A physically plausible theory of the 100 kyr climate cycle is proposed. Free oscillations between the mean ocean temperature and the marine ice-margin colatitude are shown to exist without requiring orbital forcing. It is shown that the curvature of the Earth causes two effects: (1) as the marine ice margin grows towards the equator, the net emmision of radiation (solar and terrestrial) per unit surface area increases; and (2) as the poleward extent of the ocean decreases, the net absorption of radiation per unit surface area increases. These radiation balance considerations, included with a realistic meridional transport of energy from the ocean to the marine-ice region and an atmospheric feedback process enhancing the ocean warming, are combined to form two nonlinear differential equations coupling the mean ocean temperature with the marine-ice margin colatitude. Using physically realistic parameters we are able to reproduce the major features of the 100 kyr climate cycle. This can be seen from Figure I which shows the δ18O record as given by Imbrie and others (1984), plotted against the model output. In addition we have found that the parameters used to obtain the general features of the ice-volume record also predict temperature “spikes” (1 to 2 K. above average) of relatively short duration (5 to 10 kyr) in the mean ocean temperature. We find that there is good qualitative agreement between the model's predicted mean ocean temperature and the estimation of summer sea-surface temperature at RC11-120 presented by Martinson and others (1987).

We have developed a physically-based parameterization for snow albedo, for the visible and near-infrared spectral regions used in general circulation models (GCMs). Snow albedo depends primarily on snow grain size, and also on solar zenith angle, snow thickness, impurity content, and atmospheric transmittance. This parameterization is now available as a Fortran subroutine. Simpler, but less accurate, parameterizations have also been developed which depend only on grain size or thickness. Since GCMs do not compute snow grain size, we also developed a method to estimate grain size based on the air temperature and the snow age.

Our parameterization for snow albedo is being incorporated in the NCAR Community Climate Model (CCM) in place of the existing empirical parameterization for snow albedo, to determine the effect of this improvement on the model's performance, and the results will be discussed. However, additional aspects of the treatment of the radiative properties of snow and ice were also capable of improvement and are being changed in the CCM. In particular, it is important to recognize that sea ice is often snow-covered and in that case has an albedo as high as that of snow, and that southern hemisphere sea ice is nearly always snow-covered, even through the melting season. The surface albedo for the Antarctic ice sheet should be about 0.83, but it had been set to 0.71 in the CCM, The CCM has been calculating temperatures too warm over Antarctica, and this low albedo contributed to that error.

Tidewater glaciers and large ice sheets, e.g. the Antarctic ice sheet and a late-Würm Arctic ice sheet, are complex but single dynamic systems composed of terrestrial, marine and floating parts. Morphology and dynamics of the different parts are different. The terrestrial parts are convex and their dynamics are controlled by shear stress only (the longitudinal stress is zero); the floating parts are concave and their dynamics are controlled by longitudinal stress only (the shear stress is zero). To connect the different parts we should consider transitional zones where shear and longitudinal stresses are comparable.

To describe glacier and ice-sheet dynamics, longwave approximation of the first order is used. In this approximation it is impossible to connect terrestrial and floating parts dynamically, only morphologically and kinematically. It means that the first-order longwave approximation is not sufficient.

If the transitional zone between the terrestrial and floating parts is long in comparison to ice thickness (in hydrodynamics the term “weak” is used) we can do the next step in the longwave approximation to describe the single dynamical system consisting of the terrestrial and floating parts and the weak transitional zones (ice streams). It is a purely hydrodynamical approach to the problem without ad hoc hypothesis.

The presented model is a non-stationary three-dimensional hydrodynamic model of glaciers and ice sheets interacted with ocean, involving the conditions of ice continuity and dynamic equilibrium, ice rheology, and boundary conditions on the free surface (dynamic and kinematic) and on the bed (ice freezing or sliding). Longwave approximation is used to reduce the three-dimensional model to a two-dimensional one. The latter consists of (1) evolution equations for grounded and floating parts and weak transitional zones; (2) boundary conditions on the fronts (e.g. the conditions of calving); (3) equations governing the junctions of the parts (the most important junction is the grounded line) with the conditions connecting the evolution equations.

Effects of Eurasian snow cover were first noted by Blanford (1884) who found an inverse relationship between summer monsoon rainfall over India and winter snow cover over the Himalayas. Hahn and Shukla (1976) confirmed it by using satellite-derived data and their work stimulated succeeding studies on the interaction between large-scale snow cover and atmosphere. Matson and Wiesnet (1981) showed that interannual variation of northern hemisphere snow cover is dominated by Eurasian snow cover, both showing similar trends and fluctuations during 1967–79. Recent studies (Barnett, 1988) also noted that Eurasian snow cover has a greater feedback potential than that of North America on hemispheric-scale climatic anomalies.

Though the importance has been thus recognized, not many studies have been done on the interaction between Eurasian snow cover and large-scale atmospheric circulation anomalies. Morinaga and Yasunari (1987) studied lag correlations between satellite-derived snow-cover extent over central Asia and the 500 mb-geopotential height field in the Northern Hemisphere (1967–84), and indicated that so-called Eurasian pattern (Wallace and Gutzler, 1981) in December brings large snow-cover extent in February; in turn, February snow cover has a considerable lingering effect on the atmosphere in April.

This study present further results on the time-lag teleconnections of the atmosphere associated with Eurasian snow-cover anomalies and their physical implications including the evaluation of snow-hydrological process.

Parameterization of the boundary conditions between the atmosphere and cryosphere is an important part of the general problem of modelling climatic change. It is necessary to define the mass, momentum and energy exchanges at the ice/atmosphere interface in order (i) to use atmospheric global circulation models (AGCMs) to predict future climate and (ii) to use snow, glacier or ice-sheet models to predict the corresponding response of the cryosphere. The physics of the boundary processes are fairly well known; the difficulty lies in choosing the appropriate space and time scales for modelling and in understanding the changes in the effective values of the model parameters which may be produced by spatial and temporal averaging.

Sensible heat, water vapour and momentum are tranferred vertically in the boundary layer of the atmosphere by turbulent motion. Equations for these fluxes contain parameters, the so-called scaling lengths zH, ZE and z0. Net radiation input to snow or ice is controlled by the albedo of the surface, α These four parameters play a major role in defining the boundary conditions between the atmosphere and cryosphere. It is normally assumed that their values are constants, determined by the characteristics of the snow or ice surface alone. For example, climate models may set zH = Ze = z0 = 0.1 mm and α = 0.9 for smooth, fresh snow. However, in modelling practice it is often found that the effective values of the parameters, i.e. those values that give the best simulations, are also influenced by the level of variability in the meteorological conditions.

The authors have made intensive micro-meteorological studies in the firn area of the Hintereisferner, Ötztal Alpen (Austria), on a frozen lake near Finse, Hardangervidda (Norway), and in the south-west coastal region of Greenland. Data from these field sites will be used to investigate the sensitivity of effective values of the boundary condition parameters to the choice of time scale using the Institute of Hydrology Distributed Model (IHDM).

With the intention of contributing to a better understanding of snow depth profiles used in reconstructing the southern hemisphere nitrate background we have measured C1−, , and in firn cores from two coastal Antarctic locations (GvN = Ekströmisen 70°S, 8°W, and Filchner = Filchner-Ronne Ice Shelf 79°S, 57°W). The depth resolution chosen is 2 cm per sample (i.e. 36 and 14 samples per year, respectively).

The isotopie composition of the firn cores was concurrently measured, with equally high resolution (deuterium and 18O data, W. Graf, private communication). The GvN core yields an average accumulation rate of 35 cm H2O per year during the period 1979–86, while the Filchner core gives 14 cm H2O per year during the period 1955-80. The net snow accumulation being relatively high allows precise determination of the year to year boundaries, as well as the relative contribution of individual seasons to the total net accumulation. This was achieved by combining the stable isotope data with the chemical tracers nss-sulfate (high concentration in summer) and sea salt (high C1− in autumn and winter). For the two individual locations this procedure allowed assessment of the glacial nitrate concentration seasonality as well as comparing the yearly nitrate deposition.

The Filchner location shows a distinct seasonality with maximum concentrations in summer. For nearly half of the years covered we also find higher concentrations in winter. This higher nitrate in winter is always accompanied by high sea salt concentrations. We suggest therefore a mechanism making sea salt aerosol an additional deposition pathway for nitrate in winter. This means that we are not able to link the enhanced winter nitrate deposition to stratospheric denitrification directly. The GvN core does not show significant seasonality. This is presumably due to frequent snow drift events preventing signal conservation.

Although the individual accumulation rates differ by a factor of 2.5 the yearly nitrate deposition is in the same range (4-11 kg (km-2 a-1) at both sites. This is caused by the different seasonal modulation of the net snow accumulation. At GvN low nitrate autumn precipitation prevails and keeps the yearly nitrate level constantly low (18-22 ppb). At Filchner, however, a high contribution of summer snow brings the yearly concentration average up to about 30–70 pbb.

In January 1987, an 7.8 m core, with an age at the bottom of 26 years, was collected from a site approximately 150 km inland from Halley Station (77°02.2′S, 22°32′W; altitude 1862 m a.s.l.); 10 m temperature ≈ −30°C; accumulation rate ≈ 14 g cm−2 a−1). The site lies some 140 km from the coast of the Weddell Sea and within the area bounded in winter by the polar vortex.

The core has been analysed at a frequency of ≈28 samples per accumulation year for sulphate, nitrate and chloride, and has been dated stratigraphically from the clear seasonal cycles in non sea salt sulphate (Mulvaney and Peel, 1988). With this resolution it is possible to examine the seasonal pattern of deposition of chemical species and their phase relationships.

Of particular interest is the possibility that ice cores may preserve evidence for disturbances in tropospheric chemistry, associated with the recent spring-time depletion of stratospheric ozone. It has been proposed that this is accompanied by a denitrification of the stratosphere during the winter months, implying enhanced levels of NOX in the late winter/spring troposphere and in precipitation. Our data reveal a strong seasonal signal in nitrate deposition, apparently peaking in spring. Similar behaviour has been reported by Wagenbach and others (1988) for nitrate in the atmospheric aerosol, in a 3-year sequence (1983–86) from Georg von Neumayer Station (70°S, 8°W). There does not appear to be any evidence in our data of an increase in spring-time nitrate deposition since the appearance of the Antarctic ozone hole in 1978.

Reliable reconstructions of paleoglaciers using topographic maps and aerial photographs allow calculation of effective basal shear stresses along the longitudinal profiles of these glaciers. Glacial flow theory applied to these shear stresses provides an estimate of the component of mass flux due to internal deformation. Assuming basal slip to be zero at the point where deformation mass flux is a maximum, minimum average accumulation gradients (above the equilibrium-line altitude (ELA)) and ablation gradients (below the ELA) can be calculated and minimum mass flux at the ELA can be estimated using the continuity equation. Average net winter accumulation can also be calculated by dividing the mass flux at the ELA by the accumulation area. Because local climate controls the mass balance of a glacier, and therefore the accumulation and ablation gradients, this model provides information on the climatic setting of these paleoglaciers.

This model also allows estimation of basal slip as a factor in point estimates of glacial flow. Application of the continuity model above and below the ELA generates additional estimates of mass flux at discrete points along the glacier. The difference between deformation mass flux and continuity flux at these points yields a first approximation of basal slip, which is highly variable along the glacier.

The model was tested on the Big Timber glacier of west-central Montana and applied to several other late Pleistocene glaciers in the northern Rocky Mountains of south-western Montana and north-eastern Idaho. Low ablation gradients (<4.0 mm m-1) suggest a climate during the late Pleistocene comparable to the present-day climate of the Brooks Range in Alaska. Calculated average net winter accumulation for the area is well below modern values, again indicating that the climate was much drier during the full glacial period. Basal sliding accounts for most (>90%) of the glacial flow near the terminus of each glacier but is variable along the rest of the glacier. While the mass-balance values are minima, they are assumed to be reasonable approximations of the actual values unless very high basal slip rates occurred along the entire length of each glacier.

A continuous, detailed, 200-years record of the anionic species, chloride, nitrate and sulphate, has been measured on an ice core from Dolleman Island (70°35.2′ S, 60°55.5′ W), Antarctic Peninsula. The site lies on the east coast of the Peninsula, and the chemistry of the core is dominated by the changing pattern of sea-ice distribution and storm activity in the Wed dell Sea. Strong annual cycles in chloride and non sea salt sulphate reflect the dominance of the seasonal cycle in sea-ice distribution in the Weddell Sea, observed in time series derived from satellite imagery since the early 1970s. However, in the case of chloride there is also an exceptionally strong interannual variability, which in many parts of the core dominates the seasonal cycle.

Secular variations in the sea-ice extent appear to have a strong influence on the climate of the region and may play a major role in determining how long-term climate change in the Antarctic Peninsula relates to global climate change. The paper examines documented evidence for sea-ice extent in the Weddell Sea sector, and evaluates the usefulness of ice-core data for reconstructing this parameter in the earlier period.

The Patagonian glaciers located in the southern part of the Andes between 46°30′S and 51°30′S are characterized by typical temperate conditions of heavy precipitation, rapid ice flows and high melting rates. During the austral summers of 1983–84 and 1985–86, field studies were made of the ice flow, heat balance and morphology of several glaciers in Patagonia. Coupled with aerial photographic surveys, these revealed that most glaciers had retreated extensively in the recent years, a maximum being 200 m a-1 at San Rafael Glacier from 1974 to 1986. The lower part of Soler Glacier had thinned by a rate of 5.2 m a-1 from 1983 to 1985.

This paper presents three possible mechanisms to explain the large variation of temperate glaciers during the last decade, based on analyses of mass balance and dynamics of Patagonian glaciers:

(1) The annual melting rate was estimated at about 10–15 m a-1 in water equivalent over the ablation area (from 350 to 1350 m a.s.l.) of Soler Glacier. Monthly mean air temperature in the coldest season (June through August) was estimated at about 0°–4°C near the termini of most glaciers in Patagonia. That temperature coincides with an air temperature which is critical for solid or liquid precipitation. The difference in the surface albedo, that is, 0.7–0.8 for new snow and 0.4–0.55 for bare ice (0.1–0.2 for debris-covered ice), results in different melting rates. Hence, a slight change in air temperature should cause an enhanced change in ice thickness by a positive feedback mechanism.

(2) The flow velocity was measured or estimated and was found to change daily and seasonally by factors of 3 to 5 at Soler Glacier. The large flow velocity variation was attributed to difference in the basal sliding velocity. Consequently, a change in the amount of subglacial water or the structure of the basal water system should cause a large change in the ice flow, which in turn results in a retreat or an advance of the glacier-like “mini-surge”.

(3) Frequent fluctuations of calving glaciers (e.g. San Rafael and Pio XI glaciers) have been much reported; however, information on the position of the grounding lines is very scarce. The advance or retreat of the glacier front may possibly have been affected by that of the floating terminus. The rate of calving from the ice tongue or spreading of ice shelves should mainly be controlled by the melting rate of ice in the water and by the mechanical properties of ice, and these factors are not directly related to climatic change or the surge phenomenon.

Solar radiation is one of the most important meteorological components in the heat balance of the glaciers and other snow and ice masses. Observed solar radiation values are generally discussed in relation to altitude, landform, geographical position of the observation point and cloud. However, as surface of snow and ice masses possess high albedo, the surface global solar radiation on snow and ice masses usually shows higher value than surrounding ground which is low albedo, due to multiple reflection between surface and cloud. This is an important factor in determination of solar radiation on snow and ice masses. When this process is prominent, surface global solar radiation will depend on the size of the snow and ice masses. This process will be investigated in two parts. One is the presentation of radiation data taken on an ice cap and surrounding ground in the northwestern Tibetan plateau. Observations show that mean daily global solar radiation was 26% higher on the ice cap than on the ground: especially on cloudy days this value exceeded 50%. The second is collaboration of radiation data on snow and ice masses of previous studies: those data will be discussed from the above point of view. The size effect of snow and ice masses on surface solar radiation will be stressed.

Quite high amounts of evaporation observed on glaciers in dry Central Asia are one important component in the mass balance of glaciers. During glaciological research at West Kunlun Mountains on the northwestern Tibetan plateau in 1987, high evaporation of more than 2 mm of water per day was observed near the equilibrium line (5800–5900 m a.s.l.) in mid-summer due to dry and windy climate. Annual evaporation is estimated by a semi-empirical method using the full year meteorological data taken by an automatic meteorological station. This value shows that evaporation occupies quite a high percentage within the mass balance near the equilibrium line and on the whole glacier. Discussions will centre on comparison with evaporation data obtained on other glaciers and on the effect of such evaporation in the climate-glacier relation.

Antarctic iceberg observations provide two types of climatic information: (1) the rate of iceberg calving gives the main negative term in the mass balance of Antarctica; (2) the distribution of icebergs in the Southern Ocean is related to various factors including sea-ice extent and ocean conditions. This paper discusses climatic information obtained both from modern iceberg observations, and from historical data.

The main modern data source is the international iceberg observation programme, initiated in 1981. This has generated a database which now contains observations of size and position of 150 000 icebergs in the Southern Ocean. Other recent data sources include observations from 1974 of icebergs of over 22 km length by the Navy/NOAA Joint Ice Center (which now total nearly 100 bergs). Historical sources include both scientific and commercial (whaling) expeditions.

Three conclusions can be drawn from the recent iceberg data.

• (1) For the past seven years annual calving rates of icebergs less than 22 km in length have been approximately constant both in numbers and total mass, and have exceeded the annual mean mass of bergs over 22 km calved during the same period.

• (2) The annually-calved mass of gigantic icebergs of length over 22 km, has varied more than two orders of magnitude during the past two decades. The mass of calved gigantic icebergs alone exceeded continental snow accumulation in 1967, 1986 and 1987.

• (3) Mean iceberg calving rate exceeds continental snow accumulation rate.

These results are discussed in relation to Antarctic mass balance and sea level.

Historical observations of iceberg distributions in the Southern Ocean differ from recent observations by showing: (i) higher frequencies of bergs at lower latitudes, (ii) a larger proportion of large bergs, and (iii) a more uniform iceberg distribution throughout the region. Comparisons between past and present distribution patterns indicate that there must be errors, including exaggerations and selective reporting, in the historical data. Such observations should therefore not be used uncritically to make climatic conclusions.

Bearing in mind the potential flaws in historical data there still appears to be real variations with time in iceberg distribution. These include higher frequencies of bergs at lower latitudes during the past century and the first decades of this century, than at present. Expanded northern range of the bergs would occur if one or more of the following conditions then applied: (1) calved icebergs were larger; (2) sea ice was more extensive in time and space; (3) ocean temperatures were lower; (4) ocean (wind) conditions were such that northward advection rates were higher. The importance of each of these conditions is discussed.

A stable isotope record extending back to 1795 is now available from Dolleman Island (70°35.2′S, 60°55.5′W), a small ice rise on the Weddell Sea coast of Antarctic Peninsula. An accurate chronology has been achieved by combined stratigraphic analysis of clear seasonal cycles in δ18O and excess SO4. Previous work (Peel and others, 1988) has shown that, since 1947, there is generally a satisfactory correlation between interannual variations in δ18O and air temperature (T) as recorded at weather stations in various parts of the region, suggesting that the derived δ18O/T ratio may be used to reconstruct air temperatures for the earlier period.

Taken together with previously-reported data (Aristarain and others, 1986) for an ice core from James Ross Island it is now possible to propose a regional climatic signal for the Weddell Sea coastal sector of the region. The most striking feature is a broad maximum in δ18O for the mid-19th century, implying decadal average temperature at least as high as the present. This contrasts with available evidence from elsewhere in the southern hemisphere which suggest that this period was cooler than today. Tentative explanations for the anomaly are proposed based on evidence for a period (1974–80), where climatic shifts are clearly amplified in the isotopic records.

Since 1977 ablation-season temperature has been 1.1°C above the 1930–80 mean and winter precipitation has been 14% below the 1930–80 mean. In order to identify the effect of this climatic fluctuation on North Cascade glaciers, the North Cascade Glacier-Climate Project has monitored the terminus behavior of 107 glaciers between 1983 and 1988.

The 107 glaciers examined represent six climate sensitivity groups. Each group has a different sensitivity to the four primary climatic parameters: (1) ablation-season temperature, (2) accumulation-season precipitation, (3) summer cloud cover and (4) freezing levels during May and October precipitation events. A glacier’s sensitivity to each climatic parameter is determined by its geographic location and topographic position. Each sensitivity type has specific geographic and topographic characteristics, such as degree of radiational shading, orientation, altitude with respect to the local glaciation threshold, accumulation sources, and distance from the Cascade Crest. Accumulation sources are direct snowfall, wind drifting and avalanching.

Of the 107 glaciers examined, 91 had retreated significantly between 1983 and 1988 and three had advanced. Correlation of retreat rate and climatic sensitivity type indicates that the higher a glacier’s winter balance, the smaller the retreat rate. High-altitude accumulation zones, multiple accumulation sources and a northward orientation are all associated with higher winter balances. Retreat was greatest for glaciers with poor radiational shading, and only direct snowfall accumulation. Retreat rate was slowest for glaciers with multiple accumulation sources and a northward orientation, though not necessarily good radiational shading. Retreat rate increased with distance east of the Cascade Crest and retreat rate was high for low-altitude glaciers.

The climate of the Holocene is, for continental regions from middle and low latitudes, relatively well documented from pollen studies and other sources. To obtain a global picture, these data must be supplemented by climatic series from polar regions. Such information may be extracted from δD or δ18O ice-core profiles but the interpretation of these isotopic records suffers some limitations, (1) because, expected temperature changes being small, they can be obscured by noise effects in the isotope-temperature relationship, and (2) because they can be influenced, especially in coastal regions, by changes in origin of the ice.

With this in mind, we focus our presentation on Dome C and Vostok cores drilled on the East Antarctica Plateau and essentially undisturbed by ice-flow conditions. The detailed comparison between these continuous isotopic records makes it possible to know which part of the isotopic signal is climatically significant. Spectral properties of these two records are also examined over the Holocene period. In addition, we present isotopic results obtained on a 950 m ice core drilled at Komsomolskaia (also on the East Antarctica Plateau) by the Soviet Antarctic Expedition. This core fully covers the Holocene and, although discontinuous, the new data help us to document the East Antarctica isotopic record.

From these data, an average climatic record is constructed which shows that the East Antarctica climate was fairly stable during the Holocene, marginally warmest around 10 kyear B.P. and coldest in periods around 1.5 and 6 kyear B P. These features are discussed in relation with other Antarctic data (Byrd, Law Dome, Ross Ice Shelf) and with climate records from both southern and northern hemispheres

Time scales for adjustment in the length of a glacier to changing climate may be described in terms of a relatively short time scale, TS and a longer memory time, Tm. The memory Tm represents the time scale needed for asymptotic approach of the glacier to a steady state following a climate change event. Tm is determined by simple continuity considerations concerning the total volume change that must occur to reach steady state and the balance rate that drives the change. We show that Tm is relatively independent of the size of the climate change or to details of how ice flow is related to the geometry of the glacier. The time scale TS represents the time between a climate event and the occurrence of substantial changes in the glacier length. We show that, in contrast to Tm, Ts is highly dependent on the size of the climate change and on details of ice dynamics. This dependence is investigated by several ice-flow models including a simple one in which ice transport is determined by local thickness and slope, as in the analysis of kinematic waves, and a finite element representation that fully includes longitudinal stress gradients. The ice-flow models are subjected to mass-balance perturbations of varying size — from small, for which linearization approximations are valid, to large, for which linearization breaks down.

The following behavior may be identified. Increasing the size of a mass-balance rate change causes a more rapid initial response of a glacier terminus, which tends to shorten Ts. Longitudinal stress gradients damp local variations in velocity and thereby slow the propagation and diffusion of kinematic waves and retard the response of the terminus, which tends to lengthen Ts. Longitudinal stress gradients transmit forces to the terminus region and influence the terminus motion without the necessity of redistributing mass from the glacier length into the terminus zone, which tends to shorten Ts. These various results indicate that accurate modeling of the short term responses of glaciers to climate change requires fairly sophisticated ice-flow models, However, for purposes of tracking glacier lengths (or areas) over time scales considerably great than Ts, fairly simple ice-flow models may suffice.

Observations of high particulate loads on Eurasian Basin sea ice in 1987 raise questions of consequence for sediment budgets, ice melting, ice modeling and remote sensing. Biogenic and lithogenic particles were observed in concentrations high enough to color the ice surface brown over large area (greater than 15 × 15 km2) within the Siberian branch of the Transpolar Drift stream. The sediment is most likely incorporated when ice forms on the Siberian shelf seas, and is concentrated at the ice surface after several years of summer surface melting and biological growth within the Arctic basin. Much of the particle-laden multi-year ice appears to leave the Arctic basin via Fram Strait, depositing its sediment load along the axis of the East Greenland Current.

To date, variation in sea-ice particle load has not been taken into consideration when modeling ice thickness or distribution for past or future environmental scenarios, with the exception of soot deposited from nuclear war. Naturally elevated surface-particle concentration may occur if there is increased deposition from long-range or coastal transport of aeolian material, increased sediment input into sea ice which is then exposed to surface melting, and/or increased biogenic productivity on the ice surface. Such conditions may have prevailed during the Younger Dryas. If particle loads become high enough to cause extensive sea-ice melting, changes may be expected in sea-ice concentration and distribution, sea-floor sedimentation rates, and oceanic productivity.

About 1500 surface-ice samples for δ18O analysis were collected in the 1988 field season along a 750 m profile perpendicular to the margin of the Greenland ice sheet at Pakitsoq, ca 40 km north-east of Jakobshavn, central West Greenland. The purpose of the study was to evaluate how well the continuity of the layer sequence is preserved in ice-margin records, a question of crucial importance for evaluating the potential yield of using ice margins as “mining areas” for easily accessible old ice for climate and environmental studies.

More than half of the 1500 samples were taken continuously as 20 cm samples along a 170 m section through the Wisconsinan-Holocene transition which, previously, had been located at the surface of the ice margin. Along this transition section δ18O values decrease by about 6‰ from −31.5 to 37.5‰ on an average. Detailed studies were made of surface elevations and surface structures (e.g. blue bands) along the “horizontal core” profile which, moreover, was photographed section by section, thus enabling the δ-record to be correlated with surface features.

Results of the δ18O analyses are promising. Even though ice from the blue bands has δ-values that are 7–8‰ higher than those of the surrounding white ice, there seems to be no discontinuity in the white-ice δ-record across the blue bands.

The GISS climate model is used to investigate whether the growth of ice sheets could have been initiated by solar insolation variations. Three different orbital configurations are used, corresponding to 116 000 yr B.P., 106 000 yr B.P., and a modified insolation field with greater reductions in summer insolation at high northern latitudes. The time periods chosen are those in which geophysical evidence implies that ice sheets were growing rapidly.

The results show that the model fails to maintain snow cover through the summer at locations of suspected initiation of the major ice sheets, despite the reduced summer insolation. When 10 m-thick ice was inserted in all locations where continental ice sheets existed during the last glacial maximum, the model failed to maintain it as well: the ice would melt away in less than five years. Only cooling the ocean to its full ice-age value allowed the model to keep any of the additional ice, and then only in a very restricted region in northern Baffin Island.

The experiments indicate there is a wide discrepancy between the model’s response to Milankovitch variations and the geophysical evidence of ice sheet initiation. As the model failed to grow or sustain low-elevation ice during the time of high latitude maximum solar radiation reduction (120 000–100 000 yr B.P.), it is unlikely it could have done so at any time within the last several hundred thousand years. Either the model is not nearly sensitive enough, or we do not understand the connection between Milankovitch variations and the growth of ice sheets.