Landforms indicating warm-based ice

Glacial scouring is generally accepted as an indicator of warmbased ice (cf. Reference DrewryDrewry, 1986). Reference SugdenSugden (1978) described this type of landscape, which is similar to the knock-and-lochain morphology introduced already by Reference LintonLinton (1963). Basal sliding, with plucking and abrasion under a thawed bed, is generally accepted as a process responsible for creating this terrain (e.g. Reference Andrews, Stravers, Miller and WoldenburgAndrews and others, 1985). In this study, when the individual landforms, as roche moutonées, correlate with ice movements inferred by the deglaciation pattern, the features are presumed to indicate warm-based ice during the late Weichselian deglaciation.

The exact origin of drumlins and fluting is much debated. Both accretion and erosion theories exist (Reference DrewryDrewry, 1986), as well as the theory of deforming beds (Reference Boulton, Menzies and RoseBoulton, 1987). All theories seem to presume wet-bed conditions. Some cases have been reported where flutes have been found frozen into the ice at the glacier snout (Reference Gordon, Whalley, Gellatly and VereGordon and others, 1992), which could imply a cold-based origin. However, it is also possible that the flutes were frozen on to the sole in a phase with thinning ice (Reference Benn and EvansBenn and Evans, 1998). It is assumed here that streamlining of glacier beds is only possible under a thawed glacier bed.

Subglacial eskers and meltwater channels also require a thawed water-penetrable bed. Many eskers in the northern and western part of the area are located in former glacial lake basins, which suggests they are of subglacial origin. Some of the smaller eskers on the low mountain plateaux may have been formed supraglacially and therefore cannot be used as temperature indicators.

Short meltwater channels that quickly curve downslope to form subglacial chutes can also be used to demarcate warm- based conditions. This type of channel cannot form when the ice is cold because it is then non-penetrable for melt- water drainage.

Landforms indicating cold-based ice

Surfaces with intensely weathered bedrock and well-developed patterned ground occur on many mountain summits. Lateral meltwater channels cut such surfaces at several localities, which implies that the weathered surfaces are of pre-late Weichselian age. They can thus be used as indicators of cold-based ice during deglaciation. Where the bedrock is glacially stripped of boulders as on Åreskutan, in some patches of Lunndörrsfjällen (Fig. 7) and Nipfjället, there are no signs of Holocene frost shattering. It is also supported by the fact that the patterned ground is often abruptly replaced by a homogeneous surface, with no sign of periglacial patterns. Although frost-shattered bedrock and patterned ground generally occur on mountain summits, the local appearance is one of an irregular patchiness. The distribution can not be explained by differences in topography, bedrock geology or local climate, but most likely reflects differential glacial erosion or accumulation due to a different basal regime. Sometimes the relict surfaces have distinct borders carrying erosion scarps or shear moraines against surrounding terrain that was fluted and eroded during the late Weichselian (Reference KlemanKleman and Borgström, 1994).

Thermoerosion features, mostly scarps and ridges (Fig. 3), are found on slopes in close connection to relict surfaces in the area. Similar features have been documented in supraglacial environments (Reference Paul and EylesPaul, 1983) and have also been found in permafrost regions. Reference Astakhov and IsayevaAstakhov and Isayeva (1988) described the formation of “thermocirques” in Siberia, where the ground ice is supposed to be early Weichselian glacier ice. Reference Lewkowicz, Dixon and AbrahamsLewkowicz (1992) described similar features from Canada. In both, the erosion is backwasting and initiated by running water. The thermoerosion features in the study area are always located on mountain or valley sides and often associated with flights of long meltwater channels that have drained the water that seemed to have initiated the backwasting. In at least one case, on the northern slopes of Anåfjällen, wave activity in a glacial lake initiated the thermoerosion. These landforms indicate a permafrost environment during deglaciation, and thus cold- based ice. This opinion was also put forward by Reference LagerbäckLagerbäck (1988).

Fig. 3. Thermoerosion landforms on Vedungsfjället. The slopes are aligned with thermoerosion ridges and hummocks. The summits of these features are remnants of the pre-late Weichselian surface. (Width in centre of photo is 500 m.).

Flights of long, consistent lateral meltwater channels reflect a cold ice surface. A cold ice surface is probably the only way to prevent water from penetrating down into the ice for long distances and during several melting seasons (Reference Sollid and SørbelSollid and Sörbel, 1982, Reference Sollid and Sørbel1988; Reference DykeDyke, 1993). A low crevasse frequency may also prevent surface water from draining into the ice, but this is unlikely to be the case for long distances in a mountainous area.

Glacial features related to earlier glacial stages may also indicate cold-based conditions. However, preservation of certain land- forms such as large drumlins is also possible under thawed- bed conditions (Reference LagerbäckLagerbäck, 1988; Reference ClarkClark, 1993; Reference KlemanKleman, 1994).

Landforms indicating a transition from cold-based to warm-based ice Reference Kleman and BorgströmKleman and Borgström (1990, Reference Kleman1994) described boulder blankets as boulder masses formed by the deposition of boulders, glacially transported over short distances. The source can often be related to scars in boulder fields or talus formations immediately up-ice. The direction of transportation usually corresponds with the last local ice-flow direction inferred from striations (Reference KlemanKleman, 1992). They are believed to form in a phase; the basal parts melt shortly before ice movement ceases during deglaciation. Thus, they most likely reflect a late-frozen to thawed-bed transition.

The formation of Rogen moraines (ribbed moraines) is discussed in numerous papers. A review is given by Reference Hättestrand and KlemanHättestrand and Kleman (1999). Reference HättestrandHättestrand’s (1997) model is interesting since it accounts for morphological, as well as geographical and stratigraphical, criteria. It also gives a plausible explanation of the dynamics and interactions between the glacier bed and its substratum. The ridges are believed to have been formed by detachment of frozen drift when the pressure-melting isotherm ascended through the glacier bed. The bed may have been completely thawed before deglaciation, but the ridges were only slightly remoulded or drumlinized because of stagnating ice movement close to the ice divide. However, there is still a question of how close to deglaciation the moraines were formed. As was shown in previous studies of the area (cf. Reference LundqvistLundqvist, 1969b; Reference BorgströmBorgström, 1989), many Rogen moraines were formed in high-relief topography without any sign of deviation from the overall flow direction in a fairly thick (> 1000 m) ice sheet. This implies that at least some of the Rogen moraine areas thawed a considerable time before ice-free conditions.

Summarizing assumptions

Glacial scouring, glacial lineations, subglacial eskers and chutes are assumed to indicate warm-based conditions. When individual landforms fit into the deglaciation pattern inferred from coherent marginal meltwater traces, the landform assemblages are believed to demarcate areas that had warmbased conditions during the late Weichselian deglaciation.

It is assumed the ice was cold-based during deglaciation in areas where relict surfaces dominate on mountain summits, and where thermoerosion features and flights of long lateral meltwater channels dominate on mountain and valley slopes.

Rogen moraines and boulder blankets reflect a transition from cold-based to warm-based conditions in a late stage of the late Weichselian cycle.

Thawed-bed conditions

Thawed-bed conditions appear to have been the dominating thermal regime in the northern and northwestern part of the area. This is inferred from the widespread occurrence of glacially scoured terrain. Apart from the scoured bedrock, fluting, eskers and glacial-lake sediments characterize this area. A few large drumlins are also found here. The erosion seems to have been moderate, since stoss-and-lee topography, formed by earlier ice flow from the northwest, is still present in the scoured terrain.

The most extreme examples of glacial scouring are found along the Swedish—Norwegian border and across the Åreskutan—Ottfjället mountain ridge. There is almost no surficial material, except peat, in these areas. This is also true for the summit area of Åreskutan (1420 m a.s.l.), where the lack of unconsolidated sediment implies that no frost shattering occurred here during the Holocene.

The northern and northwestern areas also show numerous short channel flights. Sometimes the lateral channels curve downslope to form chutes (Fig. 6) and sometimes the channels are replaced by eskers. An illustrative example of the latter situation is the Dörrsjöarna valley in Oviksfjällen. The thawed-bed conditions here seem to be a consequence of the high flow rates of ice converging into the valley. Subglacial eskers are frequent in the north and northwest. The larger ones are situated in valleys, the Indalsalven valley esker in the north is the largest and longest, locally more than 40 m high. Large eskers also occur in the western parts of the Ljusnan valley. There is a strong connection between large eskers and the amount of glacial lake sediments. The sediments are most widespread in the Indalsälven valley, with a thickness sometimes reaching 10 m (Reference LundqvistLundqvist, 1969a).

Fig. 6. Subglacial chutes in the Sylsjön valley. Short lateral meltwater channels and subglacial chutes indicate wet-bed ice. (Width in centre of photo is 1600 m.).

An extensive area of scouring is found in the Ljungan basin, northeast of Storsjön. Although many bedrock hills in this area reflect glacial plucking from the northwest, roche moutonées and crag-and-tail features clearly show that the last ice movement was from the southeast. This area is part of a narrow zone of glacial scouring and fluting that stretches to the southeast over the plateau east of Henvålen—Särvfjället. The last local ice divide was situated over this plateau, implying that the lineations were formed prior to the deglaciation when the ice divide was located further to the southeast, posibly during the LGM (Reference Kleman, Hättestrand, Borgström and StroevenKleman and others, 1997).

A striking feature is the lack of eskers, as well as glaciallake sediments, in the eastern areas of glacial lakes Ljungan- issjön and Ljusnan-issjön. This may be a result of low erosion rates, either due to cold-based ice or stagnant ice, near the ice divide. The scoured area northeast of Storsjön and the fluting field east of Särvfjället indicate the ice was warmbased in the Ljungan valley at an earlier stage. It seems unlikely the ice would refreeze in the valley to prevent production of eskers and glacial lake sediments. Hence, the short distance to the ice divide is a likely explanation.

Frozen-bed conditions

Frozen-bed conditions seem to have dominated in the low mountain terrain in the east, south and southwest. Boulder fields and patterned ground are almost invariably found here on topographic highs. The pre-late Weichselian origin of these surfaces is clearly demonstrated by lateral meltwater channels that are cut into boulder fields (cf. Klövsjöfjällen), boulder blankets that are drawn out from the boulder fields (cf. Sånfjället) and glacially eroded scarps in the patterned ground (cf. Häggingfjället, southwest of Sömlinghagna) (Reference KlemanKleman and Borgström, 1994).

Flights with long lateral meltwater channels also occur in the southern and eastern low mountains. They are often found high up on the summits, but can sometimes also reach all the way down to local valley bottoms, as in Klövsjöfjällen. They often completely surround the mountains. However, many channel flights have been more or less destroyed by thermoerosion processes, leaving only short remnants of individual channels intact.

Relict glacial landforms seem to occur on topographic highs in most of the area. They are especially frequent in Oviksfjällen, on Städjan—Nipfjället and on Långfjallet. A large terminal moraine is situated on the mountain ridge between Nipfjället and Städjan. Fossil ice-wedge polygons can be seen on top of the ridge. A similar ridge is located on Långfjället, near the border with Norway. Both moraines contain kettle holes. Reference KlemanKleman (1992) mapped similar moraines in the northern part of Sweden, where they were interpreted to represent a temporary halt in the recession of an older west-centred ice sheet.

Relict surfaces also occur in the northern mountains. The surfaces here are often restricted to the high mountain peaks, as in Sylarna and Helags, or to rather narrow remnants of dissected high mountain plateaux, as in Lunndörrsfjällen and Bunnerfjällen. Several localities with boulder blankets in Lunndörrsfjällen show the glacial plucking of boulder fields (Fig. 7).

Fig. 7. Boulder field on a summit in Lunndörrsfjällen, partially stripped by late Weichselian ice flow. (Photo: A. Philip). (Width in centre of photo is 400 m.).

Frozen- to thawed-bed transition

Following the hypothesis by Reference HättestrandHättestrand (1997), that Rogen moraines occur where the bed is recently thawed, there is an area mainly in the south and southwest on both sides of the ice divide where such an evolution occurred in valleys and on plateaux. In general, Rogen moraine areas and glacially scoured areas exclude each other. However, in some areas the two types occur together: around lake Rogen, most of the low mountain hills are scoured and even heavily plucked in the westernmost parts.

Boulder blankets are often found associated with relict surfaces. They form drawn-out tails from frost-shattered bedrock on Häggingfjället (southwest of Sömlinghågna), Sånfjället, Slagufjället and in Lunndörrsfjällen, and from talus formations on Lillfjället and Nipfjället. They show that the frozen patches all over the area probably were larger prior to deglaciation and that the erosion stopped because of stagnating ice movement.

Both eskers and fluting in the southeastern area indicate that the landforming processes continued after the bed had thawed, but the alteration of the terrain was small. Eskers on the plateaux and in the mountains are generally small and short, rarely more than 10 m high, and the fluting is faint.