Modelagem matemática da matéria orgânica do solo em sistema silvipastoril biofertilizado com água residuária da suinocultura

• Ano de defesa: 2019
• Autor(a) principal: Shimamoto, Giulia Faria
• Tipo de documento: Dissertação
• Tipo de acesso: Acesso aberto
• Idioma: por
• Instituição de defesa: Universidade Federal de Uberlândia; Brasil; Programa de Pós-graduação em Qualidade Ambiental
• Palavras-chave em Português: Modelos multivariados; Sistemas ambientais; Corymbia citriodora; Multivariate models; Environmental systems; Desenvolvimento sustentável; CNPQ::CIENCIAS AGRARIAS
• Link de acesso: https://repositorio.ufu.br/handle/123456789/27125; http://dx.doi.org/10.14393/ufu.di.2019.2279

Studies that prove the biofertilizing potential of swine wastewater (ARS) are interesting to reduce the effects of its incorrect disposal, which, in addition to contrary to environmental legislation, negatively impacts the soil-plant-water resources system. From this perspective, the study was conducted at Fazenda Bonsucesso, in Uberlândia, Minas Gerais; after the first year of application of ARS, in three experimental areas: single pasture - PS, and silvipastoral system (SSP), in single line - LS, and double line - LD, with consortium of species Corymbia citriodora and Urochloa decumbens, in layers 0,0-0,2m (X) and 0,2-0,4 m (Y). It is proposed in this research to evaluate the interactions between soil physicochemical characteristics and their effects on soil organic matter (MOS) content, using the combination that techniques of exploratory factor analysis, the Generalized Linear Model (GLM) and response optimization. With the exploratory factor analysis it was found that all the physicochemical characteristics analyzed correlate with each other and, distributed in four factors, explain 66,615% of the total variance of the variables: potassium, magnesium, copper, zinc, manganese, phosphorus, sulfur, calcium, as well as organic matter, and the iron and aluminum elements are the highlight, as they explain 49,06% of the total variance of the variables. The standards established by this analysis allow us to consider that the results obtained are satisfactory and consistent with the soil biofertilization technique with ARS. In GLMs, the technique of double and triple nesting of the factors Area, Depth and Treatment was used. In all double combinations, sodium, calcium and aluminum were significant and directly proportional to the behavior of MOS, as well as depth X, dose 400 m³ ha-1 and area LS, and on the other hand; potassium, depth Y, 800 m³ ha-1 dose and PS and LD areas were indirectly proportional to the MOS contents. Already in the nesting between three factors it was found that calcium, depth X and LS area are significant and positive variables, and in contrast, potassium, depth Y, LD and PS areas and dose 800 m³ ha-1 are indirectly proportional to the concentration of MOS. Therefore, it is concluded that triple nesting is more robust because it considers all experimental situations and, therefore, complement the information obtained by the associations between two factors. With the application of the response optimization technique it was determined that the scenario of maximum expression of MOS results from the interaction of the LS area with the 400 m³ ha-1 dose of ARS, at depth X which validates the results obtained in the Generalized Linear Models (GLMs) generated.