| Popis: |
The modeling of the taper and shape of the tree stem is of great interest to forest managers and can be carried out using different approaches. The objective of this study was to improve the accuracy of taper modeling, considering a methodology that combines generalization and calibration with an upper-stem diameter. An experiment was conducted with stands of Pinus taeda from southern Brazil that were unthinned and crown thinned at different levels. The destructive sampling was performed at age 30, accumulating the effect of the management regime over time. The model was developed in three parts: a) the Kozak variable exponent model was selected; b) the inclusion of variables was tested (sociological position, relative spacing index, thinning index, uniformity index, and site) in the exponent of the Kozak model of part “a”, allowing generalizations; c) upper-stem diameters at heights of 2, 3, 4, 5, 6, 7, 8, 9, 10, and 11 m were tested to calibrate the part ''b'' model, using the algebraic restriction method. The inclusion of the sociological position in the model exponent and the calibration with the upper-stem diameter at the stem height of 6 m reduced the RMSE and the MAE by 3%, with respect to volume (m³). The reduction of the error was concentrated mainly in the lower-stem section, considered the most economically valuable part. The methodology developed uses information available from a forest inventory to compute the sociological position of the trees and from harvesting for a single upper-stem diameter. Although the equation has been developed for unthinned and crown thinned stands at different intensities of loblolly pine (Pinus taeda), the extrapolation potential of the methodological approach is valid, making implementation feasible for other species and management alternatives. |
