The release of a dry-snow slab avalanche involves brittle fracture. It is therefore essentially a non-linear fracture mechanics problem. Traditional snow-stability evaluation has mainly focused on snow strength measurements. Fracture toughness describes how well a material can withstand failure. The fracture toughness of snow is therefore a key parameter to assess fracture propagation propensity, and hence snows lope stability. Fracture toughness in tension KIc and shear KIIc was determined with notched cantilever-beam experiments in a cold laboratory. Measurements were performed at different temperatures and with different snow types of density ρ = 100–300 kgm–3, corresponding to typical dry-snow slab properties. The fracture toughness in tension KIc was found to be larger (by about a factor of 1.4) than in shear KIIc. Typical values of the fracture toughness were 500–1000 Pam1/2 for the snow types tested. This suggests that snow is one of the most brittle materials known to man. A power-law relation of toughness KIc on relative density was found with an exponent of about 2. The fracture toughness in tension KIc decreased with increasing temperature following an Arrhenius relation below about –8°C with an apparent activation energy of about 0.16 eV. Above –6°C the fracture toughness increased with increasing temperature towards the melting point, i.e. the Arrhenius relation broke down. The key property in dry-snow slab avalanche release, the critical crack size under shear at failure, was estimated to be about 1 m.

A theoretical model is formulated to investigate the influence of random variations in strength of the weak layer on slab avalanche release. An equation for slab stability is derived which accounts for stress redistribution between strong and weak regions. The model is first used to re-derive results on shear fractures (shear bands) in a homogeneous slab. It is then applied to the case of a slab with randomly varying peak strength of the weak layer. It is demonstrated that such variation may have a dramatic knock-down effect on the failure strength of a slope. The nature of the critical flaw is studied and precursors to failure are investigated.

Daily weather measurements, snow stability assessments and recorded weak layers of 23 stations covering an observation area of approximately 40 000 km2 in western Canada were analyzed. The study area includes three major mountain ranges with different snow climates. All stations included assess the stability of the snow cover. However, the focus of the avalanche safety program of the different types of operation (heli-ski operation, ski resorts and parks) varies significantly. The three stations in the Coast Mountains show the highest snow stability, followed by the South Columbia Mountains and then the North Columbia and Rocky Mountains. The weather data were analyzed to try to explain some of these differences. Intensive snowfall at relatively high temperatures proved to be important for the higher snow stability over the season in the Coast Mountains. Theweak-layer data were used to complement the snow stability assessments. Most persistent weak layers were reported in the Columbia Mountains, followed by the three stations in the Coast Mountains and trailed by the Rocky Mountains. Although some weather observations indicate climatic reasons for the smaller number of weak layers in the Rocky Mountains, it cannot be excluded that these differences are also related to the different type of operations.

To complement a technique to detect internal structures of seasonal snow covers and glacier firn with ground-penetrating radar (GPR), we carried out calibration experiments and an observation of winter snow cover (5.7m thick dry snow with numerous ice layers) with an 800 MHz GPR. In particular, we aimed to discriminate periodic noise, which is inherent in GPR, from radar echoes and to obtain a relationship between the observed reflection strength and the magnitude of density contrasts. Experiments were done in air to evaluate noise levels and receiver characteristics of this system. Based on these, we removed noise from radar echoes in the snow-cover observation. We recognized numerous marked echoes in a noise-free radargram. The depths of these echoes coincided roughly with those of large density contrasts observed in the snow pit. Thus, we argue that the echoes correspond to thin ice layers. Furthermore, the minimum density contrasts detected by this GPR are found to vary from about 100 to 250 kgm–3 at 1–6m depth in the seasonal snow cover.

This paper presents a model of snow structure (model of regular grain packing) that is based on experimental determination of various geometrical characteristics of fine-, medium- and large-grained granular snow. Data analysis supports the possibility of approximating the ice-matrix configuration as a regular lattice of nearly spherical ice grains connected by rigid ice bonds. The model was successfully used for relating microstructural parameters of snow to snow density, compaction behavior and evolution of mechanical properties.

In order to study channelled snow flows over rough surfaces, a laboratory-scale experiment was installed at the “Col du Lac Blanc”, a 2800 m high pass in the French Alps, near the Alpe d’Huez ski resort. It consists of a 0.2 mwide, 10 m long channel fed with snow by a motorized hopper. Both the slope of the channel and the feeding rate of the hopper can be modified. Sensors in the channel provide measurements of the velocity profile, the flow height and the shear and normal stresses at the bottom of the flow. Velocity profiles for different slopes are presented in this paper. Results indicate the presence of a highly active layer at the bottom. This layer is mainly responsible for the avalanche velocity, while the upper layer has a much smaller velocity gradient. A first interpretation of both layers is given.

The saltation length of aeolian snow particles and a new parameter, the ejection factor, which expresses the degree of erosion due to drifting snow, were obtained as functions of friction velocity by means of wind-tunnel experiments for semi-hard snow cover. The saturated-snowdrift transport rate was also obtained experimentally as a function of friction velocity. Based on these characteristics and the parameter, the development of the saltation layer of drifting snow along the fetch was simulated under various conditions such as snow hardness, wind speed and snowfall intensity. The main results are as follows. The developing distance denoting the distance required for the saltation layer to attain saturation, X sat, is determined by saltation length, ejection factor and saturated-snowdrift transport rate, all of which depend on wind speed. It is also affected by the magnitude of snowdrift transport rate at the starting point and by the intensity of snowfall if it exists. The dependence of Xsat on wind speed is not simple in the case of semi-hard snow cover: Xsat increases with wind speed under weak to moderate wind conditions and then decreases under moderate to strong wind conditions. It is sensitive to snow hardness: it is about one order longer on hard snow cover than on semi-hard snow cover. Snowfall reduces not only the value of Xsat but also its dependence on snow hardness.

Snow, from its fall until its full melting, undergoes a structural metamorphism that is governed by temperature and humidity fields. Among the many possible mechanisms that contribute to snow metamorphism, those that depend only on curvature are the most accessible to modelling. In this paper, techniques of volume data analysis adapted to the complex geometry of snow are introduced and then applied to experimental tomographic data coming from the isothermal metamorphism of snow near 0°C. In particular, an adaptive algorithm of curvature computation is described. Present results on the evolution of specific surface area and anisotropy already show that such image-analysis methods are relevant tools for the characterization of real snow microstructures. Moreover, the evolution of the curvature distribution with time provides valuable information for the development of sintering models, in the same way as a possible quantitative calibration of snow-grain coarsening laws.

Parameters critical to slushflow release have been studied during 11 winters in Rana, North Norway. Slushflows occurred during 16 periods within the research areas and during another 13 in the district. A noticeable rise of water in snowpack was registered 49 times. Field predictors of acute-hazard and critical water influxes related to different snowpack conditions are identified. Threshold values for acute-hazard prediction based on five main combinations of winter snowpack and current rise of water in snowpack were identified and checked against the registered slushflows, with promising results. Cohesionless new snow and coarse-grained snow are most liable to start flowing. A snowpack reinforced by layers of crust and ice is normally the most stable. An automatic system for supervising slushflow hazard based on the defined criteria is presented. The system can be tailored for any location and problem owner.

The new approaches in absorption X-ray microtomography allow snow-volume image acquisition in the cm3 range without destroying snow structure, and make it possible to perform micromechanical studies with the real geometry. The main objective of this paper is to introduce the development of a new three-dimensional (3-D) geometry modeling and finite-element analysis simulation package adapted to these scales. These new code modules and procedures are briefly described using an isothermal metamorphism snow experiment (–2°C, 3 months).This sample of aged snow (3.375mm3; 3003 voxels) allows simulation of a simple hypothetical uniaxial compressive stress experiment. The constitutive equations, boundary conditions, basic assumptions and the result showing the stress field over the 3-D data are discussed. The first qualitative results show a maximum stress of 6–9MPa in some small bonds, showing the potential of these codes to simulate the micromechanical behavior of complex materials.

A new experimental software for blowing snow, SYTRON2, is presented. It takes advantage of the previous limited realization SYTRON1 and of the entire real-time operational environment of the automatic chain SAFRAN–Crocus–MÉPRA (SCM). It runs presently on a rectangular grid, with a mesh size of about 45 m, covering an area of 3.0 × 3.0km2 around the observation test site Col du Lac Blanc, French Alps (2700m a.s.l., near the Alpe d’Huez ski resort). SYTRON2 is initialized with a realistic snowpack representation derived from the SCM outputs and coupled to a wind field computed by the new SAMVER model. The final aim is to build a fully embedded tool, suitable for representing snowdrift effects with a better resolution at different specific locations. The different tools and operators used are described, as well as the field validation processes which are based on remote identification of the eroded and accumulated areas using digital pictures.

In wind transport of snow, horizontal momentum is extracted from the mean wind flow and transferred to the snow grains. Upon colliding with the surface the grains can bounce and eject further grains in a process known as splashing. How efficiently the horizontal momentum is converted to vertical momentum in the splash process is the determining factor for mass-transport rates. This paper discusses wind-tunnel experiments performed to calculate the splash function for snow particles. The data are used to develop a new splash function. Particular care is taken to include correlations in the data such as between ejection velocity and ejection angle. The new splash function includes these correlations, and its parameters are related to physical properties of the bed and snow.

This paper focuses on the surface mass balance of a horizontally homogeneous snowfield, with emphasis on the effects of snowdrift sublimation. A one-dimensional model of the atmospheric boundary layer that includes snowdrift physics and thermodynamics is used. In sufficiently strong winds, snow particles are eroded from the surface. Once airborne, they are susceptible to sublimation. Averaged over longer time periods, the net erosive flux equals sublimation of snowdrift. However, model results show that there is no such balance in the course of a snowstorm event. They also indicate that snowdrift sublimation tends to enhance net erosion, but the increase occurs more slowly than the mass transfer by snowdrift sublimation, and the maximum is smaller. This difference in temporal behaviour influences the average erosion rate owing to non-linear interactions between snowdrift sublimation, drift density and erosion. Since the increase in relative humidity due to snowdrift diminishes surface sublimation, the average change in total ablation induced by snowdrift sublimation remains small. Observations made during snowdrift episodes in Antarctica agree qualitatively with some of the model results.

This paper examines the positive and negative aspects of a range of interpretations of nearest-neighbours models. Measures-oriented and distribution-oriented verification methods are applied to categorial, probabilistic and descriptive interpretations of nearest neighbours used operationally in avalanche forecasting in Scotland and Switzerland. The dependence of skill and accuracy measures on base rate is illustrated. The purpose of the forecast and the definition of events are important variables in determining the quality of the forecast. A discussion of the application of different interpretations in operational avalanche forecasting is presented.

Engineers require impact scenarios when developing hazard mitigation strategies to protect structures against snow avalanches. Since direct measurements of avalanche impacts on large obstacles are rare, the documentation and post-event analysis of avalanche damages is essential to understand the interaction of avalanches with obstacles. The objective of this paper is to develop hazard scenarios for avalanche actions on bridges, based on a case study of a well-documented avalanche event. The 40 m long pre-stressed road bridge Ri di Rialp in the Swiss Alps collapsed after being struck by a dense-flow avalanche in 1998. The post-event analysis shows that the controlling hazard scenario was an inclined avalanche impact, not a horizontal impact as one would assume given the topographical situation. Using a failure analysis, an impact angle of 40° and a minimal impact pressure of 172 kN m−2 were found. Finally the insights are summarized so that engineers can apply the results when designing a bridge at risk from avalanches.

It has been postulated that heavy summer snowfalls have a large impact on the mass balance of mid-latitude glaciers, because they simultaneously add mass to the glacier and reduce the amount of absorbed solar radiation. An automatic weather station (AWS) on the snout of Morteratschgletscher, Switzerland, registered a large summer snowfall event on 10–11 July 2000. Sonic rangers recorded about 20 cm of new snow on the snout and about 50 cm near the equilibrium line. We have used data from the AWS to study the impact on the melt process. The data show that in the ablation zone of the glacier the snow has melted and the effect on the albedo has disappeared after about 5 days. The suppression of the melt by the high albedo of the fresh snow is an important effect. For the ablation zone we find a feedback factor of about four, that is, the total effect of the snowfall event on the annual specific balance is about four times the amount of mass added during the event. We have also used a mass-balance model with 25 m spatial resolution to assess the impact of the snowfall on the net balance of the entire glacier. We find the strongest effects just below the equilibrium line. Averaged over the glacier, the amount of snow deposited was 224 mm w.e. The calculated effect on the total mass balance of the glacier is 354 mm w.e.

Snow-courses data have been collected in order to investigate the temporal variability of snow distribution in two catchments in southern Norway during the 2002 melt season. The profiles represent different elevations, aspects and terrain types. At snow maximum the spatial distribution of snow above the tree line was positively skewed (long tail in the positive direction), whereas the spatial distribution below the tree line followed a more normal distribution. During the snowmelt season the spatial distribution of snow became increasingly skewed. By separating the datasets into two terrain classes, alpine and forest, the snow distribution could be described by a time-variant gamma distribution function, one for each terrain class. The results of the study will be used to formulate a new snow routine in the Swedish rainfall–runoff model HBV, which is used for flood forecasting in Norway.

Historical snow-depth observations in the former Soviet Union (FSU) during the period September 1960–August 1984 have been analyzed in order to understand the seasonal cycle of snow coverage in the FSU. Snow cover first appears in September in northeastern regions, and spreads over the entire territory before early January. Snowmelt begins in mid-January in the southern regions and then snow cover retreats rapidly northward until it disappears completely before late June. Northward of 60°N, the land surface is snow-covered for more than half the year. The longest snow-cover duration is observed on the central Siberian plateau (about 9.5 months) and along the Arctic coastal regions (about 8.5 months). One of the most conspicuous features of the snow coverage in the FSU is that the length of the snow-accumulation period differs considerably from region to region (2–7 months), while the length of the snowmelt period is rather short and uniform over almost the entire territory (1–2 months). Although the maximum snow depths are 20–50 cm in most regions of the FSU, they exceed 80 cm in the mountainous regions in central Siberia, Kamchatka peninsula, and along theYenisei river valley. Values for the maximum snow depth are very small along the Lena river valley in spite of the air temperature being extremely low in winter. By calculating correlation coefficients between the snowfall intensities and the sea-level pressures or 500 hPa heights, it is shown that deep snow along the Yenisei river valley is caused by frequent migration of synoptic disturbances from the Arctic Ocean. Snowfalls along the Lena river valley are also caused by traveling disturbances from the Arctic Ocean. Snow accumulation is suppressed after the Arctic Ocean has been frozen.

This paper addresses the problem of defining a proper method for formal risk analysis in avalanche-prone areas. In this study, risk is defined as the annual probability of being killed by an avalanche for someone living or working permanently in a building under a hazardous hillside. A new methodology to estimate the hazard component of avalanche risk based on the use of dynamic models is introduced. This approach seems to have some advantages over the current methods based on statistical analysis of historic avalanche data. The vulnerability component of risk is formulated as a function of avalanche velocity, according to previous formulations. However, given the lack of knowledge on how avalanche impact damages structures and causes fatalities, the effect on the resulting risk mapping of using different vulnerability relations is explored. The potential of the proposed approach for evaluating the residual risk after the implementation of defensive structural work is discussed.