The formation of segregated ice is of fundamental importance to a broad range of permafrost and periglacial features and phenomena. Models have been developed to account for the microscopic interactions that drive water migration, and predict key macroscopic characteristics of ice lenses, such as their spacings and thicknesses. For a given set of sediment properties, the temperature difference between the growing and incipient lenses is shown here to depend primarily on the ratio between the effective stress and the temperature deviation from bulk melting at the farthest extent of pore ice. This suggests that observed spacing between ice lenses in frozen soils, or traces of lenses in soils that once contained segregated ice, might be used to constrain the combinations of effective stress and temperature gradient that were present near the time and location at which the lower lens in each pair was initiated. The thickness of each lens has the potential to contain even more information since it depends additionally on the rate of temperature change and the permeability of the sediment at the onset of freezing. However, these complicating factors make it more difficult to interpret thickness data in terms of current or former soil conditions.

A. L. Washburn famously proposed and reviewed 19 hypotheses for the origin of patterned ground in periglacial environments over 50 years ago (Washburn, 1956). Of these 19 mechanisms, only a few have been analyzed starting from a fundamental description of the physics to assess their potential contribution to the initiation of patterned ground. Here, differential frost heave (DFH) is investigated as the origin of non-sorted circles in periglacial landscapes. Model results indicating that DFH can spontaneously lead to patterned ground are compared to measurements of patterned ground in the Canadian Arctic Archipelago. The characteristic size of the predicted emerging pattern depends strongly on the maximum depth of freezing but is only weakly dependent on the soil composition. The predicted emerging patterns may also be dramatically different in size with a small change in active layer when a relatively thin (~ 10 cm) insulating snow covers the ground during freezing. The model predicted trends agree with field observations of pattern size and active layer depth at two distinct sites. Although two data points cannot conclusively indicate a trend, this correlation gives support for the possibility of determining some subsurface properties using remote sensing images of surface patterned ground.

Slope hummocks, a type of nonsorted patterned ground, are composed of stratified, organic, silty sand, and develop through the interaction of niveo-eolian deposition, solifluction, slopewash, and vegetation growth. Fields of hummocks show consistent patterns: forms on convex slopes increase in height downslope until the channel is reached, whereas those on convexo-concave slopes increase on the upper convexity but are buried by niveo-eolian deposition downslope of the snowbank remnant. These trends can be reproduced using a simple numerical model based on measured slope and snow depth profiles, sediment concentrations in the snow and solifluction rates. The model indicates that hummocks transit slopes of 20–40 m in about 2–4 ka, a time-frame that is plausible given site emergence, measured rates of solifluction, and published dates for organic horizons within hummocks on northern Ellesmere Island. Sensitivity analyses show that long-term effect of climate warming on hummock heights may differ depending on whether it is accompanied by precipitation increase or decrease. The required combination of two-sided freezing to promote plug-like movement, incomplete vegetation cover and thin snow that enable eolian erosion during winter and spring, and vegetation growth in snow-bed sites to stabilize niveo-eolian deposits may explain why these forms are important regionally but apparently are not present throughout the Arctic.

Polygonal patterned ground in polar regions of both Earth and Mars has received considerable attention. In comparison with the size, shape, and arrangement of the polygons, the diverse micro-relief and topography (termed here simply “relief”) of polygonal patterned ground have been understudied. And yet, the relief reflects important conditions and processes occurring directly below the ground surface, and it can be observed readily in the field and through remote sensing. Herein, we describe the relief characteristic of the simplest and relatively young form of patterned ground in the Dry Valleys of Antarctic. We also develop a numerical model to examine the generation of relief due to subsurface material being shouldered aside contraction cracks by incremental sand wedges growth, and to down-slope creep of loose granular material on the surface. We model the longterm subsurface deformation of ice-cemented permafrost as a non-linear viscous material. Our modeling is guided and validated using decades of field measurements of surface displacements of the permafrost and relief. This work has implications for assessing the activity of surfaces on Earth and Mars, and much larger scale potential manifestations of incremental wedging in icy material, namely the distinct paired ridges on Europa.

One of the distinguished efforts of A.L. Washburn was to reconstruct mean annual air temperature using periglacial features as climate indicators. This paper reviews existing periglacial indicators and proposes a strategy to improve their thermal resolution based on recent periglacial process studies, with a focus on solifluction and thermal contraction cracking and associated landforms/structures. Landforms resulting from solifluction reflect both the depth subjected to freeze–thaw and the thickness of frost-susceptible soils. The thickness of a solifluction structure can be used to infer the dominant freeze–thaw regime and minimum seasonal frost depth. Ice-wedge pseudomorphs have limited potential as a climate indicator because (1) they mainly reflect extreme winter temperatures, (2) their thermal thresholds depend on the host material, and (3) they need to be distinguished from frost wedges of other origin produced under different thermal and/or material conditions. Monitoring studies of currently active ice wedges suggest that ice-wedge cracking requires a combination of low temperature and large temperature gradients in the frozen active layer. Further field monitoring of periglacial processes and their controlling factors under various climate conditions and in various materials are needed, however, to improve the resolution of periglacial paleoclimate indicators.

This review presents a synthesis of four decades of palsa studies based on field experiments and observations mainly in Fennoscandia, as well as laboratory measurements. Palsas are peat-covered mounds with a permanently frozen core; in Finnish Lapland, they range from 0.5 to 7 m in height and from 2 to 150 m in diameter. These small landforms are characteristic of the southern margin of the discontinuous permafrost zone. Palsa formation requires certain environmental conditions: long-lasting air temperature below 0°C, thin snow cover, and low summer precipitation. The development and persistence of their frozen core is sensitive to the physical properties of peat. The thermal conductivity of wet and frozen peat is high, and it decreases significantly as the peat dries and thaws. This affects the development of the active layer and makes its response to climate change complex. The insulating properties of dry peat during hot and dry summers moderate the thawing of the active layer on palsas. In contrast, humid and wet weather during the summer causes deep thawing and may destroy the frozen core of palsas. Ice layers in palsas have previously been interpreted as ice segregation features but because peat is not frost-susceptible, the ice layers are now reinterpreted as resulting from ice growth at the base of a frozen core that is effectively floating in a mire.

The lithalsas in the Hudson Bay region of northern Québec, Canada, are the closest modern analogs of ancient features that collapsed to form conspicuous circular depressions (“viviers”) common in the Hautes-Fagnes, a region in Belgium. Observations made in both regions are complementary and suggest that these mounds formed by frost heaving displacing soil not only upward, as previously assumed, but also laterally. This lateral displacement is consistent with diverse observations and inferences, which include (1) the simple rounded outline, either circular or oval, typical of both active and relic lithalsas; (2) evidence of local lateral extension inferred from exposures of the relic forms; (3) the relative inefficiency of solifluction in accumulating surface material to form the peripheral ramparts of remnant lithalsas due to the very gentle slopes of the mounds; and (4) the dip of ice lenses within a lithalsa in the Hudson Bay region, perhaps indicating that the freezing front dipped outward along its periphery. The growth of segregation ice is the primary driver for the vertical growth and lateral enlargement of a lithalsa.

Recent accounts suggest that periglacial processes are unimportant for large-scale landscape evolution and that true large-scale periglacial landscapes are rare or non-existent. The lack of a large-scale topographical fingerprint due to periglacial processes may be considered of little relevance, as linear process–landscape development relationships rarely can be substantiated. Instead, periglacial landscapes may be classified in terms of specific landform associations. We propose “cryo-conditioning”, defined as the interaction of cryotic surface and subsurface thermal regimes and geomorphic processes, as an overarching concept linking landform and landscape evolution in cold regions. By focusing on the controls on processes, this concept circumvents scaling problems in interpreting long-term landscape evolution derived from short-term processes. It also contributes to an unambiguous conceptualization of periglacial geomorphology. We propose that the development of several key elements in the Norwegian geomorphic landscape can be explained in terms of cryo-conditioning.

Staircases of gravelly river terrace deposits in southern England occupy valleys typically underlain by frost-susceptible and brecciated bedrocks. The valleys developed during the Quaternary by alternating episodes of (1) brecciation, incision and planation through the bedrock, forming wide low-relief erosion surfaces; and (2) aggradation in braidplains of gravel a few meters thick that bury the erosion surfaces. A conceptual model to account for some of the terraces proposes that brecciation resulted from ice segregation in the ice-rich layer in the upper meters of Pleistocene permafrost, making them vulnerable to fluvial thermal erosion and therefore predisposing the bedrock to planation. The low gradients of the valleys were adjusted such that rivers transferred fine materials out of the basins but lacked the competence to remove gravel, which therefore accumulated within floodplains. The model challenges the prevailing view of incision during climate transitions. It attributes incision and planation to very cold and arid permafrost conditions, when rivers had limited discharges and hillslopes supplied limited volumes of stony debris into valley bottoms.