Predicting the height-diameter pattern of planted Pinus kesiya stands in Zambia and Zimbabwe

The aim of the study was to define and analyse a prediction model for the relationship between tree height and diameter of Pinus kesiya (Royle ex Gordon) grown in the forest plantations of Zambia and Zimbabwe. The modelling was based on an equation of the power type. The development of the height-diameter pattern was assumed to be dependent on both tree size and stand characteristics. The tree diameter at breast height, stand age, dominant height and diameter were used as predictors of the tree height. Due to the spatially hierarchical (stands, trees) and temporal (measurement occasions) correlation structures in the data, the basic assumption about noncorrelated error terms did not hold. Therefore, the generalised least squares method was used in the parameter estimation of the random parameter model. In model applications, the development of the height-diameter pattern is predicted with the parameters of the fixed model part. The variance estimates for the random stand, measurement and tree effects can be used in model calibration by applying the standard linear prediction theory. These results lead to the recommendation that the power type of model be used as the predictor of the height-diameter relationship of P. kesiya stands in the tree plantations of southeastern Africa. (C) 2002 Elsevier Science B.V. All rights reserved.