Models for biomass prediction of Cunninghamia lanceolata tree and stands in Southeastern China

Large uncertainties still remain when using existing biomass equations to estimate total tree and forest stand scale. In this paper, we develop individual-tree biomass models for Chinese fir ( Cunninghamia lanceolata （Lamb.）Hook. ) stands in Fujian Province, southeast of China. For this, we used 74 previously established models that are most commonly used to estimate tree biomass, and selected the best fit models and modified it. The results showed the published model with ln(B) (biomass), ln(D) (diameter at breast height), (ln(H)) 2 , (total height) (ln(H)) 3 and ln(WD) (wood density) to be the best fitting model for estimating the tree biomass of Chinese fir. Furthermore, we observed that variables D, H (height), WD significantly correlated with the total tree biomass estimation model, as a result of it portraying the natural logarithm structure to be the best tree biomass structure. Finally, when a multi-step improvement on tree biomass model was performed, the analytic model with TV (tree volume), WD and BECF (biomass wood density conversion factor), achieved the highest accuracy simulation. Therefore, when combined with TV, WD and BECF to tree biomass volume coefficient bi for Chinese fir, the optimal model is the forest stand biomass (SB) estimation model, model with variables of stand volume (SV) and coefficient bi.