Predictions of species’ ranges from distribution modeling are often used to inform marine management and conservation efforts, but few studies justify the model selected or quantify the uncertainty of the model predictions in a spatial manner. This thesis employs a multi-model, multi-area SDM analysis to develop a higher certainty in the predictions where similarities exist across models and areas. Partial dependence plots and variable importance rankings were shown to be useful in producing further certainty in the results. The modeling indicated that glass sponges (Hexactinellida) are most likely to exist within the North Pacific Ocean where alkalinity is greater than 2.2 μmol l-1 and dissolved oxygen is lower than 2 ml l-1. Silicate was also found to be an important environmental predictor. All areas, except Hecate Strait, indicated that high glass sponge probability of presence coincided with silicate values of 150 μmol l-1 and over, although lower values in Hecate Strait confirmed that sponges can exist in areas with silicate values of as low as 40 μmol l-1. Three methods of showing spatial uncertainty of model predictions were presented: the standard error (SE) of a binomial GLM, the standard deviation of predictions made from 200 bootstrapped GLM models, and the standard deviation of eight commonly used SDM algorithms. Certain areas with few input data points or extreme ranges of predictor variables were highlighted by these methods as having high uncertainty. Such areas should be treated cautiously regardless of the overall accuracy of the model as indicated by accuracy metrics (AUC, TSS), and such areas could be targeted for future data collection. The uncertainty metrics produced by the multi-model SE varied from the GLM SE and the bootstrapped GLM. The uncertainty was lowest where models predicted low probability of presence and highest where the models predicted high probability of presence and these predictions differed slightly, indicating high confidence in where the models predicted the sponges would not exist.

No description available.

No description available.

Darbe yra siekiama išanalizuoti trimačio kartografinio vaizdo (žemėlapio) informatyvumo galimybes autoriaus sukurtų trimačių žemėlapių (modelių) pagrindu. Darbe apžvelgiami trimačio kartografinio vaizdo kūrimo metodai ir priemonės. Svarbų indelį trimačių kartografinių kūrinių sudaryme šiandien vaidina atskira grafinio dizaino kryptis – infografika. Tačiau trimatį kartografavimą palaikančių kompiuterinių programų nėra daug, pagrindinės iš jų leidžiančios sudarinėti erdvinius modelius yra ArcGIS ir AutoCAD. ArcGIS grafinio trimačio vizualizavimo ir Google SketchUp programinės įrangos pagrindu sukurti 4 trimačiai įvairių teminių sričių žemėlapiai. Informatyvumo nustatymo tikslui pasiekti yra vykdomas trimačių kartografinių ženklų informatyvumo, informacijos talpumo (kartografuojamų rodiklių kiekio atžvilgiu) bei panaudojimo tyrimas. Taip pat atliekama kartografijos profesionalų bei teminių sričių specialistų apklausa. Rezultatuose išryškėja pradinių kartografuojamų duomenų analizės svarba. Tinkamai parinkus kartografinį metodą leidžiantį atskleisti trečios dimensijos papildomas galimybes trimatis žemėlapis tampa informatyvesnis lyginant su įprastu planiniu kartografiniu vaizdu. Santykinių – hipotetinių paviršių žemėlapiai, kuriuose naudojami struktūriniai ženklai tinkamiausi socialinių rodiklių, o kombinuojant su tolydžiu paviršiumi – ir fiziniams reiškiniams kartografuoti. / The author is seeking to analize three-dimensional cartographical view informative opportunities created by the author 3D maps (models) basis. The paper gives an overview of 3D cartographic display methods and instruments. Important role in nowadays 3D mapping plays a separate graphic design trend – infographics. However 3D mapping computer programs are not many, the main ones allowing a spatial model are ArcGIS and AutoCAD. Basis of 3D graphics and imaging software ArcGIS together with Google SketchUp were created 4 different 3D thematic maps. Informative purpose is atchieved by research of 3D cartographic signs informative-capacity and uses potential. Also were done a poll of mapping professionals and thematic cartography specialists. Initial results revealed the importance of mapping data analysis. Properly selected mapping method allows to reveal additional opportunities of third dimension, in this way the map becomes more informative comparing it with two-dimensional map. Relative – hyphotethical surfaces maps in which are used structural cartographic signs is most appropriate for mapping of social indicators. And in the combination with continual surface (terrain model) – is appropriate for mapping of natural phenomena.

<p>The aim of the thesis is to quantitatively study the driving forces and mechanisms for regional growth from an endogenous and exogenous perspective and reveal the most important factors contributing to regional growth, by focusing on three aspects: local labour market, the supply side and the demand side of the labour market. The thesis is designed to use Swedish micro register data to develop spatial models with higher spatial resolution. It was found that endogenous factors are important and probably explain about at least one third of total regional economic growth. Among the endogenous factors, localised demographic composition, labour force and labour market, firms, and business environment have the strongest influence on regional economic growth. The findings from the Swedish context were briefly compared to China’s economic growth in the last fifty years.</p><p>The thesis consists of three related papers. The first paper studied the endogenous and exogenous factors in 108 Swedish LA regions during the 1990s. By using the SNI92 code, individual longitudinal data and an improved shift-share analysis method, it was found that the endogenous factor is important for regional economic growth because it is able to accelerate, decelerate or reverse the impact from exogenous factors during the period studied.</p><p>The second paper studied regional growth from the supply side of the labour market by focusing on population redistribution and place attractiveness. A ‘floating grid’ approach was developed to understand the factors shaping place attractiveness. The approach disregards administration zones by focusing on a small spatial unit—vicinity which is one kilometre square. Each unit has a unique set of surrounding zones that are local area and hinterland. By constructing spatial models, the total explained variance in place attractiveness was decomposed into partial explanatory effects that are assigned for physical attraction, demographic, service and labour market factors over the spatial scales. The finding is that the spatial scale of vicinity and demographic factors contribute most to place attractiveness.</p><p>The third paper studied regional growth from the demand side of the labour market by focusing on workplace and its economic performance. The ‘floating grid’ approach was once more applied while the basic analysis unit is a constructed workplace that holds working-square, local area and hinterland as surrounding zones. The economic performance of the workplace was attributed to external demand, local demand, business environment and labour force factors over different spatial scales. A method was developed to quantitatively identify intervals of partial explanatory effects that are components of the total explained variance. It was found that working-square and labour force factors contribute most to workplace economic performance.</p>

The aim of the thesis is to quantitatively study the driving forces and mechanisms for regional growth from an endogenous and exogenous perspective and reveal the most important factors contributing to regional growth, by focusing on three aspects: local labour market, the supply side and the demand side of the labour market. The thesis is designed to use Swedish micro register data to develop spatial models with higher spatial resolution. It was found that endogenous factors are important and probably explain about at least one third of total regional economic growth. Among the endogenous factors, localised demographic composition, labour force and labour market, firms, and business environment have the strongest influence on regional economic growth. The findings from the Swedish context were briefly compared to China’s economic growth in the last fifty years. The thesis consists of three related papers. The first paper studied the endogenous and exogenous factors in 108 Swedish LA regions during the 1990s. By using the SNI92 code, individual longitudinal data and an improved shift-share analysis method, it was found that the endogenous factor is important for regional economic growth because it is able to accelerate, decelerate or reverse the impact from exogenous factors during the period studied. The second paper studied regional growth from the supply side of the labour market by focusing on population redistribution and place attractiveness. A ‘floating grid’ approach was developed to understand the factors shaping place attractiveness. The approach disregards administration zones by focusing on a small spatial unit—vicinity which is one kilometre square. Each unit has a unique set of surrounding zones that are local area and hinterland. By constructing spatial models, the total explained variance in place attractiveness was decomposed into partial explanatory effects that are assigned for physical attraction, demographic, service and labour market factors over the spatial scales. The finding is that the spatial scale of vicinity and demographic factors contribute most to place attractiveness. The third paper studied regional growth from the demand side of the labour market by focusing on workplace and its economic performance. The ‘floating grid’ approach was once more applied while the basic analysis unit is a constructed workplace that holds working-square, local area and hinterland as surrounding zones. The economic performance of the workplace was attributed to external demand, local demand, business environment and labour force factors over different spatial scales. A method was developed to quantitatively identify intervals of partial explanatory effects that are components of the total explained variance. It was found that working-square and labour force factors contribute most to workplace economic performance.

Le suivi et la connaissance de la phénologie de la vigne sur plusieurs saisons sont des informations essentielles qui permettent de caractériser les régions viticoles, d’étudier le changement climatique et de planifier des pratiques agricoles telles que l'irrigation, la fertilisation, l'application de pesticides et la vendange à l'échelle parcellaire. La plupart des études menées sur de la phénologie de la vigne, considèrent la parcelle comme une unité de gestion homogène, sans tenir compte de la variabilité spatiale qui existe à cette échelle. Sur la base de travaux précédents qui mettent en évidence une forte variabilité intra-parcellaire en viticulture, ce travail de thèse vise à caractériser la variabilité spatiale de la phénologie de la vigne et les principaux facteurs qui la déterminent à l’échelle d’une parcelle viticole. Il vise aussi à en proposer un modèle de prédiction spatiale qui considère la variabilité de la phénologie de la vigne à cette échelle. Pour répondre à ces objectifs, quatre parcelles expérimentales situées dans la Vallée du Maule (Talca, Chili) ont été utilisées. Celles-ci correspondent aux principaux cépages rouges (Cabernet Sauvignon et Carménère) et blancs (Chardonnay et Sauvignon Blanc) cultivés au Chili. Ce travail a été réalisé tout au long 6 années de 2009 et 2015, au cours desquelles entre de deux et quatre séries des mesures ont été effectuées selon la parcelle viticole. Pour chaque parcelle, une grille régulière de 12 sites d'échantillonnage par hectare a été définie, sur chaque site, des mesures de la phénologie de la vigne, de la maturité des baies (accumulation des sucres), de l’état hydrique de la plante, de l’expression végétative, du rendement, des variables climatiques (température et hygrométrie), des propriétés physique du sol, entre autres, ont été réalisées. Les principaux résultats ont montré l'existence d’une variabilité spatiale de la phénologie de la vigne importante à l’échelle intra-parcellaire. Cette variabilité a été observée autant pour tous les stades phénologiques clés (débourrement, floraison et véraison) et pour l’accumulation des sucres dans les baies. Elle est structurée spatialement et stable au cours du temps. Son importance est similaire à la variabilité observée à des échelles spatiales régionale. Au niveau intra parcellaire, la topographie (différence d'altitude) a été identifiée comme le principal facteur d’influence sur la variabilité spatiale de la phénologie et de l’accumulation des sucres dans les baies. Dans le cas où la variabilité de la topographie de la parcelle viticole est faible, les caractéristiques du sol sont le second facteur expliquant la variabilité spatiale de la phénologie, tandis que les variables de la plante (expression végétative et rendement) expliquent la variabilité observée dans l'accumulation de sucre. La stabilité temporelle de la variabilité spatiale des stades phénologiques et de l'accumulation des sucres dans les baies, a permis de proposer un modèle spatial empirique qui permet d’estimer la valeur ces variables à l’échelle d’une parcelle viticole. L’originalité de l’approche proposée est d’utiliser une mesure ponctuelle de la phénologie ou de la teneur en sucres des baies, effectuée dans la parcelle viticole (site de référence) et une combinaison de coefficients site-spécifiques estimés à partir des données historiques. Ce modèle spatial permet d’obtenir des estimations de meilleure qualité en comparaison avec des méthodes classiques d’échantillonnage. Il permet d’obtenir des cartes pour les stades phénologiques clés et pour l’accumulation des sucres. Cette méthode pourrait être utilisée comme un outil pratique pour la planification de la gestion modulée des opérations au niveau intra-parcellaire, où la connaissance de la phénologie est essentielle pour la prise de décisions. / The knowledge and monitoring of grapevine phenology over several seasons are important requirements for the characterization of vine regions, climate change studies and planning of various production activities such as irrigation, fertilization, phytosanitary applications and harvesting at the vine field scale. However, the main studies developed on grapevine phenology consider the vine field as a homogeneous unit of management and do not take into account the spatial variability observed at this spatial scale. Based on previous studies highlighting a significant variability at the within field level in viticulture, this doctoral research aims to characterize the spatial variability of grapevine phenology at the vine field scale, relating the main factors that determine this variability and proposing a spatial prediction model that considers the variability of phenology. In order to answer the above objectives, experiments were carried out in four vine fields located in the Maule Valley (Talca, Chile), considering the main red (Cabernet Sauvignon and Carménère) and white (Chardonnay and Sauvignon Blanc) cultivars planted in Chile. This experiment was carried out over 6 years between years 2009 and 2015. Within each vine field, a regular grid with 12 sampling sites per hectare was defined. Measurements of grapevine phenology, berry maturity (sugar accumulation), plant water status, vegetative expression, yield, climatic variables (temperature and relative humidity), soil physical properties, among others, were performed on each site. The main results showed the existence of a significant spatial variability of the phenological development at the within-field scale, observed for both key phenological stages (budburst, flowering and veraison) and sugar accumulation in berries. This spatial variability was spatially organized and stable over seasons, being comparable even with the observed variability at important spatial scales such as regional or vine valley. At the within-field scale, topography (elevation difference) was identified as the main integrative factor affecting the spatial variability of both grapevine phenology and maturity. If there is no variability in the topography, the soil characteristics become the second factor of the spatial variability of grapevine phenology, while plant variables (vegetative expression and yield) explained the observed variability in sugar accumulation at this scale. The temporal stability observed for the spatial variability of the phenological stages and sugar accumulation of berries allowed an empirical spatial model to be proposed. The model characterizes the spatial variability of these variables at the vine field scale. The originality of the approach is to use a single measurement of the grapevine phenology or sugar accumulation in the berries performed in the field (reference site) and a combination of site-specific coefficients estimated through historical data. The spatial model presented the best results when compared with classical sampling methods, allowing maps of key phenological stages and sugar accumulation to be obtained. This methodology could be used as a practical tool for the planning site-specific management operation at the vine field scale where phenology is essential for decision-making.

Submitted by Neusa Fagundes (neusa.fagundes@unioeste.br) on 2017-09-25T18:57:03Z No. of bitstreams: 1 Regiane_Fagundes2017.pdf: 6331934 bytes, checksum: faab7007f3c7c2e91c6bf26bc30fea8e (MD5) / Made available in DSpace on 2017-09-25T18:57:03Z (GMT). No. of bitstreams: 1 Regiane_Fagundes2017.pdf: 6331934 bytes, checksum: faab7007f3c7c2e91c6bf26bc30fea8e (MD5) Previous issue date: 2017-01-17 / Fundação Araucária de Apoio ao Desenvolvimento Científico e Tecnológico do Estado do Paraná (FA) / This study aimed at a discussing problems of multivariate statistical inference and linear spatial modeling when observations are from a continuous, symmetric population, with multivariate slash distribution. Firstly, a reparametrization of slash distribution was performed, assuming the existence of the finite second moment. Thus, some iterant properties were shown. Analytical expressions were tested for the score function and Fisher information matrix of reparameterized distribution. An approach to estimate some parameters by maximum likelihood was considered based at the EM (Expectation-Maximization) algorithm. Linear hypothesis tests have been described regarding the means vector and the covariance matrix using statistics such as C(α), likelihood ratio, Wald, and score. Studies of simulation were carried out to evaluate the efficiency of the statistical tests and EM algorithm. Data related to the agricultural area illustrated the methodology developed, and the hypothesis tests for equality of means, sphericity and equicorrelation were also applied. A slash linear spatial model, with and without the use of covariates, was proposed. Were Discussed the global and local influence diagnostic analysis in order to evaluate the influence of observations on the process of parameters’estimation. The curvatures required for the local influence procedure and based on the slash model were derived, in which the perturbation scheme has been chosen properly and related to the different perturbation schemes. Spatial variability maps of chemical attributes of soil and yield were generated by kriging with external drift. Finally results of simulations and applications indicated that the slash distribution is a robust alternative when the data present high kurtosis. / O objetivo deste trabalho foi discutir problemas de inferência estatística multivariada e de modelagem espacial quando as observações são provenientes de uma população contínua, simétrica, com distribuição slash multivariada. Inicialmente, foi realizada uma reparametrização da distribuição slash supondo existência do segundo momento finito, sendo apresentadas algumas propriedades recorrentes. Provaram-se expressões analíticas para a função escore e matriz de informação de Fisher da distribuição reparametrizada. Abordou-se um enfoque para a estimação dos parâmetros por máxima verossimilhança considerando um algoritmo do tipo EM (Esperança-Maximização). Descreveu-se a prova de hipóteses lineares sob o vetor de médias e matriz de covariância com o uso das estatísticas C(α), razão de verossimilhança, Wald e score. Estudos de simulação foram realizados para avaliar a eficiência dos testes estatísticos e do algoritmo EM. Dados relacionados à área agrícola ilustraram a metodologia desenvolvida, sendo aplicado sobre os mesmos os testes de igualdade de médias, esfericidade e equicorrelação. Como ilustração da aplicação da distribuição slash multivariada na área de modelagem estatística, o modelo espacial linear slash, com e sem o uso de covariáveis, foi discutido e proposto. Com o intuito de avaliar a influência das observações no processo de estimação dos parâmetros, discussões relacionadas à análise de diagnóstico, global e local, foram apresentadas. Derivaram-se as curvaturas requeridas no procedimento de influência local para o modelo slash, adequando o esquema de perturbação a distribuição e considerando diferentes esquemas de perturbação. Mapas de variabilidade espacial de atributos químicos do solo e produtividade foram gerados utilizando krigagem com drift externo. Os resultados das simulações e aplicações indicaram que a distribuição slash é uma alternativa robusta quando os dados apresentam alta curtose.

Esta tese visa propor extensões dos modelos de regressão quantílica bayesianos, considerando dados de proporção com inflação de zeros, e também dados censurados no zero. Inicialmente, é sugerida uma análise de observações influentes, a partir da representação por mistura localização-escala da distribuição Laplace assimétrica, em que as distribuições a posteriori das variáveis latentes são comparadas com o intuito de identificar possíveis observações aberrantes. Em seguida, é proposto um modelo de duas partes para analisar dados de proporção com inflação de zeros ou uns, estudando os quantis condicionais e a probabilidade da variável resposta ser igual a zero. Além disso, são propostos modelos de regressão quantílica bayesiana para dados contínuos com um componente discreto no zero, em que parte dessas observações é suposta censurada. Esses modelos podem ser considerados mais completos na análise desse tipo de dados, uma vez que a probabilidade de censura é verificada para cada quantil de interesse. E por último, é considerada uma aplicação desses modelos com correlação espacial, para estudar os dados da eleição presidencial no Brasil em 2014. Nesse caso, os modelos de regressão quantílica são capazes de incorporar essa informação espacial a partir do processo Laplace assimétrico. Para todos os modelos propostos foi desenvolvido um pacote do software R, que está exemplificado no apêndice. / This thesis aims to propose extensions of Bayesian quantile regression models, considering proportion data with zero inflation, and also censored data at zero. Initially, it is suggested an analysis of influential observations, based on the location-scale mixture representation of the asymmetric Laplace distribution, where the posterior distribution of the latent variables are compared with the goal of identifying possible outlying observations. Next, a two-part model is proposed to analyze proportion data with zero or one inflation, studying the conditional quantile and the probability of the response variable being equal to zero. Following, Bayesian quantile regression models are proposed for continuous data with a discrete component at zero, where part of these observations are assumed censored. These models may be considered more complete in the analysis of this type of data, as the censoring probability varies with the quantiles of interest. For last, it is considered an application of these models with spacial correlation, in order to study the data about the last presidential election in Brazil in 2014. In this example, the quantile regression models are able to incorporate spatial dependence with the asymmetric Laplace process. For all the proposed models it was developed a R package, which is exemplified in the appendix.

Esta tese visa propor extensões dos modelos de regressão quantílica bayesianos, considerando dados de proporção com inflação de zeros, e também dados censurados no zero. Inicialmente, é sugerida uma análise de observações influentes, a partir da representação por mistura localização-escala da distribuição Laplace assimétrica, em que as distribuições a posteriori das variáveis latentes são comparadas com o intuito de identificar possíveis observações aberrantes. Em seguida, é proposto um modelo de duas partes para analisar dados de proporção com inflação de zeros ou uns, estudando os quantis condicionais e a probabilidade da variável resposta ser igual a zero. Além disso, são propostos modelos de regressão quantílica bayesiana para dados contínuos com um componente discreto no zero, em que parte dessas observações é suposta censurada. Esses modelos podem ser considerados mais completos na análise desse tipo de dados, uma vez que a probabilidade de censura é verificada para cada quantil de interesse. E por último, é considerada uma aplicação desses modelos com correlação espacial, para estudar os dados da eleição presidencial no Brasil em 2014. Nesse caso, os modelos de regressão quantílica são capazes de incorporar essa informação espacial a partir do processo Laplace assimétrico. Para todos os modelos propostos foi desenvolvido um pacote do software R, que está exemplificado no apêndice. / This thesis aims to propose extensions of Bayesian quantile regression models, considering proportion data with zero inflation, and also censored data at zero. Initially, it is suggested an analysis of influential observations, based on the location-scale mixture representation of the asymmetric Laplace distribution, where the posterior distribution of the latent variables are compared with the goal of identifying possible outlying observations. Next, a two-part model is proposed to analyze proportion data with zero or one inflation, studying the conditional quantile and the probability of the response variable being equal to zero. Following, Bayesian quantile regression models are proposed for continuous data with a discrete component at zero, where part of these observations are assumed censored. These models may be considered more complete in the analysis of this type of data, as the censoring probability varies with the quantiles of interest. For last, it is considered an application of these models with spacial correlation, in order to study the data about the last presidential election in Brazil in 2014. In this example, the quantile regression models are able to incorporate spatial dependence with the asymmetric Laplace process. For all the proposed models it was developed a R package, which is exemplified in the appendix.