Surface-based ice-penetrating radar profiles were made across the active north margin (the Snake) of the upper part of Whillans Ice Stream (formerly Ice Stream B, branch B2), West Antarctica, at three locations. Low frequency (about 2 MHz) and the ground deployment of the radar allowed penetration through the near-surface zone of fracturing to detect internal layering and bed reflection characteristics on continuous profiles spanning from the slow-moving ice of Engelhardt Ridge well into the chaotic zone of the shear margin. Internal layers were tracked beneath the chaotic zone, where they are warped but remain continuous. The energy returned from internal layers showed no systematic changes associated with the transition from the undisturbed surface of the slow-moving ice into the fractured surface of the shear margin, thus indicating little effect from the surface crevasses on the penetration of the radar signal. Based on this calibration of the near-surface effects and corrections for path length, spreading and attenuation, we examine the spatial variation of bed reflectivity. Low bed reflectivity found under Engelhardt Ridge extends under the chaotic zone of the margin into fast-moving ice. We argue that the fast motion in a band along the margin is mediated by processes other than deformation of thick dilated till that is the source of lubrication allowing fast motion in the interior of the ice stream.

A survey of July 1st glacier, Qilian Shan, China, was carried out in 2002. Previously, the glacier’s boundary had been recorded in 1956, and further research had been carried out in the mid- 1970s and 1980s. Our survey reveals that area shrinkage and surface lowering have accelerated in the past 15 years. Surface elevation changes can result from changes in accumulation, surface melting and emergence velocity. The contributions of these elements to surface lowering are evaluated at the lower part of the glacier from observations of surface velocity, ice thickness and precipitation, and from temperature data near the glacier. Apart from the effect of glacier ice redistribution, our analysis reveals quantitatively that the recent accelerated glacier shrinkage has been caused by increasing temperature. Furthermore, it is established that meltwater discharge from the glacier in the past 17 years has increased due to glacier shrinkage, by about 50% over that from 1975 to 1985.

Extensive observations on Nivlisen, an ice shelf on Antarctica’s Atlantic coast, are analyzed and combined to obtain a new description of its complex glaciological regime. We generate models of ice thickness (primarily from ground-penetrating radar), ellipsoidal ice surface height (primarily from ERS-1 satellite altimetry), freeboard height (by utilizing precise sea surface information) and ice-flow velocity (from ERS-1/-2 SAR interferometry and GPS measurements). Accuracy assessments are included. Exploiting the hydrostatic equilibrium relation, we infer the ‘apparent air layer thickness’ as a useful measure for a glacier’s density deviation from a pure ice body. This parameter exhibits a distinct spatial variation (ranging from ≈2 to ≈16m) which we attribute to the transition from an ablation area to an accumulation area. We compute mass-flux and mass-balance parameters on a local and areally integrated scale. The combined effect of bottom mass balance and temporal change averaged over an essential part of Nivlisen is –654 ± 170 kg m–2 a–1, which suggests bottom melting processes dominate. We discuss our results in view of temporal ice-mass changes (including remarks on historical observations), basal processes, near-surface processes and ice-flow dynamical features. The question of temporal changes remains open from the data at hand, and we recommend further observations and analyses for its solution.

The instability of hanging glaciers and more generally of avalanching glaciers is discussed on the basis of observations performed on several glaciers located in the European Alps. A classification of avalanching glaciers is proposed, which allows a primary appreciation of the danger inherent in these glaciers. On the basis of field observations and results of numerical simulations of crevassing, the fracture processes which lead to major icefalls, as well as their recurrence, are analyzed. A method for predicting the time of failure is then discussed. The disaggregation of the unstable ice masses usually observed prior to large icefalls is investigated and its implications for the forecasting of icefalls analyzed.

We recorded electrical resistivity data at the base of four boreholes drilled through Haut Glacier d’Arolla, Switzerland. The data were acquired repetitively every hour over two diurnal hydrological cycles in the late melt season, separated by 10 days. Constrained three-dimensional (3-D) data inversion allowed reconstruction of hourly variations in bulk resistivity in the subglacial sediment layer. Inverted resistivity models reflect the establishment of channelized subglacial drainage in the study area between the two hydrological cycles, in agreement with previous work. Daily variations in bulk and water resistivity are in phase, and bulk resistivity amplitudes decrease away from the subglacial channel. Using selected electrical–hydraulic relationships, we estimate metre-scale changes in the hydraulic conductivity and porosity of the subglacial sediment layer, accounting for increasing clay content and decreasing median grain radius with distance from the channel. Hydraulic conductivity and porosity were respectively calculated to decrease from (6.4 ± 2.1) × 10–2ms–1 and 0.34 ± 0.01 at the channel to (3.3 ± 2.2) × 10–2ms–1 and 0.26 ± 0.01 at a distance of 5m from it. The hydraulic conductivity estimates are in agreement with previously inferred values, and the porosity estimates fall within the expected range for unlithified subglacial sediments. We conclude that collection and inversion of repeat 3-D subglacial resistivity data is feasible and has the capacity to generate multidimensional images of subglacial hydraulic processes and properties.

We analyzed the profiles of ionic chemical species in three 500 mm sections of an ice core from Dome Fuji, Antarctica, dated 3.0, 8.9 and 13.3 kyr BP (before present), and compared the profiles to those in the surface snow. The 3.0 and 8.9 kyr sections are from the Holocene and the 13.3 kyr section slightly predates the Holocene. The analyses were done on 2 mm thick slices within each section. At each depth, the primary ionic species were Na+, H+, Cl– and SO42 A The SO42, Na+ and Mg2+ levels varied with depth in each section over distances ranging from several millimeters to several centimeters. Also, the correlation coefficients between Na+ and SO4 and between Mg2+ and SO42 for each depth were 0.90 or greater, in contrast to the value of 0.59 or less in the surface snow (defined here as 0–3.4 m from the surface). These results suggest that almost all Na+ and Mg2+ in the Holocene ice exists as Na2SO4 and MgSO4 salts, and the formation of these salts occurs not only in the atmosphere during transport, but also in the firn layer.

Puncak Jaya, Irian Jaya, Indonesia, contains the only remaining tropical glaciers in East Asia. The extent of the ice masses on Puncak Jaya has been mapped from high-resolution IKONOS satellite images acquired on 8 June 2000 and 11 June 2002. Exclusive of Southwall Hanging Glacier, the ice extent on Puncak Jaya was 2.326 km2 and 2.152 km2 in 2000 and 2002, respectively. From 2000 to 2002, the Puncak Jaya glaciers lost a surface area of 0.174 km2 or 7.48% of their 2000 ice extent. Comparison of the IKONOS-based glacier extents with previous glacier extents demonstrates a continuing reduction of ice area on Puncak Jaya. By 2000, ice extent on Puncak Jaya had reduced by 88% of its maximum neoglacial extent. Between 1992 and 2000 Meren Glacier disappeared entirely. All remaining ice masses on Puncak Jaya continue their retreat from their neoglacial maxima. Comparison of 2000/2002 ice extents with previous extents suggests that these glaciers have not experienced accelerating rates of retreat during the last half of the 20th century. If the recession rates observed from 2000 to 2002 continue, the remaining ice masses on Puncak Jaya will melt within 50 years.

Complex glacier motion in the Grove Mountains region, Antarctica, is measured using four-pass differential synthetic aperture radar (SAR) interferometry (InSAR). The components of the motion vector field are resolved using a 44 day-separation Japanese Earth Resources Satellite-1 (JERS-1) InSAR pair and a European Remote-sensing Satellite-1/-2 (ERS-1/-2) tandem InSAR pair. The 44 day temporal baseline provides the sensitivity required to observe the range of ice motion (around 8–10ma–1), and the 1 day short baseline provides the best choice for glacier digital elevation model reconstruction. It is remarkable that the scattering field of the JERS-1 pair remained coherent over the long time interval and the interferometric fringes are clear. The overall ice flow is from east to west, downslope and towards Lambert Glacier. The regional flow is obstructed by nunataks extending north–south, with two wide gaps. Two narrow glaciers flow past the nunataks and rejoin each other at the downstream end. Regional morphology, and the resolved flow in the Grove Mountains area, suggests that ice flow is channeled throughout this eastern flank of Lambert Glacier.

This paper assesses a two-dimensional, vertically integrated ice model for its numerical properties in the calculation of ice-sheet evolution on a sloping bed using the shallow-ice approximation. We discuss the influence of initial conditions and individual model parameters on the model’s numerical behaviour, with emphasis on varying spatial discretizations. The modelling results suffer badly from numerical problems. They show a strong dependence on gridcell size and we conclude that the widely used gridcell spacing of 20 km is too coarse. The numerical errors are small in each single time-step, but increase non-linearly over time and with volume change, as a result of feedback of the mass balance with height. We propose a new method for the calculation of the surface gradient near the margin, which improves the results significantly. Furthermore, we show that we may use dimension analysis as a tool to explain in which situations numerical problems are to be expected.

Glaciers occupy an area of ~1600km2 in the Caucasus Mountains. There is widespread evidence of retreat since the Little Ice Age, but an up-to-date regional assessment of glacier change is lacking. In this paper, satellite imagery (Landsat Thematic Mapper and Enhanced Thematic Mapper Plus) is used to obtain the terminus position of 113 glaciers in the central Caucasus in 1985 and 2000, using a manual delineation process based on a false-colour composite (bands 5, 4, 3). Measurements reveal that 94% of the glaciers have retreated, 4% exhibited no overall change and 2% advanced. The mean retreat rate equates to ~8ma–1, and maximum retreat rates approach ~38 m a–1. The largest (>10 km2) glaciers retreated twice as much (~12ma–1) as the smallest (<1 km2) glaciers (~6ma–1), and glaciers at lower elevations generally retreated greater distances. Supraglacial debris cover has increased in association with glacier retreat, and the surface area of bare ice has reduced by ~10% between 1985 and 2000. Results are compared to declassified Corona imagery from the 1960s and 1970s and detailed field measurements and mass-balance data for Djankuat glacier, central Caucasus. It is concluded that the decrease in glacier area appears to be primarily driven by increasing temperatures since the 1970s and especially since the mid-1990s. Continued retreat could lead to considerable changes in glacier runoff, with implications for regional water resources.

Cerro Charquini, Bolivia (Cordillera Real, 5392 ma.s.l.) was selected as a site to reconstruct glacier recession since the maximum of the Little Ice Age (LIA) in the central Andes. Five glaciers, located on differently exposed slopes, present comprehensive and well-preserved morainic systems attributed to former centuries. The moraines were dated by lichenometry and show a consistent organization on the different slopes. The past geometry of the glaciers was reconstructed using ground topography and aerophotogrammetry. Lichenometric dating shows that the LIA maximum occurred in the second half of the 17th century, after which the glaciers have receded nearly continuously. Over the last decades of the 20th century (1983–97), recession rates increased by a factor of four. On the northern and western slopes, glaciers receded more than on the southern and eastern slopes (by 78% and 65% of their LIA maximum area, respectively). The mean equilibrium-line altitude (ELA) rose by about 160 m between the LIA maximum and 1997. Recession rates were analysed in terms of climatic signal, suggesting that glacier recession since the LIA maximum was mainly due to a change in precipitation and that the 19th century may have been drier. For the 20th century, a temperature rise of about 0.6°C appears to be the main cause of glacier recession. Recent climatic conditions from 1983 to 1997 correspond to a mass deficit of about 1.36m w.e.a–1. If such conditions persist, the small glaciers below 5300ma.s.l. in the Cordillera Real should disappear completely in the near future.

The development of land-fast sea ice and overlying snow was monitored during a 4 week period, until the snow cover had completely disappeared, at a site in the Gulf of Bothnia, Baltic Sea (63.57ú N, 19.85° E). The meteorological and radiative boundary conditions were continuously recorded. During the observation period, a 15 cm thick snow layer on the ice was transformed into a 7 cm thick granular ice layer (superimposed ice) on the ice surface, contributing significantly (about 11%) to the total ice thickness. Approximately 1 cm w.e. of the snow was sublimated. Neither snow-ice formation nor basal ice growth was significant during the same period. The salinity and isotopic (δ18O) composition of the ice indicated that prior to the experiment a 7 cm layer of superimposed ice had already formed. Hence, superimposed ice layers contributed 22% of the total ice thickness by the time all snow had disappeared. The advancing spring, decrease in surface albedo, diurnal cycle in the incoming solar radiation, and synoptic-scale changes in the cloud cover and the air–ice turbulent heat fluxes caused variations in the heat budget of the snowpack. Superimposed ice formation due to refreezing of meltwater occurred during most nights of the study period, and the most important refreezing periods were under such synoptic conditions that the air and snow surface temperatures also remained below zero during daytime. In contrast to typical summer conditions in polar oceans, low snow surface temperatures acted as the primary heat sink for the refreezing of meltwater.

Brittle structures exposed in the ablation area of Pasterzenkees, Austria, were interpreted using aerial photographs and maps covering a period of 100 years. The most common structural features observed in aerial photographs are: (1) normal faults, which are particularly well developed along the lateral margins of the glacier and at the terminus; (2) large-scale tension gashes and Riedel shears that develop along the northeastern lateral margin of the glacier and between ice-flow units; (3) thrust faults, which develop at the terminus and cross-cut the full width of the glacier; and (4) band ogives. Longitudinal and transverse topographic profiles are available for the period covered in this study, and ice-flow velocity data are available from 1927. These data provide a means for interpreting the variations of observed structures in terms of ice-flow velocity. Thrust faults predominantly develop during periods of glacier retreat, when the glacier snout becomes an obstacle. Normal faults typically develop in areas of high glacier surface relief and are interpreted as gravity collapse structures. The orientation of sub-vertical, wide open crevasses along the lateral margin of Pasterzenkees varied. These variations are interpreted as reflecting changes of the flow regime and indicate a transition from simple shearing to transtension during a period of ice-flow deceleration.

Rock tunnels beneath Engabreen, northern Norway, permit access to the ice-bedrock interface beneath a 210 m thick glacier. Eight load cells have been installed in the bedrock of the glacier sole along a 22 m transect. With some interruptions, the load cells have been logged at 15 min intervals since December 1992; here we analyse the records until 2003. Load-cell signals measure stresses acting normal to the bedrock, and usually log the pressure of the thin water film between the basal ice and the bed. Occasionally there are distinct pressure events, characterized by short-lived (hours) local minima, often followed by a maximum before decaying to background load-cell pressure. The amplitudes of these pressure events are of the order of 0.01–1 MPa and depend on the placement of the sensor and the state of the subglacial drainage system. We identify winter and summer pressure regimes. The winter regime is characterized by few pressure events of large pressure amplitude influencing all load cells. A lag of 0–6 days is observed between surface forcing (rain) and pressure events during winter. The summer regime typically has periods of daily pressure events of low amplitude. No delay is seen between surface forcing and pressure events during the summer regime. In summer, the onset of a pressure event is correlated with a local maximum of the derivative of the subglacial discharge record, whereas no such relation is found during the winter regime. The transition from winter to summer (May/June) is easily detectable and is strongly correlated with a rapid increase in subglacial discharge and the transition to a dominating R-channel system. The autumn transition is less clearly defined, but has usually occurred by the beginning of November. Stress bridging, an increase in bed pressure at the edge of low-pressure channels, is recorded during the summer regime. Water pressures at the bed are connected or unconnected to the drainage system. Pressure increases in the connected system, beyond local normal stress values, lead to an uplift of the connected system and a pressure drop in the unconnected system. The occurrence of pressure events is determined by the capacity of the drainage system. Uplift is controlled by local normal stress values and not mean ice-overburden pressure.

A widely used method for investigating palaeotemperatures is to analyze local proxy records (e.g. ice cores or deep-sea sediment cores). The interpretation of these records is often not straightforward, and global or hemispheric means cannot be deduced from local estimates because of large spatial variability. Using a different approach, temperature changes over the last glacial cycle can be estimated from sea-level observations by applying an inverse method to an ice-sheet model. In order to understand the underlying physical mechanisms, we used a 1-D ice-sheet model and a 3-D coupled thermodynamic ice-sheet–ice-shelf–bedrock model to investigate the importance of several physical processes for the inverse temperature reconstructions. Results show that (i) temperature reconstructions are sensitive to the employed formulation of mass balance, (ii) excluding thermodynamics in the ice sheet leads to a smaller temperature amplitude in the reconstruction and (iii) hysteresis in the non-linear relation between sea level and temperature occurs as a consequence of ice redistribution in the process of merging and separation of ice sheets. The ice redistribution does not occur if the geometry does not support the formation of a relatively flat dome, which tends to be preserved in warming conditions.

We present a novel method of improving signal-to-noise ratio in radargrams. The method uses singular spectrum analysis (SSA) to separate each individual radar trace into orthogonal components. The components that explain most of the original trace variance contain mainly physically meaningful signal, while the components with little variance tend to be noise. Adding the largest-magnitude components together until the sum of components accounts for the variance above the noise level (typically 60–80%) of the original trace variance results in a much cleaner radargram with more easily seen internal features than in traditionally filtered data. The radargrams can be further enhanced by envelope-detecting the SSA-filtered data, as this measure of instantaneous energy minimizes the deleterious effects of innumerable phase changes at dielectric boundaries. Subsequent incoherent stacking results in far more structured radargrams than are achieved with traditionally processed radar data and amplitude stacking.

Slab avalanche release requires fracture initiation and propagation in a weak snowpack layer. While field tests of weak-layer strength are useful for fracture initiation, the challenge remains to find a verified field test for fracture propagation. We introduce the two current versions of a field test for fracture propagation propensity, and report results of testing conducted in the Columbia Mountains of British Columbia, Canada, during the winter of 2005. By extending the column of a stability test approximately 3 m in the downslope direction, the test method allows for the development of a flexural wave in the slab, and thereby maintains the contribution of this wave and the associated weak-layer collapse to the fracture process. Fracture lengths collected on a day and location where the propagation propensity of the snowpack was locally high show a bimodal distribution, with approximately 50% of observed fractures similar to those collected in stable snowpacks, and approximately 50% with much longer fracture lengths.