Modelagem da biomassa e da quantidade de carbono de clones de Eucalyptus da Chapada do Araripe-PE

• Autor(a) principal: SILVA, José Wesley Lima
• Data de Publicação: 2016
• Tipo de documento: Dissertação
• Idioma: por
• Título da fonte: Biblioteca Digital de Teses e Dissertações da UFRPE
• Texto Completo: http://www.tede2.ufrpe.br:8080/tede2/handle/tede2/4565
• topic: Modelo de regressão; Análise de regressão; Semiárido; Regression analysis; Regression model; Semiarid; CIENCIAS EXATAS E DA TERRA::PROBABILIDADE E ESTATISTICA

The objective of this study was to quantify and test different regression models to estimate the biomass and the amount of carbon from the aerial parts of Eucalyptus clones planted in the Northeastern semi-arid region, and select the best equations based on R2 aj, the Akaike information criterion (AIC), Furnival Index (FI), the graphical analysis of the residuals and through the Shapiro-Wilk, Breusch- Pagan and Durbin Watson tests. The database came from an experiment with 15 Eucalyptus spp. clones conducted at the Experimental Station of the Agricultural Research Institute of Pernambuco – IPA, located in the municipality of Araripina - PE. Through a completely random sampling process 75 trees were selected, in which were determined the fresh weight and leaf samples collected were samples of leaf, branch, bark and bole to determine the average wood density, biomass and carbon content. The most productive clone in terms of biomass and carbon was the hybrid E. urophylla natural crossing. The average plant biomass accumulation was 59.64 t h−1 and the amount of carbon 24.96 t h−1. The adjustment of the regression models showed that each partition presented particular behavior of dry biomass production and total carbon. It was not possible to select a common model that represents all of parts of the trees. For the variable biomass, the models of Schumacher and Hall, Spurr, the logistics and the exponential model 11 presented the best fits. For the amount of organic carbon the models 6 and exponential 11 presented best results.