Övergödning är ett allvarligt miljöproblem, som bland annat orsakar omfattande blomningar av alger och blågrönbakterier. I söt- och brackvatten är fosfor ofta det mest begränsande näringsämnet för dessa organismer, varför en minskning av fosfortillförseln är nödvändig för att nå förbättringar. I Sverige beräknas jordbruket bidra med 44 % av fosforbelastningen till Egentliga Östersjön, och olika åtgärder för att minska fosforförlusterna från jordbruksmark tillämpas runtom i Sverige. Våtmarker anläggs ofta för att fånga näringsämnen och partiklar från jordbruksmark innan de läcker ut i vattendrag och slutligen i Östersjön. Tidigare utvärderingar av anlagda våtmarker i Sverige har visat på en varierande och relativt låg fastläggning av fosfor. Osäkerheten kring dessa utvärderingar är dock ganska hög, och bottnar i kunskapsluckor både vad gäller processer för fastläggning och transport av fosfor från mindre jordbruksområden. I denna avhandling utreds därför hur anlagda våtmarker fungerar som fällor för jordpartiklar och partikelbunden fosfor i områden med höga fosforförluster. Sju anlagda våtmarker i jordbruksområden med mycket lerjordar studerades, och mängden fosfor och partiklar som fastlades på botten varierade mycket mellan olika våtmarker (13-108 ton partiklar/ha/yr och 11-175 kg fosfor/ha/yr). De faktorer i området uppströms som var kopplade till fosforfastläggning var lutningen i området, markens lerhalt och innehåll av växttillgänglig fosfor samt områdets djurtäthet. Resultat från fyra våtmarker visade på en hög resuspension (partiklar från botten virvlas tillbaka upp i vattnet), men en del av de uppvirvlade partiklarna kom troligtvis från erosion från våtmarkernas sidor och inte från det material som fastlades på botten. Man såg även indikationer på resuspension från vattenprover tagna i utloppet av en annan våtmark. Där var partikelbunden fosfor klart dominerande, vilket kan ha varit en konsekvens av resuspension från botten. Variationerna av fosforkoncentrationer vid in- och utlopp i sju anlagda våtmarker studerades, för att kritiskt kunna granska tidigare retentionsskattningar. Det var stora variationer i sambanden mellan vattenflöde och fosforkoncentrationer mellan de olika våtmarkerna. De faktorer som påverkade sambandet mellan flöde och koncentration var 1) om det var varm eller kall årstid (d v s sommar eller vinter), 2) om det var högt eller lågt vattenflöde, samt 3) om inflödet bestod av ett dräneringsrör eller ett öppet dike/åfåra. I våtmarker med öppet dike som inflöde var flödes-koncentrationssambandet av totalfosfor negativt vid låga flöden och positivt vid höga flöden. De olika sambanden visar hur viktig sättet att provta vatten är, då inkommande mängd fosfor både kan över- och underskattas om man inte är medveten om sådana variationer. Vid automatisk provtagning styrd av flödesmätningen sker detta ofta i utloppet, men eftersom vattnet har en viss uppehållstid i våtmarken (speciellt sommartid) kan retentionsberäkningen påverkas av att all provtagningen styrs av flödet i utloppet.     För att identifiera vilka områden som bidrar med mest näring och partiklar valdes ett stort område dominerat av jordbruksmark ut – för att undersöka hur man skulle kunna bedöma var anlagda våtmarker skulle kunna göra mest nytta. Området delades in i tio mindre områden, och vattenprover samlades in från diken och åfåror. Resultaten visade på stora skillnader i fosfordynamik mellan olika delområden med olika jordtyper, trots att de hade liknande markanvändning. Det fanns ofta en koppling mellan höga fosforkoncentrationer i vattnet och en hög andel lerjordar i området. För partiklar fanns det en tendens till samband mellan höga koncentrationer och hög andel vinterbar mark. En anlagd våtmark skulle antagligen ha högst effekt om den placerades nedströms områden som är känsliga för erosion – områden med hög andel lerjordar eller med hög andel vinterbar mark. / Eutrophication is one of the more serious current environmental problems, causing algal blooms and anoxic bottoms. In fresh and brackish water, phosphorus (P) is often the most limiting nutrient, and various mitigating strategies are used to reduce the load of P to sensitive recipients. In the agricultural sector, this includes both on-field measures (e.g. managing P inputs) and measures at the field edge (e.g. buffer zones and constructed wetlands). Previous evaluations of constructed wetlands (CWs) in Sweden have indicated a variable and relatively low P retention. However, the uncertainties in the estimates are large, and related to an incomplete knowledge about both retention processes and factors determining the P load from agricultural land. Hence, the overall aim of this thesis was to investigate possible reasons for the variation in wetland P retention estimates, and to assess the P retention in wetlands located in agricultural areas where losses are expected to be high. When comparing seven CWs located downstream small catchments with predominantly arable land, the particle and P net accumulation varied considerably (13-108 t particles ha-1 yr-1 and 11-175 kg P ha-1 yr-1, respectively). Catchment factors that were statistically correlated with accumulation of particles and P in the CWs were the slope of the arable land, the P content of the top soil, the animal density (expressed as livestock units per arable land) and the percentage clay in the topsoils. In four of the wetlands, resuspension was studied using sediment traps and plates. The results showed that up to 87 % of the settled material was resuspended, and indicated that erosion of the wetland sides and bottom probably contributed a substantial part of the trapped particles. In order to critically evaluate existing retention data from earlier investigations, the temporal dynamics of P concentrations and P retention in seven CWs were evaluated. The relationships between water flow and concentration (from grab sampling) varied, and depended on the season (warm or cold period of the year), water flow (high or low) and the inlet type (drainage pipe or open ditch). In CWs that received water through an open ditch, flow-concentration relationships were negative during low flow periods but positive during high flow periods. These differences in flow-concentration relationship have implications for water sampling, since P loads can be both over- and underestimated with grab or automatic sampling guided by clock-time. Also composite automatic sampling, regulated from the water flow at the outlet, can lead to errors in transport calculations since the same ‘water parcel’ is not measured at in and out (difference depending on how long the water retention time is in the CW). This may have an effect on estimates of P retention in both past, present and future investigations of constructed wetlands. Finally, a synoptic sampling approach with ten sampling points was used in an agriculturally dominated catchment area (160 km2) to identify differences in nutrient transport dynamics and areas with the highest losses. Spatial differences in P concentrations were strongly correlated with some of the catchment factors, for instance with soil type, and particle concentrations were weakly correlated to agricultural practices associated with bare soils during winter. This supports the practice to focus P mitigation measures – such as constructed wetlands – to erosion sensitive areas.

The first part of this thesis deals with studying the effect of the specified bathymetric resolution and ideal bathymetric form parameters on the output from the wave and hydrodynamic modules of Delft-3D. This thesis then describes the use of an optimization to effectively reduce the required bathymetric sampling for input to a numerical forecast model, by using the model’s sensitivity to this input. A genetic algorithm is developed to gradually evolve the survey path for a ship, AUV, or other measurement platform to an optimum, with the resulting effect of the corresponding measured bathymetry on the model, used as a metric. Starting from an initial simulated set of possible random or heuristic sampling paths over the given bathymetry using certain constraints like limited length of track, the algorithm can be used to arrive at the path that would provide the best possible input to the model under those constraints. This suitability is tested by a comparison of the model results obtained by using these new simulated observations, with the results obtained using the best available bathymetry. Two test study areas were considered, and the algorithm was found to consistently converge to a sampling pattern that best captured the bathymetric variability critical to the model prediction.

The first part of this thesis deals with studying the effect of the specified bathymetric resolution and ideal bathymetric form parameters on the output from the wave and hydrodynamic modules of Delft-3D. This thesis then describes the use of an optimization to effectively reduce the required bathymetric sampling for input to a numerical forecast model, by using the model’s sensitivity to this input. A genetic algorithm is developed to gradually evolve the survey path for a ship, AUV, or other measurement platform to an optimum, with the resulting effect of the corresponding measured bathymetry on the model, used as a metric. Starting from an initial simulated set of possible random or heuristic sampling paths over the given bathymetry using certain constraints like limited length of track, the algorithm can be used to arrive at the path that would provide the best possible input to the model under those constraints. This suitability is tested by a comparison of the model results obtained by using these new simulated observations, with the results obtained using the best available bathymetry. Two test study areas were considered, and the algorithm was found to consistently converge to a sampling pattern that best captured the bathymetric variability critical to the model prediction.

Regnvatten, smältvatten och spolvatten som rinner av från hårdgjorda ytor kallas dagvatten. Dagvatten är ofta förorenat av tungmetaller eller svårnedbrytbara organiska ämnen, som kan göra stor skada om vattnet inte renas innan det når recipienten. I dagsläget finns det inga nationellt fastslagna riktvärden för föroreningshalter i dagvatten. Dagvattengruppen på Vattenfall AB Värme Uppsala har tillsammans med miljökontoret i Uppsala tagit fram riktvärden för föroreningshalter från kvarteret Brännugnen, som området som studerats i detta examensarbete kallas. Vattenfall har idag problem med att klara dessa riktvärden för bland annat metaller i en del av kvarterets dagvattenbrunnar. För att reducera utsläppet av föroreningar till dagvattennätet från kvarteret Brännugnen har därför denna dagvattenutredning genomförts. Genom platsundersökningar och provtagning på dagvattnet i sex punkter har källor till föroreningar identifierats. Förslag på åtgärder har tagits fram genom litteraturstudier, platsundersökningar och samtal och diskussioner med erfaren personal. Studien har också resulterat i förslag på en förbättrad provtagningsstrategi som ger en mer korrekt bild av utsläppta mängder föroreningar per år. Målsättningen var att den nya provtagningsstrategin skulle vara mer kostnadseffektiv vilket uppnåddes. I dagvattenstudien undersöktes också om rening sker i det underjordiska magasin som dagvattnet från halva kvarteret Brännugnen leds till. Resultatet visade att ingen rening sker i magasinet, då uppehållstiden är för kort för att partiklar ska hinna sedimentera. Förutom vatten från regn- och snösmältning kommer även kondensat och processvatten till dagvattennätet från kvarteret Brännugnen. Ett av målen i utredningen var att undersöka hur mycket kondensatet bidrar till den utsläppta mängden metaller. Provtagning av kondensat och dagvatten visade tillsammans med flödesberäkningar att kondensatet har en stor påverkan på hur stor mängd förorening som släpps ut per år. Riktvärden studerades också i denna dagvattenutredning, slutsatsen är att de riktvärden som Vattenfall förhåller sig till i dag är låga. Både schablonhalterna för värmeverk och vägdagvatten samt analysresultat av dagvatten från vägen utanför området överskrider Vattenfalls riktvärden för dagvatten. Slutligen skulle dagvattenstudien resultera i förslag på framtida åtgärder för rening av dagvattnet från kvarteret Brännugnen i form av tekniska lösningar. Avsättningsmagasin under mark och filteranläggning är två lösningar som är lämpliga för rening av dagvattnet med hänsyn till föroreningar och flöden. Kostnaden beror mycket på vilket flöde som blir dimensionerande. / Stormwater is rain and snow melt that runs off from hard surfaces. Stormwater is often polluted with heavy metals and organic pollutants, which can cause great damage if the water is not treated before it reaches the recipient. There are no national guidelines for threshold values of pollutants in stormwater. The stormwater group at Vattenfall AB Heat Uppsala and the environmental department at Uppsala municipality have together established threshold values for pollutants in the stormwater from the area Brännugnen, which is the name of the investigated area. Vattenfall has difficulties to keep the concentration of pollutants in the stormwater below the threshold values. To reduce the content of contaminants in the stormwater from the area Brännugnen this study was therefore carried out. Through site investigations and sampling of stormwater in six measuring points, sources of the contaminants were identified. Through literature studies, site investigations and discussions with experienced personnel, propositions of actions to avoid contamination of the stormwater have been developed. The stormwater study has also led to suggestion of an improved sampling strategy which better represents the amount of contaminants in the stormwater and is more cost-effective. The study included an investigation with the purpose to find out if the magazine that the stormwater from half of the area runs to have the capacity to reduce the contaminants. The results showed no reduction of contaminants. The reason is that the residence time is too short for the particles to have time to settle. In the study the contribution from condensate to the total amount of emitted metals in the stormwater was investigated. The condensate passes a treatment plant for purification before being discharged to the stormwater pipes. The result shows that the condensate has a large impact on the amount of emitted metals per year. Threshold values and standard values for stormwater was also studied, the conclusion is that the threshold values for the stormwater from the area Brännugnen are low. Both standard values for stormwater from thermal power stations and roads are higher than the threshold values. The results from the analyses of stormwater from the road outside the area Brännugnen also show higher concentrations of metals than the threshold values. Finally the stormwater study resulted in suggestions of technical solutions for future treatment of the stormwater from the area Brännugnen. An underground sedimentation magazine and a filter system are two treatment processes that can be used for removal of pollutants in the stormwater from the area Brännugnen.

The thesis deals with several aspects of monitoring of persistent organic contaminants (POPs) in biological matrices, for example choice of sample, sampling design, and statistical treatment of data both for temporal and spatial trends and for compliance towards a set target value. The efficiency has been evaluated through statistical power analyses. Contaminant data from more than 4 decades from the Swedish National Monitoring Programs for monitoring of contaminants in biota (marine, freshwater and human health), has been quantitatively evaluated both temporally and spatially and for compliance. The aim was also to evaluate the suitability of different matrices, i.e. herring (Clupea harengus), guillemot (Uria aalge) egg, cod (Gadus morhua), perch (Perca fluviatilis), eelpout (Zoarces viviparous), blue mussel (Mytilus edulis), pike (Esox lucius), Arctic char (Salvelinus alpinus) and human milk, for monitoring of POPs with the overall aim to improve the assessment within monitoring programs. The results show that variation can be reduced by using pooled samples including more specimens but fewer chemical analyses, which in turn generate a higher statistical power to a lower cost, at least in cases where the cost of collection and sampling is considerably lower than the cost of chemical analysis. However, there are also a number of advantages using individual samples, such as information of sample variance and maximum value, which allows the choice of an appropriate central measure and direct adjustment of confounding factors. Generally, the levels of polychlorinated biphenyls (PCBs), dichlorodiphenyltrichloroethanes (DDTs), hexachlorocyclohexanes (HCHs) and hexachlorobenzene (HCB) have decreased both in marine and freshwater biota but concentrations are still higher in the Baltic compared to e.g. the North Sea. The levels of dioxinlike-PCBs and polychlorinated dibenzo-p-dioxins/polychlorinated dibenzofurans (PCDD/Fs) have decreased in human milk over time, but not to the same extent in fish and guillemot egg from the Baltic and the freshwater environment. This may be explained by the dietary advice developed by the Swedish Food Administration with the goal that girls, reproductive aged, and pregnant women should eat less food containing high levels of PCDD/Fs. Thus the levels in milk could continue to decrease at the same rate although the temporal trend in the environment has slowed down or leveled out. The most essential regarding the choice of species and matrices for contaminant monitoring, is that the species and organ fit the purpose of the monitoring.

Le secteur du tri des déchets est en pleine expansion. De par la nature de leurs activités, les centres de tri des déchets sont une source d’émission de bioaérosols dont l’exposition peut entrainer diverses troubles sur la santé des travailleurs. La composition des bioaérosols dans les centres de tri est peu documentée. Dans ce contexte, les objectifs de la thèse étaient de caractériser et de déployer dans un centre de tri (pour un suivi d’un an) une méthode d’analyse de la biodiversité microbienne dans les bioaérosols par séquençage à haut débit. Les travaux ont permis de connaitre les avantages et les limites de plusieurs méthodes d’analyse de la biodiversité et de mettre en lumière la nécessité de standardiser un processus de mesure dans le cadre d’études de comparaison. Le suivi de la biodiversité microbienne dans les bioaérosols émis en centre de tri a permis de mettre en évidence la grande complexité et la variabilité des taxons bactériens et fongiques à différents postes de l’entreprise. L’analyse statistique a mis en évidence le caractère multifactoriel de la variation et de la composition des bioaérosols émis en environnement professionnel. La quantité importante de données a permis d’améliorer les connaissances sur la composition des bioaérosols émis en centre de tri des déchets et sur la stratégie de prélèvement à adopter pour une évaluation du risque biologique représentative de l’entreprise / The waste sorting activities is constantly increasing. Due to the nature of their activities, waste sorting plants are a source of bioaerosol emissions whose exposure can lead to various health problems for workers. The composition of bioaerosols in sorting centers is poorly documented. In this context, the aims of the thesis were to characterize and deploy in a sorting center (for a follow- up of one year) a method of analysis of microbial biodiversity in bioaerosols using high throughput sequencing. The work provided information about the advantages and the limits of several methods of analysis of the biodiversity and highlighted the need to standardize a process of measurements for comparative studies. The monitoring of microbial biodiversity in bioaerosols emitted in a sorting plant revealed the high complexity and the variability of bacterial and fungal taxa at different places in the company. Statistical analysis highlighted the multifactorial nature of the variation and composition of bioaerosols emitted in this occupational environment. The large amount of collected data improved the knowledge on the composition of bioaerosols emitted in the waste sorting plant and the sampling strategy to be conducted for a representative biological risk assessment in the company

In order to investigate the etiology of and evaluate interventions for work-related back injuries, researchers need efficient, accurate occupational exposure assessment methods suitable for large samples. The chapters in this thesis examine critical decisions using electromyography (EMG): How should exposure be measured? For what duration? Who should be measured, and how many times? Low-back EMG, or muscle activity data, was collected during 138 full-shift field measurements over 30 different job titles at 50 different worksites in 5 heavy industries: forestry, transportation, wood products, construction, and warehousing. Observations and self-reports of posture, manual materials handling (MMH), and driving exposures were collected concurrently. 1) Variability of EMG calibration measurements was investigated on right/left sides, multiple trials, 4 positions, and pre/post-shift. Position accounts for the majority of explained variability; there is little to gain by measuring multiple trials or pre- and post-shift, but measuring both sides and multiple positions is worthwhile. 2) Observation and self-report data were easier to collect and cheaper than the EMG direct measure. Costs and successful field performance need to be weighed against the added data detail when making choices about exposure assessment techniques for epidemiological studies. 3) Observed and self-reported exposures were used to predict EMG exposure metrics using mixed multiple linear regression models. Regression models using observed variables predicted 43-50% of the variability in the EMG metrics, while self-reported variables predicted 21%-36%. The observation exposure model provides a low-cost alternative to direct measurement. The self-reported exposure model should be considered with more caution. 4) Full-shift EMG data was resampled for 4, 2, and 1 hour, and for 10 and 2 minute durations to determine the optimal sampling duration. Bias was consistently low, but shorter durations had higher absolute error, percentage error, and limits of agreement. Durations of 4 and 2 hours may be acceptable but those less than 1 hour had large errors. 5) Components of EMG variance were calculated between- and within-subject, and between- industry, company, job, and post hoc grouping. Resolution, contrast, and exposure-response relationship attenuation were calculated for each grouping scheme. The post hoc scheme had the highest contrast and lowest resolution.

In order to investigate the etiology of and evaluate interventions for work-related back injuries, researchers need efficient, accurate occupational exposure assessment methods suitable for large samples. The chapters in this thesis examine critical decisions using electromyography (EMG): How should exposure be measured? For what duration? Who should be measured, and how many times? Low-back EMG, or muscle activity data, was collected during 138 full-shift field measurements over 30 different job titles at 50 different worksites in 5 heavy industries: forestry, transportation, wood products, construction, and warehousing. Observations and self-reports of posture, manual materials handling (MMH), and driving exposures were collected concurrently. 1) Variability of EMG calibration measurements was investigated on right/left sides, multiple trials, 4 positions, and pre/post-shift. Position accounts for the majority of explained variability; there is little to gain by measuring multiple trials or pre- and post-shift, but measuring both sides and multiple positions is worthwhile. 2) Observation and self-report data were easier to collect and cheaper than the EMG direct measure. Costs and successful field performance need to be weighed against the added data detail when making choices about exposure assessment techniques for epidemiological studies. 3) Observed and self-reported exposures were used to predict EMG exposure metrics using mixed multiple linear regression models. Regression models using observed variables predicted 43-50% of the variability in the EMG metrics, while self-reported variables predicted 21%-36%. The observation exposure model provides a low-cost alternative to direct measurement. The self-reported exposure model should be considered with more caution. 4) Full-shift EMG data was resampled for 4, 2, and 1 hour, and for 10 and 2 minute durations to determine the optimal sampling duration. Bias was consistently low, but shorter durations had higher absolute error, percentage error, and limits of agreement. Durations of 4 and 2 hours may be acceptable but those less than 1 hour had large errors. 5) Components of EMG variance were calculated between- and within-subject, and between- industry, company, job, and post hoc grouping. Resolution, contrast, and exposure-response relationship attenuation were calculated for each grouping scheme. The post hoc scheme had the highest contrast and lowest resolution.

Les proliférations de cyanobactéries présentent en général d'importantes variations spatiales et temporelles de leur abondance cellulaire et de leur potentiel toxique, ce qui rend très difficile la surveillance de ces microorganismes et l'estimation des risques sanitaires associés à ces événements. Dans ce cadre général, le premier objectif de ma thèse a été de tester différentes stratégies d'échantillonnage pour suivre au mieux l'évolution des abondances cellulaires lors des proliférations de cyanobactéries. Par un échantillonnage à haute fréquence (six points tous les deux jours) réalisé sur un étang piscicole du Forez, nous avons montré que les deux espèces qui proliféraient dans cet écosystème (Microcystis aeruginosa et Aphanizomenon flos-aquae) présentaient des dynamiques spatiales et temporelles de leur abondance cellulaire très contrastées, ne permettant pas de définir une stratégie d'échantillonnage optimale commune aux deux espèces. Si pour ces deux espèces, trois points d'échantillonnage sont au minimum nécessaires pour bien prendre en compte l'hétérogénéité dans la distribution spatiale des cellules, les fréquences d'échantillonnage optimales sont en revanche très variables, allant d'un échantillonnage mensuel ou bimensuel pour Microcystis à un échantillonnage au minimum hebdomadaire pour Aphanizomenon. Le second objectif de ma thèse a été de m'appuyer sur cette stratégie d'échantillonnage à haute fréquence, pour mieux comprendre le développement de la prolifération de M. aeruginosa. Pour ce faire, nous avons estimé les changements spatiaux et temporels intervenant dans la composition génotypique de la population de cyanobactéries (par utilisation de la SSCP sur l'ITS 16S-23S) et dans sa toxicité potentielle (par le dosage des microcystines et l'évaluation de la proportion de cellules possédant le gène mcyB, un des gènes de synthèse de la microcystine). La répartition spatiale de la composition génotypique s'est révélée homogène à l'échelle de l'étang. Il est également apparu qu'au cours de la phase de croissance de la population, cette composition génotypique subissait de nombreux changements à courte échelle temporelle puis restait stable pendant plusieurs semaines lorsque le maximum d'abondance était atteint. Au niveau du potentiel toxique, la proportion de cellules possédant le gène mcyB est resté proche de 60 % durant toute la prolifération avec toutefois, des variations plus importantes pendant le développement du bloom. Aucune relation n'a pu être établie entre les variations observées dans la composition génotypique de la population et celles dans les proportions de cellules toxiques. De même, aucun lien n'a pu être établi entre les variations de diverses variables environnementales (nutriments, température, pluie) et celles de l'abondance cellulaire, de la toxicité potentielle et de la composition génotypique de la population de Microcystis. L'ensemble de ces résultats suggère que d'autres facteurs et processus interviennent probablement dans ces variations et qu'il existe sans doute des interactions très complexes entre toutes ces variables. Enfin, le troisième objectif de ma thèse a été de comparer la composition génotypique et le potentiel toxique de plusieurs populations de Microcystis plus ou moins connectées entre elles. Cette comparaison a permis de montrer l'importance fondamentale des facteurs et processus environnementaux locaux, dans la mise en place et le développement de ces évènements. / Generally, cyanobacteria proliferations show great spatial and temporal variations in their cell abundances and potential toxicity, which makes it difficult to control the development of these microorganisms and to predict the health risks associated with these events. Within this scope, the first goal of my PhD thesis was to test different sampling strategies to guarantee the best monitoring of the cell abundances during cyanobacteria proliferations. We made a high frequency sampling (six points every other day) in a shallow lake located in Forez and we evidenced that the two blooming-species (Microcystis aeruginosa and Aphanizomenon flos-aquae) showed strongly contrasted spatial and temporal patterns of their cell abundances precluding having a common optimal sampling strategy for both species. Even if three sampling points were enough to take into account the spatial heterogeneity of Microcystis and Aphanizomenon cells, a monthly or two-monthly sampling was sufficient for Microcystis whereas a weekly sampling was necessary for Aphanizomenon. The second goal was to gain a better understanding of Microcystis proliferation development. To achieve this aim, we estimated spatial and temporal changes in the genotypic composition (using the SSCP method in the 16S-23S ITS) and in the potential toxicity (by measuring the microcystin concentration and proportion of mcyB+ cells). We obtained a homogeneous spatial repartition of the genotypic composition. Moreover, during the growth phase, there were many rapid changes in the genotypic composition whereas this composition remained stable for several weeks where the maximum cell abundance was reached. As for potential toxicity, the proportion of mcyB+ cells remains at around 60 % during the proliferation but we observed higher variations during the growth phase. No relation was found between the variations of the genotypic composition and proportion of toxic cells on the one hand and the variations of several environmental factors (nutrients, temperature, rain) on the other hand, suggesting that other factors may be involved in these variations and that many complex interactions occur between these factors. Finally, the third goal of my PhD thesis was to compare the genotypic composition and the potential toxicity of different Microcystis populations, which were more or less interconnected. This comparison evidenced the great importance of local environmental factors and processes in the beginning and development of these events.

In order to investigate the etiology of and evaluate interventions for work-related back injuries, researchers need efficient, accurate occupational exposure assessment methods suitable for large samples. The chapters in this thesis examine critical decisions using electromyography (EMG): How should exposure be measured? For what duration? Who should be measured, and how many times? Low-back EMG, or muscle activity data, was collected during 138 full-shift field measurements over 30 different job titles at 50 different worksites in 5 heavy industries: forestry, transportation, wood products, construction, and warehousing. Observations and self-reports of posture, manual materials handling (MMH), and driving exposures were collected concurrently. 1) Variability of EMG calibration measurements was investigated on right/left sides, multiple trials, 4 positions, and pre/post-shift. Position accounts for the majority of explained variability; there is little to gain by measuring multiple trials or pre- and post-shift, but measuring both sides and multiple positions is worthwhile. 2) Observation and self-report data were easier to collect and cheaper than the EMG direct measure. Costs and successful field performance need to be weighed against the added data detail when making choices about exposure assessment techniques for epidemiological studies. 3) Observed and self-reported exposures were used to predict EMG exposure metrics using mixed multiple linear regression models. Regression models using observed variables predicted 43-50% of the variability in the EMG metrics, while self-reported variables predicted 21%-36%. The observation exposure model provides a low-cost alternative to direct measurement. The self-reported exposure model should be considered with more caution. 4) Full-shift EMG data was resampled for 4, 2, and 1 hour, and for 10 and 2 minute durations to determine the optimal sampling duration. Bias was consistently low, but shorter durations had higher absolute error, percentage error, and limits of agreement. Durations of 4 and 2 hours may be acceptable but those less than 1 hour had large errors. 5) Components of EMG variance were calculated between- and within-subject, and between- industry, company, job, and post hoc grouping. Resolution, contrast, and exposure-response relationship attenuation were calculated for each grouping scheme. The post hoc scheme had the highest contrast and lowest resolution. / Medicine, Faculty of / Population and Public Health (SPPH), School of / Graduate