Information required by forest managers about homogeneous, even-aged, single-species stands can be provided by whole-stand models, which are easily built with data often available in forest inventories and which represent a good compromise between generality and accuracy. The objective of the present study was to develop a dynamic growth model for I-214 clonal poplar plantations in the northern and central plateaux in Spain by use of data obtained from a network of 198 experimental plots measured between 2 and 12 times. The overall model is a hierarchical system that consists of three main components: a stand height projection function, a stand basal area projection function and a merchantable stand volume equation. In the model, the initial stand conditions at any point in time are defined by two state variables usually collected from common forest inventories: stand mean height and stand basal area. The model includes two transition functions, derived by the generalized algebraic difference approach, to enable projection of these variables at any particular time. The projected variables were incorporated into a merchantable stand volume equation to enable calculation of the commercial stand volume at a minimum diameter, which establish a merchantability threshold. Sensitive analysis of the accuracy of the predictions of the submodels according to the length of the time interval was achieved by use of the critical error statistic and the RMSE in relative proportion of the mean. Results indicated that the overall model provides satisfactory predictions for time intervals of between 1 and 4 years. (c) 2007 Elsevier B.V. All rights reserved.
