Over the past two decades, Lake Winnipeg has been experiencing increasingly rapid eutrophication, and large cyanobacterial blooms now form in the North Basin in most years in late summer or fall. Cyanobacteria are considered a relatively poor food source compared with other phytoplankton, but the impacts of these blooms upon the primary consumers in the lake have not previously been researched. A microscopic analysis of whole water samples found cyanobacteria to be scarcely present in summer 2012, with nitrogen-fixing and non-fixing cyanobacteria comprising 11.2% and 8.4% of the basin-wide biovolume, respectively, and all but absent in fall. Gut content analysis of chironomids found that cyanobacteria made up an almost negligible part of their diet. Stable isotope analysis revealed that nitrogen-fixing cyanobacteria reduced phytoplankton δ15N values, and that this same reduction could be traced through the zooplankton, but not down to the sediments or chironomids.

In a climate change perspective, increased air temperatures are already a reality and are expected to increase even more in the future, especially in areas at high latitudes. The present thesis therefore addresses the influence of climate change on the physical properties and the phytoplankton communities of typical small and oligotrophic lakes in northern Sweden (62-64˚N). In the first part of the study, we found a significant trend (10 lakes from 1916 to 2010) of ice break-ups occurring increasingly earlier. The timing of ice break-up was strongly influenced by the April air temperature indicating that expected increases in air temperature in the future will also result in an earlier ice break-up. We also used concentrations of chlorophyll a (chl a) as estimations of phytoplankton biomass and discovered a positive relationship between surface water temperature and concentrations of chl a in Lake Remmaren (from 1991 to 2008). The second part of the thesis focuses on climatic conditions and cyanobacteria abundance in three small, oligotrophic lakes in northern Sweden; Lake Remmaren, Lake S. Bergsjön and Lake Gransjön. The concentration and relative abundance of cyanobacteria differ between 2011 and 2012, with different climatic conditions. The "warm" year of 2011 had higher concentrations and relative abundance of cyanobacteria than the "cold" year of 2012. Trends in increasing surface water temperatures as well as increasing abundance of cyanobacteria in August were found in Lake Remmaren (from 1988 to 2011). The direct or indirect effects of warming had a positive effect on the cyanobacteria abundance, since nutrients (Tot N and Tot P) did not display an increasing trend in Lake Remmaren. An analysis on the composition of phytoplankton species in Lake Remmaren, Lake S. Bergsjön and Lake Gransjön revealed that the cyanobacteria Merismopedia sp. was more common in 2011 than 2012. If different cyanobacteria become more common in oligotrophic lakes in the future, the functioning of lake ecosystems may be impacted. Small zooplankton eats small phytoplankton and if smaller phytoplankton species, e.g. cyanobacteria, increase at the expense of other phytoplankton groups, an extra step in the food chain might be added. Less energy might be transferred to the upper levels because many cyanobacteria contain toxic compounds and are less edible than other phytoplankton groups. An increase of toxic containing cyanobacteria in lakes can also make lakes less attractive for recreational purposes in the future.

Productivity in large areas of the world's oceans is limited by low concentrations of dissolved iron in surface waters. Phytoplankton have adapted to persist in these environments by reducing their requirements for iron (Fe) in key metabolic pathways, in some cases by replacing Fe-containing catalysts with their iron-free functional equivalents. This thesis examines the requirements and biochemical roles for copper (Cu) and manganese (Mn) in Fe-limited centric marine diatoms. A major finding of my research is that diatoms have elevated requirements for Mn and Cu when grown in Fe-deficient seawater. Iron deficiency induces oxidative stress and increases the cellular concentrations of toxic oxygen radicals and damage products in Thalassiosira pseudonana. The increased Mn-requirement is used, in part, to activate Mn-containing isoforms of the antioxidant enzyme superoxide dismutase. Cultures co-limited by Fe and Mn exhibit high levels of oxidative stress and an inefficient detoxification pathway that further reduces cell growth. Diatoms isolated from the metal poor open ocean require more Cu to divide than related species from metal-rich coastal waters. This pattern is in stark contrast to all other known nutritive trace metals. One part of the diatom Cu requirement that is independent of provenance is for efficient Fe transport. The additional Cu requirement of oceanic species appears to be due to the constitutive expression of a Cu-containing electron transport protein, possibly plastocyanin. Coastal species, which have higher Fe-requirements for growth, retain the Fe-containing functional homologue cytochrome c6. By employing metals other than Fe within photosynthesis and antioxidant pathways, marine diatoms are able to increase their fitness in Fe-deficient environments. However, Mn and Cu also occur in low concentrations in the open ocean and thus may co-limit growth of natural populations of phytoplankton. Metal enrichment experiments i

Some species of marine phytoplankton are believed to be more tolerant of high concentrations of trace metals than others, but no conclusive test of this hypothesis has been conducted. Eleven species of phytoplankton representing 5 classes were grown in Aquil medium containing Cd$ sp{2+}$ concentrations between 10$ sp{-9.85}$ and 10$ sp{-6.84}$ M. Growth rates and intracellular concentrations of Cd, C, N and S were measured. Cadmium quotas (mol Cd/litre-cell volume) were lower in members of Bacillariophyceae than in Chlorophyceae, Prymnesiophyceae, Dinophyceae and Cyanophyceae (ANOVA, p $<$ 0.001). Cellular C:S molar ratios decreased in phytoplankton grown at high (pCd 7.37-6.84) compared to low Cd (no added Cd), as S/litre-cell volume increased. Similar results were observed for C:N molar ratios. In two species that were examined, C:S ratios decreased as a linear function of increasing Cd concentration. Mean Cd$ sp{2+}$ concentration that reduced growth rate to 50% of maximum (pCd$ sp{50})$ was not significantly different among phytoplankton classes (ANOVA, p $<$ 0.05). When these experimental data were combined with pCd$ sp{50}$s calculated from published sources, Chlorophyceae were found to be the most resistant class (ANOVA, p $<$ 0.01). Cadmium and Cu resistance (pCd$ sp{50}$ and pCu$ sp{50})$ were correlated (r = 0.52, p $<$ 0.05), suggesting co-tolerance of phytoplankton to toxic levels of these metals. Chlorophyceae were most tolerant and Cyanophyceae the least tolerant of Cu (ANOVA, p $<$ 0.01). No significant differences were observed among Bacillariophyceae, Prymnesiophyceae, and Dinophyceae, which were of intermediate sensitivity to both metals. The results confirm the existence of a phylogenetic dependence of resistance to trace metal toxicity in phytoplankton.

Over the past two decades, Lake Winnipeg has been experiencing increasingly rapid eutrophication, and large cyanobacterial blooms now form in the North Basin in most years in late summer or fall. Cyanobacteria are considered a relatively poor food source compared with other phytoplankton, but the impacts of these blooms upon the primary consumers in the lake have not previously been researched. A microscopic analysis of whole water samples found cyanobacteria to be scarcely present in summer 2012, with nitrogen-fixing and non-fixing cyanobacteria comprising 11.2% and 8.4% of the basin-wide biovolume, respectively, and all but absent in fall. Gut content analysis of chironomids found that cyanobacteria made up an almost negligible part of their diet. Stable isotope analysis revealed that nitrogen-fixing cyanobacteria reduced phytoplankton δ15N values, and that this same reduction could be traced through the zooplankton, but not down to the sediments or chironomids.

The structure of phytoplankton communities was examined seasonally across five rivers with a focus on small cells and their relative importance. Picophytoplankton (0.2-2 μm), previously considered insignificant in rivers, reached densities as high as those observed in lakes and oceans (~ 10e4-10e5 cells/mL). Their relative importance was not a function of trophic state with the highest contribution to algal biomass found in the most eutrophic river. Body size distributions were analyzed from both chlorophyll-a size fractions and taxonomic enumerations; no significant effect of river or season was detected, suggesting that phytoplankton size distribution is not a useful metric of change in rivers. Unlike lake ecosystems, the rivers were uniformly dominated by small cells (< 20 μm). Taxonomic analyses of the seasonal succession did not reveal a common periodicity of particular divisions (e.g. diatoms). However, strong dominance was more typical of eutrophic rivers even though taxa richness was similar.

Three species of photosynthetic flagellates capable of phagotrophy (mixotrophic species) were tested for their abilities to use inorganic iron colloids for growth. Ochromonas sp., Chrysochromulina ericina (a coastal strain) and C. ericina (an oceanic strain) were grown in iron-free seawater supplemented with 1 muM goethite, hematite, magnetite/maghemite or ferrihydrite (90°) in the presence and absence of desferrioxamme B, an iron-binding siderophore. Both strains of Chrysochromulina grew at 35--70% of their maximum rates with goethite, hematite, and magnetite/maghemite, but were unable to use ferrihydrite. Ochromonas, however, grew well with ferrihydrite, but could not use any of the other forms. All the flagellates were able to acquire iron from ingested bacteria. Diatoms that were known only to take up dissolved forms of iron, Thalassiosira oceanica (clone 1003) and T. pseudonana (clone 3H), were unable to use any of the colloids tested. The mechanism of iron acquisition by the flagellates appeared to involve ingestion of the iron colloids as DFB had no effect on colloidal iron availability and bacteria resident in the cultures were unable to use the iron contained in the colloids. Variations in the size of the colloids were hypothesized to account for differences in their availability, independent of colloid chemical stability. The results provide the first strong evidence for direct utilization (i.e. without prior dissolution) of colloidal iron by mixotrophic phytoplankton and document a new pathway of iron acquisition that may be important for their survival in low-iron waters of the sea.

Variable fluorescence of chlorophyll a (Fv/Fm), measured by pulse amplitude modulated (PAM) fluorometers, is an attractive target for phytoplankton-related water quality management. Lowered Fv/Fm is believed to reflect the magnitude of nutrient sufficiency or deficiency in phytoplankton. This rapid and specific metric is relevant to Lake Erie, which often experiences problematic Cyanobacteria blooms. It is unknown whether PAMs reliably measure phytoplankton nutrient status or if different PAMs provide comparable results. Water samples collected from Lake Erie and two Lake Ontario sites in July and September 2011 were analysed using alkaline phosphatase assay (APA), P-debt, and N-debt to quantify phytoplankton nutrient status and with three different PAM models (PhytoPAM, WaterPAM and DivingPAM) to determine Fv/Fm. The Lake Ontario, Lake Erie East and Central Basin sites were all N- and P-deficient in July, but only the East and Central Basin and one Lake Ontario site were P-deficient in September. The West Basin sites were P-deficient in July and one West Basin site and a river site were N-deficient in September. Between-instrument Fv/Fm comparisons did not show the expected 1:1 relationship. Fv/Fm from the PhytoPAM and WaterPAM were well-correlated with each other but not with nutrient deficiency. DivingPAM Fv/Fm did not correlate with the other PAM models, but correlated with P-deficiency. Spectral PAM fluorometers (PhytoPAM) can potentially resolve Fv/Fm down to phytoplankton group by additionally measuring accessory pigment fluorescence. The nutrient-induced fluorescent transient (NIFT) is the observation that Fv/Fm drops immediately and recovers when the limiting nutrient is reintroduced to nutrient-starved phytoplankton. A controlled laboratory experiment was conducted on a 2x2 factorial mixture design of P-deficient and P-sufficient Asterionella formosa and Microcystis aeruginosa cultures. Patterns consistent with published reports of NIFT were observed for P-deficient M. aeruginosa in mixtures; the pattern for A. formosa was less clear. This thesis showed that Fv/Fm by itself was not a reliable metric of N or P deficiency and care must be taken when interpreting results obtained by different PAM fluorometers. NIFT analysis using spectral PAM fluorometers may be able to discriminate P-deficiency in M. aeruginosa, and possibly other Cyanobacteria, in mixed communities.

Degradation of water quality, introduction of dreissenid mussels (notably <i>Dreissena polymorpha</i>) and depletion of oxygen concentrations in the hypolimnion in Lake Simcoe, Ontario prompted a study of phytoplankton primary production to inform efforts to improve the lake conditions. The characterization of algal production is critical since, as primary producers, their biomass is positively correlated with production at higher trophic levels in pelagic food webs and oxygen levels. This study was conducted from August 2010 to August 2011, including the winter season (Dec-Mar). Temporally, the lake displayed a unimodal pattern with late summer to fall production maxima. For all seasons considered, the pelagic daily areal primary production (P_int) was lower in the nearshore than offshore, consistent with the nearshore shunt hypothesis that mussels should be able to deplete phytoplankton more effectively in the nearshore. The sensitivity analysis revealed that chl a and the photosynthetic parameter P^B_max were the most influential variables for explaining such spatial differences. The size distribution of chl a and production varied where both netplankton (>20µm) and nanoplankton (2-20µm) were greatest in fall and picoplankton (<2µm) was highest in summer and early fall. A large chl a peak of nanoplankton was also found in late-winter (Mar) at offshore stations. The seasonal areal primary production (SAPP; May-Oct) and chl a:TP were significantly lower nearshore than offshore, consistent with grazing impacts from the large nearshore dreissenid mussel community. The lake as a whole is quite productive comparable to other large lakes with comparable total P concentrations and dreissenid mussel populations. The latter part of the study showed that the deep chlorophyll layer (DCL) was not as frequent as expected and was detected only 28% of time during late-spring to summer when the lake was thermally stratified (Aug-Sept 2010 and May-Aug 2011). The percent dissolved oxygen (%) did not show any indications of elevated primary production in the DCL although the production estimates suggested that there is a substantial (an average of 55%) amount of primary production occurring below thermocline when a DCL exists. Whether or not the DCL has potential to nourish the benthic filterers (dreissenids) and has ecological significance in the lake remains unclear. Overall, the factors that control phytoplankton primary production in Lake Simcoe seem to operate somewhat differently from other large lakes and further investigation is needed to elucidate them. The analysis of primary production and biomass has improved knowledge of non-summer production and can provide guidance to site-specific P and oxygen remediation.

Temporal changes in the environmental conditions upon which life depend are ubiquitous in nature, acting at every level of organisation from cells to ecosystems. Although the actions themselves are often poorly understood, they strongly depend on the temporal and hierarchical (organisational) scales at which they are measured; ecosystems are relatively stable through time while their species composition may undergo vast changes. Likewise any hierarchical level may be relatively stable in the short-term, but undergo vast long-term changes. This thesis aims to better understand the importance of these scales for mediating the impact of environmental variability on ecological systems. / The approach used herein employs both mathematical models and empirical data which represent food webs responding to environmental variability at different hierarchical scales. Within each of these representative food webs, the influence of environmental variability on the stability of the food web is determined using an approach which accounts for the effects of temporal scale. This thesis demonstrates that the stability of simple model food webs (high hierarchical scale) is tightly linked to environmental variability and the temporal scales at which these changes occur dictate which species in the model are most affected. At lower scales of organisation, empirical data indicate that environmental variability generally has a lesser impact on stability and that only certain temporal scales are responsible for this trend. At these temporal scales some species respond differently to environmental variability, negative changes in one species (or group) are offset by positive changes in another - a process known as compensation. These results highlight the importance of both temporal and hierarchical scale in mediating the response of food webs to environmental variability. Ultimately, they will serve to better understand how models and experiments should scale-up from low to high hierarchical and temporal scales.