Les marais filtrants artificiels sont des écosystèmes recréés par l’homme dans le but d’optimiser l’épuration des eaux usées. Lors de la sélection d’espèces végétales pour la mise en place de ces marais filtrants, l’utilisation d’une polyculture ainsi que d’espèces indigènes non invasives est de plus en plus recommandée. Néanmoins, la plupart des marais filtrants existants sont des monocultures utilisant des plantes envahissantes, probablement à cause du manque d’évidences scientifiques sur les avantages de la diversité végétale et de la performance des espèces locales. Ainsi, les questions de recherche autour desquelles s’oriente ma thèse sont: Les polycultures présentent-elles un potentiel épuratoire aussi ou plus grand que les monocultures, et une espèce indigène est-elle aussi efficace et performante qu’une espèce exotique envahissante dans des marais filtrants ? Trois expériences ont été conduites afin de répondre à ces questions. J’ai d’abord testé l’influence de la richesse végétale sur l’élimination des polluants en deux dispositifs expérimentaux: 1) comparant deux espèces de plantes émergentes en monoculture ou combinées séquentiellement, et 2) évaluant la performance de quatre espèces flottantes plantées en monoculture par rapport à des associations de deux (avec toutes les combinaisons possibles) et de quatre espèces. Une troisième expérience a été réalisée afin de comparer l’efficacité épuratoire de l’haplotype européen envahissant du roseau commun (Phragmites australis) et de la sous-espèce locale non-invasive (P. australis subsp. americanus). La composition en espèces végétales a produit un effet notable sur la performance des marais filtrants. La comparaison des performances en mono- et en polyculture n’a pas permis de démontrer clairement les avantages de la diversité végétale pour l’élimination des polluants dans les marais filtrants. Toutefois, les marais filtrants plantés avec une combinaison d’espèces étaient aussi efficaces que les monocultures des espèces les plus performantes. La comparaison entre les deux sous-espèces de P. australis indiquent que la sous-espèce indigène pourrait remplacer le roseau exotique envahissant, évitant ainsi les potentiels risques environnementaux sans toutefois compromettre l’efficacité du traitement. Les résultats prometteurs de la sous-espèce indigène de P. australis doivent encore être testés dans des expériences à grande échelle avant d’utiliser largement cette espèce dans les marais filtrants. Nos résultats suggèrent que, dans des conditions où la performance des macrophytes disponibles est inconnue ou ne peut être déterminée, l’utilisation d’une combinaison d’espèces présente les meilleures chances d’accomplir le plus haut niveau possible d’élimination de polluants. De plus, même si la diversité végétale ne présente pas un avantage mesurable en termes d’efficacité épuratoire, celle-ci améliore la résilience des marais filtrants et leur résistance aux stress et aux maladies. / Treatment wetlands (TWs) are complex engineered ecosystems designed to remove pollutants from wastewater. In selecting plant species for TWs, the use of polycultures as well as the choice of local non-invasive species is increasingly recommended. However, to date, the majority of TWs have been planted in monoculture using invasive species, probably because the performance of native species and the possible advantages associated with plant diversity have not been clearly demonstrated. The research questions of my thesis are: Are polycultures as good as or better than monocultures for pollutant removal, and could a native species be as efficient as an exotic invasive species in TWs? Three experiments were performed to answer this question. The influence of plant richness on pollutant removal was tested in two mesocosm experiments: 1) comparing two emergent plant species in monoculture and planted in sequence, and 2) evaluating four free-floating species, each in monoculture versus pairs of species (all possible combinations) and the four species planted together. A third experiment was carried out to compare the removal efficiency of the invasive European haplotype of common reed (Phragmites australis) and the native non-invasive subspecies (P. australis subsp. americanus). Plant species composition influenced the performance of TWs. The comparison of the performance of monocultures versus polycultures showed that, overall, TWs planted with a combination of species were as efficient as monocultures of the best performing species. Comparing the performance of native and invasive P. australis strongly suggests that the native could replace the invasive species in TWs, thereby avoiding possible environmental risks without compromising treatment efficiency. The promising results of the native P. australis subspecies need to be further tested in full-scale experiments in order to encourage the use of this species in TWs. The results obtained comparing plant species combinations suggest that under circumstances in which the performance of available macrophytes is unknown or cannot be determined, using a mixture of species offers the best chances of achieving the highest possible level of pollutant removal. Furthermore, it has been suggested that plant diversity contributes to TWs in terms of their resilience to stress and disease.

In Australia, significant effort goes into reducing the amount of nitrogen and phosphorus entering inland waters from point sources. However, little is known of the extent to which riparian organic matter may act as a source of these nutrients. Also, whilst the relationships between the nitrogen, phosphorus and carbon cycles are broadly known, there is little quantitative data regarding the release of these elements from Australian riparian organic matter and their subsequent microbial mineralisation within aquatic environments. In particular, comparatively little is known of their comparative role in nutrient and organic matter cycling within anoxic zones, and the influence that different riparian organic matter may have on stream water quality. This lack of such data presently hampers the ability of water managers to make educated decisions regarding the management of riparian zones in Australia. In order to improve understanding in this area, a combination of laboratory and in situ experiments were carried out in order to compare the abiotic release and aerobic/ anaerobic mineralisation of leaf derived dissolved organic carbon (DOC), dissolved nitrate/nitrite (NOx) and soluble reactive phosphorus (SRP) under different environmental conditions. Four plants common to Australian riparian zones were investigated: two native species, Eucalyptus camaldulensis (gum) and Phragmites australis (common reed), and two exotic species, Salix babylonica (willow) and Lolium multiflorum (rye grass). After 30 days, formaldehyde inhibited 1g willow and rye grass extracts contained the most SRP (0.7 mg/L), whilst gum extracts contained 0.3 mg/L and common reed 0.1 mg/L of SRP.Willow and rye grass abiotically released twice as much NOx than gum and common reed, although concentrations were only between 0.05-0.1 mg/L. Gum and common reed released the most DOC per gram of leaf matter (14 and 12 mmol/g of leaf matter respectively), but based on the initial carbon content of each leaf type, the largest percentage contributor of DOC under abiotic conditions was common reed and rye grass (both 38% mass/mass), with gum (33% mass/mass) and willow (30% mass/mass) being smaller contributors. The most bioavailable DOC was released by rye grass and common reed, with between 83 and 94% of this DOC microbially mineralised after 30 days in oxic conditions. When conditions were not inhibited, microbial growth was evident almost immediately in willow, rye grass and common reed leaf extracts. However, microbial growth was suppressed for the first 48 hours in gum leaf extracts. After this suppression period, the rate of DOC mineralisation was equal in willow and gum leaf extracts (0.1 day-1). Under anoxic conditions, the rate and extent of DOC mineralisation of willow and gum leaves depended on the type of electron acceptor provided. Added nitrate and iron III enhanced the mineralisation of both willow and gum leaves relative to no terminal electron acceptors (from zero to 0.01-0.04 and 0.002- 0.004 moles/day respectively), but added sulphate only enhanced the mineralisation of gum leaves (0.04 moles/day). When no additional electron acceptors were provided, particulate leaf mineralisation was more extensive under oxic than anoxic conditions. However, the mineralisation of leaf derived DOC were the same regardless of oxygen availability, and after 35 days in either condition the percentage of leaf DOC mineralised for each leaf type was of the order common reed > rye grass > willow > gum. All the leaf types tested were able to sustain the caddis fly larvae Triplectides australis under controlled laboratory conditions, and survival rates were high using all four leaf types as a food source. Triplectides australis did not significantly increase the amount of DOC released from each type of leaf matter, but they did consistently increase the proportion of simple carbohydrates present within the DOC fraction. The results of these experiments suggest that changes to riparian vegetation, particularly from the native to exotic species used in this study, will inherently alter in-stream concentrations of dissolved carbon and nutrients (particularly SRP). This potentially will affect in-stream, hyporheic and subsurface processes, particularly in areas where surface water flow is low and riparian leaf inputs are high.

Thesis (doctoral)--Swedish University of Agricultural Sciences, 2007. / Thesis documentation sheet inserted. Appendix reproduces four papers and manuscripts co-authored with others. Includes bibliographical references. Also issued electronically via World Wide Web in PDF format; online version lacks appendix.

The objective of this work was to understand microbial population density and diversity, both spatially and temporally, in wetland mesocosms to gain a better fundamental understanding for use in the optimization and design of constructed wetlands (CWs). A standardized community level physiological profiling (CLPP) data analysis protocol was adapted and utilized for CW mesocosms. A new one-dimensional metric was developed to track community divergence using BIOLOGTM ECO plate data. The method proved easy to use, did not require a background in multivariate statistics, and accurately described community divergence in mesocosm systems. To study mesocosm biofilm-bound bacterial communities an appropriate detachment protocol was required. Various shaking protocols were evaluated for their effectiveness in the detachment of bacteria from mesocosm pea gravel, with a focus on detachment of viable and representative bacterial communities. A protocol based on mechanical shaking with buffer and enzymes was identified as an optimal approach and used further in this study. The bacterial communities associated with the interstitial water, pea gravel media, and rhizospheric regions from both planted and unplanted CW mesocosms were profiled using the CLPP method and compared. Vertical community stratification was observed for all mesocosm systems. Rhizospheric communities were found to be significantly more active than their gravel-associated counterparts, suggesting that although rhizospheric bacteria were less abundant in the mesocosms they may play a more significant role in the removal and fate of water born contaminants. The start-up dynamics of CW mesocosms was investigated using the CLPP and standard CW characterization methods over an eight month period. All mesocosms showed a steep increase in interstitial community divergence until day 75-100, at which point a steady-state was reached. The interstitial communities were also characterized in terms of similarity based on experimental design treatments (planted/unplanted and origin of seeding inoculum). Four stages were identified during the start-up consisting of an initial stage where mesocosm communities were differentiated based on origin of the inoculum, a period where adjustments and shifts occurred in all mesocosm, a time where all mesocosm communities were quite similar, and a final state where community differentiations were made based plant presence in the mesocosms.

The objective of this work was to understand microbial population density and diversity, both spatially and temporally, in wetland mesocosms to gain a better fundamental understanding for use in the optimization and design of constructed wetlands (CWs). A standardized community level physiological profiling (CLPP) data analysis protocol was adapted and utilized for CW mesocosms. A new one-dimensional metric was developed to track community divergence using BIOLOGTM ECO plate data. The method proved easy to use, did not require a background in multivariate statistics, and accurately described community divergence in mesocosm systems. To study mesocosm biofilm-bound bacterial communities an appropriate detachment protocol was required. Various shaking protocols were evaluated for their effectiveness in the detachment of bacteria from mesocosm pea gravel, with a focus on detachment of viable and representative bacterial communities. A protocol based on mechanical shaking with buffer and enzymes was identified as an optimal approach and used further in this study. The bacterial communities associated with the interstitial water, pea gravel media, and rhizospheric regions from both planted and unplanted CW mesocosms were profiled using the CLPP method and compared. Vertical community stratification was observed for all mesocosm systems. Rhizospheric communities were found to be significantly more active than their gravel-associated counterparts, suggesting that although rhizospheric bacteria were less abundant in the mesocosms they may play a more significant role in the removal and fate of water born contaminants. The start-up dynamics of CW mesocosms was investigated using the CLPP and standard CW characterization methods over an eight month period. All mesocosms showed a steep increase in interstitial community divergence until day 75-100, at which point a steady-state was reached. The interstitial communities were also characterized in terms of similarity based on experimental design treatments (planted/unplanted and origin of seeding inoculum). Four stages were identified during the start-up consisting of an initial stage where mesocosm communities were differentiated based on origin of the inoculum, a period where adjustments and shifts occurred in all mesocosm, a time where all mesocosm communities were quite similar, and a final state where community differentiations were made based plant presence in the mesocosms.

De nombreuses études en phytoremédiation visent à accroître le prélèvement des métaux par les plantes pour dépolluer les sols. Ce travail porte sur une nouvelle approche de phytoremédiation appelée Jardins Filtrants qui consiste à traiter le sol dans des casiers végétalisés par des plantes de milieux humides (Phragmites australis, Iris pseudacorus et Salix viminalis) et irrigués de manière à imposer une alternance des conditions hydromorphie-assèchement afin d'accroître la solubilité de métaux dans le sol et de les extraire par lixiviation. Dans une expérience pilote de seize mois, cette approche a été appliquée pour la phytoremédiation de Zn, Cu et Pb d'un sol agricole fortement pollué par l'épandage d'eaux usées. Le bilan de masse des métaux dans les systèmes sol-plante a montré que seule une quantité non-significative des métaux a été accumulée dans la biomasse des plantes. Une quantité importante des métaux a été éliminée du sol via la phytolixiviation résultant de l'interaction de l'activité racinaire avec l'irrigation. Un traitement chimique complémentaire au citrate permet d'augmenter la lixiviation. Les mécanismes de transformations de Zn et Cu impliqués dans cette phytoremédiation ont été mis en évidence par combinaison des techniques analytiques sur la source synchrotron à micro- (µXRF, µXRD, µEXAFS) et macro-échelle (EXAFS) couplée aux analyses chimiques, permettant d'identifier et quantifier les formes des métaux dans le sol. Dans le sol initial, le zinc a été majoritairement sous formes de minéraux secondaires (Zn-ferrihydrite, Zn-phosphate et Zn-phyllosilicate modélisé par Zn-kérolite) et le cuivre a été associé essentiellement à la matière organique. L'activité racinaire dans les conditions hydromorphie-assèchement a profondément modifié la spéciation des métaux. Zn-ferrihydrite, une des formes majoritaires de Zn, a été complètement dissoute. La dissolution réductive de cet oxyhydroxyde de fer, favorisée par les conditions d'hydromorphie, a induit la lixiviation de Zn. Une partie de Zn solubilisé a coprécipité avec Fe en un autre oxyhydroxyde de fer zincifère moins soluble, Zn-goethite substituée, dans les conditions oxydantes et avec assistance des racines formant des plaques de goethite en défense contre l'excès de métaux dissous. De plus, les nouvelles particules de Zn métallique et ZnO ont été découvertes dans la rhizosphère, en faible quantité. L'oxydation de la matière organique a induit l'excès de Cu cationique toxique. En réponse au stress oxydant, ce cuivre a été biotransformé par les racines en association avec des mycorhizes en nanoparticules de Cu métallique, en quantité importante. Ce nouveau mode de biominéralisation peut être typique des plantes de milieux humides. Cette nouvelle voie de phytoremédiation implique principalement la phytolixiviation induisant la solubilisation des métaux et leur lixiviation et la phytotransformation, due pour une part à la phytodétoxication, conduisant la conversion des métaux toxiques en formes peu solubles.

Many plant species are closely related to river biotopes or to biotopes influenced by rivers. River systems create important linear corridors in ecosystems and directly or non-directly influence spatial spread of species in these environments. This offer many questions about species spread in this system. We can answer these questions by using molecular methods. Using 10 microsatellite (SSRs) primers, 202 individuals of Phragmites australis from 60 populations were analysed. Those analyses allowed reveal kinship of individuals, obtain information about spatial spread of populations and about spreading of common reed. Phragmites australis creates both - monoclonal and polyclonal - populations. Dependence of rate of clonality on environment was revealed. Populations from river banks are more monoclonal, populations from pond shores are more polyclonal. Populations are isolated. The highest percentage of variability was explained on among-population level. This is common for anemochoric species. Communication between populations is present, more on shorter distances. Evidence for vegetative spread was found on short distance. Generative spread is much more common. Long-distance spreading is mediated by generative diaspores - seeds. This spreading includes within-river spread, among-river spread and spreading...