Simulação geoestatística e classificação numérica para o delineamento de unidade de mapeamento de atributos e erosão do solo

• Ano de defesa: 2014
• Autor(a) principal: Silva Júnior, João Fernandes da [UNESP]
• Tipo de documento: Tese
• Tipo de acesso: Acesso aberto
• Idioma: por
• Instituição de defesa: Universidade Estadual Paulista (Unesp)
• Palavras-chave em Português: Geologia - Métodos estatísticos; Solos - Conservação; Mineralogia do solo; Levantamentos do solo; Solos - Erosão; Soil surveys
• Link de acesso: http://hdl.handle.net/11449/121988

Agriculture calls for attributes mapping as well as soil classification data leading to an accurate presentation. The objectives of this study were: (i) to evaluate the morphologic geostatistical potential for the mapping of soil losses due to erosion; (ii) to assess two geostatistical methodologies; ordinary kriging and gaussian sequential simulation in mapping of the average diameter of goethite crystal; and (iii) to evaluate the potential of the usage of dissimilarity between mobile windows analysis from multivariated attribute data from the soil attributes of plants in a transection for the allocation of the mapping unity. The study (i) was performed in Tabapuã, a municipality in São Paulo State, in a 530 ha area, within a sampling net of 626 points, being soil samples collected, for which physical and chemical attributes were determined, and soil erosion calculated by means of the universal equation of soil losses. Based on those values, descriptive and geostatistical analysis were performed, while soil attribute values were esteemed from a morphological approach of indicator variables (dummy) function. Those resulted in morphological maps from 100 simulations. Some uncertainties of these maps were noticed by means of Shannon entropy and greater uncertainty came for the transition between erosion classes. The simulation is efficient in spatial characterization and in the definition of the lateral limits of soil losses by erosion classes. The study (ii) was in the same former area, having a net sampling of 113 points at each 10 meters. At these points the content of goethite clay fraction was determined, together with the width at halfheight, from which the goethite crystal diameter values were obtained. Models were adjusted to the variograms, followed by estimation and simulation by ordinary kriging and gaussian sequential simulation. Ordinary kriging map does not show extreme values (minimum and maximum) at specific sites ...