Mountain regions are cold environments that are hostile to human occupation and widely regarded as places where the air is clean, water is pure and ecosystems are pristine. Yet many mountain regions, especially in Europe, are far from pristine. In the 1980s, research showed that mountain lakes were especially vulnerable to acid deposition and sediment core studies at many sites demonstrated that some mountain lakes had become acidified over the course of the last century. Since then, studies of the water chemistry, biology and history of lakes across the different European mountain regions have increased our understanding of the processes within these systems, their ecological condition and the threats facing them. These studies have demonstrated that:
i) gradients of sulphur, nitrogen, metals and persistent organic pollutant deposition occur from relatively uncontaminated regions in Spain and Central Norway to regions of heavy pollution loading in central and eastern Europe;
ii) nitrate and sulphate concentrations in lake water reflect the gradients in atmospheric deposition;
iii) concentrations of mercury, lead and cadmium in fish tissue show marked regional differences;
iv) the spatial pattern of organochlorine concentrations in fish and in sediments follow the pattern for other pollutants, although there is also good evidence for the selective cold trapping of some compounds both at high latitudes and at high altitudes;
v) fish suffer physiological stress in mountain lake-water of low ionic strength;
vi) climate change over the last century is likely to have induced significant changes in lake ice-cover and water column stratification and mixing in some regions and these have in turn influenced the structure and productivity of biological communities;
This paper provides a review of the research undertaken on European mountain lakes over the past 30 years. In particular, it highlights the stresses faced by these vulnerable systems and the effects these have had, continue to have, and are likely to have on mountain lake ecosystems in future.
Mountain lakes have been less studied in the past than other lakes in Europe, partly because of their remoteness and partly because they were perceived to be unpolluted and undisturbed. However, in the 1980s several research projects concerned with problems of surface water acidification showed that mountain lakes were especially vulnerable to acid deposition, and sediment core studies at many sites demonstrated that some lakes had become acidified over the course of the last century. Since then, as a result of both national and pan European programmes, a general knowledge has been acquired of the water chemistry of a large number of sites and a more detailed knowledge of the chemistry, biology (algae, invertebrates, zooplankton, fish) and history (from sediment records) of a smaller number of sites located in different European mountain regions. An understanding has been developed of the dynamics of just a few sites that have been instrumented (automatic weather stations, deposition collectors, data loggers, sediment traps etc.) and sampled intensively both in the ice-on and ice-free seasons. This paper provides a review of the research undertaken on European mountain lakes over the past 30 years, focusing on the effects of sulphur and nitrogen deposition and contamination by airborne toxic substances on mountain lake ecosystems. Now climate change brings an added threat both directly through changes in temperature and precipitation and indirectly as a result of interactions between climate, mountain biogeochemistry and pollution. Here we consider the effects these have had, continue to have, and are likely to have on mountain lake ecosystems in future.
Mountain lakes and their sensitivity to ecological change
Lakes situated in mountain regions above the timber line are especially sensitive to stresses from human activity owing to their:
i) Cold, wet and windy climate. Mountain regions are cold and volatile persistent organic pollutants (POPs) derived from industrial regions migrate towards and condense in mountains, in some cases increasing in concentration with altitude (Grimalt et al., 2001). Cold temperatures also reduce the re-volatilisation of POPs from lake surfaces, enhancing trapping. Higher precipitation and high wind speeds enhance pollutant deposition (Fowler and Battarbee, 2005).
ii) Snow and ice-cover. Snow cover reflects solar radiation, limiting its warming effect. As winter temperatures increase, snow cover decreases and mountain catchments warm more rapidly. Reduced ice cover also means lake water temperatures will increase more rapidly. Changes in snow and ice regimes will also alter hydrological regimes (i.e. reduced spring melt) and affect plants adapted to, and dependent on, snow-beds.
iii) Glaciated catchments. Some catchments contain rock or ice glaciers. Warming leads to major changes in meltwater discharge and causes the release of solutes and pollutants previously frozen in the ice (e.g. Thies et al., 2007).
iv) Thin and poorly developed catchment soils. Rates of weathering increase with increased temperature and steep, unvegetated and highly erodible slopes of mountain catchments can amplify the effects of extreme precipitation events, as runoff occurs much more quickly and with much higher energy. Pollutants deposited in such catchments have short residence time and are effectively transported to streams and lakes.
v) Dilute water. Although some mountain lakes occur in limestone regions, e.g. the Julian Alps, and are characterised by alkaline waters, the majority of lakes in the mountains have very dilute waters, sometimes with exceptionally low conductivity close to the chemistry of precipitation (The MOLAR Water Chemistry Group, 1999) Such lakes have very little ability to neutralise acid deposition.
vi) Clear water. Mountain lakes are often exceptionally clear, with very low concentrations of coloured dissolved organic matter (CDOM). CDOM is derived from humic substances produced in catchment soils as a result of the decay of organic matter. In lake water CDOM gives colour to the water and reduces the penetration of solar radiation with depth. Clear water allows potentially damaging UV radiation to penetrate more deeply than in lowland lakes. Furthermore, as a result of the thinner atmosphere in the mountains incident UV radiation is higher. Whilst many aquatic organisms are adapted to high UV environments, some biological groups are sensitive to high UV and future climate change may bring about changes in the exposure of aquatic biota to UV (Sommaruga, 2001).
vii) Low nutrient status. Mountain lakes usually have very low nutrient status (Catalan et al., 2002a). Both N and P concentrations are exceptionally low and even N and P from long-distance transported pollution is sufficient to cause symptoms of eutrophication (e.g. Wolfe et al., 2001). There is increasing evidence that N concentrations especially have been increasing in remote mountain lakes through this transport pathway (e.g. Monteith et al., 2007).
viii) Simple food webs. Mountain lakes have simple food webs (Straškrábová et al., 1999) and relatively low overall biodiversity. They are extremely susceptible to losses in biodiversity and to invasion by alien species. Once damaged, communities are very slow to recover partly due to geographical isolation.
Assessing the status of mountain lakes
Early research (e.g. Pechlaner, 1971; Light, 1975) on high mountain lakes focussed on purely scientific questions rather than on issues of pollution and environmental protection. Today there is an awareness not only of the threat of human activity to these systems, but also that these sensitive systems can be used as indicators of the state of the wider environment both in the mountains and globally. Within Europe it has been possible under the auspices of the European Union (EU) to design large-scale projects (e.g. the AL:PE (Acidification of Mountain Lakes:Palaeolimnology and Ecology) and MOLAR (Mountain Lake Research) projects) to characterise in detail, the chemistry, biology and history of representative mountain lakes in the different mountain systems across Europe and to assess their status with respect to their exposure to long-distance transported pollutants and, more recently, to climate change (e.g. Psenner et al., 2002; Battarbee, 2002). However these research efforts concentrated on a relatively small number of sites and, until recently there was no systematic evaluation of the variability within and between European mountain regions. A unique pan-European survey of remote mountain lakes was undertaken under the auspices of the EMERGE project (European Mountain lake Ecosystems: Regionalisation, diaGnostic & socio-economic Evaluation) with the aim of evaluating the regional variability found in those remote mountain lakes together with an assessment of their ecological status. Physical, chemical and biological data from over 350 remote mountain lakes across Europe were collated allowing comprehensive multivariate analyses to be carried out, the generation of ecological classifications for these lakes and a comprehensive evaluation of the status of remote mountain lake ecosystems throughout Europe to be undertaken (Catalan and Kernan, Centre for Advanced Studies of Blanes—CSIC, Blanes, Spain in press; Kernan et al., Environmental Change Research Centre, University College London, Gower Street, London, in press (a)). Wide-ranging datasets now exist which have allowed pan-European assessments to be made of species composition in remote mountain lakes and the environmental factors which determine the distribution of species (e.g. Catalan et al., Centre for Advanced Studies of Blanes—CSIC, Blanes, Spain, in press; Kernan et al., Environmental Change Research Centre, University College London, Gower Street, London, in press (b)).
Sampling remote lakes, however, can be a daunting task, especially during the winter season, yet in the successive EU projects it has been possible, partly with the help of automatic sampling systems, to obtain high quality data from a large number of remote mountain lakes. Today we have a large environmental database that allows comprehensive analysis of mountain lake chemistry and biology, and for some sites, where sampling has been maintained over successive projects a number of decadal-scale time series are now available (Mosello et al., 2002). Where these are supplemented by palaeolimnological records we can reconstruct trends in mountain lake ecology representing changes that have occurred over the last 100 or more years.
The quality of the sediment record in high mountain lakes has been one of the main discoveries of recent research. Almost all the lakes studied have been found to contain good records with relatively organic (5–20%) sediment, containing an abundance of microfossils, especially diatoms, cladocera and chironomids (Battarbee, 2002). Sediment accumulation rates are usually low (less than 1 mm per year and sometimes much less; Appleby, 2000), but to some extent these are offset by the relatively low degree of bioturbation due to benthic invertebrates. With fine interval sampling (e.g. 2 mm), therefore, it is usually possible to achieve a decadal, or finer time, scale resolution sufficient to show clear trends in the impacts of human activity over the last 150 years. Additionally, for studies of atmospheric pollutant or climate change impacts, mountain lake systems offer clear advantages in that the relationships between lake response and external forcing is not obfuscated by direct catchment influences.
Evidence for ecological change
Sulphur and nitrogen deposition
Evidence for the contamination of mountain lakes by sulphur and nitrogen can be demonstrated from the large concentration of non-marine sulphate and nitrate in the water column and the presence of fly-ash particles (e.g. spheroidal carbonaceous particles; SCPs) in lake sediments (Rose et al., 2002). The spatial variation in concentration of these substances is in good agreement with the distribution of industrial regions within Europe and the temporal variation in the concentration of SCPs measured in sediment cores reflects the progressive industrialisation of Europe beginning in the nineteenth century.
Whilst many mountain lakes in Europe remain un-acidified either because they have adequate natural alkalinity (including dust supply) to neutralise the acidity or because they occur in regions of low acid deposition, lakes in high acid deposition regions that also have low natural alkalinity have been acidified. This is indicated most clearly by changes in the composition of diatom assemblages preserved in recent lake sediments (Figure 1) (Jones et al., 1993). The data from the AL:PE project show that the most severe acidification has taken place in Central and Western Europe (e.g. Fott et al., 1994; Camarero et al., 1995). In these regions macro-invertebrate populations are impoverished and lack species sensitive to acidification such as Baetis rhodani (Raddum and Fjellheim, 2002) and fish populations (mainly brown trout (Salmo trutta) or arctic char (Salvelinus alpinus)) show signs of acid stress (Rognerud et al., 2002). Critical loads modelling applied to 300 mountain lakes across Europe demonstrated the potential extent of the acidification problem (Curtis et al., 2005a). Subsequent analysis of surface (representing the then contemporary conditions) and pre-industrial sediments on a sub-set of these sites showed major biological changes but multiple drivers were thought to be responsible (Curtis et al., Environmental Change Research Centre, University College London, Pearson Building, Gower Street, London, in press) and it was not possible to distinguish the effects of acidification from climatic and other pollutant drivers (Clarke et al., 2005).
The threat of increased lake acidification from sulphur in mountain regions, however, is now receding. Following international agreements on the reduction of acidifying gases in Europe, acid deposition in mountain regions has begun to decline and the pH and alkalinity of some, but not all sites, has begun to recover (Mosello et al., 2002). However, this is almost entirely driven by reductions in S emissions.
Previous work during the EU-MOLAR project identified the high levels of nitrate present in remote mountain lakes, especially on the southern slopes of the Alps and in the Tatra Mountains (Kopáček et al., 2005; Vreča and Muri, 2006) and other studies have shown that nitrogen contamination from long-distance transport processes occurs in almost all remote mountain regions of the Northern Hemisphere (e.g. Wolfe, 2001; Sickman et al., 2003). Despite attempts to reduce N emissions in Europe nitrate concentrations in surface waters show no tendency to decrease and in some cases nitrate concentrations are increasing (Monteith et al., 2007). Hence, in acidified lakes nitrate rather than sulphate is becoming the dominant acid anion (Curtis et al., 2005b). Unlike sulphur, nitrogen can also be a limiting nutrient (e.g. Bergström et al., 2005). There is concern that soils in many mountain regions may be becoming progressively saturated by N, potentially leading to increased N leaching to freshwaters, not only causing acidification, or delaying recovery by offsetting the effects of S reduction, but also causing eutrophication. Convincing evidence for this has been demonstrated by Wolfe (2001) from studies of the sediments in mountain lakes in Colorado, where increases in planktonic diatom abundance are correlated with a shift towards more depleted values of δ15N in the organic sediment record (Figure 2). Camarero et al. (Centre for Advanced Studies of Blanes-CSIC. Accés, Blanes, Spain in press, a) suggest that the main sources of chemical variation in European remote mountain lakes are related to weathering of bedrock, a maritime influence and the deposition and processing of nitrogen within the catchment.
The contamination of mountain lakes by toxic trace metals is most clearly seen in lake sediment records (e.g. Yang et al., 2002). In Lochnagar, Scotland, increases in Pb and Hg began at the end of the 19th century and increased through the 20th century to a maximum in the 1970s. The long-range transport of the metals occurs as aerosols, predominantly in the size range 0.1 to 1.0 μ m and the enhanced accumulation of metals in mountain ecosystems occurs as a consequence of two effective deposition processes. First, wet deposition provides an efficient mechanism for the removal of small particles and second, dry deposition, which is an inefficient removal mechanism for small particles at low altitudes, becomes a much more efficient deposition mechanism in the mountains due to the much higher wind speeds, and the increase in particle size as the particles are lifted to higher and cooler elevations (Fowler and Battarbee, 2005). Camarero et al. (Centre for Advanced Studies of Blanes—CSIC, Blanes, Spain in press, b) in an analysis of surface and pre-industrial sediment samples show that the concentration of trace metals and metalloids is much higher in contemporary sediments than in pre-industrial samples, particularly for Pb, Hg, and As with the Tatra Mountains and Scotland being particularly affected.
Although high concentrations of metals can be found stored in lake sediments, of greater concern is the high concentration of toxic metals in fish tissue. Fish from lakes in the central parts of Europe have especially high concentrations of Pb (Figure 3) and Cd (Fernandez et al., 1999).
Persistent organics pollutants (POPs)
Sediment records also show contamination of mountain lakes by persistent organic pollutants (POPs). Polycyclic aromatic hydrocarbons (PAHs) show a history very similar to those of Pb and Hg reflecting a common source from fossil fuel combustion (although PAHs are also produced by natural combustion of organic matter—e.g. forest fires), whereas the organochlorine pesticides (OCPs) and polychlorinated biphenyls (PCBs), as expected, only occur in more recent sediments reflecting their manufacture and use mainly since the 1950s (Grimalt et al., 2004). Polybromodiphenyl ethers (PBDEs), introduced in the 1970s are found only from that date (Gallego et al., 2007).
The presence of POPs in high mountain lakes indicates contamination from long-distance transported air pollution, and the sediment record of compounds such as toxaphene, can provide evidence of transport from North America to Europe indicating a trans-continental component to long-range transport (Rose et al., 2001). The accumulation pattern depends on local climatic conditions whereas their atmospheric fallout is quite uniform and seasonally dependent (Carrera et al., 2002). For some POPs concentrations in lake sediments and fish tissue increase with altitude as these substances become progressively redistributed to colder and more remote regions by volatilization and cold trapping processes (Figure 4, Grimalt et al., 2001) and contaminate lakes that are distant (> 1000 km) from production and use of the compounds. Temperature is also the main factor controlling the deposition fluxes of lower molecular weight PAHs but particle deposition and wet precipitation determine the atmospheric deposition fluxes of the higher molecular weight PAH homologues. Some of the largest concentrations of both metals and POPs in fish are found in Svalbard, one of the most remote regions of the world, where levels of Hg, and a number of PCB congeners are double those at other sites as a result of food chain biomagnification. Rognerud et al. (2002) have shown that these higher values are the result of a progressive shift to cannibalism in the diet of arctic char between the ages of 11 and 20 years. Concentrations of some organic contaminants in fish from mountain lakes are often high enough that toxic effects might be expected to occur in piscivorous birds and mammals (Schindler, personal communication).
Future climate change will have direct and indirect consequences on water in mountain regions through impacts on water quantity, water quality and the structure and functioning of aquatic ecosystems. Although the primary impacts will occur at high elevation in headwater systems, many of the consequences will be transmitted downstream to lowland regions changing conditions both for natural ecosystems and for human use.
Climate trends in mountain regions
Most mountain lakes in Europe have clearly experienced significant variability in climatic conditions over the last 200 years. Instrumental temperature reconstructions by Agustí-Panareda and Thompson (2002) show that decadal-scale fluctuations in mean annual temperature with amplitude of up to 2°C have taken place at these sites. These are sufficient to cause ecologically important changes in lake heat balance and ice-cover to occur. At most sites there has been a rapid recent increase in mean annual temperature consistent with the global warming hypothesis. Whereas temperature trends show coherent patterns in time and space, precipitation trends are less clear. Nevertheless significant changes are expected. IPCC 2007 catalogues changes in precipitation and clouds (IPCC, 2007a) and importantly stresses the very large regional variability in expected trends with time (IPCC, 2007b). Mountain regions are especially sensitive to changes in clouds and precipitation since they amplify the precipitation amounts and wet deposition by orographic processes.
Climate change impacts on mountain hydrology include potential changes in annual and inter-annual water balance, the melting of mountain glaciers and permafrost, the amount of precipitation falling as rain rather than snow, changes in the seasonality of precipitation and runoff regimes, including extreme events, and the seasonal distribution of precipitation and runoff (Jenkins et al., 2007). The consequences for aquatic ecosystems include direct impacts on stream and lake habitats and indirect effects through changes in soil erosion and sediment transport. Variations in lake-levels are likely to be more extreme (some mountain lakes in Southern Europe may even disappear) and upland water bodies draining to higher-order rivers at lower elevations may become disconnected.
Water quality response
The recent pronounced increase in temperature especially in the Alpine region (Auer et al., 2006) has enhanced melt processes in high mountain environments (Harris et al., 2003; Gruber et al., 2004). Depending on altitude and exposure, rock glaciers especially are vulnerable to melting (Krainer and Mostler, 2006). The melting of rock glaciers induces the release of ions and trace metals (Thies et al., 2007). As rock glaciers are a widespread geomorphological features in high mountain regions around the world (e.g. Humlum, 1998) similar impacts are likely to occur elsewhere. In the same way there is a growing concern, yet to be studied, about the potential release of POPs and trace metals that have accumulated in mountain glaciers (Blais et al., 2001) and are now being released into headwater streams and lakes, potentially bio-accumulating in aquatic food chains and causing decreases in water quality for potable supply.
More universally climate change will have a major impact on the water quality of mountain lakes through its impact on catchment biogeochemical processes, especially with respect to weathering rates and nutrient loading. The primary impact of increased weathering rates associated with higher temperatures is on the acid-base chemistry of surface waters. Sommaruga-Wograth et al. (1997) have shown that temperature is a key driver of lake water pH. At acidified sites decadal-scale change in diatom-inferred pH is strongly correlated with instrumental temperature records and at some non-acidified sites recent diatom-inferred pH changes follow trends in temperature over the last few decades (Sommaruga-Wograth et al. 1997, Larsen et al. 2006). As sulphur deposition continues to decrease temperature change is likely to emerge as the a key driver of lake acidity at all sites and pH values may reach levels only previously experienced in the warmer early Holocene (Larsen et al., 2006).
Increased temperatures are also likely to have an impact on nutrient loading and productivity. Changes in diatom plankton at sites in the Pyrenees (Catalan et al., 2002b), Finland (Sorvari et al., 2002, Austria (Koinig et al., 2002), Norway (Larsen et al., 2006), northern Russia (Solovieva et al., 2005) and the Canadian Arctic (Smol et al., 2005) are consistent with this hypothesis. Increases in the amount of organic matter accumulating in mountain lake sediments at almost all sites over recent decades (Battarbee et al., 2002) also indicate a primary production response to warming. However, warming may not be the only explanation for these observations as an increase in diatom plankton has also been reported at remote sites where warming has not occurred, but where there is evidence from 15N records of eutrophication by long-range transported nitrogen (cf. Wolfe 2001; Sickman et al., 2003). Understanding the relative roles of warming and N deposition in driving eutrophication processes in remote lakes is an important research priority.
Water quality may also be influenced by climate change in the future more indirectly through interactions with land-cover change. Changing soils and vegetation as a result of warming can promote long-term changes in nutrient loading and in dissolved and particulate organic carbon loadings to lakes, as thicker, more organic soils develop, especially in catchments close to the current timber line. Consequences could be negative, where nutrient increase lead to eutrophication or positive in the case of DOC as increased DOC may offset the damaging effects of UV radiation.
Climate change will have a major impact on freshwater biodiversity in mountain regions. Increased temperature may cause changes to organism life-cycles (Nauwerck 1994) and species distribution. A particular concern is the potential decrease of cold stenothermic taxa (especially salmonid fish) in streams fed by glaciers, from individual lakes and from mountain regions as a whole (Simčič and Brancelj, 2002). The most cold-adapted taxa may have no opportunity for migration, with the possibility of local and regional extinctions occurring. Given the degree of endemic species in some mountains, local extinction may result in a worldwide species extinction. In addition, although suitable habitats may not disappear completely for most species, the altitudinal upward displacement of low temperature isotherms will result in a shrinking of habitats for some species. Because of their shorter life-span and ease of migration through the connected drainage network, changes in the distribution of freshwater biota will precede apparent changes in the surrounding terrestrial ecosystems. Therefore, changes in freshwater ecosystems may serve as early warning indicators of wider changes taking in place in mountain environments. Finally, milder weather will allow low altitude species, including alien species, to expand their distribution upwards and invade currently unsuitable habitats. In arid areas mountains may become a refuge for some previously lowland species.
Mountain lake ecosystems are still less understood than most lake types, but recent research has shown that few if any mountain lakes are pristine. Almost all are contaminated in some way by atmospherically transported pollutants, and in some cases the level of contamination is sufficiently high to have caused significant ecological change. Whilst some recovery from acidification might be expected in the future there are remaining threats from increased nitrogen deposition and trace metal and persistent organic pollutant bioaccumulation. Some of these threats may become greater if climate change causes a redistribution and/or remobilisation of these toxic substances.
Understanding how climate change influences mountain lakes both directly and indirectly by modifying catchment processes and the behaviour of pollutants is central to future research. Emphasis especially needs to be placed on problems associated with the melting of rock and ice glaciers, the interactions between climate change and biogeochemical processes with respect to enhanced weathering rates, pollutant release and nutrient generation, the role of long-range transported nitrogen and the threats to aquatic biodiversity posed by future climate change. Continued research and monitoring is essential to understand and protect mountain environments themselves and to manage better downstream goods and services provided by aquatic ecosystems in mountains. Such understanding is of wider benefit as changes in mountain lake ecosystems can forewarn impending change in areas far beyond the mountains themselves.
The data used in this paper have been generated through a succession of EU-funded projects between 1991 and the present day, principally: AL:PE 1: 1991–1993 Acidification of Mountain Lakes: Palaeolimnology and Ecology; AL:PE 2: 1993–1995 Acidification of Mountain Lakes: Palaeolimnology and Ecology; MOLAR:1996–1999 Measuring and modelling the dynamic response of remote m ountain lake ecosystems to environmental change, a programme of mountain lake research (http://www.natur.cuni.cz/hydrobiology/molar/), EMERGE: 2000–2003 European mountain lake ecosystems: regionalisation diagnostics & socio-economic evaluation (http://www.mountain-lakes.org/) and Euro-limpacs: 2004–2009 Global change impacts on European freshwater ecosystems (http://www.eurolimpacs.ucl.ac.uk/). An earlier version of this paper was written as a contribution to a workshop on Mountain environments that took place in Obergurgl, May 2007. We are grateful to the very many colleagues for their contributions to mountain lake research and to their scientific help and support over the last 15 years.