This paper describes fine-resolution measurements of glacier surface strain rates very close to the margin of Russell Glacier, West Greenland. Measurements at a small scale make possible detailed analysis of strain patterns close to the glacier margin, and suggest that strain rates vary over small areas. The strain pattern is determined by ice flexure over subglacial obstacles as well as by seasonally variable marginal retardation and by the orientation of the ice margin relative to the flow direction.

Temperature compensation built into many conductivity meters becomes inaccurate in low-temperature waters typical of glacial melt streams. Experiments using simulated glacial waters show a linear dependence of electrical conductivity on temperature from 0.3° to 25°C. The slope of the temperature–conductivity relation is linearly dependent on the conductivity of the solution, allowing a numerical or analytical temperature correction to be made. To minimize error introduced by temperature compensation, measurements of electrical conductivity in glacial streams should be corrected to a low standard temperature, and 0°C is suggested.

Laws of snow metamorphism have been introduced in a numerical model which simulates the evolution of temperature, density and liquid-water profiles of snow cover as a function of weather conditions.

To establish these laws, the authors have summarized previous studies on temperature gradient and on wet-snow metamorphism and they have also conducted metamorphism experiments on dry or wet fresh-snow samples. An original formalism was developed to allow a description of snow with parameters evolving continuously throughout time.

The introduction of laws of metamorphism has improved significantly the derivation of the settlement of internal layers and of snow-covered albedo, which depend on the simulated stratigraphy, i.e. the type and size of snow grains of different layers of the snow cover.

The model was tested during a whole winter season without any re-initialization. Comparison between the simulated characteristics of the snow cover and the observations made in the field are described in detail. The model proved itself to be very efficient in simulating accurately the evolution of the snow-cover stratigraphy throughout the whole winter season.

Remote-sensing methods are the primary ones used for ice mapping in the Baltic Sea. A major methodological improvement is now being introduced by satellite radars due to their weather independency and high resolution. To learn how to use ERS-1 synthetic aperture radar (SAR) data, an extensive field programme BEPERS (Bothnian Experiment in Preparation for ERS-1) with airborne SARs has been arranged. The BEPERS pilot study was undertaken in 1987 using the French VARAN-S X-band SAR. The SAR was flown on 1 day over four study areas of size approximately 10 km x 50 km, and intensive validation observations were made. The data were most useful for the education they provided on how to work with SAR in sea-ice mapping. They have been used for developing SAR image-analysis methods, back-scatter modelling investigations and geophysical validation of SAR imagery. Cleaning-up of images consisted of speckle reduction and segmentation. Back-scatter characteristics of undeformed ice and ridges were examined. Ice-type classification was based on the box-classification method. Eight ice types were defined but basically only two types, undeformed ice/open water and deformed ice, could be discriminated. Two basic problems of high practical importance remained: how to discriminate between (1) open water and undeformed ice, and (2) ridged ice and brash ice. The data further showed illustrative examples of SAR imagery over sea ice.

The long-term monitoring of Lewis Glacier on Mount Kenya serves as a basis for the study of glacier evolution in response to climatic forcing through modeling of its ice flow and mass budget. Following up on an earlier modeling and prediction study to 1990, this paper examines the ice-mass and flow changes in relation to the net-balance conditions over 1986–90. A model experiment using as climatic forcing the observed 1978–86 vertical net-balance profile yielded a volume loss and slow down of ice flow more drastic than observed during 1986–90. The causes of this discrepancy were examined in successive model experiments. Realistic simulations of mass-budget and thickness changes over 1986–90 are obtained using as input the net-balance forcing for the same period rather than for the preceding 1978–86 interval, and approximate flow velocities. With the same net-balance forcing and a completely stagnant Lewis Glacier, the elimination of mass redistribution by ice flow acts to mitigate the loss of volume and thickness in the upper glacier, and to accentuate it in the lower glacier. Accordingly, the observed 1986–90 net-balance profile along with the 1990 ice-flow velocities provide suitable input for the modeling of Lewis Glacier changes to 1994. Under continuation of the 1986–90 climatic forcing, ice thinning ranging from less than 1 m in the upper glacier to more than 7 m in the lower glacier, and a total volume loss of order 57 × 104 m3, are anticipated over the 1990–94 time interval.

Systematic isotopic studies based on natural and artificial radio-isotopes (32Si, 137Cs), stable isotopes (δ18O) and total β activity measurements have been carried out on Chhota Shigri glacier, Himachal Pradesh, central Himalaya, to study the dynamics of the ice, meltwater composition and to identify the deposition of the Chernobyl fall-out in the Himalayan region.

Using 32Si concentrations, the snout ice has been dated at ~ 250 years, based on which the past average surface ice-flow rate has been estimated as ~ 28 m year−1. Based on δ18O variations, in a shallow ice core, the accumulation rate of the ice has been estimated at ~520kgm−2 year−1. 32Si measurements of snout ice and englacial meltwaters indicate that at least 55% of the snow meltwater mixed with 45% of the old ice-melt water that emerged from englacial streams in the month of August 1987. Deposition of the artificial radionuclide (137Cs) and the very high total β activity observed in snow samples on Chhota Shigri glacier give the first evidence of Chernobyl fall-out deposition in the Indian Himalaya.

The subglacial drainage system of South Cascade Glacier is inferred by examining how discharges of water, solutes and artificial tracers in streams draining the glacier respond to meltwater and precipitation on the glacier surface. Results indicate that the glacier is divided into three (two large and one small) drainage basins, each drained by a single stream. The positions of drainage divides on the glacier surface do not always lie directly over those at the base of the glacier. Comparisons of water discharge and cation load in each of the two main streams indicate that subglacial hydraulic processes differ between drainage basins. One stream drains from a conduit that is isolated in its lower reach from the surrounding subglacial region and receives water routed englacially from the surface. The upper reach of the conduit also receives water routed englacially from the surface as well as from a distributed subglacial flow system. The other main stream drains from a conduit coupled to a debris layer beneath the glacier. Observations of the layer in natural ice tunnels indicate that the water may flow within a thin layer of debris. A one-dimensional model of flow through the debris layer can explain both the base-flow and diurnal variations of the second main stream.

Triaxial and uniaxial compression tests have been carried out at –10°C on granular ice in order to study the role of microcracking on failure in the ductile-brittle transition zone. In the triaxial tests, the effect of confining pressure and strain rate on the crack population, as well as on strength and strain at the peak stress, was investigated. In the uniaxial tests, we measured the evolution of elastic and non-elastic components of deformation with the stress-strain history. The concept of effective stress, with a single scalar damage variable, was used to calculate the effect of microcracking on the strain components.

The classic sliding theories usually assume that the sliding motion occurs frictionlessly. However, basal ice is debris-laden and friction exists between the substratum and rock particles embedded in the basal ice. The influence of debris concentration on the sliding process is investigated. The actual conditions where certain types of friction apply are defined, the effect for the case of bed separation due to a subglacial water pressure is studied and consequences for the sliding law are formulated. The numerical modelling of the sliding of an ice mass over an undulating bed, including the effect of both the subglacial water pressure and the friction, is done by using the finite-clement method. Friction, seen as a reduction of the driving shear stress due to gravity, can be included in existing sliding laws which should contain the critical pressure as an important variable. An approximate functional relationship between the sliding velocity, the effective basal shear stress and the subglacial water pressure is given.

Non-linear acoustic techniques have been used to investigate the properties of a pond-ice cover. The field tests showed that the position of the lower and upper surfaces, as well as some internal bubble layers, could be located with centimetre-scale resolution. These results are presented, along with an outline of parametric array theory and a description of the instrumentation. The data indicate that dynamic modelling of the acoustic response of ice will enable remote estimation of some of the bulk physical properties of the ice such as thickness, roughness and possibly brine volume. It follows that this approach to non-destructive imaging of ice properties may find application in a wide range of ice-engineering problems.

The accumulation area ratio (AAR) for Himalayan glaciers representing zero mass balance is substantially lower than for North America and Europe. Regression analysis suggests 0.44 for the AAR representing zero mass balance in the western Himalaya. A good correlation was observed when this method was applied to individual glaciers such as Gara and Gor-Garang in Himachal Pradesh, India. The correlation coefficients (r), using 6 and 7 years of data, respectively, were 0.88 and 0.96 for Gara and Gor-Garang Glaciers, respectively. However, when data from six western Himalayan glaciers were correlated, the correlation was 0.74. The AAR was also estimated by using Landsat images which can be useful in obtaining a trend in mass balance for a large number of Himalayan glaciers for which very little information exists.

A higher correlation was observed between equilibrium-line altitude (ELA) and mass balance. The field data from Gara and Gor-Garang Glaciers shows a high correlation coefficient, i.e. −0.92 and −0.94, respectively. The ELA values obtained from the Landsat satellite images combined with topographic maps suggest positive mass balance for the year 1986–87 and negative for 1987–88.

The likely genesis of low-relief ice-sheet lobes is episodic subglacial floods which occur in the form of water sheets hundreds of kilometers wide. During the few weeks prior to termination, when the flood discharge is large, the average thickness of a water sheet beneath the Lake Michigan lobe likely exceeded 1 m and reached thicknesses in excess of 6 m. The release of a part of the basal shear stress during this period produced an elongation of the lobe of tens of kilometers.

Borehole inclinometers are standard equipment for field glaciologists and are commonly used for investigating the flow law of ice and for measuring the spatial position of englacial and sub-surface sensors. The recent development, at the University of British Columbia (UBC), of a prototype inclinometer that employs a three-component fluxgate magentometer to obtain a compass bearing has stimulated our interest in borehole inclinometry. Following a review of various approaches to glacier inclinometry, we present a unified theory of data interpretation that can be applied to all inclinometers, discuss the application of the theory to the UBC inclinometer, and discuss the sensitivity of the theory to error in the data.

Gravity wants to pull an ice sheet to the center of the Earth, but cannot because the Earth’s crust is in the way, so ice is pushed out sideways instead. Or is it? The ice sheet “sees” nothing preventing it from spreading out except air, which is much less massive than ice. Therefore, does not ice rush forward to fill this relative vacuum; does not the relative vacuum suck ice into it, because Nature abhors a vacuum? If so, the ice sheet is not only pulled downward by gravity, it is also pulled outward by the relative vacuum. This pulling outward will be most rapid where the ice sheet encounters least resistance. The least resistance exists along the bed of ice streams, where ice-bed coupling is reduced by a basal water layer, especially if the ice stream becomes afloat and the floating part is relatively unconfined around its perimeter and unpinned to the sea floor. Ice streams are therefore fast currents of ice that develop near the margins of an ice sheet where these conditions exist. Because of these conditions, ice streams pull ice out of ice sheets and have pulling power equal to the longitudinal gravitational pulling force multiplied by the ice-stream velocity. These boundary conditions beneath and beyond ice streams can be quantified by a basal buoyancy factor that provides a life-cycle classification of ice streams into inception, growth, mature, declining and terminal stages, during which ice streams disintegrate the ice sheet. Surface profiles of ice streams are diagnostic of the stage in a life cycle and, hence, of the vitality of the ice sheet.

Short-term variations in ice-flow velocity were obtained at intervals of a few hours and a few days in the ablation area of Glaciar Soler, Patagonia, Chile, in November 1985. A maximum flow rate was measured at about four times the minimum value. A good correlation, with a time lag of 7.5 h, was found between the ice-flow velocity in the lower reaches and the amount of water discharge from the glacier terminus. It was concluded, therefore, that the velocity variations should have resulted from the variations in basal sliding velocity which is strongly controlled by the subglacial water pressure.

The position of the UpB camp on Ice Stream B, West Antarctica, has been measured 14 times during three field seasons. TRANSIT (doppler) satellite receivers were used. At the resolution of our measurements (2 m), no significant velocity variation was detected over time intervals ranging from 1 d to 2 years.

Two deep-drilling projects (GISP2 and GRIP) in central Greenland will provide ice cores for paleoclimate studies. Drilling decisions and preliminary interpretations require age-depth curves (time-scales). Using a finite-element momentum-balance model, we calculate the modern ice-flow pattern on the flowline through the two drill sites. Our model appears to require relatively soft ice either throughout the ice sheet or below the Wisconsinan-Holocene transition in order to match the modern geometry and mass balance. By scaling the ice velocity to an assumed mass-balance history throughout the past 200 000 years, we estimate the time-scales at both sites. At GISP2, a flank site, we place the 10 000 years BP isochrone (representing the Wisconsinan-Holocene transition) at 1535 m ice-equivalent depth. At GRIP, on the ice divide, the corresponding depth is 1377 m. Our calculations show ice older than 200 000 years at 100 m above the bed at both coring sites. The time-scale calculation can be used for drilling decisions and preliminary interpretations. It should be refined as more regional-survey and ice-core data become available.

Accumulation studies along the lowermost 100 km of Jakobshavns Isbræ show that the local net balance above the equilibrium line (1210 m elevation in 1986) is significantly less than that measured along the EGIG line about 100 km further north. This indicates the presence of a precipitation low in this region which will affect any global mass-balance assessment for the Jakobshavns Isbræ drainage basin. Comparison of the estimated calving and ablation fluxes shows that calving removes about twice as much mass from this drainage basin as does melting. Basal melting over the entire basin accounts for about 20% of the total ice loss by ablation. Temperature measurements at 12 m depth along the same section of the Isbræ show the warming effects of refreezing meltwater and cooling effects of severe crevassing. In addition, there is a significant variation in temperature across the fast-moving ice stream which is probably caused by deformation heating in the shear margins which delineate the ice stream within the ice sheet. This lateral temperature gradient could be important in ice-stream dynamics through its effects on ice rheology. Detailed measurements within the percolation fades show that surface melt can penetrate up to 3 m by piping in cold firn, and, upon refreezing, can cause significant warming at these depths.

Observations at Athabasca Glacier and elsewhere suggest that basal sliding can account for a very significant part of total glacier motion, and that sliding rates vary significantly across a glacier section. The ability to model such spatial variations in basal velocities is important in understanding flow in valley glaciers, as well as in predicting spatial patterns of glacial erosion that drive land-form development models. With a sliding law in which the basal velocity is dependent on the basal shear stress and inversely dependent on the effective pressure at the bed, it is possible to predict an overall flow pattern that is consistent with the empirical data, if it is assumed that friction increases close to the margin of a glacier.

Snow dams have been observed in many stream channels in the Arctic and the sub-Arctic, but there is no theoretical analysis of their decay processes. These processes include snowmelt, seepage erosion, down-cutting, snow-slope failure and flotation of the snow dam. Snowmelt can be determined by the energy balance. Seepage of water will cause serious erosion at the point where the stream or pond level in front of the dam intersects the dam face. Erosion by water overflowing the dam is due to down-cutting, under-cutting and thermal erosion. Down-cutting increases with discharge but decreases with the shear strength of the snow. Where a hydraulic jump occurs downstream of the dam, under-cutting can accelerate dam decay. Thermal erosion, depending mainly on water temperature, may be less significant than the previous two processes. Slope failure occurs when the driving forces exceed the resisting forces, and these are affected by snow property and snow load. When the stream or pond level downstream of the dam rises rapidly, the dam is prone to float. As peak flow often occurs during the break-up period, results of the theoretical study of the mechanisms of snow-dam decay will improve flood forecasting for Arctic streams.