METHODS

ICE-WEDGE POLYGONS DISTRIBUTION
Ice-wedge polygons zones were identified using Digital Globe high-resolution (0.5 m) satellite imagery. Location of polygons was related to surficial deposits, elevation and slope.

FIELDWORK
3 study sites were visited in June 2019. Very high-resolution UAV imagery was captured. 1 m-deep soil cores were extracted.

LABORATORY ANALYSES
Soil cores were cut in 5 cm sections and analyzed for particle size, gravimetric water content, organic matter and carbon. 

POLYGONS MORPHOMETRY
 A semi-automated method based on the principle of watershed segmentation, where the ice-wedge troughs are treated as high elevations (i.e. mountains) and the polygon centres as low elevations (i.e. valleys), was applied to the UAV imagery to delineate polygons geometry.

ICE-WEDGE POLYGONS DISTRIBUTION

Ice-wedge polygons preferentially form in woody sedge peat (WSP) and glaciofluvial deposits (IGF, WGF), at elevations ranges between 985 and 1045 m and between 1065 and 1135 m, and on hillslopes of <1°.

Normalizing the area occupied by polygons by the area occupied by each surficial deposit indicate that the distribution of surficial deposits in the TTP does not determine the probability of ice wedge occurrence.

Similarly, the normalized polygons frequency of elevations and slopes shows that the frequency distribution of ice-wedge polygons is not a function of the frequency distribution of landscape parameters and that the distribution of these parameters does not determine the probability of ice wedge occurrence.

Accordingly, polygonal terrain are mainly located in the northern part of the two main river valleys (Blackstone and East Blackstone rivers).

MORPHOMETRY OF POLYGONS

The polygons morphology at the 3 study sites is rather subdued, with a subtle centre-trough relief but no evident ridges on either side of the troughs, except for the low-centered polygons of site KM93.

The troughs depth vary between 5 and 25 cm.

Mean area of polygons for all three sites vary between 119.5 and 132 m-2 and the perimeter between 43.2 and 46.9 m. The largest polygons are at site KM95 and the smallest polygons are at site KM92.

The general slope of each patch vary between 0.8 and 5.8% and is steepest at KM92.

The intersections between polygons form a net with a combination of orthogonal (i.e. right angle) and hexagonal (120°) angles. KM92 and KM95 both have a majority of orthogonal intersections, whereas KM93 has a majority of hexagonal intersections. The predominance of orthogonal angles is described as a manifestation of the succession of primary and secondary cracks which further divide the polygon net (Burn and O'Neill, 2015; French, 2017).

PHYSICAL PARAMETERS OF SOIL CORES

Site KM92 is very ice-rich, with occurrence of ice lenses in the polygons centre and trough above the ice wedge, which start at ca. 70 cm below the surface. It is located in glacial deposits which are characterized as medium to coarse silts. The active layer is located at 35 cm below the surface and is followed by a very ice-rich layer, with occurrence of ice lenses in the polygons centre and in the trough above the ice wedge.

Site KM93 is very organic (organic matter content <88%) and has the highest gravimetric water content, although no ice lenses were observed. The ice wedge starts at 30 cm below the surface. The thaw layer in June 2019 was shallower in the trough (10 cm) compared to the low center (25 cm). The ice wedge starts at 40 cm below the surface, and is topped with a 15 cm thick ice-rich layer.

Site KM95 is located in alluvial deposits also characterized as medium to coarse silts. It has a very high carbonate content with concentrations up to 9% in the centre and 5% in the trough, compared to maximums of less than 3% at sites KM92 and KM93. Coring was done up to 105 cm below surface without sampling an ice wedge. Only ice lenses were observed between ca. 45 and 80 cm below surface in the trough.

ICE-WEDGE POLYGONS DISTRIBUTION

Analysis of the distribution of ice-wedge polygons indicate that they preferentially form in woody sedge peat and glaciofluvial deposits, at elevations ranges between 985 and 1045 m and between 1065 and 1135 m, and on hillslopes of <1°.

Theoretically, IW polygons are usually smaller and more abundant in poorly-drained, fine material lowlands, and larger and less abundant in well-drained, coarse material uplands and slopes, mainly because the coefficient of thermal contraction is lower in better-drained soils. Here, we find a similar signature regarding the material, elevation and slope. Polygons occurrence is predominant in glaciofluvial deposits and woody sedge peat, which both have the potential to have a low hydraulic conductivity , and in lower elevations valleys with gentle slopes (<1°), which are by definition poorly-drained compared to uplands and steeper slopes. 

RELATIONSHIP BETWEEN MORPHOMETRY, PHYSICAL PROPERTIES AND STATE

The morphometry of polygons and the physical properties of soil at site KM93 indicate that ice wedges are young and still developing. This evident trough the presence of ridges on either side of the narrow troughs, the abundance of hexagonal intersections, the inundated low centers, the presence of a 15 cm thick ice-rich intermediate layer above the ice-wedge and the presence of dwarf shrubs in both the polygon centers and troughs.

The ice wedges at site KM95 and KM92 show signs of having degraded and stabilized with permafrost re-aggrading. At site KM95, this is evident trough failure from sampling an ice wedge despite coring to a depth of 105 cm, the presence of ice lenses which form an ice-rich intermediate layer, a high carbonate content in the upper portion of the depths that characterize the ice lenses, and the predominance of water-tolerant vegetation in the troughs, some of which are still very wet, especially at polygons intersections. At site KM92, this is evident trough the presence of a 50 cm thick ice-rich intermediate layer containing ice lenses, abundance of water-tolerant vegetation and signs of ice wedge degradation in the 2009 imagery (see "Study Sites" section).