Song repertoire structure, organization, and use were studied in 68 male Canada Warblers (Cardellina canadensis) in a breeding population in New Hampshire in 2010-2011. On average, males had complex repertoires of 12 phrases and 55 variants. Repertoire sharing was negatively related to distance between territories, and positively related to longer territory tenure, evidence that males learn songs from neighbours. Males used two singing modes: (I) slow, regular delivery of less variable songs, and (II) fast, intermittent delivery of more variable songs interspersed with chips. Males used Mode I when unpaired and when near females, and Mode II at dawn and during territory disputes, a pattern similar to other warbler species with two song categories. Detectability (whether a male sang) differed little between 1-, 3-, 5-, and 10-min count intervals. Song output and detectability were highest at dawn and in unpaired males, and lowest in paired males late in the season.

During my time as a PhD student at the Physics of Medical Imaging group at KTH, I have taken part in the work of developing a photon counting spectrally resolved silicon detector for clinical computed tomography. This work has largely motivated the direction of my research, and is the main reason for my focus on certain issues. Early in the work, a need to quantify and optimize the performance of a spectrally resolved detector was identified. A large part of my work have thus consisted of reviewing conventional methods used for performance quantification and optimization in computed tomography, and identifying which are best suited for the characterization of a spectrally resolved system. In addition, my work has included comparisons of conventional systems with the detector we are developing. The collected result after a little more than four years of work are four publications and three conference papers. This compilation thesis consists of five introductory chapters and my four publications. The introductory chapters are not self-contained in the sense that the theory and results from all my published work are included. Rather, they are written with the purpose of being a context in which the papers should be read. The first two chapters treat the general purpose of the introductory chapters, and the theory of computed tomography including the distinction between conventional, non-spectral, computed tomography, and different practical implementations of spectral computed tomography. The second chapter consists of a review of the conventional methods developed for quantification and optimization of image quality in terms of detectability and signal-to-noise ratio, part of which are included in my published work. In addition, the theory on which the method of material basis decomposition is based on is presented, together with a condensed version of the results from my work on the comparison of two systems with fundamentally different practical solutions for material quantification. In the fourth chapter, previously unpublished measurements on the photon counting spectrally resolved detector we are developing are presented, and compared to Monte Carlo simulations. In the fifth and final chapter, a summary of the appended publications is included. / <p>QC 20150303</p>

Submitted by Lara Oliveira (lara@ufersa.edu.br) on 2017-03-28T20:22:19Z No. of bitstreams: 1 ClarisseCOS_DISSERT.pdf: 4409173 bytes, checksum: ddfe0c2c141a91a0fc8fb6c388923cb2 (MD5) / Approved for entry into archive by Vanessa Christiane (referencia@ufersa.edu.br) on 2017-04-13T14:41:28Z (GMT) No. of bitstreams: 1 ClarisseCOS_DISSERT.pdf: 4409173 bytes, checksum: ddfe0c2c141a91a0fc8fb6c388923cb2 (MD5) / Approved for entry into archive by Vanessa Christiane (referencia@ufersa.edu.br) on 2017-04-13T15:07:13Z (GMT) No. of bitstreams: 1 ClarisseCOS_DISSERT.pdf: 4409173 bytes, checksum: ddfe0c2c141a91a0fc8fb6c388923cb2 (MD5) / Made available in DSpace on 2017-04-13T15:08:37Z (GMT). No. of bitstreams: 1 ClarisseCOS_DISSERT.pdf: 4409173 bytes, checksum: ddfe0c2c141a91a0fc8fb6c388923cb2 (MD5) Previous issue date: 2016-03-31 / Coordenação de Aperfeiçoamento de Pessoal de Nível Superior / The semi-arid tropical environments exhibit high seasonality, including wide variation in the rainfall amount throughout the year. This triggers fluctuations in environmental conditions and the resources availability and therefore should influence the variation in species richness of birds in time and space. In this study we used capturerecapture data from birds to evaluate issues such as: (1) species richness vary seasonally with a high change rate, as in other semi-arid; (2) species richness variation is mainly due to the temporary departure of species in the wet-dry transition and return on dry-wet cycle; and (3) the environment with a lower degree of disturbance is associated with less seasonal variation in species composition and richness than higher disturbed environment. The study was carried out in the Caatinga (semi-arid Neotropical) over three years (2012-2015), with two annual capture season (dry and wet) in two environments (natural and disturbed). We analyzed the data through MARK program, with Robust design models and CJS to estimate parameters of the communities’ dynamics (estimates of species input and output (permanent and temporary), species richness and richness change rate). The analysis showed that seasonality and environment are factors decisive for the temporary emigration, the largest output of species in the wet-dry transition and greater input in the dry-wet. We estimate that approximately 40% of the species left the sampling area after the rainy season and 85% of the outside species entered the area before the rainy season. The average of species richness estimated was 33.4 in the dry season and 53.2 in the rainy season. The change rate of species richness varied from 0.6 in the wet-dry transition to 1.5 in the dry-wet, being similar between environments. The estimates of permanent output of species varied from one to three species per transition, while permanent input varied between four and six species per season. Our results showed that the seasonal variation of water regime, as well as the phytophysiognomy change were important factors in determining the community dynamics / Os ambientes semiáridos tropicais apresentam alta sazonalidade climática, incluindo grande variação pluviométrica ao longo do ano. Isto desencadeia flutuações nas condições ambientais e na disponibilidade de recursos e, consequentemente, deve influenciar a variação na riqueza de espécies de aves no espaço e tempo. Neste estudo utilizamos dados de captura-recaptura de aves para avaliar questões como: (1) a riqueza de espécies varia sazonalmente e a taxa de mudança é alta, assim como em outros semiáridos; (2) a variação na riqueza é consequência principalmente da saída temporária de espécies na transição chuva-seca e retorno na seca-chuva; e (3) o ambiente com menor grau de perturbação está associado à menor variação sazonal em riqueza e composição de espécies do que o mais perturbado. O estudo ocorreu na Caatinga (semiárido neotropical), ao longo de três anos (09/2012 a 08/2015), com dois períodos anuais de captura (seca e chuva) em dois ambientes (natural e perturbado). Analisamos os dados no programa Mark, com modelos do tipo Desenho Robusto e CJS visando estimar parâmetros da dinâmica de comunidades (estimativas de entrada e saída espécies (permanente e temporária), riqueza de espécies e taxa de mudança na riqueza). As análises mostraram que a sazonalidade e o tipo de ambiente são fatores determinantes para a emigração temporária, sendo a maior saída de espécies na transição chuva-seca e maior entrada na seca-chuva. Estimamos que aproximadamente 40% das espécies saíram da área de amostragem após o período chuvoso e 85% das espécies que estavam fora, entraram na área antes do período chuvoso. A riqueza de espécies estimada foi em média de 33,4 na época seca e 53,2 na época de chuva. A taxa de mudança na riqueza de espécies variou entre 0,6 na transição chuva-seca e 1,5 na seca-chuva, sendo semelhantes entre ambientes. As estimativas de saída permanente de espécies variou de uma a três espécies por transição, enquanto que de entradas permanentes variou entre quatro e seis espécies por período. Os nossos resultados mostraram que a variação sazonal do regime hídrico, assim como, a variação na fitofisionomia foram fatores importantes na determinação da dinâmica da comunidade / 2017-03-28

The main objectives of the thesis are to learn about the detection of a hypothetical marine cloud brightening (MCB) field experiment and to assess the benefits and risks of an actual implementation using model simulations. The first aim of the thesis is to assess the detectability of MCB from the natural variability of clouds using satellite data. The analysis uses two approaches, i) an analytical method, assessing the radiative forcing by aerosol-cloud interactions for an idealised perturbation of the droplet concentration and ii) a stochastic method, including the radiative forcing and the cloud adjustment effects of perturbing the cloud droplet number concentration. One of the main findings of the study is that in the analytical method, detection of an MCB experiment in the North Pacific region requires longer duration, larger domain and larger intensity of the Nd perturbation than in the South Pacific and the South Atlantic regions. In the Stochastic method, larger domain size as well as longer duration are required for the detection of the hypothetical field experiments compared to the analytical method. We found that increasing the duration of the experiment and intensity of Nd perturbation has more influence in detectability than increasing the domain size. Secondly, we analyse the detection of MCB using ECHAM GCM simulations. Two methods are used for the analysis, i) a temporal method, using temporal reference of nine years for the analysis, and ii) a spatial method, where a surrounding unperturbed regions are used as a reference for the analysis. In the temporal method, the detectability of the MCB experiment over the North Pacific region is most difficult. Our calculations suggest that for an experiment to be cost effective, it is preferable to design an experiment of longer duration with smaller domain sizes as well as less intensity of perturbation of Nd. Finally, we investigate the effectiveness of MCB in alleviating the changes in the mean and extremes for surface air temperature and precipitation. We compare these results with another climate engineering technique, namely stratospheric aerosol injection. The study is based on an existing multi model simulation of three Earth system models. The main conclusion from the study is that the two solar radiation management techniques are rather effective in mitigating the climate change driven increases in the mean and extremes of temperature and precipitation according to the climate simulations. However, the potential to mitigate differs around the globe and seasonally. The strong increases in lower temperature extremes, especially in the Arctic, are not well dampened with the solar radiation management implementation.

京都大学 / 新制・課程博士 / 博士(理学) / 甲第24423号 / 理博第4922号 / 新制||理||1703(附属図書館) / 京都大学大学院理学研究科地球惑星科学専攻 / (主査)准教授 伊藤 喜宏, 准教授 西村 卓也, 教授 王 功輝 / 学位規則第4条第1項該当 / Doctor of Science / Kyoto University / DGAM

This thesis contains two parts: the detectability of convex sets and the study on regression models In the first part of this dissertation, we investigate the problem of the detectability of an inhomogeneous convex region in a Gaussian random field. The first proposed detection method relies on checking a constructed statistic on each convex set within an nn image, which is proven to be un-applicable. We then consider using h(v)-parallelograms as the surrogate, which leads to a multiscale strategy. We prove that 2/9 is the minimum proportion of the maximally embedded h(v)-parallelogram in a convex set. Such a constant indicates the effectiveness of the above mentioned multiscale detection method. In the second part, we study the robustness, the optimality, and the computing for regression models. Firstly, for robustness, M-estimators in a regression model where the residuals are of unknown but stochastically bounded distribution are analyzed. An asymptotic minimax M-estimator (RSBN) is derived. Simulations demonstrate the robustness and advantages. Secondly, for optimality, the analysis on the least angle regressions inspired us to consider the conditions under which a vector is the solution of two optimization problems. For these two problems, one can be solved by certain stepwise algorithms, the other is the objective function in many existing subset selection criteria (including Cp, AIC, BIC, MDL, RIC, etc). The latter is proven to be NP-hard. Several conditions are derived. They tell us when a vector is the common optimizer. At last, extending the above idea about finding conditions into exhaustive subset selection in regression, we improve the widely used leaps-and-bounds algorithm (Furnival and Wilson). The proposed method further reduces the number of subsets needed to be considered in the exhaustive subset search by considering not only the residuals, but also the model matrix, and the current coefficients.

Les insectes sont encore aujourd’hui peu pris en compte dans les programmes de conservation de la biodiversité, malgré leur rôle majeur dans le fonctionnement des écosystèmes terrestres. Du fait de leur grande diversité, et d’une grande méconnaissance des patrons de structuration des communautés, la conservation de ces organismes se heurte à des problématiques liées à l’évaluation et à la hiérarchisation des enjeux ainsi qu’à l’estimation de la richesse spécifique des communautés. Parce qu’ils constituent deux groupes d’insectes particulièrement impactés par les changements environnementaux, les Rhopalocères et les Odonates constituent une part importante des insectes protégés en Europe. Dans ce contexte de faible prise en compte de la diversité entomologique en biologie de la conservation, cette thèse a eu pour objectifs (1) de mesurer les dynamiques récentes (déclin, stabilité ou essor) de ces deux taxa en France, et d’identifier les facteurs climatiques et écologiques susceptibles de les conditionner, (2) pour les Rhopalocères, d’utiliser une description fine des réseaux d’interaction entre les larves et leurs plantes-nourricières pour questionner les statuts de conservation selon une approche systémique (3) et enfin, dans un objectif résolument opérationnel, d’évaluer les liens entre les gains d’informations et les efforts/coûts d’échantillonnage mobilisés au cours des études d’impacts.L’étude comparative des dynamiques temporelles des populations de ces deux taxons a été réalisée à l’échelle de trois pays d’Europe de l’Ouest grâce à une analyse diachronique des patrons de répartition des espèces sur un pas de temps de 35 ans. Ce travail a mis en évidence une forte corrélation entre l’artificialisation des paysages (urbanisation, intensification agricole, régression des zones humides) et le déclin de nombreuses espèces, majoritairement caractérisées par des exigences écologiques strictes et actuellement non protégées. A cette même échelle, l’analyse de l’architecture du réseau d’interactions entre les larves de Rhopalocères et leurs plantes nourricières a quant à elle révélé une structure modulaire en relation avec la taxonomie des partenaires, ainsi qu’un lien entre le degré de spécialisation et la vulnérabilité des espèces (sans que la similarité du régime alimentaire des espèces ne semble conditionner leur statut de conservation). Une dernière partie a été ciblée sur l’évaluation du degré d’exhaustivité des inventaires entomologiques réalisés lors d’études d’impact réalisées en amont d’aménagements. Une procédure de repasse systématique et calibrée, sur des sites échantillonnés en garrigue méditerranéenne, a permis de mettre en évidence les limites des protocoles actuels, en termes d’estimation de la richesse spécifique et de détection d’espèces à enjeux. Sur cette base, des recommandations sont formulées pour optimiser les protocoles d’échantillonnage à mettre en œuvre pour la détection des espèces à enjeux.Cette thèse contribue à l’articulation entre problématiques fondamentale et besoins opérationnels, en permettant à la fois une meilleure compréhension des mécanismes de structuration des communautés d’insectes et la formulation de recommandations pour une meilleure application locale des politiques de conservation. / Insects are still poorly considered in biodiversity conservation programs, despite their major role in the functioning of terrestrial ecosystems. Because of their great diversity, and a great lack of knowledge of community structuring patterns, the conservation of these organisms faces major challenges related to the evaluation and prioritization of issues as well as the estimation of the species richness of communities. Because they constitute two groups of insects particularly impacted by environmental changes, butterflies and dragonflies are an important part of protected insects in Europe. In this context of low consideration of entomological diversity in conservation biology, this thesis aims to (1) measure the recent dynamics (decline, stability or expansion) of these two taxa in France, and to identify the climatic and ecological factors likely to condition them, (2) for butterflies, to use a fine description of the interactions network between the larvae and their hostplants to question the conservation status according to a systemic approach (3) and finally, with a clearly operational objective, to evaluate the links between the information gains and the efforts / sampling costs mobilized during environmental impact assessments.The comparative study of the temporal dynamics of the populations of these two taxa was carried out at the scale of three countries of Western Europe through a diachronic analysis of the species distribution patterns on a time step of 35 years. This work has highlighted a strong correlation between the artificialisation of landscapes (urbanization, agricultural intensification, regression of wetlands) and the decline of many species, mainly characterized by strict ecological requirements and currently unprotected. On this same scale, the analysis of the architecture of the interactions network between butterfly larvae and their food plants revealed a modular structure in relation to the taxonomy of the partners, as well as a link between the degree specialization and species vulnerability (although the similarity of the species' diet did not seem to condition their conservation status). A final part was focused on the evaluation of the completeness of entomological surveys carried out during environmental impact assessments. A systematic and calibrated return procedure, on sites sampled in Mediterranean scrublands, allowed highlighting the limits of the current protocols in the estimation of species richness and the detection of protected or vulnerable species.This thesis contributes to the articulation between fundamental issues and operational needs, by allowing both a better understanding of the structuring mechanisms of insect communities and the formulation of recommendations for a better local application of conservation policies.

Orientador: André Victor Lucci Freitas / Tese (doutorado) - Universidade Estadual de Campinas, Instituto de Biologia / Made available in DSpace on 2018-08-18T19:14:01Z (GMT). No. of bitstreams: 1 Ribeiro_DaniloBandini_D.pdf: 21049242 bytes, checksum: aa5d5c4cafb6f3c74669788c029701d9 (MD5) Previous issue date: 2011 / Resumo: A Amazônia representa mais da metade da área de todas as florestas tropicais do mundo, e vem sendo ameaçada por diversas atividades antrópicas, incluindo vários tipos de extração de madeira. O corte seletivo de impacto reduzido (em inglês "Reduced Impact Logging - RIL") é considerado um método menos destrutivo de exploração madeireira que promove uma menor alteração na estrutura da floresta comparado com métodos convencionais. No entanto, existem poucas informações sobre os efeitos do RIL em invertebrados, incluindo borboletas. Desta forma, investigamos o efeito do RIL na estrutura da vegetação florestal e sobre as borboletas frugívoras, comparando a fauna do dossel e do sub-bosque entre uma área explorada (RIL) e uma área não explorada (controle). Devido à pouca informação disponível sobre protocolos de amostragem para as borboletas tropicais, alguns aspectos metodológicos relativos à amostragem dos ninfalídeos frugívoros foram investigados. Analisamos como a detectabilidade varia entre diferentes espécies, habitats (Amazônia x Mata Atlântica), estratos da floresta e também qual o esforço adequado de amostragem para detectar uma certa quantidade espécies em uma determinada área, no prente estudo esta quantidade foi estabelecida como 25% do total de espécies estimado no local. Implicações biológicas Uma floresta não explorada apresenta árvores adultas e juvenis maiores do que as de uma floresta explorada (RIL). O número de plântulas e mudas é maior na floresta explorada (RIL) e a distribuição de freqüências de tamanhos (SDFslope) não foi diferente entre as áreas de RIL (-2,61) e não explorada (-2,31). A abertura do dossel foi maior na floresta não explorada, provavelmente devido a um aumento das plantas do sub-bosque na floresta RIL. A área basal e altura das árvores foram maiores na floresta não explorada. Em relação às borboletas frugívoras, a fauna do dossel é diferente e significativamente mais rica do que a fauna do sub-bosque, mostrando que amostrar apenas o estrato mais baixo pode subestimar a diversidade de borboletas. Os efeitos do RIL foram detectados principalmente na assembleia de borboletas do sub-bosque, onde foram observadas diferenças significativas na composição de espécies entre as áreas. Os impactos do RIL, que incluem o corte de árvores, a abertura de trilhas de arraste e de estradas, são mais intensos no sub-bosque do que no dossel. Estas diferentes intensidades de impacto no dossel e no sub-bosque podem explicar os diferentes efeitos do RIL nas borboletas presente nesses estratos. Considerando o grande potencial das borboletas frugívoras para prever as respostas de vários outros grupos taxonômicos à perturbação da floresta na Amazônia, esperamos que padrões semelhantes sejam encontrados em outros táxons. No entanto, apesar dos efeitos detectáveis do RIL em borboletas frugívoras, não pudemos perceber espécies que foram excluídas ou que invadiram a área de RIL, os nossos resultados mostram principalmente diferenças nas abundâncias das espécies entre as áreas. Este resultado é similar ao encontrado em estudos feitos com outros taxa mostra que o RIL em geral não altera a diversidade de espécies podendo ser uma alternativa para preservar uma parcela significativa da fauna em áreas com este tipo de exploração. A criação de áreas protegidas na Amazônia apesar de desejável é bastante complicada e nem sempre é efetiva na preservação da cobertura vegetal, assim sendo o corte seletivo de impacto reduzido pode ser considerado uma boa alternativa para preservar borboletas frugívoras na floresta amazônica e, certamente, muitos outros táxons. Além disso, devido ao alto valor da madeira produzida neste sistema, esta poderia ser uma alternativa econômica desejável para a região. Implicações metodológicas Quase todas as borboletas e mariposas amostrados no presente estudo foram mais facilmente amostradas em um estrato específico (dossel ou sub-bosque). No presente estudo, mesmo as espécies mais comuns do dossel raramente foram amostradas no subbosque. Assim, fazer uso de um protocolo de amostragem que não utiliza armadilhas em ambos os estratos irá aumentar os erros de detecção de muitas borboletas e pode conduzir a inferências incorretas sobre a riqueza e diversidade em uma determinada área. As diferenças na detectabilidade entre os meses no conjunto de dados da Amazônia mostraram que mesmo quando as borboletas são amostradas durante o período do ano com maior probabilidade de captura, existem diferenças importantes na detectabilidade entre os meses. A baixa detectabilidade e a grande variação entre os estratos e meses nos levam a propor que, para borboletas frugívoras, a amostragem deve ser feita na época correta e em florestas altas os diferentes estratos devem ser considerados de modo a reduzir os erros de detecção e possíveis vieses nos resultados. O esforço amostral mínimo para a detecção de 25% das espécies presentes nas florestas tropicais é de 130 armadilhas / dia para a Mata Atlântica e 510 dias para Amazônia Central. Além disso, a amostragem deve ser feita com réplicas temporais em um curto espaço de tempo para aumentar o poder de interpretação dos dados coletados / Abstract: The Amazon region represents more than half of the area of all tropical forests in the world, and has been threatened by many anthropogenic activities, including several kinds of timber harvesting. Reduced-Impact Logging (RIL) is considered a less destructive method of timber harvesting that promotes a smaller change in forest structure than conventional logging. However, there is a general lack of information about the effects of RIL on Amazonian invertebrates, including butterflies. We therefore investigated the effect of RIL on forest vegetation structure and on fruit-feeding butterflies by comparing their distribution on canopy and understory between an area under RIL and a control area without RIL. Because of the relative lack of information about sampling protocols for tropical butterflies, some methodological aspects of fruit-feeding Nymphalidae sampling were investigated. We analyzed the variation in detectability among species, habitats (Amazon x Atlantic Forest), layers and the adequate sampling effort need for detect an specific amount of species in a given area were analyzed in this thesis, in the present study this amount was established as 25% of the total estimated species richness. Biological implications An unlogged forest has bigger juveniles and adult trees, and less seedlings and saplings than a RIL forest, and the Size Frequency Distribution (SDF) slope was not different from those of logged (-2.61) and unlogged (-2.31) areas. The canopy openness was greater in the unlogged forest, probably due to an increase of understory plants in the RIL forest. The basal area was wider and the height was taller in unlogged forest trees. In relation to the fruit-feeding butterflies, the canopy fauna is different and significantly richer than the understory fauna, showing that sampling only the lower strata underestimates the diversity of fruit-feeding butterflies. The effects of RIL were mainly detected in the understory butterfly assemblage, as significant differences were observed in species composition within this stratum. Effects of the RIL regime, which include tree cutting, skid trails and road openings, are stronger in the understory than in the canopy, explaining the reported differences. Despite the detectable effects of RIL on the composition of fruit-feeding butterfly's assemblages the overall diversity was not affected, this pattern is very similar for many other taxa indicating that a noticeable part of the diversity of many taxa could be preserved in areas under RIL management. Given the problems of creating protected areas in the Amazon, RIL is a good alternatives to preserve fruit-feeding butterflies and surely many other taxa, and it might be a desirable economic alternative for the region. Methodological implications Almost all butterflies and moths sampled in the present study were more readily trapped in one specific stratum. Indeed, in the present study, even the most common canopy species were rarely sampled in the understory. Thus, using a sampling protocol that does not locate traps in both layers will increase the imperfect detection of many butterflies and could lead to incorrect inferences about the richness and diversity in a given area. The differences in detectability between months in the Amazon dataset showed that even with an experimental design planned for sampling butterflies during the period that enhances capture probability, there are important differences in butterfly detectability across months. The low detectability and great variation among strata and months in fruit-feeding butterflies lead us to assume that sampling designs must address sampling effort to the correct season and strata reducing imperfect detections and biases in the results. The minimal sampling effort for detecting 25% of the species present in tropical forests is 130 trap/days in Atlantic Forest and 510 days in Central Amazon. Additionally, such sampling should use temporal replication over a short period to improve the interpretability of the data collected / Doutorado / Doutor em Ecologia

Les souches de Legionella pneumophila sérogroupe 1 ST47 Lorraine sont actuellement responsables d'un nombre croissant de cas de légionelloses particulièrement dans le Nord de l'Europe. Cependant, elles ne sont retrouvées que très rarement dans l'environnement quand des investigations sont menées pour rechercher la source de cas de légionellose. Son environnement naturel reste donc méconnu. Notre objectif était donc d'étudier le phénomène de non détectabilité de la souche ST47 dans l'environnement. Nous avons observé que les traitements pré-analytiques (acidification, chauffage) pouvaient interférer avec les paramètres de culture pour certaines souches Lorraine environnementales, et que le milieu GVPC utilisé dans la norme NF T90-431 n'était pas recommandé pour la recherche des légionelles dans les prélèvements respiratoires, et donc potentiellement dans les prélèvements environnementaux. Ainsi, la PCR spécifique ST47, développée en collaboration avec l'équipe de Tim Harrison (Londres, Royaume-Uni), peut constituer un outil de détection intéressant dans les différentes matrices environnementales pour identifier les réservoirs de la souche. Nous avons cherché à cibler ces réservoirs. Nous avons montré, d'une part, que les robinets pouvaient être des réservoirs potentiels de légionelles. D'autre part, les données clinicoépidémiologiques et géographiques des souches Lorraine ST47 ont montré que les cas survenaient essentiellement hors institution (hôpital, maison de retraite) et principalement dans le quart Nord-Est du pays. L'amélioration des connaissances clinico-épidémiologiques et l'utilisation de la PCR spécifique ST47 devraient permettre, en s'affranchissant ou en améliorant les paramètres culturaux, d'identifier les réservoirs environnementaux de cette souche / The recently identified strains of Legionella pneumophila serogroup 1 belonging to Sequence Type (ST) 47 and pulsotype Lorraine are now responsible for an increasing number of Legionnaires’ disease (LD) cases mainly in the North of Europe. The major paradox is that ST47 strains are extremely rarely detected in environmental samples when investigations to identify the source of ST47 associated LD cases are undertaken. Thus its environmental habitat is unknown. The aim of our work was to study the non-detectability of this strain in environmental samples. Firstly, we have observed that pre-treatments (heat, acid) could interfere with recovery of an environmental ST47 Lorraine strain, and secondly, that GVPC medium imposed by the French standard NF T90-431, was not recommended for detection of Legionella spp. In clinical samples and potentially in environmental samples. Thus, to detect and identify ST47 Lorraine strains from clinical and ultimately, environmental samples in order to find the potential source of infection, we developed a strain specific realtime PCR method, in collaboration with Tim Harrison’s team (London, United Kingdom). We have also sought to target these sources. Then, we have shown that L. pneumophila could be detected in aerosols of a tap water. Moreover, epidemiologic French data have demonstrated that most of the LD cases associated with ST 47 strains occurred mainly in North East and were mainly community-acquired. Improved epidemiologic knowledge and the use of the strain specific real-time PCR should enable identification of environmental sources of ST47 Lorraine strain without culturing

Les systèmes avioniques s'étoffent et se complexifient de plus en plus. Avec l'augmentation des capacités de calcul, de nouvelles architectures basées sur le partage de ressources émergent. Effectuer le diagnostic d'un système n'est désormais plus une opération anodine. L'enjeu actuel est donc de mettre en place des techniques de diagnostic performantes tout en optimisant les capacités de monitoring nécessaires.Ce mémoire donne une caractérisation basée modèle d'un système sous diagnostic, puis propose des techniques pour en évaluer les performances de diagnostic, ainsi que celles de son monitoring (relativement à ces performances). Le contexte industriel dans lequel s'inscrit cette thèse amène d'autres contraintes, notamment la prise en compte de la taille des systèmes avioniques à analyser. Cette thèse étudie alors l'applicabilité des techniques introduites dans ce contexte et en propose une adaptation. / Avionics systems become more and more complex. With the improvment of computing possibilities, new architectures based on resources sharing are growing up. Perform diagnosis of a system is no longer a trivial operation. The challenge is to develop efficient techniques of diagnosis while optimizing capabilities of monitoring required.This thesis give a model-based characterization of a system under diagnosis, and proposes techniques to assess diagnostic performances, as well as its monitoring ones (with respect to these diagnostic performances). The industrial context of this thesis brings other constraints, and in particular the need to handle the size of avionics systems to analyze. That thesis then examines the applicability of the introduced techniques to this particular context, and proposes an adaptation.