Modeling of morphological variables in Eucalyptus dunnii seedlings.

Abstract

The morphological variables of forest seedlings may be indicative of its quality, having great importance to the success of forest enterprises. In this context, the aim of this study was to correlate the operationally measurable variables (stem diameter and height) with a destructive measurement (leaf area and root length) of E. dunnii seedlings in order to formulate equations that provide accurate estimates of these variables. For this, 77 seedlings were evaluated at 100 days old. The equations were formulated using the methodology "stepwise" for selecting the best combination of mathematical transformations of 45 operationally measurable variables. These equations were evaluated by statistical adjustment and precision, and subsequent validation by means of adherence test, with a confidence level of 99%, using 50 independent seedlings. The results indicated high correlation between the variables. For leaf area, the obtained adjusted coefficient of determination was 0.91 and standard error of estimate 15.2%. To root length, the adjusted coefficient of determination was 0.76 and standard error estimate of 22.5%. The two cases presented no residual bias and had been validated. as result, it is possible to formulate equations to obtain estimates of leaf area and root length in E. dunnii.