Conditional auto-regressive (CAR) models are frequently used with spatial data. However, the likelihood of such a model is expensive to compute even for a moderately sized data set of around 1000 sites. For models involving latent variables, the likelihood is not usually available in closed form. In this thesis we use a Monte Carlo approximation to the likelihood (extending the approach of Geyer and Thompson (1992)), and develop two strategies for maximising this. One strategy is to limit the step size by defining an experimental region using a Monte Carlo approximation to the variance of the estimates. The other is to use response surface methodology. The iterative procedures are fully automatic, with user-specified options to control the simulation and convergence criteria. Both strategies are implemented in our R package mclcar. We demonstrate aspects of the algorithms on simulated data on a torus, and achieve similar results to others in a short computational time on two datasets from the literature. We then use the methods on a challenging problem concerning forest restoration with data from around 7000 trees arranged in transects within study plots. We modelled the growth rate of the trees by a linear mixed effects model with CAR spatial error and CAR random e ects for study plots in an acceptable computational time. Our proposed methods can be used for similar models to provide a clearly defined framework for maximising Monte Carlo approximations to likelihoods and reconstructing likelihood surfaces near the maximum.

The aim of this research was to assess three hierarchical aspects of alga-grazer interactions in intertidal communities on a small scale: spatial heterogeneity, grazing effects and spatial stability in grazing effects. First, using semivariograms and cross-semivariograms I observed hierarchical spatial patterns in most algal groups and in grazers. However, these patterns varied with the level on the shore and between shores, suggesting that either human exploitation or wave exposure can be a source of variability. Second, grazing effects were studied using manipulative experiments at different levels on the shore. These revealed significant effects of grazing on the low shore and in tidal pools. Additionally, using a transect of grazer exclusions across the shore, I observed unexpected hierarchical patchiness in the strength of grazing, rather than zonation in its effects. This patchiness varied in time due to different biotic and abiotic factors. In a separate experiment, the effect of mesograzers effects were studied in the upper eulittoral zone under four conditions: burnt open rock (BOR), burnt pools (Bpool), non-burnt open rock (NBOR) and non-burnt pools (NBpool). Additionally, I tested spatial stability in the effects of grazing in consecutive years, using the same plots. I observed great spatial variability in the effects of grazing, but this variability was spatially stable in Bpools and NBOR, meaning deterministic and significant grazing effects in consecutive years on the same plots. Both the significance in grazing effects and spatial stability depended on the level of resolution (species, functional, biomass) at which the algal assemblage was evaluated, suggesting hierarchical variability. In order to be able to predict spatial variability in the effects of grazers in the upper eulittoral zone using biotic and abiotic micro- and macrofactors, a conceptual model was proposed, based on data from several multiple-regressions. This linked the interactions among three elements: idiosyncratic heterogeneity, micro and macrofactors. This suggests that spatial variability can be a product of these factors, while spatial stability can be caused by the same or different combinations of factors. In conclusion, grazing and other ecological phenomena must be studied hierarchically, not only through spatiotemporal scales, but also at different levels of resolution, as these also influence our perception of patterns.

Made available in DSpace on 2016-06-07T17:12:15Z (GMT). No. of bitstreams: 0 Previous issue date: 2015-05-21. Added 1 bitstream(s) on 2016-06-07T17:16:51Z : No. of bitstreams: 1 000864053.pdf: 3409653 bytes, checksum: 79a04fc4e3a4aa2ac8982655c947c380 (MD5) / O objetivo deste estudo foi Identificar o padrão da distribuição espacial da hanseníase no município de Rondonópolis e sua relação com a situação clínico - epidemiológica da doença nos anos de 2000 a 2010. Foi utilizado um Sistema de Informação Geográfica (SIG), o Arcgis 9.3, para determinar a localização espacial dos 1.832 casos novos e elaborar os mapas temáticos. Utilizou-se o programa MapInfo, versão 7.8, para geocodificação dos casos. A análise de agrupamento foi desenvolvida a partir da metodologia de análise espacial, a estatística de varredura (Scan). Foram utilizadas as bases cartográficas e os dados populacionais dos censos de 2000 e 2010 do IBGE. Portanto, as análises dos dados foram dividas em dois períodos, sendo de 2000 a 2005 e 2006 a 2010. Para a elaboração dos mapas com agrupamentos espaciais de casos novos de hanseníase paucibacilar, multibacilar, em menores de 15 anos e na população total, foram utilizados a base cartográfica digital de endereços do município de Rondonópolis e o software ARCGIS 9.3. Os indicadores socioeconômicos foram mapeados segundo a técnica cartográfica coroplética. A partir da análise desses dados, observou-se em todas as populações estratificadas redução do coeficiente de detecção geral, aumento dos agrupamentos espaciais de alto risco em todas as populações estratificadas, acentuada alteração na distribuição dos agrupamentos de alto e baixo risco, detecção do agrupamento de alto risco em menores de 15 anos no período de 2006 a 2010, sinalizando doença recente e presença de foco ativo, sobreposição do agrupamento de alto risco da população de multibacilar e menor de 15 anos. As conclusões mostram que em Rondonópolis a hanseníase continua sendo um problema de saúde pública em todo o município, apresentando áreas espaciais de alto risco, que demandam intensificar ainda mais as ações e estratégias prioritárias de busca ativa para... / The objective of this study was to identify a pattern of spatial distribution of leprosy cases from 2000 to 2010, in the Rondonópolis municipality and its relationship with clinical epidemiological aspects of the disease. The Geographic Information System (GIS), ArcGIS 9.3, was used to determine the spatial location of 1,832 new cases and prepare thematic maps. MapInfo program, version 7.8, was used for geocoding the leprosy cases. Cluster analysis was carried out using the spatial analysis methodology, scan statistics (Scan). The cartographic databases and population census data of 2000 and 2010 from IBGE were used. Therefore, the analysis was performed in two periods, 2000 to 2005 and 2006 to 2010. The digital cartographic database of Rondonópolis addresses and the software ArcGIS 9.3 were used in order to build the new cases spatial clusters maps stratifying the cases as paucibacillary leprosy, multibacillary leprosy and disease in children under 15 years and the total population. Socioeconomic indicators were mapped according to coropletic cartographic technique. The analysis in all stratified populations revealed an overall reduction in the detection rate and increase in high-risk spatial clusters in all stratified populations. In addition a marked change was observed in the distribution of high and low risk groups and in detection high risk cluster in children under 15 years in the period of 2006 to 2010, signaling recent illness and presence of active focus of disease, besides an overlap of high-risk group of the multibacillary population with disease in children under 15 years. The findings show that leprosy remains a public health problem in the entire municipality, with demonstration of spatial high-risk areas that evidence the need of intensifying actions and definition of priority strategies for active search of leprosy cases

Orientador: Joel Carlos Lastória / Coorientador: Lígia Vizeu Barrozo / Banca: Maria Stella de Mello Ayres Putinatti / Banca: Hélio Amante Miot / Banca: Susilene Maria Tonelli Nardi / Banca: Ida Foschiani Dias Baptista / Resumo: O objetivo deste estudo foi Identificar o padrão da distribuição espacial da hanseníase no município de Rondonópolis e sua relação com a situação clínico - epidemiológica da doença nos anos de 2000 a 2010. Foi utilizado um Sistema de Informação Geográfica (SIG), o Arcgis 9.3, para determinar a localização espacial dos 1.832 casos novos e elaborar os mapas temáticos. Utilizou-se o programa MapInfo, versão 7.8, para geocodificação dos casos. A análise de agrupamento foi desenvolvida a partir da metodologia de análise espacial, a estatística de varredura (Scan). Foram utilizadas as bases cartográficas e os dados populacionais dos censos de 2000 e 2010 do IBGE. Portanto, as análises dos dados foram dividas em dois períodos, sendo de 2000 a 2005 e 2006 a 2010. Para a elaboração dos mapas com agrupamentos espaciais de casos novos de hanseníase paucibacilar, multibacilar, em menores de 15 anos e na população total, foram utilizados a base cartográfica digital de endereços do município de Rondonópolis e o software ARCGIS 9.3. Os indicadores socioeconômicos foram mapeados segundo a técnica cartográfica coroplética. A partir da análise desses dados, observou-se em todas as populações estratificadas redução do coeficiente de detecção geral, aumento dos agrupamentos espaciais de alto risco em todas as populações estratificadas, acentuada alteração na distribuição dos agrupamentos de alto e baixo risco, detecção do agrupamento de alto risco em menores de 15 anos no período de 2006 a 2010, sinalizando doença recente e presença de foco ativo, sobreposição do agrupamento de alto risco da população de multibacilar e menor de 15 anos. As conclusões mostram que em Rondonópolis a hanseníase continua sendo um problema de saúde pública em todo o município, apresentando áreas espaciais de alto risco, que demandam intensificar ainda mais as ações e estratégias prioritárias de busca ativa para... / Abstract: The objective of this study was to identify a pattern of spatial distribution of leprosy cases from 2000 to 2010, in the Rondonópolis municipality and its relationship with clinical epidemiological aspects of the disease. The Geographic Information System (GIS), ArcGIS 9.3, was used to determine the spatial location of 1,832 new cases and prepare thematic maps. MapInfo program, version 7.8, was used for geocoding the leprosy cases. Cluster analysis was carried out using the spatial analysis methodology, scan statistics (Scan). The cartographic databases and population census data of 2000 and 2010 from IBGE were used. Therefore, the analysis was performed in two periods, 2000 to 2005 and 2006 to 2010. The digital cartographic database of Rondonópolis addresses and the software ArcGIS 9.3 were used in order to build the new cases spatial clusters maps stratifying the cases as paucibacillary leprosy, multibacillary leprosy and disease in children under 15 years and the total population. Socioeconomic indicators were mapped according to coropletic cartographic technique. The analysis in all stratified populations revealed an overall reduction in the detection rate and increase in high-risk spatial clusters in all stratified populations. In addition a marked change was observed in the distribution of high and low risk groups and in detection high risk cluster in children under 15 years in the period of 2006 to 2010, signaling recent illness and presence of active focus of disease, besides an overlap of high-risk group of the multibacillary population with disease in children under 15 years. The findings show that leprosy remains a public health problem in the entire municipality, with demonstration of spatial high-risk areas that evidence the need of intensifying actions and definition of priority strategies for active search of leprosy cases / Doutor

Orientadores: Lucí Hidalgo Nunes, Armando Zaupa Remacre, Michael James Friedel / Tese (doutorado) - Universidade Estadual de Campinas, Instituto de Geociências / Made available in DSpace on 2018-08-17T10:14:48Z (GMT). No. of bitstreams: 1 Koga-Vicente_Andrea_D.pdf: 5761562 bytes, checksum: 071a6fa472f20045e34a37442cb5d89b (MD5) Previous issue date: 2010 / Resumo: A incerteza é inerente ao funcionamento do sistema hidrometeorológico. Neste trabalho buscou-se incorporar esse conceito no estudo sobre as precipitações que causam impactos nas regiões da Baixada Santista e Litoral Norte paulista. Considerando que as incertezas epistêmicas podem ser diminuídas a partir do melhor conhecimento sobre o fenômeno, foram realizadas investigações sobre as chuvas que deflagraram impactos, e admitindo que as incertezas aleatórias possam ser minimizadas por meio de ferramentas que demonstrem possíveis cenários para o desenrolar do evento, foram elaborados mapas de probabilidade, de desvio padrão e propostas modelagens preditivas. Com base em informações levantadas sobre os impactos associados à chuva que ocorreram em doze municípios entre 1994/95 e 2003/04, foi investigada a distribuição espacial e temporal dos registros e da pluviometria associada. Verificou-se que os eventos pluviais que causaram problemas são compostos por chuvas contínuas com um pico mais intenso, sendo que em 1999/00 e 1995/96 houve maior ocorrência de impactos registrados. Embora os maiores totais pluviométricos tenham ocorrido em Bertioga, os municípios de Santos e Guarujá apresentaram mais problemas; por outro lado, o maior número de vítimas foi registrado em Ubatuba, indicando que embora a vulnerabilidade seja o cerne do impacto, a suscetibilidade determinou a magnitude das pessoas afetadas. A partir da simulação estocástica foram elaborados cenários de distribuição espacial das precipitações com e sem a incorporação de variáveis físicas relacionadas, resultando em mapas que apontam locais com maior probabilidade de serem atingidos por chuvas acima de 80,0 mm/24h durante episódios pluviais impactantes e de mapas que representaram a incerteza associada à variabilidade espacial, fortemente associada à escarpa da Serra do Mar. Estes mapas podem ser usados como ferramenta no planejamento regional, bem como os modelos de predição de impactos propostos a partir da utilização de duas abordagens: (1) o Mapa Auto-Organizável, uma rede neural artificial não supervisionada, indicada para relações não-lineares e (2) uma técnica empírica de Regressão Linear Múltipla, tradicional em estudos preditivos. Os modelos propostos apresentaram entre 71% e 82% de acurácia na predição de ocorrências de impactos deflagrados por precipitações / Abstract: Uncertainty is part of the hydrometeorological system working. This research looked for ways to incorporate this concept in the study about the precipitations that cause impacts in Baixada Santista area and the North Coast of São Paulo state. Regarding that the epistemic uncertainties can be decreased when better knowledge about the phenomenon is acquired, investigations about the rain which cause impacts were made, and assuming that the random uncertainties can be decreased by using tools that show possible scenarios to develop the event, maps of probability and standard deviation were made, and also predictive modelings were proposed. Based on information raised about the impacts associated to rain that happened in twelve municipalities between 1994/95 and 2003/04, the spatial and time record distribution and the associated rainfall were studied. The rainfall events that had caused problems were mostly compound of continuous rain with a stronger peak, being in 1999/00 and 1995/96 the largest impacts recorded. Although the greatest total rainfall amounts happened in the city of Bertioga, major impacts were recorded in Santos and Guarujá. However, most victims were recorded in the city of Ubatuba showing that, although vulnerability is the core of the impact, the susceptibility determined the magnitude of people affected. By applying the stochastic simulation, scenarios of spatial distribution of precipitations were created with and without the incorporation of related physical variables. It resulted in maps that show locations with great probability of being reached by rain over 80.0 mm/24h during impacting rainfall episodes and maps that represent the uncertainty associated to the spatial variability strongly associated to the Serra do Mar escarpment. These maps can be used as tools to a regional planning, as well as the prediction models of impacts proposed from the use of two approaches: (1) the Self-Organized Map, an artificial neural net unsupervised, indicated to non-linear relations, and (2) an empirical technique of Multiple Linear Regression, traditional in predictive studies. The proposed models showed between 71% and 82% of accuracy in predicting impact caused by rainfalls / Doutorado / Análise Ambiental e Dinâmica Territorial / Doutor em Ciências

Orientador: Marcos Cesar Ferreira / Tese (doutorado) - Universidade Estadual de Campinas, Instituto de Geociencias / Made available in DSpace on 2018-08-13T02:47:13Z (GMT). No. of bitstreams: 1 Gomes_PriscilaBrochado_D.pdf: 6950162 bytes, checksum: fc83241a2e03189e1a609b95c290e928 (MD5) Previous issue date: 2009 / Resumo: O surgimento de sensores de alta resolução espacial permitiu que o nível de detalhamento das imagens de satélite chegasse ao ponto em que estudos de ecologia florestal pudessem ser realizados, através deste instrumento, com base em informações de espécies e indivíduos arbóreos. Tal tema de pesquisa é um campo promissor, já que a possibilidade de se estudar árvores de uma floresta à distância, pode contribuir para maior agilidade e menor custo do manejo florestal, minimizando custos de boa parte dos levantamentos de campo. Os objetivos deste trabalho foram investigar alguns procedimentos metodológicos utilizando imagens do satélite Quickbird, combinadas a dados de campo e a ferramentas computacionais para detecção, identificação e mapeamento de árvores de uma floresta semidecidua e de um cerradão no estado de São Paulo. O método de classificação orientada a objetos foi adequado para a identificação das espécies Sclerolobium paniculatum e Mabea fistulifera nas imagens Quickbird. A baixa resolução espectral destas imagens pode ser compensada pelo uso de informações espaciais dos indivíduos de interesse na classificação orientada a objetos. O método utilizado foi apropriado para a separação entre árvores decíduas e perenefólias da floresta semidecidua, porém, para esta formação florestal, a identificação de espécies não foi possível. Índices de vegetação obtidos através da álgebra de bandas de imagens Quickbird e Aster se mostraram boas fontes de informação para a elaboração de mapas de distribuição potencial de espécies arbóreas. Estes resultados indicam que, dependendo da formação florestal estudada e da espécie de interesse, a resolução espacial do satélite Quickbird pode ser satisfatória para a identificação de espécies arbóreas e que essas imagens têm grande aplicação em estudos biogeográficos. / Abstract: New high resolution satellite images, like Ikonos and Quickbird, provide great potential to monitor and study trees of tropical forest through species identification and mapping. This topic of research is a promising field since remote sensing has potential to provide, at lower cost, forest information with great coverage than is attainable using field sampling. The objective of this thesis was to investigate some methodological proceedings using Quickbird images, combined with field data and informatics tools, to detection, identification and mapping of trees from a cerradão (Savanna Forest) and a semi-deciduous seasonal forest in the state of São Paulo. Sclerolobium paniculatum and Mabea fistulifera had good results in their identification by object oriented classification in Quickbird images. The low spectral resolution of Quickbird images is somewhat compensated by the spatial information available for the species of interest defined in our object-oriented classification. The species identification for semi-deciduous forest was not satisfactory, however in this kind of vegetation it was possible the separation among deciduous and evergreen trees. Vegetation indexes from Quickbird and Aster band math were good information source for tree species potential distribution maps. The results show that, depending on the forest type and species of interest, the spatial resolution of Quickbird images is sufficient for the identification of individual tree crowns. These images have big application in biogeography studies. / Doutorado / Análise Ambiental e Dinâmica Territorial / Doutor em Ciências

Orientador: Ricardo Cordeiro / Tese (doutorado) - Universidade Estadual de Campinas, Faculdade de Ciencias Medicas / Made available in DSpace on 2018-08-14T19:13:07Z (GMT). No. of bitstreams: 1 Zangirolani_LiaThiemeOikawa_D.pdf: 1534879 bytes, checksum: 1b381837b0cecac6a7352641ffe10179 (MD5) Previous issue date: 2009 / Resumo: A estreita relação da obesidade com a ocorrência de agravos crônicos não transmissíveis e o grau epidêmico de sua expansão faz com que esta figure entre os problemas de saúde pública mais graves em todo o mundo. O objetivo do estudo foi estimar a distribuição espacial do excesso de peso, em adultos, no Distrito Sul de Campinas-SP. Trata-se de um estudo transversal de base populacional com adultos, selecionados por meio de amostra aleatória em dois estágios. Foram caracterizadas variáveis demográficas, de condições de moradia e saneamento, do estado nutricional, das práticas alimentares e do estilo de vida. Foi ajustado um modelo de Regressão Logística Ordinal, tendo como variável resposta os três níveis de estado nutricional (eutrofia=1, sobrepeso=2, obesidade=3) e como preditoras as variáveis selecionadas nos ajustes logísticos simples. Para a distribuição espacial do excesso de peso, e para analisar se o aumento do Índice de Massa Corporal (IMC) se dá de forma homogênea no território, utilizou-se ferramentas do geoprocessamento e da estatística espacial. Os resultados apontaram a idade, as condições precárias de vida, a qualidade da alimentação e o estilo de vida no lazer como aspectos associados positivamente ao aumento do IMC. Foi possível mapear a topologia do excesso de peso, em adultos, além de verificar que os aspectos associados positivamente ao aumento do IMC se distribuem de forma heterogênea no território / Abstract: The close relationship between obesity and the occurrence of chronic non-communicable diseases and the epidemic degree of its expansion makes this, amongst the public health problems, one of the most serious in the world. The main goal of this study was to estimate the spatial distribution of weight excess in adults in the South District of Campinas-SP. This is a population-based cross-sectional study with adults, selected through random sampling in two stages. In this study, it was characterized demographic variables in terms of housing and sanitation, nutritional status, food habits and lifestyle. Considering the ordinal nature of scales used to Body Mass Index (BMI) it was adjusted an Ordinal Logistic Regression model, known as Proportional Odds, having as the dependent variable the three levels of nutritional status (normal weight=1, overweight=2, obesity=3) and as predictive, variables selected in simple logistics settings. For the spatial distribution of weight excess, and to examine whether the increase in BMI occurs homogeneously in the territory, we make use of Geographic Information System (GIS) and spatial statistics. The results indicated the age, poor living conditions, quality of food and sedentary leisure time as aspects positively associated with increased BMI. It was possible to map the spatial distribution of weight excess in adults and to identify which aspects positively associated with increased BMI are distributed heterogeneously in the area / Doutorado / Epidemiologia / Mestre em Saude Coletiva

M.A. (Geography) / The provision of open space m cities is important in creating a pleasant atmosphere, in dividing incompatible land zones and m providing recreation facilities. Researchers on this topic agree that with rapid population growth and urbanization it is necessary that, in future, city planning should be done to enhance the quality of life of city dwellers ...

Disease mapping and spatial statistics have become an important part of modern day statistics and have increased in popularity as the methods and techniques have evolved. The application of disease mapping is not only confined to the analysis of diseases as other applications of disease mapping can be found in Econometric and financial disciplines. This thesis will consider two data sets. These are the Georgia oral cancer 2004 data set and the South African acute pericarditis 2014 data set. The Georgia data set will be used to assess the hyperprior sensitivity of the precision for the uncorrelated heterogeneity and correlated heterogeneity components in a convolution model. The correlated heterogeneity will be modelled by a conditional autoregressive prior distribution and the uncorrelated heterogeneity will be modelled with a zero mean Gaussian prior distribution. The sensitivity analysis will be performed using three models with conjugate, Jeffreys' and a fixed parameter prior for the hyperprior distribution of the precision for the uncorrelated heterogeneity component. A simulation study will be done to compare four prior distributions which will be the conjugate, Jeffreys', probability matching and divergence priors. The three models will be fitted in WinBUGS® using a Bayesian approach. The results of the three models will be in the form of disease maps, figures and tables. The results show that the hyperprior of the precision for the uncorrelated heterogeneity and correlated heterogeneity components are sensitive to changes and will result in different results depending on the specification of the hyperprior distribution of the precision for the two components in the model. The South African data set will be used to examine whether there is a difference between the proper conditional autoregressive prior and intrinsic conditional autoregressive prior for the correlated heterogeneity component in a convolution model. Two models will be fitted in WinBUGS® for this comparison. Both the hyperpriors of the precision for the uncorrelated heterogeneity and correlated heterogeneity components will be modelled using a Jeffreys' prior distribution. The results show that there is no significant difference between the results of the model with a proper conditional autoregressive prior and intrinsic conditional autoregressive prior for the South African data, although there are a few disadvantages of using a proper conditional autoregressive prior for the correlated heterogeneity which will be stated in the conclusion.

From Introduction: In geography during the 1950's there was a definite move away from the study of unique phenomena to the study of generalized phenomena or pattern (Mather and Openshaw, 1974). At the same time interrelationships between phenomena distributed in space and time became the topic of much interest among geographers, as well as members of other disciplines. The changing emphasis initiated acceptance of certain scientific principles (Cole, 1973), and mathematical techniques became the recognized and respected means through which objective analysis of pattern, structure, and interrelationships between a really distributed phenomena could be achieved (Ackerman, 1972; Burton, 1972; Gould, 1973). Geographers, as do members of other disciplines, frequently borrow mathematical techniques developed for problems encountered in the pure sciences and apply these techniques to what are felt to be analogous situations in geography.