Glacier Characteristics and Sediment Transfer System of Longyearbreen and Larsbreen, Western Spitsbergen

Norsk geogr. Tidsskr. Vol. 54, 157–168. Oslo. ISSN 0029-1951

Glacier characteristics and sediment transfer system of Longyearbreen and Larsbreen, western Spitsbergen

BERND ETZELMU¨ LLER, RUNE STRAND ØDEGA˚ RD, GEIR VATNE, RØNNAUG SÆGROV MYSTERUD, TORE TONNING & JOHAN LUDVIG SOLLID

Etzelmu¨ller, B., Ødega˚rd, R. S., Vatne, G., Mysterud, R. S., Tonning, T. & Sollid, J. L. 2000. Glacier characteristics and sediment transfer system of Longyearbreen and Larsbreen, western Spitsbergen. Norsk Geografisk Tidsskrift–Norwegian Journal of Geography Vol. 54, 157–168. Oslo. ISSN 0029-1951. Two small high-Arctic glaciers (Longyearbreen and Larsbreen) on Svalbard (788N158E) were studied with respect to glaciological and hydrological characteristics. Fieldwork during the melting season of 1993 and 1994 was coupled with digital map analysis based on high-resolution digital elevation models (DEM) to reveal the dynamics and temperature regime of small glaciers in a high-Arctic environment, and its relationship to the material transport and sedimentation of these glaciers. The study showed Long- yearbreen and Larsbreen to be low activity glaciers, cold-based with temperate patches, and thus having a low potential of basal erosion. The transport of ions and suspended solids in the glacial meltwater implies storage of material in and around the glacier which comes into contact with the meltwater. The study suggests that small Arctic glaciers couple the slope system with the fluvial system and therefore build a highly effective denudation system. Small polythermal glaciers are therefore important in understand- ing Pleistocene and Holocene landform development in cold regions. Keywords: cold glaciers, DEM, GIS, high-Arctic glaciers, map analysis, sediment budget Bernd Etzelmu¨ller, Johan Ludvig Sollid, Department of Physical Geography, University of Oslo, P.O. Box 1042, Blindern, N-0316 Oslo, Norway. E-mail: bernd.etzelmuller@geografi.uio.no; Rune Strand Ødega˚rd, Gjøvik College, P.O. Box 191, N-2801 Gjøvik, Norway; Geir Vatne, Department of Geography, NTNU-Dragvoll, N-7491 Trondheim, Norway; Tore Tonning, Statens Kartverk i Sogn og Fjordane, P.O. Box 9, N-6861 Leikanger, Norway; Rønnaug Sægrov Mysterud, Hydro Energi, Norsk Hydro ASA, N-0246 Oslo, Norway

Introduction Setting More than 60% of the land area of Svalbard is covered by The study site lies at 788N158E on the island of Spitsbergen glaciers. Most of the glacierized areas consist of plateau in the Svalbard archipelago (Figs 1 and 2). Mean annual air glaciers drained by large tidewater ice streams. On the temperature at sea level was 6.18C during the period 1976– western and northwestern parts of Spitsbergen, valley and 98 (Norwegian Meteorological Institute – DNMI). The cirque glaciers are abundant, with a surface area of up to 50 meteorological records show a mean annual precipitation km2 (Hagen et al. 1993). These glaciers often terminate on of 400 mm at the airport of Longyearbyen, which, however, land. As Svalbard lies in the zone of continuous permafrost, is known to underestimate the actual precipitation (Hanssen- at least the outer margins of these glaciers tend to be cold- Bauer et al. 1990). Permafrost thickness from 200 m to based (cf. Ødega˚rd et al. 1992, Bjo¨rnsson et al. 1996). 450 m is known from mining activity and boreholes in the Glaciers below 5 km2 are abundant in the relatively arid area (Liestøl 1977, Isaksen et al. 2000). Tertiary, flat-lying central and northern parts of Spitsbergen. They have an even sedimentary bedrock, mainly shale, siltstone and sandstone or regular surface with few or no crevasses, except in the (Major & Nagy 1972), dominates the lithology of the bergschrund and randkluft areas. They are characterized by study area. Mechanically, the bedrock is soft and fine- large, ice-cored terminal moraine complexes compared to grained. glacier size. Depending on glacier thickness, these glaciers Longyearbreen and Larsbreen are 2.7 km2 and 3.0 km2 in only have small temperate areas (Bjo¨rnsson et al. 1996), or size, respectively. Large, ice-cored moraines terminate both they are cold-based (Hodsen et al. 1997, Hodkins 1997). In glacier marginal areas. The surface layer covering the ice- the literature, the thermal regime of glaciers is often assumed core is between 0.5 m and 1.5 m in thickness and consists of to be an important system component regulating erosion, angular gravel to stones in a fine matrix. On Larsbreen the transport and sedimentation of debris (Weertman 1961, moraine texture is considerably more fine-grained than on Boulton 1972, Etzelmu¨ller & Sollid 1996, Benn & Evans Longyearbreen as Larsbreen has eroded in a more silt- 1998). The small valley glaciers can give valuable informa- rich bedrock. In the glacier marginal areas, ground-ice tion about the sediment transfer system in Arctic regions and slumps and thermokarst processes are common, exposing thus the geomorphological significance of cold or near-cold the ice core of the terminal moraines in many places. The glacier systems as a relief-forming agent. The main moraine area is morphologically sharply distinguished from objectives of this study are therefore to reveal the dynamics the surroundings, defining the maximum postglacial and temperature regime of small glaciers in a high Arctic extension of the glaciers, a pattern that is commonly environment, and their relationship to the material transport observed on Svalbard (Sollid & Sørbel 1988a, Etzelmu¨ller and sedimentation. et al. 1996). 158 B. Etzelmu¨ller et al. NORSK GEOGRAFISK TIDSSKRIFT 54 (2000)

Methods

Fieldwork The fieldwork was carried out during the melting seasons of 1993 and 1994. Ice-flow velocity on Longyearbreen and Larsbreen was measured using a conventional geodetic survey with a theodolite and electro-optical distance meter (Fig. 3). As the velocities were low and the fixed points could only be established on side moraines, annual movement was registered from 1993 to 1995. This resolution was satisfac- tory for the objectives of this paper. Glacier thickness and the transition between cold and temperate ice were mapped using radio-echo sounding with the two different frequency bands at 320–370 MHz and 30– 80 MHz (Figs 3 and 4). The radar system used is a time-gated synthetic pulse radar (Hamran & Aarholt 1993, Hamran et al. 1995). The radar and antennae were mounted on a sledge pulled by a snowmobile. Interpretation of the internal echo as an indicator of the thermal structure of a polythermal glacier followed that of Holmlund & Eriksson (1989), Ødega˚rd et al. (1992) and Bjo¨rnsson et al. (1996). Comparison of radio- echo sounding at 320–370 MHz and accurate borehole Fig. 1. Keymap of Svalbard. The black circle denotes the location of the temperature measurements at Finsterwalderbreen showed study site. Black areas show the glacier coverage. that the depth to temperate ice can be mapped within an error range of Æ4 m (Ødega˚rd et al. 1997). Meltwater discharge, electrical conductivity, and the concentration of suspended sediment and solutes were measured in the glacier marginal area (Fig. 3) at several downstream locations. Stage was measured using both a

Fig. 2. The Longyear valley, western Spitsbergen, view towards south. LoB = Longyearbreen, LaB = Larsbreen (photo: Trond Eiken). The small glacier located west of Longyearbreen has no ofﬁcial name. NORSK GEOGRAFISK TIDSSKRIFT 54 (2000) Glacier characteristics, western Spitsbergen 159

Fig. 3. Topographic map of the Longyear valley catchment and position of measurement sites on Longyearbreen and Larsbreen. standard analogue hydrograph and a pressure transducer. geometrical description of the glaciers were calculated using Water samples for determining suspended sediment Zevenbergen & Thorne’s method (1987). The index of concentration were collected using an ISCO automatic roughness calculates the 3-dimensional resultant vector and sampler and filtered through Whatman GF/C glass vector strength in a local neighbourhood (window kernel) as microfibre filters. Samples for water chemistry analyses a measure of changes in surface slope and aspect. The were collected manually and filtered in field through algorithm used is proposed in Mark (1975) and the Millipore 0.45 mm HA filters and transferred to pre-cleaned mathematics are outlined in detail in Mardia (1972). polypropylene bottles. Temporal resolution of these measurements varied between twice a day and hourly series. Map analysis GIS spatial analysing capability based on map modelling Digital elevation model and glacier geometry concepts (Tomlin 1990, Berry 1993) was used to calculate ice deformation velocities, basal shear stress and potential The digital elevation model of the subglacial bedrock drainage system of both glaciers. topography was compiled by drawing manual contours based The deformation of ice is described by a flow law, e = Atn, on the radio-echo profiles, digitising of the contours and where e is the strain rate and t is the shear stress (Glen 1955). gridding using the Hutchinson (1989) procedure. A high- A depends on temperature and crystal orientation, while n is a resolution (1:10000) digital map of the drainage basins of constant, normally 3. Kamb & Echelmeyer (1986) take both glaciers was compiled using a WILD A7 photogram- account of variations of the longitudinal stress along the flow metric instrument equipped with a digital encoder. Altitude direction, which is caused by varying subglacial topography information for the 10 m DEM was sampled by digitising and thus glacier thickness. These stress gradients are 5 m contours, break lines and prominent points. described by a longitudinal flow coupling equation contain- Relief parameters such as slope and aspect in the ing weighting functions for averaging that depend on the 160 B. Etzelmu¨ller et al. NORSK GEOGRAFISK TIDSSKRIFT 54 (2000)

Fig. 4. Example of a radio-echo sounding proﬁle with frequency bands of 30–80 MHz along a transect across the lower part of Longyearbreen. The bottom reﬂectors are indicated by the line of black pixels within the image. The gray solid line displays our interpretation of the subglacial relief.

glacier thickness and position on the glacier. The basal shear The shape factor F is a number between 0 and 1, and is stress is calculated as: related to the ratio between the cross-channel area and its Z1 perimeter (hydraulic radius) of a glacier (Nye 1965, Paterson j 0 j 1 0 1=n 0 x x 0 1994). x x h x e l dx 1 b 2lh1=n L The driving force for the water flow in and under glaciers 1 is the hydraulic potential. Calculations of the hydraulic potential are based on the ice-surface and ice-thickness where tL = Frgh sina was calculated at the local centre position x’ of the weighting function, l is the longitudinal gradients. The water pressure potential at the base of a Φ coupling length, h is the glacier thickness, is the density of glacier ( b) is the sum of the elevation potential and the ice and a is the glacier surface gradient. Surface velocity is pressure potential (Shreve 1972, Bjo¨rnsson 1988) defined as Φ then calculated in a two-step process. First, a quantity A at b = w gZb Pw, where Pw is the subglacial water pressure location x is Pw =k ig(Zs Zb), i.e. proportional to the ice-overburden pressure. Z and Z are the elevations of glacier surface and Z1 s b 0 bedrock, respectively, while is the density of ice and is jx xj 0 i w A x nln F n1lnh 1 dx 2 the density of water. k is a factor between 0 and 1 2l 1 representing the range of the water pressure in subglacial conduits from atmospheric pressure (k = 0) to full ice- where x’ is an integration variable and displays a local overburden pressure (k = 1) when the channels are com- longitudinal co-ordinate position. Second, the velocities (U) pletely filled with water. The subglacial meltwater then flows are calculated as normal to lines with equal potential (Shreve 1972, Bjo¨rnsson

A xA x0 1988). On a discrete elevation model containing information U xUobs x0e 3 of subglacial water potential, the solution for this problem is where Uobs (x0) is an observed surface velocity vector at a recursive algorithm which sums up the number of cells that location x0, and A(x0) is a corresponding value obtained from drain to each individual cell (Etzelmu¨ller & Bjo¨rnsson 2000). equation (2). The calculated velocity is calibrated with the The map analysis methods are described in more detail in surface velocity measured in the ﬁeld, thus including Etzelmu¨ller & Bjo¨rnsson (2000). The spatial analysis possible sliding movement. describes a glacier stage at a deﬁned time and does not NORSK GEOGRAFISK TIDSSKRIFT 54 (2000) Glacier characteristics, western Spitsbergen 161

Table 1. Statistical descriptors of glacier surface relief parameter and average expressed by the F-factor (Nye 1965). Thus, Longyear- values of glaciological measurements and calculations based on the 1990- breen’s narrow, but relatively deep, channel results in lower surface for Longyearbreen and Larsbreen. Italic numbers in parentheses are calculated along the central ﬂow lines of the glaciers investigated. F-values compared to Larsbreen.

Longyearbreen Larsbreen

Geometry Thermal regime Area (1990) 2.52 km2 2.96 km2 Perimeter 13.4 km 13.0 km The interpretation of the radio-echo data shows that neither Length (L) 3600 m 2800 m Longyearbreen nor Larsbreen display a typical polythermal Width (W) 520 m 700 m structure. In most places, the backscatter from the 30– Elongateness (L/W) 6.9 3.5 80 MHz frequency band interpreted as the bed echo (cf. Fig. Orientation 458 358 Volume 0.13 km3 0.15 km3 4) was difﬁcult to distinguish from the 320–370 MHz echo, Avg. thickness 53 m (88 m) 49 m (82 m) which was expected to detect temperate ice. The backscatter Glaciological characteristics from both frequency bands showed differences in the 1 1 Velocity ELA 3ma 4ma positions of the main reﬂectors, but this difference could Avg. shear stress 0.40 (0.38 bar) 0.57 (0.56 bar) 1 1 also have been caused by an increase in debris content close Avg. calc. velocities 0.6 (0.6 ma ) 1.2 (2.0 ma ) F-factor (Nye 1965) 0.76 0.90 to the bed of the glaciers. The only area that shows clear Altitude distribution (m) indications of temperate ice is the western upper area of Mean 542 600 Longyearbreen. The western ice-stream of Longyearbreen Median 535 603 originates from high altitude on Nordenskio¨ldfjellet. De- Std 135 89 Skew 0.3 –0.5 tailed studies of the bergschrund/randkluft area also show Slope distribution (8) accumulation of temperate ice in the upper accumulation Mean 9.7 8.8 area of Longyearbreen. At Larsbreen there are no clear Median 6.6 5.8 indications of temperate ice from the radio-echo sounding. Std 7.6 6.8 Except for the areas of Longyearbreen mentioned, the Index of roughness Mean 23.4 22.8 radio-echo sounding shows that these glaciers are either Std 16.0 14.8 cold-based or cold with a thin layer of temperate ice at the bed. Former coal mining beneath the front area of Larsbreen revealed subglacial bedrock temperatures of about 28C (O. Liestøl, pers. comm. 1988). This means that the front area of integrate changes of mass balance or water pressures in Larsbreen is clearly cold-based. This is most likely to be the channels during the year. case also for Longyearbreen.

Dynamics Results The ice-flow velocities are low on both glaciers, with values Mass balance and glacier geometry between 2 ma 1 and 4 ma 1 for the stakes situated in the accumulation area of Longyearbreen (560 m a.s.l.) and Long-term mass balance measurements show that most of Larsbreen (650 m a.s.l.), respectively. In the lowermost the smaller glaciers in western and central Spitsbergen have ablation area, flow velocities decreased to below 1 ma1 decreased in volume, at least in the last 25 years (Hagen & for both glaciers. Liestøl 1990, Etzelmu¨ller & Sollid 1996). Mass balance The calculated basal shear stress values are low, below measurement on Longyearbreen during the period 1977 to 60 kPa along the central flow line (Fig. 5). Owing to higher 1992 displayed an annual mass loss of 0.5 m w.e. (Hagen & F-factors and partly to greater glacier thickness values for Liestøl 1990), a number which was roughly measured also in Larsbreen, both basal shear stress and calculated surface the balance year 1993–94 along a limited number of stakes. velocities are considerably higher on Larsbreen than on The mass-balance gradient for glaciers on western Spitsber- 1 Longyearbreen. The surface velocity measurements on gen is in the order of 0.3 m (100m) , which describes Longyearbreen display higher velocities than modelled, relatively low mass turnover rates. indicating sliding movement at the bed during summer. For Longyearbreen and Larsbreen are of approximately the Larsbreen, both the calculated and the measured flow same size and volume (cf. Table 1). Surface geometry differs velocities are in the same order. by almost twice the elongateness values for Longyearbreen in relation to Larsbreen. As glacier thickness is about 50 m on average for both glaciers (Table 1), the channel geometry Hydrology is different. For Longyearbreen the radio-echo backscatter from the 30–80 MHz frequency band revealed a deep and Similar to most small valley and cirque glaciers in Spitsber- near V-shaped channel profile in the lower part of the glacier. gen, supraglacial drainage is an important mode of drainage Larsbreen displayed a wider, U-shaped channel geometry on Longyearbreen and Larsbreen. This is mainly due to the (Figs 3 and 4). The channel geometry is an important factor cold surface layer and lack of crevasses. Close to the for estimating the subglacial stress conditions, and is usually marginal areas, supraglacial meltwater is routed to englacial 162 B. Etzelmu¨ller et al. NORSK GEOGRAFISK TIDSSKRIFT 54 (2000)

Fig. 5. Calculated shear stress (a) and deformation velocity (b) for Longyearbreen and Larsbreen, based on Kamb & Echelmeyer’s model (1986). The diagrams (c), (d) and (e) are sampled along the central ﬂow lines of the glaciers. The ﬁlter kernel for spatial averaging was 500 m in diameter. Solid line = Larsbreen; dotted line = Longyearbreen.

or subglacial conduits. Observations in the portal show that under atmospheric pressure (Fig. 6). The spatial analysis the channels are of Ro¨tlisberger type (R-channels) (Ro¨thlis- shows that both supraglacial drainage and channel pressure berger 1972) developed in the transition area of clean ice and close to ice-overburden pressures are most realistic in debris-rich basal ice. On both Longyearbreen and Larsbreen describing the drainage pattern. Moreover, a large area of most meltwater is routed through tunnels within the ice- the Larsbreen accumulation area is drained towards Long- cored moraine. On Larsbreen, some meltwater runs on the yearbreen under both conditions. However, no moulins were surface of the ice-cored moraine, eroding in the ice-core found in the indicated areas of subglacial temperate ice, material cover. On both glaciers, lateral meltwater channels suggesting no signiﬁcant subglacial drainage in this area. are well developed, collecting glacier meltwater, snow melt Hence, most of the meltwater drainage is believed to be and melting permafrost from the adjacent slopes. supraglacial in the central and upper parts of the glaciers. The position of conduits was modelled using map- The discharge from Longyearbreen varied from 2 m3 s1 modelling principles (Etzelmu¨ller & Bjo¨rnsson 2000). to 4 m3 s 1 over the monitoring period. Discharge measure- Conduit shape and position are highly dependent on the ment below the conjunction between the streams from both water pressure conditions. For the glacier studied, three cases glaciers indicates that about 60% of the water came from of glacier drainage were modelled: (1) supraglacial drainage, Longyearbreen and 40% from Larsbreen. The higher run-off (2) drainage under ice-overburden pressure, and (3) drainage from Longyearbreen is explained by higher ablation rates NORSK GEOGRAFISK TIDSSKRIFT 54 (2000) Glacier characteristics, western Spitsbergen 163

Fig. 6. The potential discharge system of Longyearbreen and Larsbreen. (a) Potential supraglacial drainage, (b) subglacial drainage under near atmospheric conditions (k = 0.2) and (c) subglacial drainage under ice-overburden conditions (k = 1).

there due to lower surface altitude, and the capture of sediment concentration increased throughout the ablation meltwater from both Larsbreen and the small cirque glacier season (Fig. 8). west of Longyearbreen (Figs 2 and 3). The grain size distribution of the suspended material was significantly finer on Larsbreen than on Longyearbreen, which seems to reflect differences in till properties found in the proglacial areas of the two glaciers. Observations suggest Sediment transport the till of Longyearbreen to be coarser than at Larsbreen. Suspended sediment transport from both glaciers was high, Total solute transport was in the order of 20% to 30% of with average values of 0.5 gl1 and maximum values above suspended transport, with generally lower values on Lars- 1.0 gl1. Specific suspension transport was in the order of breen than on Longyearbreen (Fig. 8). During the melting 200–400 gs1 km2 for both glacier streams, corresponding season a decrease in solute concentration was observed for to 1500 ta1 km2. At Longyearbreen, the increase was Larsbreen, and an increase for Longyearbreen. A change parallel to an increase in discharge, especially towards to the from a dominance of alkali ions toward earth alkali ions and end of the melting season (Fig. 7). On Larsbreen this chloride towards bicarbonate and sulphate was found for relationship was not clearly identified. The suspended both stream systems during the summer (Fig. 8). This 164 B. Etzelmu¨ller et al. NORSK GEOGRAFISK TIDSSKRIFT 54 (2000)

value which is commonly observed for smaller Spitsbergen glaciers ending on land (cf. Hallet et al. 1996).

Discussion The glaciology of small high-Arctic glaciers The ice-flow velocity and thus mass flux on Longyearbreen and Larsbreen are low and mainly attributed to internal deformation of ice. The expected limited sliding, or even absence of sliding, for the two glaciers studied results in low subglacial erosion potential under recent glaciological conditions. Analysis of the subglacial relief, especially below Longyearbreen, shows that the preglacial valley forms were preserved (cf. Fig. 4). This means that at least Longyearbreen did not have much basal erosion in the lower Fig. 7. Relationship between discharge and suspension sediment transport for part, which would have caused a more U-shaped valley the Longyearbreen meltwater stream. profile. Liestøl (1988) claims that most of the Spitsbergen glaciers are of surge type. In contrast, other glaciologists argue that indicates that meltwater is in constant contact with solid only a third of Svalbard glaciers show surge behaviour material towards the end of the melting season. (cf. Dowdeswell et al. 1991, Hamilton 1992). They used Total material transport out of the glacier catchment 1 morphological indicators to identify surge-type glaciers, corresponds roughly to a denudation rate of 0.2 mm a ,a such as folded medial moraines, thrusts in the ice and push-moraine forms (cf. Hambrey et al. 1996, 1999). On Longyearbreen and Larsbreen, these features were not identified. Large ice-cored moraines cover the glacier fore- land area. The development of these moraines is due to advances, followed by stagnation and long retreat phases (cf. Etzelmu¨ller et al. 1996) under permafrost conditions, and are not necessarily indicative of surge behaviour. For Longyearbreen and Larsbreen the mass balance and flow velocity measurements were not sufficient to detect a possible surge build-up.

Controls of sediment transfer in cold glaciers Meltwater drainage. On both glaciers meltwater drainage is mostly supraglacial in the accumulation area and the higher parts of the ablation area. Towards the glacier front, meltwater is routed, allowing contact to bedrock or till debris. The observed englacial and subglacial conduits in the marginal areas seem to be formed by a gradual down-melting from the ice surface due to enhanced melt rates in the meltwater channels compared to the surrounding ice surface. Under thin ice, with temperatures well below pressure melting point, closure rates of englacial and subglacial conduits are low (cf. Paterson 1994). Direct observations in such glaciers (cf. Pulina 1984, Pulina & Rehak 1991) and studies of internal drainage systems using tracer techniques indicate the drainage system to be more static within cold ice than those found in temperate ice.

Suspended sediment source. – Several studies of glacier hydrology on Spitsbergen glaciers observed high values for suspended sediment transport. This is in spite of a short melting season and low melting rates in relation to Fig. 8. Solute and solid material transport measured during the melting temperate glaciers at lower latitudes (Vatne et al. 1995, season on Longyearbreen and Larsbreen in 1993. Bogen 1996, Hallet et al. 1996, Hodsen et al. 1997, NORSK GEOGRAFISK TIDSSKRIFT 54 (2000) Glacier characteristics, western Spitsbergen 165

Hodkins 1997). This has been attributed to the presence of gradients at sub-zero temperatures (Hallet 1983). Freeze– unconsolidated subglacial material, serving as a sediment thaw cycles were not observed in the bergschrund except for source to be evacuated by fluvial processes (Vatne et al. seasonal freeze–thaw. In the case of Longyearbreen, debris 1995). Some of these Svalbard valley glaciers described in from the rock wall did not enter the bergschrund, but the literature are temperate over large areas, allowing accumulated on the glacier within a range of c. 150 m from subglacial meltwater to evacuate unconsolidated sediment. the glacier margin. The relative importance of extra-glacial Also from this study, the sustained glaciofluvial-suspended material input in the accumulation area is a function of the sediment transport indicates the presence of sediment surrounding mountain head wall area (M) in relation to the sources from which meltwater can evacuate sediment. For glacier size (G). The quotient M/Gs are 1.7 and 0.5 for Longyearbreen and Larsbreen the source may either be Longyearbreen and Larsbreen, respectively. This number material released from the ice-cored frontal and lateral expresses, the high potential influence of extra-glacial moraines, or material stored subglacially. material input, at least for Longyearbreen, where the Material released from ice-cored areas is called thermo- surrounding head wall area is considerably larger than the karst and is described in more detail in, for example, Driscoll glacier size. (1980), Pickard (1983, 1984), King & Volk (1994) and Etzelmu¨ller (2000). Studies of thermokarst on glaciers in the Liefdefjorden and Ny-A˚ lesund area show that a large Influence of glacier thermal regime on the sediment cas- quantity of material evacuation out of the catchment can cade system. – The specific sediment evacuation rate for attributed to the decay of ice-cored moraines. There, the whole glacier area is in the order of 0.1 ma1 to Etzelmu¨ller (2000) estimated a specific denudation rate due 0.2 ma1. These values are comparable to other studies on to thermal erosion in the order of 0.05 mm a1. This process Svalbard (Bogen 1990, Vatne et al. 1995, Bogen 1996, is active in the frontal and lateral areas of both the glaciers Hodsen et al. 1997, Vatne 1997, Hodkins 1997) and investigated, but was not quantified in this study. demonstrate high catchment denudation rates despite short On Longyearbreen, the high ion transport and carbonate/ melting seasons and often low dynamic activity. Compared sulphate compositions during the end of the melting season to glaciers on the Scandinavian mainland, the values are in (Fig. 8) are indicative of the presence of subglacial material. the same order (Hallet et al. 1996), despite the temperate As the present dynamics inhibit subglacial material produc- glacier types dominating there. This implies that the tion, it may stem from an earlier period with higher glacier estimated denudation rate based on fluvial material trans- activity such as, for example, a surge event, from preglacial port is not related to present glacier activity for the two weathering or from extra-glacial material input. glaciers studied. Hence, the control of cold or near-cold A surge-event is possible, but the hypothesis cannot be high-Arctic glaciers on the sediment cascade system within verified in the case of the glaciers investigated. Preglacial a catchment is different from that of temperate glaciers weathering material is obviously a factor. Liestøl (1994) terminating in a non-permafrost environment. reports a continuous transition from ice to deeply weathered An example of a sediment cascade system for two bedrock in a mining tunnel below the glacier Foxfonna, different settings in terms of glacier thermal regime is about 10 km south of the study site. The surrounding conceptually displayed in Fig. 9. Polythermal or cold valley summits and ice-free valleys in central Spitsbergen also and cirque glaciers have high supraglacial material input show clearly deep weathering (Sollid & Sørbel 1988b), from adjacent slopes. Low fluvial evacuation of material and which would be available for subglacial evacuation if possible surging of the glacier produce a subglacial sediment glaciated. magazine. This magazine is constantly evacuated during However, extra-glacial material might be the most melting, giving a high correlation between discharge and important factor on small Arctic glaciers. Andre´ (1991) suspension transport throughout the year. During periods measured material input by rock fall and snow avalanches on with high discharges, for example, after a period with higher Austre Loveńbreen on Brøggerhalvøya in western Spitsber- summer temperatures, sediment transport and evacuation gen. Andre´ (1991) roughly estimated the erosion in the also increase as long as a sediment magazine exists in the mountain walls consisting of mica schist to the magnitude of course of the meltwater streams. In contrast, meltwater 0.08 to 016 ma1. In the case of Longyeardalen this number streams from temperate glaciers often only transport the may be a good estimate, as highly frost-sensitive sedimentary winter production of material during the summer, and rocks dominate the bedrock there. Extra-glacial material, subglacial sediments are exhausted during autumn (cf. which falls onto the accumulation area, will reach a Lawsen 1993). A period with higher summer temperatures subglacial position and be a supply to a subglacial material would increase meltwater production, but could lead to a magazine. In this respect, the bergschrund area of glaciers is decrease in glacier volume. This in turn reduces glacier flow of high importance. During field studies in 1996 mechanical velocity and thus subglacial material production due to weathering was observed in the bergschrund of Long- reduced subglacial erosion. Hence, total sediment transport yearbreen. Temperature measurements in the bergschrund and evacuation during the melting season would decrease showed a strongly buffered thermal regime during the melt also. season due to snow cover. The thermal regime is similar to Therefore, the same climate signal in terms of increased what has been observed in Arctic coastal rock cliffs (Ødega˚rd summer temperature would result in different sediment et al. 1995). This could be an environment of mechanical transfer signals depending on the thermal regime of the weathering due to moisture migration along temperature glaciers. 166 B. Etzelmu¨ller et al. NORSK GEOGRAFISK TIDSSKRIFT 54 (2000)

Fig. 9. Simpliﬁed sediment transport system for small valley glacier systems in a cold-polar and temperate environment, respectively. The thickness of the arrows indicates the relative importance of sediment transport paths.

Significance of small polythermal and cold glaciers present glacier erosion is less important than the non-glacial on landscape development processes. This has the consequence that material production and storage may be more periglacially than glacially The relationship between slope processes and river sys- governed on Longyearbreen and Larsbreen. Thus, the tem. – Caine (1986) and Caine & Swanson (1989) report sediment cascade system in such catchments is strongly from mountainous areas in Colorado that the fluvial system influenced by the periglacial system environment. is only loosely coupled to the slope system with respect to material transfer. Without material removal from the slopes by, for example, rivers, the slope will develop to a steady- Land-form preserving characteristics of cold glaciers. – A state and thus stable profile (cf. Ahnert 1994, 1996). This striking feature in areas dominated by sedimentary rocks is the case in the large unglaciated valleys of central on Svalbard is the abundance of steep V-shaped valleys. Spitsbergen. These have wide, flat valley bottoms, which These have been described earlier by, for example, Bu¨del normally do not reach the valley sides in the lower valley (1950), Rudberg (1988), and Sollid & Sørbel (1988b). areas. Hence, there the slope system is de-coupled from the Obviously, there are high weathering rates, and material is present non-glacier stream system (Sollid & Sørbel 1988b, effectively evacuated by the streams. Bu¨del (1950) dis- L. Sørbel, pers. comm. 1999). These large valleys can only cusses this phenomenon in terms of an area of ‘extensive be emptied during glaciations, which would probably valley generation’, where there is high material production remove the talus deposits. For example, the frequent due to frost processes and material removal, which keeps existence of rock glaciers on the slopes indicates several the slopes steep and preserves an effective coupling of the thousand years without removal of material. However, valley and fluvial system. episodic events like spring slush avalanches result in high- It seems likely, however, that the formation of these fluvial magnitude transport into the valley (cf. Barsch et al. 1993, valley cuttings may be pre-Holocene, and that many of these 1994). valleys may have survived the last glaciation. On Long- Through the small valley glaciers, such as those studied in yearbreen the V-shaped valley profile seems to have been this paper, the slope system is effectively coupled with the preserved under the ablation area. As we can assume that fluvial system. Comparing the mountain wall erosion most Svalbard cirque and valley glaciers have existed for estimates by Andre´ (1991) and the mobilisation of material most of Holocene time (cf. Furrer 1992, 1994), this profile is due to thermal erosion (cf. Etzelmu¨ller 2000) with the likely to have been formed in pre-Weichselian time. Our estimated erosion rates from the Longyear valley catchment, observations support Sollid & Sørbel’s (1988b) suggestion NORSK GEOGRAFISK TIDSSKRIFT 54 (2000) Glacier characteristics, western Spitsbergen 167 that the subglacial bedrock topography is preserved in the Bogen, J. 1990. Erosion and sediment transport in Svalbard. Proceedings presence of cold ice. Arctic Hydrology. Present and Future Tasks, 14–17 September 1999. NHK Report 23. Longyearbyen, Svalbard. Bogen, J. 1996. Erosion rates and sediment yields of glaciers. Annals of Conclusions Glaciology 22, 48–52. Boulton, G. S. 1972. The role of thermal regime in glacial sedimentation. The following conclusions might be drawn from this study: Inst. Br. Geogr. Spec. Publ. 4, 1–19. . Bu¨del, J. 1950. Das System der klimatischen Morphologie. Deutscher Longyearbreen and Larsbreen are low activity glaciers, Geographentag Mu¨nchen, 1948 27, 65–100. cold-based with temperate patches and a low potential for Caine, N. 1986. Sediment movement and storage on alpine slopes in the basal erosion. Colorado Rocky Mountains. Abrahams, A.D. (ed.) Hillslope Processes, . Transport of ions and suspended solids in the glacial 115–137. Allen & Unwin, London. Caine, N. & Swanson, F. J. 1989. Geomorphic coupling of hillslope and meltwater implies storage of material in and around the channels systems in two small mountain basins. Zeitschrift fu¨r Geomor- glacier which comes into contact with the meltwater. phologie 33, 189–203. . Small Arctic glaciers couple the Periglacial slope system Dowdeswell, J. A., Hamilton, G. S. & Hagen, J. O. 1991. The duration of the with the fluvial system and therefore build a highly active phase on surge-type glaciers: contrast between Svalbard and other effective denudation system. regions. Journal of Glaciology 37, 388–400. . Driscoll, F. G. 1980. Wastage of the Klutlan ice-cored moraines, Yukon V-shaped subglacial valley profiles beneath Longyear- Territory, Canada. Quaternery Research 14, 31–49. breen indicate the low subglacial activity and the Etzelmu¨ller, B. 2000. Quantification of thermo-erosion in pro-glacial areas – possibility of preglacial form preservation under cold examples from Spitsbergen. Zeitschrift fu¨r Geomorphologie, NF. In press. thermal conditions. Etzelmu¨ller, B. & Sollid, J. L. 1996. Long-term mass balance of selected poly-thermal glaciers on Spitsbergen, Svalbard. Norsk Geografisk Tids- Small polythermal glaciers with subdued dynamics are skrift 50, 55–66. likely to be important with respect to Pleistocene and Etzelmu¨ller, B. & Bjo¨rnsson, H. 2000. Analysis of digital elevation models for glaciological applications. International Journal of Geographical Holocene landform development in Polar regions. Quantita- Information Science 14, 567–581. tive studies are needed to investigate this role and to assess Etzelmu¨ller, B., Hagen, J. O., Vatne, G., Ødega˚rd, R. S. & Sollid, J. L. 1996. their geomorphic impact during the Weichselian glaciation Glacier debris accumulation and sediment deformation influenced by and on post-Weichselian landscape development. permafrost, examples from Svalbard. Annals of Glaciology 22, 53–62. Furrer, G. 1992. Zur Gletschergeschichte des Liefdefjords. Stuttgarter Geographische Studien 117, 267–278. Acknowledgements. – The study was initiated by Johan Ludvig Sollid, Furrer, G. 1994. Zur Gletschergeschichte des Liefdefjords/NW-Spitzbergen. Department of Physical Geography University of Oslo, partly in collabora- Zeitschrift fu¨r Geomorphologie, Suppl. Bd 97, 43–48. tion with Svalbard Samfunnsdrift, Longyearbyen, Svalbard. The Norwegian Glen, J. W. 1955. The creep of polycrystalline ice. Proceedings Royal Society Polar Institute (scholarship no. 46/93 and 35/94), the Store Norske Series A 238, 519–538. Kullkompani and the Faculty of natural science, University of Oslo, Hagen, J. O. & Liestøl, O. 1990. Long-term glacier mass-balance investiga- financially funded the fieldwork. Analysis of the water samples was carried tions in Svalbard 1950–88. Annals of Glaciology 14, 102–106. out at the Department of Geology, University of Oslo. Production of the Hagen, J. O., Liestøl, O., Roland, E. & Jørgensen, T. 1993. Glacier atlas of topographic map of the area and of the digital spatial analysis were carried Svalbard and Jan Mayen. Norsk Polarinstitutt Meddelelser 129. out at the Laboratory for Geographical Information Technology and Remote Hallet, B. 1983. The breakdown of rock due to freezing: a theoretical model. Sensing, Department of Physical Geography, University of Oslo. Lothar Proceedings. 4th International Conference on Permafrost, Fairbanks, Schrott, University of Bonn, and an anonymous reviewer made considerable Alaska, 433–438. improvements to the manuscript. We thank the aforementioned persons and Hallet, B., Hunter, L. & Bogen, J. 1996. Rates of erosion and sediment institutions. evacuation by glaciers: a review of field data and their implications. Global Manuscript submitted 1 November 1999; accepted 1 July 2000 and Planetary Change 12, 213–235. Hambrey, M. J., Dowdeswell, J. A., Murray, T. & Porter, P. R. 1996. Thrusting and debris entrainment in a surging glacier: Bakaninbreen, References Svalbard. Annals of Glaciology 22, 241–248. Hambrey, M. J., Bennett, M. R., Dowdeswell, J. A., Glasser, N. F. & Huddart, Ahnert, F. 1994. Equilibrium, scale and inheritance in geomorphology. D. 1999. Debris entrainment and transfer in polythermal valley glaciers. Geomorphology 11, 125–140. Journal of Glaciology 45, 69–86. Ahnert, F. 1996. Einfu¨hrung in die Geomorphologie. UTB, Stuttgart. Hamilton, G. S. 1992. Investigations of Surge-type Glaciers on Svalbard. Andre´, F. M. 1991. Geomorphic impact of spring avalanches in Northwest PhD thesis. University of Cambridge, Cambridge. Spitsbergen (798N). Permafrost and Periglacial Processes 1, 97–110. Hamran, S.-E. & Aarholt, E. 1993. Glacier study using wavenumber domain Barsch, D., Gude, M., Maüsbacher, G., Schuhkraft, G. & Strauch, D. 1993. synthetic aperture radar. Radio Science 28, 559–570. Slush stream phenomena – process and geomorphic impact. Zeitschrift fu¨r Hamran, S.-E., Gjessing, D. T., Hjelmstad, J. & Aarholt, E. 1995. Ground Geomorphologie, Suppl. Bd. 92, 39–54. penetrating synthetic pulse radar: dynamic range and modes of operation. Barsch, D., Gude, M., Maüsbacher, R., Schukraft, G. & Schulte, A. 1994. Journal of Applied Geophysics 33, 7–14. Recent fluvial sediment budgets in glacial and periglacial environments, Hanssen-Bauer, I., Kristensen Sola˚s, M. & Steffensen, E. L. 1990. The NW Spitsbergen. Zeitschrift fu¨r Geomorphologie, Suppl. Bd. 97, 111–122. climate of Spitsbergen. Report 39/90, . Benn, D. I. & Evans, D. J. A. 1998. Glaciers and Glaciations. Arnold, Hodkins, R. 1997. Glacier hydrology in Svalbard, Norwegian High Arctic. London. Quaternary Science Review 16, 957–973. Berry, J. K. 1993. Cartographic modelling: the analytical capabilities of GIS. Hodsen, A. J., Tranter, M., Dowdeswell, J. A., Gurnell, A. M. & Hagen, J. O. Goodchild, M. F., Parks, B. O. & Steyaert, L. T. (eds) Environmental 1997. Glacier thermal regime and suspended-sediment yield: a comparison Modeling with GIS, 56–74. Oxford University Press. of two high-Arctic glaciers. Annals of Glaciology 24, 32–37. Bjo¨rnsson, H. 1988. Hydrology of Ice Caps in Volcanic Regions. Dr. Philos. Holmlund, P. & Eriksson, M. 1989. The cold surface layer on Storglacia¨ren. Thesis. Reykjavik, University of Iceland. Geografiska Annaler 71A, 241–244. Bjo¨rnsson, H., Gjessing, Y., Hamram, S. E., Hagen, J. O., Liestøl, O., Hutchinson, M. F. 1989. A new procedure for gridding elevation and stream Palsson, F. & Erlingsson, B. 1996. The thermal regime of subpolar glaciers line data with automatic removal of spurious pits. Journal of Hydrology mapped by multi-frequency radio echo sounding. Journal of Glaciology 106, 211–232. 42, 23–32. Isaksen, K., Vonder Mu¨hll, D., Gubler, H., Kohl, T. & Sollid, J. L. 2000. 168 B. Etzelmu¨ller et al. NORSK GEOGRAFISK TIDSSKRIFT 54 (2000)

Ground surface temperature reconstruction based on data from a deep Paterson, W. S. B. 1994. The Physics of Glaciers. Pergamon/Elsevier borehole in permafrost at Janssonhaugen, Svalbard. Annals of Glaciology Science, Oxford. 31. In press. Pickard, J. 1983. Surface lowering of ice-cored moraine by wandering lakes. Kamb, B. & Echelmeyer, K. A. 1986. Stress-gradient coupling in glacier Journal of Glaciology 29, 338–342. flow: I. Longitudinal averaging of the influence of ice thickness and Pickard, J. 1984. Retreat of ice scarps on an ice-cored moraine, Vestfold surface slope. Journal of Glaciology 32, 267–284. Hills, Antarctica. Zeitschrift fu¨r Geomorphologie 28, 443–453. King, L. & Volk, M. 1994. Glaziologie und Glazialgeomorphologie des Pulina, M. 1984. Glacier karst phenomena in Spitsbergen. Norsk Geografisk Liefde- und Bockfjordgebietes, NW-Spitzbergen. Zeitschrift fu¨r Geomor- Tidsskrift 38, 163–168. phologie, Suppl. Bd 97, 145–160. Pulina, M. & Rehak, J. 1991. Glacier caves in Spitsbergen. Proceedings. 1st Lawsen, D. E. 1993. Glaciohydrologic and glaciohydraulic effects on runoff International Symposium of Glacier Caves and Karst in Polar Regions, and sediment yield in glacierized basins. CRREL Monograph. 93-2. Madrid, 93–117. Liestøl, O. 1977. Pingos, springs, and permafrost in Spitsbergen. Norsk Rudberg, S. 1988. High arctic landscapes: comparison and reflections. Norsk Polarinstitutt A˚ rbok 1975, 7–29. geografisk Tidsskrift 42, 255–264. Liestøl, O. 1988. The glaciers in the Kongsfjorden area, Spitsbergen. Norsk Ro¨thlisberger, H. 1972. Water pressure in intra- and subglacial channels. Geografisk Tidsskrift 42, 231–238. Journal of Glaciology 11, 177–203. Shreve, R. L. 1972. Movement of water in glaciers. Journal of Glaciology 11, Liestøl, O. 1994. Kompendium i glasiologi. Rapportserie i naturgeografi 3. 205–214. Geografisk institutt, Universitetet i Oslo. Sollid, J. L. & Sørbel, L. 1988a. Influence of temperature conditions in Major, H. & Nagy, J. 1972. Geology of the Adventdalen area. Norsk formation of end moraines in Fennoscandia and Svalbard. Boreas 17, 553– Polarinstitutt Skrifter 138. 558. Mardia, K. V. 1972. Statistics of Directional Data. Academic Press, London. Sollid, J. L. & Sørbel, L. 1988b. Utbredelsesmønsteret av løsmateriale og Mark, D. M. 1975. Geomorphometric parameters: a review and evaluation. Landformer pa˚ Svalbard – noen hovedtrekk. Norsk Geografisk Tidsskrift Geografiske Annaler 57A, 165–177. 42, 265–270. Nye, J. F. 1965. The flow of a glacier in a channel of rectangular, elliptic or Tomlin, C. D. 1990. Geographic Information Systems and Cartographic parabolic cross-section. Journal of Glaciology 5, 661–690. Modeling. Prentice Hall, Englewood Cliffs, N.J. Ødega˚rd, R., Hamran, S. E., Bø, P. H., Etzelmu¨ller, B., Vatne, G. & Sollid, J. Vatne, G. 1997. Glacial hydrology of the polythermal glaciers Erikbreen and L. 1992. Thermal regime of a valley glacier, Erikbreen, northern Hannabreen, Liefdefjorden, northern Spitsbergen, Svalbard. Dr. scient Spitsbergen. Polar Research 11, 69–80. thesis. Department of Physical Geography. Rapportserie i naturgeografi 7. Ødega˚rd, R., Etzelmu¨ller, B., Vatne, G. & Sollid, J. L. 1995. Near-surface Vatne, G., Etzelmu¨ller, B., Ødega˚rd, R. & Sollid, J. L. 1995. Subglacial spring temperatures in an Arctic coastal rock cliff: possible implications drainage and sediment evacuation at Erikbreen, nothern Spitsbergen. for rock breakdown. Slaymaker, O. (ed.) Steepland Geomorphology, 89– Nordic Hydrology 26, 169–190. 102. Wiley, Chichester. Weertman, J. 1961. Mechanism for the formation of inner moraines found Ødega˚rd, R. S., Hagen, J. O. & Hamram, S.-E. 1997. Comparison of radio- near the edge of cold ice caps and ice sheets. Journal of Glaciology 3, 965– echo sounding (30–1000 MHz) and high-resolution borehole-temperature 978. measurements at Finsterwalderbreen, southern Spitsbergen, Svalbard. Zevenbergen, L. W. & Thorne, C. R. 1987. Quantitative analysis of land Annals of Glaciology 24, 262–271. surface topography. Earth Surface Processes and Landforms 12, 47–56.