Verification of two- and three-parameter simple height-diameter models for birch in the European part of Russia

Abstract

The accuracy of determining the height of trees is essential both in forestry and in scientific research. Height is usually determined using specific models, where it is a function of the diameter at breast height. On the materials of 23 sample plots with the measurement of model trees in birch stands, the parameters were determined for 29 two-parameter and three-parameter models that are most often found in literary sources. The following metrics evaluated the quality of the models: root mean square error, mean absolute percentage error, coefficient of determination, adjusted coefficient of determination, Akaike information criterion, and Bayesian information criterion. Three-parameter models of the dependence of height on diameter by a set of metrics show somewhat better quality than two-parameter models. Nevertheless, in general, the differences between most models are minor. Along with the models selected as the best, the Näslund and Chapman-Richards equations, which are often used in the literature as the most flexible, showed good quality. The methodology of this study allows you to repeat the same work for tree species and forest conditions, for which information on the nature of the relationship of height with diameter is incomplete or missing.

Keywords

• simple model
• selection of models