Conditioning a Segmented Stem Profile Model for Two Diameter Measurements

Abstract

Abstract The stem profile model of Max and Burkhart (1976) is conditioned for dbh and a second upper stem measurement. This model was applied to a loblolly pine data set using diameter outside bark at 5.3m (i.e., height of 17.3 foot Girard form class) as the second upper stem measurement, and then compared to the original, unconditioned model. Variance of residuals was reduced; however, bias was approximately the same for both the conditioned and unconditioned models. Benchmark prediction problems, which reflect common multiproduct utilization criteria for southern pines, were used to judge the practical importance of these reductions. Square root of the average squared residuals from the conditioned model were 10 to 25% less than those from the unconditioned model for most benchmark evaluations. These reductions might be of practical importance, depending on the particular objectives and the relative cost of an upper stem measurement compared to the cost of measuring more trees. For. Sci. 34(2):512-522.