Crystal size variations in Eemian-age ice from the GRIP ice core, Central Greenland

Continuous measurements of ice crystal size have been carried out on an 80 m sequence between 2790 and 2870 m depth in the GRIP ice core from Central Greenland. The ice in this interval is at present considered to orginate from the Eemian interglacial period. The record reveals that the crystal size in ice older than 100,000 yr is highly dependent on climatic conditions at the time of snowfall. This dependence shows up as a strong correlation between ?18O values and crystal size throughout the Eemian, as well as a negative correlation between crystal size and several soluble and insoluble impurities. Although high-resolution impurity records are available from selected parts of the Eemian ice, the study is not conclusive on which impurities are most effective in slowing grain growth. It is shown that the normal grain-growth process, commonly observed in the upper few hundred metres of polar ice sheets, does not yield grain sizes compatible with observed ones at this depth in the ice sheet, even in those parts of the Eemian ice where impurity drag effects are not present. Polygonization of crystals within the ice sheet and the nucleation and rapid growth of new grains at relatively high temperatures in the lowest part probably play an important role in producing the observed grain-size variations. The relevance of possible flow disturbances of the GRIP Eemian climatic record for the results presented is discussed briefly.

Age models and summer sea surface temperature and winter sea ice concentration for the EPILOG-LGM time slice in the Pacific Southern Ocean

Sea surface temperatures and sea-ice extent are the most critical variables to evaluate the Southern Ocean paleoceanographic evolution in relation to the development of the global carbon cycle, atmospheric CO2 variability and ocean-atmosphere circulation. In contrast to the Atlantic and the Indian sectors, the Pacific sector of the Southern Ocean has been insufficiently investigated so far. To cover this gap of information we present diatom-based estimates of summer sea surface temperature (SSST) and winter sea-ice concentration (WSI) from 17 sites in the polar South Pacific to study the Last Glacial Maximum (LGM) at the EPILOG time slice (19,000-23,000 cal. years BP). Applied statistical methods are the Imbrie and Kipp Method (IKM) and the Modern Analog Technique (MAT) to estimate temperature and sea-ice concentration, respectively. Our data display a distinct LGM east-west differentiation in SSST and WSI with steeper latitudinal temperature gradients and a winter sea-ice edge located consistently north of the Pacific-Antarctic Ridge in the Ross sea sector. In the eastern sector of our study area, which is governed by the Amundsen Abyssal Plain, the estimates yield weaker latitudinal SSST gradients together with a variable extended winter sea-ice field. In this sector, sea-ice extent may have reached sporadically the area of the present Subantarctic Front at its maximum LGM expansion. This pattern points to topographic forcing as major controller of the frontal system location and sea-ice extent in the western Pacific sector whereas atmospheric conditions like the Southern Annular Mode and the ENSO affected the oceanographic conditions in the eastern Pacific sector. Although it is difficult to depict the location and the physical nature of frontal systems separating the glacial Southern Ocean water masses into different zones, we found a distinct temperature gradient in latitudes straddled by the modern Southern Subtropical Front. Considering that the glacial temperatures north of this zone are similar to the modern, we suggest that this represents the Glacial Southern Subtropical Front (GSSTF), which delimits the zone of strongest glacial SSST cooling (>4K) to its North. The southern boundary of the zone of maximum cooling is close to the glacial 4°C isotherm. This isotherm, which is in the range of SSST at the modern Antarctic Polar Front (APF), represents a circum-Antarctic feature and marks the northern edge of the glacial Antarctic Circumpolar Current (ACC). We also assume that a glacial front was established at the northern average winter sea ice edge, comparable with the modern Southern Antarctic Circumpolar Current Front (SACCF). During the glacial, this front would be located in the area of the modern APF. The northward deflection of colder than modern surface waters along the South American continent leads to a significant cooling of the glacial Humboldt Current surface waters (4-8K), which affects the temperature regimes as far north as into tropical latitudes. The glacial reduction of ACC temperatures may also result in the significant cooling in the Atlantic and Indian Southern Ocean, thus may enhance thermal differentiation of the Southern Ocean and Antarctic continental cooling. Comparison with temperature and sea ice simulations for the last glacial based on numerical simulations show that the majority of modern models overestimate summer and winter sea ice cover and that there exists few models that reproduce our temperature data rather well.

Marine ice-rafted debris records and age determinations along the northern Eurasian margin

Ice-rafted debris (IRD) (>2 mm), input in eight sediment cores along the Eurasian continental margin (Arctic Ocean), have been studied over the last two glacial/interglacial cycles. Together with the revised chronologies and new micropaleontological data of two cores from the northern Barents Sea (PS2138) and northeastern Kara Sea (PS2741) spanning Marine Isotope Stages (MIS) 6 to 1, the IRD data give new insights into the glacial history of northern Eurasian ice-sheets over the last 150 ka. The chronologies of the cores are based on stable isotope records, AMS 14C datings, paleomagnetic and biostratigraphic data. Extensive episodes of northern Barents Sea ice-sheet growth, probably to the shelf edge, occurred during the late Weichselian (MIS 2) and the Saalian (MIS 6). Major IRD discharge at the MIS 4/3-transition hints to another severe glaciation, probably onto the outer shelf, during MIS 4. IRD-based instabilities of the marine-based ice margin along the northern Barents Sea between MIS 4 and 2 are similar in timing with North Atlantic Heinrich events and Nordic Seas IRD events, suggesting similar atmospheric cooling over a broad region or linkage of ice-sheet fluctuations through small sea-level events. In the relatively low-precipitation areas of eastern Eurasia, IRD peak values during Termination II and MIS 4/3-transition suggest a Kara Sea ice-sheet advance onto the outer shelf, probably to the shelf edge, during glacial MIS 6 and 4. This suggests that during the initial cooling following the interglacials MIS 5, and possibly MIS 7, the combined effect of sustained inflow of Atlantic water into the Arctic Ocean and penetration of moisture-bearing cyclones into easterly direction supported major ice build-up during Saalian (MIS 6) and Mid-Weichselian (MIS 4) glaciation. IRD peak values in MIS 5 indicate at least two advances of the Severnaya Semlya ice-sheet to the coast line during the Early Weichselian. In contrast, a distinct Kara Sea ice advance during the Late Weichselian (MIS 2) is not documented by the IRD records along the northeastern Kara Sea margin.