Análise geoestatística para geração de superfícies a partir de dados de clorofila-a adquiridos em transectos

• Ano de defesa: 2015
• Autor(a) principal: Ribeiro, Gabrielle Gomes dos Santos [UNESP]
• Tipo de documento: Dissertação
• Tipo de acesso: Acesso aberto
• Idioma: por
• Instituição de defesa: Universidade Estadual Paulista (Unesp)
• Palavras-chave em Português: Cartografia; Amostragem (Estatistica); Armazenamento de dados; Inferência espacial; Espaço urbano; Krigagem; Cartography
• Link de acesso: http://hdl.handle.net/11449/131874; http://www.athena.biblioteca.unesp.br/exlibris/bd/cathedra/24-11-2015/000854398.pdf

The arrangement of sampling units in the study area and its influence on the results of spatial analysis is something that has been frequently discussed by researchers of the area, since the quality of a spatial inference will depend on sample size and spatial distribution of sample points. In this sense, this work aims to analyze the impact that different sampling designs may cause in the results of spatial inference by ordinary kriging. For this, first we used a dataset collected in the form of transects in a part of New Avanhandava Reservoir, consisting of 978 observations. This set suffered systematics reductions, with the aim to analyze what these reductions would cause in the results of spatial inferences. In order to analyze different sampling designs, simulated up a dense amount of data and it was applied the Simple Sampling, Systematic Sampling and Stratified Sampling techniques. To complement, it was used a data set with a small size (70 observations) and collected randomly, in order to analyze the results obtained by ordinary kriging when using a set considered small, from a statistical point of view. Then, it was possible to perform the process of Ordinary Kriging and obtain mappings of the variable chlorophyll-a in the region of interest, for different types and sizes of samples. The validation of inference processes was carried out from two methods, the Mean Squared Error and the Kappa Index...