Prognose do uso da terra em ambientes sujeitos à desertificação utilizando Cadeias de Markov e Autômatos Celulares

• Ano de defesa: 2012
• Autor(a) principal: Barros, Kelly de Oliveira
• Orientador(a): Ribeiro, Carlos Antônio Alvares Soares
• Banca de defesa: Assis, Igor Rodrigues de
• Tipo de documento: Dissertação
• Tipo de acesso: Acesso aberto
• Idioma: por
• Instituição de defesa: Universidade Federal de Viçosa
• Programa de Pós-Graduação: Mestrado em Ciência Florestal
• Departamento: Manejo Florestal; Meio Ambiente e Conservação da Natureza; Silvicultura; Tecnologia e Utilização de
• País: BR
• Palavras-chave em Português: Geoprocessamento; Degradação da terra; Mata seca
• Palavras-chave em Inglês: Geoprocessing; Land degradation; Dry forest
• Área do conhecimento CNPq: CNPQ::CIENCIAS AGRARIAS::RECURSOS FLORESTAIS E ENGENHARIA FLORESTAL::MANEJO FLORESTAL
• Link de acesso: http://locus.ufv.br/handle/123456789/3108

The foremost objective of this study was to a analyze the performance of a Markov Chain/Cellular Automata model for predicting land use/land cover changes in environments subjected to desertification. The study area is the Vieira river basin, located in Montes Claros, state of Minas Gerais, Brazil, bounded from a SRTM dataset. A temporal analysis of land use/land cover changes and the classifications overall accuracy, measured by the Kappa index, was held in ArcGIS 10 and Idrisi Selva software. A set of Landsat-5 TM images, for Aug.1990, Sep.1995, Sep.2000 and Jul.2005 were classified into 4 classes of land use/land cover: Urban Area, Forest, Cropland/Pasture, and Bare Soil. The Markov matrices, which describe the transition probabilities between land use changes, were derived for the following pair of images: 1995-2000, 2000-2005, 1995-2005, 1995-2000, and then used to forecast land use for these respective years: 2005, 2010, 2015 and 2020. The CA_Markov model from Idrisi was applied to each one of these pairs in order to provide a spatial depiction of the land use changes estimated by the Markov Chain model. Next, the results were compared to the previously classified images from the corresponding years, using the Kappa index as change indicator. A quantitative temporal analysis for each land use/land cover class in the study area pointed out that both Urban Area and Bare Soil classes faced a steady increase over the years, while Forest and Cropland/Pasture presented an opposite trend. Results from Cellular Automata should be evaluated separately from those of the Markov Chain model. The latter proved to be efficient in the quantitative prediction of changes in land use/land cover. Regarding the Cellular Automata it was possible to notice an average performance in the spatial distribution of classes. Specifically, with regard to desertification, the use of CA_Markov model was effective to estimate the total area of the most susceptible class to this process, the Bare soil, however, was inefficient in its spatialization. Even with the caveats relating to the performance of Cellular Automata can be considered as good overall capacity prediction CA_Markov model.