Search

Seasonal or degradational thaw subsidence of permafrost terrain affects the landscape, hydrology, and sustainability of permafrost as an engineering substrate. We perform permafrost thaw sensitivity prediction via supervised classification of a feature set consisting of geological, topographic, and multispectral variables over continuous permafrost near Rankin Inlet, Nunavut, Canada. We build a reference classification of thaw sensitivity using process-based categorization of seasonal subsidence as measured from differential interferometric synthetic aperture radar whereby categories of thaw sensitivity are reflective of ground ice conditions. Classification is performed using a neural network trained on both dispersed and parcel-based reference data. For Low, Medium, High, and Very High thaw sensitivity categories, generalized classification accuracy is 70.8% for 20.6 [km.sup.2] of dispersed training data. In all cases, the majority classes of Low and Medium thaw sensitivity are predicted with higher accuracy and more certainty, while the minority classes of High and Very High thaw sensitivity are underpredicted. Minority classes can be combined to improve accuracy at the expense of a reduced level of discrimination. The two-class problem can be classified with an accuracy of 81.8%, thereby effectively distinguishing between stable and unstable ground. The method is applicable to similar Low-Arctic permafrost terrain with geological and topographical controls on thaw sensitivity. However, generalized accuracy is reduced for parcel-based training, indicating that reference samples are not totally representative for inference beyond the parcel, and any deployment of the network to other geographical regions would benefit from full or partial retraining with local data. Key words: permafrost, thaw sensitivity, DInSAR, supervised classification. La subsidence associée au dégel saisonnier ou à la dégradation du pergélisol a une incidence sur le paysage, l'hydrologie et la pérennité du pergélisol comme substrat d'ingénierie. Nous prédisons la sensibilité au dégel du pergélisol par l'entremise d'une classification supervisée d'un ensemble de caractéristiques comprenant des variables géologiques, topographiques et multispectrales sur du pergélisol continu près de Rankin Inlet (Nunavut, Canada). Nous établissons une classification de référence de la sensibilité au dégel en utilisant une catégorisation basée sur les processus de la subsidence saisonnière mesurée par interférométrie radar différentielle à synthèse d'ouverture, dans laquelle les catégories de sensibilité au dégel reflètent les conditions de la glace de sol. La classification est réalisée à l'aide d'un réseau neuronal entraîné avec des données de référence dispersées et parcellaires. Pour les catégories de sensibilité au dégel faible, intermédiaire, forte et très forte, l'exactitude de la classification généralisée est de 70,8 % pour 20,6 [km.sup.2] de données d'entraînement dispersées. Dans tous les cas, les classes majoritaires de sensibilité au dégel intermédiaire et faible sont prédites avec une plus grande exactitude et plus de certitude, alors que les classes minoritaires de sensibilité au dégel forte et très forte sont sousprédites. Les classes minoritaires peuvent être combinées pour améliorer l'exactitude au détriment du degré de discrimination, qui s'en trouve réduit. Le problème des deux classes peut être classé avec une exactitude de 81,8 %, ce qui permet donc, dans les faits, de distinguer sols stables et instables. La méthode peut s'appliquer à d'autres régions de pergélisol du bas Arctique où la géologie et la topographie modulent la sensibilité au dégel. L'exactitude généralisée est toutefois réduite pour l'entraînement parcellaire, ce qui indique que les échantillons de référence ne sont pas complètement représentatifs pour l'inférence au-delà de la parcelle et qu'un entraînement complet ou partiel avec des données locales serait pertinent pour tout déploiement du réseau dans d'autres régions. [Traduit par la Rédaction] Mots-clés : pergélisol, sensibilité au dégel, DInSAR, classification supervisée., Introduction Ground ice content is an important property of permafrost terrain that influences the thermal response of permafrost to warming (Anderson and Morgenstern 1973; Westermann et al. 2016). Thawing of [...]

Coupled thermo-hydro-mechanical finite element (FE) modelling of thaw consolidation is presented. One-dimensional FE analyses are performed for thaw consolidation of a soil column due to self-weight and with a combination of self-weight and surcharge, with the linear and nonlinear void ratio - effective stress - hydraulic conductivity relationships of thawed soil. The nonlinear behaviour of thawed soil is modelled using a modified Drucker-Prager cap model, while the hydraulic conductivity is varied with the void ratio. Finally, two-dimensional FE modelling of thaw consolidation around a warm pipeline buried in permafrost is performed. The rapid reduction of the void ratio with consolidation, especially at the low-stress level, results in a wide variation of hydraulic conductivity within the thawed zone. The significantly large hydraulic conductivity of soil elements along the curved thaw front, as compared to that of thaw consolidated soil, causes the flow of water along the thaw front, instead of a vertical flow, as assumed in previous 1D thaw consolidation modelling of buried pipelines. Key words: thaw consolidation, finite element modelling, warm pipeline, permafrost, nonlinear thawed soil behaviour. Cette etude presente une modelisation thermo-hydro-mecanique par elements finis (FE) de la consolidation par degel. La consolidation par degel d'une colonne de sol due au poids propre et a une combinaison de poids propre et de surcharge, avec les relations lineaires et non lineaires entre le taux de vide, la contrainte effective et la conductivite hydraulique du sol degele, fait l'objet d'analyses FE unidimensionnelles. Le comportement non lineaire du sol degele est modelise par un modele modifie de Drucker-Prager cap, alors que la conductivite hydraulique varie avec le taux de vide. La modelisation par elements finis bidimensionnelle de la consolidation du degel entourant un pipeline chaud enterre dans le pergelisol est enfin realisee. La reduction rapide du taux de vide avec la consolidation, en particulier au niveau des faibles contraintes, entraine une grande variation de la conductivite hydraulique dans la zone degelee. La conductivite hydraulique significativement plus importante des elements du sol le long du front de degel incurve, comparee a celle du sol consolide par le degel, entraine l'ecoulement de l'eau le long du front de degel, au lieu d'un ecoulement vertical, tel que suppose dans les precedentes modelisations 1D de consolidation par le degel de pipelines enterres. [Traduit par la Redaction] Mots-cles: consolidation par degel, modelisation par elements finis, canalisation chaude, permafrost, comportement non lineaire du sol degele., Introduction Thawing of permafrost is one of the design issues for some infrastructure in cold regions, including roadway embankments, impoundment reservoirs, fuel storage tanks, building foundations, and warm pipelines. In [...]

To link to full-text access for this article, visit this link: http://dx.doi.org/10.1016/j.earscirev.2008.12.002 Byline: Charles Harris (a), Lukas U. Arenson (b), Hanne H. Christiansen (c), Bernd Etzelmuller (d), Regula Frauenfelder (d), Stephan Gruber (e), Wilfried Haeberli (e), Christian Hauck (f), Martin Holzle (e), Ole Humlum (d), Ketil Isaksen (g), Andreas Kaab (d), Martina A. Kern-Lutschg (a), Michael Lehning (h), Norikazu Matsuoka (i), Julian B. Murton (j), Jeanette Notzli (e), Marcia Phillips (h), Neil Ross (k), Matti Seppala (l), Sarah M. Springman (m), Daniel Vonder Muhll (n) Keywords: European permafrost; climate change; geothermal monitoring; geothermal modelling; geophysics; slope stability; permafost hazards; permafrost engineering Abstract: We present a review of the changing state of European permafrost within a spatial zone that includes the continuous high latitude arctic permafrost of Svalbard and the discontinuous high altitude mountain permafrost of Iceland, Fennoscandia and the Alps. The paper focuses on methodological developments and data collection over the last decade or so, including research associated with the continent-scale network of instrumented permafrost boreholes established between 1998 and 2001 under the European Union PACE project. Data indicate recent warming trends, with greatest warming at higher latitudes. Equally important are the impacts of shorter-term extreme climatic events, most immediately reflected in changes in active layer thickness. A large number of complex variables, including altitude, topography, insolation and snow distribution, determine permafrost temperatures. The development of regionally calibrated empirical-statistical models, and physically based process-oriented models, is described, and it is shown that, though more complex and data dependent, process-oriented approaches are better suited to estimating transient effects of climate change in complex mountain topography. Mapping and characterisation of permafrost depth and distribution requires integrated multiple geophysical approaches and recent advances are discussed. We report on recent research into ground ice formation, including ice segregation within bedrock and vein ice formation within ice wedge systems. The potential impacts of climate change on rock weathering, permafrost creep, landslides, rock falls, debris flows and slow mass movements are also discussed. Recent engineering responses to the potentially damaging effects of climate warming are outlined, and risk assessment strategies to minimise geological hazards are described. We conclude that forecasting changes in hazard occurrence, magnitude and frequency is likely to depend on process-based modelling, demanding improved understanding of geomorphological process-response systems and their impacts on human activity. Author Affiliation: (a) School of Earth and Ocean Sciences, Cardiff University, CF10 3YE, UK (b) Department of Civil and Environmental Engineering, University of Alberta, Edmonton, Alberta, Canada T6G 2W2 (c) Department of Geology, The University Centre in Svalbard, 9171 Longyearbyen, Norway (d) Department of Geosciences, University of Oslo, Blindern, NO-0316 Oslo, Norway (e) Department of Geography, University of Zurich, CH-8057 Zurich, Switzerland (f) Institute for Meteorology and Climate Reasearch, Karlsruhe Institute of Technology, 76131 Karlsruhe, Germany (g) The Norwegian Meteorological Institute, Blindern, 0313 Oslo, Norway (h) WSL,Swiss Federal Institute for Snow and Avalanche Research, SLF Davos, CH-7260 Davos Dorf, Switzerland (i) Graduate School of Life and Environmental Sciences, University of Tsukuba, Ibaraki 305-8572, Japan (j) Department of Geography, University of Sussex, Brighton, BN1 9SJ, UK (k) School of Geosciences, University of Edinburgh, West Mains Road, Edinburgh EH9 3JW, UK (l) Department of Geography, University of Helsinki, Helsinki, FIN-00014, Finland (m) Geotechnical Institute, ETH Zurich, CH-8093 Zurich, Switzerland (n) The Swiss Initiative in Systems Biology, ETH-Zurich, CH-8092 Zurich, Switzerland Article History: Received 13 July 2007; Accepted 5 December 2008

We have developed physically based, diurnal, and annual models for freezing/thawing moist soils subject to annual insolation, radiant heating, and cooling, and sensible and latent heat exchanges with the atmosphere. Both models have the same weather forcing, numerical scheme, and soil constitutive properties. We find that surface temperature differences over a diurnal cycle between the annual and diurnal models are as much as -5 K in March, -7 K in June, -4 K in September, and 5 K in December for 38% (by volume fraction) moist soil. This difference occurs because the annual model includes the history of energy fluxes at the surface of the soil. The annual model is linked to microwave emission models for predictions of temporal radiobrightness signatures. The model predicts a relatively weak decrease in diurnal differences in soil temperature with increased moisture content, but a significant decrease in diurnal differences in radiobrightness. It also exhibits notable perturbations in radiobrightness when soils freeze and thaw. The moisture dependent, day-to-night radiobrightness difference is enhanced by as much as -42 K at 19.35 GHz horizontal polarization for frozen soil if daytime thawing occurs.

A one-dimensional heat flow model is developed to estimate depths of soil freezing and thawing using a daily time step. This physically based model assumes near-saturated soil moisture conditions and simulates freezing under bare soil and sod using only those meteorological variables observed at cooperative network stations (i.e., daily temperature, precipitation, and snow depth). Therefore, using this relatively dense observation network, climatological model-derived frost depth statistics can be developed for stations in regions where soil conditions are typically near saturation during the winter. Soil-freezing estimates should be useful in a variety of applications, especially those concerned with agricultural and engineering design specifications. Based on comparisons with measured soil temperature data, the model successfully captures the seasonal progression, rate, and maximum depth of soil freezing and thawing at several climatically diverse locations. Verification results indicate that model-derived maximum frost penetration values were typically within 0.04 m of those observed over a three-winter period at Ithaca, New York. On a daily basis, frost depth estimates were typically within 0.10 m of the observed Ithaca values. The magnitudes of these errors are unaffected by the presence or absence of snow cover and the depth of soil freezing. However, more reliable estimates tend to occur under bare soil than sod. In the case of bare soil, verification analyses indicate a small but consistent bias toward overestimation of the depth of soiling freezing. For sod, a consistent bias could not be identified. This condition may result from representing different, but unreported, sod conditions at various sites with a single fixed parameter.

In this paper, we develop a numerical solution to equations derived in the first part of this study to simulate the thaw subsidence of an unsaturated frozen soil due to surface heating. Solutions obtained in terms of pore pressure, temperature, degree of unfrozen water content, ice content, and porosity provide the spatial and temporal variations of these quantities in a thawing soil column. A one-dimensional form of governing equations is discretized by employing a Crank-Nicolson finite-difference scheme. These highly nonlinear equations are then solved by an iterative procedure at every time step for predescribed positive surface temperatures that produce thawing. The numerical solutions are obtained for a 1-m silt column. Results show that, above the frozen region, unfrozen water saturation increases due to low hydraulic conductivity, which prevents the melting ice from draining, thus causing accumulation of water above it. It has also been observed that most of the settlement at the surface is due to thaw consolidation rather than settlement due to drainage. The former is completed in seven days of a 12-day simulation period. Furthermore, a sensitivity analysis was carried out to determine the response of the simulation to deviations in the values of certain parameters. It is found that the settlement estimations are very sensitive to the thaw-settlement parameter and phase-composition curve, thus requiring accurate determination of these parameters for a reliable simulation of the thaw-consolidation process.

Construction and maintenance of engineering structures in arctic regions demands an understanding of and the ability to cope with environmental problems produced by permafrost. Where thawing of the permafrost cannot be prevented, the range, amount, and extent of degradation are factors that must be taken into account in the design of engineering structures. Predictions of these factors are facilitated by the use of a mathematical model designed to predict the physical system under consideration. This study addresses the problem by developing a predictive model to describe permafrost thaw subsidence based upon a complete formulation of the problem. This problem is one of multiphase (air, ice, water, and solid) transport and deformation in a porous medium. The conservation of mass equations for liquid water, ice with phase change, and deforming soil solids are developed. Darcy's law is extended to include moisture movement due to thermal gradients. The quasi-static equilibrium equations and stress-strain relations that assume a perfectly elastic solid matrix are also employed. Simultaneous variations of moisture retention and phase-composition curves with temperature and pore pressure are also incorporated in the model. The energy-conservation equation, which includes terms due to viscous dissipation and compression effects, along with the other equations expressed, is found to satisfactorily represent the physical processes occurring in thawing permafrost soils. Approximations to simulate the thawing of a one-dimensional soil column are then developed from the general model.

Numerically-based, two-dimensional computer simulations explain the natural formation of stone stripes on bare arctic and alpine hillsides such as Hawaii's Mauna Kea. The numerical model shows that the puzzling regularity of these stripes arises from the nightly growth of needle ice. This recurrent freezing tends to move only the larger stones, which results in alternating rows of stones and fine-grained material. Soil texture, degree of hillslope and extent of ice growth produce feedback effects that govern the development of stripes.

Computer cellular automata explained the natural formation of remarkably regular stone stripes in arctic and alpine terrain. B.T. Werner and B. Hallet applied the cellular-automaton method, which simplified the physical laws governing the movements of individual grains in a system, to the pattern of stripes found on the slopes of Hawaii's Mauna Kea. The resulting model indicated that the stripes arose when nightly freezing of ice needles separated fine-grain material from coarse-grained. Cellular-automaton is thus a promising technique in the field of geomorphology.