USO DE REGRESSÃO MÚLTIPLA NO ESTUDO DA DINÂMICA DE CARBONO, NITROGÊNIO E FÓSFORO EM UM SISTEMA SOLO-PLANTA IRRIGADO COM ÁGUA RESIDUÁRIA

• Ano de defesa: 2012
• Autor(a) principal: Lima, Joaquim José Honorio de
• Orientador(a): Fonseca, Adriel Ferreira da
• Banca de defesa: Leal, Rafael Marques Pereira , Vaz, Maria Salete Marcon Gomes
• Tipo de documento: Dissertação
• Tipo de acesso: Acesso aberto
• Idioma: por
• Instituição de defesa: UNIVERSIDADE ESTADUAL DE PONTA GROSSA
• Programa de Pós-Graduação: Programa de Pós Graduação Computação Aplicada
• Departamento: Computação para Tecnologias em Agricultura
• País: BR
• Palavras-chave em Português: modelo de regressão; estatística multivariada; água residuária; reciclagem
• Palavras-chave em Inglês: model regression; multivariate statistic; wastewater; recycling
• Área do conhecimento CNPq: CNPQ::CIENCIAS EXATAS E DA TERRA::CIENCIA DA COMPUTACAO
• Link de acesso: http://tede2.uepg.br/jspui/handle/prefix/155

The search for alternative water sources for agriculture activities have been an actual concern due to decreasing of available potable water. The agricultural use of secondary treated sewage effluent (STSE) for crops irrigation has been an attractive practice in several soil and climate conditions. However, the changes in soil-plant system after STSE application is not well comprehended in tropical environment, specially under pasture crops. The aim of this work, to model scenarios, including soil-pasture systems subjected to irrigation and rates of nitrogen by mineral fertilization (NMF), aiming mathematics models to understand of the dynamic attributes of the total carbon (TC), total nitrogen (TN) and available phosphorus (Pavai) soil. The linear regression models (LRM) were used to approach. The variable selection methods of stepwise, forward and backward were used to LRM development and compared to standard method, through the adjusted coefficient of determination and variance influence factor. The models developed using variable selection methods were the most indicated. The CT has most of its concentration explained by the variables block, time, temperature and precipitation. The concentration of NT and P were not explained by the same group of variables in the different scenarios. The organic carbon from Tifton 85 grass and TN from STSE inputs appeared not to be enough to mitigate the TC and TN losses of soil, respectively. The grass was, in adequate conditions for growth and above ground dry mass accumulation, an important Pavai sink and, moreover, the input of P from STSE few changed the concentration of soil Pavai.