Uranium mining activities have the potential to impact aquatic systems through mine drainage (runoff) and the release of treated effluent into nearby watersheds. Such anthropogenic exposure can lead to elevated concentrations of metals and major ions, which may impact aquatic biota. Previous studies have looked at the effects of water quality on aquatic biota within flooded pit lakes and natural lakes that have been exposed to various mechanisms of mining exposure. However, the literature often only examines the effects of a limited number of contaminants on a limited number of species. Researchers have rarely looked at the effects of multiple contaminants on species composition, biodiversity and ecosystem function in aquatic systems. This study uses a multivariate approach to look for relationships between water quality (24 variables), plankton species composition and abundance, biodiversity (richness and evenness) and ecosystem function among lakes exposed to mining activities (n = 18) and non-exposed reference lakes (n = 8). Lake water quality data was used to cluster lakes into groups. Lake groups were then overlain onto multivariate ordinations derived from species composition-abundance data to determine if species composition was related to water quality. Ecosystem function variables included planktonic phosphorus cycling and planktonic respiration. The classified lake groups clustered well on ordinations derived from species composition-abundance data suggesting that relationships exist between water quality and plankton species composition. However, ecosystem function was similar among the majority of lakes and flooded pits despite differences in species richness, species composition and species abundance. Only a small number of aquatic systems had ecosystem function properties that were different from the majority of lakes and pits. These systems had the greatest concentrations of contaminants and had very low biodiversity (richness and evenness) compared to the other systems. Despite having differences in plankton species composition and species richness, all lake groups were functionally similar. This suggests that functional redundancy in species composition may be present in the majority of lakes and pits in such a way that ecosystem function is maintained.

Community-based ecosystem management (CBEM) is increasingly advocated as a way to conserve biodiversity, monitor, and maintain ecosystem functions in the context of local land use practices through an inclusive management approach. However, while CBEM is based in principles of inclusion, there is very little attention in environmental management and education literature directed to the role of youth in stewardship activities, and the environmental learning outcomes and other meanings that may result from these practices.<p>The purpose of this thesis is to describe participatory and experiential environmental learning carried out in the Frenchman River Basin, Southwestern Saskatchewan. Here, I investigated how students participation in an ecological monitoring program contributed to their understanding of their local environment and to their sense of place, and considered how the development of a learning community among students, teachers, community members, and academic researchers influenced these processes.<p>This research adopts a mixed methods approach, employing knowledge-based tests to explore student learning outcomes and using interpretations of place through student photographs and interviews to examine their sense of place. I take a phenomenological approach to defining what constitutes place for students, as well as how sense of place is formed for them, elucidating how their experiences participating in the ecological monitoring program entered the process of meaning construction.<p>This case study found that both experiential and participatory approaches to learning helped foster environmental understanding as well as place appreciation and attachment. The Frenchman River, previously described as a taken-for-granted feature of the familiar landscape and largely associated with its agricultural importance, was re-negotiated as a social space, a place of play, learning, and biological significance. Research findings also suggest that place meanings are deeply rooted in students rural identity, and that this influenced their participant experience, independent of environmental learning outcomes. <p>The creation of a learning community was a mobilizing force for school-based ecological monitoring and information sharing, while acting as a source of symbolic significance for student participants, helping students to see their place from the perspective of an outsider.

The boreal forest is one of the worlds largest forest biomes and comprises a major portion of the terrestrial carbon (C) sink. Quantifying the net C change in forest ecosystems is an important step in understanding and modeling the global C cycle. The goals of this project were: to estimate and compare the total change in ecosystem C over a 10-year period in two boreal forest stands using biometric and eddy-covariance approaches, and to evaluate the year-to-year changes in C uptake. This study utilized 10 years of eddy-covariance data and ecosys model data from the Old Aspen (OA) and Old Jack Pine (OJP) sites in central Saskatchewan, part of the Boreal Ecosystem Research and Monitoring Sites (BERMS). According to the eddy-covariance and C stock approaches, between 1994 and 2004 the net change in C storage at OA was 15.6 ± 4.0 and 18.2 ± 8.0 Mg C ha-1, respectively. At OJP, the 10-year net change in C storage from eddy-covariance was 5.8 ± 2.0 Mg C ha-1 in comparison to 6.9 ± 1.6 Mg C ha-1 from the carbon stock approach. While both sites were sinks of C between 1994 and 2004, the greatest increase in C occurred in different components - the forest floor at OA (14.6 Mg C ha-1) and in the living vegetation at OJP (8.0 Mg C ha-1). In 2004, total ecosystem C content was greater at OA (180.6 Mg C ha-1) than OJP (78.9 Mg C ha-1), with 50% (OA) and 39% (OJP) of the C in the detritus and mineral soil pools. During the 10-year period of eddy-covariance measurements, there was a positive correlation between both annual and growing season gross ecosystem photosynthesis (GEP) and live stem C biomass increment at OA, whereas no significant relationships were found at OJP. Stem C increment accounted for 30% of total net primary productivity (NPP) at both sites, and NPP/GEP ratios were 0.36 and 0.32 at OA and OJP, respectively. Overall, this study found good agreement between eddy-covariance and biometric estimates of ecosystem C change at OA and OJP between 1994 and 2004. Over that period at OA, eddy-covariance estimates of photosynthesis captured the inter-annual variability in C uptake based on the growth of tree rings.

In aquatic ecosystems, dissolved organic matter (DOM) is an important source of detrital energy on which microorganisms rely. However, its dynamics are not well understood in an ecological context. By isolating watershed sources, the work reported in this thesis has attempted to characterize the seasonal patterns of DOM in the hyporheic zone of a temperate stream and to find the likely sources that contribute to this pool of organic carbon. Hyporheic DOM characteristics described by UV spectroscopy indicated temporal rather than spatial dependence. Excitation-emission matrices (EEMs) showed that hyporheic DOM was mainly comprised of fulvic- and humic-like fluorescence with small amounts of protein-like fluorescence. Increases in dissolved organic carbon (DOC) concentrations from birch litter isolates were greater than those from cedar litter in early autumn, but less in late autumn. Although streambed biofilm was not significant in increasing DOC concentrations, it was also a source of protein-like fluorescence.

The projected increase of shrubs across the Arctic is expected to alter patterns of snow cover, which may affect the phenology and survival of arthropods such as spiders. In this study, we simulated shrub encroachment on a series of tundra plots and examined the effects on the spider assemblages during the following growing season. Our simulated shrub treatment did not affect the abundance or composition of spider communities over the season; however, adults from the dominant genus Pardosa (Lycosidae) had significantly higher body mass on treatment plots. This difference in mass was observed following snow melt and persisted until halfway through the growing season. Given the importance of spiders as arthropod predators and as food sources for breeding birds, such a change in summer body mass could represent a shift in spiders’ functional contributions to Arctic ecosystems.

The Convention on Biological Diversity (CBD) promotes the Ecosystem Approach (EA). In this thesis it is analyzed how the EA approach could be applied to regional management of coastal areas in Sweden. The aim of this report is to compare management for preserving biodiversity on a landscape level in two regional coastal areas in Uppsala and Västra Götaland respectively. In Uppsala, one of the plans consists of a previous nature reserve and in Västra Götaland, a Regional landscape strategy has been established. The ecosystem approach is used as a theoretical framework in this thesis. More precisely, five EA operational guidelines have been used as analytical tools in the comparative approach adopted. The results show that the Västra Götaland Regional landscape strategy has been most consistent with the EA framework, since it fulfills three of the five operational guidelines while Gårdsskärskusten only accomplish two of the criteria. The social aspects are more central in the landscape strategy than in Gårdskärskusten, which is the main difference between the two plans. One area of inadequacy detected in the landscape strategy was the few identified aspects on maintenance of landscape connectivity features. The Regional landscape strategy could, if established in every county in Sweden, be a way to prevent the loss of biodiversity. However the plan needs to be supported by actions where preservation and utilization are integrated in an economic context. By considering nature protection as a development opportunity in strategic planning rather than as only a cost, important steps towards a sustainable future can be taken. / För att implementera målsättningarna i FN:s Konvention om Biologisk Mångfald har en ekosystemansats (EA) antagits och en viktig ekosystemtyp i Sverige där EA kan appliceras är kuster. Syftet med examensarbetet är att, med ekosystemansatsen som analysverktyg, studera hur och varför två planer för att bevara biologisk mångfald på landskapsnivå i två regionala kustområden i Uppsala län respektive Västra Götalands län skiljer sig åt. Planen består i Uppsala län av ett tilltänkt naturreservat och i Västra Götalands län av en Regional landskapsstrategi. Uppsatsens teoretiska förankring sker i ekosystemansatsen, med fokus på dess fem vägledande punkter och metodvalet är en jämförande design. Inom ramen för den jämförande designen har sedan en innehållsanalys samt kompletterande semistrukturerade intervjuer utförts. Resultatet visar att den Regionala landskapsstrategin i dagsläget är mer i linje med ekosystemansatsen, eftersom den uppfyller tre av fem vägledande punkter medan processen med Gårdsskärskusten enbart uppfyller två av fem vägledande punkter. I landskapsstrategin är sociala aspekter centrala och det är en förklaring till varför de olika planerna skiljer sig åt. Intressant är dock att betoningen på aktörsinvolvering verkar leda till att vissa ekologiska grundprinciper som konnektivitet i landskapet tonas ned, vilka är framträdande i processen med Gårdskärskusten. Regionala landskapsstrategier kan enligt min mening innebära ett stort steg mot att hindra förlusten av biologisk mångfald i Sverige om de uppförs i varje län. Inom landskapsstrategierna är det vidare essentiellt att skyddade områden, som exempelvis Gårdsskärskusten, också inkorporerar ett landskapsstrategitänk och i många fall innebär det att skötselåtgärder fortsätter att utföras. Svårigheten kan dock vara hur finansieringen av skötseln skall ske men där finns alternativa inkomstmöjligheter, från exempelvis naturturism och stresshantering på företag, som behöver utvecklas ytterligare.

The projected increase of shrubs across the Arctic is expected to alter patterns of snow cover, which may affect the phenology and survival of arthropods such as spiders. In this study, we simulated shrub encroachment on a series of tundra plots and examined the effects on the spider assemblages during the following growing season. Our simulated shrub treatment did not affect the abundance or composition of spider communities over the season; however, adults from the dominant genus Pardosa (Lycosidae) had significantly higher body mass on treatment plots. This difference in mass was observed following snow melt and persisted until halfway through the growing season. Given the importance of spiders as arthropod predators and as food sources for breeding birds, such a change in summer body mass could represent a shift in spiders’ functional contributions to Arctic ecosystems.

A three dimensional time-dependent coupled ecosystem model is applied to Izmir Bay for the first time. Delft3D modelling suite&rsquo / s FLOW and ECO modules are adapted and tuned for the region. A reference model with a time frame of three years is produced that represents the current physical and biogeochemical status of the bay. Model skill assessment methods are used as a measure of model performance and to address the shortcomings of it. The hydrodynamics model is able to produce physical features in terms of seasonality and spatial distribution within reasonable ranges, whereas the ecosystem model has certain discrepancies which can be reduced with improved quality of model inputs, such as open boundary conditions, and fresh water and nutrient fluxes. The reference model is used as a tool with predictive capacity to assess the ecosystem response of the bay to possible changes it may undergo in the future. Five nutrient enrichment/reduction scenarios are constructed to predict the reactions of the bay to changing external inputs of DIN and PO4. Results suggest that both physical and biogeochemical properties of the bay show strong horizontal gradients between outer and inner regions in which both natural and anthropogenic influences are effective. It is revealed that Outer bays are mostly occupied by waters originating from the oligotrophic Aegean Sea, while eutrophicated inner regions are mainly controlled by local influences such as increased fresh water inputs and excessive wastewater discharges. Results of the nutrient enrichment/reduction scenarios suggest that the N-limited Inner and Middle bays and the P-limited Outer bays, give contrasting reactions to changes in inputs of DIN and PO4 such that the former is more sensitive to DIN input whereas the latter is more sensitive to PO4 input. Due to the existence of these two contrasting environments in the bay, availability of one nutrient is dependent on the availability of the other, therefore treatment of both should be considered in parallel. Among the scenarios tested in this study, the best possible option to reduce eutrophication in Izmir Bay is to prevent the increase of PO4 input and to reduce the DIN input simultaneously. These outcomes are aimed to provide a scientific insight for coastal policy makers and environmental managers on how changes in anthropogenic influences can impact the marine ecosystem of the bay.

In aquatic ecosystems, dissolved organic matter (DOM) is an important source of detrital energy on which microorganisms rely. However, its dynamics are not well understood in an ecological context. By isolating watershed sources, the work reported in this thesis has attempted to characterize the seasonal patterns of DOM in the hyporheic zone of a temperate stream and to find the likely sources that contribute to this pool of organic carbon. Hyporheic DOM characteristics described by UV spectroscopy indicated temporal rather than spatial dependence. Excitation-emission matrices (EEMs) showed that hyporheic DOM was mainly comprised of fulvic- and humic-like fluorescence with small amounts of protein-like fluorescence. Increases in dissolved organic carbon (DOC) concentrations from birch litter isolates were greater than those from cedar litter in early autumn, but less in late autumn. Although streambed biofilm was not significant in increasing DOC concentrations, it was also a source of protein-like fluorescence.

This thesis presents a rough set approach to reinforcement learning. This is made possible by considering behaviour patterns of learning agents in the context of approximation spaces. Rough set theory introduced by Zdzisław Pawlak in the early 1980s provides a ground for deriving pattern-based rewards within approximation spaces. Learning can be considered episodic. The framework provided by an approximation space makes it possible to derive pattern-based reference rewards at the end of each episode. Reference rewards provide a standard for reinforcement comparison as well as the actor-critic method of reinforcement learning. In addition, approximation spaces provide a basis for deriving episodic weights that provide a basis for a new form of off-policy Monte Carlo learning control method. A number of conventional and pattern-based reinforcement learning methods are investigated in this thesis. In addition, this thesis introduces two learning environments used to compare the algorithms. The first is a Monocular Vision System used to track a moving target. The second is an artificial ecosystem testbed that makes it possible to study swarm behaviour by collections of biologically-inspired bots. The simulated ecosystem has an ethological basis inspired by the work of Niko Tinbergen, who introduced in the 1960s methods of observing and explaining the behaviour of biological organisms that carry over into the study of the behaviour of interacting robotic devices that cooperate to survive and to carry out highly specialized tasks. Agent behaviour during each episode is recorded in a decision table called an ethogram, which records features such as states, proximate causes, responses (actions), action preferences, rewards and decisions (actions chosen and actions rejected). At all times an agent follows a policy that maps perceived states of the environment to actions. The goal of the learning algorithms is to find an optimal policy in a non-stationary environment. The results of the learning experiments with seven forms of reinforcement learning are given. The contribution of this thesis is a comprehensive introduction to a pattern-based evaluation of behaviour during reinforcement learning using approximation spaces. / May 2006