Environ. Sci. Technol. 2007, 41, 7424-7429
Unexpected Response of High Alpine Lake Waters to Climate Warming HANSJO ¨ R G T H I E S , * ,† U L R I K E N I C K U S , ‡ VOLKMAR MAIR,§ RICHARD TESSADRI,| DANILO TAIT,⊥ BERTHA THALER,⊥ AND ROLAND PSENNER† Institute of Ecology, University of Innsbruck, Technikerstrasse 25, A-6020 Innsbruck, Austria, Institute of Meteorology and Geophysics and Institute of Mineralogy and Petrography, University of Innsbruck, Innrain 52, A-6020 Innsbruck, Austria, Service for Geology and Material Testing of the Autonomous Province Bozen, South Tyrol, Via Val d’Ega 48, I-39053 Cardano, Italy, and Biological Laboratory of the Autonomous Province Bozen, South Tyrol, Via Sottomonte 2, I-39055 Laives, Italy
TABLE 1. Site Characteristics of RAS and SOS longitude east latitude north watershed altitude (m asl) watershed areaa (km2) rock glacier area (km2) lake surface area (km2) maximum depth (m) lake volume (m3) residence timeb (months) a
Includes lake surface area.
b
Introduction High altitude lakes in the European Alps are known to be sensitive indicators of environmental and climatic change (1-8). Furthermore, a pronounced increase in mean annual air temperature of more than 1 °C is reported for the Alps for the period of 1980-2003 with respect to the reference period 1901-2000 (9). Since the late 19th century, this region has warmed twice as much as the global or Northern Hemispheric average (9). The Alps have experienced a pronounced increase in temperature both during winter and summer since the 1970s, and the warmest summers during the past 500 years were 1994, 2000, 2002, and particularly 2003 (10). Impacts of climate warming include the substantial warming of permafrost in several European mountains (11), the permafrost thaw in the Upper Engadin of the Eastern Swiss Alps (12), and the related destabilization of Alpine rock walls in the hot summer of 2003 (13). Permafrost and rock glaciers are common to high mountain areas (14) and have been investigated comprehensively (15), but studies on the * Corresponding author phone: +43 512 507 6123; fax: +43 512 507 6190; e-mail:
[email protected]. † Institute of Ecology, University of Innsbruck. ‡ Institute of Meteorology and Geophysics, University of Innsbruck. § Service for Geology and Material Testing of the Autonomous Province Bozen. | Institute of Mineralogy and Petrography, University of Innsbruck. ⊥ Biological Laboratory of the Autonomous Province Bozen. 7424
9
ENVIRONMENTAL SCIENCE & TECHNOLOGY / VOL. 41, NO. 21, 2007
SOS 10° 56′46′′ 46° 57′57′′ 2796-3016 0.22 0.012 0.035 18 340000 18-23
Estimated, cf. text.
TABLE 2. Elemental Composition in Gneiss and Micaschist Rock Samples from the Catchments of RAS and SOS, Mean Values, and Relative Standard Deviationsa RAS (mg kg-1)
RAS rel SD (%)
SOS (mg kg-1)
N)5
Over the past two decades, we have observed a substantial rise in solute concentration at two remote high mountain lakes in catchments of metamorphic rocks in the European Alps. At Rasass See, the electrical conductivity increased 18-fold. Unexpectedly high nickel concentrations at Rasass See, which exceeded the limit in drinking water by more than 1 order of magnitude, cannot be related to catchment geology. We attribute these changes in lake water quality to solute release from the ice of an active rock glacier in the catchment as a response to climate warming. Similar processes occurred at the higher elevation lake Schwarzsee ob So¨ lden, where electrical conductivity has risen 3-fold during the past two decades.
RAS 10° 27′23′′ 46° 44′50′′ 2682-2870 0.22 0.038 0.015 7 72000 5-7
Ca Mg Fe Mn Zn Ni
11 428 13 870 44 064 774 80 29
SOS rel SD (%)
N ) 13 22 20 15 13 48 21
9286 13 810 41 616 620 171 32
Clarke number (mg kg-1) range
67 61 50 57 86 84
29 600-36 000 18 700-23 500 46 000-51 000 800-1000 40-80 60-80
a Clarke numbers give a range of corresponding mean values in the earth’s crust (from ref 31).
impact of active rock glaciers on hydrology and water chemistry of the adjacent surface waters are rare (16, 17). Here, we discuss changes in the water chemistry of two high altitude lakes in the Alps under the impact of increasing air temperature.
Experimental Procedures Rasass See (RAS) and Schwarzsee ob So¨lden (SOS) are high alpine lakes in headwater catchments of the Central Eastern Alps in Europe (Table 1). RAS is located in the upper Vinschgau in South Tyrol (Italy), south of the main alpine divide near the Swiss and Austrian borders. SOS is located in the Oetztal Alps in North Tyrol (Austria), north of the main alpine divide about 45 km northeast of RAS. The bedrock in both catchments consists of bare rock and talus of the Oetztal metamorphic complex (i.e., paragneisses, micaschists, and orthogneisses). Soil coverage in both catchments is sparse (SOS: ∼5% and RAS: ∼10% of catchment area) and characterized by alpine grass vegetation. A more detailed site description is summarized in Tables 1 and 2. Both lakes were sampled during the autumn overturn at several depths along vertical profiles (at RAS in 1986, 1990, 2000, 2001, and 2005 and at SOS every year since 1985 to date except 1987 and 1988). Details on analytical methods and analytical quality control have been reported elsewhere (4, 5, 18). Lake water samples (RAS: 2005 and SOS: 2006) were additionally analyzed for trace elements by ICP-OES. A geological survey at RAS in autumn 2005 was performed by V.M., and collected catchment rocks were analyzed with WDXRFA after calibration by standard rock reference samples. Rocks from the SOS catchment were analyzed in 1994 with EDXRFA. RAS and SOS lake water samples of autumn 2005 were analyzed for tritium by liquid scintillation spectrometry to estimate the mean age of the lake water. A one-box exponential model 10.1021/es0708060 CCC: $37.00
2007 American Chemical Society Published on Web 09/21/2007
FIGURE 1. Conductivity, sulfate, calcium, and magnesium in lake water of RAS (black triangles) and SOS (open circles) (1985-2005). Values for each lake represent mean values of four to seven discrete samples along the lake vertical profile taken during holomixis. Variability among single values is 2700 µequiv L-1), more than 26-fold for sulfate (167 to >4400 µequiv L-1), and more than 13-fold for calcium (133 to >1700 µequiv L-1) (Figure 1). During the past two decades, the silica concentration at RAS increased by almost 2-fold (44-82 µmol L-1), which is a similar rate as compared to SOS (19-41 µmol L-1). A ternary diagram showing the relative contribution of calcium, magnesium, and sulfate (Figure 2)sthese three ions comprise 80-99% of the ion balancesreveals for SOS during the past 20 years a 10% increase in the share of sulfate at the expense of calcium, while the share of magnesium remained almost constant. In 1986 and 1990, the ion ratios at RAS equaled the recent ones at SOS (i.e., Ca/Mg/SO4 ) ∼40:10:50%). From 2000 onward, the relative share of magnesium at RAS increased by 20% at the expense of calcium, while the share of sulfate remained almost constant (Figure 2). We discuss potential factors that may have contributed to the pronounced solute increase at RAS and SOS: atmospheric deposition, rising air temperature, weathering, and the impact of active rock glaciers. Atmospheric Deposition. To evaluate the impact of atmospheric deposition, we selected two sites in the vicinity of RAS and SOS, one located north (Reutte) and one south (Innervillgraten) of the main alpine divide, where daily wet deposition has been measured since the mid 1980s. Concentrations of major ions in atmospheric deposition (Figure 3) have either remained stable (e.g., magnesium) or declined (e.g., sulfate), in agreement with trends observed elsewhere in Europe (20-22). A decreasing sulfate concentration in precipitation reflects the reduction in sulfur dioxide emission in recent decades (23, 24). Current levels of calcium, magnesium, and sulfate in atmospheric deposition at our VOL. 41, NO. 21, 2007 / ENVIRONMENTAL SCIENCE & TECHNOLOGY
9
7425
FIGURE 2. Ternary plot of relative contribution of calcium, magnesium, and sulfate in lake water of RAS (open circles) and SOS (black dots). Arrows indicate the temporal change in ionic ratios at RAS and SOS between 1986 and 2005.
FIGURE 3. Volume weighted annual mean concentration of atmospheric deposition (µequiv L-1) at Reutte and Innervillgraten (1985-2005). selected sites are less than 20 µequiv L-1, and their ranges are well below the corresponding lake water concentrations at RAS and SOS (Figures 1 and 3). Only episodic Saharan dust events may result in annual volume-weighted calcium concentrations in deposition of up to 40 µequiv L-1 in the alpine region (21, 25). We believe that the impact of atmospheric deposition on the recent strong rise of base cations and sulfate at RAS and SOS is negligible. Air Temperature and Weathering. In the Alps, the mean annual air temperature has increased by more than 1 °C since 1980 (9). As the air temperature increases, the average height of the freezing line moves upward. The major changes occurred in May and June (Figure 4a). While the elevation of the mean freezing line showed an interannual variability up to several hundred meters during the past century, the 5 year running averages for May and June have moved upward by about 150 and 280 m during the two last decades. The running average of the 0 °C line for May reached the elevation of RAS during the mid 1990s and was well above the SOS catchment in June during the past decade (Figure 4a), thus 7426
9
ENVIRONMENTAL SCIENCE & TECHNOLOGY / VOL. 41, NO. 21, 2007
favoring earlier and more intense melt processes in both lake catchments than had previously occurred. Weathering rates for granite and gneiss are reported to increase with rising temperature and water availability (2628). A temperature increase from 0 to 25 °C is expected to cause weathering rates to increase by about 1 order of magnitude (27). Summer temperatures of up to 20 °C are reported for alpine bedrock at a depth of 10 cm (29) and for shallow running waters (18), both at altitudes above 2500 m. From 1986 to 2005, the silica concentration in RAS and SOS increased at similar rates (RAS: 1.9-fold and SOS: 2.2-fold), which indicates a comparable increase of silica release in both catchments. The observed increase in sulfate at RAS and SOS was well-balanced with the increase in the calcium and magnesium concentration (i.e., the ratio Σ(Ca + Mg)/ SO4 ranged between 0.9 and 1.1). These results agree with those reported for 57 alpine lakes in Tyrol and Carinthia (Austria) (2, 30), where the increase in conductivity, sulfate, calcium, and magnesium at the majority of the studied lakes between 1985 and 1995 was attributed to enhanced weathering. However, we believe that the pronounced rise of solutes and conductivity at RAS and SOS since the mid-1990s (Figure 1) cannot be explained by bedrock weathering. Impact of Active Rock Glaciers. While both catchments have a similar geological composition, as confirmed by rock analysis in 1994 and 2005 (Table 2, includes ref 31), they differ in altitude and in the size of recently detected active rock glaciers. At RAS, the active rock glacier extends on a north-facing slope of the catchment down to the lake shore (Figure 4b). With an area of 0.038 km2, the active rock glacier is more than twice the lake area and almost 20% of the total catchment area. Three smaller active rock glaciers are found in the SOS catchment covering a total area of 0.012 km2, which corresponds to only one-third of the lake area and to about 5% of the total catchment area. The discharge of active rock glacier outflows is known to vary with season both in magnitude and in chemical composition. A seasonal increase in the conductivity of meltwater streams from three active rock glaciers in the Austrian Alps (from 20 to 200 µS cm-1 during May to October 2000) has been reported (16). Maximum conductivity concentrations in these streams occur at a base flow in autumn when water percolates slowly through active rock glaciers and adjacent sediments (16). A similar seasonality of conductivity has been reported (17) for a rock glacier on schist and gneiss bedrock in the front ranges of Colorado. During 2003, the seasonal solute increase in the outflow of the rock glacier RG5 (17) was more than 60-fold for sulfate (40-120 to >6000 µequiv L-1), more than 30-fold for magnesium (30 to >900 µequiv L-1), more than 20-fold for calcium (200 to >4000 µequiv L-1), and about 4-fold for silica (25-100 µmol L-1). Base flow from the rock glacier consisted predominantly of an internal ice melt with enriched geochemical concentrations (17). Significant chemical enrichment in glacial meltwaters is reported in a review on glacial meltwater hydrochemistry (32). The seasonality in solute enrichment from active rock glacier outflows is expected to intensify with increasing mean air temperature and subsequent enhanced melt processes. This is supported by reports of higher flow velocities of active rock glaciers in the Austrian Alps since the 1990s, with the greatest displacements between 2002 and 2004 (33). The increase in flow velocities of distinct active rock glaciers is attributed to a slightly increased ice temperature and greater discharge of meltwater during summer (33). We therefore attribute the strong increase in solutes at RAS to an intensified melting and release of substances from the rock glacier in the RAS lake catchment. The rising solute concentrations at RAS may have been additionally stimulated by a period of higher than normal rainfall. From 1999 to 2002, summer and
FIGURE 4. (a) Calculated height of the 0 °C level (in m asl) (1901-2005) for the catchments of RAS and SOS in May (blue) and June (red). Bold lines: 5 year running average. Dotted lines give elevation boundaries of RAS and SOS catchments. (b) RAS and adjacent shallow pond (P); the active rock glacier is indicated by the white ellipsis. Photo by V.M.
TABLE 3. Calcium, Magnesium, and Conductivity in Water Samples of RAS and the Adjacent Pond, September 2005a
L-1)
Ca (mg Mg (mg L-1) conductivity (µS cm-1) a
TABLE 4. Trace Elements in Water Samples from Lake Vertical Profiles of RAS, September 2005 and SOS, October 2006a
RAS
RAS
pond
RAS (µg L-1)
RAS (µg L-1)
SOS (µg L-1)
SOS (µg L-1)
N)5
SD
N)1
Mean
SD
mean
SD
34.0 27.4 450.8
0.5 0.2 0.5
6.5 1.7 110
143 12 1 559 243 8 181
2 1 0 3 2 1 4
25
