Stem volume models with random coefficients for Pinus kesiya in Tanzania, Zambia, and Zimbabwe

The aim of the study was to estimate stem volume and taper models for Pinus kesiya (Royle ex Gordon). The volume function provides a simple prediction model for the stem volume. Taper models were developed for over- and under-bark diameters. The under-bark taper curve was determined with the variable-exponent taper equation, whereas the over-bark taper curve was derived from the predicted under-bark taper model using the variable-exponent form of the bark-thickness model. Because of the spatial correlation structures of the data, the general assumption of uncorrelated residuals did not hold. In addition, the models were assumed to contain random parameters that vary from stand to stand and from tree to tree. Therefore, the fixed and random parameters of the models were estimated with the generalized least squares technique. The results of the study show that the mixed models for stem volume and taper are more reliable volume and diameter predictors for P. kesiya than earlier taper and volume functions.