This study combines Optical Plankton Counter data on zooplankton abundance and Acoustic Doppler Current Profiler data on water currents to test hypotheses about spatial zooplankton distributions near a major point of land affecting local wind and water current patterns. Data were collected by repeatedly sampling a 2 km linear transect in the South Arm of Lake Opeongo, Ontario, Canada during July, 2008. Moving Split Window techniques were used to identify breakpoints in both zooplankton biomass concentration and a measure of water turbulence, and confirmed an effect of the topographic feature on local zooplankton distributions. Using additional data collected in 2001 and 2003, zooplankton abundance distributions were also tested under three wind speed ranges predicted from physical principles to correspond to varying stability of the water column. Significant differences were found in the variability and patch sizes of the distributions, indicating the importance of wind speeds on zooplankton patchiness.

This study combines Optical Plankton Counter data on zooplankton abundance and Acoustic Doppler Current Profiler data on water currents to test hypotheses about spatial zooplankton distributions near a major point of land affecting local wind and water current patterns. Data were collected by repeatedly sampling a 2 km linear transect in the South Arm of Lake Opeongo, Ontario, Canada during July, 2008. Moving Split Window techniques were used to identify breakpoints in both zooplankton biomass concentration and a measure of water turbulence, and confirmed an effect of the topographic feature on local zooplankton distributions. Using additional data collected in 2001 and 2003, zooplankton abundance distributions were also tested under three wind speed ranges predicted from physical principles to correspond to varying stability of the water column. Significant differences were found in the variability and patch sizes of the distributions, indicating the importance of wind speeds on zooplankton patchiness.

Nous avons étudié l’application de plusieurs mesures d’isotopes stables afin de caractériser les processus du cycle de l’azote et les sources d’azote dans les lacs tempérés à diverses échelles spatiales et temporelles. Les résultats d’une étude à travers 65 lacs sur un gradient trophique ont démontré que le ratio d’isotopes stables d’azote (δ15N) des sédiments de surface est un indicateur de l’importance relative des sources d’azote anthropique, mais que ce ratio peut être altéré par la diagenèse. La mesure du δ15N des sédiments demeure néanmoins un outil permettant de déterminer à long terme le changement des charges en azote anthropique aux écosystèmes lacustres et les causes de l’eutrophisation de ces systèmes. Nos résultats d’une étude sur la variation saisonnière de plusieurs isotopes stables dans trois lacs peu profonds situés sur un gradient trophique et ayant différents régimes de stratification ont démontré que cette approche est prometteuse dans les lacs mésotrophes et stratifiés. Dans ces systèmes, le δ15N de la matière organique particulaire (MOP) aurait le potentiel de déterminer les sources d’azote assimilées par le phytoplancton. Cependant les mesures d’isotopes stables du carbone (δ13C) et du ratio C:N indiquent que les apports de matières organiques du bassin versant peuvent altérer les relations observées. Nous avons également constaté une déviation de la relation 1:1 entre les isotopes stables d’azote et d’oxygène (δ18O) du nitrate (NO3-) indiquant son assimilation et sa nitrification simultanée. Cette application est particulièrement prometteuse puisque la nitrification est méconnue dans les lacs et peut exacerber les effets de l’eutrophisation. / We studied the application of multiple stable isotopes measures with the overall objective of improving our understanding of nitrogen cycling and sources in temperate lakes over different spatial and temporal scales. Results from our study across 65 lakes on a trophic gradient demonstrated that surface sediment nitrogen stable isotope ratio (δ15N) is an indicator of the relative importance of anthropogenic N loads, but that diagenesis can alter this ratio. Nevertheless, sediment core δ15N is a powerful proxy for the determination of longterm changing anthropogenic N loads to lake ecosystems and the causes of lake eutrophication. Results from our second study on multiple stable isotopes seasonal variation in three shallow lakes along a trophic gradient and with different stratification regimes have demonstrated that such an approach is particularly promising in mesotrophic and stratified lakes. In these systems, our results showed that the suspended particulate organic matter (SPOM) δ15N could be used to assess the nitrogen sources assimilated by phytoplankton. However, measurement of carbon stable isotopes (δ13C) and C:N ratios from the SPOM showed that the observed relationships can be altered by watershed derived organic matter. We have also found a deviation from the 1:1 relationship between nitrogen and oxygen stable isotopes (δ18O) of nitrate (NO3-) indicating a simultaneous NO3- assimilation and nitrification in our stratified system. The application of dual nitrate isotopes is promising in the study of nitrification since this process is not well understood in lakes and can increase the severity of eutrophication symptoms.

We deployed depth-specific drifters in the western and eastern parts of the South Arm basin of Lake Opeongo and collected zooplankton samples at west and east fixed stations and at additional up- and downwind locations at three depths of the epilimnion under a range of wind conditions. Water currents had highest association with the immediate wind direction and the direction they travelled was dependent on wind strength. Along the main west-east fetch large zooplankton resided high in the epilimnion and were transported eastwards by strong surface currents where they accumulated. Small zooplankton were more uniformly distributed with depth and their accumulation patterns and transport mechanisms are less clear. Along shorter fetches oriented off-angle with the main one accumulations of zooplankton occurred at all downwind locations under heavy winds although the patterns are more variable and complex. These downwind accumulations likely create high quality habitat for warm water fish.

Master of Science / Department of Biology / Craig Paukert / I investigated the influence of anthropogenic threats and hydrologic alteration on fish assemblages within the Lower Colorado River Basin (LCRB). Life history traits of fish assemblages for individual stream segments were summarized by species presence/absence data of current (1980-2006) records. To assess anthropogenic threats, I developed a series of ecological risk indices at various scales (e.g., catchment, watershed, aquatic ecological system and upstream of aquatic ecological system) and related each index to fish life-history traits to determine the method and scale that best related to biotic metrics. Hydrologic alteration was quantified using the Indicators of Hydrologic Alteration software to calculate hydrologic alteration values using the range of variability approach. Ecological risk indices within all scales were strongly correlated (r[superscript]2>0.54, p<0.0001) to one another. Relationships between fish life history traits and ecological risk indices occurred only at the catchment and watershed scales. Strongest relationships were at the watershed scale where increased levels of anthropogenic risk were related to reduced occurrences of native, fluvial dependent species (r[superscript]2=0.12, p<0.0001) and increased occurrences of nonnative generalist species (r[superscript]2=0.22, p<0.0001). The percent agriculture was positively related to indices of alteration of low flows (r=0.401, p=0.006) while forested land cover was negatively related to alteration of low flow events (r=-0.384, p=0.008). Relationships between indices of hydrologic alteration and fish traits indicate the occurrence of piscivorous, nonnative fishes increased with alteration of low flow events whereas occurrence of fluvial dependent fishes that preferred rubble substrate decreased with alteration of low flow events (r=0.64, p=0.001). Our analysis suggests that ecological risk indices and hydrologic alteration in the LCRB are related to composition of biotic communities. Incorporating cost-effective risk indices into conservation planning will likely increase the effectiveness of conservation efforts while understanding biotic responses to modified flow regimes are a necessity in sustainable development of water resources as human populations grow and water resources decrease in the LCRB.

Master of Science / Department of Biology / Walter K. Dodds / Human-induced stressors such as increased nitrogen (N) loadings, altered watershed land-use, and biodiversity losses are a few of the numerous threats to aquatic systems. Prairie streams experience natural disturbances, such as flooding and desiccation, which may alter responses to anthropogenic stressors. Denitrification, the dissimilatory reduction of NO3- to N gas (N2O or N2), is the only permanent form of N removal from terrestrial or aquatic ecosystems, and is important in mitigating N pollution to streams and downstream waters. Little is known about the relationships between denitrification and riparian prairie vegetation or large consumers. In the first chapter, I used outdoor mesocosms to determine the impact of a grazing minnow, Campostoma anomalum, on structural and functional responses of prairie streams to a simulated flood, focusing on denitrification. In terrestrial ecosystems, grazing can stimulate denitrification, but this has not been studied in streams. Ammonium (NH4+) enrichments, used to simulate fish excretion, alleviated N limitations on denitrification. Both fish and NH4+ affected algal biomass accrual, but only fish affected algal filament lengths and particulate organic matter. In a second experiment, I examined the impact of woody vegetation expansion, a primary threat to tallgrass prairie, on riparian and benthic denitrification. Expansion of woody vegetation in these grasslands is due primarily to altered fire regimes, which historically inhibited woody vegetation growth. To determine the effect of woody vegetation expansion on benthic and riparian denitrification, woody vegetation was removed from the riparian zone of a grazed and an ungrazed watershed. Both soil and benthic denitrification rates from this removal buffer were compared to rates in grassy or woody riparian zones. Riparian soil denitrification was highly seasonal, with greatest rates occurring during early spring, and rates being low throughout the remainder of the year. Benthic denitrification was also temporally variable but did not exhibit seasonal trends, suggesting benthic denitrification is driven by factors other than water temperature. Removal of woody vegetation stimulated soil and benthic denitrification rates over rates found in naturally vegetated riparian zones. Elevated N loadings will continue to affect aquatic ecosystems, and these effects may be exacerbated by biodiversity losses or changing riparian vegetation.

Nous avons cherché des relations empiriques entre l’abondance des macrophytes submergés et le développement résidentiel du bassin versant, les propriétés du lac et la présence de milieux humides dans 34 lacs de la région des Laurentides et de Lanaudière sélectionnés à travers un gradient de développement résidentiel. Les macrophytes submergés ont été échantillonnés par méthode d’échosondage à l’intérieur de la zone littorale. L’abondance moyenne des macrophytes a ensuite été estimée à l’intérieur de quatre zones de croissance optiquement définies (profondeur maximale = 75 %, 100 %, 125 % et 150 % de la profondeur de Secchi) ainsi qu’à l’intérieur de toute la zone littorale. L’occupation humaine a été considérée selon trois échelles spatiales : celle présente 1- dans un rayon de 100 mètres autour du lac, 2- dans la fraction du bassin versant qui draine directement vers le lac et 3- dans le bassin versant en entier. Nous avons aussi testé, lac par lac, l’effet de la pente locale sur l’abondance des macrophytes. Nous avons observé des corrélations positives et significatives entre l’abondance des macrophytes submergés et l’occupation humaine de l’aire de drainage direct (r > 0.51). Toutefois, il n’y a pas de relation entre l’abondance des macrophytes submergés et l’occupation humaine de la bande de 100 mètres entourant le lac et du bassin versant entier. Les analyses de régression multiple suggèrent que l’abondance des macrophytes submergés est faiblement corrélée avec l’aire du lac (+) et avec la présence de milieux humides dans le bassin versant entier (-). Localement, l’abondance des macrophytes est reliée à la pente et à la profondeur qui expliquent 21% de la variance. Les profondeurs de colonisation maximale et optimale des macrophytes submergés sont corrélées positivement au temps de résidence et à la profondeur de Secchi et négativement à l’occupation humaine et à l’importance des milieux humides. / We searched for empirical relationships between whole-lake submerged macrophyte abundance and residential development, lake properties and wetlands in 34 lakes from the Laurentides-Lanaudière region subjected to a wide range of residential development. Submerged macrophyte abundance was assessed by echosounding of the entire littoral zone. Human encroachment was measured at three spatial scales: Within 1- a 100 m-wide buffer strip circling the lake, 2- the fraction of the watershed that drains directly into the lake and 3- the entire watershed. Single and multiple regression were used to test the effect of residential development on macrophyte abundance in the entire littoral zone and within four optical growth zones defined by Secchi depth. The local effect of slope on local macrophyte abundance was also tested in individual lakes. We found significant positive correlations between submerged macrophyte abundance and dwelling density present within the direct drainage area (r > 0.51). In contrast, there was no significant relation between macrophyte abundance and human occupation of the 100 m-wide buffer circling lake and the entire watersheds. Multiple regression analysis suggest that submerged macrophyte abundance is weakly correlated with lake area (+) and wetland importance within the entire watershed (-). Locally, macrophyte abundance is related to slope and depth which explained together 21% of the variance. Maximum and optimum colonization depths of submerged macrophytes were positively correlated to residence time and Secchi depth and negatively to human occupation and to wetland importance.

Master of Science / Department of Biology / Keith B. Gido / Nonnative species represent a major threat to the continued persistence of native fishes globally, especially in the Colorado River Basin of western North America, where there are now more nonnative than native fishes. In the upper Gila River, a tributary of the Colorado, numerous nonnative fishes have established populations, and predation by these nonnatives has been linked to extirpation of native fishes under low-flow conditions at some locations. Historically, the upper Gila lacked a top piscivore, and it is unclear what mechanisms have allowed the establishment of nonnative piscivores and resultant extension in food chain length. To investigate the phenomenon of increased food chain length through nonnative introductions we explored the influence of autochthonous energy availability on nonnative predator abundance, food chain length, and abundance of other trophic levels. Predictions were that increased basal energy availability would lead to increased nonnative predator abundance and thus increased food chain length, based upon predictions from food web theory. Annual production and biomass of four trophic levels measured across six longitudinally-positioned sites were calculated between June 2008 and June 2009 to test these predictions. In addition, energy demand of trophic levels relative to energy supply was compared across sites using a quantitative food web approach, to evaluate energy limitation across trophic levels. Primary production was found to vary considerably across the upper Gila (1,677-16,276 kcal m-2 yr-1), but production and biomass of other trophic levels was not related to this gradient as predicted. In addition, food chain length demonstrated a marginally-significant negative relationship with primary production (R[superscript]2=0.42, d.f.=5, p=0.16), which was in contrast with predicted responses. These results suggest that energy availability does not appear to be a limiting factor to the production or biomass of consumers. The influence of other mechanisms on food chain length in the upper Gila River, in particular disturbance frequency and intensity, deserve further investigation.

This research aims to integrate mathematical water quality models with Bayesian inference techniques for obtaining effective model calibration and rigorous assessment of the uncertainty underlying model predictions. The first part of my work combines a Bayesian calibration framework with a complex biogeochemical model to reproduce oligo-, meso- and eutrophic lake conditions. The model accurately describes the observed patterns and also provides realistic estimates of predictive uncertainty for water quality variables. The Bayesian estimations are also used for appraising the exceedance frequency and confidence of compliance of different water quality criteria. The second part introduces a Bayesian hierarchical framework (BHF) for calibrating eutrophication models at multiple systems (or sites of the same system). The models calibrated under the BHF provided accurate system representations for all the scenarios examined. The BHF allows overcoming problems of insufficient local data by “borrowing strength” from well-studied sites. Both frameworks can facilitate environmental management decisions.

Les thiols et le sélénium peuvent jouer un rôle important dans la méthylation du mercure des environnements aquatiques. Pour démontrer la présence des thiols et du sélénium et leur relation avec le mercure dans certains écosystèmes d'eau douce québécois, une campagne d’échantillonnage fut réalisée durant l’été 2010, dans le parc national du Mont-Tremblant (Laurentides, Québec). Il existe une corrélation significative entre le sélénium et le mercure total dans l’eau des lacs du parc. Cependant, les concentrations de sélénium sont très faibles dans les lacs, les étangs de castor et les ruisseaux. Par ailleurs, les lacs du parc national du Mont-Tremblant ont des concentrations relativement élevées de méthylmercure avec une moyenne de 0,33 ng L-1 et des maximums allant jusqu’à 3,29 ng L-1. Les étangs de castor peuvent aussi être considérés comme des lieux de contamination au méthylmercure, avec une concentration moyenne de 0,95 ng L-1. Toutefois, la présence d’une colonie de castors sur le bassin versant d’un lac ne semble pas influencer les concentrations de mercure que l’on y retrouve. Deux thiols sont détectables dans l’eau de surface des Laurentides, soit le glutathion et l’acide thioglycolique. La concentration de ce dernier thiol est corrélée significativement avec celle du mercure total et du méthylmercure. Les thiols peuvent jouer un rôle important dans les processus de méthylation en favorisant le transport du mercure inorganique à l’intérieur des bactéries sulfato-réductrices. Afin de mieux comprendre l’action antagoniste entre le sélénium et le mercure, des études devraient être réalisées au niveau des tissus des organismes vivants dans ces zones pauvres en sélénium. / Thiols and selenium may influence mercury methylation in aquatic environments. To demonstrate the occurrence of both elements with mercury in selected aquatic systems of Quebec, a sampling campaign was conducted during the summer of 2010 in Mont-Temblant national park (Laurentians, Quebec). There is a significant correlation between selenium and total mercury in surface water. However, concentrations of selenium are very low in lakes and beavers ponds. In addition, the park’s lakes have a relatively high concentrations of methyl mercury with an average of 0,33 ng L-1 and maximum values up to 3,29 ng L-1. Beaver ponds can also be considered hotspots of methylmercury contamination, with an average concentration of 0,95 ng L-1. However, the presence of a colony of beavers in the watershed of a lake does not appear to influence the levels of mercury in that lake. Two species of thiols are present in the surface water of the Laurentian ecosystems: glutathione and thioglycolic acid. The latter is significantly correlated with total mercury and methyl mercury. Thiols may play an important role in the methylation process by promoting the transport of inorganic mercury within the sulfate-reducing bacteria. To better undestand the antagonistic action between selenium and mercury, more research should be conducted in the tissue of organisms living in areas of low concentration of selenium to know if they are protected against mercury toxicity.