A method to attain uniform resolution of the competition variable Basal-Area-in-Larger Trees (BAL) during forest growth projections of small plots

• Published in: Ecological modelling 2004, Vol.171 (1), p.195-206
• Main Authors: Ledermann, Thomas, Eckmüllner, Otto
• Format: Article
• Language: English
• Subjects: Animal and plant ecology; Animal, plant and microbial ecology; Biological and medical sciences; Fundamental and applied biological sciences. Psychology; General aspects. Techniques; Individual-tree growth modelling; Methods and techniques (sampling, tagging, trapping, modelling...); Sampling; Scenario analysis; Simulation; Synecology; Terrestrial ecosystems; Uneven-aged forestry
• ISSN: 0304-3800; 1872-7026
• DOI: 10.1016/j.ecolmodel.2003.08.005

The Basal-Area-in-Larger Trees (BAL) is a measure of competition which is often used in individual-tree growth modelling. However, depending on the inventory design used for sampling the parameterization data, this competition variable is characterized by a specific resolution which has an effect on the model coefficients. If such an individual-tree growth model is used for growth projections, differing BAL-resolution might cause biased predictions of stand development. In this analysis, we used the individual-tree growth simulator PROGNAUS to display the source of this problem, and to test the simple method of tree record splitting to reduce this bias. Initial conditions for these growth simulations were obtained from sample plots of different size which were located in a larger stand. A growth period of 150 years was simulated for all sample plots and for the full stand applying the tree record splitting method alternatively. The results show that the change in the BAL-resolution, and therefore, the bias in the predicted stand development depends on plot size and on the length of the simulation period. Furthermore, the final diameter-distributions of the sample plots are significantly different from the diameter-distribution of the full stand showing greater deviations on small plots. However, if the sample tree splitting method is applied, the BAL-resolution can be held more or less constant during the course of growth projection and the bias in the predicted stand development is reduced. Moreover, the sample tree splitting method ensures that the final diameter-distributions at the end of the growth period closely match the diameter-distribution of the full stand. Thus, the proposed tree record splitting method is concerned with the problem of making a model more realistic from an ecological point of view, representing exactly the front line of today’s research in ecological modelling.