John T. Andrews

Forcing of the cold event of 8,200 years ago by catastrophic drainage of Laurentide lakes

The sensitivity of oceanic thermohaline circulation to freshwater perturbations is a critical issue for understanding abrupt climate change. Abrupt climate fluctuations that occurred during both Holocene and Late Pleistocene times have been linked to changes in ocean circulation, but their causes remain uncertain. One of the largest such events in the Holocene occurred between 8,400 and 8,000 calendar years ago,, (7,650–7,200 14C years ago), when the temperature dropped by 4–8 °C in central Greenland and 1.5–3 °C at marine, and terrestrial, sites around the northeastern North Atlantic Ocean. The pattern of cooling implies that heat transfer from the ocean to the atmosphere was reduced in the North Atlantic. Here we argue that this cooling event was forced by a massive outflow of fresh water from the Hudson Strait. This conclusion is based on our estimates of the marine 14C reservoir for Hudson Bay which, in combination with other regional data, indicate that the glacial lakes Agassiz and Ojibway, (originally dammed by a remnant of the Laurentide ice sheet) drained catastrophically ∼8,470 calendar years ago; this would have released >1014 m3 of fresh water into the Labrador Sea. This finding supports the hypothesis,, that a sudden increase in freshwater flux from the waning Laurentide ice sheet reduced sea surface salinity and altered ocean circulation, thereby initiating the most abrupt and widespread cold event to have occurred in the past 10,000 years.

The Laurentide and Innuitian ice sheets during the Last Glacial Maximum.

The Late Wisconsinan advance of the Laurentide Ice Sheet started from a Middle Wisconsinan interstadial minimum 27–30 14 C ka BP when the ice margin approximately followed the boundary of the Canadian Shield. Ice extent in the Cordillera and in the High Arctic at that time was probably similar to present. Ice advanced to its Late Wisconsinan (stage 2) limit in the northwest, south, and northeast about 23–24 14 C ka BP and in the southwest and far north about 20–21 14 C ka BP. In comparison to some previous reconstructions of ice extent, our current reconstruction has substantially more Late Wisconsinan ice in the High Arctic, where an Innuitian Ice Sheet is generally acknowledged to have existed, in the Atlantic Provinces, where ice is now thought to have extended to the Continental Shelf edge in most places, and on eastern Baffin Island, where ice probably extended to the fiord mouths rather than to the fiord heads. Around most of the ice margin, the Late Wisconsinan maximum ice extent either exceeded the extent of earlier Wisconsinan advances or it was similar to the Early Wisconsinan advance. Ice marginal recession prior to 14 14 Ck a BP occurred mainly in deep water and along the southern terrestrial fringe. However, Heinrich event 1 probably drew down the entire central ice surface at 14.5 14 C ka BP sufficiently to displace the Labrador Sector outflow centre 900 km eastward from the coast of Hudson Bay. The onset of substantial ice marginal recession occurred about 14 14 C ka BP in the northwest, southwest, and south but not until about 10–11 14 C ka BP in the northeast and in the High Arctic. Thus, the period of maximum ice extent in North America generally encompasses the interval from B24/21 to 14 14 C ka BP, or considerably longer than the duration of the LGM defined as occurring during a period of low global sea level as well as during a time of relative climate stability B18 14 C ka BP. The interval of advance of much of the Laurentide Ice Sheet to its maximum extent (between B27 14 C ka BP and B24 14 C ka BP) coincides with a suggested interval of rapid fall in global sea level to near LGMlevels. r 2001 Elsevier Science Ltd. All rights reserved.

Detrital carbonate-rich sediments, northwestern Labrador Sea: Implications for ice-sheet dynamics and iceberg rafting (Heinrich) events in the North Atlantic

Much of the bed of the central and eastern sectors of the Laurentide Ice Sheet was underlain by Paleozoic carbonates. We propose that pulses of detrital carbonate-rich sediments in two cores from the northwestern Labrador Sea reflect episodes when an ice stream from the Hudson Strait extended to the shelf break and delivered sediment onto the slope and deep-sea plain. Atomic mass spectroscopy 14 C dating of planktonic foraminifera in cores HU75-009-IV-55 (2.4 km water depth) and HU87-033-009 (1.4 km water depth and 150 km north-northwest), indicates that the youngest event occurred between ca.14 and 15 ka and another occurred between 19 and 21 ka. The two carbonate intervals in northwestern Labrador Sea cores are coeval with Heinrich events 1 and 2 in the eastern North Atlantic (lat 45°-50°N, long 20°W), where they are associated with an increase in lithic fragments and a drastic reduction in the numbers of foraminifera. These changes have been linked with the massive production of icebergs associated with ice-sheet surges. Our evidence indicates that Heinrich events 1 and 2 are associated with the dynamics of the Hudson Strait ice stream and denote considerable glaciological instability.

Iceberg production, debris rafting, and the extent and thickness of Heinrich layers (H-1, H-2) in North Atlantic sediments

The pattern of Heinrich-layer distribution for the last two events (H-1, ∼14.5 and H-2, ∼21.1 ka), mapped from magnetic susceptibility analysis of more than 50 North Atlantic Ocean cores, provides the most detailed information to date on their extent and thickness. An integrated spatial average thickness for the layers is 10–15 cm, and there is a strong distance decay eastward. The pattern of deposition over the North Atlantic is similar for events H-1 and H-2, indicating that icebergs followed similar drift tracks. Rates of iceberg production and sediment flux from the Hudson Strait drainage basin of the North American Laurentide ice sheet, the major iceberg source for the events, were calculated by using a mass-balance approach. This provides an envelope of sedimentation rates and the prediction that it would take between 50 and ∼1250 yr of iceberg sediment delivery to accumulate a Heinrich layer averaging 10 cm thick over the North Atlantic, depending on the model assumptions used. The most likely duration of Heinrich events is 250–1250 yr.

The use of hypsometry to indicate long-term stability and response of valley glaciers to changes in mass transfer

A simple equation is derived relating the net mass-balance and hypsometric curves of a steady-state valley glacier. It is used to examine how valley shape is linked to disparate extents and responses of glaciers subjected to similar climatic conditions. Examples are given which show that area-based indices (e.g. AAR) for estimating the equilibrium line altitude (ELA) may be subject to a substantial built-in variance because they implicitly rely upon similarity of glacier shape and regimen over a region. If accurate topographic maps are available, the equation may be used to infer the regimen of modern glaciers in the form of a dimensionless ratio of net mass-balance gradients. Alternatively, if similar information is available concerning regional glacier regimen, disparate extents and responses may be collectively utilized to estimate values of ELA or to infer climatic influence, taking glacier hypsometry into account.

Chronology of late Wisconsin ice retreat from the western Ross Sea, Antarctica

Lithologic data from marine sedimentary cores and accelerator mass spectrometer (AMS) radiocarbon dates indicate that grounded ice did not advance to the western Ross Sea continental shelf edge during the last glacial maximum (LGM). A chronology of the timing of ice retreat was provided by 26 AMS dates, obtained from 12 cores. Dates ranging from 20 to 29 ka suggest that the outer continental shelf (beyond ca. lat 74°S) was not covered by grounded glacial ice prior to and during the early stages of the LGM. 14C dates just above transitions from subglacial diamictons to marine muds indicate that the area around the Drygalski ice tongue was deglaciated by at least 11.5 ka. The Ross Ice Shelf reached its present-day position near Ross Island by about 7 ka. A hiatus of about 15 ka is apparent in radiocarbon dates in cores from the outer continental shelf. The hiatus is interpreted to represent an absence of sedimentation caused by the presence of an ice shelf.

Preboreal oscillation caused by a glacial Lake Agassiz flood

The Preboreal oscillation (PBO) has been attributed to increased meltwater, but the source of the meltwater and causative mechanism of the PBO has remained elusive. Here we attribute the source to a massive meltwater discharge event from an abrupt drainage of glacial Lake Agassiz, Canada, via the Mackenzie River into the Arctic Ocean. A maximum volume of 21,000 km 3 was discharged over a 1.5–3 yr period with a peakdischarge of 0.500 Sverdrups (Sv), equivalent to a 6 m rise in the Arctic Ocean (or 0.062 m rise in global sea level). The flood occurred at about 11,335 cal yr BP, and was followed by a B0.042 Sv flow until 10,750 cal yr BP when the southern outlet of Lake Agassiz reopened and diverted drainage to the Mississippi River system. We estimate that only 2–4% of the flood water would have frozen into sea ice within the Beaufort region, but coupled with increased river ice production during winter, and thicker pack ice growth throughout the Arctic Ocean, a thicker, longer lasting and more extensive packice may have been flushed through Fram Strait. The thicker and more extensive packice, and freshened sea surface, may have triggered the PBO by increasing albedo, and generating a low salinity anomaly upon melting in the North Atlantic, thus decreasing the formation of North Atlantic Deep Water. r 2002 Elsevier Science Ltd. All rights reserved.

Abrupt changes (Heinrich events) in late Quaternary North Atlantic marine environments: a history and review of data and concepts

In the last several years evidence has mounted that a series of abrupt changes in sediment delivery and palaeoceanography effected the North Atlantic; they are recorded in a variety of proxy records but most dramatically in changes of grain-size and mineralogy associated with postulated iceberg rafting events, specifically Heinrich (H) events. This review paper examines the evidence for such evidence prior to the development of the ‘Heinrich event’ concept (1988–1992), and then examines the explosion of data and correlations that stemmed from the acceptance and use of this paradigm. Specific attention is focused on ideas pertaining to the cause(s) of glaciological mechanisms, sediment delivery, and sediment source, and major gaps in our understanding of the underlying glaciological, glacial geological, and glacial marine processes are stressed. Abrupt changes across boundaries associated with H-2 in the Labrador Sea are illustrated from core HU87033-009LCF. The case is made that correlations between records need to be more rigorous, especially within the range of radiocarbon dating, but it appears that abrupt changes in the delivery of iceberg rafted debris, which characterize these intervals, are multisourced. It is unclear how global ‘climate’ can simultaneously influence the abrupt behaviour of ice margins in different parts of the world—for marine-based margins, changes in relative sea-level are one possibility. A qualitative model is presented for triggering H events, which includes global changes in mass balance, changes in relative sea-level at ice margins due to glacial isostasy, and changes in the basal thermal regime of tidewater ice-streams.

Distinguishing subglacial till and glacial marine diamictons in the western Ross Sea, Antarctica: Implications for a last glacial maximum grounding line

Analyses of lithology, stratigraphy, and tephra from marine sediment cores collected from the western Ross Sea during cruises Eltanin 32 and 52 and Deep Freeze 80 and 87 indicate that subglacial till does not extend to the continental shelf edge. Subglacial till occurs as the lowest unit in most cores landward (south) of approximately 74°S, while seaward of approximately 74°S, the lowest diamicton units are glacial marine diamictons. Glacial marine diamictons are distinguished from subglacial tills by the presence of higher and more variable total organic carbon content downcore, distinct tephra layers, stratification, higher diatom and foraminifera abundances, higher sand content, and radiocarbon dates in chronological order downcore. Sand-sized tephra layers from two cores on the outer continental shelf are interpreted as single eruptive events, one likely to have been derived from the Mount Melbourne volcano and the other from the Pleiades volcano. Radiocarbon dates from sediment above and below the tephra layer in one of these cores (Df87-32) show that deposition indicative of open-water conditions occurred between 22 and 26 ka in the western Ross Sea.

Growth rate of the Laurentide Ice Sheet and sea level lowering (with emphasis on the 115,000 BP sea level low)

Abstract A physically plausible three-dimensional numerical ice flow model is used to examine the rate at which the Laurentide Ice Sheet could spread and thicken using as input likely values for the rate of fall of snowline and the amount of net mass balance over the growing ice sheet. This provides then both a test of the hypothesis of “instantaneous glacierization” and of the suggested rapid fall of world sea level to between −20 and −70 m below present at 115,000 BP. Two experiments are described: The first terminated after 10,050 years of model run with ice sheets centered over Labrador-Ungava and Baffin Island with a total volume of 3.0 × 106 km3 of ice, whereas the second was completed after 10,000 years and resulted in a significantly larger ice sheet (still with two main centers) with a volume of 7.78 × 106 km3 of ice. This latter figure is equivalent to the mass required to lower world sea level by 19.4 m. Our results indicate that large ice sheets can develop in about 10,000 years under optimum conditions.