Made available in DSpace on 2016-08-17T15:00:42Z (GMT). No. of bitstreams: 1 LAIS_ DE_ MORAIS_ REGO _SILVA.pdf: 832865 bytes, checksum: c498c083ec0f78b354b2e99bdf847b57 (MD5) Previous issue date: 2007-02-15 / PETRÓLEO BRASILEIRO S.A. / The objective of this work was to characterize use of habitat and seasonality of shorebirds in Raposa s Channel, North of São Luís Island, Gulf of Maranhão. The censuses of shorebirds occurred during two consecutive weeks of each month between February and November 2006. Three areas (each one with 3ha) were selected for censuses, considering the type of flat (mud, sand-mud, sand). The censuses were taken during one spring tide and during one neap tide of each month. In foraging, 17407 shorebirds were observed during the study. The highest density of shorebirds feeding was in the mudflat (180.95 ind/ha), then sand-mudflat (100.25 ind/ha) and sandflat (27.10 ind/ha). Calidris pusilla was the most abundant species in Raposa s Channel (n = 8922) and in the mudflat (n = 7578). In sand-mudflat, Tringa semipalmata was the most abundant species (n = 1270) and in the sandflat it was Pluvialis squatarola (n = 130). Tringa semipalmata was the most abundant species in the roost area (n = 788). The highest density of shorebirds feeding was observed during neap tides. The high density of shorebirds observed probably might be due to the study area to form a mosaic of habitats still conserved. Pluvialis squatarola, Numenius phaeopus, Tringa semipalmata, Calidris canutus, A. interpres and Charadrius semipalmatus presented this seasonal pattern: high density between February and April, decreases between May and July and increase in density between August and November. Calidris pusilla and differed for presenting increases in density in July; Limnodromus griseus had presented high density only since August and Charadrius collaris had increases in density between February and July, and decreases between August and November. The seasonal patterns demonstrated the importance of Raposa s Channel as stopover area for some species and winter area for others shorebirds species. Thus, the identification and characterization of Raposa s Channel as important wintering area for migratory shorebirds is fundamental to create conservation plans for north brazilian coast, in order to maintaining the shorebirds populations at long term in nature. / Este trabalho teve como objetivo caracterizar o uso de habitats e a sazonalidade de aves limícolas no Canal da Raposa, norte da Ilha de São Luís, Golfão Maranhense. Os censos de aves limícolas ocorreram durante duas semanas consecutivas de cada mês entre fevereiro e novembro de 2006. Foram selecionadas três áreas (cada uma com 3ha) para a realização dos censos, considerando o tipo de substrato (lamoso, areno-lamoso e arenoso). Os censos foram realizados durante uma maré de sizígia e durante uma maré de quadratura de cada mês. Foram observadas 17407 aves limícolas forrageando durante o período de estudo. A maior densidade de aves limícolas em forrageio foi verificada no substrato lamoso (180,95 ind/ha), seguido pelo areno-lamoso (100,25 ind/ha) e pelo arenoso (27,10 ind/ha). Calidris pusilla foi a espécie mais abundante (n = 8922) no Canal da Raposa, sendo também mais abundante no susbtrato lamoso (n = 7578). No Areno-lamoso, Tringa semipalmata foi mais abundante (n = 1270) e no arenoso foi Pluvialis squatarola (n = 130). Tringa semipalmata também foi a espécie mais abundante (n = 788) na área de descanso. Houve maior densidade de aves limícolas forrageando durante as marés de quadratura. A alta densidade de aves limícolas observada provavelmente se deve ao fato da área estudada se tratar de um mosaico de habitats ainda conservados. Pluvilais squatarola, Numenius phaeopus, Tringa semipalmata, Calidris canutus, Arenaria interpres e Charadrius semipalmatus apresentaram o seguinte padrão sazonal: alta densidade entre fevereiro e abril, decréscimos entre maio e julho e acréscimos de agosto a novembro. Calidris pusilla diferiu por apresentar acréscimos no mês de julho; Limnodromus griseus apresentou altas densidades somente a partir de agosto e Charadrius collaris obteve maior densidade de fevereiro a julho e decréscimos de agosto a novembro. Os padrões sazonais mostraram a importância do Canal da Raposa como área de parada para algumas espécies de limícolas e como área de invernada para outras. Portanto, a identificação e caracterização do Canal da Raposa como uma importante área de invernada para aves limícolas migratórias é fundamental para traçar planos de conservação para a costa norte brasileira, a fim de manter as populações de aves limícolas em longo prazo na natureza.

Le travail présenté dans cette thèse a pour sujet la mesure du seul mécanisme de production électrofaible de quark top qui n'ait pas encore été observé au collisionneur hadronique du CERN (LHC): la voie-s. L'intérêt réside dans la confirmation des prédictions du Modèle Standard, mais le but est aussi de contraindre plusieurs scénarios de nouvelle physique à partir de la combinaison de la section efficace voie-s avec celle des autres canaux de production single top, la voie-t et Wt. Après une introduction générale sur la physique du quark top et les dispositifs expérimentaux visant à la détection et la reconstruction des objets physiques, on présente deux analyses réalisées des données de collisions proton-proton enregistrées par le détecteur ATLAS `a une énergie dans le centre de masse de 7 et 8 TeV. Etant donné que la production de quark top en voie-s est caractérisée par une tr`es faible pureté et que, en même temps, elle est défavorisée au LHC puisque elle a lieu par annihilation quark-antiquark, une technique multi variée a été appliquée dans les deux cas afin de discriminer le signal. Pour l' étude à 7 TeV, relatif à une luminosité intégrée de 4.7 $fb^{-1}$, deux classificateurs basés sur des arbres de décision boostés sont entrainés contre les événements de bruits de fond principaux; une coupure appliquée sur un des discriminateurs met en oeuvre une sélection des événements, la distribution issue de l'autre discriminateur est ensuite employée dans un fit de maximum de vraisemblance. Cela permet de d'eriver une limite supérieure sur la section efficace voie-s de 21.5 pb (14.3 pb attendu), correspondant à une portée de 0.6 déviations standard (0.8 attendu). L'analyse à 8 TeV, qui est basée sur une luminosité intégrée de 20.3 $fb^{-1}$, repose sur une strat'egie plus simple: seulement un classificateur BDT est optimisé pour isoler le signal, après une préalable sélection des événements grâce `a de simples coupures. Sa distribution discriminante est ensuite ajustée aux données selon la même approche qu' `a 7 TeV, en considérant les incertitudes systématiques et statistiques pendant la génération de pseudo expériences. La portée de la mesure étant encore inférieure à trois déviations standard (1.5 $sigma$ observé et 1.4 attendu), seule une limite `a 95% C.L. peut être établie sur la section efficace; le résultat correspond `a 15.7 pb (7.9 pb attendu). / The work reported in this thesis is aimed at measuring with the ATLAS detector theonly mechanism of top quark electroweak production which has not yet been observedat the Large Hadron Collider: the s-channel. Its interest relies in the confirmationof the Standard Model predictions but, as well, in the possibility of constraining severalnew physics scenarios when comparing the s-channel cross section with the oneof the other single top production modes (the Wt and t-channel). After a generalintroduction upon top physics and the experimental setup employed for the detectionand the reconstruction of the physics objects, we present two analyses realizedwith proton-proton collisions collected by the ATLAS detector at a center of massenergy of 7 and 8 TeV. Since the s-channel production is characterized by a verylow purity and, at the same time, is not favoured at the LHC because it proceedsvia quark-antiquark annihilation, a multivariate approach is applied in both cases todiscriminate the signal. For the 7 TeV study, based on an integrated luminosity of4.7 fb−1, two boosted decision trees classifiers are optimized against the main sourcesof background; a cut realized on one of such discriminants implements a tight eventselection, after which the second output distribution is used for a maximum likelihoodfit to the data. This allows to set an upper limit on the s-channel cross section of21.5 pb (14.3 pb expected), corresponding to a significance of 0.6 standard deviations(0.8 expected). The 8 TeV analysis, based on an integrated luminosity of 20.3 fb−1,has a simpler strategy: only one boosted decision trees classifier is trained to isolatesignal after a preliminary cut-based selection; its output distribution is then fitted tothe data with the same approach than at 7 TeV, taking into account the systematicand statistical uncertainty via the generation of pseudoexperiments. The significanceof the s-channel measurement being lower than 3 σ (1.3 standard deviation observed,1.4 expected), only a 95% C.L. limit can be set on the cross section; the result ishowever improved and corresponds to 14.6 pb (7.9 pb expected).3

Diese Arbeit präsentiert die Messung der Produktion eines einzelnen top quarks im s-channel. Dazu wird der 13 TeV ATLAS Datensatz, welcher von 2015 bis 2018 aufgenommen wurde und insgesamt 139 fb^−1 integrierte Luminositat hat, analysiert. Die große Schwierigkeit dieser Analyse ist die überwältigende Menge an Untergrund. Nach einer optimisierten Auswahl von Ereignissen stammen nur 3 % der Ereignisse von dem gesuchten Prozess. Demzufolge muss der Untergrund so gut wie möglich von dem Signal getrennt werden. In dieser Analyse wird die Matrixelement Methode verwendet. Diese Methode wendet das Neyman-Pearson Lemma, das die optimale Teststatistik beschreibt, auf dieses Teilchenphysik Experiment an. Die Likelihood Funktionen werden durch eine Faltung der Matrixelemente, die die harte Streuung beschreiben, mit den Transferfunktionen, die für die Modelierung der Detektoreffekte zuständig sind, berechnet. Wegen begrenzter Rechenkapazitäten kann dabei nur die erste Ordnung der Näherung in die Berechnung einbezogen werden. Diese Arbeit stellt außerdem Studien vor, die das Ziel haben diese Methode zu verbessern. Dazu wurden die Transferfunktion neu bestimmt. Außerdem wird ein Neuronales Netzwerk zur Identifizierung des Signals angewendet. Um das Resultat zu extrahieren wurde eine statistisches Modell der diskriminierenden Variable der Matrixelement Methode erstellt und mit Hilfe des profile likelihood Fits wurde der Parameter μ, der als Quotient aus dem gemessenen s-channel Wirkungsquerschnitt zu dem vom Standard Modell vorhergesagten Wirkungsquerschnitt definiert ist, bestimmt. Das Resultat der Analyse ist μ = 1.28 ^+0.36_− 0.31. Dies stimmt mit der Standard Modell Vorhersage überein. Die gemessene Signifikanz beträgt 4.19 σ. Dies ist die höchste jemals gemessene Signifikanz für die s-channel Produktion eines einzelnen top quarks in Proton-Proton Kollisionen. / This thesis presents the measurement of the single top s-channel production with the ATLAS detector. The analysis is based on data collected between 2015 and 2018, at a center-of-mass energy of 13 TeV and with an integrated luminosity of 139 fb^−1. The main challenge of this analysis is the overwhelming background. After an optimized selection of events the signal amounts to only approximately 3 % of all events in the signal region. This necessitates the best possible separation of signal and background. In this thesis, the Matrix Element is utilized. The underlying principle for this method is the Neyman-Pearson lemma which states the optimal test statistic. The likelihoods are computed by a convolution of the matrix elements describing the hard scattering process and the transfer functions which model the detector effects. Due to limited computing resources, only the leading order approximation can be calculated. This thesis further investigates methodologies aimed at enhancing the effectiveness of the signal separation process. For that reason, the transfer functions are updated and a study in which a neural network uses the event likelihoods of the Matrix Element Method as in- put to produce a different discriminating variable is shown. To extract the final result, a statistical analysis of the discriminating variable result- ing from the Matrix Element Method is preformed. A statistical model is built and the profile likelihood fit is used to extract the parameter of interest μ which is the ratio of the observed s-channel single top production cross-section to the Standard Model expectation. The result of the analysis is μ = 1.28 ^+0.36_− 0.31 which is in agreement with the Standard Model. The observed significance of this analysis is 4.19 σ. This is the highest measured significance of the s-channel single top production in proton-proton collisions.