Vision research has uncovered computational principles that generalize across species and brain area. However, these biological mechanisms are not frequently implemented in computer vision algorithms. In this thesis, models suitable for application in computer vision were developed to address the benefits of two biologically-inspired computational principles: multi-scale sampling and active, space-variant, vision. The first model investigated the role of multi-scale sampling in motion integration. It is known that receptive fields of different spatial and temporal scales exist in the visual cortex; however, models addressing how this basic principle is exploited by species are sparse and do not adequately explain the data. The developed model showed that the solution to a classical problem in motion integration, the aperture problem, can be reframed as an emergent property of multi-scale sampling facilitated by fast, parallel, bi-directional connections at different spatial resolutions. Humans and most other mammals actively move their eyes to sample a scene (active vision); moreover, the resolution of detail in this sampling process is not uniform across spatial locations (space-variant). It is known that these eye-movements are not simply guided by image saliency, but are also influenced by factors such as spatial attention, scene layout, and task-relevance. However, it is seldom questioned how previous eye movements shape how one learns and recognizes an object in a continuously-learning system. To explore this question, a model (CogEye) was developed that integrates active, space-variant sampling with eye-movement selection (the where visual stream), and object recognition (the what visual stream). The model hypothesizes that a signal from the recognition system helps the where stream select fixation locations that best disambiguate object identity between competing alternatives. The third study used eye-tracking coupled with an object disambiguation psychophysics experiment to validate the second model, CogEye. While humans outperformed the model in recognition accuracy, when the model used information from the recognition pathway to help select future fixations, it was more similar to human eye movement patterns than when the model relied on image saliency alone. Taken together these results show that computational principles in the mammalian visual system can be used to improve computer vision models.

North American populations of grassland birds have been declining consistently for the past several decades. Grassland birds respond to multiple scales which encompass a spectrum of habitat variables, and the habitat scale of importance may depend on the response of interest. For practitioners, having knowledge of the effect of scale is useful because conservation efforts can be focused at the most appropriate scale. However, previous multi-scale studies of grassland birds and other taxa have rarely incorporated on-the-ground habitat management while simultaneously investigating site-specific species turnover dynamics (Chapter 1) and daily nest survival (Chapter 2). Also, habitat management-related studies often suffer when not accounting for inherent variation between fields, field landscapes, and study year; thus, to disentangle specific effects of management, it is important to account for this variation by using these factors as random effects within mixed-effects models (Chapter 3). Understanding multi-scale habitat relationships affecting site turnover, also known as dynamic occupancy, and daily nest survival rate (nest survival or DSR), as well as how multiple avian responses vary with grassland management would further benefit conservation decision making for focal species. I investigated dynamic occupancy and nest survival of both obligate and facultative grassland species relative to multi-scale habitat factors on private lands (Conservation Reserve Program) in northwest Illinois during 2011-2014. I also conducted a separate analysis focusing on how multiple avian responses, ranging from species presence/absence (P/A) to species-specific nest survival, are influenced by non-fire grassland management treatments. For dynamic occupancy, a combination of ≥2 scales always outperformed single-scale models for all species. Three of 7 species responded to either cumulative habitat management (proportion of field managed over the study period) or yearly management (33% of field managed before a particular breeding season) regardless of dominant grass type. Of the 48 covariates appearing in top models across species for both dynamic and single-season occupancy, microhabitat covariates (42%) were represented most often, followed by patch-scale (33%) and landscape-scale (25%) factors. Covariates with the most consistent effects and the greatest frequencies appearing in competitive (∆AIC ≤ 4) dynamic and single-season occupancy models included landscape forest cover (n = 10), surrounding patch grassland cover (n = 7), and field size (n = 6). In general, increasing levels of forest cover adjoining fields had consistently negative effects on occupancy, colonization, and persistence across species, while grassland surrounding fields and field size had positive effects on these responses. Microhabitat covariates better explained colonization and extinction across the focal species. Of the 22 habitat covariates in top colonization and extinction models, microhabitat variables were represented 50% of the time, compared to 32% and 18% representation for patch-scale and landscape-scale covariates, respectively. I recommend that wildlife biologists utilize habitat management techniques to ensure grassland birds have different successional stages within their range of preferences. My results suggest that grassland disturbance, while influencing turnover dynamics of different species, may be less influential within smaller fields and landscapes dominated by forest cover. I analyzed the daily nest survival rate for 2 species (red-winged blackbirds and dickcissels) and 2 groups of nesting communities (ground and above-ground nesters). Temporal variables such as Julian date and stage of nest were significant predictors of nest survival for red-wings (50.58 ∆AICc = best temporal model) and dickcissels (2.28 ∆AICc = best temporal model) in addition to habitat covariates. In 3 of 4 analyses ≥2 habitat scales were better predictors of nest survival over one scale. For ground nesters the patch scale was be best predictor of nest survival. However, the best model for ground nesters did not have overwhelming support compared to the random model (1.43 ∆AICc). The ground-nesting community appeared to suffer decreased nest survival with increasing proportion of surrounding grassland. The blackbird top model included nest- and landscape-scale covariates, and top models for dickcissels and the above-ground nesting community included the nest- and patch-scale covariates. Blackbirds had a significant increase in nest survival when nests were placed in areas with higher vegetation density and height (greater visual obstruction). Comparatively, the predictive ability of habitat covariates was not as strong for the other 3 analyses; however, notable patterns include dickcissel nest survival decreasing with increasing nest distance to edge and above-ground nests had increased nest survival with increasing patch perimeter-area-ratio. My results suggest different species are responding to different scales, but finer-scale habitat covariates generally help predict nest survival over course-scale habitat features, such as landscape covariates. Songbird nest survival maybe more influenced by fine-scale habitat characteristics such as nesting cover and field vegetation complexity and density, which can deter nest predators and reduce their search efficiency. In a way, this is positive news for practitioners working in grasslands patches located in less than ideal patch or landscape configurations, suggesting more effort could be focused on improving conditions for colonization and persistence of focal species (Chapter 1). By focusing management on dynamic occupancy responses, increasing potential nesting habitat and territory quality for focal species will likely follow. When investigating multiple avian responses to management, within a consistent mixed-effect modeling framework, it appeared that avian survey related responses were best supported, having consistently larger ∆AICc values for top models, when compared to nesting-related data. One explanation of this relative difference could be attributed sample size difference between analyses. Focusing on the effects of management, this analysis accounted for inherent variation across fields, year, and potentially field landscape, as random effects within all models. As suggested from Chapter 2, and despite a large sample of nests, explanation of DSR across all species and focal species showed little improvement with management covariates. Brome fields cumulatively managed with spray or spray/seed treatments (Chapter 1) appeared to respond most and likely explained general positive effects for red-winged blackbird and dickcissel abundance; however, this was at a cost to species abundance for bobolink (Dolichonyx oryzivorus), Henslow’s sparrow (Ammodramus henslowii), eastern meadowlark (Sturnella magna), field sparrow (Spizella pusilla), and northern bobwhites (Colinus virginianus), due to the release of tall forbs within the seed bank after spraying brome dominated fields. Dramatic vegetation responses in sprayed brome fields likely led to an overall decrease in obligate-grassland bird richness and conservation value on brome field types that were cumulatively management over the course of the study. These results suggest the importance of non-native brome CRP fields for certain grassland-obligate bird species, while also revealing the apparent decoupling of nest survival and other avian responses measured commonly. I suggest managers and researchers pay closer attention to variables affecting nesting abundance/density in grassland systems when possible and recognize the potential importance of non-native grasses for grassland-obligate birds in some regions.

Les fibres végétales utilisées comme renfort dans les matériaux composites présentent des propriétés mécaniques spécifiques concurrentielles par rapport à celles des fibres de verre. De plus, elles ont l'avantage d'être renouvelables et recyclables. Toutefois, leur principal inconvénient est leur sensibilité à l'humidité, ce qui a pour conséquence d'induire une baisse des propriétés mécaniques ainsi d'une décohésion de l'interface fibre/matrice. L'objectif principal de cette thèse est d'étudier l'influence de l'humidité sur le comportement hygro-mécanique de fibres de lin. La première partie de mes travaux a été consacrée à la caractérisation des propriétés hygroscopiques et mécaniques de la fibre et du composite. Dans la deuxième partie, deux modèles multi-échelles, l'un analytique et l'autre numérique, ont été développés pour l'estimation des propriétés hygro-mécaniques des fibres élémentaires de lin. Ils utilisent en partie pour données d'entrée, les propriétés identifiées dans la première partie. / Natural fibres used as reinforcement in composite materials present specific mechanical properties, which are comparable to glass fibres. In addition, they have the advantage of being renewable and recyclable. But, their main drawback is their inherent susceptibility to moisture expansion, which has the effect of inducing a decrease in mechanical properties, and of debonding and fracturing interface in the composite. The main aim of this thesis is to study the influence of humidity on hygro-mechanical behavior of flax fibres. The first part of my work was deal with the characterization of mechanical and hygroscopic properties of the fibre and the composite. The second part is dedicated to the development of two multiscale models, one analytical and one numerical. They have been developed for the estimation of hygro-mechanical properties of elementary flax fibres. Properties identified in the first part of the work are used as input data.

abstract: Commercially pure (CP) and extra low interstitial (ELI) grade Ti-alloys present excellent corrosion resistance, lightweight, and formability making them attractive materials for expanded use in transportation and medical applications. However, the strength and toughness of CP titanium are affected by relatively small variations in their impurity/solute content (IC), e.g., O, Al, and V. This increase in strength is due to the fact that the solute either increases the critical stress required for the prismatic slip systems ({10-10}<1-210>) or activates another slip system ((0001)<11-20>, {10-11}<11-20>). In particular, solute additions such as O can effectively strengthen the alloy but with an attendant loss in ductility by changing the behavior from wavy (cross slip) to planar nature. In order to understand the underlying behavior of strengthening by solutes, it is important to understand the atomic scale mechanism. This dissertation aims to address this knowledge gap through a synergistic combination of density functional theory (DFT) and molecular dynamics. Further, due to the long-range strain fields of the dislocations and the periodicity of the DFT simulation cells, it is difficult to apply ab initio simulations to study the dislocation core structure. To alleviate this issue we developed a multiscale quantum mechanics/molecular mechanics approach (QM/MM) to study the dislocation core. We use the developed QM/MM method to study the pipe diffusion along a prismatic edge dislocation core. Complementary to the atomistic simulations, the Semi-discrete Variational Peierls-Nabarro model (SVPN) was also used to analyze the dislocation core structure and mobility. The chemical interaction between the solute/impurity and the dislocation core is captured by the so-called generalized stacking fault energy (GSFE) surface which was determined from DFT-VASP calculations. By taking the chemical interaction into consideration the SVPN model can predict the dislocation core structure and mobility in the presence and absence of the solute/impurity and thus reveal the effect of impurity/solute on the softening/hardening behavior in alpha-Ti. Finally, to study the interaction of the dislocation core with other planar defects such as grain boundaries (GB), we develop an automated method to theoretically generate GBs in HCP type materials. / Dissertation/Thesis / Doctoral Dissertation Mechanical Engineering 2014

abstract: Self-heating degrades the performance of devices in advanced technology nodes. Understanding of self-heating effects is necessary to improve device performance. Heat generation in these devices occurs at nanometer scales but heat transfer is a microscopic phenomena. Hence a multi-scale modeling approach is required to study the self-heating effects. A state of the art Monte Carlo device simulator and the commercially available Giga 3D tool from Silvaco are used in our study to understand the self heating effects. The Monte Carlo device simulator solves the electrical transport and heat generation for nanometer length scales accurately while the Giga 3D tool solves for thermal transport over micrometer length scales. The approach used is to understand the self-heating effects in a test device structure, composed of a heater and a sensor, fabricated and characterized by IMEC. The heater is the Device Under Test(DUT) and the sensor is used as a probe. Therefore, the heater is biased in the saturation region and the sensor is biased in the sub-threshold regime. Both are planar MOSFETs of gate length equal to 22 nm. The simulated I-V characteristics of the sensor match with the experimental behavior at lower applied drain voltages but differ at higher applied biases. The self-heating model assumes that the heat transport within the device follows Energy Balance model which may not be accurate. To properly study heat transport within the device, a state of the art Monte Carlo device simulator is necessary. In this regard, the Phonon Monte Carlo(PMC) simulator is developed. Phonons are treated as quasi particles that carry heat energy. Like electrons, phonons obey a corresponding Boltzmann Transport Equation(BTE) which can be used to study their transport. The direct solution of the BTE for phonons is possible, but it is difficult to incorporate all scattering mechanisms. In the Monte Carlo based solution method, it is easier to incorporate different relevant scattering mechanisms. Although the Monte Carlo method is computationally intensive, it provides good insight into the physical nature of the transport problem. Hence Monte Carlo based techniques are used in the present work for studying phonon transport. Monte Carlo simulations require calculating the scattering rates for different scattering processes. In the present work, scattering rates for three phonon interactions are calculated from different approaches presented in the literature. Optical phonons are also included in the transport problem. Finally, the temperature dependence of thermal conductivity for silicon is calculated in the range from 100K to 900K and is compared to available experimental data. / Dissertation/Thesis / Masters Thesis Electrical Engineering 2016

Les images et données volumiques sont devenues importantes dans notre vie quotidienne que ce soit sur le plan artistique, culturel, ou scientifique. Les données volumiques ont un intérêt important dans l'imagerie médicale, l'ingénierie, et l'analyse du patrimoine culturel. Ils sont créées en utilisant la reconstruction tomographique, une technique qui combine une large série de scans 2D capturés de plusieur points de vue. Chaque scan 2D est obtenu par des methodes de rayonnement : Rayons X pour les scanners CT, ondes radiofréquences pour les IRM, annihilation électron-positron pour les PET scans, etc. L'acquisition des images et données volumique est influencée par le bruit provoqué par différents facteurs. Le bruit dans les images peut être causée par un manque d'éclairage, des défauts électroniques, faible dose de rayonnement, et un mauvais positionnement de l'outil ou de l'objet. Le bruit dans les données volumique peut aussi provenir d'une variété de sources : le nombre limité de points de vue, le manque de sensibilité dans les capteurs, des contrastes élevé, les algorithmes de reconstruction employés, etc. L'acquisition de données non bruitée est iréalisable. Alors, il est souhaitable de réduire ou d'éliminer le bruit le plus tôt possible dans le pipeline. La suppression du bruit tout en préservant les caractéristiques fortes d'une image ou d'un objet volumique reste une tâche difficile. Nous proposons une méthode multi-échelle pour lisser des images 2D et des données tomographiques 3D tout en préservant les caractéristiques à l'échelle spécifiée. Notre algorithme est contrôlé par un seul paramètre – la taille des caractéristiques qui doivent être préservées. Toute variation qui est plus petite que l'échelle spécifiée est traitée comme bruit et lissée, tandis que les discontinuités telles que des coins, des bords et des détails à plus grande échelle sont conservés. Nous démontrons les données lissées produites par notre algorithme permettent d'obtenir des images nettes et des iso-surfaces plus propres. Nous comparons nos résultats avec ceux des methodes précédentes. Notre méthode est inspirée par la diffusion anisotrope. Nous calculons nos tenseurs de diffusion à partir des histogrammes continues locaux de gradients autour de chaque pixel dans les images et autour de chaque voxel dans des volumes. Comme notre méthode de lissage fonctionne entièrement sur GPU, il est extrêmement rapide. / Two-dimensional images and three-dimensional volumes have become a staple ingredient of our artistic, cultural, and scientific appetite. Images capture and immortalize an instance such as natural scenes, through a photograph camera. Moreover, they can capture details inside biological subjects through the use of CT (computer tomography) scans, X-Rays, ultrasound, etc. Three-dimensional volumes of objects are also of high interest in medical imaging, engineering, and analyzing cultural heritage. They are produced using tomographic reconstruction, a technique that combine a large series of 2D scans captured from multiple views. Typically, penetrative radiation is used to obtain each 2D scan: X-Rays for CT scans, radio-frequency waves for MRI (magnetic resonance imaging), electron-positron annihilation for PET scans, etc. Unfortunately, their acquisition is influenced by noise caused by different factors. Noise in two-dimensional images could be caused by low-light illumination, electronic defects, low-dose of radiation, and a mispositioning tool or object. Noise in three-dimensional volumes also come from a variety of sources: the limited number of views, lack of captor sensitivity, high contrasts, the reconstruction algorithms, etc. The constraint that data acquisition be noiseless is unrealistic. It is desirable to reduce, or eliminate, noise at the earliest stage in the application. However, removing noise while preserving the sharp features of an image or volume object remains a challenging task. We propose a multi-scale method to smooth 2D images and 3D tomographic data while preserving features at a specified scale. Our algorithm is controlled using a single user parameter – the minimum scale of features to be preserved. Any variation that is smaller than the specified scale is treated as noise and smoothed, while discontinuities such as corners, edges and detail at a larger scale are preserved. We demonstrate that our smoothed data produces clean images and clean contour surfaces of volumes using standard surface-extraction algorithms. In addition to, we compare our results with results of previous approaches. Our method is inspired by anisotropic diffusion. We compute our diffusion tensors from the local continuous histograms of gradients around each pixel in image

Le développement de systèmes de mesures et de suivi non-invasifs, fiables, robustes et utilisablesen autonomie est de première importance pour le confort et l’implication de la personne prise en charge dansson parcours de santé. En collaboration avec la société IDS SA, ce travail a été initié et motivé par la volontéde développer des méthodes de traitement et d’analyse dédiées à l’exploration fonctionnelle de mesures phy-siologiques non-invasives d’une part, et le développement de solutions pour la santé et l’autonomie, d’autre part.Les aspects fondamentaux de ce travail doctoral ont pour objectifs la mise au point et la validation de méthodesde décomposition modale pour : (1) l’extraction et la reconstruction des composantes cardiaques d’un signalpléthysmographique et (2) l’estimation d’un indice de complexité témoignant de modifications de stratégiesde contrôle postural. Il est attendu que ces méthodes soient : (1) locales pour gérer les non-stationnarités, (2)lisibles pour permettre l’identification des composantes et (3) adaptatives au sens d’un paramétrage a minima.`A cette fin, la connaissance du contexte physiologique et des modèles spectraux attenants sont employés toutautant comme un guide dans le choix et l’utilisation qu’une grille de lecture dans l’exploitation des méthodesengagées.Dans la première partie de ce travail, à travers le développement d’un banc de test sur signaux simulés, nousavons démontré la supériorité d’une variante de la décomposition modale empirique en comparaison à sa formeoriginale. Par la suite, nous avons établi l’utilisabilité, avec des hypothèses a minima, d’une représentationmodale récente, la transformée par Synchrosqueezing (SQT).Dans la seconde partie de ce travail, nous avons mis au point un indice de complexité en échelles basé sur la SQTet dont le fenêtrage temporel est entièrement déterminé par la dynamique spectrale de la représentation. Aupréalable, nous avons mis en place un protocole expérimental pour identifier les limites des méthodes existanteset légitimer une approche spectrale de l’entropie. Notre indice a été validé sur signaux simulés et testé sur si-gnaux réels où il a pu mettre en évidence le phénomène de repondération sensorielle à la suite d’une perturbation.Les aspects appliqués de ce travail doctoral, quant à eux, s’articulent autour du développement de solutionsdédiées : (1) à la prévention du risque de chute et (2) au suivi des activités de la vie quotidienne.Dans cette dernière partie, nous nous sommes intéressés : (1) à l’évaluation des troubles latéralisés de l’équilibre,(2) à l’évaluation de la prise en charge en charge de troubles posturaux, (3) à la conception et la validationd’un outil de mesure et de rééducation des capacités de contrôle postural intégré dans un smartphone, et (4) àla détection de dérives comportementales par le développement d’un indice de persévération et son applicationau suivi du rythme nycthéméral des activités de la vie quotidienne. / L'auteur n'a pas fourni de résumé en anglais

Apesar de vários estudos terem demonstrado que a presença de abelhas resulta em aumento da produtividade de culturas agrícolas, os mecanismos que determinam a eficácia da polinização em escalas espaciais mais amplas ainda são desconhecidos. Avaliamos a influência da composição e configuração da paisagem sobre a composição da comunidade de abelhas e a polinização do café (Coffea arábica). Nosso estudo foi realizado em uma das regiões cafeeiras mais importantes do Brasil, em nove paisagens compostas de cafezais e diferentes quantidades de remanescentes de Mata Atlântica. Utilizando experimentos de exclusão floral, avaliamos a polinização em 15 pés de café por paisagem. A eficiência da polinização foi medida através da contagem e pesagem dos frutos formados. Também coletamos e identificamos os visitantes florais. Nossas análises foram feitas no nível de paisagem, com 1 e 2 km de raio, e no nível do pé de café, com 300 m de raio em torno de cada árvore. Foram coletados 241 indivíduos de abelhas e identificadas um total de 22 espécies. A abelha Apis mellifera (Apini) foi a espécie mais abundante, seguida de Trigona spinipes (Meliponini). A abundância A. mellifera foi afetada pela composição da matriz na paisagem e a composição da comunidade de abelhas nativas mudou com a distância ao fragmento florestal mais próximo. A presença das abelhas, por sua vez, resultou em um aumento 28% na produção de frutos. A abundância de A. mellifera afetou positivamente a frutificação, enquanto a composição da comunidade de abelhas nativas afetou a frutificação de modo diferente no nível da paisagem e do pé de café. Nossos resultados indicam que a estrutura da paisagem afeta a composição da comunidade de visitantes florais e que as flores expostas às abelhas apresentam um aumento na frutificação. Embora nossos resultados não permitam separar claramente os mecanismos responsáveis pelo aumento observado na frutificação, nosso trabalho mostra que o serviço de polinização das abelhas pode ser afetado pela estrutura da paisagem. Estes resultados podem ser utilizados na conservação e planejamento agrícola a fim de maximizar a produção de culturas e, ao mesmo tempo, conservar a biodiversidade e o serviço de polinização. / Although several studies have shown that the presence of bees results in increased crop yields, the mechanisms that determine pollination effectiveness on broader spatial scales are still largely unknown. We evaluated the influence of landscape composition and configuration over bee community composition and coffee (Coffea Arabica) pollination. Our study was undertaken in one of the most important coffee-producing regions of Brazil, and comprised nine landscapes of sun coffee plantations surrounded by different amounts of Atlantic Forest remnants. Using floral exclusion experiments we evaluated pollination effectiveness in 15 coffee trees per landscape. Pollination effectiveness was measured by counting and weighing the fruit set. We also sampled the bees visiting coffee flowers. Our analyses were made at the landscape level, with 1 and 2 km radii, and at the coffee tree level, with 300 m radius around each tree. We collected 241 bee individuals and identified a total of 22 species. The honeybee Apis mellifera (Apini) was the most abundant species followed by Trigona spinipes (Meliponini). Honeybee abundance was affected by landscape matrix composition while native bee community composition changed with distance to the nearest forest fragment. The presence of bees resulted in an increase in coffee fruit set of 28%. A. mellifera abundance positively affected fruit set at both levels, while the composition of the native bee community affected fruit set differently at the landscape level than at the tree level. Landscape structure was found to affect the community composition of floral visitors, and coffee flowers exposed to bees showed an increase in fruit set compared to flowers excluded from pollinators. Although our results do not allow disentangling the mechanisms responsible for the observed increase in fruit set, our work shows that bee pollination services can be affected by landscape structure. These findings can be used in conservation and agricultural planning to maximize crop production while safeguarding biodiversity and the provision of pollination services.

A Bacia Amazônica é compartilhada por sete países (Bolívia, Brasil, Colômbia, Equador, Guiana, Peru e Venezuela) e tem sido foco de diversos projetos de infraestrutura e de atividades de exploração econômica que causam contaminação e degradação, demonstrando a necessidade de governança e gestão de seus recursos hídricos. Este trabalho busca compreender a governança multi-escalar como forma de regulação do uso dos recursos hídricos transfronteiriços na Bacia Amazônica, por meio da análise das ações dos Estados e demais atores sociais em duas bacias compartilhadas: a Bacia do rio Acre e a Bacia do rio Napo. A governança de uma bacia transfronteiriça envolve diversos atores sociais na escala local, nacional e internacional que deveriam coordenar suas ações no intuito de regular o uso da água. Na escala local encontram-se as regiões de fronteira onde foram estudas a governança na bacia do rio Acre e na bacia do rio Napo. Na escala nacional buscou-se compreender o arcabouço institucional para a gestão dos recursos hídricos na Bolívia, no Brasil, no Equador e no Peru, pois são os países que compartilham as duas bacias estudadas. Na escala internacional foram analisadas as instituições regionais que abarcam os países amazônicos e desenvolvem projetos que influenciam a governança dos recursos hídricos transfronteiriços na Bacia Amazônica, em especial nas bacias do Acre e do Napo. Ao analisar a cooperação entre os Estados e entre os demais atores sociais envolvidos na regulação do uso dos recursos hídricos transfronteiriços nas diferentes escalas geográficas que se sobrepõem na Bacia Amazônica demonstrou-se que apesar da existência de cooperação nas bacias analisadas a governança ainda é incipiente. Portanto, a cooperação não é o único fator que garante a emergência de um processo de governança dos recursos hídricos transfronteiriços na Amazônia. / Amazon River Basin is shared by seven countries (Bolivia, Brazil, Colombia, Ecuador, Guyana, Peru and Venezuela) and has been the focus of infrastructure projects and economic activities that cause contamination and degradation demonstrating need for water resources governance and management. This work aims to comprehend the multi-scalar governance, understood as a mean for regulating the use of transboundary water resources, in the Amazon Basin analyzing actions taken by states and other social actors in two shared basins: the Acre River Basin and the Napo River Basin. Transboundary river basin governance involves different social actors in local, national and international scales that must coordinate their actions in order to regulate water use. Border regions are located at the local scales where the Acre River Basin and the Napo River Basin were studied. At the national scale it was analyzed the institutional framework of water resources management in Bolivia, Brazil, Ecuador and Peru, which are the countries that share the two river basins studied. At the international scale it was analyzed the cooperation and regional institutions of Amazon countries that have projects in the Amazon Basin which influence water governance in the basin, especially in the Basins of Rivers Napo and Acre. By analyzing the cooperation among Amazon countries and other social actors involved in the regulation of transboundary water resources uses in different scales overlapped in the Amazon Basin, this study demonstrated that despite the existence of cooperation in the basins analyzed, water governance is still incipient. Therefore, cooperation is not the only factor that ensures the emergence of a process of transboundary water resources governance in the two River Basins studied.

This thesis reports a study of the effects on processing and properties of incorporating nano-scale reinforcements (multiwall carbon nanotubes, MWCNTs) in the matrix of epoxy- carbon fibre (CF) laminate composites to produce multi-scale composites (M-SC). The main aim of this research was to study the effects of MWCNTs on matrix toughening and the through-thickness properties of M-SCs based on a commonly used aerospace grade epoxy resin — triglycidyl-p-aminophenol (TGPAP) cured with diaminodiphenyl sulphone (DDS). In order to improve resin processing, diglycidyl ether of bisphenol F (DGEBF) was added into the TGPAP/DDS system as a reactive diluent. Factorial experimental design (FED) was used to optimize the composition of this tri-component system to obtain high Tg and low resin viscosity, which gave a TGPAP/DGEBF/DDS system with 30.56 wt.% of DGEBF and a chemical stoichiometry of 0.5. Three types of MWCNTs were used; as-received (AR-), base-washed (BW-) and amine functionalized (NH2-). These were shear-mixed with both the bi- and tri-component systems using a 3-roll mill to produce nanocomposite matrices (NCM). The curing behaviour, dispersion state of MWCNTs in the resin and processability of NCMs were studied to decide upon the preparation method for the final M-SC. The fracture toughness (KIC) and the flexural properties of NCM were affected by both MWCNTs and the matrix type; thus KIC increased by up to 8 % in TGPAP/DDS NCM but decreased by 23% in TGPAP/DGEBF/DDS NCM with 0.5 wt.% AR-CNTs. The addition of both non-functionalized and functionalized MWCNTs increased the flexural modulus. The failure mechanism of NCMs was found to be dominated by the size and distribution of CNT aggregates and the behaviour of MWCNTs, both those dispersed in the matrix and in aggregates. The addition of functionalized MWCNTs increased the interfacial bonding between MWCNT and epoxy resin and thus improved the mechanical properties. All the NCM systems were taken forward to manufacture M-SC using a hybrid resin film infusion (RFI)/hot press process. The fibre volume fraction and the void content could be controlled at 43 ± 5 % for M-SC with TGPAP/DDS NCM and 60 ± 6 % for M-SC with TGPAP/DGEBF/DDS NCM. M-SCs were characterised using a range of tests, including flexural, interlaminar shear strength (ILSS), mode-II interlaminar fracture toughness (GIIC), low velocity impact and compression after impact (CAI). The most obvious improvement occurred for the M-SC with tri-component system with 0.5 wt.% CNTs, whereILSS increased by 16 % upon adding NH2-CNTs and GIIC increased significantly on addition of 0.5 wt.% AR-CNTs and NH2-CNTs, by 85% and 184% respectively. However the effect of MWCNTs on other properties was at best marginal. For example, for the M-SC with TGPAP/DDS, the flexural modulus and ILSS only increased by 4.1 % and 2.3 % with 0.5 wt.% AR-CNT.