Análises de trilha em caracteres relacionados à produção de etanol a partir de sorgo sacarino (Sorghum bicolor L. Moench)

Abstract

This work used data from an experiment carried out with 80 sorghum sorghum genotypes for ethanol production in the Zona da Mata in Pernambuco, conducted in a randomized block scheme with 3 replications, in the municipality of Vitória de Santo Antão - PE, in agricultural year 2013. The objective of this work was to quantify the direct and indirect effects of technological variables (total dry matter production, stalk production, BRIX percentage at harvest and juice extraction efficiency) on ethanol production and quantify the direct and indirect effects exerted by agronomic variables (average plant height, flowering, total green matter production and percentage of stalk in the PMV) on stalk production per hectare, based on the analysis of trail. Prior to the analysis, a collinearity diagnostic test was performed, which indicated the presence of such an effect in both analyzes. In the analysis of ethanol production, an approach applied to circumvent this effect was to exclude a variable causing a high degree of multicollinearity (%MST). In the analysis performed with the agronomic variables, to circumvent the possible adverse effects caused by the presence of multicollinearity detected in the diagnostic test, a crest analysis was performed with addition of constant k = 0.09 in the X'X correlation matrix. The BRIX variable showed, respectively, the greatest total effect (0.8844) and the greatest direct effect (0.5969) on the theoretical production of ethanol, indicating great dependence on the main variable for the explanatory one. The greatest indirect effect was via PCL (0.2820). The four dependent variables as dependent on the total effect on PCL. The greatest direct effect (0.8010) was exerted by the PMV.