The annual mean oxygen-isotope content of Antarctic snow is strongly influenced by the seasonal variability of accumulation. Since the annual mean δ18O value is frequently used to derive mean annual temperatures from ice cores, changes in atmospheric circulation pattern can lead to large errors in the deduced temperature record. At the German Antarctic wintering base, Neumayer, accumulation measurements have been carried out continuously over the last 16 years. Weekly readings of accumulation stakes combined with snow pits and shallow firn cores are used to investigate the influence of the seasonal variability of accumulation on the annual mean δ18O values and to estimate the possible error in the determination of annual mean temperatures from ice cores by using the oxygen-isotope record.

Snow-pit analysis in the tow-accumulation region in northeast Greenland indicated a mean annual accumulation of 12 ± 1.6 cm. It also revealed numerous faceted layers that were not related to the summer horizon of previous years. In general, the faceted crystals were found under a wind crust. These crusts appear to originate during the winter. Snow- and air-temperature data from automatic weather stations suggest that the katabatic storms are responsible for the formation of the wind crusts and the faceted layers. Frequent, abrupt temperature changes up to 30°C were observed within two days, due to turbulent mixing of cold-air inversions. The near-surface air-temperature variations are correlated with wind speed, the latter leading the former by 6–9 hours for increasing wind, and 1–4 hours for decreasing wind. Because of low accumulation, katabatic storms affect the same snow surface for several months, causing repealed air-temperature fluctuations. This leads to strong vapor-pressure gradients and faceted crystal growth in the top snow layer. On average we found two faceted layers (in addition to the summer layer) per year for a 4 year record. In addition to its climatological significance in the snow record, the formation of these layers is especially important to understand for the interpretation of microwave satellite observations of the firn.

This paper describes the current state of a complete automatic system of three numerical models that simulate snow-cover stratigraphy and avalauche risks for operational avalanche forecasting. The first model, SAFRAN, estimates relevant meteorological parameters affecting snowpack evolution. The second Crocus, is a snow numerical model which simulates the physical processes inside the snowpack and its stratigraphy. The last model, MÉPRA, is an expert system; based on an assessment of snowpack stability, it deduces natural and accidental avalanche risks. To describe the great Variability of the snowpack and the associated avalanche risks, this automatic system simulates snow-cover evolution and its stability for many typical slopes, elevations and aspects representative of the different French massifs. To achieve this result, different kinds of validations have been carried out since winter 1981; they are mainly based on comparisons with different sets of measurements and on the opinion of users.

Although the routinely obtained results do not yet take into account all small-scale effects such as wind transport, they have been considered as valuable information by avalanche forecasters since 1992 93 and used operationally since then.

The total air content (V) of ice has been measured along the Dome Summit South (DSS) core from Law Dome, East Antarctica. The main features of this record are the very well-preserved sub-annual fluctuations of V (down to at least 900 m depth) and the significant increase of the V values during the last deglaciation. The sub-annual variations reflect changes in close-off porosity, and we interpret the V seasonal peaks as tracers of depth-hoar layers. For the longer time-scale, the large V fluctuations observed are interpreted in terms of elevation and/or close-off porosity changes under different assumptions. We analyze the possible influence of a different global pressure field and/or a change in seasonal temperature and precipitation cycles during the last glacial period. Our estimates of surface elevation changes derived from the V data are then compared with independent reconstructions of past elevations.

As polar ice descends from the free surface to depth in a large ice sheet, it undergoes deformations which give rise to the formation and subsequent evolution of a fabric and associated anisotropy. In this paper two orthotropic models of such strain-induced anisotropy are considered. Model A is based on analysis of the microscopic behaviour of an individual ice crystal with transversely isotropic response and assumed uniform stress in a polycrystal. The macroscopic response of the ice aggregate is then derived by applying the concept of an orientation distribution function, and the resulting viscous law relates the strain rate to the stress and three structure tensors. In model B it is assumed that the macroscopic response of ice is determined by the fabric induced entirely by macroscopic deformations, and all microprocesses taking place at the grain level are ignored. A constitutive relation is derived from a general frame-indifferent law for orthotropic materials, and expresses the stress in terms of the strain rate, strain and throe structure tensors. The two models are applied to determine the viscous response of ice to continued uniaxial compression and simple shearing in order to compare the predictions of both theories.

Observations on the spatial distribution and on the strain dependence of the crack density are given for cracks formed during compressive, unidirectional, constant-strain-rate deformation of columnar-grain ice. Specimens, in the grain-size range of about 2 9 mm, were strained at the nominal rates of 10−3, 10−4 and 10−5 s−1 at 10°C. The axis of hexagonal crystallographic symmetry of each specimen grain tended to be in the plane perpendicular to the long direction of the grains and to have a random orientation in that plane. For stress applied perpendicular to the long direction of the grains, the deformation was practically two-dimensional. It was found that the log-normal distribution function provided a good approximation to the strain dependence of the crack density. Statistical characteristics of the distribution had a maximum in the same range of strain rate as found for the strength of columnar-grain ice. Analysis of the spatial distribution of the cracks indicated some deviation from randomness for specimens of grain-size less than 5 mm and total strain less than 50 × 10−4. The observations provide further evidence that crack formation under the experimental conditions is a random process.

Modes of debris entrainment and subsequent transfer in seven “normal” and five surge-type glaciers in Svalbard (76–79° N) are outlined in the context of the structural evolution of a glacier as the ice deforms during flow. Three main modes of entrainment and transfer are inferred from structural and sedimentological observations: (i) The incorporation of angular rockfall material within the stratified sequence of snow/firn/superimposed ice. This debris takes an englacial path through the glacier, becoming folded. At the margins and at the boundaries of flow units the stratified ice including debris is strongly folded, so that near the snout the debris emerges at the surface on the hinges and limbs of the folds, producing medial moraines which merge towards the snout. The resulting lines of debris are transmitted to the proglacial area in the form of regular trains of angular debris. (ii) Incorporation of debris of both supraglacial and basal character within longitudinal foliation. This is particularly evident at the surface of the glacier at the margins or at flow unit boundaries. It can be sometimes demonstrated that foliation is a product of strong folding, since it usually has an axial planar relationship with folded stratification. Foliation-parallel debris thus represents a more advanced stage of deformation than in (i). Although the presence of basal debris is problematic, it is proposed that this material is tightly folded ice derived from the bed in the manner of disharmonie folding. The readily deformed subglacial sediment or bedrock surface represents the plane of décollement. (iii) Thrusting, whereby debris-rich basal ice (including regelation ice) and subglacial sediments are uplifted into an englacial position, sometimes emerging at the ice surface. This material is much more variable in character than that derived from rockfalls, and reflects the substrate lithologies; diamicton with striated clasts and sandy gravels are the most common facies represented. Thrusting is a dynamic process, and in polythermal glaciers is probably linked mainly to the transition from sliding to frozen bed conditions. It is not therefore a solely ice-marginal or proglacial process.

In the Jutulgryta area of Dronning Maud Land, Antarctica, subsurface melting of the ice sheet has been observed. The melting takes place during the summer months in blue-ice areas under conditions of below-freezing air and surface temperatures. Adjacent snow-covered regions, having the same meteorological and climatic conditions, experience little or no subsurface melting. To help explain and understand the observed melt-rate differences in the blue-ice and snow-covered areas, a physically based numerical model of the coupled atmosphere, radiation, snow and blue-ice system has been developed. The model comprises a heat-transfer equation which includes a spectrally dependent solar-radiation source term. The penetration of radiation into the snow and blue ice depends on the solar-radiation spectrum, the surface albedo and the snow and blue-ice grain-sizes and densities. In addition, the model uses a complete surface energy balance to define the surface boundary conditions. It is run over the full annual cycle, simulating temperature profiles and melting and freezing quantities throughout the summer and winter seasons. The model is driven and validated using field observations collected during the Norwegian Antarctic Research Expedition (NARE) 1996–97. The simulations suggest that the observed differences between subsurface snow and blue-ice melting can be explained largely by radiative and heat-transfer interactions resulting from differences in albedo, grain-size and density between the two mediums.

We employed a commercial wireline drill rig to investigate the subglacial conditions of Black Rapids Glacier, a well-studied surge-type glacier in the central Alaska Range. The four main goals were: to assess the capabilities of the commercial drilling industry for sampling subglacial material, to investigate the basal morphology, to determine the subglacial geology and to emplace borehole instruments. The drilling was done in an area where seasonal and secular variations in speed are large, and where seismic studies suggested the presence of a till layer. Four holes were drilled at three locations to a maximum depth of 620 m. Three holes yielded samples of basal ice and till, although recovery of the latter was generally poor. Bedrock was sampled in one or possibly two of the holes. In the area sampled, t he glacier is underlain by a till layer some 4–7 m thick, confirming the seismic interpretation. It consists of a sandy matrix at least 20–30% of which comprises larger clasts. Limited samples of the matrix indicate that near the top of the till the porosity is 40%, and t hat some of the pore water is frozen. Geologic studies suggest that the drilling area lies to the north of the Denali Fault, a major tectonic boundary followed by the glacier, and that most of the till is locally derived with transport distances of <2 km.

A detailed study of the major-ion chemistry of the Dokriani glacier meltwaters, feeding the Bhagirathi river, Ganga basin, Garhwal Himalaya, India, has been carried out to assess the role of active glaciers in the higher chemical-denudation rate (CDR) in this area. The CDR of the Dokriani glacier catchment is 321.6 t km-2 a-1, higher than in other glacierized catchments of the world, indicating intense chemical erosion in the glacierized catchments of the Himalaya. The dominance of Ca2+, HCO3- and SO42- in meltwaters throughout the 6 month (May–October) ablation period suggests that the chemical weathering is dominated by coupled reactions involving sulphide oxidation and carbonate dissolution. Good positive correlation between SO42- and suspended-sediment concentration during July and August (r2 = 0.72 and 0.67, respectively) suggests that the southwest monsoonal rainfall enhances the weathering of supraglacial moraine and contributes significant amounts of sulphate to the high meltwater discharges. The sulphate flux, as a proportion of combined (SO42- + HCO3-) flux, also increased from 45% at 2 m3 s-1 to 63% at 10 m3 s-1.

Of several problems associated with theoretical explanations of the jökulhlaups which emerge from the outlet glacier Skeiðarárjökull of the ice cap Vatnajökull in southeast Iceland, the mechanism of flood initiation is one that has hitherto defied explanation. We provide such an explanation based on a careful analysis of the classical Nye-Röthlisberger model; near the subglacial lake Grímsvötn, the hydraulic potential gradient is towards the lake, and there is therefore a catchment boundary under the ice, whose location depends on the subglacial meltwater drainage characteristics. As the conditions for a flood approach, we show that the water divide migrates towards the lake, while at the same time the lake pressure increases. When the hydraulic potential gradient towards the lake is low and the refilling rate is slow, the seal will "break" when the catchment boundary reaches the lake, while the lake level is still below flotation pressure, whereas if refilling is rapid, flotation can be achieved before a flood is initiated. This theory can thus explain why the seal is normally broken when the lake level at Grímsvötn is still some 60 m below flotation level. In addition, we are able to explain why the jökulhlaup following the 1996 eruption did not occur until flotation level was achieved, and we show how the cyclicity and magnitude of jökulhlaups can be explained within this theory.

High-resolution radar images of Glaciar Moreno, Argentina, acquired by the Shuttle Imaging Radar C (SIR-C) on 9 and 10 October 1994 at 24 cm wavelength (L-band), are utilized to map the glacier velocity both interferometrically and using the phase correlation method. The precision of the interferometric ice velocities is 1.8 cm d-1 (6 m a-1) (1σ). The phase correlation method measures ice velocity with a precision of 14 cm d-1 (50 m a-1) with image data at a 6 m sample spacing acquired 1 day apart. Averaged strain rates are measured with a precision of 10-4 d-1 at a 240 m sample spacing with the phase correlation method, and 10-5 d-l with radar interferometry. The phase correlation method is less precise than radar interferometry, but it performs better in areas of rapid flow, is more robust to temporal changes in glacier scattering and measures the glacier velocity in two dimensions with only one image pair. Using this technique, we find that Glaciar Moreno flows at 400 m a-1 in the terminal valley and 800 m a-1 at the calving front, in agreement with velocities recorded a decade ago. Assuming steady-state flow conditions, the vertical strain rates measured by SIR-C are combined with prior data on mass ablation to estimate the glacier thickness and ice discharge. The calculated discharge is 0.6 ± 0.2 km3 ice a-1 at 300 m elevation, and 1.1 ± 0.2 km3 ice a-1 at the equilibrium-line elevation (1150 m), which yields a balance accumulation of 6 ± 1 m ice a-1.

The hydraulic properties of the firn on Storglaciären, Sweden, were investigated in firn cores by water-table measurements and pumping tests. The mean density of the firn was 800 850 kg m3, giving an effective porosity of 0.073. The lower part of the firn layer was saturated with water, producing a maximum saturated layer of 5 m in late July. Hydraulic conductivity of the firn aquifer was determined from pumping tests to be 4.9 × 105 m s1. Percolation velocity, calculated from the time lag of maximal water input at the glacier surface and the water-level peaks, was 0.25 m h1. Percolation velocity increased over the ablation season, indicating a widening of the percolation pathways. A decrease in percolation velocity with percolation depth was found, reflecting decreasing permeability. The firn–water table responded to water input at the glacier surface with a delay of about 3 days. No diurnal variations were found in an area which was not influenced by fast drainage, indicating a diffusion of diurnal variations in meltwater production. One borehole intersected a water-filled cavity. Water level in this cavity showed diurnal variations, which probably were caused by diurnally produced meltwater waves moving fast through englacial conduits.

Hourly melt and discharge of Storglaciären, a small glacier in Sweden, were computed for two melt seasons, applying temperature-index methods to a 30 m resolution grid for the melt component. The classical degree-day method yielded a good simulation of the seasonal patient of discharge, but the pronounced melt-induced daily discharge cycles were not captured. Modelled degree-day factors calculated for every hour and each gridcell from melt obtained from a distributed energy-balance model varied substantially, both diurnally and spatially. A new distributed temperature-index model is suggested, attempting to capture both the pronounced diurnal melt cycles and the spatial variations in melt due to the effects of surrounding topography. This is accomplished by including a radiation index in terms of potential clear-sky direct solar radiation, and thus, without the need for other data besides air temperature. This approach improved considerably the simulation of diurnal discharge fluctuations and yielded a more realistic spatial distribution of melt rates. The incorporation of measured global radiation to account for the reduction in direct solar radiation due to cloudiness did not lead to additional improvement in model performance.

Temperature, density and accumulation data were obtained from shallow firn cores, drilled during an overland traverse through a previously unknown part of Dronning Maud Land, East Antarctica. The traverse area is characterised by high mountains that obstruct the ice flow, resulting in a sudden transition from the polar plateau to the coastal region. The spatial variations of potential temperature, near-surface firn density and accumulation suggest that katabatic winds are active in this region. Proxy wind data derived from firn-density profiles confirm that annual mean wind speed is strongly related to the magnitude of the surface slope. The high elevation of the ice sheet south of the mountains makes for a dry, cold climate, in which mass loss owing to sublimation is small and erosion of snow by the wind has a potentially large impact on the surface mass balance. A simple katabatic-wind model is used to explain the variations of accumulation along the traverse line in terms of divergence/convergence of the local transport of drifting snow. The resulting wind- and snowdrift patterns are closely connected to the topography of the ice sheet: ridges are especially sensitive to erosion, while ice streams and other depressions act as collectors of drifting snow.

Stable-oxygen-isotope ratios (δ18O) of precipitation and ice-core samples collected from the headwaters of the Urümqi river, Tien Shan, China, were used to test the relationship between δ18O and contemporaneous surface air temperature (Ta). A strong temporal relationship is found between δ18O in precipitation and Ta, particularly for the monthly averages which remove synoptic-scale influences such as changes in condensation level, condensation temperature and moisture sources ( Yao and others, 1996). Linear fits as high as 0.95‰° C-1 for precipitation events and 1.23‰° C-1 for monthly averages are found. Although the δ18O amplitude in ice cores drilled at the nearby Ürümqi glacier No.1(~2 km from the precipitation sampling site) decreased dramatically compared to the precipitation samples, the ice-core records of annually averaged δ18 are still positively correlated with contemporaneous air temperature, especially summer air temperature, at the nearby Daxigou meteorological station. Nevertheless, the relationship between the ice-core δ18O records and contemporaneous air temperature is less significant than that for the precipitation samples due to depositional and post-depositional modification processes, which are highlighted by the successive snow-pit δ18O profiles from the Ürümqi glacier No.1. Our results might extend the application of high-altitude and subtropical ice-core δ18O —Ta records for paleoclimate reconstruction.

The hydrological method is used to compute the mass balance of Siachen Glacier in the Nubra Valley, Karakoram Himalaya, India. The method was chosen because of its suitability in rugged terrain. Analysis of precipitation data from the glacier as a function of altitude shows a power-law relationship. Accumulation estimates based on this relationship show good agreement with the recorded precipitation. The discharge data from the glacier show high runoff in the ablation seasons of 1986–87, 1987–88, 1989–90 and 1990–91. The mass-balance calculations for these years are negative, the lowest being in 1990–91 (–1084 mm). Positive mass balance in 1988–89 (+358 mm) is attributed to comparatively heavy winter snowfall amounts and comparatively low temperatures during the ablation season. Significantly lower runoff was measured during this season. Negative mass-balance values during 1989–90 and 1990–91 are a result of comparatively dry winters and comparatively warm ablation periods, with monthly mean air temperatures 1.4–5.1 °C higher at the beginning of the ablation season (June and July) than the mean of the last 5 years.

Detailed ground-penetrating radar (GPR) surveys at 50 and 200 MHz on Hansbreen, a polythermal glacier in southern Svalbard, are presented and interpreted. Comparison of the variations in character of the radar reflections with borehole thermometry and water levels in moulins suggests that GPR can be used to study the hydrothermal properties of the glacier. The high resolution of the GPR data shows that the hydrothermal structure of the glacier is highly variable both along the centre line and on transverse profiles. Water contents for many places and depths within the glacier were calculated by estimating radar-wave velocities to point reflectors. We find typical water contents of 1-2% for the temperate ice, but wetter ice associated with surface crevassing and moulins (typically 4% water content). There is evidence that wet ice sometimes overlays drier ice. The hydrothermal structure is thus shown to be very complex. Temperature gradients in the cold ice indicate freezing rates of temperate ice below cold ice of 0.1-0.5 ma-1, while isolated point reflectors within the cold ice indicate large water-filled bodies that are probably related to the regular drainage structure of the glacier.

In this paper we present a simulation approach to mapping avalanche risk with application to settlements in Iceland. Two simulation models are developed to calculate the probability of avalanches travelling a certain distance, and of the flow being a specific width. These two simulation models, in combination with knowledge of the average frequency of avalanche occurrence, the variability in avalanche direction and the degree of loss caused by an avalanche, permit risk values to be determined for the areas of concern.

The conductive heat flux through the snow cover (Fa) is used as a proxy to examine the hypothesis that there is a significant heat flow from the Alaskan North Slope to the atmosphere because of the large number of lakes in the region. Fa is estimated from measurements of snow depth, temperature and density on tundra, grounded ice and floating ice in mid-April 1997 at six lakes near Barrow, northwestern Alaska. The mean Fa values from tundra, grounded ice and floating ice are 1.5, 5.4 and 18.6 W m2, respectively. A numerical model of the coupled snow/ice/water/soil system is used to simulate Fa and there is good agreement between the simulated and measured fluxes. The flux from the tundra is low because the soils have a relatively low thermal conductivity and the active layer cools significantly after freezing completely the previous autumn. The flux from the floating ice is high because the ice has a relatively high thermal conductivity, and a body of relatively warm water remains below the growing ice at the end of winter. The flux from the grounded ice is intermediate between that from the tundra and that from the floating ice, and depends on the timing of the contact between the growing ice and the lake sediments, and whether or not those sediments freeze completely. Using the estimated Fa values combined with the areal fractions of tundra, grounded ice and floating ice derived from synthetic aperture radar images, area-weighted Fa values are calculated for six areas. Fa values for the ice vary between 9.8 and 13.8 W m−2, and those from the ice plus tundra vary between 3.9 and 5.3 W m−2. The Fa values are similar to those observed in the sea-ice-covered regions of the south and north polar oceans in winter. The North Slope of Alaska may thus make a significant contribution to the regional energy budget in winter.