No abstract

A short‐term, small‐scale ice motion model was developed in order to facilitate the comparison of various conceptual and numerical formulations. This ice motion model is comprised of two basic components: a momentum balance or equation of motion component and a thickness distribution component. The equation of motion component includes the standard terms relating to air stress, water stress, Coriolis force, sea surface tilt, and internal ice stress. In order to describe the internal ice stress, three different constitutive laws were implemented: a standard linear viscous constitutive law, a closed‐form viscous plastic constitutive law, and a piecewise linearized viscous plastic constitutive law. For the thickness distribution component, a choice of either a two‐level or a multicategory model may be used to characterize the spatial and temporal changes in the thickness distribution of the ice cover. The governing equations for both of these basic ice motion model components are presented. The finite element method based on the Galerkin method of weighted residuals is used to spatially integrate the resulting set of nonlinear partial differential equations while a standard recurrence relation is employed for the temporal integration. The selection of an Eulerian or Lagrangian description for both the equations of motion and the thickness distribution equations is also included in the model development. In order to investigate the response of the developed ice motion model and its various options, the model was applied to a 4‐day ice motion episode in the southern Beaufort Sea. The response of the ice motion model was quantified in terms of the magnitude and the spread of the difference between calculated and observed ice displacements. Comparison between the various model options is included and their effect on the overall model performance is discussed.

Tests in the field and full-scale experience with arctic structures show that the crushing of ice is accompanied by large fluctuations in load. Field experiments show that, in addition to variations of load in time, significant spatial variations across the contact surface also occur. The deformation is observed to take place in a thin layer of damaged ice, which appears near the structure or indenter surface. It is important to model the deformation and strength of ice in this zone. Various aspects of modelling are discussed in the paper, in particular, measures of damage and the relation to the deformation of ice. The relevance of various components of deformation (elastic, viscous, delayed elastic) is outlined, and two mathematical formulations for the deformation are discussed. The behaviour was investigated by a series of tests at constant strain rate as well as tests in which the strain response to stress of damaged and undamaged ice was measured. The creep rate in damaged ice is shown to be significantly enhanced, even for short-term loading. Comparisons of theory and experiment are given for constant strain-rate tests. The models have been calibrated to the experimental data described in the paper. It is a matter for future research to generalize the models to all damage levels and stress states.Des essais en nature et I'expCrience grandeur nature acquise avec les structures arctiques dimontrent que le broyage de la glace est accompagnC de fortes fluctuations de charge. Des expCriences sur le terrain montrent qu'en plus de variations de charge en fonction du temps, des variations spatiales significatives se produisent le long de la surface de contact. L'on observe que la dCformation se produit dans une couche mince de glace endommagke qui apparait prks de la structure ou de la surface de l'entaille. I1 est important de modCliser la deformation et la resistance de la glace dans cette zone. DiffCrents aspects de la modClisation sont discutCs dans cet article, et en particulier les mesures des dommages et la relation a la dCformation de la glace. La pertinence de diverses composantes de dkformation (Clastique, visqueuse, Clastique retardCe) est mise en Cvidence et deux formulations mathCmatiques pour la dCformation sont discutCes. Le comportement a Ct C CtudiC par une sCrie d'essais a vitesse de dkformation constante de mCme que des essais dans lesquels la rCaction de la dkformation a la contrainte de la glace endommagtie et intacte Ctait mesurCe. La vitesse de fluage dans la glace endommagie apparait Ctre accrue de facon significative, mCme pour un chargement a court terme.Des comparaisons entre la thCorie et llexpCrience sont donnCes pour des essais a vitesse de dkformation constante. Les modkles ont Ct C CtalonnCs avec les donnCes expCrimentales dCcrites dans l'article. La gCnCralisation des modkles a tous les niveaux de dommage et d'Ctats de contrainte devrait faire I'objet de futures recherches.

Wave-induced ice motions measured during the Labrador Ice Margin Experiment (LIMEX '89) IntroductionThe ice floes encountered during the Labrador Ice Margin Experiment (LIMEX) in March and April 1989 were small by most standards; most floes were less than 15 m across and were approximately 1 m thick. Collisions between the floes are of interest because they influence the lateral deterioration of the floes and contribute to wave attenuation within the ice-pack, particularly when there is continuous contact between adjacent floes within each wave cycle. On a broader scale, collisions between floes have a role in determining ambient noise levels beneath the ice and in the drift of the pack-ice, but neither of these is pursued in the present work. The wave-induced motions and the contact between floes are also of importance, for example, in predicting the dispersal of an oil slick in this environment. Several features of the marginal ice zone environment are expected to influence the nature and frequency of collisions between floes. The present analysis concentrates on the characteristics of the wave field, ambient temperature and local winds. A characteristic feature of small floes is that they tend to follow the motion of the sea surface and are therefore prone to collide on a regular basis according to the wave cycle, in contrast to floes that are larger relative to the wavelength, which have more damped motions.In McKenna and Crocker (1990), the influence of wave amplitude on the relative speed between floes at the point of contact was investigated. Only direct collisions in the direction of wave propagation were considered. The contact was assumed to be inelastic and ceased when the floes drifted apart. Such a situation is likely to cause the largest collision energies for small floes relative to the wavelength (e.g. 10-m floes with a 10-s wave period), but contact with other neighbouring floes is probable when floes are not circular or are offset. Furthermore, collisions are more likely when different wave frequencies and directions are superimposed. For several periods during the LIMEX '89 experiment, the air temperature dropped well below freezing and ice was formed between the floes, occasionally filling the entire space. This meant that there was continuous energy transfer between floes, with each wave cycle, until the ice melted with the heat of the day.Rottier (1991) developed a model for predicting collision frequencies under various wave conditions and floe dimensions in an effort to simulate floe collisions in marginal ice zones. Based on observations in Fram Strait and the Barents Sea, three types of collisions were characterized: (i) a collision that is caused by a direct hit, (ii) compression of broken ice pieces or newly grown ice in the interstitial spaces, and (iii) shear between adjacent floes transverse to the direction of wave propagation. The more irregular the floe sizes and shapes are, the more likely that type (iii) interactions will occur. Collisions may not fall uniquely into...

No abstract