Numerous studies have looked at biotic succession at the aboveground level; however, there are no studies describing fungal community change associated with long-term ecosystem development. To understand ecosystem development, the organisms responsible for shaping and driving these systems and their relationships with the vegetation and soil factors, it is critical to provide insight into aboveground and belowground linkages to ultimately include this new information into ecosystem theory. I hypothesized that fungal communities would change with pedogenesis, that these changes would correlate with vegetation community change, and that they should show change of composition and diversity as the seasons change. Chapter 1 discusses the main topics related to this dissertation. Chapter 2 includes a publication draft that describes a study of sand-dune soil samples from northern Michigan that were analyzed to pinpoint the structural change in the fungal community during the development of the ecosystem. The samples were analyzed by pyrosequencing the soil DNA, targeting the internal transcribed spacer region. Chapter 3 contains a coauthored published paper that describes plant invasion of fields in Virginia to determine how they impact soil bacterial and fungal communities. The bacterial and fungal communities that were invaded by 3 different plant species exhibited similar changes, regardless of plant species, suggesting that some functional traits of invasives may have similar impacts on belowground communities. Chapter 4 remarks the conclusions of this research. / Master of Science / Ecosystems, including the soils underneath, are the environments that surround us perform a large number of critical human-relevant functions (playing roles in production of food, filtration of water for drinking, sequestration of carbon and nitrogen to build soil organic matter, and buffer against flooding). Yet, how these systems naturally develop over time are still in need of detailed study. One particular area of interest and need is the study of belowground fungal communities. It is not commonly known, but plants and ecosystems are highly dependent on the underground web of fungal hyphae that transform nutrients and provide water to plants. A first step in gaining this understanding utilized a natural ecosystem development gradient known as a chronosequence. It was expected that fungal communities would change as soil and ecosystem development progressed and that they would mimic changes in soil and vegetative properties. Discerning if these linkages occur is the first step to assessing how they work together to create ecosystems and their valuable environmental services. Chapter 1 provides a discussion of the main topics in this dissertation. Chapter 2 is at the heart of the dissertation via a study of fungal communities in a developmental soil ecosystem in northern Michigan in addition, in Chapter 3, I include a coauthored published paper that describes plant invasion of fields in Virginia. Chapter 4 remarks on the major conclusions of this Master thesis, supporting the role that vegetation and fungal community change in soil are associated with one another.

Soil microbial communities mediate important ecological processes and play essential roles in biogeochemical cycling. Ecosystem disturbances such as surface mining significantly alter soil microbial communities, which could lead to changes or impairment of ecosystem functions. Reforestation procedures were designed to accelerate the reestablishment of plant community and the recovery of the forest ecosystem after reclamation. However, the microbial recovery during reforestation has not been well studied even though this information is essential for evaluating ecosystem restoration success. In addition, the similar starting conditions of mining sites of different ages facilitate a chronosequence approach for studying decades-long microbial community change, which could help generalize theories about ecosystem succession. In this study, the recovery of microbial communities in a chronosequence of reclaimed mine sites spanning 30 years post reforestation along with unmined reference sites was analyzed using next-generation sequencing to characterize soil-microbial abundance, richness, taxonomic composition, interaction patterns and functional genes. Generally, microbial succession followed a trajectory along the chronosequence age, with communities becoming more similar to reference sites with increasing age. However, two major branches of soil microbiota, bacteria and fungi, showed some contrasting dynamics during ecosystem recovery, which are likely related to the difference in their growth rates, tolerance to environmental change and relationships with plants. For example, bacterial communities displayed more intra-annual variability and more complex co-occurrence networks than did fungi. A transition from copiotrophs to oligotrophs during succession, suggested by taxonomic composition shifts, indicated that the nutrient availability is one important factor driving microbial succession. This theory was also supported by metagenomic analysis of the functional genes. For example, the increased abundance of genes involved in virulence, defense and stress response along ages indicated increased competition between microorganisms, which is likely related to a decrease of available nutrients. Metagenomic analysis also revealed that lower relative abundances of methanotrophs and methane monooxygenase at previously-mined sites compared with unmined sites, which supports previous observations that ecological function of methane sink provided by many forest soils has not recovered after 30 years. Because of the difficulty identifying in situ functional mechanisms that link soil microorganisms with environmental change, modeling can be a valuable tool to infer those relationships of microbial communities. However, the extremely high richness of soil microbial communities can result in extremely complicated models that are difficult to interpret. Furthermore, uncertainty about the coherence of ecological function at high microbial taxonomic levels, grouping operational taxonomic units (OTUs) based on phylogenetic linkages can mask trends and relationships of some important OTUs. To investigate other ways to simplify soil microbiome data for modeling, I used co-occurrence patterns of bacterial OTUs to construct functional groups. The resulting groups performed better at characterizing age-related microbial community dynamics and predicted community structures and environmental factors with lower error. / PHD / Disturbances to ecosystems are known to largely impact important ecological functions such as soil carbon loss, decreased nutrient retention and increased greenhouse gas emission. As a result, surface mining, which totally removes the topsoil and original vegetation, has severe negative influences on forest ecosystem function. Reforestation is performed on reclaimed mined sites to accelerate the return of forest vegetation and ecosystem functions. Although considerable research has shown that the plant community can be well developed after 30 years, little is known on whether ecosystem functions are also recovered during a similar time period. As direct mediators of many ecological processes in the environment, soil microorganisms are important for understanding the restoration progress of ecosystems. They could also provide early indications of restoration progress compared to plants. Historically, most soil microorganisms have been difficult to study because they are highly diverse and the majority cannot be cultured in lab, making it difficult to understand changes in the total soil microbiota. However, technological advances such as DNA sequencing have made it feasible to study soil microorganisms in detail. In this work, we studied soil microbial communities from reclaimed mined sites ranging from 5 to 30 years post-reforestation. We found that overall the microbial community was recovering from the disturbances of surface mining, but many differences from unmined soils still remain after 30 years, such as the unrecovered function as methane sink. Two major groups of soil microorganisms, bacteria and fungi, showed different characteristics during recovery, which are likely due to differences between the two groups with regard to growth rates, tolerance to environmental change and relationships with plants. Mathematical modeling is a useful tool for simulating changes and impacts on microbial communities under different conditions, given that actual interactions between microorganisms and their environment can be difficult to measure. However, the high complexity of soil microbial communities becomes an obstacle for modeling that needs to be addressed by simplifying data describing soil microbial community. One approach is grouping organisms based on their natural evolutionary relationships, but this can mask the trends of some microorganisms since all organisms in these groups do not always respond the same to environmental change. Here we used a method of grouping microorganisms based on their co-occurrence patterns, which resulted in better predictions of changes in community structure and environmental factors when applied in modeling.

Phosphorus (P) is an essential element for life on our planet. It is central in numerous biochemical processes in terrestrial and aqueous ecosystems including food production; and it is the primary growth-limiting nutrient in some of the world’s biomes. The main source of P for use as agricultural fertilizer is mining of non-renewable mineral phosphate. In terrestrial ecosystems the main source is soil P, where the largest fraction is organic P, composed of many species with widely differing properties. This fraction controls the utilization of P by plants and microorganisms and influences ecosystem development and productivity. However, there is only scarce knowledge about the molecular composition of the organic P pool, about the processes controlling its bioavailability, and about its changes as soils develop. Therefore, the aim of this thesis was to develop robust solution- and solid-state 31P nuclear magnetic resonance spectroscopy (NMR) methods to provide molecular information about speciation of the organic P pool, and to study its dynamics in boreal and tropical soils. By studying humus soils of a groundwater recharge/discharge productivity gradient in a Fennoscandian boreal forest by solution- and solid-state NMR, it was found that P speciation changed with productivity. In particular, the level of orthophosphate diesters decreased with increasing productivity while mono-esters such as inositol phosphates increased. Because the use of solution NMR on conventional NaOH/EDTA extracts of soils was limited due to severe line broadening caused by the presence of paramagnetic metal ions, a new extraction method was developed and validated. Based on the removal of these paramagnetic impurities by sulfide precipitation, a dramatic decrease in NMR linewidths was obtained, allowing for the first time to apply modern multi-dimensional solution NMR techniques to soil extracts. Identification of individual soil P-species, and tracking changes in the organic P pools during soil development provided information for connecting P-speciation to bioavailability and ecosystem properties. Using this NMR approach we studied the transformation of organic P in humus soils along a chronosequence (7800 years) in Northern Sweden. While total P varied little, the composition of the soil P pool changed particularly among young sites, where also the largest shift in the composition of the plant community and of soil microorganisms was observed. Very old soils, such as found Africa, are thought to strongly adsorb P, limiting plant productivity. I used NMR to study the effect of scattered agroforestry trees on P speciation in two semi-arid tropical woodlands with different soil mineralogy (Burkina Faso). While the total P concentration was low, under the tree canopies higher amounts of P and higher diversity of P-species were found, presumably reflecting higher microbial activity. / <p>I delarbete III har titel och författaruppgifter förändrats.</p>

Submitted by Erika Demachki (erikademachki@gmail.com) on 2015-01-16T17:09:22Z No. of bitstreams: 2 license_rdf: 23148 bytes, checksum: 9da0b6dfac957114c6a7714714b86306 (MD5) Dissertação - Diego Oliveira Ribeiro - 2012.pdf: 696702 bytes, checksum: 93b9aedefecff8fe2650183e7d1a7c73 (MD5) / Approved for entry into archive by Erika Demachki (erikademachki@gmail.com) on 2015-01-16T17:34:29Z (GMT) No. of bitstreams: 2 license_rdf: 23148 bytes, checksum: 9da0b6dfac957114c6a7714714b86306 (MD5) Dissertação - Diego Oliveira Ribeiro - 2012.pdf: 696702 bytes, checksum: 93b9aedefecff8fe2650183e7d1a7c73 (MD5) / Made available in DSpace on 2015-01-16T17:34:29Z (GMT). No. of bitstreams: 2 license_rdf: 23148 bytes, checksum: 9da0b6dfac957114c6a7714714b86306 (MD5) Dissertação - Diego Oliveira Ribeiro - 2012.pdf: 696702 bytes, checksum: 93b9aedefecff8fe2650183e7d1a7c73 (MD5) Previous issue date: 2012-07-03 / ABSTRACT - The conversion of natural ecosystems for agriculture results in the initial entropy of the system affecting the dynamics of soil organic matter. Several areas were incorporated into the production process from the 70 encouraged by the government, among them more sensitive areas, without any previous study. Thus studies are needed on the impacts of agricultural activities in these environments. Therefore, the purpose of this study was to evaluate the impact of conversion of native Coval for agriculture in the dynamics of organic matter and its fractions.The study was conducted in areas belonging to the Fazenda Boa Vista in Haplic Plinthosol, subjected to a chronosequence of 7, 11 and 16 years of cultivation, managed under no-tillage system, and a reference area characterized by top and bottom of the murundus. In July 2010 the soil was sampled at depths from 0 to 0,025; 0,025 to 0,05; to 0,05 to 0,075; 0,075 to 0,10; 0,10 to 0,15 and 0,15 to 0,20 m. The carbon stocks were similar to the native top with 7 years of implementation no tillage having a greater tendency to increase with the increase of conducting this management system. However when compared to the native range of bottom of the murundus there is a reduction of carbon stocks by presenting different moisture conditions of farming systems and native range top.The weighted mean diameter of soil aggregates accompanied the stocks of soil organic matter, higher than in native areas / RESUMO – A conversão de ecossistemas naturais para a atividade agrícola resulta numa entropia inicial do sistema afetando a dinâmica da matéria orgânica do solo. Diversas áreas foram inseridas ao processo produtivo a partir da década de 70 pelo incentivo do governo, estando entre elas áreas mais sensíveis, sem nenhum estudo prévio. Assim são necessários estudos sobre os impactos da atividade agrícola nestes ambientes. Portanto o objetivo deste estudo foi avaliar o impacto da conversão de áreas nativas de coval para a atividade agrícola, em plantio direto, sobre a matéria orgânica do solo e suas frações. O estudo foi desenvolvido em áreas pertencentes à Fazenda Boa Vista, em Plintossolo Háplico, submetido a uma cronossequência de 7, 11 e 16 anos de cultivo, manejadas sob sistema de plantio direto, e uma área referência caracterizada pelo topo e base dos murundus. As amostras de solo foram coletadas em julho de 2010 nas profundidades 0 a 0,025; 0,025 a 0,05; 0,05 a 0,075; 0,075 a 0,10; 0,10 a 0,15 e 0,15 a 0,20 m. Os estoques de carbono foram semelhantes a área nativa de topo com 7 anos de implantação do plantio direto, tendo maior tendência de elevação com o aumento de condução deste sistema de manejo. Entretanto quando comparado à base dos murundus ocorre uma redução dos estoques de carbono por apresentar condições de umidade diferentes dos sistemas agrícolas e do topo dos murundus. O diâmetro médio ponderado dos agregados acompanhou os estoques de matéria orgânica do solo, sendo superiores nas áreas nativas.

Succession is often studied by using a chronosequence. When using a chronosequence we study a set of sites with different ages at the same time and by comparing them we conclude what kind of changes occurred during time (space for time substitution). Only a few studies however compare how results obtained by using a chronosequence differ from those obtained by long-term studies. In my theses I repeated a study that investigated succession of ant communities on brown coal mining spoil dumps in Sokolov district after 19 years. There are chronosequences of two types of sites (spontaneous succession and recultivation) in Sokolov coal mining district. By repeating the original study I could compare changes that occurred du- ring time with changes along a chronosequnce. Relationship between occurrence of ant groups with different ecological requirements and age of site was also investigated. RDA model and variation partitioning were used to find out statistical significance between sites and their age. An increase in number of species was recorded on the spoil dumps. 22 ant species were found in the year 2020, from which 5 species were new on the dumps. All the new species are specialists, two of them are dendrophilous. A statistically significant increase in abundance of forest species with site age was...

No description available.

The thesis was conducted to investigate soil fertility changes and assess the sustainability of agroecological management in different land-use systems of the southern Ecuadorian Andes using quantitative and qualitative methods. Ecuador still holds the highest deforestation rate of all Latin American countries which also has a large impact in the research area by forest conversion into agricultural land. Agricultural land-use systems in the research area are multifaceted due to heterogeneous biophysical and socio-economic conditions. To map this diversity, land-use systems were investigated in Yantzaza (low-external-input), El Tambo (irrigated cash crops) and San Lucas (integrated nutrient management). Yet, management effects on soil fertility have not been assessed systematically in Ecuadorian farming systems which hampers the evaluation whether agroecological management is sustainable. Therefore, the present study used a set of quantitative and qualitative approaches to assess soil fertility changes at plot and farm scale with a nutrient balance/chronosequence approach and local expert knowledge. Nutrient balances were modeled with Nutmon after adaptation of difficult-to-quantify flows to the local conditions facilitating area and land-use specific calculation. Soil nutrient balances in the research area were diverse and varied between −151 to 66 kg ha-1 a-1, −4 to 33 kg ha-1 a-1 and −346 to 39 kg ha-1 a-1 for NPK, respectively. The evaluation of socio-economic and soil fertility explanatory variables revealed that up to 70% of the balances’ variability could be explained. Land-uses with a strong market orientation such as annual crops in El Tambo received large amounts of external inputs which were often focused on mineral N fertilization causing strongly negative PK balances. In contrast, P balances were mainly positive after the application of organic fertilizers and nutrient recycling as was found in perennial crops of San Lucas. NP balances in annual crops of Yantzaza were most negative due to the low-external-input system with nonexistent fertilization as well as leaching and burning of crop residues. Highest soil nutrient stocks were found in land-uses benefiting from a surplus of within-farm flows. The quantification of soil nutrient stocks and their temporal changes were carried out with a chronosequence approach in Yantzaza. SOC stocks in annual/perennial crops and pastures decreased between 14% and 19% after forest conversion by slash-and-burn. Annual sites were abandoned not later than five years after forest conversion due to a shortage of available N and P closely linked to low-external-input management. Stocks for TN, TP, TS and exchangeable bases increased above forest level in perennial crops and pastures 6-20 years after forest conversion. Yet a strong decrease in SOC and soil nutrient stocks was found in oldest perennial and pasture sites compared to medium aged sites. This was traced back to adverse site processes such as the decay of clay humus complexes, leaching as well as poor pasture management. To assess sustainability of the agroecological management, a set of sustainability indicators was implemented including N balances, yearly N stock change and SOC stocks as well as total (TN) and available (PO4-P) soil nutrient stocks. Sustainability assessment took place based on individual land-uses and nutrients within each pilot study since soil fertility change did not show a consistent trend within one research area. Despite mainly negative soil nutrient balances, the impact on the yearly soil nutrient stock change was often negligible due to large soil nutrient pools. Annual and perennial crops of Yantzaza and pastures of San Lucas exceeded the threshold value of 1% for yearly TN stock losses. Yet, only annual crops in Yantzaza, having the highest yearly TN stock losses of 4.9%, also showed severe TN and SOC losses between 15-25% below those of the forest reference area. Therefore, the present agroecological management of annuals in Yantzaza is not sustainable which was also indicated by the abandonment of these sites not later than 5 years after forest conversion due to soil fertility decline. Hence, it is proposed to install an integrated agricultural management in annual crops of Yantzaza using nutrient recycling and fertilization for the replenishment of soil nutrient stocks. Nutrient balance studies indicated an average N-fertilizer application of more than 200 kg ha-1 a-1 for annual crops in El Tambo and low SOC stocks in soils of the colluvial foot slopes. Therefore, a laboratory incubation experiment was conducted to investigate fertilization effects of urea and newly introduced guinea pig manure on the microbial activity in colluvial and eroded soils of El Tambo. While urea fertilization induced an acceleration of SOM mineralization, a combined fertilization (urea + GPM) increased the amount of microbial biomass and provided mineral nitrogen for immediate plant uptake. SOM stocks in colluvial soils were 40% below those of eroded soils which was partly due to the positive priming effect after urea fertilization. A participatory appraisal with local farmers resulted in the adaptation of the present harvest residue management aiming at SOM maintenance in colluvial soils. Yet, the calculation of the potential for SOM replenishment indicated that only the maize residue biomass had the potential to compensate for SOM mineralization losses. Therefore, it is recommended to support SOM replenishment by additional organic inputs since SOM has to be maintained in the long-term to enable agricultural productivity.

O número de projetos de restauração de ecossistemas florestais no Brasil aumentou nas últimas décadas e, consequentemente, aumentaram as exigências na avaliação e na busca por indicadores de sucesso. Apesar dos avanços na ecologia da restauração, existem lacunas no conhecimento em relação ao funcionamento e à autossustentabilidade dos ecossistemas restaurados. Esta pesquisa teve o objetivo de responder às seguintes questões: (1) É possível detectar padrões nos atributos de riqueza, estrutura e grupos funcionais nas matas ciliares nativas da Floresta Estacional Semidecidual (FES), que podem ser utilizados como metas da restauração florestal nesta região? (2) Quais as variáveis de biodiversidade, estrutura florestal e grupos funcionais, nas áreas restauradas, que seguem trajetórias previsíveis com o tempo? (3) Quanto tempo é necessário para que estas variáveis atinjam os valores de referência das matas ciliares nativas? (4) Quais são os fatores bióticos, abióticos, de técnica de plantio e de manutenção pós-plantio que influenciam nas trajetórias das matas ciliares restauradas? O estudo foi desenvolvido em região de FES, onde amostramos quatro matas nativas (referências) e 26 plantios de restauração, com idade entre quatro e 53 anos, que utilizamos para modelar as trajetórias em cronoseqüência. Em cada área, dez parcelas de 100 \'M POT.2\' foram aleatoriamente distribuídas, dentro de uma faixa de até 50 m de largura a partir da margem. Identificamos e contabilizamos os indivíduos de espécies arbóreas e arbustivas, plantadas ou regenerantes, a partir de 50 cm de altura e os agrupamos em três classes de tamanho segundo o diâmetro à altura do peito (DAP): DAP < 1 cm, 1 cm \'< ou =\' DAP < 5 cm e DAP \'> ou =\' 5 cm. Na classe de maior tamanho registramos o DAP e estimamos a altura de cada indivíduo. Classificamos as espécies de acordo com a sua síndrome de dispersão, ritmo de crescimento, tolerância à sombra, capacidade de fixar N, grau de ameaça e raridade. Para cada parcela, medimos a abertura do dossel, contabilizamos o número de lianas, pteridófitas e árvores com epífitas e coletamos uma amostra de solo, para formar uma amostra composta para cada local de estudo. Calculamos a biomassa acima do solo para árvores com DAP \'> ou =\' 5 cm, a riqueza total estimada por Jackknife e a riqueza rarefeita para 100 indivíduos. Por meio de entrevistas, observações em campo e análises da paisagem, obtivemos dados representativos de fatores abióticos, bióticos e de manejo que poderiam influenciar as trajetórias das comunidades arbóreas restauradas. Modelamos a trajetória de cada variável relativa à comunidade arbustivo-arbórea em função do tempo e estimamos o tempo que será necessário para igualar os ecossistemas de referência para cada variável. Entre as variáveis analisadas, selecionamos, como indicadores para monitoramento da evolução e sustentabilidade do ecossistema restaurado, as que fossem mais previsíveis (melhor qualidade dos modelos) e que representassem processos ecológicos mais relevantes. Os atributos que não variaram entre florestas nativas e que puderam ser utilizados como referência foram: riqueza observada, riqueza estimada para 100 indivíduos com DAP \'> ou =\' 5 cm, riqueza estimada por Jackknife para indivíduos com DAP \'> ou =\' 1 cm, densidade de indivíduos com DAP \'> ou =\' 1 cm, área basal, biomassa, cobertura de copas, proporção de indivíduos tolerantes à sombra e proporções de espécies zoocóricas, de crescimento lento, tolerantes à sombra, fixadoras de nitrogênio e de distribuição rara. No geral, as trajetórias foram mais bem ajustadas aos modelos logarítmicos, mas a riqueza de regenerantes de maior tamanho (DAP \'> ou =\' 5) ajustou-se melhor ao modelo linear. A riqueza nesta classe de tamanho não atingiu o nível de regenerantes dos ecossistemas de referência no período estudado. As trajetórias seguiram padrões semelhantes, independentes das técnicas de plantio ou do número de espécies plantadas. Variáveis de estrutura florestal foram homogêneas entre as matas nativas. Nos plantios, são facilmente medidas e evoluem rapidamente. Assim, metas da restauração podem ser estabelecidas com base nos parâmetros estruturais das matas nativas e tais variáveis seriam bons indicadores inclusive em plantios de restauração mais jovens. As variáveis que tendem a alcançar mais rapidamente os valores de referência são: área basal (12 anos), biomassa (13 anos), altura média das maiores árvores (26 anos), cobertura de gramíneas (33 anos), cobertura de copas (35 anos), riqueza de regenerantes com DAP \'> ou =\' 1 cm (52 anos), densidade de regenerantes com DAP \'> ou =\' 1 cm (52 anos) e riqueza total (53 anos). A área basal e a biomassa ultrapassaram em muito os valores de referência, o que pode indicar um filtro restritivo para a regeneração natural, caso a densidade dos indivíduos na classe de maior tamanho (geralmente plantados) não diminua naturalmente com o tempo. Para alguns grupos de espécies (zoocóricas, fixadoras de N, espécies raras e ameaçadas), as proporções em termos de riqueza e de densidade não apresentaram trajetórias previsíveis, mas algumas matas ciliares restauradas atingiram e ultrapassam os valores de referência. Excluímos o fator idade para identificar os fatores que influenciaram as trajetórias das variáveis selecionadas como indicadores, calculando um índice de sucesso (proporção entre o dado observado em cada local e o estimado para a mesma variável com base no modelo). Entre as seis variáveis selecionadas como indicadores, uma foi influenciada somente pela distância de fragmento-fonte (riqueza de regenerantes com DAP \'> ou =\' 5) e outra apena pela proporção de partículas finas no solo (área basal). Apesar da proporção do terreno ocupada por gramíneas não influenciar nas trajetórias, o controle da mato-competição com herbicida resultou em maiores valores de riqueza de plantas em regeneração. Independentemente das técnicas de plantio e dos tipos de manejo dos projetos de restauração, a estrutura florestal e os processos funcionais dos ecossistemas restaurados tendem a se tornar mais semelhantes aos ecossistemas de referência com o tempo, desde que haja fontes de propágulos e disponibilidade de água e nutrientes no solo. Os resultados obtidos apontam para uma influência muito pequena ou nula das técnicas de plantio e manejo, se comparados com a influência dos fatores ambientais e da paisagem sobre as trajetórias das matas ciliares em restauração. / A remarkable increase in the number of forest restoration projects has been recorded in Brazil in the last decade. As a consequence of this increase, the demand for monitoring and indicators of restoration success has also increased. Despite the advances in restoration ecology, there are still gaps in knowledge about the functioning and self-sustainability of restored ecosystems. In order to fill some of these gaps, in this study we aimed at to answer the following questions: (1) Are there patterns in the attributes of richness, structure, and functional groups on the native riparian Seasonally Semideciduous Forest (SSF), which may be used as goals to forest restoration in this region? (2) What are the variables representing biodiversity, forest structure and functional groups in the restored areas, which follow predictable paths over time? (3) How long it takes for these variables to achieve the reference values of native riparian forests? (4) Among the biotic, abiotic and technical factors analyzed, which can be considered as drivers of the successional trajectories of the restored forests? We assessed four native forests (as references) and 26 restored forests in the SSF region (a sub-type of the Atlantic Forest) in a chronosequence from four to 53 years since planting. At each site, the sample area comprised ten plots of 100 \'M POT.2\' randomly distributed within a range of 50 m away from the margin. We identified and counted all individuals of tree and shrub species from 50 cm in height, planted or regenerating, categorized in three size classes according to diameter at breast height (DBH): DBH < 1 cm, 1 cm \'< ou =\' DBH < 5 cm DBH \'> ou =\' 5 cm. We measured DBH and estimated the height of each individual from the upper class. We categorized the species according to their dispersion syndrome, growth rate, shade tolerance, Nitrogen-fixing ability, degree of threaten, and rarity. For each plot, we measured the canopy openness, and recorded the number of lianas, pteridophytes and number of trees with epiphytes. We also collected composite soil samples for chemical and physical analyses, from each study site. We estimated the aboveground biomass for trees with DBH \'> ou =\' 5 cm, total richness and richness rarefied to 100 individuals. By interviews, field observations, and landscape studies, we obtained information on other ecological factors and management practices that could affect the paths of the restored plant communities. We modeled the trajectory of each variable of the plant community over time, and estimated the period of time required for the restored forests to reach the reference ecosystems. Amongst the analyzed variables, we selected as indicators for monitoring the evolution and sustainability of the restored ecosystem those that were more predictable (best quality models) and that represent the most relevant ecological processes. The attributes that did not change among the native forests and could be used as references were: observed richness, estimated richness for 100 individuals with DBH \'> ou =\' 5 cm, total richness estimated by Jackknife DBH \'> ou =\' 1 cm, density DBH \'> ou =\' 1 cm, basal area, biomass, canopy cover, proportion of shade-tolerant individuals, and proportion of zoochoric species, slow growing, shade tolerant, nitrogen fixers, and rare distribution. Overall, the trajectories were best fitted to the logarithmic models, but richness (DBH \'> ou =\' 5 cm) was better adjusted to the linear model. In the upper size class, richness did not reach the \"species pool\" of the reference ecosystems in the 53 years period analyzed. The trajectories followed similar patterns, regardless of the planting techniques or the number of planted species. Variables representing forest structure were similar among the native forests and quickly recovered in comparison to those related to biodiversity. Thus, the restoration goals can be established based on native forests parameters and such variables would be good indicators even in young restoration plantings. The variables reaching more quickly the reference values were: basal area (12 years), biomass (13 years), average height of the largest trees (26 years), grass cover (33 years), canopy cover (35 years), richness of regenerants with DBH \'> ou =\' 1 cm (52 years), density of regenerants with DBH 1 cm (52 years), and total richness (53 years). The basal area and biomass greatly exceeded the reference values, which may indicate a restrictive filter for the forest dynamics if the density of individuals in the upper class (most planted) does not decrease naturally over time. For some functional groups (zoochoric, N-fixing, rare, and endangered species), the proportions of species and relative densities did not follow predictable trajectories, but some of the restored riparian areas reached and exceeded the reference values. In order to identify the factors driving the paths of the variables selected as indicators, we excluded the age factor, calculating an index of success (the ratio between observed and estimated values for the variable, the first directly measured for each site and the last obtained by the linear models). Among the six variables selected as indicators, some were influenced only by the distance of the source fragment (seedling richness of DBH \'> ou =\' 5 cm) and others only by the proportion of fine particles in the soil (basal area). While the ground cover by grasses does not influence the trajectories, weed control with herbicides resulted in higher plant richness in regeneration. Regardless of planting techniques and management practices, the forest structure and functional processes of restored ecosystems tend to become more similar to reference ecosystems over time, as long as there are propagule sources in the vicinity and availability of soil water and nutrients. Our findings pointed to a weak, if any, effect from the planting techniques or management and a high influence of environmental and landscape factors on the trajectories of a riparian forest restoration.

Ce premier travail de thèse sur les plantes non vasculaires de La Réunion visait à mesurer et à expliquer la diversité des communautés de bryophytes à différentes échelles de perception. Les résultats ont confirmé que les bryophytes constituent un compartiment majeur de la biodiversité à La Réunion représentant actuellement 753 espèces. Cette étude en trois ans a mis en évidence la présence d'environ 100 taxons nouveaux pour l'île et une espèce nouvelle pour la Science ce qui souligne le besoin d'échantillonnage d'un tel groupe biologique. L'utilisation du site naturel expérimental que sont les coulées de lave du Piton de la Fournaise a permis d'étudier pour la première fois la dynamique de la végétation des bryophytes et de mettre en évidence une forte diversité spécifique le long d'une chronoséquence de six coulées de laves à basse altitude (~ 304 ans) et d'un gradient altitudinal (250 - 850 m) sur la coulée 1986. Cette diversité est principalement représentée par les hépatiques, plus particulièrement la famille des Lejeuneaceae. L'analyse de la distribution spatiale des bryophytes à une échelle fine le long de ces gradients a montré un turn over des microhabitats et de sa bryoflore associée en fonction de l'age des coulées et de l'altitude, et du rôle prédominant de la présence et la disponibilité de ces microhabitats dans la structuration des communautés bryophytiques. Vingt six groupements de bryophytes ont été caractérisés pour ces coulées. Le patron biogéographique des espèces recensées sur ces coulées de laves reflète principalement une origine africaine (66.5 % des espèces) suivi par des espèces à plus large distribution (pantropicale, paléotropicale, subcosmopolite et disjoncte entre l'Amérique et l'Afrique). A plus large échelle, cette étude s'inscrit dans un programme de recherches visant à obtenir une meilleure connaissance de la diversité bryophytique et de l'écologie des espèces pour la zone ouest de l'océan Indien (Réunion, Comores, Seychelles, Maurice, Madagascar) en vue d'une meilleure stratégie de conservation et de gestion du patrimoine naturel de ce hot spot de biodiversité.

Why evaluations of the ecological outcomes of stream and river restoration have largely reported inconclusive or negative results has been the subject of much debate over the last decade or more. Understanding the reasons behind the lack of positive results is important for bettering future restoration efforts and setting realistic expectations for restoration outcomes. This thesis explores possible explanations for why researchers have failed to find clear and predictable biotic responses to stream restoration: recovery time has been too short, that restoration of habitat complexity is not clearly linked to instream biodiversity, that one monitored organism group is not representative of the entire community, that restoration effort was not intense enough to restore the potential habitat complexity of a system, and that reach-scale restoration done in the presence of catchment-scale degradation obscures restoration results. The overarching goal of this thesis is to study the holistic effect of reach-scale restoration of historic reach-scale simplification, due to timber floating in northern Swedish streams, thus avoiding the added pressure of catchment-scale degradation typically found at most restoration sites (e.g., non-point-source pollution and impervious cover). Using this model system, I was able to show that it took 25 years for riparian plant species richness at restored sites to increase above that of channelized sites. Furthermore, it was clear that restoration of these streams caused a large and rapid change in N-processing in the riparian zone and this alteration persists for at least 25 years. Additionally, multiple metrics of geomorphic complexity were needed to explain some of the more subtle responses of organism groups. Macroinvertebrates, diatoms, and macrophytes did not respond concordantly and cannot serve as surrogates or indicators for each other. I found that older best practice methods of restoration rarely restored the large-scale features needed to bring the sites up to their potential complexity because these elements were destroyed or removed from the system. Advanced restoration techniques used in more recent restorations added big boulders and instream wood and increased complexity to a level that elicited a biological response. By combining surveys of multiple metrics of structure, diversity of multiple organism groups, and process in this thesis I was able to get a holistic view of the effects of restoration of streams after timber floating. We now know that it takes at least 25 years for riparian plants and N-cycling to recover, we understand that multiple metrics of geomorphic complexity should be measured to be able to explain biotic responses, and that restored complexity should better match the potential complexity of the site in order to elicit a biological response. Finally, we know that multiple organism groups need to be assessed when evaluating the response of biodiversity to restoration.