some aspects of the CLIMAP reconstruction have been questioned, and new methods developed since the CLIMAP project offer the opportunity for a more complete understanding of the ice-age climate system, including the distribution and climatic effects of ice sheets. For the ice sheets, CLIMAP presented two optional reconstructions: a minimum model in which ice margins were largely restricted to continental margins and accounted for 127 m of sea-level lowering, and a "maximum" model which included significantly expanded marine ice sheets and accounted for 163 m of sea-level lowering. The maximum reconstruction, which included a high-elevation single-domed Laurentide Ice Sheet (LIS), became the standard ice-sheet boundary condition used in many model simulations of LGM climate. Nevertheless, significant challenges were mounted against this reconstruction based initially on the interpretation of field evidence for less extensive ice, particularly in the high Arctic, and then on the basis of Earth models that infer the distribution of ice loading by inverting relative sea level records. These approaches suggested substantially thinner ice sheets that, by one estimate, accounted for equivalent sea-level lowering of 105 m.

Other evidence of global ice volume has presented similarly conflicting results. The growth of land-based ice causes lower sea level and changes in ocean chemistry. The best sea-level records come from drilling coral reefs that can then be precisely dated with U/Th and ¹⁴C. These records have yet to sample the full LGM interval, however, and currently only constrain sea level to have been at least 125 m lower at ~19,000 calendar years ago. The growth of continental ice sheets also causes a change in the d¹⁸O composition of the global ocean, but other factors (temperature, local salinity) that affect this signal partially obscure the ice-volume component measured in carbonate fossils.

The first issue addressed by workshop participants concerned the glacial geologic record of ice sheet extent and height. Extensive geophysical surveys of the Antarctic continental shelf have improved our understanding of the LGM extent of the Antarctic Ice Sheet. These data suggest that ice on the shelf rested largely on a widespread deformable substrate, implying that ice, while more extensive than present, was also relatively thin. The extent of the three ice sheets that existed on North America (Laurentide, Cordilleran, and Innuitian) is now reasonably well known and compares favorably to the CLIMAP maximum reconstruction. Radiocarbon dating indicates that the LIS, which was the largest of the former Northern Hemisphere ice sheets, may have advanced rapidly to its maximum extent as early as 23-25 ¹⁴C ka, or well before the global LGM, while the Cordilleran Ice Sheet, which was small at 25 ¹⁴C ka, reached its maximum extent up to 4000 years after the LGM. The outlines of the LGM Scandinavian and British Ice Sheets are well established. Like the Innuitian Ice Sheet, the Barents Ice Sheet was a component of the CLIMAP maximum model that many glaciologists initially questioned, but it is now widely accepted. A serious debate over ice-sheet extent concerns the Kara Sea Ice Sheet in northwestern Russia. Geomorphic evidence demonstrates that an extensive ice sheet once covered this region, but new chronological data seem to indicate that the LGM ice sheet was much smaller than this ancient maximum. Another debate centered on ice extent in eastern Siberia, supported based on geomorphic evidence, but questioned by ¹⁴C-dated sediment records. Because the relative contribution of these ice sheets to the total extent of ice is small (perhaps 5% of the total area), a general consensus that global ice sheet extent is reasonably well known.

Records of past sea level provide the most direct means of determining changes in ice volume, but many of these records are incomplete or poorly dated. Drilling tropical coral reefs offers the best opportunity to develop a well-dated and relatively continuous sea-level record. Although no record of dated corals spans the full LGM interval, corals from Tahiti, New Guinea, and Barbados provide coherent sea level records of the last 19,000 yr that reveal important information about the rates of sea-level rise during the last deglaciation.

There is considerably less glacial geologic information on ice sheet height, and most of these data are restricted to the former ice-sheet margins, revealing little about the interior of the ice sheets where most mass is found. Glaciological models provide a powerful means of evaluating former ice sheet height, particularly when ice sheet extent is well constrained. Ice sheet models are also able to partition the global ice suggested by sea-level records among the areas suggested by the glacial geologic record. Current three-dimensional thermodynamic ice-sheet models have improved significantly since CLIMAP. However, processes at the base of ice sheets that are important in controlling ice movement and can strongly influence ice thickness are still poorly represented. With the exception of the West Antarctic Ice Sheet, these processes likely only play a relatively minor role in the dynamics of the Greenland and East Antarctic Ice Sheets, which may explain why model simulations are able to reproduce many of the general features of modern ice sheets. These processes may have played a much larger role in the dynamics of former Northern Hemisphere ice sheets, however, so that a better representation of them in models is required in order to address LGM issues. Other factors, such as the extent of iceberg calving and sensitivity of ice sheets to climate, also contribute to uncertainties in modeling former ice sheets.

Geochemical records of oxygen isotopes (d¹⁸O) from deep-sea cores and ice cores provide supporting evidence for the volume of ice at the LGM, and for the timing of key changes in the ocean-climate system. Detailed records of d¹⁸O reveal that the last isotopic maximum (LIM, near 18 ka calendar) is younger than the LGM as defined by sea level lowstand (~21 ka calendar). This implies either deep-sea cooling in the three-thousand year interval between these maxima (to maintain high d¹⁸O in calcite shells in spite of reduced seawater d¹⁸O), or conversion of land ice to floating ice (which would raise sea level without lowering seawater d¹⁸O), or an exceptionally long lag in the propagation of the d¹⁸O signal of ice melting through the ocean system. After the LIM interval, the first rapid shift to lower d¹⁸O values in deep-sea foraminifera began near 17 ka calendar, earlier than well-known meltwater pulse 1A which occurred near 14 ka calendar. This mismatch suggests early warming of the deep sea, and implies that deep ocean circulation may play a key role in the termination of the last ice age.

Significant debate ensued regarding isotopic constraints on the total volume of LGM ice sheets. A growing data set of pore water d¹⁸O measurements, when deconvolved, suggest a total range of sea-water d¹⁸O change of about 1o/oo. This finding seems to require substantial cooling of the deep sea during the LGM, to nearly the freezing point, and based on comparisons to glaciological reconstructions suggests that glacier ice was isotopically less depleted than previously suggested by atmospheric-isotopic models, and some data sets on fossil groundwaters and subglacial calcite. Resolving this issue will require a precise, non-isotopic, temperature proxy. Progress here has come from precise measurements of Mg/Ca ratios in planktonic foraminifera. It remains uncertain whether such proxies can be applied with confidence to deep-sea foraminifera, but present data support the concept of substantial temperature changes that affect the d¹⁸O record in calcite shells.

Data from d¹⁸O in molecular oxygen, preserved in polar ice cores, was debated. A key disagreement concerns how to remove the imprint of changing isotopic character of oxygen produced by land and marine plants, the Dole Effect, from changes in the isotopic budget of surface seawater on atmospheric oxygen. In one view, the Dole Effect is dominated by long-term cycles near 100,000 years (similar to orbital eccentricity) and 23,000 years (similar to orbital precession). This implies that sea-water d¹⁸O changes had little or no 100,000 year cycle, and challenges long-held views that this cyclic climate changes that dominates most records of climate was associated with the global ice-sheet system. An alternative view, that temperature-corrected marine d¹⁸O records can constrain the history of the Dole Effect, suggests that the 100,000 year cycle was prevalent in the ice-sheet system, but that each glacial termination may have had some unique features unrelated to direct orbital forcing of the climate system. Many uncertainties remain, including the need for better proxies of biological production that influence the Dole Effect on atmospheric oxygen, and for better ways of linking the chronologies of paleoclimatic timeseries developed in ice cores, in marine sediment cores, and from deposits preserved on land.