Despite the sparseness and the exceedingly slow growth rate of its vegetation, the Arctic terrestrial ecosystem is known to be a site of net accumulation of organic matter within the global biosphere, and at the same time, an important source of organic matter feeding the Arctic Ocean. These two great ecosystems interact through the hydrosphere, represented by surface water runoff and its temporary passage through streams, ponds, lakes and rivers. As part of the Swedish Canadian Tundra Northwest 1999 expedition to the Arctic parts of Canada in July and August, we set out to study the role of freshwater bacterioplankton, i.e. free-living bacteria, in organic matter decomposition, as well as to characterize their growth substrate, the dissolved organic matter (DOM). The unique circumstances of the Arctic hydrological cycle (low precipitation inputs but even lower losses through evapotranspiration) result in a wealth of shallow lakes, ponds, and bogs spread across the tundra. Although all three of these aquatic systems were sampled, the emphasis in this study was placed largely on lakes and ponds. In this report we introduce and briefly discuss some preliminary data from the expedition.

The large General Circulation Models (GCMs) that are currently being developed to predict the potential for climate change in different ecosystems as a result of the build-up of greenhouse gases, point unanimously to the Arctic as the region most likely to experience large shifts in temperature and precipitation over the coming decades. If these changes come to pass, what will be the consequences for the communities living here? The better we know how they function today, the better we will be able to understand and predict future conditions. Although Arctic lakes and ponds have received appreciable limnological interest, there has been no synoptic study encompassing such a large west-east and south-north gradient as the Tundra Northwest 1999 expedition made possible. On the other hand, the study was necessarily limited in that it did not include possible seasonal variability, as each system was sampled only once. Hence we cannot know to what extent values are representative for the total ice-free period. However, if the character of lakes and ponds can be shown to be similar over the whole, vast sampling area, this would be an indication that it might be possible to generalize from our spot samples. These studies of bacteria, their activity, and their reactions to changes in temperature and organic matter availability were carried out within Theme D of Tundra Northwest 1999: ”Freshwater ecosystems in the high Arctic”.

Both conceptually and methodologically aquatic microbiology has made rapid advances in recent years. Thus, studies of bacteria and their growth substrates during Tundra Northwest 1999 were made with methods not extensively used in the Arctic before. The older view of bacteria as mineralizers of detrital organic matter, primarily in the sediments, has been substituted by the new view that pelagic (water-column) bacteria are numerous and active, that they compete with phytoplankton for nutrients (phosphorus) and yet that they depend on algal exudates as a carbon source. A whole new food-chain, the microbial loop, has been conceptualized. Also the idea that lakes are generally net autotrophic systems (primary production is greater than community respiration; i.e. the lake acts like a plant) has been revised, as a majority of lakes are undersaturated with oxygen and supersaturated with respect to carbon dioxide (indicating net heterotrophy; i.e. the lake functions like an animal that needs to feed on inputs from the drainage basin). Accordingly, it has been shown that bacteria may use allochthonous carbon sources coming from outside the lake, e.g. humic matter from plant breakdown, which can then be followed up the food-chain of temperate lakes all the way to the top predators, by measuring the abundance, at each level of the chain, of stable isotopes of carbon and nitrogen. So in this new understanding, we see that bacteria are not only decomposers but that they also play a central role in the development and maintenance of the entire food-web of aquatic organisms, including the fish and their consumers.

Along with the emerging understanding of the ecological importance of freshwater bacteria, new developments have shown that they are not the only ecosystem component responsible for organic matter breakdown. In recent years it has also been shown that purely abiotic, photochemical reactions involving DOM, or a combination of photochemical and bacterial mineralization pathways of DOM, are of great importance for the carbon turnover of lakes and surface layers in marine systems.

This has emerged as a ”by-product” of research dealing with the influence of solar ultraviolet-B (UV-B) radiation on aquatic organisms and ecosystems, motivated by stratospheric ozone thinning, with its associated increase in UV-B radiation reaching the surface of the earth.

The new paradigms described above have emerged from research in temperate waters, while few studies have been performed of tropical and polar systems, where conditions may be dramatically different. Yet Arctic/Antarctic lakes are especially well suited for testing general ecological theories, e.g. with respect to feedback controls on food-chain dynamics, and the importance of allochthonous organic matter. The characteristics that confer this advantage can be summarized as follows: their simple trophic structure, their short growing seasons, the harsh climate setting which produces feeble levels of vegetation, the presence/absence of organic matter accumulation, and the direct dependence of all aquatic life on small changes in temperature.

Firstly then, tundra lakes and ponds have a very simple biotic structure, due to harsh environmental conditions, lack of fish, and zooplankton communities being poor in species. Under such conditions it is easier to analyse the influence of various biotic and a biotic factors on the systems, especially since lakes/ponds with or without fish may be found close to each other. Secondly, many of the Arctic ponds and lakes have a very short ice-free period. Although phytoplankton can survive even under a thick ice cover, and benthic primary production through attached algae may be appreciable, it may be hypothesized that biota in Arctic lakes to a large extent depends on allochthonous matter. Wetlands and peat deposits, sources of allochthonous DOC for surface waters, are conspicuous elements in many tundra regions, and tundra ponds may be highly influenced by such material. Furthermore, due to the shallow depth of these ponds (usually<1m), UV-radiation penetrates through the entire water column. In addition, many high-latitude lakes are located in extremely poor and barren areas, with little soil organic matter. Waters draining such areas are poor in dissolved organic compounds, which are generally responsible for most of the absorption of UV-radiation in natural waters. Hence, damaging UV-radiation may penetrate to substantial depths. On the other hand, it is perhaps more common in Arctic than in lower latitude lakes that unstable and unvegetated soils, as well as influence from glacial melt water, cause mineral turbidity which reduces the penetration of solar radiation.

Another issue that is in the focus of this project is the temperature dependence of heterotrophic bacterial activity in the water. Bacterial metabolism is strongly temperature dependent. In cold, high latitude marine waters, bacterial degradation of organic matter has been demonstrated to be slower than the production of organic matter by primary producers (phytoplankton). Accordingly, during periods of high phytoplankton production, there is a build-up of DOM in the water that does not occur in warmer waters. This is because the bacteria, due to limitation by low temperatures, respond to the input of organic substrates with a longer time lag in high latitude seas than in areas using warmer water. It is reasonable to hypothesize that the same phenomenon occurs in fresh waters, although this has so far not been studied systematically. To address this question, we performed experiments with water from several of the ponds and lakes visited during Tundra Northwest 1999. Bacterioplankton were incubated in gradients of temperature and DOM concentration, in order to determine their ability to react to increased substrate availability at different temperatures. Our aim for the Tundra Northwest 1999 expedition was to sample a wide selection of Arctic lakes and ponds for investigations of bacterioplankton and dissolved organic matter and to shed light on the biogeochemistry and microbial ecology of freshwater tundra ecosystems; in general to obtain basic information about these systems, and in particular with regard to the discussion above concerning UV radiation and temperature.

The following questions were addressed:

1. How abundant are the bacteria in surface waters and what is their activity (protein production)? Are there systematic differences between lakes and ponds?
2. What is the quantity and quality of dissolved organic matter of autochthonous and allochthonous origin in Arctic freshwaters?
3. To what extent can photochemical processes contribute to abiotic and biotic turnover of dissolved organic matter in Arctic lakes and ponds?
4. Which is the limiting nutritional factor for bacterial growth potential in Arctic ponds and lakes, carbon, nitrogen or phosphorus, or a combination of these elements?
5. What are the short and long-term effects of temperature fluctuations on bacterial substrate utilization and growth rate in Arctic freshwaters?
6. To what extent does mesozooplankton (copepods, cladocerans) community composition influence bacterial production and abundance in Arctic ponds and lakes?
7. Are Arctic ponds and lakes supersaturated with CO², indicating net heterotrophy of the system?
8. What is the relation between bacterial growth potential, bacterial growth and abundance in situ, and various physicochemical (dissolved organic carbon concentration, quality of DOM, temperature, pH, water residence time) and biological parameters (virioplankton, zooplankton, phytoplankton)?

Methods

During the Swedish-Canadian Tundra Northwest 1999 expedition with the ice-breaker Louis S. St-Laurent, 17 sites were visited, in temporal order: Ungava Peninsula, Melville Peninsula, Somerset Island, Bathurst Island South, Bathurst Island North, King William Island, Wollaston Peninsula, Amundsen Gulf South, Banks Island South, Ivvavik National Park (close to the Canadian border with Alaska), Cape Bathurst, Banks Island N, Melville Island S, Ellef Ringnes Island N (N magnetic pole), Ellesmere Island S, Devon Island S and Baffin Island S. Sites were distributed in a west-east and south-north transect, from 63° to 140°W, and from 62° to 78°N. The first site was visited on 2 July, and the last on 30 August, 1999. The first lakes were sampled during early summer when many lakes were still ice-covered, while at the last sites autumn had arrived.

At each site at least one lake and one tundra pond were chosen for in-depth investigations, while several other ponds and lakes were visited for survey-type sampling.

Sampling in the main lake was performed from a Zodiac, while all other sampling was carried out from the shore, using rubber boats or waders. Polyethylene or polycarbonate  containers of between 0.5 and 10 l capacity (depending on the amount of water needed for various experiments and analyses) were filled with surface water. Containers were stored cool, usually by immersing them in lake water (generally around 5°C). In most water bodies, samples were taken for bacterial abundance (glutaraldehyde fixed samples), bacterial protein production (leucine incorporation method), dissolved organic carbon (high temperature combustion and infrared-detection of CO₂), UV/visual absorbance spectra (spectrophotometric characterization), and average molecular weight of coloured DOM (gel filtration using high-pressure liquid chromatography). Temperature was recorded at the surface with an electronic thermometer. In addition water was always taken for basic chemical parameters (conductivity, pH, total inorganic carbon, nutrients).

Bacterial and viral numbers in situ and in experimental incubations will be counted after the cruise using a flow-cytometric protocol, with selective staining of nucleic acids using the stains Syto-13® and SYBR Green I. Bacterial production analysis, assessed by tracking the incorporation of ³H-leucine into newly synthesized proteins was carried out on board the ship. Absorbance scanning, total IC analysis using gas chromatography and other more simple chemical analyses were also performed on board. All other analyses will be conducted after the cruise on cold-stored or frozen water samples.

In addition to quantification of in situ bacterial and chemical parameters, several experimental studies were made:

1. Limiting nutrients for bacterial growth. These experiments were carried out with water from a total of 8 lakes and ponds. The water was filtered to remove bacterial grazers (GF/F glass fibre filters) and carbon (glucose, 1 mg C/l), nitrogen (nitrate, 100 μg N/l) and phosphorus (phosphate, 10 μg P/l) were added separately and in all possible combinations. Bacterial growth response to these additions was measured at intervals of several days as leucine uptake and bacterial abundance.
2. Response of bacteria to fine-scale change in temperature, and the interaction of temperature and organic matter availability in the control of growth rate. Experiments to assess the short term response of bacterial production were performed on subsamples incubated with radioactive leucine for 4-6 hours in a temperature gradient block (0-12 °C). Experiments on the long-term response were incubated in a temperature gradient for 6 to 10 days. To specifically study the interaction of temperature and access to organic substrate (DOM), a two dimensional experimental design was employed, including different additions of organic matter at different temperatures. The organic matter was added as concentrates, either obtained as described below (G), or extracted from DOM that had been coagulated in to lake foam. The bacterial response was followed in terms of abundance, production (leucine incorporation), respiration (carbon dioxide formation), and cell replication activity. Actively replicating cells were identified as those with detectable DNA production, as indicated by the uptake of BrdU, an analogue of the nucleic acid precursor thymidine, in to the DNA, or detectable respiratory activity, as indicated by the reduction of CTC, a redox sensitive dye that is converted into a fluorescent precipitate upon interaction with the bacterial respiratory chain. The presence of BrdU can be detected immunologically with a fluorescent monoclonal antibody; CTC reduction is quantified by flow cytometry.
3. The influence of solar radiation on DOM quantity and quality and the impact on subsequent bacterial growth potential. Experiments were performed for 3 sites on leg 2, each time with water from one lake and one pond. Water was first sterile filtered (0.2 μm) and then exposed to 1-2 full days of solar radiation in quartz tubes, and incubated on the deck of the ship. Al-foil wrapped tubes served as dark controls. The incident radiation, i.e. PAR (photosynthetically active radiation), UV-A and UV-B were logged during the entire irradiation period. After exposure, water was sampled for DOC and IC analyses, and was then inoculated with a natural bacterial assemblage (GF/F filtered water, 10% addition). N and P were added to ensure carbon limitation. Bacterial response was measured as abundance and leucine uptake, with intervals of several days.
4. Growth potential of bacteria in relation to water chemistry (DOM quantity and quality, nutrients, pH). For this study, a larger set of lake and pond-samples were used. Water was sterile filtered and a 5% (volume) mixed bacterial inoculum from the complete set of waters was added. The resulting cultures were inoculated under identical conditions (incubation bottles, temperature etc.) and growth was followed by quantification of bacterial numbers until a growth plateau had been reached (<10 days).
5. Influence of copepods and daphnia on bacteria. This study was based on the hypothesis that the very large biomasses of mesozooplankton in fish-free ponds would strongly influence bacterial abundance and production, either through excretion of nutrients or through direct ingestion. Two experiments were performed. One assessing the impact of various numbers of zooplankton (copepods, cladocerans) on bacteria and phytoplankton abundance. The other study surveyed the impact of copepods and cladocerans on bacterial abundance at various temperatures.
6. For each site a test was carried out of the concentration of leucine where uptake saturation occurred. For one pond and one lake, subsamples were incubated for 4-6 hours with a mixture of cold and radioactive (³H) leucine, with concentrations ranging from 5-200 nM. These experiments were made to get an idea of what leucine concentrations should be used in measurements of bacterial protein production, and also to test for differences in leucine uptake capacity between lake and pond bacteria, as well as between bacterial populations from different sites.
7. For post-cruise detailed studies of bacterial communities and their interaction with the natural substrates available in the tundra lakes and ponds, bacteria and DOM were collected on leg 1 from at least one lake or pond at each site. Bacteria were stored frozen (-80°C), and DOM was collected using two independent methods, in most cases both were performed at each site. Large volumes of water (100-400 L) were prefiltered (0.4 μm) and the DOM was collected by reverse osmosis, or by adsorption on to an anion exchange resin (DEAE). Samples were frozen (-20°C) and will be further processed upon return to the laboratory in Uppsala.

Provisional results

For each site it was generally possible to find a lake of suitable dimensions, from a few hundred meters up to over a km in length, and with maximum depths of between 1.5 and >50 m. Most lakes we sampled had a char population (usually anadromous), except the most shallow ones (<2 m), where the presence of fish was probably prohibited by complete winter freezing of the water column. Ponds on the other hand were distinctly different: smaller in size (a few m up to 100m), shallow (generally <1 m depth, often 20-30 cm), and without fish. There were often dramatic differences in the zooplankton community between lakes and fish-free ponds, with small and transparent mesozooplankton in char lakes, and large, pigmented Daphnia in the fish-free ponds. In addition, these systems often had dense populations of larger crustaceans.

Only a few of the bacterial numbers have been counted to date and we will therefore mostly report bacterial production levels. These are at this stage relative, since on board liquid scintillation values have to be calibrated against quench-corrected CPM values, which require recounting on a more advanced liquid scintillation counter in Lund.

Bacterial leucine in corporation was generally saturated at between 50 and 100 nM leucine. Uptake in lake water was typically very low. However, by extending incubations to 6 hours, a CPM value significantly different from the blank was obtained. Initial tests demonstrated a linear incorporation of leucine up to an incubation time of at least 7 hours.

When comparing lakes and ponds with respect to bacterial production, it is evident that ponds had a much higher bacterial production. This is most likely an effect of a combination of higher temperatures and higher DOM and nutrient availability, although ponds also showed a surprisingly strong phosphorus limitation of bacterial growth (see below).

In most cases there was a strong positive short-term effect of increased temperature on bacterial protein production. At present, we have preliminary results from only one of the long-term experiments on temperature effects. After 8 days, it was evident that bacteria had responded positively to the addition of organic substrates, even at temperatures close to freezing, although lower bacterial density was achieved than at higher temperatures. The short-term temperature response indicates that bacteria in Arctic lakes and ponds are not a truly cold-adapted community, but that they rather have the ability to rapidly respond to temperature fluctuations. This should be an advantage in tundra ponds, where diurnal temperature fluctuations may be large, due to the shallowness of the systems. For bacteria in a 20 m deep lake, temperature fluctuations should be much more gradual. However, a preliminary analysis did not reveal any marked differences between bacteria from lakes versus ponds with respect to the reaction to rapid (a few hours) temperature changes. The ability of bacteria to take advantage of enhanced substrate concentration even at low temperatures suggests that their growth is not severely limited by low temperatures. Further data analysis will furnish insights into the regulation of bacterial response to temperature and substrate fluctuations, e.g. whether changes in activity are due to alteration between dormant and active states in fractions of the total bacterial community, or whether the response is due to changes in metabolic rates more evenly distributed among all cells.

In most cases bacterial production was strongly phosphorus limited. An addition of phosphate alone in most cases caused a several fold increase in bacterial production.  Neither nitrate, nor glucose had any effect, often not even in combination with phosphate. This result was surprising, since it could be expected that at least bacterioplankton in lakes with low DOM content should be carbon, rather than phosphorus limited.

Although solar radiation altered both the average molecular weight of coloured DOM and the absorbance spectrum of DOM somewhat, the impact on bacterial growth seemed to be non-significant or at least slight. An initial inhibition of bacterial growth in the first photo-experiment (performed with water from a lake and a pond in the Ivvavik National Park) indicates a temporary build-up of inhibiting concentrations of reactive species, detrimental to bacterial growth. A more complete evaluation of these experiments has to be based on bacterial abundance.

Grazing and growth potential: These experiments rely fundamentally on analyses of bacterial numbers and will not be discussed here.

The amount of dissolved organic carbon varied between 0.43 and 19 mg C/l for the different lakes, while the corresponding figure for ponds was 0.17 to 40 mg C/l. With few exceptions, both lakes and ponds were rather clear and preliminary results indicate that the average molecular weight of coloured DOM did not vary much between the different systems. For all systems, except the Baffin Island site, total inorganic carbon was very high, both in lakes and ponds (5-100 mg C/litre). This is most likely due to a strong influence from alkaline bedrock and soils in the respective drainage areas.

Preliminary conclusions

Our results indicate a surprisingly large difference in bacterial activity between tundra lakes, even small ones, and tundra ponds. Short-term changes in temperature appear to have a direct positive impact on the activity of bacteria in Arctic ponds and lakes, but the bacteria are able to respond to increased substrate concentration at temperatures close to the freezing point of lake water. Bacterioplankton showed strong signs of phosphorus limitation, also somewhat unexpectedly. Indirect solar radiation had comparatively little impact on bacterial growth (through phototransformations of DOM to more bacteria available forms), even though solar radiation was intense and dark hours were few.