Biomass functions applicable to Scots pine

This study describes parameterization of biomass functions applicable to Scots pine (Pinus sylvestris, L.) in the conditions of Central Europe. Fifty-two sample trees from seven sites in different regions of the Czech Republic were used for destructive measurements. The observed aboveground biomass (B (AB)) and its individual components were examined by different types of non-linear regression models using one to five independent variables including stem diameter (D), tree height (H), tree age (A), crown length (C) and altitude (Z). The best single-equation approximation of B (AB) was a three-parameter model using D, H and Z, which was also best suited for bark. Including altitude into classical multiplicative exponential function with three independent variables (DHZ) and four parameters yielded the best model for stem over or under bark. Age was important for assessment of dead branches within DHA model, whereas living branches were best approximated using four variables including C (DHAC model). The most complex model (DHACZ) worked best for needle biomass. The comparison of B (AB) assessment by single regression equation and from sum of individual components showed a negligible difference for full-grown trees, whereas the additive assessment yielded considerably higher B (AB) for small or young trees. This corresponds to the assessed confidence intervals for individual trees that were larger for smaller trees. The paper also discusses the issue of commonly used linearization of biomass equations and application of linear regression, and provides comparative examples with nonlinear approach used here. The paper presents the parameter sets for the tested equations for B (AB) and individual biomass components.