Geochemical constraints on the Laurentide Ice Sheet contribution to Meltwater Pulse 1A

Planktonic and benthic δ¹⁸O records adjacent to the runoff outlets of the Laurentide Ice Sheet (LIS) indicate that the LIS contributed to the abrupt ～20 m rise in sea level ～14.6 ka, Meltwater Pulse 1A (MWP-1A). However, the magnitude of the LIS contribution still remains unresolved. Here, I use a freshwater runoff–ocean mixing model to calculate the LIS meltwater required to explain the decreases in planktonic and benthic δ¹⁸O observed during MWP-1A at the southern, eastern and northern runoff outlets of the LIS. Maximum LIS contributions in equivalent sea level rise for a 500-year long MWP-1A are 2.7 m discharged into the Gulf of Mexico as a combined hyperpycnal and hypopycnal flow, 2.1 m discharged into the North Atlantic, and 0.5 m into the Arctic Ocean, for a total LIS contribution of ≤5.3 m. A LIS contribution of <30% to MWP-1A supports the hypothesis that a significant component of this MWP was sourced from the Antarctic Ice Sheet.     

Climate-induced fluvial dynamics in tropical Africa around the last glacial maximum?

The Greenland and East and West Antarctic ice sheets are assessed as being the source of ice that produced an Eemian sea level 6 m higher than present sea level. The most probable source is total collapse of the West Antarctic Ice Sheet accompanied by partial collapse of the adjacent sector of the East Antarctic Ice Sheet in direct contact with the West Antarctic Ice Sheet. This conclusion is reached by applying a simple formula relating the “floating fraction” of ice along flowlines to ice height above the bed. Increasing the floating fraction lowered ice elevations enough to contribute up to 4.7 m to global sea level. Adding 3.3 m resulting from total collapse of the West Antarctic Ice Sheet accounts for the higher Eemian sea level. Partial gravitational collapse that produced the present ice drainage system of Amery Ice Shelf contributes 2.3 m to global sea level. These results cast doubt on the presumed stability of the East Antarctic Ice Sheet, but destabilizing mechanisms remain largely unknown. Possibilities include glacial surges and marine instabilities at the respective head and foot of ice streams.     

Deglacial dynamics of the Vestfjorden – Trænadjupet palaeo‐ice stream,northern Norway

Few well‐dated records of the deglacial dynamics of the large palaeo‐ice streams of the major Northern Hemisphere ice sheets are presently available, a prerequisite for an improved understanding of the ice‐sheet response to the climate warming of this period. Here we present a transect of gravity‐core samples through Trænadjupet and Vestfjorden, northern Norway, the location of the Trænadjupet – Vestfjorden palaeo‐ice stream of the NW sector of the Fennoscandian Ice Sheet. Initial ice recession from the shelf break to the coastal area (~400 km) occurred at an average rate of about 195 m a⁻¹, followed by two ice re‐advances, at 16.6–16.4 ka BP (the Røst re‐advance) and at 15.8–15.6 ka BP (the Værøy re‐advance), the former at an estimated ice‐advance rate of 216 m a⁻¹. The Røst re‐advance has been interpreted to be part of a climatically induced regional cold spell while the Værøy re‐advance was restricted to the Vestfjorden area and possibly formed as a consequence of internal ice‐sheet dynamics. Younger increases in IRD content have been correlated to the Skarpnes (Bølling – Older Dryas) and Tromsø – Lyngen (Younger Dryas) Events. Overall, the decaying Vestfjorden palaeo‐ice stream responded to the climatic fluctuations of this period but ice response due to internal reorganization is also suggested. Separating the two is important when evaluating the climatic response of the ice stream. As demonstrated here, the latter may be identified using a regional approach involving the study of several palaeo‐ice streams. The retreat rates reported here are of the same order of magnitude as rates reported for ice streams of the southern part of the Fennoscandian Ice Sheet, implying no latitudinal differences in ice response and retreat rate for this ~1000 km² sector of the Fennoscandian Ice Sheet (~60–68°N) during the climate warming of this period.     

Predicted relative sea-level changes (18,000 years B.P. to present) caused by late-glacial retreat of the Antarctic Ice Sheet

Predictions of global changes in relative sea level caused by retreat of the Antarctic Ice Sheet from its 18,000 yr B.P. maximum to its present size are calculated numerically. When combined with the global predictions of relative sea-level change resulting from retreat of the Northern Hemisphere ice sheets, the results may be compared directly to observations of sea-level change on the Antarctic continent as well as at distant localities. The comparison of predictions to the few observations of sea-level change on Antarctica supports the view that the Antarctic Ice Sheet was larger 18,000 years ago than at present. The contribution of the Antarctic Ice Sheet to the total eustatic sea-level rise is assumed to be 25 m (25% of the assumed total eustatic rise). If as little as 0.7 m of this 25-m rise occurred between 5000 yr B.P. and the present, few mid-oceanic islands would emerge. If the Antarctic Ice Sheet attained its present dimensions by 6000 yr B.P., however, and if the volume of the ocean has remained constant for the past 5000 years, numerous islands throughout the Southern Hemisphere would emerge. It is suggested that a thorough study of Pacific islands, believed by some to have slightly emerged shorelines of Holocene age, would yield useful information about ocean volume changes during the past 5000 years, and hence on the glacial history of the Antarctic Ice Sheet.     

Younger Dryas cirque glaciers in western Spitsbergen: smaller than during the Little Ice Age

The outermost moraines in front of the Scottbreen glacier in Spitsbergen date from c . AD 1900. These moraines rest on top of a marine shoreline radiocarbon-dated to about 11 200 ¹⁴C yr BP and demonstrate that the AD-1900 moraines show the maximum glacier extent since late Allerød time. This means that Scottbreen was smaller during the Younger Dryas than at AD 1900, in contrast with glaciers on mainland western Europe, which were all much larger during the Younger Dryas. The explanation is probably starvation of precipitation on western Spitsbergen during the Younger Dryas. In contrast, ice sheets and glaciers in Spitsbergen reacted more or less in concert with glaciers in western Europe, during the global Last Glacial Maximum and the Little Ice Age.     

The subglacial thermal organisation (STO) of ice sheets

This paper examines ice-sheet wide variations in subglacial thermal regime and ice dynamics using the landform record exposed on the beds of former mid-latitude ice sheets (the Laurentide, Cordilleran, Fennoscandian and British-Irish Ice Sheets). We compare the landform patterns beneath these former ice sheets to the flow organisation beneath parts of the contemporary Antarctic Ice Sheet inferred from RADARSAT-1 Antarctic Mapping Project (RAMP) data. The evidence preserved in the landform record and observed on contemporary ice masses can be grouped into four major ice-dynamical components that collectively define the subglacial thermal organisation (STO) of ice sheets. These ice-dynamical components are frozen-bed patches, ice streams, ice-stream tributaries and lateral shear zones. Frozen-bed patches appear at a wide range of spatial scales, spanning four orders of magnitude. In some areas, frozen-bed zones comprise large proportions of the bed (e.g. near the ice divide in continental areas), whilst in other areas they constitute isolated “islands” in areas dominated by thawed-bed conditions. Ice streams, narrow zones of fast flow in ice sheets that are otherwise dominated by slow sheet flow, are also common features of Quaternary ice sheets. Tributaries to ice streams flow at velocities intermediate between full ice-stream and sheet flow, and may divert ice drainage from one primary ice-stream corridor to an adjacent one. Sharp lateral boundaries between landforms indicate sliding and non-sliding conditions, respectively. These lateral boundaries represent important discontinuities in the glacial landscape and mark the location of shear zones between thawed-bed ice streams and intervening frozen-bed areas. We use the landform evidence in the area around Great Bear Lake, Canada to trace the evolution of an ice-stream web through time, demonstrating that frozen-bed patches are integral components of this complex system. We conclude that frozen-bed patches are important for the stability of ice sheets because they laterally constrain and isolate peripheral drainage basins and their ice streams.     

Southwest African climate independent of Atlantic sea surface temperatures during the Younger Dryas

To investigate land–sea interactions during deglaciation, we compared proxies for continental (pollen percentages and accumulation rates) and marine conditions (dinoflagellate cyst percentages and alkenone-derived sea surface temperatures). The proxies were from published data from an AMS-radiocarbon-dated sedimentary record of core GeoB 1023-5 encompassing the past 21,000 years. The site is located at ca. 2000 m water depth just north of the Walvis Ridge and in the vicinity of the Cunene River mouth. We infer that the parallelism between increasing sea surface temperatures and a southward shift of the savanna occurred only during the earliest part of the deglaciation. After the Antarctic Cold Reversal, southeast Atlantic sea surface temperatures no longer influenced the vegetation development in the Kalahari. Stronger trade winds during the Antarctic Cold Reversal and the Younger Dryas period probably caused increased upwelling off the coast of Angola. A southward shift of the Atlantic anti-cyclone could have resulted in both stronger trade winds and reduced impact of the Westerlies on the climate of southwestern Africa.     

The Bølling‐age Blomvåg Beds,western Norway: implications for the Older Dryas glacial re‐advance and the age of the deglaciation

Blomvåg, on the western coast of Norway north of Bergen, is a classical site in Norwegian Quaternary science. Foreshore marine sediments, named the Blomvåg Beds and now dated to the Bølling‐Allerød from 14.8 to 13.3 cal. ka BP, contain the richest Lateglacial bone fauna in Norway, numerous mollusc shells, driftwood, and flint that some archaeologists consider as the oldest traces of humans in Norway. The main theme of this paper is that the Blomvåg Beds are overlain by a compact diamicton, named the Ulvøy Diamicton, which was interpreted previously as a basal till deposited during a glacial re‐advance into the ocean during the Older Dryas (c. 14 cal. ka BP). Sediment sections of the Blomvåg Beds and the Ulvøy Diamicton were exposed in ditches in a cemetery that was constructed in 1941–42 and have subsequently not been accessible. A number of radiocarbon and cosmogenic ¹⁰Be exposure ages demonstrate that the diamicton is not likely to be a till because minimum deglaciation ages (14.8–14.5 cal. ka BP) from the vicinity pre‐date the Ulvøy Diamicton. We now consider that sea ice and icebergs formed the Ulvøy Diamicton during the Younger Dryas. The Scandinavian Ice Sheet margin was located on the outermost coastal islands between at least c. 18.5 and 14.8 cal. ka BP; however, no ice‐marginal deposits have been found offshore from this long period. The Older Dryas ice margin in this area was located slightly inside the Younger Dryas margin, whereas farther south it was located slightly beyond the Younger Dryas margin.     

Timing of the first drainage of the Baltic Ice Lake synchronous with the onset of Greenland Stadial 1

Glacial varves can give significant insights into recession and melting rates of decaying ice sheets. Moreover, varve chronologies can provide an independent means of comparison to other annually resolved climatic archives, which ultimately help to assess the timing and response of an ice sheet to changes across rapid climate transitions. Here we report a composite 1257‐year‐long varve chronology from southeastern Sweden spanning the regional late Allerød–late Younger Dryas pollen zone. The chronology was correlated to the Greenland Ice‐Core Chronology 2005 using the time‐synchronous Vedde Ash volcanic marker, which can be found in both successions. For the first time, this enables secure placement of the Lateglacial Swedish varve chronology in absolute time. Geochemical analysis from new varve successions indicate a marked change in sedimentation regime accompanied by an interruption of ice‐rafted debris deposition synchronous with the onset of Greenland Stadial 1 (GS‐1; 12 846 years before AD 1950). With the support of a simple ice‐flow/calving model, we suggest that slowdown of sediment transfer can be explained by ice‐sheet margin stabilization/advance in response to a significant drop of the Baltic Ice Lake level. A reassessment of chronological evidence from central‐western and southern Sweden further supports the hypothesis of synchronicity between the first (penultimate) catastrophic drainage of the Baltic Ice Lake and the start of GS‐1 in Greenland ice‐cores. Our results may therefore provide the first chronologically robust evidence linking continental meltwater forcing to rapid atmosphere–ocean circulation changes in the North Atlantic.     

A simple holistic hypothesis for the self-destruction of ice sheets

Ice sheets are the only components of Earth’s climate system that can self-destruct. This paper presents the quantitative force balance for bottom-up modeling of ice sheets, as first presented qualitatively in this journal as a way to quantify ice-bed uncoupling leading to self-destruction of ice sheets (Hughes, 2009a). Rapid changes in sea level and climate can result if a large ice-sheet self-destructs quickly, as did the former Laurentide Ice Sheet of North America between 8100 and 7900 BP, thereby terminating the last cycle of Quaternary glaciation. Ice streams discharge up to 90 percent of ice from past and present ice sheets. A hypothesis is presented in which self-destruction of an ice sheet begins when ubiquitous ice-bed decoupling, quantified as a floating fraction of ice, proceeds along ice streams. This causes ice streams to surge and reduce thickness by some 90 percent, and height above sea level by up to 99 percent for floating ice, so the ice sheet undergoes gravitational collapse. Ice collapsing over marine embayments becomes floating ice shelves that may then disintegrate rapidly. This floods the world ocean with icebergs that reduce the ocean-to-atmosphere heat exchange, thereby triggering climate change. Calving bays migrate up low stagnating ice streams and carve out the accumulation zone of the collapsed ice sheet, which prevents its recovery, decreases Earth’s albedo, and terminates the glaciation cycle. This sequence of events may coincide with a proposed life cycle of ice streams that drain the ice sheet. A first-order treatment of these life cycles is presented that depends on the longitudinal force balance along the flowbands of ice streams and gives a first approximation to ice-bed uncoupling at snapshots during gravitational collapse into ice shelves that disintegrate, thereby removing the ice sheet. The stability of the Antarctic Ice Sheet is assessed using this bottom-up approach.     

南极冰盖与中国风尘堆积

通过南极冰盖消长和风尘堆积物的气候事件对比,说明在4.3—2.5 Ma B.P,间南极冰盖增长和北极冰盖发展,中国第三纪风尘红壤土沉积可能是中国东部季风气候雏形形成的反映。2.6Ma以来,随着北极冰盖的迅速扩展,中国内陆进一步干旱化并促成黄土堆积。南极极锋带北移与黄土区极寒等事件的发生具有同步性,据此建立极寒事件的序列,可能反映大气环流强度变化情况。南极东方站冰岩芯气候曲线说明,黄土高原晚更新世以来的干冷期可能形成于0.115Ma B.P.。我国东部地区转暖大致发生在13000aB.P.,此后不久可能出现欧洲新仙女木寒冷事件。此外,大约在40000aB.P.可能有一寒冷事件。     

On the hemispheric origins of meltwater pulse 1a

During the glacial–interglacial transition that began subsequent to the Last Glacial Maximum approximately 21,000 calendar years ago, globally averaged (eustatic) sea-level rose by approximately 120 m as climate warmed to its current (Holocene) state. This rise of relative sea-level (RSL) did not occur smoothly, however, but was characterized by the occurrence of one or more episodes of extremely rapid increase. The most extreme of these events has come to be referred to as meltwater pulse 1a, and was initially identified in the coral based record of RSL history from the island of Barbados in the Caribbean Sea. Although it has usually been assumed that this episode of rapid RSL rise was derivative of a partial collapse of the northern hemisphere ice sheets, it has recently been suggested that this pulse could have originated in a dramatic melt-back of the Antarctic Ice Sheet. In this paper the arguments presented in favour of the southern hemisphere source are revisited in order to assess the plausibility of this alternative scenario.Based upon the analyses presented, it is concluded that the evidence previously provided in support of the southern hemisphere scenario is in fact unable to rule out an entirely northern hemisphere source for the meltwater pulse 1a. Since explicit evidence does exist that both the Laurentide and Fennoscandian ice sheets contributed to this event and that Antarctic ice sheet melting occurred significantly later, the southern hemisphere appears not to have been a prime mover of northern hemisphere events.     

A model for Holocene retreat of the West Antarctic Ice Sheet

Marine ice sheets are grounded on land which was below sea level before it became depressed under the ice-sheet load. They are inherently unstable and, because of bedrock topography after depression, the collapse of a marine ice sheet may be very rapid. In this paper equations are derived that can be used to make a quantitative estimate of the maximum size of a marine ice sheet and of when and how rapidly retreat would take place under prescribed conditions. Ice-sheet growth is favored by falling sea level and uplift of the seabed. In most cases the buttressing effect of a partially grounded ice shelf is a prerequisite for maximum growth out to the edge of the continental shelf. Collapse is triggered most easily by eustatic rise in sea level, but it is possible that the ice sheet may self-destruct by depressing the edge of the continental shelf so that sea depth is increased at the equilibrium grounding line.Application of the equations to a hypothetical “Ross Ice Sheet” that 18,000 yr ago may have covered the present-day Ross Ice Shelf indicates that, if the ice sheet existed, it probably extended to a line of sills parallel to the edge of the Ross Sea continental shelf. By allowing world sea level to rise from its late-Wisconsin minimum it was possible to calculate retreat rates for individual ice streams that drained the “Ross Ice Sheet.” For all the models tested, retreat began soon after sea level began to rise (～15,000 yr B.P.). The first 100 km of retreat took between 1500 and 2500 yr but then retreat rates rapidly accelerated to between 0.5 and 25 km yr^?1, depending on whether an ice shelf was present or not, with corresponding ice velocities across the grounding line of 4 to 70 km yr^?1. All models indicate that most of the present-day Ross Ice Shelf was free of grounded ice by about 7000 yr B.P. As the ice streams retreated floating ice shelves may have formed between promontories of slowly collapsing stagnant ice left behind by the rapidly retreating ice streams. If ice shelves did not form during retreat then the analysis indicates that most of the West Antarctic Ice Sheet would have collapsed by 9000 yr B.P. Thus, the present-day Ross Ice Shelf (and probably the Ronne Ice Shelf) serves to stabilize the West Antarctic Ice Sheet, which would collapse very rapidly if the ice shelves were removed. This provides support for the suggestion that the 6-m sea-level high during the Sangamon Interglacial was caused by collapse of the West Antarctic Ice Sheet after climatic warming had sufficiently weakened the ice shelves. Since the West Antarctic Ice Sheet still exists it seems likely that ice shelves did form during Holocene retreat. Their effect was to slow and, finally, to halt retreat. The models that best fit available data require a rather low shear stress between the ice shelf and its sides, and this implies that rapid shear in this region encouraged the formation of a band of ice with a preferred crystal fabric, as appears to be happening today in the floating portions of fast bounded glaciers.Rebound of the seabed after the ice sheet had retreated to an equilibrium position would allow the ice sheet to advance once more. This may be taking place today since analysis of data from the Ross Ice Shelf indicates that the southeast corner is probably growing thicker with time, and if this persists then large areas of ice shelf must become grounded. This would restrict drainage from West Antarctic ice streams which would tend to thicken and advance their grounding lines into the ice shelf.     

Climatic and environmental changes during the Weichselian Lateglacial Interstadial in the Weerterbos region,the Netherlands

Lake sediment records from the Weerterbos region, in the southern Netherlands, were studied to reconstruct summer temperature and environmental changes during the Weichselian Lateglacial Interstadial. A sediment core obtained from a small lacustrine basin was analysed for multiple proxies, including lithological changes, oxygen isotopes of bulk carbonates, pollen and chironomids. It was found that the oxygen isotope record differed strongly from the other proxies. Based on a comparison with three additional lake sediment records from the same region, it emerged that the oxygen isotope records were strongly affected by local environmental conditions, impeding the distinction of a regional palaeoclimate signal. The chironomid‐inferred July air temperature reconstruction produced inferred interstadial temperatures ranging between ～15° and 18°C, largely consistent with previously published results from the northern part of the Netherlands. A temporary regressive phase in the pollen record, which can be tentatively correlated with the Older Dryas, preceded the expansion of birch woodland. Despite differences between the four pollen records from the Weerterbos region, a comparable regressive vegetation phase that was possibly the result of a shift to drier conditions could be discerned in all of the profiles. In addition, a temporary temperature decline of ～1.5°C was inferred from the chironomid record during this regressive phase. The multi‐proxy approach used here enabled a direct comparison of inferred changes in temperature, vegetation and environmental conditions at an individual site, while the multi‐site approach provided insight into the factors influencing the pollen and isotope records from these small‐scale depressions.     

Lateglacial summer temperatures in the Northwest European lowlands: a chironomid record from Hijkermeer,the Netherlands

Lateglacial environments at Hijkermeer, northwest Netherlands, were reconstructed by means of chironomid, diatom and pollen analyses. Diatom assemblages indicate that Hijkermeer was a shallow, oligo- to mesotrophic lake during this period. Pollen assemblages reflect the typical northwest European Lateglacial vegetation development and provide an age assessment for the record from the beginning of the Older Dryas (ca 14 000 calibrated ¹⁴C yr BP) into the early Holocene (to ca 10 700 calibrated ¹⁴C yr BP). The chironomid record is characterized by several abrupt shifts between assemblages typically found in mid-latitude subalpine to alpine lakes and assemblages typical for lowland environments. Based on the chironomid record, July air temperatures were reconstructed using a chironomid-temperature transfer-function from central Europe. Mean July air temperatures of ca 14.0–16.0 °C are inferred before the Older Dryas, of ca 16.0–16.5 °C during most of the Allerød, of ca 13.5–14.0 °C during the Younger Dryas, and of ca 15.5–16.0 °C during the early Holocene. Two centennial-scale decreases in July air temperature were reconstructed during the Lateglacial interstadial which are correlated with Greenland Interstadial events (GI)-1d and -1b. The results suggest that vegetation changes in the Netherlands may have been promoted by the cooler climate during GI-1d, immediately preceding the Older Dryas biozone, and GI-1b. The Hijkermeer chironomid-inferred temperature record shows a similar temperature development as the Greenland ice core oxygen isotope records for most of the Lateglacial and a good agreement with other temperature reconstructions available from the Netherlands. This suggests that chironomid-based temperature reconstruction can be successfully implemented in the Northwest European lowlands and that chironomids may provide a useful alternative to oxygen isotopes for correlating European lake sediment records during the Lateglacial.     

Basaltic subglacial sheet-like sequences: Evidence for two types with different implications for the inferred thickness of associated ice

Subglacially-erupted volcanic sequences provide proxies for a unique range of palaeo-ice parameters and they are potentially highly useful archives of palaeoenvironmental information, particularly for pre-Quaternary periods. They can thus be incorporated by climate and ice sheet modellers in the same way as other environmental proxies, yet they remain largely under-utilised. Basaltic volcanic sequences erupted subglacially consist empirically of two major types, corresponding to eruptions under “thick” and “thin” ice, respectively. The latter are called subglacial sheet-like sequences and only one generic type of sequence has been described so far. However, there is now evidence that there are at least two generic types, with significantly different implications for interpretations of associated palaeo-ice sheet thicknesses. One type, which is relatively well described, is believed to be a diagnostic product of eruptions associated with a relatively thin glacial cover (< c. 150–200 m), probably corresponding most commonly to mountain glaciers but also conceivably thin ice caps or sheets, of any thermal regime (temperate, sub-polar, polar). It is here called the Mount Pinafore type. By contrast, a second subglacial sheet-like sequence, described in this paper for the first time and called the Dalsheidi-type, represents products of eruptions under much thicker ice (probably > 1000 m). Eruptions that form the Dalsheidi-type of sequence commence with the injection and inflation of a sill along the ice:bedrock interface. Such “interface sills” were predicted theoretically but had no known geological example, until now. Subsequent evolution commonly involves floating of the ice cover, catastrophic meltwater drainage and emplacement of widespread sheets of hyaloclastite, as cohesionless mass flows and hyperconcentrated flows. The water-saturated hyaloclastite is characteristically intruded by apophyses sourced in the underlying “interface sill”. Eruptions are commonly not explosive until their later stages. Dalsheidi-type deposits are outflow sequences probably linked to subglacial pillow volcanoes, which in Iceland were erupted along fissures. They only provide an indication of minimum thicknesses of the associated overlying ice, although theoretical considerations suggest substantial ice thicknesses in excess of 1000 m. However, they are likely to be characteristic products of eruptions under the thick West Antarctic Ice Sheet, but are currently inaccessible. Such eruptions may be capable of destabilising that ice sheet.     

Modeling ice sheets from the bottom up

Three facts should guide ice-sheet modeling. (1) Ice height above the bed is controlled by the strength of ice-bed coupling, reducing ice thickness by some 90 percent when coupling vanishes. (2) Ice-bed coupling vanishes along ice streams that end as floating ice shelves and drain up to 90 percent of an ice sheet. (3) Because of (1) and (2), ice sheets can rapidly collapse and disintegrate, thereby removing ice sheets from Earth's climate system and forcing abrupt climate change. The first model of ice-sheet dynamics was developed in Australia and applied to the present Antarctic Ice Sheet in 1970. It treated slow sheet flow, which prevails over some 90 percent of the ice sheet, but is the least dynamic component. The model made top-down calculations of ice velocities and temperatures, based on known surface conditions and an assumed basal geothermal heat flux. In 1972, Joseph Fletcher proposed a six-step research strategy for studying dynamic systems. The first step was identifying the most dynamic components, which for Antarctica are fast ice streams that discharge up to 90 percent of the ice. Ice-sheet models developed at the University of Maine in the 1970s were based on the Fletcher strategy and focused on ice streams, including calving dynamics when ice streams end in water. These models calculated the elevation of ice sheets based in the strength of ice-bed coupling. This was a bottom-up approach that lowered ice elevations some 90 percent when ice-bed coupling vanished. Top-down modeling is able to simulate changes in the size and shape of ice sheets through a whole glaciation cycle, provided the mass balance is treated correctly. Bottom-up modeling is able to produce accurate changes in ice elevations based on changes in ice-bed coupling, provided the force balance is treated correctly. Truly holistic ice-sheet models should synthesize top-down and bottom-up approaches by combining the mass balance with the force balance in ways that merge abrupt changes in stream flow with slow changes in sheet flow. Then discharging 90 percent of the ice by ice streams mobilizes 90 percent of the area so ice sheets can self-destruct, and thereby terminate a glaciation cycle.     

Response of a marine ice sheet to changes at the grounding line

A numerical model was designed to study the stability of a marine ice sheet, and used to do some basic experiments. The ice-shelf/ice-sheet interaction enters through the flow law in which the longitudinal stress is also taken into account. Instead of applying the model to some (measured) profile and showing that this is unstable (as is common practice in other studies), an attempt is made to simulate a whole cycle of growth and retreat of a marine ice sheet, although none of the model sheets is particularly sensitive to changes in environmental conditions. The question as to what might happen to the West Antarctic Ice Sheet in the near future when a climatic warming can be expecied as a result of the CO₂ effect, seems to be open for discussion again. From the results presented in this paper one can infer that a collapse, caused by increased melting on the ice shelves, is not very likely.     

Record of a Mid-Pleistocene depositional anomaly in West Antarctic continental margin sediments: an indicator for ice-sheet collapse?

Modern global warming is likely to cause future melting of Earth's polar ice sheets that may result in dramatic sea-level rise. A possible collapse of the West Antarctic Ice Sheet (WAIS) alone, which is considered highly vulnerable as it is mainly based below sea level, may raise global sea level by up to 5–6 m. Despite the importance of the WAIS for changes in global sea level, its response to the glacial–interglacial cycles of the Quaternary is poorly constrained. Moreover, the geological evidence for the disintegration of the WAIS at some time within the last ca. 750 kyr, possibly during Marine Isotope Stage (MIS) 11 (424–374 ka), is ambiguous. Here we present physical properties, palaeomagnetic, geochemical and clay mineralogical data from a glaciomarine sedimentary sequence that was recovered from the West Antarctic continental margin in the Amundsen Sea and spans more than the last 1 Myr. Within the sedimentary sequence, proxies for biological productivity (such as biogenic opal and the barium/aluminum ratio) and the supply of lithogenic detritus from the West Antarctic hinterland (such as ice-rafted debris and clay minerals) exhibit cyclic fluctuations in accordance with the glacial–interglacial cycles of the Quaternary. A prominent depositional anomaly spans MIS 15–MIS 13 (621–478 ka). The proxies for biological productivity and lithogenic sediment supply indicate that this interval has the characteristics of a single, prolonged interglacial period. Even though no proxy suggests environmental conditions much different from today, we conclude that, if the WAIS collapsed during the last 800 kyr, then MIS 15–MIS 13 was the most likely time period. Apparently, the duration rather than the strength of interglacial conditions was the crucial factor for the WAIS drawdown. A comparison with various marine and terrestrial climate archives from around the world corroborates that unusual environmental conditions prevailed throughout MIS 15–MIS 13. Some of these anomalies are observed in the pelagic Southern Ocean and the South Atlantic and might originate in major ice-sheet drawdown in Antarctica, but further research is required to test this hypothesis.     

Last glacial ice‐rafted debris off southwestern Europe: the role of the British–Irish Ice Sheet

Ice‐rafted debris (IRD) seeded into the ocean from Northern Hemisphere ice sheets is found in ocean cores along the southwestern European margin through the last glacial period. It is known that the origin of this IRD, especially off Iberia, can vary between North America and western Europe during short‐lived episodes of greatly enhanced iceberg flux, known as Heinrich events, although in most Heinrich events the IRD has a North American source. During the longer times of much lower IRD fluxes between Heinrich events, use of an intermediate complexity climate model, coupled to an iceberg dynamic and thermodynamic model, shows that background levels of IRD most likely originate from western Europe, particularly the British–Irish Ice Sheet. Combining modelling with palaeoceanographic evidence supports reconstructions of a short‐lived, but substantial, Celtic and Irish Sea Ice Stream around 23 ka. Copyright © 2010 John Wiley & Sons, Ltd.