Atmospheric haze, which builds up over South Africa including our study areas, Cape Town and the Mpumalanga Highveld under calm weather conditions, causes public concern. The scope of this study was to determine the concentration and composition of atmospheric aerosol at Khayelitsha (an urban site in the Western Cape) and Ferrobank (an industrial site in Witbank, Mpumalanga). Particulate matter was collected in Khayelitsha from 18 May 2007 to 20 July 2007 (i.e. 20 samples) using a Partisol-plus sampler and a Tapered Element Oscillating Microbalance (TEOM) sampler. Sampling took place at Ferrobank from 07 February 2008 to 11 March 2008 (6 samples) using a Partisol-plus sampler and an E-sampler. The gravimetric mass of each exposed sample was determined from pre- and post-sampling weighing. The elemental composition of the particulate matter was determined for 16 elements at Khayelitsha using Proton Induced X-ray Emission (PIXE). The concentration of the elements Al, Si, S, Cl, K, Ca, Ti, Cr, Mn, Fe, Cu, Zn, As, Br, Sn, and Pb was determined by analysing the PIXE spectra obtained. In similar manner, the elemental composition of the particulate matter was determined for 15 elements at Ferrobank (Al, Si, S, Cl, K, Ca, Ti, Cr, Mn, Fe, Cu, Zn, As, Br and Pb). The average aerosol mass concentrations for different days at the Khayelitsha site were found to vary between 8.5 μg/m3 and 124.38 μg/m3. At the Khayelitsha site on three occasions during the sampling campaign the average aerosol mass concentrations exceeded the current South African air quality standard of 75 μg/m3 over 24 h. At the Ferrobank site, there are no single days that exceeded the limit of the South African air quality standard during the sampling campaign. Enrichment factors for each element of the particles sampled with an aerodynamic diameter of less than 10 μm (PM10) samples have been calculated in order to identify their possible sources. The analysis yielded five potential sources of PM10 : soil dust, sea salt, gasoline emissions, domestic wood and coal combustion. Interestingly, enrichment factor values for the Khayelitsha samples show that sea salt constitutes a major source of emissions, while Ferrobank samples, the source apportionment by unique ratios (SPUR) indicate soil dust and coal emission are the major sources of pollution. The source apportionment at Khayelitsha shows that sea salt and biomass burning are major source of air pollution.

O presente trabalho envolveu o projeto e construção de um amostrador seqüencial automático de chuva considerando a deposição úmida. Para isso, o amostrador seqüencial fornece um volume escolhido em função de análises químicas previamente estabelecidas. O amostrador seqüencial faz a coleta da deposição úmida, que é a chuva precipitada sobre um coletor inicialmente tampado. Este procedimento adotado é de fundamental importância para a avaliação das condições de poluição presentes na atmosfera. Através deste amostrador seqüencial, também é possível determinar a intensidade da chuva para o evento considerado, ou seja, é possível avaliar se a chuva precipitada é forte ou fraca. O planejamento do amostrador seqüencial automático de chuva foi baseado em diversos amostradores encontrados na literatura, os quais visavam o monitoramento para diferentes objetivos. O processo de desenvolvimento e construção foi bem sucedido, uma vez que todas as fases foram executadas de modo que o processo de coleta das amostras ocorreu com sucesso. Foram coletados alguns ventos de chuva possibilitando estimar as respectivas intensidades pluviométricas, ou seja, avaliar para o campo considerado qual foi o tipo de chuva, se foi forte ou fraca. Para uma coleta específica, o amostrador foi exposto a um evento de chuva muito intenso, funcionando conforme o esperado. Os registros foram 290 ml de volume de amostra, coletados em uma área de 1230 cm2, em um tempo de 1,12 min apresentando intensidade de 2,1 mm/min. Esse resultado indica um valor muito alto de intensidade. / The present work involved the design and construction of an automatic sequential rainfall sampler, with the goal of evaluating the quality of rainfall water. For such, the sequential sampler supplies a chosen volume taking into account the chemical analysis previously defined. The sequential sampler performs the collection of humid deposition, which is the precipitated rain that falls on a collector initially covered. This adopted procedure is of fundamental importance for the evaluation of the pollution conditions present in the atmosphere. Through this sequential sampler, it is also possible to define the rainfall intensity for the event considered, i.e., it is possible to evaluate if the precipitated rain is heavy or light. The planning of the automatic sequential rainfall sampler was based on several samplers found in the literature, which \'aimed the monitoring due to different goals. The design and construction processes were successful, once all the phases were performed and the sample collection process occurred with success. Some events were collected and their respective pluviometric intensities were measured, i.e., evaluate for the considered time which was the rain type, if it was heavy or light. For a specific collection the sampler was exposed to a very intense event of rain, working as expected. The records were 290 of sample volume, collected in an area of 1230 cm2 during a period of 1.12 min, with intensity of 2.1 mm/min. This result indicate a value of very high intensity.

No presente trabalho apresentamos o algoritmo de simulacão perfeita CFTP, proposto em Propp & Wilson (1996). Seguindo o trabalho de Philippe & Robert (2003) implementamos o CFTP gerando amostras da distribuicão normal bivariada truncada no quadrante positivo. O algoritmo proposto e comparado com o amostrador de Gibbs e o método de rejeição. Finalmente, apresentamos sugestões para a implementação do CFTP para gerar amostras da distribuição normal truncada em dimensões maiores que dois e a geração de amostras em conjuntos diferente do quadrante positivo. / This project will display the CFTP perfect simulation algorithm presented at Propp & Wilson (1996). According to Philippe & Robert (2003) will be implemented the CFTP providing samples of the bivariate normal distribution truncated at the positive quadrant. The proposed algorithm is compared to the samples generated by Gibbs Sampler and by the rejection sampling ( or acceptance rejection method or \"accept-reject algorithm\"). Finally, suggestions to the implementation of CFTP in order to produce truncated normal distribution samples at bigger dimensions than two and the provide a diferent set of samples from the positive quadrant.

The No-U-Turn Sampler (NUTS) is a relatively new Markov chain Monte Carlo (MCMC) algorithm that avoids the random walk behavior that common MCMC algorithms such as Gibbs sampling or Metropolis Hastings usually exhibit. Given the fact that NUTS can efficiently explore the entire space of the target distribution, the sampler converges to high-dimensional target distributions more quickly than other MCMC algorithms and is hence less computational expensive. The focus of this study is on applying NUTS to one of the complex IRT models, specifically the two-parameter mixture IRT (Mix2PL) model, and further to examine its performance in estimating model parameters when sample size, test length, and number of latent classes are manipulated. The results indicate that overall, NUTS performs well in recovering model parameters. However, the recovery of the class membership of individual persons is not satisfactory for the three-class conditions. Also, the results indicate that WAIC performs better than LOO in recovering the number of latent classes, in terms of the proportion of the time the correct model was selected as the best fitting model. However, when the effective number of parameters was also considered in selecting the best fitting model, both fully Bayesian fit indices perform equally well. In addition, the results suggest that when multiple latent classes exist, using either fully Bayesian fit indices (WAIC or LOO) would not select the conventional IRT model. On the other hand, when all examinees came from a single unified population, fitting MixIRT models using NUTS causes problems in convergence.

The analysis of big data is of great interest today, and this comes with challenges of improving precision and efficiency in estimation and prediction. We study binary data with covariates from numerous small areas, where direct estimation is not reliable, and there is a need to borrow strength from the ensemble. This is generally done using Bayesian logistic regression, but because there are numerous small areas, the exact computation for the logistic regression model becomes challenging. Therefore, we develop an integrated multivariate normal approximation (IMNA) method for binary data with covariates within the Bayesian paradigm, and this procedure is assisted by the empirical logistic transform. Our main goal is to provide the theory of IMNA and to show that it is many times faster than the exact logistic regression method with almost the same accuracy. We apply the IMNA method to the health status binary data (excellent health or otherwise) from the Nepal Living Standards Survey with more than 60,000 households (small areas). We estimate the proportion of Nepalese in excellent health condition for each household. For these data IMNA gives estimates of the household proportions as precise as those from the logistic regression model and it is more than fifty times faster (20 seconds versus 1,066 seconds), and clearly this gain is transferable to bigger data problems.

The eCommerce industry introduced new business principles, as well as new strategies for achieving these principles, and as a result some traditional measures of success are no longer valid. We classified and ranked the performance of twenty business-to-consumer eCommerce companies by developing critical benchmarks using the Balanced scorecard methodology. We applied a Latent class model, a statistical model along the Bayesian framework, to facilitate the determination of the best and worst performing companies. An eCommerce site's greatest asset is its customers, which is why some of the most valued and sophisticated metrics used today evolve around customer behavior. The results from our classification and ranking procedure showed that companies that ranked high overall also ranked comparatively well in the customer analysis ranking, For example, Amazon.com, one of the highest rated eCommerce companies with a large customer base ranked second in the critical benchmark developed towards measuring customer analysis. The results from our simulation also showed that the Latent class model is a good fit for the classification procedure, and it has a high classification rate for the worst and best performing companies. The resulting work offers a practical tool with the ability to identify profitable investment opportunities for financial managers and analysts.

L’utilisation répétée de la bouillie bordelaise pour la lutte contre le mildiou a entrainé une accumulation de cuivre dans les sols viticoles. Afin d’évaluer l’impact de cette pollution, les capteurs passifs sont fréquemment utilisés pour renseigner sur la biodisponibilité du cuivre et donc sur sa toxicité. Toutefois, au vu de l’ensemble des principes de diffusion qui régissent l’échantillonnage par capteur passif, plusieurs modifications, notamment sur la géométrie, ont pu être apportées et constituent l’objet de cette étude au travers du développement de nouveaux capteurs passifs : les DMG (Diffusive MilliGel). Leur synthèse par un système millifluidique leur confère une géométrie unique, constituée de billes ellipsoïdales de dimension millimétrique. Leurs paramètres physico-chimiques ont été caractérisés. Différents paramètres tels que le temps d’exposition et les méthodes d’analyses ont été optimisés. Les DMG ont été testés au cours d’applications environnementales et leur aptitude à échantillonner la fraction labile du cuivre a été mise en évidence. Une première étude dans des eaux douces a permis de mettre en évidence l’influence de la matière organique sur la fraction retenue par DMG, ainsi que l’aptitude de ce capteur à évaluer la toxicité du cuivre. Pour cela, un organisme modèle, Ceriodaphnia dubia, a été exposé dans les mêmes conditions au cours d’une étude écotoxicologique. La comparaison entre les DMG, les DGT et une méthode par colonne de chelex a mis en évidence les avantages qu’offre ce capteur par rapport aux autres techniques. Enfin, une série d’expériences sur les sols viticoles a révélé les aptitudes des DMG à échantillonner du cuivre dans cette matrice complexe. / Repeated use of copper in Bordeaux mixture to fight against mildew has led to an accumulation of this metal in the vineyard soils. To assess the impact of this pollution, the quantification of this element in the soil is essential and more precisely the quantification of the bioavailable fraction which give an indication of the copper toxicity. For that, the passive samplers appear as reliable and efficient tools. However, due to diffusion principles which control the uptake by passive samplers, an optimization of the geometry seems to be necessary and induced this study through the development of a new passive sampler: the DMG (Diffusive MilliGels). The synthesis by millifluidic process confers a unique geometry; DMG are composed of ellipsoidal beads of around 1 mm diameter. First of all, their physico-chemical parameters were characterized and analytical methods and the exposure time were optimized as well. DMG were then tested for an environmental application and their ability to sample the labile fraction of copper was thus demonstrated. Namely, analysis of freshwater by DMG allowed highlighting the influence of organic matter on the fraction retained by this passive sampler and their ability to evaluate the copper toxicity was confirmed. For that, a living organisms Ceriodaphnia dubia was exposed in the same condition during a comparative ecotoxicological study. Moreover, the comparison between DMG, DGT and a Chelex column method showed the advantages of our developed method particularly for in situ application. Finally, a series of experiments on vineyard soils was performed to reveal the abilities of DMG in copper sampling in this complex matrix.

No presente trabalho apresentamos o algoritmo de simulacão perfeita CFTP, proposto em Propp & Wilson (1996). Seguindo o trabalho de Philippe & Robert (2003) implementamos o CFTP gerando amostras da distribuicão normal bivariada truncada no quadrante positivo. O algoritmo proposto e comparado com o amostrador de Gibbs e o método de rejeição. Finalmente, apresentamos sugestões para a implementação do CFTP para gerar amostras da distribuição normal truncada em dimensões maiores que dois e a geração de amostras em conjuntos diferente do quadrante positivo. / This project will display the CFTP perfect simulation algorithm presented at Propp & Wilson (1996). According to Philippe & Robert (2003) will be implemented the CFTP providing samples of the bivariate normal distribution truncated at the positive quadrant. The proposed algorithm is compared to the samples generated by Gibbs Sampler and by the rejection sampling ( or acceptance rejection method or \"accept-reject algorithm\"). Finally, suggestions to the implementation of CFTP in order to produce truncated normal distribution samples at bigger dimensions than two and the provide a diferent set of samples from the positive quadrant.

O presente trabalho envolveu o projeto e construção de um amostrador seqüencial automático de chuva considerando a deposição úmida. Para isso, o amostrador seqüencial fornece um volume escolhido em função de análises químicas previamente estabelecidas. O amostrador seqüencial faz a coleta da deposição úmida, que é a chuva precipitada sobre um coletor inicialmente tampado. Este procedimento adotado é de fundamental importância para a avaliação das condições de poluição presentes na atmosfera. Através deste amostrador seqüencial, também é possível determinar a intensidade da chuva para o evento considerado, ou seja, é possível avaliar se a chuva precipitada é forte ou fraca. O planejamento do amostrador seqüencial automático de chuva foi baseado em diversos amostradores encontrados na literatura, os quais visavam o monitoramento para diferentes objetivos. O processo de desenvolvimento e construção foi bem sucedido, uma vez que todas as fases foram executadas de modo que o processo de coleta das amostras ocorreu com sucesso. Foram coletados alguns ventos de chuva possibilitando estimar as respectivas intensidades pluviométricas, ou seja, avaliar para o campo considerado qual foi o tipo de chuva, se foi forte ou fraca. Para uma coleta específica, o amostrador foi exposto a um evento de chuva muito intenso, funcionando conforme o esperado. Os registros foram 290 ml de volume de amostra, coletados em uma área de 1230 cm2, em um tempo de 1,12 min apresentando intensidade de 2,1 mm/min. Esse resultado indica um valor muito alto de intensidade. / The present work involved the design and construction of an automatic sequential rainfall sampler, with the goal of evaluating the quality of rainfall water. For such, the sequential sampler supplies a chosen volume taking into account the chemical analysis previously defined. The sequential sampler performs the collection of humid deposition, which is the precipitated rain that falls on a collector initially covered. This adopted procedure is of fundamental importance for the evaluation of the pollution conditions present in the atmosphere. Through this sequential sampler, it is also possible to define the rainfall intensity for the event considered, i.e., it is possible to evaluate if the precipitated rain is heavy or light. The planning of the automatic sequential rainfall sampler was based on several samplers found in the literature, which \'aimed the monitoring due to different goals. The design and construction processes were successful, once all the phases were performed and the sample collection process occurred with success. Some events were collected and their respective pluviometric intensities were measured, i.e., evaluate for the considered time which was the rain type, if it was heavy or light. For a specific collection the sampler was exposed to a very intense event of rain, working as expected. The records were 290 of sample volume, collected in an area of 1230 cm2 during a period of 1.12 min, with intensity of 2.1 mm/min. This result indicate a value of very high intensity.

Historical change in the geomorphology, hydrodynamics, and surficial sediment of the tidal delta system of Tauranga Harbour are investigated with the general aim of analysing The general aims of this thesis are: firstly to analyse historical changes to inlet delta system geomorphology using historical hydrographic charts, secondly, to conduct hydrodynamic numerical modelling using historical bathymetries to access changes in peak spring flow and potential net tidal sediment transport, and thirdly, to analyse historical changes in surficial sediment and bedforms. Geomorphic change was investigated through plotting difference in bathymetry graphs and conducting cross sections taken from digisitied bathymetries obtained from historical hydrographic charts from 1852, 1879, 1901, 1927, 1954 and a modern bathymetry from 2006. Two-dimensional hydrodynamic numerical modelling was conducted to investigate the changes in peak tidal current flow and potential net sediment transport between 1852 and 2006. Changes in surficial sediment patterns were determined through completing a side scan sonar survey with associated sediment samples for ground truthing of grain size and underwater videography to gather surficial shell coverage information. This was used to produce a surficial sediment coverage map which was compared to historical studies Major geomorphological findings include that the shipping channel appears to have induced minor change in the geomorphology of the FTD but such changes are similar to those identified in the historical bathymetries of 1852, 1879, 1901, 1927, 1954 prior to dredging. Significant changes have occurred on the ETD, with the majority of the ETD showing scour of 1 m while the terminal lobe has extended seawards. This is associated with historical (since 1852) narrowing of the inlet from Panepane Point to Mt Maunganui by ~ 900 m. Hydrodynamic numerical modelling has shown a significant increase in potential net tidal sediment transport in the Cutter Channel due to dredging, while the Maunganui Roads Channel shows a reduction of net potential tidal sediment transport that is associated with the dredging of this channel. The area surrounding Panepane Point undergoes significant increases and decreases in net potential tidal sediment transport both before and after dredging Investigation of the surficial sediment patterns over the FTD and ETD from sidescan sonar and bottom samples show that between 1983 and 2007 there has been a northwards extension of the area of major shell (greater than 50 %) converge in the main ebb channel as well as reduction in major shell converge in flood tidal delta ebb shield region. The Maunganui Roads Channel changes from sitly sands to medium and fine sands.