Abstract
Tree diameter distribution is a key descriptive parameter of forest stands, yet studies on its modelling for major Central European forest types remain limited, particularly in near-natural forests. This study aimed to i) identify the best modelling method and ii) test the applicability of concentric fixed-area sample plots for modelling diameter distribution in even-aged, near-natural European beech stands. Diameter distribution was modelled using the two-parameter Weibull function based on data from 118 concentric fixed-area sample plots (200–400 m2 each). Several modelling methods were tested, including the parameter prediction method, parameter recovery method and combinations of the two. Truncated samples were addressed using three approaches: applying expansion factors, incorporating truncation into the function formulation and ignoring truncation. Models were evaluated based on the ability to predict relative frequencies of trees per 5-cm diameter classes and key plot parameters (quadratic mean diameter, basal area and dominant diameter). Some of the 14 developed models performed similarly well; however, the parameter recovery model incorporating expansion factors was identified as the most suitable for forest management. The selected plot parameters were predicted with an acceptable bias, precision and accuracy, while prediction of relative frequencies within diameter classes was less reliable, particularly for thinner trees. The study also revealed considerable heterogeneity in tree diameters in near-natural beech forests. Overall, the findings demonstrate that concentric fixed-area sample plots are a feasible data source for modelling diameter distribution in near-natural beech forests. However, limitations related to heterogeneity, plot size and truncated samples must be considered in future applications.