I have, in this thesis, studied the interactions between gray-sided voles (Clethrionomys rufocanus) and tundra vegetation, on islands in, and mainland sites close to the lake Iešjávri, in northern Norway. As isolated islands are virtually free of predation, I have been able to compare plant-herbivore interactions in the presence and absence of predators. I transplanted vegetation from an island with predators and voles, to predator-free islands with and with out voles. The results reveal the existence of a terrestrial trophic cascade as voles had a severe impact on the transplanted vegetation on the predator-free islands, but only minor effects on the mainland where predators are present. Moreover, this study shows that plant defence was only a successful strategy when predators were present. Voles reduced the abundance of all available plants during winter on the predator-free islands. The results imply that cascading effects of predation are most important for well-defended plants with grazing-sensitive morphology as these plants escape herbivore impacts in the presence of predators but are vulnerable in their absence I studied the recovery of intensively grazed vegetation by building exclosures on islands that have been heavily grazed by voles for almost a decade.This study shows that the collective biomass of vascular plants recovered completely on three years, when voles were excluded. Although most species that are dominating the ungrazed vegetation recovered rapidly in the exclosures, the vegetation did not simply return to its ungrazed state. Herbaceous plants increased and there were pronounced differences in response among evergreen dwarf-shrub species. The semi-prostrate and tannin rich crowberry (Empetrum nigrum), showed the strongest recovery of all species, while the erect lingonberry (Vaccinium vitis-idaea) only showed weak signs of recovery. Thus, growth form determined the potential of plants to recover, whereas the trade-off between defensive investments and capacity to recover was weak or absent. I studied the interaction between gray-sided voles and their main winter food plant, bilberry (Vaccinium myrtillus) on islands in and mainland sites close to the lake Iešjávri. I compared the abundance, population structure and palatability of bilberry ramets between vole-free islands, islands with voles but no predators and mainland sites with both voles and predators. Voles decreased the abundance of bilberry on the mainland, but the effect was much stronger on predator-free islands. Bilberry was fairly tolerant to grazing as it partially compensated for the lost tissue by producing more new ramets. Moreover, a cafeteria experiment showed that voles preferred the ramets from predator-free islands, which is inconsistent with conjectures emphasizing inducible plant defenses. The vole-bilberry interaction lacks features of delayed density dependence that could explain the vole cycles. I conducted a clipping and fertilization experiment to further investigate the effects of herbivory on palatability of bilberry shoots. Fertilization decreased the concentration of condensed tannins in shoots of bilberry and voles preferred fertilized and clipped shoots. I found no indication of induced defense that could reduce the palatability of bilberry twigs in response to herbivory. The relationships between gray-sided vole densities, levels of invertebrate herbivory and chemical quality of leaves of Northern willow (Salix glauca) were studied on islands and mainland sites with contrasting vole densities. I found a positive correlation between level of invertebrate herbivory and vole density. The number of leaves per shoot, leaf size and leaf nitrogen content were also positively correlated with vole densities, while leaf C/N ratios were negatively correlated with vole densities. The positive correlation between vole densities and level of invertebrate herbivory is probably due to a facilitative effect of voles on invertebrate herbivores, mediated through changes in plant chemistry.

The distribution of plant species and functional traits in alpine and arctic environments are determined by abiotic conditions, but also by biotic interactions. In this thesis, I investigate interactions among plants and herbivory effects on plant community composition and plant functional traits in three different regions: Swedish Lapland, Beringia (USA/Russia) and Finnmark (Norway). Reindeer grazing was found to be extensive in southern Lapland and had limited effects on plant community composition and seedling germination. However, reindeer presence was found to influence plant functional traits, particularly in the subalpine birch forest. Tall herbs were lower and had lower SLA when reindeer were present, while small herbs showed an opposite pattern. The contrasting effects on the two herb groups are probably explained by a competitive release for small herbs when the tall herbs are suppressed by reindeer. Rodents had the largest relative impact on plant community composition in southern Lapland and this is consistent with the study from Finnmark, where rodents heavily affected dwarf shrubs on predator-free islands. With no predators present, vole densities increased profoundly and almost depleted some dwarf shrub species. These results support the idea that small mammals in arctic and alpine tundra are controlled by predators (i.e. top-down). However, a decrease in the nutritional quality in a sedge after defoliation gives support for the idea that small mammals are regulated by plant quality (i.e. bottom-up). In Beringia, small and large herbivores differed in the relation to plant community composition, since large herbivores were related to species richness and small herbivores were related to plant abundance. Plant functional traits were related only to large herbivores and standing crop of vascular plants.

Phosphorus is the main limiting nutrient for plant growth in large areas of the world and the availability of phosphorus to plants and microbes can be strongly affected by soil properties. Even though the phosphorus cycle has been studied extensively, much remains unknown about the key processes governing phosphorus availability in different environments. In this thesis the complex dynamics of soil phosphorus and its availability were studied by relating various phosphorus fractions and soil characteristics to microbial respiration kinetics. The soils used represent a range of aluminium, iron, carbon and total phosphorus content, and were located in four different biomes: subtropical forest, warm temperate forest, boreal forest and tundra. The results showed that NaOH extractable phosphorus, a fraction previously considered to be available to plants only over long time scales, can be accessed by microbes in days or weeks. Microbial phosphorus availability was not related to aluminium or iron content in any of the studied systems, not even in highly weathered soils with high aluminium and iron content. This is in contrast with other studies of soils with high sorption capacity and shows the variability of factors that govern phosphorus availability in different environments. In the boreal forest chronosequence, no difference could be seen with age in total phosphorus content or concentrations of occluded phosphorus forms. However, there were lower concentrations of labile phosphorus forms in older systems, which were correlated with a decrease in microbial respiration. This was most likely related to organic matter quality in the system, and not to geochemical factors. Phosphorus availability was linked to differences in topography (water regime) and vegetation in the tundra ecosystems. The results suggest that the availability of phosphorus, both for microbes and plants, was lower on the meadow vegetation sites compared to the two types of heath vegetation. Many factors are important for phosphorus availability in soils, but these results suggest that microbes can access less available phosphorus if not restricted by carbon, and this may be important in regard to forest management practices as well as effects of environmental change.

Oil and gas exploration conducted in the 1970s left behind a legacy of abandoned well sites in the Mackenzie Delta region of northern Canada, including several in the Kendall Island Migratory Bird Sanctuary and surrounding areas. Evidence of 30 year-old well sites is present in the form of drilling mud sumps, which are mounds of disturbed tundra that contain frozen drilling-wastes. One to two years after the wells were decommissioned some of the sites were seeded with non-native grass species and fertilized to test whether these treatments could accelerate vegetation recovery and prevent erosion. The main objective of this research was to examine the long-term impact of post-disturbance seeding treatments on the vegetation recovery of drilling mud sumps.<p> Surveys of vegetation composition and environmental conditions at 12 sump sites (6 seeded and 6 unseeded) showed that, after over 30 years of recovery, seeded sumps in the Mackenzie Delta did not significantly differ from those left for natural recovery. However, seeded and previously introduced grasses <i>Festuca rubra</i> and <i>Poa pratensis</i> were found on both seeded and unseeded sumps. The undisturbed surrounding tundra seems to be resistant to invasion by these introduced grasses. However, these species could become invasive in the future, particularly in the context of warming in the North and increasing anthropogenic disturbance. The results of this study contribute valuable information on the long-term effects of revegetation treatments that is critical for making informed management decisions about the rehabilitation of industrial disturbances in the Arctic.

Oil and gas exploration conducted in the 1970s left behind a legacy of abandoned well sites in the Mackenzie Delta region of northern Canada, including several in the Kendall Island Migratory Bird Sanctuary and surrounding areas. Evidence of 30 year-old well sites is present in the form of drilling mud sumps, which are mounds of disturbed tundra that contain frozen drilling-wastes. One to two years after the wells were decommissioned some of the sites were seeded with non-native grass species and fertilized to test whether these treatments could accelerate vegetation recovery and prevent erosion. The main objective of this research was to examine the long-term impact of post-disturbance seeding treatments on the vegetation recovery of drilling mud sumps.<p> Surveys of vegetation composition and environmental conditions at 12 sump sites (6 seeded and 6 unseeded) showed that, after over 30 years of recovery, seeded sumps in the Mackenzie Delta did not significantly differ from those left for natural recovery. However, seeded and previously introduced grasses <i>Festuca rubra</i> and <i>Poa pratensis</i> were found on both seeded and unseeded sumps. The undisturbed surrounding tundra seems to be resistant to invasion by these introduced grasses. However, these species could become invasive in the future, particularly in the context of warming in the North and increasing anthropogenic disturbance. The results of this study contribute valuable information on the long-term effects of revegetation treatments that is critical for making informed management decisions about the rehabilitation of industrial disturbances in the Arctic.

Net ecosystem carbon dioxide exchange (NEE) was measured using the eddy covariance (EC) technique at a wetland tundra-sedge fen near Churchill, Manitoba, Canada during two consecutive growing seasons (2007 and 2008). Mean daily NEE at the fen (DOY 157-254) was -3.5 (± 0.26 S.E.) g CO2 m-2 d-1 in 2007 and -4.6 (± 0.36) g CO2 m-2 d-1 in 2008. The fen was a net carbon dioxide (CO2) sink during both the 2007 and 2008 growing seasons of -343 (± 79) and -450 (± 87) g CO2 m-2, respectively. Mean air temperature during the summer (June 1-August 31) was about 1°C greater than the historical average (1971-2000) in 2007 and about 2°C greater in 2008. Growing season precipitation was 107.5 mm below normal in 2007 and 359.5 mm above normal in 2008. These data suggest that if future climate change brings warmer temperatures and near-to-above average precipitation maintaining the water table near the surface, similar subarctic ecosystems will experience increased gross ecosystem productivity enhancing CO2 sequestration during the growing season.

This study examines the impact of snowmelt runoff on nutrient transfer from the terrestrial to the aquatic (lake) portion of a subarctic catchment 6 km WSW from Schefferville, Quebec. / Statistically significant differences in snow chemistry were recorded among the tundra, woodland and forest snowpacks. Significant overland substantial scouring of nutrients from the organic horizons were recorded in the tundra, woodland and forest. A calculation designed to generate, on a daily basis, 70% of the meltwater from the terrestrial catchment to the lake indicates overland flow is a predominant route of meltwater during the spring. This calculation made possible the determination of daily nutrient mass transfer from land to lake. The daily nutrient mass balance of input and output to Elizabeth Lake during springmelt indicates significant retention of land source nutrients. Deuterium/hydrogen measurements aided in determining mixing of snowmelt and lake water. The data indicate snowmelt water is undistinguishable from lake water at depths greater than 1.0 m. / It is concluded that snowmelt runoff is an important event for nutrient transfer from terrestrial to aquatic systems in this environment.

Net ecosystem carbon dioxide exchange (NEE) was measured using the eddy covariance (EC) technique at a wetland tundra-sedge fen near Churchill, Manitoba, Canada during two consecutive growing seasons (2007 and 2008). Mean daily NEE at the fen (DOY 157-254) was -3.5 (± 0.26 S.E.) g CO2 m-2 d-1 in 2007 and -4.6 (± 0.36) g CO2 m-2 d-1 in 2008. The fen was a net carbon dioxide (CO2) sink during both the 2007 and 2008 growing seasons of -343 (± 79) and -450 (± 87) g CO2 m-2, respectively. Mean air temperature during the summer (June 1-August 31) was about 1°C greater than the historical average (1971-2000) in 2007 and about 2°C greater in 2008. Growing season precipitation was 107.5 mm below normal in 2007 and 359.5 mm above normal in 2008. These data suggest that if future climate change brings warmer temperatures and near-to-above average precipitation maintaining the water table near the surface, similar subarctic ecosystems will experience increased gross ecosystem productivity enhancing CO2 sequestration during the growing season.

Up to 59 small tundra lakes were sampled in the uplands east of the Mackenzie Delta, NWT, Canada, in order to assess the impact of permafrost thaw on pelagic nutrient concentrations and biota. Permafrost thaw did not affect the concentrations of nitrogen, phosphorus or organic carbon in the water column of the study lakes. Instead, nitrogen and organic carbon concentrations were positively related to relative catchment size, and phosphorus concentrations were negatively related to maximum lake depth. Lakes affected by permafrost thaw did have lower water colour. In lakes not affected by thaw, where plankton production could have been light limited due to high water colour, the TP-chlorophyll a relationship was weaker than in lakes affected by thaw, where light limitation was probably weaker. A model selection analysis for chlorophyll a concentration indicated water colour as the best predictor variable in unaffected lakes, but nitrogen and phosphorus as the best predictors in thaw-affected lakes. This result, in particular, suggested a significant shift in the processes governing productivity in thaw-affected lakes. In a smaller subset of lakes, chlorophyll a concentrations were lower in lakes affected by actively degrading permafrost than in lakes affected by stabilized thaw scars or in unaffected lakes. In contrast, zooplankton abundance was lowest in lakes with stabilized thaw scars. Bacterioplankton abundance was not different across the gradient of permafrost thaw. The differences in phytoplankton and zooplankton abundance between active and stable thaw scar lakes did not display a gradient of response that mirrored the degree of permafrost thaw activity. Because sampling did not include higher trophic levels, including macroinvertebrates and fish, it was difficult to determine how these biomass patterns arose. However, detectable differences existed in the morphometry of the lakes that might have affected habitat conditions for several species. Deep, near-shore lake-bottom pits occurred in lakes affected by permafrost thaw, while unaffected lakes were generally deepest at their centre. The pits increased the effective depth of the thaw-affected lakes, and allowed for thermal stratification where it might not have otherwise occurred. Future research should explore the habitat-related impacts of permafrost thaw on adjacent and higher trophic levels in order to better understand the fundamental shifts in trophic structures that appear in lakes affected by permafrost thaw.

Diss. (sammanfattning) Umeå : Sveriges lantbruksuniv., 2006. / Härtill 4 uppsatser.