Influência de modelos de dependência espacial na definição de mapas temáticos
Arquivos
Data
2012-07-24
Autores
Título da Revista
ISSN da Revista
Título de Volume
Editor
Universidade Federal de Viçosa
Resumo
Nos últimos anos a Geoestatística vem sendo amplamente utilizada na área de agricultura de precisão, isso se deve ao fato de ser uma ferramenta que permite analisar a variabilidade espacial existente na área de produção agrícola, possibilitando avaliar a necessidade de criação de subáreas ou zonas onde serão realizados manejos de forma diferenciada. Um dos focos da agricultura de precisão é a identificação de zonas de manejo dentro do campo baseadas na variabilidade existente, e por meio da Geoestatística é possível produzir os mapas temáticos que auxiliam no estabelecimento das zonas de manejo através de ajustes de modelos de dependência espacial. Assim, o objetivo geral deste estudo foi analisar a influência de modelos de dependência espacial na definição de mapas temáticos de zonas de manejo, usando diferentes modelos de semivariogramas e diferentes grades de amostragem. Para a realização do estudo foram analisados dados simulados gerados no software SAS, onde se considerou diferentes estruturas de dependência espacial (DE) e diferentes grades de amostragens (grid), com diferentes tamanhos e densidades de pontos, sendo um total de 9 conjunto de dados simulados. Para a análise da estrutura de dependência espacial foram utilizados semivariogramas experimentais e ajustados três modelos teóricos ao semivariograma experimental: exponencial, esférico e gaussiano, para cada conjunto de dados em estudo, conforme
os parâmetros estipulados pelo auto-ajuste do software ArcGis 9.3. Foram ajustados 27 diferentes modelos de semivariogramas através do método da krigagem ordinária e através da técnica de validação-cruzada verificados os ajustes dos modelos semivariográficos. Mapas temáticos de zonas de manejo foram gerados a partir da técnica de krigagem ordinária, e com o auxilio do software Idrisi foi realizada a comparação dos mapas através do Índice Kappa de concordância, o qual foi testado sua significância através do teste Z. A partir das análises realizadas, concluindo-se que o modelo de dependência espacial não influencia na definição de mapas temáticos de zonas de manejo.
In recent years Geostatistics has been widely used in the field of precision agriculture, this is due to being a tool to analyze the spatial variability exists in the area of agricultural production, enabling assess the need for creation of subareas or zones where they are handlings performed differently. One focus of precision agriculture is to identify management zones within the field based on the variability, and through Geostatistics is possible to produce thematic maps that assist in the establishment of management zones through adjustments of models of spatial dependence. The objective of this study was to analyze the influence of models of spatial dependence in the definition of thematic maps of management zones, using different models and different semivariogram sampling grids. To conduct the study, we analyzed simulated data generated in SAS software, where he held various structures of spatial dependence (DE) and different sampling grids (grid), with different sizes and densities of points, with a total of 9 data set simulated. To analyze the spatial dependence structure experimental semivariograms were used and adjusted three theoretical models to experimental semivariogram: exponential, spherical and Gaussian for each data set under study, according to the parameters set forth by autotune software ArcGis 9.3. Were adjusted 27 different semivariogram models using the method of ordinary kriging and through cross-validation technique checked the fit of the models semivariográficos. Thematic maps management zones were generated from the technique of kriging, and with the aid of software Idrisi was performed to compare the maps using Kappa Index of agreement, which was tested by testing their significance Z. From the analyzes, it was concluded that the model of spatial dependence does not influence the definition of thematic maps of management zones.
In recent years Geostatistics has been widely used in the field of precision agriculture, this is due to being a tool to analyze the spatial variability exists in the area of agricultural production, enabling assess the need for creation of subareas or zones where they are handlings performed differently. One focus of precision agriculture is to identify management zones within the field based on the variability, and through Geostatistics is possible to produce thematic maps that assist in the establishment of management zones through adjustments of models of spatial dependence. The objective of this study was to analyze the influence of models of spatial dependence in the definition of thematic maps of management zones, using different models and different semivariogram sampling grids. To conduct the study, we analyzed simulated data generated in SAS software, where he held various structures of spatial dependence (DE) and different sampling grids (grid), with different sizes and densities of points, with a total of 9 data set simulated. To analyze the spatial dependence structure experimental semivariograms were used and adjusted three theoretical models to experimental semivariogram: exponential, spherical and Gaussian for each data set under study, according to the parameters set forth by autotune software ArcGis 9.3. Were adjusted 27 different semivariogram models using the method of ordinary kriging and through cross-validation technique checked the fit of the models semivariográficos. Thematic maps management zones were generated from the technique of kriging, and with the aid of software Idrisi was performed to compare the maps using Kappa Index of agreement, which was tested by testing their significance Z. From the analyzes, it was concluded that the model of spatial dependence does not influence the definition of thematic maps of management zones.
Descrição
Palavras-chave
Dependência espacial, Semivariograma, Índice Kappa, Spatial dependence, Semivariogram, Ordinary kriging, Kappa index
Citação
BATISTA, Flávia Ferreira. Influence of model spatial dependence in the definition of thematic maps. 2012. 70 f. Dissertação (Mestrado em Estatística Aplicada e Biometria) - Universidade Federal de Viçosa, Viçosa, 2012.