Chemical denudation in arctic-alpine Latnjavagge (Swedish Lapland) in relation to regolith as assessed by radio magnetotelluric-geophysical profiles

Geomorphology 57 (2004) 303 – 319 www.elsevier.com/locate/geomorph

Chemical denudation in arctic-alpine Latnjavagge (Swedish Lapland) in relation to regolith as assessed by radio magnetotelluric-geophysical profiles Achim A. Beylich a,*, Else Kolstrup a, Tage Thyrsted b, Niklas Linde c, Laust B. Pedersen c, Lars Dynesius c a

Department of Earth Sciences, Environment and Landscape Dynamics, Uppsala University, Villava¨gen 16, Uppsala SE-752 36, Sweden b Harbacken-Stavby, Alunda, Sweden c Department of Earth Sciences, Geophysics, Uppsala University, Uppsala, Sweden Received 13 December 2002; received in revised form 18 March 2003; accepted 21 March 2003

Abstract In Latnjavagge, a 9-km2 drainage basin with homogeneous lithology in periglacial northern Swedish Lapland, water balance, water chemistry and radio magnetotelluric geophysical investigations along selected profiles were integrated with assessment of regolith thickness as well as of ground frost conditions within the basin. In combination with direct field observations, the geophysical profiles demonstrated presence of relatively thin regolith in most of the investigated area, yet in some parts, the bedrock was located deeper and locally was not detected at 40-m depth. TDS values of the water were generally very low. The areas that contributed with the lowest ion concentrations were cold and had a thin regolith, whereas there were higher concentrations in water that drained radiation exposed slopes with earlier thaw and thicker regolith. The low resistivities found along the profiles in the geophysical investigations in combination with the relatively higher TDS values found in related runoff and subsurface water samples showed that larger volumes of ice-rich frozen ground were not found along the investigated profiles in late August. D 2003 Elsevier B.V. All rights reserved. Keywords: Chemical denudation; Regolith thickness; Frozen ground; Radiomagnetotellurics; Periglacial; Lapland

1. Introduction For many years, chemical weathering and denudation were believed to be of relatively minor importance in periglacial environments (e.g. Campbell et

* Corresponding author. Tel.: +46-18-471-2511; fax: +46-18555-920. E-mail address: [email protected] (A.A. Beylich). 0169-555X/$ - see front matter D 2003 Elsevier B.V. All rights reserved. doi:10.1016/S0169-555X(03)00162-4

al., 2001). Von Lozinski (1909, 1912) was the first to postulate a strong dominance of mechanical over chemical weathering for this type of environment and also, Peltier (1950) suggested that there was only limited chemical weathering in cold environments. In contrast, Rapp (1960) found that chemical weathering contributed to about half of the total denudation in Ka¨rkevagge, northern Swedish Lapland. Later studies in the same area (Thorn, 1975; Dixon et al., 1995; Darmody et al., 2000; 2001; Campbell et al., 2001,

304

A.A. Beylich et al. / Geomorphology 57 (2004) 303–319

2002; Thorn et al., 2001) have also found high chemical denudation values. Except for Ka¨rkevagge, comparatively few investigations of chemical denudation in periglacial areas have been done over the years. Yet, there are studies, e.g. in alpine Colorado (Caine, 1979, 1995; Caine and Thurman, 1990), the northern Cascade Mountains (Reynolds and Johnson, 1972), Alaska (Dixon et al., 1984) and eastern Iceland (Beylich, 1999, 2000a,b,c) according to which it seems as if chemical weathering and denudation are significant processes in cold environments. However, dilution in the Ka¨rkevagge drainage basin seems to be particularly high, possibly due to local lithological factors (Thorn et al., 2001; Campbell et al., 2002). To cast further light upon present denudation rates as well as the mutual relationship between mechanical and chemical weathering and denudation in periglacial environments, an investigation was started in 1999 in Latnjavagge, a representative drainage basin of the arctic – oceanic mountain area in northernmost Swedish Lapland (Beylich, 2001a,b, 2003a,b; Beylich and Gintz, in press; Beylich et al., submitted for publication (a), submitted for publication (b). In estimating chemical denudation rates, several factors play a major role. Some of these, such as lithology, slope angle and climate are reasonably well known for the area, but it soon became clear that knowledge of regolith thickness and distribution of permafrost was needed since they largely determine the contact surface between water and particles during subsurface flow. To obtain information on regolith thickness and frozen ground conditions within the Latnjavagge drainage basin, an investigation was conducted in August 2001, using a newly developed radio magnetotelluric instrument, EnviroMT (Bastani, 2001). Theoretically, electromagnetic techniques like this provide a good resolution of electrically conducting sediments. Practically, the RMT technique is favourable for studying the geometry of the regolith and variations therein because it is fast and has low power consumption owing to the fact that existing radiotransmitters are used as a source, the latter being of particular importance in isolated areas like northern Sweden. In this paper, selected geophysical profiles and related runoff and TDS data from Latnjavagge are presented and used with assessment of regolith thick-

ness and ground frost conditions in the associated areas.

2. The study area The Latnjavagge drainage basin is located in northernmost Swedish Lapland at 68j20VN, 18V30VE (Figs. 1 and 2). It covers approximately 9 km2 with a length in N – S direction of 4600 m, and elevation ranging from 950 to 1440 m asl. The bedrock consists of metamorphosed Ordovician– Silurian deposits forming part of the Caledonian Ko¨li Nappe (Stephens et al., 1994). Garnet mica schists dominate, thus, giving the catchment a rather homogeneous lithology, but also graphitic schists and amphibolites have been mapped in the area (Holmquist, 1910; Kulling, 1964). Also, in the northern part of the valley, inclusions of marble and acidic granites are found (Kling, 2003). Structurally, the rocks are gently dipping towards the north. Tectonic faults dissect the area and can be related to shearing of the bedrock; for example, the abrupt change of direction in the valley north of Latnjajaure (Fig. 2) might be explained by tectonic faults (Kling, 2003). The lower part of the valley floor at ca. 980 m asl is dominated by the 0.73 km2 large and 43 m deep lake, Latnjajaure, which is surrounded by both gentle and steep slopes up to 1440 m asl. (Fig. 2). The basin has its outlet to the south into the lower lying and larger valley, Ka˚rsavagge. During the Weichselian (and presumably also during previous glaciations), the valley became glacially sculptured with the main axis of glacier movement in a N –S direction. After deglaciation during the early Holocene (Andre´, 1995) glacially eroded, over-steepened rock walls remained around the much flatter central part of the basin, so that today, a relatively flat infilled valley floor is found between altitudes of 980 and 1020 m surrounded in most parts by steeper slopes for the next 200 –300 m before more gentle slopes are found in the highest parts of the area. The transition between the flatter bottom part and the higher lying steeper slope is sharp. Locally, both in the northern and southern parts of the basin relatively flat areas can also be found above 1020 m. Along steeper slope segments exposed bedrock is usually found alternating, both vertically and laterally, with slopes of accumulated mass movement products. The

A.A. Beylich et al. / Geomorphology 57 (2004) 303–319

305

Fig. 1. Location of the study area in northernmost Swedish Lapland.

regolith in the flatter valley bottom consists of a stony, poorly sorted mixture of sediments resulting from a combination of process types: along and beneath steeper parts of the slopes mass movement and slope wash products are important; in the flattest parts, fluvial sediments interfinger with deposits resulting from mass movement, in particular from slush flows, as well as with glacial deposits. The latter include series of transverse moraine ridges, which stand out in the terrain across the valley also in the lowest and flattest part of the basin. The soils are mainly regosols and lithosols. Where relatively fine-grained sediments are present, surfaces of flatter areas usually show patterned ground with sorted polygons. Permafrost has not been

ascertained within the basin because of difficulties of coring but in exposed areas with only a thin or shortlived snow cover the temperature conditions are sufficiently cold to allow for it (see below) and besides, a coring 6 km north of Latnjavagge demonstrated permafrost down to 80 m below the ground surface at 1200 m asl (Kling, 1996, p.5; 2003). Meteorological data have been collected by an automatic weather station at the Latnjajaure Field Station (LFS) (Fig. 2) at 981 m asl since April 1992 (Molau, 2001, 2003). The mean annual air temperature at LFS is 2.3 jC (1993 – 2001) with a mean of the warmest month (July) of 8.0 jC and of the coldest (February) of 10.1 jC. The mean annual precipitation at LFS is 818 mm (1990 – 2001) of which about

306

A.A. Beylich et al. / Geomorphology 57 (2004) 303–319

Fig. 2. Locations of sampling sites and geophysical profiles within the Latnjavagge drainage basin.

66% is temporarily stored as snow during the winter. Of the summer months June – August, August has the highest mean precipitation, 82 mm, and also the highest frequency of extreme rainfall events. The hydrological regime is thus nival with runoff limited to the period between mid/end May to October/ November (Beylich, 2003b). Parts of the valley are

covered by perennial snow and ice patches (Fig. 2). Winds predominantly come from northern and western directions and the climate can be characterised as arctic– oceanic. The vegetation in the area is a continuous mosaic of dwarf-shrub heath, alpine meadows and bogs and belongs to the mid-alpine zone (Molau, 2001; Molau

A.A. Beylich et al. / Geomorphology 57 (2004) 303–319

et al., in press). Direct anthropogenic impact on the environment is limited to extensive reindeer grazing, some hiking tourism, and research at LFS, yet, altogether, direct human impact on the natural system is small (Beylich et al., submitted for publication (c).

3. Climatic conditions for 2000, 2001 and 2002 The periglacial mountain area of northernmost Swedish Lapland is characterized by some interannual variability of the climate. Most of the data presented in this paper was collected during the field season of 2001, which was a year characterized by little snow (little winter precipitation) before the beginning of snowmelt in early June, a main snowmelt period until end of June, much precipitation and low air temperatures in July (108.2 mm, mean of 7.43 jC), much precipitation in August (107.3 mm) and a snow storm event in the beginning of that month. Compared to that, the 2000 field season was characterized by normal snow conditions (see also above) before the beginning of snowmelt in late May, a long lasting snowmelt period until late July, very little precipitation in July (46.1 mm) and much precipitation in August (110.5 mm). In turn, the 2002 season had normal snow conditions before a very early and quick snowmelt with high air temperatures and radiation inputs in late May and June (monthly mean in June 7.3 jC), high air temperatures in July (mean 8.9 jC), a very high mean air temperature for the area in August with 10.4 jC, which was the highest mean since the start of the temperature measurements at LFS in 1990; there was very little precipitation in May (37.6 mm), June (26.0 mm) and August (42.0 mm) but relatively more in July (90.2 mm) (Beylich et al., submitted for publication (c)).

4. Approach and methods The main part of the fieldwork for this paper was conducted between the 28th of May and the 19th of August, 2001. The period was chosen so that it covers the runoff season from its start through most of the summer. The geophysical profiles were made from the 11th to the 18th of August 2001 at the end of the field

307

season at a time when annual frost in the ground had time to melt in as far as possible before the risk of snowfall had become too high. To obtain a more representative record for the area and thus eliminate the effect of extreme events of 2001, the mean seasonal TDS values of different subcatchments within Latnjavagge and chemical denudation rates have been calculated on the basis of daily measurements during the summers of 2000, 2001 and 2002. Discharge measurements were done three times daily by an Ott-propeller C2 at the outlet of the Latnjavagge drainage basin and selected subcatchments within Latnjavagge as well as at the inlet and outlet of the lake, Latnjajaure (Fig. 2). Daily runoffs were calculated by interpolation of the measurements (see Beylich, 1999) and specific discharges [mm/day] were calculated. Installation of fixed gauge stations was not possible because of the characteristics of the channels (bedrock and/or blocks, shifting channels during snowmelt, slush flows) and the remote location of the measuring sites (Beylich et al., submitted for publication (b)). Immediately after the discharge measurements, the electric conductivity (AS/cm; reference 25 jC) of the surface water was measured with a temperature-corrected portable instrument and subsequently, daily discharge weighted mean values were calculated (Beylich, 1999). For practical reasons and also because the temporal variability of the conductivity values was rather small, three measurements per 24 h were done and regarded as a good approximation to a more continuous recording series. At the start of the field campaign, snow samples were collected in a plastic tube (10 cm diameter) from the previous winter snow along chosen profiles (Fig. 2) and melted in buckets at LFS (Beylich et al., submitted for publication (c)). Total dissolved solids (TDS) [mg/l] in surface water and in precipitation samples, as well as in the melted snow samples were estimated by multiplying the electric conductivity values by 0.7 (see Stro¨mquist and Rehn, 1981; Darmody et al., 2000), a factor that is representative for Ka¨rkevagge a few kilometers to the NW and it was therefore expected to provide reliable values for Latnjavagge, too. Ion concentrations from surface water, precipitation and annual snow pack were analysed later (see also Beylich et al., submitted for publication (a)). Snow depths at grids A – F (Fig. 2) were provided by Latnjajaure Field Station

308

A.A. Beylich et al. / Geomorphology 57 (2004) 303–319

(Molau, 2001, 2003; Molau et al., in press). The snow depth data are not used in detail in this paper but they provide background information with the discussion of duration of snow cover and ground frost in subareas within Latnjavagge. By means of a 1.5-m long steel rod, it was attempted to detect the depth to the frozen layer in the area and to possibly assess the presence of permafrost but there were too many stones to rely on the method. The geophysical profiles (see Figs. 2 and 3) were chosen so that they are close to and can be related to water measurement and sampling sites for integration of information. The radio magnetotelluric (RMT) method makes use of the electromagnetic field from distant radio transmitters working in the frequency range 10 –250 kHz. The EnviroMT system (Bastani, 2001) is a modern implementation of the method. The earth’s response is given by unique transfer functions (the impedance tensor) relating the horizontal electric field to the horizontal magnetic field for a given frequency. By measuring these transfer functions for a number of different frequencies, it is possible to estimate the distribution of electrical resistivity in the ground as a function of depth and profile direction. A comprehensive description of EM geophysics is found in Nabighian (1987, 1991). Inversion is the process of finding a model that fits the data to a certain level under certain constraints (e.g. smoothness). Even though the earth is generally 3-D, 2-D or even 1-D interpretations are often used because of a lack of data or negligible 3-D/2-D effects. Five profiles were measured, of which line 4 is a continuation of line 3. Normally, 10 – 15 transmitter stations were received. The spacing between measurement points was 10 m. A TSVD processing scheme (Bastani and Pedersen, 2001) was used to estimate reliable transfer functions based on the data. Fig. 4 shows the apparent resistivities and phases of the determinant data, which clearly indicate a resistor overlaying a strong conductor with considerable variations in the depth to the conductor. A 2-D inversion was carried out using the determinant of the impedance tensor because it is believed to be less affected by 3-D effects than other choices for inversion, using the Rebocc code (Siripunvaraporn and Egbert, 2000). The root mean square (rms) data fits of the resulting models are shown in Fig. 5. The data fits are generally high, giving confidence to the models, except for the

beginning of line 1 and a few areas indicated by high misfit for certain stations. The resulting models are shown in Fig. 3a – e. Since it was not possible to relate these profiles to surface elevation for an easy readable relation to topography, hand-made profiles (Fig. 3f– j) combine altitude along the profiles based upon contour maps with the location of the transition to the 0.5 exponent (i.e. a resistivity of ca. 3 V-m). The RMT method is ideal when the target is a strong conductor. However, the resolution decreases with a decrease in resistivity contrast. DC geoelectric

Fig. 3. Geophysical profiles (a – e) and the location of the 0.5 exponent (i.e. a resistivity of ca. 3 V-m) in relation to altitude along the profiles (f – j). Lines 2 and 5 are oriented with west to the left and east to the right, line 1 left side is to the south and lines 3 and 4 are oriented with the left hand side to the north.

A.A. Beylich et al. / Geomorphology 57 (2004) 303–319

309

Fig. 3 (continued).

measurements are a better choice when studying resistive features (e.g. permafrost), but it would be difficult to get contact with the ground in this rocky terrain. Refraction seismics using explosives as the source is an alternative for mapping the bedrock. However, the RMT method is non-invasive and measurements can be made by non-geophysicists. Ground Penetrating Radar (GPR) would not penetrate the regolith and the interpretation would be difficult because of scattering from large boulders.

5. Results and interpretation 5.1. Geophysical investigations Geophysical Lines 2 and 5 (see Figs. 2 and 3) are located across the valley that widens downstream and Lines 1 and 3 + 4 are located in the gently sloping/ undulating length direction of the valley. The inverted models show a top resistor (higher than 100 V-m and generally lower than 300 V-m,

310

A.A. Beylich et al. / Geomorphology 57 (2004) 303–319

ductor are, therefore, only artefacts of the initial model.

Fig. 4. The apparent resistivities and phases of the determinant data. The left column shows the apparent resistivities for lines 1 – 5 (a, c, e, g and i) and the right column shows the phases for lines 1 – 5 (b, d, f, h and j).

even though some areas of higher resistivity occur). An extremely conductive basement (less than 3 V-m) is shown for all profiles, but it is absent between 170 and 210 m at Line 3 and along parts of Line 2 (e.g. between 170 and 270 m). The transition from the resistor to the conductor is very sharp to be a smooth 2-D model (in the inversion process, the data misfit is minimised while keeping the model smooth), and no intermediate layer (the green areas above the red) should be interpreted. The upper layer is not homogeneous, most clearly seen at Lines 3 and 4, where more conductive regions are interbedded in the resistor. Features below the conductor should not be interpreted, as the EM-waves are strongly alternated by the conductor. The more resistive region below the con-

5.1.1. Interpretation of the geophysical data There are only few natural materials with so low resistivity as found at deeper levels in the profiles. It is possible that the low resistivities represent a very clayrich material, such as a saprolite from strongly and deeply weathered mica shist. Another possibility is that it represents graphitic schist intermixed with garnet-mica schist, or maybe even a combination of the two explanations. Considering the distribution of outcropping bedrock in the area in relation to the location of the geophysical profiles, such as it is seen for example in parts of line 3 and in particular in the western part of Line 2 (Fig. 3) where bedrock outcrops at the start of the profile and alternates with what seemed to be a very thin debris cover in the westernmost part, it is suggested that the low-resistivity base represents the transition between regolith and bedrock. The lateral continuity of resistivity in Lines 1, 4 and 5 suggests that there are no major gaps in the underlying bedrock along these lines, but in some places there may have been erosion across the profile, for example at 230 m across Line 4 where there is a brook. The large discontinuity in profile 2 probably represents the erosional activity of glaciers (the depth of the bottom of the lake is beneath the deepest part of this profile) possibly in combination with that of meltwater, but the profile is too shallow to ascertain whether the erosion has followed a previous bedrock lineament/zone of weakness in continuation of the contour-line knick to the southeast of this profile. Permafrost has a high resistivity as it is also shown by, e.g. Dramis et al. (1997) for Swiss rock glaciers and Isaksen et al. (2000) for layers interpreted as icesaturated (5– 70 kV-m) and ice-rich (100 – 900 kV-m) in Svalbard. Thus, the resistivity model from Latnjavagge suggests that ice-rich permafrost is not present along the geophysical profiles, not even in the western part of Line 2, which is at 1030 m asl in the shadow of a steep east-facing slope, and the likelihood of dry permafrost seems little in this moist environment (see also below). According to the table in Reynolds (1997, p. 422f), very dry clay has 50 – 150 V-m whereas the resistivity in wet clay can be as low as 1 V-m. Dry gravel has a

A.A. Beylich et al. / Geomorphology 57 (2004) 303–319

311

Fig. 5. The rms data fit for the 2-D models. The data fit of the apparent resistivities (a) and phases (b) are shown along the profiles. Error floors of 2.5j and 10% on the apparent resistivities were used.

resistivity of 1400 V-m whereas its resistivity is typically 100 V-m when saturated; and for sand, the values are between those of clay and gravel. As a consequence, when seen in combination with field observations, it is likely that the sediments within the area generally have high water contents, with the possible exception of a few well-drained sites, primarily in the upper layers, as suggested by the highest resistivity values being within the upper few metres of the profiles. This interpretation would mean that the subsurface flow from areas around Latnjajaure takes place within a few metres of mainly water-rich sediments without permafrost. As the lake is 43 m deep, the runoff and subsurface flow from the slopes, and also the water crossing Line 2, can be expected to pass through the lake at some time. 5.2. Geophysical data and information on runoff and water chemistry The geophysical investigation was meant to support the interpretation of discharge and TDS within the Latnjavagge drainage basin. The basis for the assessment is that the bedrock is rather homogeneous over the whole area and the precipitation can also be assumed to be uniform. Differences in TDS values and ion concentrations within various parts of the drainage basin could therefore to a large

extent be expected to reflect differences in local conditions including regolith thickness and frozen ground conditions. 5.3. Precipitation, runoff and TDS values Fig. 6a shows daily air temperatures measured at the Latnjajaure Field Station (LFS) between May 2001 and August 2001 and Fig. 6b gives the precipitation at LFS for the same period together with daily specific runoffs at the outlet of the Latnjavagge catchment. The precipitation comes mainly as rain, but occasionally also as snow (snow storm event in the beginning of August) and taken for the whole recording period, the mean weighted TDS value of the precipitation is 4.7 mg/l (Table 1) (Beylich et al., submitted for publication (b)). The daily discharges at the outlet of Latnjavagge were low at the start of the field season but increased rapidly once the daily minimum temperature exceeded 0 jC and there was a peak of discharge in June when the main snowmelt period occurred. In July, the discharge became more dependent on precipitation, but during high-temperature spells, melting of the remaining snow was accelerated. At the end of the season, precipitation is clearly the steering factor for discharge values even if the precipitation can be stored as snow for a few days, and even if there are some permanent snow and ice patches.

312

A.A. Beylich et al. / Geomorphology 57 (2004) 303–319

Fig. 6. Daily air temperatures (min., max. and mean) (a) and daily precipitation (b) at the Latnjajaure Field Station and daily discharges (b) at the outlet of the Latnjavagge drainage basin between May and August 2001.

Daily specific runoffs [mm/day], daily TDS values [mg/l] and daily yields of dissolved solids [kg/km2/ day] at the inlet and outlet of Latnjajaure, for subcatchment A and at the outlet of the entire Latnjavagge drainage basin (see Fig. 2), are given in Fig. 7a– d. The inlet and outlet of Latnjajaure have the most pronounced discharge peaks during the snowmelt period closely followed by the values from the whole basin whereas, owing to the smaller amount of snow at the beginning of the monitoring campaign, the peaks are more modest and short lived in subcatchment A (Beylich, 2003b). Particularly interesting

are the TDS values: at the inlet of Latnjajaure, there is only little change over the summer (Fig. 7a), and during snowmelt the values are only slightly lower (6 – 8 mg/l) than during the remaining part of the season (7– 10 mg/l). At the outlet of Latnjajaure, the water has about the same concentrations as that of the inlet during the 2– 3 week snowmelt period, but both before and after that time, the values are fairly constant at a higher level of ca. 12 –13 mg/l (Fig. 7b). In subcatchment A, the concentrations are between ca. 10 and 13 mg/l during the snowmelt period, values

A.A. Beylich et al. / Geomorphology 57 (2004) 303–319

313

Table 1 Water chemistry data from selected sites in Latnjavagge, 2001 field season Sample site

6 n=9 7 n = 10 3 n = 10 5 n=9 Prec. n = 10 Snow cores n=8

Description

Outlet SC A

Outlet Latnjavagge Inlet Latnjajaure Outlet Latnjajaure LFS

Profiles

GPS Position

68j21.094VN 18j29.735VE 972 m asl 68j20.973VN 18j29.827VE 956 m asl 68j22.231VN 18j29.278VE 1000 m asl 68j21.263VN 18j29.558VE 981 m asl

TDS (mg/l)

Ca (mg/l)

Mg (mg/l)

Na (mg/l)

K (mg/l)

Fe (mg/l)

Mn (mg/l)

Cl (mg/l)

NO3 (mg/l)

SO4 (mg/l)

PO4 (mg/l)

Mean

Mean

Mean

Mean

Mean

Mean

Mean

Mean

Mean

Mean

Mean

Min

Min

Min

Min

Min

Min

Min

Min

Min

Min

Min

Max

Max

Max

Max

Max

Max

Max

Max

Max

Max

Max

15.46 9.10 26.95 12.14 8.26 18.20 7.44 5.60 10.36 10.67 5.46 14.07 4.67 2.10 14.84 3.64

2.00 1.60 3.00 1.70 1.30 2.10 0.98 0.59 2.10 1.58 1.20 1.90 0.39 0.00 1.09 0.10

0.56 0.39 0.91 0.40 0.32 0.55 0.17 0.13 0.23 0.34 0.24 0.46 0.05 0.02 0.11 0.01

0.52 0.46 0.64 0.66 0.50 0.88 0.57 0.35 0.86 0.61 0.45 0.76 0.32 0.06 0.81 0.17

0.82 0.40 2.66 0.78 0.38 1.82 0.78 0.24 2.72 0.68 0.32 1.74 0.79 0.14 4.02 0.07

0.00

0.00

0.00

0.00

0.00

0.02 0.00 0.10 0.01 0.00 0.10 0.03

0.00

0.04 0.0 0.1 0.03 0.0 0.2 0.07 0.0 0.2 0.1 0.0 0.5 0.40 0.0 1.4 0.27

5.2 3.7 6.8 4.5 3.4 5.5 2.1 1.2 2.7 4.5 3.4 5.4 1.0 0.5 1.4 0.3

0.0

0.00

0.9 0.6 1.2 1.1 0.6 1.6 1.0 0.5 1.7 1.3 0.7 2.6 1.1 0.4 2.6 0.7

3.21 4.26

0.05 0.20

0.01 0.01

0.12 0.20

0.03 0.08

0.00 0.10

0.5 0.9

0.0 0.6

0.2 0.4

that more or less remain the same until mid-July when they gradually increase until the start of August to values well above 20 mg/l (Fig. 7c). During the rest of August, concentrations remain relatively high for this subcatchment although they are lower than during the maximum. For the drainage basin as a whole, this combination of the discharges from the different subareas adds up to 9 –10 mg/l for most of the snowmelt period (Fig. 7d), but after that, the values increase, with fluctuations, to between 14 and 18 mg/l in August. Altogether, the ranges of temporal variation in solute concentration are comparatively small, which is common in periglacial fluvial systems (Clark, 1988). A comparison of the mean discharge weighted TDS values for 2001 with the values from the previous and succeeding years (Table 2) shows that in spite of the differences in precipitation, snow cover and temperature between the years as outlined above, the TDS values show only little interannual variations at the same locality during the 3 different years (Beylich et al., submitted for publication (b)) and the year 2001 can therefore be taken to be representative.

0.00

0.00

0.0

0.0

0.0

0.02 0.0 0.4 0.0

The varying mean TDS values between different localities reflect differences in the chemical composition of the water. Table 1 gives mean concentrations of different ions in surface water, precipitation and snowpack as well as at selected sample sites. In addition, their minimum and maximum values are given to indicate the range of values. At the inlet of Latnjajaure, there are relatively lower ion concentration values of Ca2 +, Mg2 + and SO42 and higher concentrations of NO3 than at the outlet of the lake and also as compared with subcatchment A (see also Beylich et al., submitted for publication (a)). A comparison of these values with those of rainwater and snow shows that the water in the upper part of the basin is more similar in its composition to that of rainwater than the water further downstream. The outlet of the basin represents a mean for the whole area. The mean annual chemical denudation rates based on data from 2000, 2001 and 2002 are 2333 kg/km2 year at the inlet of Latnjajaure (sampling site 3 in Fig. 2), 4554 kg/km2 year at the outlet of Latnjajaure (site 5), 7894 kg/km2 year at the outlet of subcatchment A (site 6) and for the entire Latnjavagge drainage basin,

314

A.A. Beylich et al. / Geomorphology 57 (2004) 303–319

the value is 5382 kg/km2 year (site 7) (for further detail, see Beylich et al., submitted for publication (b)). 5.3.1. Interpretation of runoff and TDS values The differences in TDS values over the Latnjavagge area and during the summer months point to local differences in chemical denudation conditions within the catchment and over the season, yet, in all areas presented here, the concentrations are higher

than the mean salt concentration in the precipitation (4.7 mg/l) at any time of the season. As compared to the salt concentrations before and after snowmelt, the snowmelt discharge peak generates a diluting effect that can be recognised at all sites although it is relatively weak at the inlet of the lake and in subcatchment A. Nevertheless, days with higher discharges generated by snowmelt also show higher yields of dissolved substances, thus the lower concentrations of dissolved salts caused by diluting effects during thaws

Fig. 7. Daily specific runoff, TDS values and yields of dissolved solids (a) at the inlet of Latnjajaure (sampling site 3) and (b) at the outlet of Latnjajaure (sampling site 5), (c) the outlet of subcatchment A (sampling site 6) and at (d) the outlet of the Latnjavagge drainage basin (sampling site 7) (2001 field season).

A.A. Beylich et al. / Geomorphology 57 (2004) 303–319

315

Fig. 7 (continued).

are more than compensated for by the increased runoffs (see Collins and Young, 1981; Walling and Webb, 1983; Barsch et al., 1994; Gude et al., 1996). During the main peak of the snowmelt, the concentrations are particularly low because the underlying ground is still frozen. The frozen ground prevents infiltration of ion poor melt- and rainwater and the associated contact between water and the regolith is highly reduced (Beylich et al., submitted for publication (a), submitted for publication (b)). In contrast, during rainfall generated peaks later in the season the rainwater could percolate through the (partly) unfrozen ground and no

comparable diluting effects were noted. Relatively, high salt concentrations at the beginning of the main snowmelt period due to an ionic pulse (e.g. Johannessen and Henriksen, 1978), as reported by several authors from different periglacial environments (Leser et al., 1992; Potschin and Leser, 1994; Gude et al., 1996), could also be observed in Latnjavagge. In the area above, the inlet of Latnjajaure, the constant concentration values over the post-snowmelt season, suggest that the chemical denudation conditions did not vary much over the summer. In this area, the main part of the water comes directly from the

316

A.A. Beylich et al. / Geomorphology 57 (2004) 303–319

Table 2 Precipitation and mean discharge weighted TDS values at different sites during the field seasons of 2000, 2001 and 2002 Field season

Catchment

Precipitation (mm)

Mean TDS (mg/l)

26/05/2000 – 31/08/2000

Latnjavagge Outlet Lake Inlet Lake Subcatchment Subcatchment Subcatchment Subcatchment Latnjavagge Outlet Lake Inlet Lake Subcatchment Subcatchment Subcatchment Subcatchment Latnjavagge Outlet Lake Inlet Lake

226

11.9 11.1 8.3 15.5 10.7 6.8 6.6 12.1 10.7 7.4 15.5 9.4 6.3 6.6 12.0 10.9 7.9

29/05/2001 – 18/08/2001

28/05/2002 – 31/08/2002

A B C D 264

A B C D 158

plateau areas around 1300 m asl (Fig. 2) where there is only a very thin regolith cover or it is completely lacking. Furthermore, the area has larger permanent snow and ice patches and snow fields as well as a partly sheltered position from radiation. It is therefore likely that the ground remained frozen in most of this subcatchment during the field season. Consequently, the low concentration values of this area (see, e.g. discharge weighted TDS values from the inlet to Latnjajaure and at the outlets of subcatchments C and D in Table 2) are thought to reflect a combined result of thin regolith and frozen ground conditions, i.e., there is only little contact and reaction time between the ion poor melt- and rainwater and the regolith during the runoff season. In subcatchment A, the clearly highest weighted TDS values within the basin were found. Even at the season of lowest concentrations, the values are higher than the highest values recorded at the inlet to Latnjajaure. This subcatchment is exposed to intensive radiation and the snowmelt period is up to 1.5 months earlier and also much shorter than in other parts of the basin. At the start of the postsnowmelt season, the concentrations are still only moderate so probably there was still frozen ground during the early part of July, but the higher values later during the summer suggest that the ground

gradually melted and gave way to deeper infiltration of water into the ground and further, the relatively high concentrations point towards a fairly thick regolith of this subcatchment. The area above the inlet to Latnjajaure and subcatchment A thus form two contrasting situations. At the outlet of Latnjajaure, the low concentrations during snowmelt suggest a strong influence from snowmelt water during a few weeks before slightly higher but constant concentrations around 12 mg/l take over from the start of July for the remaining part of the season. This is interpreted as the result of outflow from a well-mixed, large water body. A comparison between concentration values from the inlet and the outlet of the lake shows that, apart from the snowmelt season, the concentrations are highest at the outlet thus there must have been an addition of salts brought into the lake by surface runoff and subsurface flow from the western and eastern slopes along the lake. So the difference between in- and outlet represents dilution of the regolith on the slopes, most likely with a strong dominance of salts from the eastern slope, which is snow-free about 1 –1.5 half months earlier than the western one, and in addition, it is characterised by both a much earlier thawing of ground frost and a thicker regolith as compared to the very steep western one (Beylich et al., submitted for publication (a)). Taken as a whole, the TDS values of the surface water in Latnjavagge are low. In some subareas above the delta they are not much higher than that of the precipitation that falls upon it and water from such subareas contribute to the relatively low value at the inlet to the lake. As a consequence, dissolution of regolith and bedrock is particularly low in such subareas (Beylich et al., submitted for publication (b)). On the other hand, TDS values from subcatchment A for the time after mid-July are of a sufficient magnitude to be clearly reflected in the total outflow.

6. Discussion and conclusions Geophysical data in combination with water discharge and water chemistry data can give rather

A.A. Beylich et al. / Geomorphology 57 (2004) 303–319

detailed information on regolith thickness and frozen ground conditions within various parts of a restricted area such as the Latnjavagge drainage basin. Geophysical data provide detailed estimates of regolith thickness along profiles. In large parts of the area, the thickness seems to be only a few metres as it was also locally estimated from field observations. Yet, in areas where deep erosion can be expected to have taken place during the Weichselian and/or where major moraines or other deposits are present, the regolith is thicker, but the thickness seems generally to be independent of topography. In August, the conductivity values in all geophysical profiles show that the ground had (mostly) thawed to the bottom of the regolith; and the water chemistry values show the highest TDS values during this part of the year indicating a maximum content of ions. The situation of thawed ground had gradually developed during the summer because during the early part of the season, the ground was frozen in the whole basin. After snowmelt, the ground thawed at different rates and to different extent: in relation to the drainage area above the inlet of Latnjajaure, the unfrozen delta had low TDS values that had not changed appreciably during the summer until midAugust, so the ground above had probably remained frozen, and this shaded area with thin regolith makes up the coldest part of the Latnjavagge drainage basin. For subcatchment A, the geophysical investigation suggests a relatively thick regolith (Fig. 3d and i). Since the ion concentrations in the drainage water from this area became gradually higher during the summer to reach relatively high concentrations from the beginning of August, it seems that in this area the thick regolith had time to thaw and that water might therefore have percolated through the whole regolith during the late part of the season. Also the slope along the eastern shore of the lake was thawed in August and the higher TDS and ion values at the outlet of Latnjajaure as compared to the inlet suggests that it had thawed just after snowmelt, as it is also suggested by ground temperatures above 0 jC measured at LFS directly after the melting of the snow cover (Beylich et al., submitted for publication (c)). The combined outflow from the whole basin reflects the various subcatchments and slopes with their

317

different regolith and ground temperature conditions: Pure water with constantly low ion concentrations over the summer came from the (partly) frozen upper catchment with a thin regolith cover. Further, downstream and around the lake this water became mixed with water from the slopes and from subareas with higher chemical denudation owing to thawed or gradually thawing ground, which locally, such as in subcatchment A, flowed through a sufficiently thick thawed regolith to make a clear imprint on the total outflow from the drainage basin. Chemical denudation in Latnjavagge (5382 kg/km2 year) is much lower than in Ka¨rkevagge (19200 kg/km2 year after Darmody et al., 2000; see also Campbell et al., 2002), but it seems to be at a similar level as in a number of other subarctic, arctic and alpine environments (Beylich, 1999, 2000a,b; Beylich et al., submitted for publication (a),submitted for publication (b); Darmody et al., 2000). It is therefore tempting to suggest that the regolith thickness as well as the ground temperature conditions over the summer in Latnjavagge reflect a fairly normal situation for arctic – oceanic high latitude mountain areas. Although the concentrations of dissolved substances are very low in this periglacial environment, chemical denudation is higher than mechanical fluvial denudation (Beylich et al., submitted for publication (b)). The dominance of chemical denudation over mechanical fluvial denudation is in line with the opinion of several authors that chemical denudation is a comparatively important process in periglacial environments; and more information on regolith conditions may greatly help towards a better understanding and quantification of the importance of underlying local conditions.

Acknowledgements Research in Latnjavagge was funded by a DAAD Post-Doc grant (‘‘Stipendium des DAAD im Rahmen des Gemeinsamen Hochschulsonderprogramms III von Bund und La¨ndern’’, 1999 – 2001; grant to Achim A. Beylich). Since 2002 research has been funded by the Deutsche Forschungsgemeinschaft (DFG, Emmy Noether-Programm; grant to Achim A. Beylich). The fieldwork was logistically supported by the Abisko Scientific Research Station (ANS), the Latnjajaure Field Station (LFS) (Ulf Molau) and by the Depart-

318

A.A. Beylich et al. / Geomorphology 57 (2004) 303–319

ment of Earth Sciences, Uppsala University. Ulf Molau kindly provided data on snow cover and meteorological data from LFS and Karin Luthbom (Uppsala/Lulea˚) gave helpful support in the field. Behrooz Oskooi kindly helped the first three authors to get acquainted with the geophysical equipment. The support from the mentioned persons and institutions is gratefully acknowledged. We thank Nel Caine and an anonymous reviewer for the critical comments on the paper.

References Andre´, M.-F., 1995. Postglacial microweathering of granite roches moutonnees in Northern Scandinavia. In: Slaymaker, O. (Ed.), Steepland Geomorphology. Wiley, London, pp. 103 – 127. Barsch, D., Gude, M., Ma¨usbacher, R., Schukraft, G., Schulte, A., 1994. Recent fluvial sediment budgets in glacial and periglacial environments, NW Spitsbergen. Zeitschrift fu¨r Geomorphologie. Supplementband 97, 111 – 122. Bastani, M., 2001. EnviroMT—a new controlled source/radio magnetotelluric system. Acta Universitatis Upsaliensis. Dissertations Faculty of Science and Technology 32. 179 pp. Bastani, M., Pedersen, L.B., 2001. Estimation of magnetotelluric transfer functions from radiotransmitters. Geophysics 66, 387 – 394. Beylich, A.A., 1999. Hangdenudation und fluviale Prozesse in einem subarktisch-ozeanisch gepra¨gten, permafrostfreien Periglazialgebiet mit pleistoza¨ner Vergletscherung-Prozessgeomorphologische Untersuchungen im Bergland der Austfirdir (Austdalur, Ost-Island). Berichte aus der Geowissenschaft. Shaker, Aachen. 130 pp. Beylich, A.A., 2000a. Untersuchungen zum gravitativen und fluvialen Stofftransfer in einem subarktisch-ozeanisch gepra¨gten, permafrostfreien Periglazialgebiet mit pleistoza¨ner Vergletscherung (Austdalur, Ost-Island). Zeitschrift fu¨r Geomorphologie. Supplementband 121, 1 – 22. Beylich, A.A., 2000b. Geomorphology, sediment budget, and relief development in Austdalur, Austfir dir,  East Iceland. Arctic, Antarctic, and Alpine Research 32 (4), 466 – 477. Beylich, A.A., 2000c. Morphoklima und rezente Morphodynamik im periglazialen Bergland der Austfir dir  (Austdalur, Ost-Island). Zeitschrift fu¨r Geomorphologie. Supplementband 123, 57 – 78. Beylich, A.A., 2001a. Slope denudation, streamwork and relief development in two periglacial environments in East Iceland and Swedish Lapland. Transactions, Japanese Geomorphological Union 22 (4), C-23. Beylich, A.A., 2001b. Recent morphoclimates and recent geomorphodynamics in periglacial environments in East Iceland, Swedish Lapland and Finnish Lapland. Transactions, Japanese Geomorphological Union 22 (4), C-24. Beylich, A.A., 2003a. Sediment budgets and relief development in

present periglacial environments—a morphosystem analytical approach. Hallesches Jahrbuch fu¨r Geowissenschaften. A (in press). Beylich, A.A., 2003b. Present morphoclimates and morphodynamics in Latujavagge the northern Swedish Lapland and Austdaluv, east Iceland. Jo¨kull 52, 33 – 54. Beylich, A.A., Gintz, D., 2003. Characteristics of high-magnitude/ low-frequency fluvial events generated by intense snowmelt or heavy rainfall in arctic periglacial environments in northern Swedish Lapland and northern Sibiria. Geografiska Annaler A (in press). Beylich, A.A., Kolstrup, E., Thyrsted, T., Gintz, D., subm.a. Water chemistry and its diversity in relation to local factors in the Latnjavagge drainage basin, arctic – oceanic Swedish Lapland. Geomorphology. Beylich, A.A., Molau, U., Luthbom, K., Gintz, D., subm.b. Rates of chemical and mechanical fluvial denudation in an arctic – oceanic periglacial environment in Swedish Lapland. Arctic, Antarctic, and Alpine Research. Beylich, A.A., Lindblad, K., Molau, U., subm.c. Direct human impacts on mechanical denudation in an arctic – oceanic periglacial environment in northern Swedish Lapland (Abisko mountain area). Zeitschrift fu¨r Geomorphologie, Supplementband. Caine, N., 1979. Rock weathering rates at the soil surface in an alpine environment. Catena 6, 131 – 144. Caine, N., 1995. Temporal trends in the quality of stream water in an alpine environment: Green Lakes Valley, Colorado Front Range, U.S.A. Geografiska Annaler 77A, 207 – 220. Caine, N., Thurman, E.M., 1990. Temporal and spatial variations in the solute content of an alpine stream, Colorado Front Range. Geomorphology 4, 55 – 72. Campbell, S.W., Dixon, J.C., Darmody, R.G., Thorn, C.E., 2001. Spatial variation of early season surface water chemistry in Ka¨rkevagge, Swedish Lapland. Geografiska Annaler 83A (4), 169 – 178. Campbell, S.W., Dixon, J.C., Thorn, C.E., Darmody, R.G., 2002. Chemical denudation rates in Ka¨rkevagge, Swedish Lapland. Geografiska Annaler 84A (3 – 4), 179 – 185. Clark, M.J., 1988. Periglacial hydrology. In: Clark, M.J. (Ed.), Advances in Periglacial Geomorphology. Wiley, Chichester, pp. 415 – 462. Collins, D.N., Young, G.J., 1981. Meltwater hydrology and hydrochemistry in snow- and icecovered mountain catchments. Nordic Hydrology 12, 319 – 334. Darmody, R.G., Thorn, C.E., Harder, R.L., Schlyter, J.P.L., Dixon, J.C., 2000. Weathering implications of water chemistry in an Arctic-Alpine environment, northern Sweden. Geomorphology 34, 89 – 100. Darmody, R.G., Allen, C.E., Thorn, C.E., Dixon, J.C., 2001. The poisonous rocks of Ka¨rkevagge. Geomorphology 41, 53 – 62. Dixon, J.C., Thorn, C.E., Darmody, R.G., 1984. Chemical weathering processes on the Vantage Peak Nunatak, Juneau Icefield, southern Alaska. Physical Geography 5, 111 – 131. Dixon, J.C., Darmody, R.G., Schlyter, P., Thorn, C.E., 1995. Preliminary investigation of geochemical process responses to potential environmental change in Ka¨rkevagge, Northern Scandinavia. Geografiska Annaler 77A, 259 – 267.

A.A. Beylich et al. / Geomorphology 57 (2004) 303–319 Dramis, F., Haeberli, W., Vonder Mu¨ll, D., Ka¨a¨b, A., Phillips, M., Wegmann, M., Krummenacher, B., Mittaz, C., Matsuoka, N., Guglielmin, M., Cannone, N., Smiraglia, C., 1997. Guide for the excursion Mountain Permafrost and Slope Stability in the Periglacial Belt of the Alps. IPA Italy. Fourth International Conference on Geomorphology. Supplementi di Geografia Fisica e Dinamica Quaternaria Suppl. III, vol. 2, pp. 181 – 203. Gude, M., Ma¨usbacher, R., Schukraft, G., Schulte, A., 1996. Untersuchung hydrologischer und hydrochemischer Prozesse in hocharktischen Permafrost-Einzugsgebieten mit Hilfe natu¨rlicher Tracer. Heidelberger Geographische Arbeiten 104, 460 – 472. Holmquist, P.J., 1910. Die Hochgebirgsbildungen am Torne Tra¨sk in Lappland. GFF 32 (4), 913 – 983. ¨ dega˚rd, R.S., Eiken, T., Sollid, J.L., 2000. ComposiIsaksen, K., O tion, flow and development of two tongue-shaped rock glaciers in the permafrost of Svalbard. Permafrost and Periglacial Processes 11, 241 – 257. Johannessen, M., Henriksen, A., 1978. Chemistry of snow melt water: changes in concentration during melting. Water Resources Research 14 (4), 615 – 619. Kling, J., 1996. Sorted circles and polygons in northern Sweden, Distribution and Processes. PhD thesis. Department of Physical Geography, Go¨teborg University. Kling, J., 2003. March 2003, www.systbot.gu.se/research/latnja/ latnja.html. ¨ versikt o¨ver Norra Norrbottensfja¨llens KaleKulling, O., 1964. O donberggrund. Sveriges Geologiska Underso¨kning: Serie Ba ¨ versiktskartor med beskrivningar 19, 166. O Leser, H., Dettwiler, K., Do¨beli, C., 1992. Geoo¨kosystemforschung in der Elementarlandschaft des Kvikaa-Einzugsgebietes (Liefdefjorden, Nordwestspitzbergen). Stuttgarter Geographische Studien 117, 105 – 122. Molau, U., 2001. Tundra plant responses to experimental and natural temperature changes. Memoirs of the National Institute of Polar Research, Special Issue 54, 445 – 466. Molau, U., 2003. March 2003, www.systbot.gu.se/research/latnja/ latnja.html. Molau, U., Kling, J., Lindblad, K., Bjo¨rk, R., Da¨nhardt, J., Liess, A., 2003. A GIS assessment of alpine biodiversity at a range of scales. In: Nagy, L., Grabherr, G., Ko¨rner, C., Thompson, D.B.A. (Eds.), Alpine Biodiversity in Europe. Ecological Studies, vol. 167 Springer, Berlin-Heidelberg. In press. Nabighian, M.N. (Ed.), 1987. Electromagnetic Methods in Applied Geophysics. Theory, Society of Exploration Geophysicists, vol. 1. Society of Exploration Geophysicists.

319

Nabighian, M.N. (Ed.), 1991. Electromagnetic Methods in Applied Geophysics. Applications, Society of Exploration Geophysicists, vol. 1. Society of Exploration Geophysicists. Peltier, L.C., 1950. The geographic cycle in periglacial regions as it is related to climatic geomorphology. Annals of the Association of American Geographers 40, 214 – 236. Potschin, M., Leser, H., 1994. Saisonaler Verlauf des Vorfluterchemismus im Kvikaa-Einzugsgebiet (Liefdefjorden, NW-Spitzbergen). Zeitschrift fu¨r Geomorphologie. Supplementband 97, 161 – 174. Rapp, A., 1960. Recent development of mountain slopes in Ka¨rkevagge and surroundings, northern Scandinavia. Geografiska Annaler 42, 71 – 200. Reynolds, J.M., 1997. An Introduction to Applied and Environmental Geophysics. Wiley, Chichester. Reynolds Jr., R.C., Johnson, N.M., 1972. Chemical weathering in the temperate glacial environment of the Northern Cascade Mountains. Geochimica et Cosmochimica Acta 36, 537 – 554. Siripunvaraporn, W., Egbert, G., 2000. An efficient data-subspace inversion method for 2-D magnetotelluric data. Geophysics 65, 791 – 803. Stephens, M., Beckholmen, M., Zachrisson, E., 1994. The Caledonides. In: Frede´n, C. (Ed.), National Atlas of Sweden. Geology, pp. 22 – 25. Stro¨mquist, L., Rehn, J., 1981. Rainfall measurements, runoff and sediment transport in Ka¨rkevagge, Swedish Lapland. Transactions, Japanese Geomorphological Union 2, 211 – 222. Thorn, C.E., 1975. Influences of late-lying snow on rock weathering rinds. Arctic and Alpine Research 7, 373 – 378. Thorn, C.E., Darmody, R.G., Dixon, J.C., Schlyter, P., 2001. The chemical weathering regime of Ka¨rkevagge, arctic-alpine Sweden. Geomorphology 41, 37 – 52. ¨ ber die mechanische Verwitterung der Von Lozinski, W., 1909. U Sandsteine im gema¨ssigten Klima. Bulletin International de L‘Academie des Sciences de Cracovie class des Sciences Mathematique et Naturalles 1, 1 – 25. Von Lozinski, W., 1912. Die periglaziale Facies der mechanischen Verwitterung. Comptes rendus, XI Congres Internationale Geologie, Stockholm 1910, 1039 – 1053. Walling, D.E., Webb, B.W., 1983. The dissolved load of rivers: a global overview. International Association of Hydrological Sciences 141, 3 – 20.