Mapeamento digital de solos por regress?o log?stica m?ltipla em ambiente de Mar de Morros em Pinheiral, RJ

• Ano de defesa: 2015
• Autor(a) principal: Costa, Elias Mendes
• Orientador(a): Anjos, L?cia Helena Cunha dos
• Banca de defesa: Anjos, L?cia Helena Cunha dos, Lepsch, Igo Fernando, Menezes, Carlos Eduardo Gabriel
• Tipo de documento: Dissertação
• Tipo de acesso: Acesso aberto
• Idioma: por
• Instituição de defesa: Universidade Federal Rural do Rio de Janeiro
• Programa de Pós-Graduação: Programa de P?s-Gradua??o em Agronomia - Ci?ncia do Solo
• Departamento: Instituto de Agronomia
• País: Brasil
• Palavras-chave em Português: Pedologia; Pedometria; Sensoriamento Remoto; Rela??o solo-paisagem
• Palavras-chave em Inglês: Pedology; Pedometry; Remote sensing; Soil landscape relationship
• Área do conhecimento CNPq: Agronomia
• Link de acesso: https://tede.ufrrj.br/jspui/handle/jspui/3152

The soil survey comprises the description of the morphological, physical and chemical soil characteristics, description of the surface, and determination of soil distribution in the landscape of a certain area, and it may be used for various purposes. The knowledge about soil characteristics and their distribution in the landscape are essential for planning the land use and conservation. The digital soil mapping (DSM) techniques may improve the soil surveying and the products derived from that through the use of geographic information systems tools (GIS). The overall objective of the study was to carry out a DSM in the environment of Mar de Morros in Pinheiral (RJ), applying the multiple logistic regression technique; in addition to evaluating if this method allows to obtain good accuracy of maps and other soils information, which may constitute basic material for environmental studies and land usage planning. The soil mapping involved 99 sampling points, among complete profiles and simple field observations, in the Pinheiros farm and adjacent areas, in the municipality of Pinheiral (RJ). Profiles in selected sites were described and sampled for analytical characterization of the soils. Since, the quality of the soil survey is dependent on the base material, the initial phase comprehended the detailed evaluation of the representation of the altimetry in the area (evaluation of the digital elevation model - DEM), and corrections in the satellite image (Rapid Eye) and evaluation of the positional error. The most adequate DEM was selected as base material and, with the support of satellite images and geology and legacy soil maps, it was established the mathematical relationship between soils and the discriminant variables for the most part closely related to the genesis and distribution of soils. For that, 5 models were adjusted to predict soil classes, and 6 models for the prediction of mapping units. The results indicated that the DSM obtained using the soil mapping unit as dependent variable increased the predictive power of the logistic models (84% accuracy and kappa 82%), when compared to the prediction of soil classes (accuracy 54% and kappa 43%). The geoprocessing technologies, the digital image processing, and the digital mapping proved to be good tools to aid in the soil survey; besides they provide products of more quantitative nature, which allow evaluating the error. The variety of products generated by these techniques may help future studies and the planning and usage of the land, to assist in the decision making about the rational use of natural resources.