Deformation beneath soft-bedded glaciers may be a physical mechanism that contributes to flow instabilities such as surging. If the role of bed deformation is to be understood, a rheological description is required, but the development of a rheology is hampered by a lack of in situ stress and strain measurements. In this paper, we describe four techniques for measuring subglacial strain. Three of these give continuous strain measurements, a capability that permits calculation of instantaneous strain rates and allows comparison of strain data with other time series. To demonstrate the practicability of the techniques, sample results from three summers of experimentation beneath Trapridge Glacier. Yukon Territory, are presented. The data show that subglacial strain rate can vary in amplitude and polarity on an hourly time-scale,and that the instantaneous strain rate can exceed the mean strain rale by an order of magnitude. Observed negative strain rates suggest extrusive flow within basal sediments.

The transition from a dry, sub-freezing snow cover to isothermal snow freely conducting water was observed near timberline in the Sierra Nevada, California, U.S.A. Although there were few major stratigraphic boundaries in the snowpacks observed, minor textural discontinuities were sufficient to both route water laterally downslope and later create ice layers at these strata. During the first few days of snowmelt, downslope movement of water parallel to the strata allowed water to bypass lower layers of the snow cover on slopes and converge in microtopographic depressions on level terrain, creating isolated drains through the snowpack. In addition to these drains, flow fingers associated with holes in the ice sheets delivered water to the base of the snow several days prior to the thorough wetting of the snowpack. Formation of a series of ice lenses just below the snow surface every night released much more latent heat than did ice-layer development within the bulk of the snowpack or at its base.

Five statistical techniques to determine peaks in ice-core time series are presented and compared. The ice-core time series, representing different signal characteristics, comprise electrical conductivity measurements (ECM), dielectric properties (DEP) and sulphate. Three techniques (I–III) utilize all the data in the time series to estimate significant thresholds for identifying peaks. Technique IV applies a moving window and conducts a statistical inference within the defined window. In technique V, a family of smoothed estimates of the ice-core time series is produced, and statistical tests are performed on the significant changes in the derivative of the estimates. The correction of the significance level, α, due to multiple tests is introduced and implemented in techniques II–V. The threshold obtained by techniques I–III is determined by the influence of the error term on the global variance estimate, whereas the threshold of IV is determined by the data within the window. The success of identifying peaks with technique V is dependent on the redundancy in the data, i.e. the sampling rate. It is concluded that techniques II and III are superior to the other techniques due to their simplicity and robustness.

Аn excellent photographic database, comprising 12 sets of vertical aerial photographs taken between 1948 and 1993, has enabled the structural glaciology of surge-type Variegated Glacier, Alaska, U.S.A., tо be analyzed for a period that spans two surge cycles and includes the effects of three surges. The glacier-wide patterns of crevasse development at the end of each of these surges, which occurred in 1947–48, 1964–65 and 1982–83, were remarkably similar, despite differences in surge intensities and extents. At the end of each surge a three-fold longitudinal zonation of crevasse orientation characterised that part of the glacier that had been affected by surge activity. This zonation comprises: (i) transverse crevasses in a narrow upper zone, (ii) superimposed longitudinal and transverse crevasses in an extensive middle zone, and (iii) longitudinal crevasses in a narrow lower zone. The main difference between patterns produced by successive surges was small variations in the extent оf the three zones. During quiescence, the most intense crevasse development occurs in the upper glacier in the latter part of quiescence, and is a function of increasingly steep down-glacier velocity gradients developing in the buildup to the impending surge. Some crevassing also develops in the central part of the glacier in mid-quiescence that is not related to the surge nature of the glacier, but to normal flow processes over bedrock. There is no structural evidence to indicate that any of the tributaries of Variegated Glacier surge. This observation suggests that the bulb-like loops in the medial moraines at Variegated Glacier form during quiescence as a result of the steady flow of the tributaries into the stagnant trunk glacier, rather than as a result of tributary surges.

A sub-sea permafrost drilling program was conducted near Prudhoe Bay, Alaska. Sub-sea permafrost was found at all offshore drill sites and was characterized as being ice bonded or unbonded. The unbonded sub-sea permafrost occurred in a thin layer at the sea bed; the near-shore thickness of this layer appears to be controlled by the bathymetry. The mean annual sea-bed temperatures were about —3.4°C at 203 m offshore, — 1.1°C at 481 m offshore and —0.7°C at 3370 m offshore. The thermal diffusivity was about 21 m2 a-1 for unbonded sandy gravel with silt. The sub-sea permafrost soils were sandy gravel with some silt overlain by a thin layer of silty sand which increased in thickness from a few meters near shore to about 14 m at 3370 m offshore. A few small ice lenses were found in a hole 195 m offshore but no massive ice was observed. Pore-water salt concentrations at the sea bed were 3-4 times that of normal sea-water where ice was frozen to the sea bed and 1–2 times that of normal sea-water otherwise. Preliminary laboratory experiments showed that the saturated hydraulic conductivity of the unbonded sub-sea permafrost was about 10–6–10–7 m s–1. The permeability of the subterranean permafrost was zero.

In this study, we explore small-scale (~1 to 20 km) thermal-refractive effects on basal geothermal heat flux (BGHF) at subglacial boundaries resulting from lateral thermal conductivity contrasts associated with subglacial topography and geologic contacts. We construct a series of two-dimensional, conductive, steady-state models that exclude many of the complexities of ice sheets in order to demonstrate the effect of thermal refraction. We show that heat can preferentially flow into or around a subglacial valley depending on the thermal conductivity contrast with underlying bedrock, with anomalies of local BGHF at the ice–bedrock interface between 80 and 120% of regional BGHF and temperature anomalies on the order of ±15% for the typical range of bedrock conductivities. In the absence of bed topography, subglacial contacts can produce significant heat flux and temperature anomalies that are locally extensive (>10 km). Thermal refraction can result in either an increase or decrease in the likelihood of melting and ice-sheet stability depending on the conductivity contrast and bed topography. While our models exclude many of the physical complexities of ice behavior, they illustrate the need to include refractive effects created by realistic geology into future glacial models to improve the prediction of subglacial melting and ice viscosity.

The mechanical behavior and microstructural evolution of laboratory-prepared, particle-free fresh-water ice and ice with 1 wt.% (~0.43 vol.%) silt-sized particles were investigated under creep with a stress level of 1.45 MPa at −10°C. The particles were present both within the grains and along the grain boundaries. The creep rates of specimens with particles were always higher than those of particle-free ice. Dynamic recrystallization occurred for both sets of specimens, with new grains nucleating along grain boundaries in the early stages of creep. The ice with particles showed a higher nucleation rate. This resulted in a smaller average grain-size for the ice with particles after a given creep strain. Fabric studies indicated that ice with particles showed a more random orientation of c axes after creep to ~10% strain than the particle-free ice.

The diffusion of stable water isotopes in snow was measured in two controlled laboratory experiments. Two batches of snow of different isotopic composition were stacked alternately with varying layer thicknesses. The stack was stored in a freezer room at constant temperature for several months, and sampled at regular intervals to analyse the diffusion. Measured isotope profiles were fitted to a theoretical model with diffusion length as the fit parameter. In the first experiment, we observed a difference in diffusion rates between layers of different thicknesses, which is likely caused by layers of snow not being in proper contact with each other. In the second experiment we found very good agreement between measurements and model results. The measured diffusivity is compared with theory, in which we mainly focus on the temperature dependence of the ice–vapour fractionation factors. This temperature dependence is slightly different for the different isotopes of water, which leads to a difference in diffusion rates. We illustrate how our set-up can be used to measure the ratio between ice–vapour fractionation factors of oxygen-18 and deuterium, which determine the relation between the difference in diffusion and the firn temperature.

Basal ice at the glacier Østerdalsisen commonly deforms under stress within cavities at the glacier bed. Some ice subject to rapid strain, however, fails in a brittle manner. The creep-rate of basal ice is influenced by sediment particles within it, and by ice grain-size and structural anisotropy; a flow law for such ice has not been derived experimentally. Near the glacier bed, particles enclosed in ice may migrate as a result of stress concentration, bubbles may be flattened, and anisotropic structures may develop. Recrystallization leads to grain-size changes. Both regelation spicules and layers of regelation ice, generally thin, form within subglacial cavities but their occurrence is limited. Regelation-ice formation has a significant influence on the distribution of rock debris at the ceilings of subglacial cavities.

Seismic refraction and reflection shooting was carried out along three profiles about 10 km long, angled 60° to one another, near “Byrd” station, Antarctica, during the 1970–71 field season. No dependence of velocity upon azimuth was found, but velocities at 200 or 300 m depth were slightly greater than at a site 30 km away where measurements were made in 1958. The difference can probably be attributed to different ice fabrics arising from a 50% difference in snow accumulation rates at the two sites. The velocity depth and density–velocity functions at the two sites are also significantly different, but close agreement was found at each site between the depths to significant changes in the velocity gradient and the depths of fundamental change in the densification process. Such agreement may permit density–depth curves, and consequently accumulation rates, to be measured by seismic refraction shooting alone.

The reflection shooting on a common reflection-point profile led to a good determination of mean velocity through the ice as a function of angle of incidence. The results agree closely with similar measurements at the 1958 site, and with an anisotropic model based on glaciological and sonic logging observations in the deep drill hole. The mean vertical velocity of 3.90–3.93 km/s through the solid ice is about 2% higher than has commonly been used for determinations of ice thickness from seismic reflection shooting.

Laboratory experiments on the growth of sea ice in a very thin plastic tank filled with salt water, cooled from above and insulated with thermopane, clearly show the formation and development of brine drainage channels. The sea-water freezing cell is 0.3 cm thick by 35 cm wide by 50 cm deep; the thermopane insulation permits the ice interior to be photographed. Experimentally, we observe that vertical channels with diameters of 1 to 3 mm and associated smaller feeder channels extend throughout the ice sheet. Close examination of the brine channels show that their diameter at the ice-water interface is much narrower than higher up in the ice, so that the channel has a “neck” at the interface. Further, oscillations occur in the brine channels, in that brine flows out of the channel followed by a flow of sea-water up into the channel. Theoretically, a qualitative theory based on the difference in pressure head between the brine inside the ice and the sea-water provides a consistent explanation for the formation of the channels, and the onset of a convective instability explains the existence of the neck. Finally, an analysis based on the presence of the brine-channel neck provides an explanation for the observed oscillations.

Methods used in measuring the mass balance of the Devon Island ice cap are described. The use of dyes and melt trays is recommended in the superimposed-ice and firn zones of sub-polar glaciers. The north-west part of the ice cap was studied in most detail and has had a slightly negative net balance for the period 1961-66. An inverse relationship between mean net balance (bn) and elevation of the equilibrium line in the north-west part of the ice cap indicates that the mean net balance there would be zero with an equilibrium line at 920 m (±80 m) elevation. Accumulation on the ice cap is greatest in the south-east but the measurements suggest that the mean net balance there is similar to the mean net balance on the rest of the ice cap. It is concluded that the present accumulation pattern must have existed for several hundreds, and possibly thousands of years. A study of firn stratigraphy and of variations in the elevation of the firn and equilibrium lines indicates that between 1961 and 1966 only 1962 had a more negative mean net balance than the average value for the period 1934-60. During the same 26 year period the net balance at 1 787 m elevation has varied, but summer conditions do not appear to have changed significantly.

Glaciers depicted on old maps reveal their historical extents, before the advent of aerial and satellite remote sensing. Digital glacier inventories produced from these maps can be employed in assessments of centennial-scale glacier change. This study reconstructs the ~1899 (covering the period 1882–1916) glacier extent in Nordland, northern Norway, from historical gradteigskart maps, with an emphasis on examining the accuracy of the mapped glaciers. Glacier outlines were digitised from georectified scans of the analogue maps in a raster graphics editor and were subsequently inventoried in a GIS. The accuracy of the historical glacier extent was established from written descriptions and landscape photographs created during the original field surveys, and further validated against independent glacier outlines of (1) the maximum Little Ice Age extent derived from geomorphological evidence, and (2) the 1945 extent derived from vertical aerial photographs. An overall uncertainty of ±17% is associated with our inventory. Nordland's glaciers covered an area of 1712 ± 291 km2 in 1899. By 2000, total ice cover had decreased by 47% (807 ± 137 km2) at a rate of 6% 10 a−1 (80 ± 14 km2 10 a−1). The approach presented here may serve as a blueprint for future studies intending to derive glacier inventories from historical maps.

A temperate glacier is defined as a glacier containing liquid inclusions in which the concentration of salts is not too high. Nevertheless these salts suffice to produce a depression in temperature comparable with that due 10 the pressure, and much greater than that due to interfacial energies. Because of this a large part of the liquid water present in the ice is not mobile, contrary to the theory of Nye and Frank. Deformation and recrystallization is bound to close off capillary intergranular channels, for glacier ice is usually impermeable. An explanation is given of why firn, at a depth which the annual cold wave does not reach, is nevertheless transformed relatively suddenly into practically impermeable ice. Saline inclusions will migrate with a velocity inversely proportional to the potential temperature (difference from the melting point of pure ice at the pressure in question) and proportional to the gradient of this potential temperature. This velocity, the salinity, the liquid water content, and the ice temperature, parameters which are all functions of the depth, are calculated for a steady state in a stagnant or moving glacier. Under the action of anisotropic stresses, isolated inclusions perpendicular to the maximum compressive stress will enlarge at the expense of their neighbours. If however the two inclusions are connected by a capillary channel, no enlargement occurs, but instead the salt content decreases, evacuated to the other inclusion. This process would constitute an objection to the theory of glacier sliding by melting and refreezing around small obstacles, unless new subglacial mechanisms were to occur.

The paper considers how to calculate the budget history of a glacier from the historical record of its advance and retreat. The linearized differential equations developed in previous papers are used to compute a series of coefficients g(n). The g(n) operate directly on a record of annually spaced observations of the position of the glacier terminus, to give the time variation of net budget. The g(n) are calculated for South Cascade Glacier, Washington, U.S.A. and for Storglaciären, Kebnekaise, Sweden, and are applied to the terminus records of these two glaciers. The results indicate that the figures for the retreat during individual years are not sufficiently reliable to deduce annual net budget changes with confidence. But the coarser features of the terminus records give net budget figures that agree well with the 9 yr. or 10 yr. means of recent observations. The theory extends the generalized budget record of Storglaciären back by several decades. and shows an increase in net budget in the 1930’s that is not immediately apparent in the terminus record.

Internal reflections are due to changes in electrical admittance between adjacent depositional layers. Reflection coefficients are given for discontinuous changes in either the permittivity or loss tangent. The observed strengths of internal echoes rule out the possibility that they are caused by isolated layers containing “foreign” material, but suggest instead that they are due to systematic fluctuations of density, anisotropy, or loss tangent. The electrical behaviour of ice from polar ice sheets is reviewed and compared with that of ice grown in controlled laboratory conditions. We suggest that the impurity distribution in polycrystalline ice is dependent on the impurity content and the temperature of freezing, and the conductivity is essentially determined by the intrinsic and impurity defects within the crystal lattice. In a polar ice sheet, density fluctuations decrease with depth, whereas loss tangents (and hence their fluctuations) increase since the ice becomes warmer towards bedrock. Echo strengths in central Antarctica are compared with those calculated for a boundary where either all bubbles disappear or the loss tangent changes by 50%. Assuming a constant layering geometry to 2 700 m depth, density fluctuations account for echoes above 1 500 m, but deeper echoes are best explained by variations in the ice conductivity.

To fix the limits of different hazards in the avalanche-hazard maps one uses criteria pertaining to avalanche dynamics. These criteria are at present the velocity and the run-out distance of a given avalanche for a given place. In 1955 A. Voellmy published his theory of avalanche dynamics which has widely been used in practical map preparation. Since 1962 his equations have also been used by the Eidg. Institut für Schnee- und Lawinenforschung (EISLF) to calculate avalanche pressures and run-out distances. Furthermore B. Salm (EISLF) developed another equation for the calculation of run-out distances in 1978. Both the equations of Voellmy and of Salm contain two friction coefficients, µ and ξ. Little is known about them and opinions, even among specialists, differ on what values should be given to them.

This paper presents field observations on very long run-out distances. These observations are used to calculate values for pairs of µ and ξ. For avalanche zoning, only extreme values are of interest, i.e. very low values for µ and very high values for ξ. For the calibration of those coefficients, ten avalanches from the winters 1915–16, 1967–68, 1974–75, and 1977–78 have been used. Those avalanches occurred during heavy and intense snowfalls. For those avalanches, the pair µ= 0.155, ξ = 1 120 m/s2 was found for the Voellmy equation and the pair µ = 0.157, ξ =1 067 m/s2 for the Salm equation. These values only partially agree with those used up to date by EISLF. It is recommended for example that for extreme flowing avalanches (newly fallen snow, soft slabs) the pair µ = 0.16,ξ = 1 360 m/s2 be used.

Variations in annual maximum and accumulated snow depths, snow-cover duration, precipitation and air temperature have been analyzed using daily snow depth, monthly air temperature and monthly precipitation data from 1960 to 1999 from six meteorological stations in the source regions of the Yangtze and Yellow Rivers in China. Annual maximum snow depth, snow-cover duration and precipitation increased by ~0.23, ~0.06 and ~0.05% a–1, respectively, during the study period, while annual accumulated snow depth increased by ~2.4% a–1. Annual mean air temperature increased by ~0.6°C over the study period. An unusually heavy snow cover in 1985 coincided with historically low air temperatures. Data from Tuotuohe and Qingshuihe meteorological stations are used to examine inter-station variability. The annual maximum and accumulated snow depths increased by ~0.35 and ~10.6% a–1 at Tuotuohe, and by ~0.42 and ~2.3%a–1 at Qingshuihe. However, from the late 1980s until 1999 the climate in the study region has become warmer and drier. The precipitation decrease in the 1990s (and not the rapid rise in measured temperature) is thought to be the primary cause of the decrease in snow depth in those years.