Models for predicting above-ground biomass of Betula pubescens spp. czerepanovii in mountain areas of southern Norway
Abstract:The objectives of this study were (1) to develop models for estimation of total above-ground biomass, tree crown biomass and stem biomass of mountain birch (Betula pubescens spp. czerepanovii), and (2) to test the stability of the relationships between biomass and biophysical tree properties across geographical regions and tree size ranges. The models were developed using a mixed modelling approach accounting for the hierarchical structure of the data that originated from sample plots. Diameter at breast height, tree height, and the ratio between height and diameter were candidate explanatory variables, but only diameter was statistically significant (p<0.05). The model fit values (pseudo-R 2) were 0.91, 0.60 and 0.85 for the three respective models. A substantial part of the model random errors could be attributed to between-plot variations. The conclusion related to objective (1) was that the models are well suited for biomass prediction of mountain birch in the mountain areas of southern Norway. Furthermore, models reported in previous research that had been calibrated on data from other regions were applied on the current data set. The results indicate that models calibrated for small trees produced predictions diverging from the observed values of the current data set. The differences between predicted and observed values also seem to vary along a site productivity gradient. Still, even though the differences between predicted and observed values using the different models varied quite a lot, the relationships were relatively stable within certain limits. The conclusion related to objective (2) was that biomass models can be applied outside the region for which they were developed, which in many cases is necessary because local models do not exist. However, the properties of the model development data related to tree size range and site productivity should be similar to those of the area for which predictions are being made.
Document Type: Research Article
Publication date: 2009-08-01