Process Form Relationships Towards a Landsystems Model for the Permafrost Zone

Regional mapping of Quaternary geology by the Geological Survey of Canada has provided a comprehensive overview of the major landform-sediment assemblages associated with the last glacial cycle in northern Canada. This has enabled attempts to reconstruct ice sheet behaviour based upon landform zonation and process-form relationships, although interpretations of certain landform elements and their palaeo-glaciological implications may differ (cf. Hodgson and Vincent, 1984; Dyke, 1987; Sharpe, 1988; Rampton, 1988; Dyke and Savelle, 2000). For example, landform and surface material mapping generally has a high degree of reproducibility but the processes responsible for the landforms that are mapped can be the subject of dispute (e.g. Fyles, 1963; Sharpe, 1988, 1992; Dyke and Savelle, 2000; Rampton, 2001).

7.6.1 Glacially Deformed Permafrost

Glacially thrust and deformed beds of permafrost outcrop in numerous coastal sections along the western Canadian Arctic between the Tuktoyaktuk region and the glacial limit. All are evidently overlain by till and have been described mainly by J.R. Mackay (see summary and references in Mackay, 1971). Internal structures (folds and thrust faults) everywhere accord with glacier flow directions as inferred from up-ice flutes and end-moraine configurations. The largest known transported mass of pre-Wisconsinan permafrost comprises Herschel Island (170 m high and 110 km2), part of the Laurentide terminal moraine just off the Yukon Coast. The island is directly down-ice from Herschel Basin, which is similar to the island in size and is therefore the likely source of the detached permafrost mass. Where large upstanding, deformed permafrost masses like Herschel Island and Nicholson Island form distinct ridges (Rampton 1988), they constitute thrust moraines. Other probable thrust moraines, which are known to contain deformed permafrost, are the so-called fingers of the Eskimo Lakes (see Fig. 62 in Mackay, 1963). The deformed permafrost, comprising a variety of fossiliferous terrestrial and marine sediments and massive ice bodies, is in effect deformation till or glacitectonite (Elson, 1981; Benn and Evans, 1996, 1998).

The widespread occurrences of deformed and displaced permafrost illustrate two important points. First, some of the relict frozen cores of moraines pre-date the last glaciation and are of nonglacial origin except in their deformation and displacement. To date, however, such occurrences are known only from the vicinity of the glacial limit. They have probably not survived elsewhere and could not have survived under warm-based ice. Hypothetically, similar features could have been generated by a readvance across permafrost during deglaciation, provided sufficient time was available for proglacial permafrost aggradation. However, no such occurrences are known from the permafrost zone of Canada. Second, the glacier ice that entrained and (or) deformed the permafrost bodies must have been cold-based from time of entrainment through to deglaciation. The extensive distribution of deformed permafrost in the vicinity of the glacial limit indicates that much (or all) of the marginal fringe was cold based.

7.6.2 The Major Belts of Ice-Cored Terrain: End Moraines Versus Thermokarst Terrain

As northern Canada lies within a zone of continuous permafrost, large bodies of glacier ice are likely to be preserved in the deglaciated landscape wherever melting snouts generated a surface debris cover thicker than the active layer (~1 m thick). The common occurrence of modern debris-covered glacier snouts, along with the fact that thousands of glaciers in this region and in the adjacent Cordillera formed ice-cored moraines during the Neoglacial period, makes it exceedingly unlikely that similar features did not form during Late Wisconsinan deglaciation.

Considerable exposures of ground ice have been observed, encountered in boreholes, or inferred from geophysical and geomorphic data in the lowlands of the western Canadian Arctic. Some ice exposures occur within features that have been mapped as end or hummocky moraine and have been interpreted as the remnants of glacier ice now entombed within the permafrost (e.g. Lorraine and Demeur, 1985; French and Harry, 1988, 1990; Dyke et al, 1992; Sharpe, 1992; St Onge and McMartin, 1995, 1999). This interpretation implies that many morainic landforms are still supraglacial in character and that deglaciation continues albeit episodically today (Worsley, 1999; Dyke and Savelle, 2000). Other exposures of massive ground ice occur in a similar till-mantled terrain with a ridged and hummocky character in the Tuktoyaktuk region east of the Mackenzie Delta. However, these ice bodies have been interpreted as segregation ice (e.g. Mackay, 1971; Rampton, 1974, 1988). A serious interpretive difficulty thus arises in distinguishing kettled end moraines from extensive segregation ice-cored terrain that has later been subjected to thermokarst and is essentially identical morphologically. This is the case where a landform that looks like a moraine apparently is not. If correctly interpreted, these thermokarst terrains might be termed pseudo-moraines. Note that we use 'moraine' here in a morphological sense, rather than as a synonym for till. Rampton's use was the latter or both.

Mackay (1971) and Rampton and Walcott (1974) among others have convincingly shown that massive ice and ice-debris mixtures underlie much of the moraine-like terrain of the western Arctic mainland. Rampton (1974, 1988) hypothesized that most of the ice is not buried glacier ice, as might be expected in true end moraines, but is of segregation origin. Nevertheless, the isotopic composition of the ice (Mackay, 1983) indicates a Wisconsinan glacier ice origin. Rampton inferred, therefore, that segregation occurred by freezing of subglacial meltwater as it encountered the aggrading base of permafrost, located in his model at the ice margin and in the immediate glacier forefield during deglaciation. Thus the glacial isotopic signature was preserved in nonglacial ice (Fig. 7.8). In this model, the subglacial meltwater flowed beneath an existing till layer and generated a 20—30 m thick layer of segregation ice. This uplifted the till layer at or just behind the receding ice margin and thereby separated the till from underlying sediment. It is not evident why the beginning of the permafrost wedge is placed at the ice margin rather than behind it, nor therefore is it evident what might have prevented the subglacial meltwater from simply exiting at or escaping to the surface at the ice front. Similarly, it is not evident why the considerable pore water pressures, sufficient to cause tens of metres of ground uplift, did not rupture the warm, incipient, ice-marginal permafrost and allow water escape. Nevertheless, and taking all elements of the model at face-value, the rate of ice segregation kept pace with the rate of ice-marginal recession, and thus generated a regional sheet of massive ground ice. The massive ice was later pocked by thermokarst, which left a terrain resembling, and perhaps indistinguishable from, ice-cored hummocky and ridged moraine. Rampton mapped this terrain as 'rolling moraine'. His map shows that there is considerable internal ridging within this terrain and that it is bounded by prominent ice-front positions and ice-marginal deposits. The putative thermokarst lakes in the moraine-like region are identical in form to features interpreted as kettles elsewhere (Fig. 7.9).

In effect, Rampton invokes the formation of features that resemble morainal belts as proglacially formed ground-ice injection structures. In a broad sense, ice-cored terrains formed in this way might still be considered to be end moraines in that they formed precisely along former ice margins. It is important to bear in mind, however, that the hummocky and ridged moraine-like features in this model developed long after deglaciation and that no morainal relief will survive removal of the segregation ice core. Only an extensive thermokarst lake plain will remain, and hence the model has no potential application to moraines south of permafrost. Rampton's model has no basis in observations at modern ice margins. For example, regional ground ice sheets are not known to be forming and uplifting ground in front of modern glaciers and the model has not been applied to similar terrain elsewhere in arctic Canada. However, Rampton (2001) has recently argued its explicit application to the morainal belt on Wollaston Peninsula of Victoria Island (see below) in response to the contention by Dyke and Savelle (2000) that these moraines are cored extensively by glacier ice. Its applicability is, therefore, considered below. But first we need to consider the proposed chronology of development of moraine-like topography in the Tuktoyaktuk type area.

Rampton's model for the ridged and hummocky moraine of the Tuktoyaktuk region supports and may be essential to his interpretation of regional glacial history (Fig. 7.8). In that interpretation, the massive, segregated, ground ice sheet was formed during Early Wisconsinan deglaciation (c. 115—64 ka BP). However, the lake basins, and hence the moraine-like topography, developed in Late Wisconsinan time (mainly 13—10 ka BP), as shown by numerous radiocarbon dates on basal lake sediments. Before that time, the massive ice was apparently left undisturbed under the cold climate of intervening Late and Middle Wisconsinan time (64—13 ka BP). However, if the lake basins in the hummocky moraine in the Tuktoyaktuk region are instead kettles, formed by melting of glacier ice as outlined above, the radiocarbon dates on these lake basins indicate that the moraine belt is also of Late Wisconsinan age. Support for Rampton's model in its type area would therefore lie in a demonstration of an Early Wisconsinan age for the till overlying the massive ice. The lack of any convincing evidence that the till is Early Wisconsinan and the fact that correlative or older deposits overlie beds of Middle Wisconsinan age (e.g. Hughes et al., 1981; Morlan et al., 1990; Hill et al. , 1985), have allowed most reviewers of regional glacial history to prefer a Late Wisconsinan age for the till overlying the massive ice in the Tuktoyaktuk region (e.g. Hughes et al., 1981; Denton and Hughes, 1981; Dyke and Prest, 1987; Dyke et al., 2002). Nevertheless, demonstration of a Late Wisconsinan age for the till at or close to the glacial limit, as is the case at Tuktoyaktuk, would not invalidate the general process model advanced by Rampton; it could still be claimed that the ground ice is of proglacial segregation origin and that the lake basins and moraine-like topography formed well after deglaciation, early in Late Wisconsinan time.

However, any application of Rampton's model as an alternative interpretation of younger Late Wisconsinan moraine-like belts (e.g. Rampton, 2001) should take into consideration the finer chronological implications of the model. One should demonstrate, for example, that sufficient time was available to exhaust the latent heat released by freezing meltwater at the freezing front upward through the overlying permafrost during the postulated interval of ice segregation. Rampton did not develop or apply his model quantitatively and hence did not evaluate its chronological implications. Nevertheless, the accumulation of segregation ice as proposed in his model is a problem analogous to that of pingo growth, both processes involving addition of massive ice at the base of aggrading permafrost. If relief of moraine-like areas is entirely due to cores of massive segregation ice and if these areas become flat upon ice removal, as Rampton (1988, 2001) suggests, then ice under the moraine-like ridges (ignoring for the moment the inter-ridge depressions, which presumably originated by thermokarst) must have formed in lock-step with ice recession across the moraine-like belt. This stepwise formation of segregation ice is necessary because the plunging permafrost base extending down from the retreating glacier front (Fig. 7.8) would have prevented groundwater from reaching far into or below the glacier forefield.

The moraine-like belt on Wollaston Peninsula is well behind the Late Wisconsinan glacial limit and hence is of deglacial age. It is as much as 40 km wide (Sharpe, 1992b, Map 1650A) and a typical transect would cross 20 or so moraine-like ridges or very large hummocks of 20—100 m relief (Fig. 7.5). The simplest calculation of the rate of segregated ice formation by basal accretion is provided by Stefan's solution:

where: z is depth of the freezing front t is time b is a constant (ice is 1.4 — incorporating latent heat of fusion, thermal conductivity, and temperature appropriate to the region; Mackay 1971, 1979).

The time taken to form an ice layer by basal accretion thus increases as the square of ice thickness. A 10 m thick layer would form in about 50 years (20 m in ~200 years; 30 m in ~460 years; 50 m in ~1,275 years; 75 m in ~2,870 years; 100 m in ~5,100 years). These are minima because the solution ignores geothermal heat flux, the lower conductivity of the capping till layer, and the insulating effect of any surface water or snow bodies, and because an unlimited supply of basal meltwater is assumed at all times. If moraine-like relief is due totally to excess segregated ice, the largest ridges would each have required thousands of years to form and the moraine-like belt probably would have required more than 10,000 years, bearing in mind that the basins in the topography were produced by removal of a volume of segregated ice by thermokarst at least equal to that remaining under the ridges. However, the entire moraine belt on Wollaston Peninsula formed in an interval of about 1,000 years, as shown by radiocarbon dating of marine deposits on distal and proximal sides of the belt (Dyke and Savelle, 2000; Dyke et al., in press). This interval allows for an average of only 50

PRE-WISCONSINAN Permafrost aggrading to 300 m

-cooling climate

-glaciation and deformation



Permafrost to 300 m


Permafrost to 300m, except under lakes

Segregated ice Glacial ice Water

Sand and gravel (Pleistocene)

Clay and silt (Pleistocene)

Till or reworked till (Toker Point Stade)

Edge of permafrost

Figure 7.8 Rampton's model of thermokarst terrain development in the western Canadian Arctic (From Rampton, 1988).

years to form each of the estimated 20 ridges. Therefore, we suggest that Rampton's model for the origin of moraine-like belts is not applicable to Wollaston Peninsula, nor to any large moraine-like accumulation formed during deglaciation. Realistically it is not applicable to any individual moraine-like ridge more than about 10 m high. Moreover, the model embodies the irony that the incorporated deglacial process operates so slowly that end moraines would almost inevitably have formed by normal processes of sediment delivery to the ice margin in the long time intervals required for thick ice segregation.

We now consider whether there might not be simple geomorphic criteria to distinguish kettle lakes formed within the permafrost zone from thermokarst lakes. The distinction has significance beyond proper identification of moraine belts. For example, kettles that formed in permafrost terrain have no particular palaeo-climatic significance, having formed chiefly in response to the same warming that caused deglaciation. On the other hand, thermokarst lakes that formed long after deglaciation are taken as evidence of a warming event. The best-known

Figure 7.9 Extensively pitted terrain typical of the ice-cored areas of Tuktoyaktuk Peninsula, illustrated in part as figure 3 in Rampton (1988). The basins, interpreted herein as kettles in an ice-cored end-moraine belt, are interpreted as Late Wisconsinan thermokarst features that formed in Early Wisconsinan drift by Rampton. (NAPL AI2902-48.) Reproduced with the permission of Natural Resources Canada 2010, courtesy of the National Air Photo Library.

Figure 7.9 Extensively pitted terrain typical of the ice-cored areas of Tuktoyaktuk Peninsula, illustrated in part as figure 3 in Rampton (1988). The basins, interpreted herein as kettles in an ice-cored end-moraine belt, are interpreted as Late Wisconsinan thermokarst features that formed in Early Wisconsinan drift by Rampton. (NAPL AI2902-48.) Reproduced with the permission of Natural Resources Canada 2010, courtesy of the National Air Photo Library.

inference of climatically induced thermokarst in North America is the formation of the numerous lake basins in the hummocky moraine-like belt near Tuktoyaktuk where this event is thought to reflect warmer conditions of the last Milankovitch insolation maximum (e.g. Rampton, 1988; Burn, 1997). If our interpretation is correct, this inference is incorrect because the lakes are kettles. Classic thermokarst lakes, on the other hand, are fairly shallow and are commonly wind oriented (e.g. Mackay, 1963). Lakes of this kind occur in the Tuktoyaktuk region (see Fig. 4 or 21 in Rampton, 1988) adjacent to, but evidently not in, the belt of moraine-like topography, where lakes are kettle-like. No explanation has been offered as to why thermokarst lakes in terrain that is non-morainal should have become wind-aligned whereas those in moraine-like terrain are not. We suggest that kettle-like lakes in areas of hummocky and ridged moraine-like topography are best interpreted as kettles unless a stadial-scale age difference between ground ice formation and lake basin development can be demonstrated.

7.6.3 Hummocky and Ridged Moraine Belts: Active Versus Stagnant Ice

In the largest morainal belts, the most common topography is hummocky or chaotically moundy in character, particularly where the surface material is ice-contact stratified drift and where the debris accumulated in the deep-cut recesses between major ice lobes. However, subparallel linear ridges impart a broad organization to the belts, strongly suggestive of multiple ice fronts or of large englacial structures or supraglacial debris-covered ridges parallel to a margin. The latter are commonly arranged in nested suites lying transverse to former glacier flow. The longest continuous moraine ridge on Wollaston Peninsula extends for 100 km with no break wider than a meltwater channel. Cross-cutting relationships of both ice-marginal and ice-flow features indicate that the more prominent ridges, both here and in the Bluenose Lake region on the adjacent mainland, represent culminations of readvances, several of which are recognized (Dyke et al., in press).

Figure 7.10 A ground ice slump exposing massive ground ice on upper slope of an ice-cored moraine on western Victoria Island. The slump was triggered by heavy rain the day before the photograph was taken, and the ice face was inaccessible because material in the floor of the slump remained liquid. The capping colluviated till approximates the thickness of the maximum active layer development (~l m).

Figure 7.10 A ground ice slump exposing massive ground ice on upper slope of an ice-cored moraine on western Victoria Island. The slump was triggered by heavy rain the day before the photograph was taken, and the ice face was inaccessible because material in the floor of the slump remained liquid. The capping colluviated till approximates the thickness of the maximum active layer development (~l m).

Extensive areas of this moraine type are clearly underlain by buried ice that is undergoing a very slow melting from the top down. The dominant ablation-triggering process is the development of flow-slides (ground ice slumps) and active layer detachments that expose the ice cores (Fig. 7.10). Fresh slumps are rare compared with the numerous old slump scars. In all probability, slumping and ice-core degradation was at a maximum during earliest postglacial time, when moraine slopes were steepest and climate was generally warmer than present. However, even today an unusually warm summer or unusually heavy summer rain will trigger new slumps or reactivate older slumps. Exposures of the buried ice in places display considerable debris concentrations including large, commonly striated boulders, debris bands and folded folia (French and Harry, 1990). A considerable amount of ice, presumably of glacial origin, has also been detected beneath morainic topography by gravity profiling (e.g. Kotler et al., 1998), and surface kettles and large ice wedge polygons on moraine surfaces have been used to infer ice cores (Dyke and Savelle, 2000). Additionally, Dyke and Savelle (2000; also Dyke et al., 1992) point out that moraine volume is considerably reduced where a moraine extends below the marine limit or into a glacial lake basin. This reduction is explained by the fact that the water either prevented formation of an ice core or quickly destroyed one that did form. Debris-rich bands within glacier ice, regardless of origin (e.g. regelation, apron overriding, thrusting; see O Cofaigh et al., Chapter 3), control the distribution of sediment melting out on the glacier surface and therefore often produce linear ridges lying transverse to glacier flow. The resulting moraine is therefore strictly defined as hummocky moraine and controlled hummocky moraine, being moraine 'deposited during the melt-out of debris-mantled glaciers' (Benn and Evans, 1998), though we stress that within the permafrost zone, melting occurs only from the top down. However, by most common definitions a moraine is a landform produced after complete removal of glacier ice and therefore most of the expansive areas of hummocky glacial terrain in northern Canada are in fact supraglacial accumulations of debris or debris-mantled, relict glacier snouts. Nonetheless, the buried glacier ice is now part of the permafrost of the region, and if the ice is now in equilibrium with the environmental conditions, that is as far as deglaciation and deposition progress.

The extensively kettled, ice-cored, ridged and hummocky moraine of northern Canada has, not unlike in many other regions, been interpreted as the product of regional glacier ice stagnation (e.g. Sharpe, 1988, 1992). However, several characteristics question this mode of formation. The moraine possesses numerous linear ridges arranged in wide belts, the belts often bordered or separated by large 'end moraines'. Dyke and Savelle (2000) consider the individual linear ridges within the belts of hummocky moraine to be a record of ice marginal, supraglacial debris accumulation, or ice-cored end-moraine formation, and therefore indicative of repetitive moraine building and readvances during the recession of active glaciers. Proglacial meltwater channels and outwash trains and fans emanate from numerous individual ice-marginal positions, which could only have formed in sequence as opposed to being formed randomly or contemporaneously by regionally stagnant ice. These characteristics of ice-cored moraine belts are important, because were the ice cores to be removed, these belts would more closely resemble vast hummocky terrains typically interpreted as products of regional ice stagnation, and successive ice-marginal positions would be more difficult to discern.

7.6.4 Ice-Shelf Landforms

Ice-shelf landforms warrant special mention because they are better documented from this region than anywhere else in glaciated North America. This limited distribution accords with the evident constraint that ice shelves can be sustained only where the feeding ice is cold-based, warm-based ice having too little tensile strength to prevent calving (Benn and Evans, 1998).

The simplest evidence of large former ice shelves comes from low-to-negligible moraine gradients over significant distances. Bearing in mind that low gradients are at least in part due to differential glaci-isostatic rebound, and hence could have been zero at time of formation, large ice shelves are postulated on this basis to have fringed the Late Wisconsinan Laurentide ice limit in Amundsen Gulf and M'Clure Strait (Vincent, 1982; Dyke and Prest, 1987). Similarly, a horizontal moraine on the north coast of Bylot Island (Klassen, 1993) may have been formed by an ice shelf extending to the mouth of Lancaster Sound. The latter was proposed as the last glacial maximum position by Dyke and Prest (1987) but is now thought to be of early deglacial age (Dyke et al., 2002). Small horizontal moraines of few kilometres length in valleys on Ellesmere Island have similarly been attributed to ice shelves (England et al., 1978), and many other small horizontal moraine segments in the region may have a similar origin.

The Viscount Melville Sound Ice Shelf formed during a readvance of the northwest Laurentide margin in the late Pleistocene (Hodgson and Vincent, 1984), probably during Younger Dryas time. It was approximately 60,000 km2 in extent and is recorded geologically by the distinctive Winter Harbour Till, which was deposited where the shelf edge grounded in water shallower than the freeboard of the ice shelf. The till is thin and nearly featureless, its distinctive characteristics being a near-horizontal elevational limit along hundreds of kilometres of coastline and its relationship to both pre- and post-Winter Harbour marine limits. Weakly inscribed striae below the till consistently trend normal to shoreline regardless of shoreline orientation. Far-travelled erratics were commonly deposited during the event, indicating that the ice shelf was laden with debris from its previous grounded phase or that debris was efficiently transferred through it. Associated morainal forms are rare and evidently of ice-push origin. Lateral drainage channels are also rare and possibly associated with lakes that were impounded along the shelf edge.

Viscount Melville Sound is a nearly ideal location for the formation of an ice shelf. Its broad eastern end was the access route for inflowing ice from the main body of the ice sheet, and it is otherwise nearly enclosed by land. Thus the ice shelf was of the 'confined' type, with grounded distal margins (Benn and Evans, 1998). The lack of similar evidence of deglacial ice shelves in other large marine basins in the region suggests that conditions suitable for ice shelf formation were rarely met at that time. Theoretically, conditions should have been more suitable for ice shelf formation during ice advance phases, and indeed this process may have been essential for regional ice sheet establishment in the Canadian Arctic Archipelago.

Project Management Made Easy

Project Management Made Easy

What you need to know about… Project Management Made Easy! Project management consists of more than just a large building project and can encompass small projects as well. No matter what the size of your project, you need to have some sort of project management. How you manage your project has everything to do with its outcome.

Get My Free Ebook

Post a comment