Convenor: UIBK

One of the main aims of the second phase of the AL:PE project was to explore the extent to which atmospheric pollutants other than sulphur and nitrogen compounds were present in remote mountain lakes. The pollutants considered were trace metals (especially lead, cadmium and mercury), persistent organic compounds (PAHs and PCBs) and fly-ash particles (especially spheroidal carbonaceous material). The results demonstrated that all sites, even the most remote site on Svalbard, were contaminated, that many of these substances were present in the kidneys and livers of fish, and that the history of pollution at each site could be clearly traced from the sediment record (Wathne et al. in prep., Battarbee et al. in prep., Vilanova et al 1995).

In addition, recent research has highlighted the role of natural dusts, especially those of Saharan origin, not only in buffering acidity in mountain lakes (Camarero et al. 1995, Psenner & Nickus 1986) but also causing a significant increase in sulphate, nitrate, ammonium and trace metal deposition (Maupetit 1992, EUROTRAC 1991, 1992, 1993).

These pollutants and dusts all have differing impacts on mountain lake systems. Alkaline dusts may be beneficial in helping to neutralise acidity, but may be pollutant carriers; fly-ash particles are excellent indicators of fossil-fuel combustion but do not impair the functioning of the ecosystem; whilst some metals and organic compounds enter food chains and accumulate in fish. Deposition of these substances is often highly episodic and their relative roles are likely to vary from site to site depending on a range of lake and catchment characteristics.

Consequently, there is a need for detailed studies of the deposition and transport of pollutants through high altitude catchment - lake systems that can be used to develop pollutant transfer models including interactions between pollutants, their impact on biota and their accumulation in sediments. Such models can be best developed using naturally occurring (²¹⁰Pb) and fallout (¹³⁷Cs) radioisotopes and chemically inert carbonaceous particles as pollutant tracers.

The high frequency and number of field and laboratory analyses needed in this work package require a sub-set of MOLAR sites which combine remoteness with reasonable ease of access and a minimum of installations. The sites are well distributed over the study area along a low (Ovre Neadalsvatn) to high (Starolesnianske Pleso) pollution gradient (Figure 1, Table 1).

The more specialised measurements, such as air-borne dust sampling and speciation of lead, organochlorinated compounds and PAH, that require on-site electricity are done only at Gossenkollesee, Estany Redo and Ovre Neadalsvatn where nationally funded installations are available.

Protocols for sampling and analysis for many determinands are available from the AL:PE project. This includes frequent analytical quality control tests organised by NIVA and CNR. For the analysis of mercury, lead and cadmium the certified procedures of NIVA are used. Organochlorinated compounds and PAH are analyzed according to the protocols of CSIC, and ²¹⁰Pb and ¹³⁷Cs are measured according to the protocol of ULIV. SCP analyses follow Rose (1994) and Rose & Juggins (1994), and fish physiology and histology arene according to Hofer et al. (1994). Methods are continually developing and new analyses are reviewed and harmonized in order to increase the performance of the participating laboratories and to create a consistent database.

The consistency of results from complex and highly specialized analyses (e.g. mercury, PCB, radiotracers, SCP, histopathology, enzymes) is guaranteed by allocating analysis to one laboratory. Participating national projects have been peer reviewed in order to guarantee the highest analytical standards.

Lead laboratories: UIBK, NIVA, FBG, CSIC

Bulk deposition of nitrogen and sulphur compounds is carried out in work package 1 (see 1.4.1). In this work package more detailed and comprehensive sampling including meteorological data is required. Automatic weather stations have been installed at all sites (except those already installed by national projects) to measure temperature, humidity, radiation, precipitation, wind speed and direction. Dry and wet deposition is automatically sampled at daily intervals at the stations where electricity is available. At all sites bulk deposition is sampled at weekly intervals. Snow is collected monthly and analysed for nutrients, SCPs, PAH, and organochlorinated compounds. Rain is sampled in special samplers (buckets and funnels made from Teflon). Soil cores for the total inventory of ²¹⁰Pb and ¹³⁷Cs have been taken from undisturbed sites in the lake catchments.

The sampling techniques used for analysis of organic pollutants encompass a wet & dry sampler, a high volume pumping system equipped with glass fibre filtration and organic resin adsorption, and other devices for the collection of snow. Sample handling involves Soxhlet extraction and column chromatography fractionation into PAH and organochlorinated compounds (clean up of fish tissue extracts will be with sulphuric acid). These are then identified and quantified by gas chromatography coupled to mass spectrometry (electron impact, selective ion monitoring) and by gas chromatography coupled to electron capture detection, respectively.

Lead laboratories: UIBK, FBG, CSIC

Although the main focus of this work package is the transfer of pollutants within the lake - catchment system, it is also important to identify the sources (local, regional or long-range) of pollutants and dusts. This is being carried out in two ways. First, event sampling and back trajectory calculation for the air masses during specified time periods is available at the Austrian, Spanish and Norwegian stations from the work of national programmes. And second, microscopic and chemical analysis of the components of dust (minerals, pollen, organic debris, SCPs) help to identify the ultimate source.

Atmospherically deposited dust may also be a neutralizing agent for acid snow and rain, a conveyor of nutrients and a carrier for organochlorinated compounds and PAHs. The collection of airborne dust is being done at the Austrian and Spanish sites using large volume air samplers, where special laboratory facilities are available on site. Collected dust particles are analysed for mineralogical and chemical composition (on cellulose acetate papers) as well as for organic carbon and nitrogen (on glass fibre filters). Organochlorinated compounds and PAHs are also analysed when the concentrations are sufficiently high. This part of the project is funded by the National Science Foundation in Austria and by the national project in Spain.

Lead laboratories: UIBK, FBG, NIVA, ULIV, CSIC, UCL

Pollutants arriving in lake - catchment systems are retained in catchment soils or transported to the lake where they are taken up by biota, lost via the outflow or stored in the sediments. In mountain lakes, snow/ice accumulation and melting play an important role in catchment - lake transfers, whilst the behaviour of the pollutants in the water column varies according to, for example, the type of pollutant, lake-water chemistry, lake residence time. SCPs are relatively large and inert and are either lost to the outflow or retained by the sediments. On the other hand, the rate of transport of metal ions and radiotracers through the water column and uptake by biota is largely determined by the competing processes of particle adsorption of the inorganic species and complexation by the dissolved organic material (Young & Harvey 1992).

Similarly, the behaviour of the organic pollutants and their interaction with inorganic and biogenic particles is determined by the partitioning between dissolved, colloidal and particulate phases. This process depends primarily on water solubility (Pavlou & Dexter 1979) but is also influenced by colloid composition and biota. Conversely, biota may bio-concentrate diffuse organic pollutants enhancing their presence in the particulate phase by incorporation into fast-sinking faeces (Elder & Fowler 1977), which will be colonized by bacteria and accumulate in the sediment.

In order to quantify these processes lake-water is sampled with a modified Schindler-Patalas-Sampler weekly at snow-melt, bi-weekly in summer and monthly in winter. Large volumes (>100 l) of water are collected through glass fibre filters (organic material) and cellulose acetate filters (inorganic components) for microscopic, mineralogical, elemental and radiochemical investigation of particles. For speciation of dissolved substances, large volumes are filtered through resins. These is done at Estany Redo and Gossenkollesee three times in spring, summer and winter.

Chemical speciation and biogeochemistry of copper, zinc, iron, lead and cobalt is carried out using Cathodic Stripping Voltametry (CSV) with ligand competition (van den Berg 1985, Gledhill & van den Berg 1994). Measurements are carried out on 1 litre samples collected from various depths in the water column of each lake during early spring, summer, autumn and winter. Comparisons will be made with samples from lowland lakes. Large volume water samples (~200-500 l) are to be collected and filtered in situ using a tangential flow pump.

Radionuclides on the particle fraction is determined by gamma spectrometry. Radionuclides in the filtrate are pre-concentrated by precipitation or resin exchange and analysed either by gamma spectrometry, or, if activity levels are too low, by alpha spectrometry. Sampling is carried out three times per year at two sites (Estany Redo and Gossenkollesee).

Sedimenting particles are collected at all sites at monthly intervals by means of an array of four sediment traps with a diameter of ca. 7 cm at two different depths. Sediments are to be sampled in those non-AL:PE lakes where the information is missing and cut into 2.5 mm slices for chemical analysis including metals, organochlorinated compounds, PAH, SCPs and radiotracers. On the basis of these analyses pollutant budgets will be established, including losses via the outflow and to the bottom sediments.

Lead laboratories: UIBK, NIVA, CSIC

The concentration of toxic substances in fish tissues is a sensitive measure of air pollution. In acidic mountain lakes they are excellent indicators for two reasons: (i) due to the soft water they accumulate toxic substances at a higher rate than fish in hard water; and (ii) in most of these lakes they have a long life (sometimes exceeding 20 years).

Fish muscle and liver are important target organs for mercury, cadmium and persistent organics. Formation of glycogen vacuoles in hepatocyte nuclei was one of the most striking acid-specific impacts found during the AL:PE project. This metabolic problem significantly increases during the period of extensive feeding at ice-break (Hofer et al. 1994). However, the relationship between pollution deposition, pollutant concentrations in the water column and pollutant concentrations in fish tissue can be complex. In this activity we separately assess the roles of trace metals and of organics.

Mercury bioaccumulates strongly in fish, and human exposure to methylmercury is almost wholly due to consumption of fish (Clarkson 1992). Although there has been little study of the potential toxicological effects of mercury accumulation on fish populations, new data seem to indicate effects on behaviour, reproduction and feeding ability (Wiener & Spry 1994, Fjeld et al. in prep.). Methylation of inorganic mercury(II) in the environment is primarily a microbial process taking place in lakes and catchment areas, but wet deposition also contains methylmercury (Bloom & Watras 1989). The accumulation of mercury in fish is usually positively correlated to age, atmospheric deposition, water temperature, organic content in water, and negatively to pH, buffering capacity and selenium deposition (Wren & Stokes 1988, Fjeld & Rognerud 1993, Wiener & Spry 1994).

One of our objectives in this study is to examine how changes in atmospheric concentration and deposition affect mercury concentrations in fish from remote lakes. These lakes receive different atmospheric loadings and are well suited to elucidate the effect of this variable due to the following:

- the same fish species (brown trout, and arctic char) are present in most of the lakes and comprise self-contained in-lake populations;

- the populations are feeding only on zooplankton and/or zoobenthos, allowing statistical comparisons between concentrations in a "standard fish" from different sites;

- mercury enters the food chains rapidly in lakes with a high littoral-to-pelagic-area ratio (Harrison et al. 1990) and low-alkalinity;

Because of these factors, most of the observed variability in mercury concentration in fish will be caused by differences in atmospheric deposition and temperature/duration of ice-free periods. The data will be used in statistical modelling to evaluate the importance of these factors. Future mercury concentrations in fish can then be predicted as a function of different developments in atmospheric deposition and temperature caused by climatic changes.

Unlike mercury, cadmium and lead do not biomagnify in aquatic food chains (Wren & McCrimmon 1983). However, the AL:PE project has shown that concentration in target organs such as kidneys and liver can increase significantly with size and age (Wathne et al. 1995). In addition, the bioavailability of lead and cadmium is increased in low-alkalinity waters (Wiener & Stokes 1990), making concentrations in fish more susceptible to changes in loadings of metals. These lakes are therefore also suitable for evaluating the importance of lead and cadmium deposition on concentrations in fish.

In this study analyses of all three trace metals are being carried out on a large range of individuals and tissue types from lakes sampled in this work package and work package 1. To obtain an estimate of the adjusted mean with an acceptable statistical significance it is necessary to collect 25 fish of different length in the interval 15-30 cm from each lake. Total length, weight and sex are recorded and scales or otoliths collected for age estimates. Approximately 40 g of bone and skinless dorsal axial muscle, liver and kidneys are removed from each fish and placed separately in polyethylene bags and stored in a freezer. All samples are carefully packed and shipped frozen to NIVA where the metal analyses are performed according to the techniques of Fjeld & Rognerud (1993).

In addition, the activity of the mixed function oxidase (MFO), which is specifically induced by many lipophilic organics, is to be measured in fish livers along an European north - south transect. Bile analyses will also be carried out. The fish reproductive capacity (analysed at all sites by NIVA), histological analyses of gills, liver and kidney (analysed at all sites by UIBK) and morphological and physiological parameters of gills and blood (analysed by CNRS) will be correlated with the concentration of trace metals in the tissues.

Schindler (1995) showed that even in remote sub-arctic locations, fish could contain such high concentrations of PCB, toxaphens, DDT and other organochlorines that aboriginal fisheries for lake trout had to be closed. Concentrations in rain, snow and surface waters may be scarcely detectable, but biomagnification of several orders of magnitude by food chains has caused fish to be a threat to humans and other consumers of fish (Kidd et al. 1995). Results from the AL:PE project (e.g. Battarbee et al. in prep.) show that fish from high altitude lakes in Europe are contaminated by organic pollutants and that a record of this pollution is contained in the bottom sediments.

In order to assess fish uptake of organic pollutants (PAH, PCB, DDTs and hexachlorobenzene) the livers from 20 fish sampled at the same time and from the same sites as trace metals are to be analyzed. The results will be correlated with the ecotoxicological results from trace metal analysis.

Lead laboratories: ULIV, UCL

The value of the research proposed in this work package is not only in understanding pollutant fluxes at these specific, intensively monitored sites but also in calibrating a model for pollutant transfer that can be used for other remote sites where only sparse information is available, and where it is important to be able to predict the extent to which specific pollutants are taken up in the food chain, lost through the lake outflow or retained in sediments. They are also essential if sediment records are to be accurately and quantitatively interpreted (Appleby et al. 1995).

Existing models of pollutant deposition and transport through catchment - lake systems (Appleby & Smith 1993) are being reviewed and developed to accommodate the particular features of high mountain lakes. Attention is being paid to the significance of transport from the snow pack during the annual thaw. Transport through the water column takes into account factors such as water residence time, chemical speciation, particle size distribution and concentration. Model outputs will include accumulation in fish, and the sediment record. Accumulation in bottom sediments will include estimation of the influence of post-depositional re-distribution (both horizontal and vertical).

The models will be calibrated using data on deposition and distribution of pollutants from the other AL:PE and MOLAR sites. Validation will be carried out using fallout ²¹⁰Pb and ¹³⁷Cs isotopes as tracers. Unlike the pollutants to be modelled these radionuclides have relatively simple and well known atmospheric fluxes (at least from a qualitative point of view), and are ideal tracers for studying transport processes. Quantitative estimates of the atmospheric radionuclide fluxes can be made from measurements on rain water samples and soil core archives, and the large quantity of data on ²¹⁰Pb and ¹³⁷Cs generated by the palaeolimnological component of the AL:PE project can be used to test models in a wider range of contexts.

In addition, inert SCPs derived from fossil-fuel combustion can also be used in the validation process. SCPs are transported conservatively so their comparative fluxes can be used to evaluate the relative importance of transport via the particulate as opposed to the aqueous phase.

Rain water and snow pack samples are being analysed for ²¹⁰Pb activity. Soil cores from undisturbed sites have been sectioned and analysed for ²¹⁰Pb and ¹³⁷Cs and the inventories used to determine the mean annual ²¹⁰Pb flux from the atmospheric and historical ¹³⁷Cs fallout record. Data from the sediment dating programme is used to quantify outputs to the sediment record. SCPs are being measured in deposition gauges, sediment traps and sediment cores. The model will finally be tested using data for other sites from the AL:PE/MOLAR project database.