Thesis (M.S.)--Ohio University, November, 1991. / Title from PDF t.p.

Thesis (Ph.D)--Electrical and Computer Engineering, Georgia Institute of Technology, 2008. / Committee Chair: Loh, Gabriel H.; Committee Co-Chair: Lee, Hsien-Hsin S.; Committee Member: Lim, Sung Kyu; Committee Member: Prvulovic, Milos; Committee Member: Yalamanchili, Sudhakar; Committee Member: Yoder, Douglas.

Στην διπλωματική εργασία παρουσιάζεται μια νέα δομή δεδομένων ειδικά σχεδιασμένη για δίκτυα μεταφορών ευρείας κλίμακας τα οποία αλλάζουν δυναμικά. Η νέα δομή δεδομένων γραφημάτων μας παρέχει ταυτόχρονα τρία μοναδικά χαρακτηρισ τικά: 1. Σύμπτυξη(Compactness): ικανότητα να προσπελάσει αποδοτικά διαδοχικές κορυφές και ακμές, μια απαίτηση όλων των αλγορίθμων γραφημάτων). 2. Ευκινησία (Agility): ικανότητα να αλλάξει και να ρυθμίσει εξαρχής την εσωτερική της διάταξη με σκοπό να βελτιώσει την τοπικότητα των αναφορών των στοιχείων, σύμφωνα με έναν δεδομένο αλγόριθμο. 3. Δυναμικότητα (Dynamicity): ικανότητα να ενθέσει ή να διαγράψει αποδοτικά κορυφές και ακμές. Όλες οι προηγούμενες γνωστές δομές γραφημάτων δεν υποστήριζαν τουλάχιστον ένα από τα προηγούμενα χαρακτηριστικά ή/και δεν μπορούσαν να εφαρμοστούν σε δυναμικά δίκτυα μεταφορών ευρείας κλίμακας. Σε αυτή τη διπλωματική εργασία, παρουσιάζεται η πρακτικότητα της νέας δομής γραφημάτων εκτελώντας μια εκτενή πειραματική μελέτη για δρομολόγηση συντομότερων διαδρομών σε Ευρωπαϊκά οδικά δίκτυα ευρείας κλίμακας με μερικές δεκάδες εκατομμύρια κορυφές και ακμές. Χρησιμοποιώντας κλασικούς αλγόριθμους εύρεσης συντομότερων διαδρομών, επιτυγχάνονται εύκολα χρόνοι ερωτημάτων από μια αρχική κορυφή σε μια τελική κορυφή της τάξης των milliseconds, ενώ η νέα δομή γραφημάτων μας μπορεί να ενημερωθεί σε μόλις μερικά microseconds μετά από μια ένθεση ή διαγραφή μιας κορυφής ή ακμής. / We present a new graph data structure specifically suited for large scale transportation networks in dynamic scenario. Our graph data structure provides tree unique characteristics, namely compactness, agility and dynamicity. All previous data structures were lacking support in at least one of the aforementioned characteristics. We demonstrate the practicality of the new graph data structure by conducting experiments on large scale European road networks, achieving query times of classical routing algorithms in the order of milliseconds and update times in the order of a few microseconds.

In this thesis I study and measure the spatial distribution of galaxies selected in optical and near-infrared surveys over cosmic time. By measuring the clustering of these sources, valuable insight can be gained into the role of environment in shaping galaxy evolution over the history of the Universe. I present a series of results from a clustering analysis of the first data release of the Visible and Infrared Survey Telescope for Astronomy (VISTA) Deep Extragalactic Observations (VIDEO) survey. VIDEO is the only survey currently capable of probing the bulk of stellar mass in galaxies at redshifts corresponding to the peak of star formation on degree scales. Galaxy clustering is measured with the two-point correlation function, which is calculated using a non para- metric kernel based density estimator. I use my measurements to investigate the connection between the galaxies and the host dark matter halo using a Halo Occupation Distribution (HOD) methodology, deriving bias, satellite fractions, and typical host halo masses for stellar masses between 10^9.35M_⊙ and 10^10.85M_⊙, at redshifts 0.5 < z < 1.7. I show that the typical halo mass increases with stellar mass (with moderate scatter) and bias also increases with stellar mass and redshift, consistent with previous studies. I find the satellite fraction increases towards low redshifts, from ∼ 5% at z ∼ 1.5, to ∼ 20% at z ∼ 0.6, also increasing for lower mass galaxies. I combine my results to derive the stellar mass to halo mass ratio for both satellites and centrals over a range of halo masses and find the peak corresponding to the halo mass with maximum star formation efficiency to be ∼ 2 × 10¹²M_⊙, finding no evidence for evolution. It has long been known that environment has a large effect on star formation in galaxies. There are several known plausible mechanisms to remove the cool gas needed for star formation, such as strangulation, harassment and ram-pressure stripping. It is unclear which process is dominant, and over what range of stellar mass. In this thesis, I find evidence for suppression of the cross-correlation function between massive galaxies and less massive star-forming galaxies, giving a measure of how less likely a galaxy is to be star-forming in the vicinity of a more massive galaxy. I develop a formalism for modelling environmental quenching mechanisms within the HOD formalism. I find that at z ∼ 2 environment is not a significant factor in determining quenching of star-forming galaxies, and that galaxies are quenched with similar probabilities in group environments as they are globally. However, by z ∼ 0.5 galaxies are much less likely to be star forming when in a group environment than when not. This increased probability of being quenched does not appear to have significant radial dependence within the halo, supportive of the quenching being caused by the halting of fresh inflows of pristine gas, as opposed to by tidal stripping. Furthermore, by separating the massive sample into passive and star-forming, I find that this effect is further enhanced when the central galaxy is passive, a manifestation of galactic conformity. Hydrodynamical cosmological simulations, with advances in computing power over the last decade, have recently made great advances in reproducing the galaxy population and understanding the underlying physical processes behind galaxy evolution. There is extensive research in the literature comparing predicted stellar mass functions from hydrodynamical simulations to observed stellar mass functions in data. In this thesis I extend these results, comparing clustering of galaxies in mock catalogues from the hydrodynamical cosmological simulation Horizon-AGN to clustering measurements from the VIDEO observations. Clustering and HOD modelling in the Horizon-AGN mock catalogue qualitatively recreates clustering measurements from the VIDEO data, but reflects the known excess stellar mass to halo mass ratio for low mass haloes in Horizon-AGN. This reinforces the need for stronger regulation of star formation in low mass haloes in the simulation. I extend my results into the high redshift regime by studying the large-scale structure of the bright high-redshift Lyman-break galaxy (LBG) population - gaining insight into the role of environment in galaxy formation physics in the early Universe. I measure the clustering of a sample of bright (−22.7 < M_UV < −21.125) LBGs at z ∼ 6 and use a HOD model to measure their typical halo masses. I find that the clustering amplitude and corresponding HOD fits sug- gests that these sources are highly biased (b ∼ 10) objects in the densest regions of the high-redshift Universe. Coupled with the observed rapid evolution of the number density of these objects, my results suggest that the shape of high lu- minosity end of the luminosity function is related to feedback processes or the onset of dust obscuration - as opposed to a scenario where these sources are pre- dominantly rare instances of the much more numerous M_UV ∼ −19 population of galaxies caught in a particularly vigorous period of star formation. Despite investigating several variations on the model, it was not possible to simultaneously fit both the number densities and clustering measurements. I interpret this as a signal that a refinement of the model halo bias relation at high redshifts or the incorporation of quasi-linear effects may be needed for future attempts at modelling the clustering and number counts. Finally, the difference in number density between the fields (UltraVISTA has a surface density ∼ 1.8 times greater than UDS) is shown to be consistent with the cosmic variance implied by the clustering measurements. Finally I discuss future data sets that will become available in the coming years, and future approaches to modelling large-scale structure. In summary I have shown that measuring the spatial distribution of galaxies on large-scales is a vital probe of galaxy evolution and an essential tool for understanding the connection between galaxies and their dark matter haloes over cosmic time.

Increasingly, online computer applications rely on large-scale data analyses to offer personalised and improved products. These large-scale analyses are performed on distributed data processing execution engines that run on thousands of networked machines housed within an individual data centre. These execution engines provide, to the programmer, the illusion of running data analysis workflows on a single machine, and offer programming interfaces that shield developers from the intricacies of implementing parallel, fault-tolerant computations. Many such execution engines exist, but they embed assumptions about the computations they execute, or only target certain types of computations. Understanding these assumptions involves substantial study and experimentation. Thus, developers find it difficult to determine which execution engine is best, and even if they did, they become “locked in” because engineering effort is required to port workflows. In this dissertation, I first argue that in order to execute data analysis computations efficiently, and to flexibly choose the best engines, the way we specify data analysis computations should be decoupled from the execution engines that run the computations. I propose an architecture for decoupling data processing, together with Musketeer, my proof-of-concept implementation of this architecture. In Musketeer, developers express data analysis computations using their preferred programming interface. These are translated into a common intermediate representation from which code is generated and executed on the most appropriate execution engine. I show that Musketeer can be used to write data analysis computations directly, and these can execute on many execution engines because Musketeer automatically generates code that is competitive with optimised hand-written implementations. The diverse execution engines cause different workflow types to coexist within a data centre, opening up both opportunities for sharing and potential pitfalls for co-location interference. However, in practice, workflows are either placed by high-quality schedulers that avoid co-location interference, but choose placements slowly, or schedulers that choose placements quickly, but with unpredictable workflow run time due to co-location interference. In this dissertation, I show that schedulers can choose high-quality placements with low latency. I develop several techniques to improve Firmament, a high-quality min-cost flow-based scheduler, to choose placements quickly in large data centres. Finally, I demonstrate that Firmament chooses placements at least as good as other sophisticated schedulers, but at the speeds associated with simple schedulers. These contributions enable more efficient and effective use of data centres for large-scale computation than current solutions.

Restos de moluscos acumulados na superfície do sedimento marinho (i.e., assembleias mortas, retrabalhadas ou não) podem ser os principais testemunhos do efeito de fatores ambientais (oceanográficos, sedimentares, climáticos) sobre o modo como os moluscos fósseis são preservados. Ou seja, o quanto o meio ambiente é responsável pela fossilização de restos esqueléticos só pode ser medido, com máxima fidelidade, através do estudo de assembleias modernas que possuam restos refratários (e.g., conchas). Fatores ambientais atuais podem influenciar o dano tafonômico em moluscos de depósitos retrabalhados. Embora muitos estudos atualísticos tenham estabelecido relações entre determinados fatores ambientais e danos causados às conchas, poucos i) enfocam a fração de microclastos (< 4 mm), são comuns em amostras de subsuperfície, ou ii) analisam essas relações utilizando escalas espaciais maiores que local (e.g., regional, diferentes províncias marinhas), especialmente com o propósito de testar certos fatores ambientais em escala local e ampla (e.g., temperatura local vs. faixas de temperatura). Os raros estudos existentes tratam-se meta análises. No registro fóssil, muitas vezes, apenas gradientes ambientais amplos são recuperados. Por isso é importante avaliar a relação destes fatores (modernos) com as assinaturas tafonômicas em moluscos. O objetivo desta tese foi determinar até que ponto variações em fatores ambientais (salinidade, temperatura, velocidade de corrente, granulometria e composição do sedimento) geralmente observados em escala local, influenciam o tipo de dano tafonômico em micro e macroclastos de moluscos marinhos, se analisados em escala mais ampla. Grande parte do material analisado esteve representado por clastos de remaniés existentes na Plataforma Sul Brasileira (PSB). Desta forma, o presente estudo também investigou se fatores ambientais atuais podem ser deduzidos a partir de perfis tafonômicos de assembleias mortas retrabalhadas. Os resultados obtidos mostraram que, na PSB, i) apesar de diferentes em termos absolutos, o perfil de dano de cada amostra analisada não varia entre macro e micromoluscos, e que ii) quando o estudo é conduzido em escala espacial mais ampla, até 60% da variação no perfil tafonômico pode ser explicada pela variação ambiental. Ainda nesta região, os principais fatores ambientais que afetaram o dano em microclastos de bivalves são: i) salinidade, que reflete o aporte de água doce (e da circulação oceânica como um todo), e ii) proporção de carbonato no sedimento, que reflete, primariamente, a existência de depósitos bioclásticos (constituídos de conchas de moluscos retrabalhadas, ou remaniés biológicos). Os depósitos bioclásticos mais proximais (ao sul da desembocadura da Lagoa dos Patos) apresentaram maior evidência de dissolução e retrabalhamento físico, enquanto que os bioclastos do depósito mais distal (entre as profundidades de ~ 100 e 200 m) apresentaram melhor preservação tafonômica. Além disso, também pôde ser determinado que a alteração de cor dos bioclastos na PSB (bastante ubíquo em amostras concentradas em cascalho e carbonato) está vinculada, significativamente, a estados de oxidação/redução do sedimento. Estes, por sua vez, estão relacionados à presença de elementos metálicos (como ferro e manganês), importantes na mineralização do carbono de origem biológica, e na geoquímica da zona tafonomicamente ativa, em ambiente marinho. Na análise conduzida em escala espacial geográfica, incluindo material do Plataforma do Nordeste do Brasil, pode ser determinado que profundidade, um fator ambiental com influência significativa sobre o perfil de dano na PSB, perde importância. Em diferentes províncias geográficas, os principais fatores que influenciaram as assinaturas tafonômicas estão vinculados à velocidade da corrente, salinidade e proporção de carbonato no sedimento. Estes refletiram a influência de diferentes massas d‟água (Corrente Norte do Brasil e Corrente das Malvinas) e diferentes regimes de sedimentação (siliciclástica relíquia e carbonática atual). / Dead molluscan remains (i.e., death assemblages, reworked or not), accumulated on the marine sedimentary surface, can preserve most of the environmental factors (oceanographic, sedimentary, climatic) which control fossil mollusk preservation. That is, how much the environment is responsible on skeletal remain fossilization can only be measured, with the highest fidelity, by studying modern assemblages possessing refractory parts (such as shells). Present-day environmental factors can influence the taphonomic damage on reworked mollusk remains. Although atualistic studies had constrained some environmental factors and shell damage, a few i) focuses mesh sizes lower than 4 mm (microclasts), which are common in subsurface samples, or ii) analyze these relationships using spatial scales larger than local (e.g., regional, different marine provinces), especially with the explicit purpose of to test some environmental factors on local and large scale (e.g., local temperature vs. temperature ranges). The rare studies available are meta-analysis. In the fossil record, sometimes only large environmental gradients can be recovered. That is why it is important to evaluate the relationship between modern environmental factors and taphonomic signatures in mollusks. The subject of this thesis is to determine to which degree the variation in environmental factors (salinity, temperature, current velocity, sediment granulometry and composition), usually observed in local scale, do influence the kind of taphonomic alteration in micro and macroclasts of marine mollusks, when analyzed in large scale. Most of the studied material was represented by clasts from remaniés localized in Brazilian South Shelf (PSB). In this way, the present study also investigated whether modern environmental factors can be deduced from taphonomic profiles of reworked death assemblages. The results here obtained show that, in the PSB, (i) despite absolute differences, the taphonomic profile of micro and macroclasts does not vary, and that (ii) when the study is conducted in larger spatial scales, up to 60% of variation in the taphonomic profile can be explained by the environmental variation. Still in this region, the main environmental factors influencing the damage in bivalve microclasts are: i) salinity, which regards the freshwater input (and also the oceanic circulation as a whole), and ii) the proportion of carbonate in the sediment, which reflects, primarily, the existence of bioclastic deposits (constituted mostly by reworked molluscan shells, or biological remaniés). The proximal bioclastic deposits (south of Lagoa dos Patos outflow) presented the higher evidence of dissolution and physical reworking, while the distal bioclastic deposits (located within the batimetric range of 100 and 200 m) showed better taphonomic preservation. Beyond this, it was also possible to determine that color alteration in bioclasts from PSB (a ubiquitous feature in samples concentrated with gravel and carbonate) is significantly linked to sediment redox status. The color alteration, by turn, are related to the presence of metallic elements (such as iron and manganese), which are important in the biological carbon mineralization, and in the TAZ geochemistry in marine environment. The analysis conducted at geographic spatial scale, including material from Northeastern Brazilian Shelf allowed determining that depth, a significant environmental factor influencing taphonomic damage in PSB, lost importance in large scale. In different geographic provinces, the main factors influencing taphonomic signatures are linked to current velocity, salinity and sedimentary carbonate. These environmental factors reflect the influence of differing water masses (North Brazilian Current and Malvinas Current) and different sedimentary regimes (relict siliciclastics and modern carbonatic).

Ce mémoire de thèse porte sur l’élaboration de systèmes de reconnaissance d’image qui sont robustes à la variabilité géométrique. La compréhension d’une image est un problème difficile, de par le fait qu’elles sont des projections en deux dimensions d’objets 3D. Par ailleurs, des représentations qui doivent appartenir à la même catégorie, par exemple des objets de la même classe en classification, peuvent être visuellement très différentes. Notre but est de rendre ces systèmes robustes à la juste quantité de déformations, celle-ci étant automatiquement déterminée à partir des données. Nos deux contributions sont les suivantes. Nous montrons tout d’abord comment utiliser des exemples virtuels pour rendre les systèmes de classification d’images robustes et nous proposons ensuite une méthodologie pour apprendre des descripteurs de bas niveau robustes, pour la recherche d’image.Nous étudions tout d’abord les exemples virtuels, en tant que transformations de vrais exemples. En représentant une image en tant que sac de descripteurs transformés, nous montrons que l’augmentation de données, c’est-à-dire le fait de les considérer comme de nouveaux exemples iid, est la meilleure manière de les utiliser, pourvu qu’une étape de vote avec les descripteurs transformés soit opérée lors du test. Du fait que les transformations apportent différents niveaux d’information, peuvent être redondants, voire nuire à la performance, nous pro-posons un nouvel algorithme capable de sélectionner un petit nombre d’entre elles,en maximisant la justesse de classification. Nous montrons par ailleurs comment remplacer de vrais exemples par des virtuels, pour alléger les couts d’annotation.Nous rapportons de bons résultats sur des bancs d’essai de classification.Notre seconde contribution vise à améliorer les descripteurs de régions locales utilisés en recherche d’image, et en particulier nous proposons une alternative au populaire descripteur SIFT. Nous proposons un nouveau descripteur, appelé patch-CKN, appris sans supervision. Nous introduisons un nouvel ensemble de données liant les images et les imagettes, construit à partir de reconstruction3D automatique d’images récupérées sur Internet. Nous définissons une méthode pour tester précisément la performance des descripteurs locaux au niveau de l’imagette et de l’image. Notre approche dépasse SIFT et les autres approches à base d’architectures convolutionnelles sur notre banc d’essai, et d’autres couramment utilisés dans la littérature. / This dissertation focuses on designing image recognition systems which are robust to geometric variability. Image understanding is a difficult problem, as images are two-dimensional projections of 3D objects, and representations that must fall into the same category, for instance objects of the same class in classification can display significant differences. Our goal is to make systems robust to the right amount of deformations, this amount being automatically determined from data. Our contributions are twofolds. We show how to use virtual examples to enforce robustness in image classification systems and we propose a framework to learn robust low-level descriptors for image retrieval. We first focus on virtual examples, as transformation of real ones. One image generates a set of descriptors –one for each transformation– and we show that data augmentation, ie considering them all as iid samples, is the best performing method to use them, provided a voting stage with the transformed descriptors is conducted at test time. Because transformations have various levels of information, can be redundant, and can even be harmful to performance, we propose a new algorithm able to select a set of transformations, while maximizing classification accuracy. We show that a small amount of transformations is enough to considerably improve performance for this task. We also show how virtual examples can replace real ones for a reduced annotation cost. We report good performance on standard fine-grained classification datasets. In a second part, we aim at improving the local region descriptors used in image retrieval and in particular to propose an alternative to the popular SIFT descriptor. We propose new convolutional descriptors, called patch-CKN, which are learned without supervision. We introduce a linked patch- and image-retrieval dataset based on structure from motion of web-crawled images, and design a method to accurately test the performance of local descriptors at patch and image levels. Our approach outperforms both SIFT and all tested approaches with convolutional architectures on our patch and image benchmarks, as well as several styate-of-theart datasets.

Submitted by MARIA ANGELA OLIVEIRA DE SÁ RUBINI null (angela.rubini@yahoo.com.br) on 2017-08-24T18:38:36Z No. of bitstreams: 1 dissertacao_angela_final (4).pdf: 1668114 bytes, checksum: 8550f99f5ce36e398de2238547739756 (MD5) / Approved for entry into archive by LUIZA DE MENEZES ROMANETTO (luizamenezes@reitoria.unesp.br) on 2017-08-25T20:57:20Z (GMT) No. of bitstreams: 1 rubini_maos_me_prud.pdf: 1668114 bytes, checksum: 8550f99f5ce36e398de2238547739756 (MD5) / Made available in DSpace on 2017-08-25T20:57:20Z (GMT). No. of bitstreams: 1 rubini_maos_me_prud.pdf: 1668114 bytes, checksum: 8550f99f5ce36e398de2238547739756 (MD5) Previous issue date: 2017-03-31 / O presente trabalho, vinculado à linha de pesquisa “Formação dos Profissionais da Educação, Políticas Educativas e Escola Pública” do Programa de Pós-Graduação em Educação na Faculdade de Ciências e Tecnologia da Universidade Estadual “Júlio de Mesquita Filho” (FCT/UNESP) campus de Presidente Prudente, tem como tema a avaliação em larga escala e suas implicações na organização escolar. O objetivo foi identificar e analisar os efeitos do IDEB na organização de uma escola da rede municipal de Presidente Prudente que atende as séries iniciais de ensino fundamental. Com o propósito de responder a esses questionamentos optamos pela pesquisa de abordagem qualitativa e como instrumento de coleta de dados a entrevista semiestruturada. Para a análise das políticas educacionais recorremos ao ciclo de políticas elaborado por Stephen Ball e seus colaboradores. Tal abordagem nos permite analisar criticamente as políticas inter-relacionando as dimensões global e local. Os sujeitos participantes do estudo foram: o Diretor, a Orientadora Pedagógica, uma professora de cada ano (1º, 2º, 3º e 4º) e duas professoras dos quintos anos da escola. Os resultados da pesquisa revelaram que a Prova Brasil e a divulgação dos resultados do IDEB influenciam na organização da escola, pois esse índice emerge no chão da escola como um direcionador do trabalho pedagógico, um referencial para a análise das competências e habilidades desenvolvidas com os alunos. Os sujeitos participantes anunciaram as avaliações externas como positivas e necessárias para o direcionamento do trabalho escolar, contudo, no decorrer da entrevista evidenciaram-se elementos favoráveis e elementos desfavoráveis relacionados às mesmas. Os dados revelaram que o IDEB também interfere na formação continuada dos profissionais da educação. O resultado do índice é subsídio para a organização dos estudos que fundamentam as práticas docentes; contudo, observamos uma ênfase nas áreas de língua portuguesa e matemática. Esse aspecto indica a necessidade de aprofundamento nos estudos sobre as origens das avaliações externas e do papel da escola pública para a melhoria da qualidade da educação básica. / This study, linked to the research area "Formation of Education Professionals, Educational Policies and Public School" from the Graduate Studies in Education Program at the Faculty of Science and Technology of "Júlio de Mesquita Filho" State University (FCT / UNESP) Presidente Prudente campus, has as subject the evaluation on a large scale and its implications for school organization. In order to answer these questions, qualitative research and semistructured interview were chosen as data collection tools. For the analysis of educational policies the policy cycle elaborated by Stephen Ball and his collaborators was used. Such an approach allows us to critically analyze policies interrelating the global and local dimensions. The subjects that participated in the study were: the Director, the Pedagogical Advisor, a teacher from each year (1st, 2nd, 3rd and 4th) and two teachers from the school's fifth grade. The results of the research revealed that "Prova Brasil" and the dissemination of IDEB results influence the organization of the school, because such index emerges from the school's daily life as a driver of the pedagogical work, a reference for analyzing the skills and abilities developed with the students. The collaborators announced the external evaluations as positive and necessary for guiding school work, however, during the interview they highlighted favorable and unfavorable elements related to such evaluations. The data revealed that IDEB also interferes with the continuous training of education professionals. The result of the index is a subsidy for the organization of studies which underlie teaching practices; however, an emphasis on the areas of Portuguese and Mathematics are noticed. This aspect indicates the need for further study in the origins of external evaluations and the role of public school in improving the quality of basic education.

Restos de moluscos acumulados na superfície do sedimento marinho (i.e., assembleias mortas, retrabalhadas ou não) podem ser os principais testemunhos do efeito de fatores ambientais (oceanográficos, sedimentares, climáticos) sobre o modo como os moluscos fósseis são preservados. Ou seja, o quanto o meio ambiente é responsável pela fossilização de restos esqueléticos só pode ser medido, com máxima fidelidade, através do estudo de assembleias modernas que possuam restos refratários (e.g., conchas). Fatores ambientais atuais podem influenciar o dano tafonômico em moluscos de depósitos retrabalhados. Embora muitos estudos atualísticos tenham estabelecido relações entre determinados fatores ambientais e danos causados às conchas, poucos i) enfocam a fração de microclastos (< 4 mm), são comuns em amostras de subsuperfície, ou ii) analisam essas relações utilizando escalas espaciais maiores que local (e.g., regional, diferentes províncias marinhas), especialmente com o propósito de testar certos fatores ambientais em escala local e ampla (e.g., temperatura local vs. faixas de temperatura). Os raros estudos existentes tratam-se meta análises. No registro fóssil, muitas vezes, apenas gradientes ambientais amplos são recuperados. Por isso é importante avaliar a relação destes fatores (modernos) com as assinaturas tafonômicas em moluscos. O objetivo desta tese foi determinar até que ponto variações em fatores ambientais (salinidade, temperatura, velocidade de corrente, granulometria e composição do sedimento) geralmente observados em escala local, influenciam o tipo de dano tafonômico em micro e macroclastos de moluscos marinhos, se analisados em escala mais ampla. Grande parte do material analisado esteve representado por clastos de remaniés existentes na Plataforma Sul Brasileira (PSB). Desta forma, o presente estudo também investigou se fatores ambientais atuais podem ser deduzidos a partir de perfis tafonômicos de assembleias mortas retrabalhadas. Os resultados obtidos mostraram que, na PSB, i) apesar de diferentes em termos absolutos, o perfil de dano de cada amostra analisada não varia entre macro e micromoluscos, e que ii) quando o estudo é conduzido em escala espacial mais ampla, até 60% da variação no perfil tafonômico pode ser explicada pela variação ambiental. Ainda nesta região, os principais fatores ambientais que afetaram o dano em microclastos de bivalves são: i) salinidade, que reflete o aporte de água doce (e da circulação oceânica como um todo), e ii) proporção de carbonato no sedimento, que reflete, primariamente, a existência de depósitos bioclásticos (constituídos de conchas de moluscos retrabalhadas, ou remaniés biológicos). Os depósitos bioclásticos mais proximais (ao sul da desembocadura da Lagoa dos Patos) apresentaram maior evidência de dissolução e retrabalhamento físico, enquanto que os bioclastos do depósito mais distal (entre as profundidades de ~ 100 e 200 m) apresentaram melhor preservação tafonômica. Além disso, também pôde ser determinado que a alteração de cor dos bioclastos na PSB (bastante ubíquo em amostras concentradas em cascalho e carbonato) está vinculada, significativamente, a estados de oxidação/redução do sedimento. Estes, por sua vez, estão relacionados à presença de elementos metálicos (como ferro e manganês), importantes na mineralização do carbono de origem biológica, e na geoquímica da zona tafonomicamente ativa, em ambiente marinho. Na análise conduzida em escala espacial geográfica, incluindo material do Plataforma do Nordeste do Brasil, pode ser determinado que profundidade, um fator ambiental com influência significativa sobre o perfil de dano na PSB, perde importância. Em diferentes províncias geográficas, os principais fatores que influenciaram as assinaturas tafonômicas estão vinculados à velocidade da corrente, salinidade e proporção de carbonato no sedimento. Estes refletiram a influência de diferentes massas d‟água (Corrente Norte do Brasil e Corrente das Malvinas) e diferentes regimes de sedimentação (siliciclástica relíquia e carbonática atual). / Dead molluscan remains (i.e., death assemblages, reworked or not), accumulated on the marine sedimentary surface, can preserve most of the environmental factors (oceanographic, sedimentary, climatic) which control fossil mollusk preservation. That is, how much the environment is responsible on skeletal remain fossilization can only be measured, with the highest fidelity, by studying modern assemblages possessing refractory parts (such as shells). Present-day environmental factors can influence the taphonomic damage on reworked mollusk remains. Although atualistic studies had constrained some environmental factors and shell damage, a few i) focuses mesh sizes lower than 4 mm (microclasts), which are common in subsurface samples, or ii) analyze these relationships using spatial scales larger than local (e.g., regional, different marine provinces), especially with the explicit purpose of to test some environmental factors on local and large scale (e.g., local temperature vs. temperature ranges). The rare studies available are meta-analysis. In the fossil record, sometimes only large environmental gradients can be recovered. That is why it is important to evaluate the relationship between modern environmental factors and taphonomic signatures in mollusks. The subject of this thesis is to determine to which degree the variation in environmental factors (salinity, temperature, current velocity, sediment granulometry and composition), usually observed in local scale, do influence the kind of taphonomic alteration in micro and macroclasts of marine mollusks, when analyzed in large scale. Most of the studied material was represented by clasts from remaniés localized in Brazilian South Shelf (PSB). In this way, the present study also investigated whether modern environmental factors can be deduced from taphonomic profiles of reworked death assemblages. The results here obtained show that, in the PSB, (i) despite absolute differences, the taphonomic profile of micro and macroclasts does not vary, and that (ii) when the study is conducted in larger spatial scales, up to 60% of variation in the taphonomic profile can be explained by the environmental variation. Still in this region, the main environmental factors influencing the damage in bivalve microclasts are: i) salinity, which regards the freshwater input (and also the oceanic circulation as a whole), and ii) the proportion of carbonate in the sediment, which reflects, primarily, the existence of bioclastic deposits (constituted mostly by reworked molluscan shells, or biological remaniés). The proximal bioclastic deposits (south of Lagoa dos Patos outflow) presented the higher evidence of dissolution and physical reworking, while the distal bioclastic deposits (located within the batimetric range of 100 and 200 m) showed better taphonomic preservation. Beyond this, it was also possible to determine that color alteration in bioclasts from PSB (a ubiquitous feature in samples concentrated with gravel and carbonate) is significantly linked to sediment redox status. The color alteration, by turn, are related to the presence of metallic elements (such as iron and manganese), which are important in the biological carbon mineralization, and in the TAZ geochemistry in marine environment. The analysis conducted at geographic spatial scale, including material from Northeastern Brazilian Shelf allowed determining that depth, a significant environmental factor influencing taphonomic damage in PSB, lost importance in large scale. In different geographic provinces, the main factors influencing taphonomic signatures are linked to current velocity, salinity and sedimentary carbonate. These environmental factors reflect the influence of differing water masses (North Brazilian Current and Malvinas Current) and different sedimentary regimes (relict siliciclastics and modern carbonatic).

Restos de moluscos acumulados na superfície do sedimento marinho (i.e., assembleias mortas, retrabalhadas ou não) podem ser os principais testemunhos do efeito de fatores ambientais (oceanográficos, sedimentares, climáticos) sobre o modo como os moluscos fósseis são preservados. Ou seja, o quanto o meio ambiente é responsável pela fossilização de restos esqueléticos só pode ser medido, com máxima fidelidade, através do estudo de assembleias modernas que possuam restos refratários (e.g., conchas). Fatores ambientais atuais podem influenciar o dano tafonômico em moluscos de depósitos retrabalhados. Embora muitos estudos atualísticos tenham estabelecido relações entre determinados fatores ambientais e danos causados às conchas, poucos i) enfocam a fração de microclastos (< 4 mm), são comuns em amostras de subsuperfície, ou ii) analisam essas relações utilizando escalas espaciais maiores que local (e.g., regional, diferentes províncias marinhas), especialmente com o propósito de testar certos fatores ambientais em escala local e ampla (e.g., temperatura local vs. faixas de temperatura). Os raros estudos existentes tratam-se meta análises. No registro fóssil, muitas vezes, apenas gradientes ambientais amplos são recuperados. Por isso é importante avaliar a relação destes fatores (modernos) com as assinaturas tafonômicas em moluscos. O objetivo desta tese foi determinar até que ponto variações em fatores ambientais (salinidade, temperatura, velocidade de corrente, granulometria e composição do sedimento) geralmente observados em escala local, influenciam o tipo de dano tafonômico em micro e macroclastos de moluscos marinhos, se analisados em escala mais ampla. Grande parte do material analisado esteve representado por clastos de remaniés existentes na Plataforma Sul Brasileira (PSB). Desta forma, o presente estudo também investigou se fatores ambientais atuais podem ser deduzidos a partir de perfis tafonômicos de assembleias mortas retrabalhadas. Os resultados obtidos mostraram que, na PSB, i) apesar de diferentes em termos absolutos, o perfil de dano de cada amostra analisada não varia entre macro e micromoluscos, e que ii) quando o estudo é conduzido em escala espacial mais ampla, até 60% da variação no perfil tafonômico pode ser explicada pela variação ambiental. Ainda nesta região, os principais fatores ambientais que afetaram o dano em microclastos de bivalves são: i) salinidade, que reflete o aporte de água doce (e da circulação oceânica como um todo), e ii) proporção de carbonato no sedimento, que reflete, primariamente, a existência de depósitos bioclásticos (constituídos de conchas de moluscos retrabalhadas, ou remaniés biológicos). Os depósitos bioclásticos mais proximais (ao sul da desembocadura da Lagoa dos Patos) apresentaram maior evidência de dissolução e retrabalhamento físico, enquanto que os bioclastos do depósito mais distal (entre as profundidades de ~ 100 e 200 m) apresentaram melhor preservação tafonômica. Além disso, também pôde ser determinado que a alteração de cor dos bioclastos na PSB (bastante ubíquo em amostras concentradas em cascalho e carbonato) está vinculada, significativamente, a estados de oxidação/redução do sedimento. Estes, por sua vez, estão relacionados à presença de elementos metálicos (como ferro e manganês), importantes na mineralização do carbono de origem biológica, e na geoquímica da zona tafonomicamente ativa, em ambiente marinho. Na análise conduzida em escala espacial geográfica, incluindo material do Plataforma do Nordeste do Brasil, pode ser determinado que profundidade, um fator ambiental com influência significativa sobre o perfil de dano na PSB, perde importância. Em diferentes províncias geográficas, os principais fatores que influenciaram as assinaturas tafonômicas estão vinculados à velocidade da corrente, salinidade e proporção de carbonato no sedimento. Estes refletiram a influência de diferentes massas d‟água (Corrente Norte do Brasil e Corrente das Malvinas) e diferentes regimes de sedimentação (siliciclástica relíquia e carbonática atual). / Dead molluscan remains (i.e., death assemblages, reworked or not), accumulated on the marine sedimentary surface, can preserve most of the environmental factors (oceanographic, sedimentary, climatic) which control fossil mollusk preservation. That is, how much the environment is responsible on skeletal remain fossilization can only be measured, with the highest fidelity, by studying modern assemblages possessing refractory parts (such as shells). Present-day environmental factors can influence the taphonomic damage on reworked mollusk remains. Although atualistic studies had constrained some environmental factors and shell damage, a few i) focuses mesh sizes lower than 4 mm (microclasts), which are common in subsurface samples, or ii) analyze these relationships using spatial scales larger than local (e.g., regional, different marine provinces), especially with the explicit purpose of to test some environmental factors on local and large scale (e.g., local temperature vs. temperature ranges). The rare studies available are meta-analysis. In the fossil record, sometimes only large environmental gradients can be recovered. That is why it is important to evaluate the relationship between modern environmental factors and taphonomic signatures in mollusks. The subject of this thesis is to determine to which degree the variation in environmental factors (salinity, temperature, current velocity, sediment granulometry and composition), usually observed in local scale, do influence the kind of taphonomic alteration in micro and macroclasts of marine mollusks, when analyzed in large scale. Most of the studied material was represented by clasts from remaniés localized in Brazilian South Shelf (PSB). In this way, the present study also investigated whether modern environmental factors can be deduced from taphonomic profiles of reworked death assemblages. The results here obtained show that, in the PSB, (i) despite absolute differences, the taphonomic profile of micro and macroclasts does not vary, and that (ii) when the study is conducted in larger spatial scales, up to 60% of variation in the taphonomic profile can be explained by the environmental variation. Still in this region, the main environmental factors influencing the damage in bivalve microclasts are: i) salinity, which regards the freshwater input (and also the oceanic circulation as a whole), and ii) the proportion of carbonate in the sediment, which reflects, primarily, the existence of bioclastic deposits (constituted mostly by reworked molluscan shells, or biological remaniés). The proximal bioclastic deposits (south of Lagoa dos Patos outflow) presented the higher evidence of dissolution and physical reworking, while the distal bioclastic deposits (located within the batimetric range of 100 and 200 m) showed better taphonomic preservation. Beyond this, it was also possible to determine that color alteration in bioclasts from PSB (a ubiquitous feature in samples concentrated with gravel and carbonate) is significantly linked to sediment redox status. The color alteration, by turn, are related to the presence of metallic elements (such as iron and manganese), which are important in the biological carbon mineralization, and in the TAZ geochemistry in marine environment. The analysis conducted at geographic spatial scale, including material from Northeastern Brazilian Shelf allowed determining that depth, a significant environmental factor influencing taphonomic damage in PSB, lost importance in large scale. In different geographic provinces, the main factors influencing taphonomic signatures are linked to current velocity, salinity and sedimentary carbonate. These environmental factors reflect the influence of differing water masses (North Brazilian Current and Malvinas Current) and different sedimentary regimes (relict siliciclastics and modern carbonatic).