Selecting a level of conditioning for the segmented polynomial taper equation

Abstract

Using tree data collected from Ioblolly pine thinning study plots, different levels of conditioning were examined for the segmented polynomial taper equation presented by Max and Burkhart (1976). An eight-parameter model with minimum constraints (diameter at the tip of a tree is zero and the adjacent functions are continuous at the join points) did not perform better than a six-parameter model with an additional smoothness constraint in terms of fit and predictive ability. The join points where the adjacent segments of a tree meet, the inflection points, were assumed known to further reduce the number of parameters in the model. A four-parameter model with inflection points at 11 and 75% of total tree height was slightly superior to the six- and eight-parameter models in estimating tree diameters. The fit statistics and predictive ability of the model were not sensitive to the lower and upper inflection points in the range of 9-12% and 70-80%, respectively.