Models for Prediction of Basal Area Mean Diameter and Number of Trees for Forest Stands in South-eastern Norway

Abstract

Basal area mean diameter and number of trees ha -1 are important stand variables for projections made with large-scale forestry scenario models in Norway. The aim of this work was to develop models for an initial description of these variables. Multiplicative models, adapted to available variables from different types of inventories (relascope, photo and visual inventories) were developed. R 2 ranged from 0.85-0.87 in models where mean height by basal area was available to 0.62-0.68 without this variable. Validations provided few statistically significant differences between predicted and observed values, while the standard deviations for the differences were relatively large (7-43%). There are numerous sources of uncertainty in large-scale forestry analyses, but the developed models will probably produce a satisfactory standard of accuracy at an aggregated level. However, the large standard deviations, and large consequential losses in terms of net present value applying the models, indicate that one should avoid relying solely on the models with respect to decisions at stand level.