Additive biomass model for Heritiera fomes (Buch.-Hum.) in the Sundarbans Reserved Forest, Bangladesh

Abstract

Mangroves are recognised as an important carbon sequester and therefore demand accurate biomass and carbon stock estimations. This study aimed to develop additive biomass models for Heritiera fomes, the most dominant tree species of the Sundarbans Reserved Forest in Bangladesh. Using a non-destructive method, 219 small branches (diameter < 7 cm) were harvested from 97 individual trees to develop biomass models for leaves and smaller branches. The biomass of bigger branches (diameter > 7 cm) and stem was calculated from the volume and mean wood density value after debarking while the biomass of all other components was derived from the determined fresh to oven dry weight conversion ratio. Finally, the biomass of one individual tree was calculated by adding the biomass of trimmed and untrimmed leaves, small and large branches, foliage and stem. An independent data set was used to validate the best-fit model. A component-wise (leaves, branches, bark and stem) biomass model was developed by recovering subsequent cross-component correlations which were then aggregated using the weighted Gaussian maximum likelihood estimation method. Among the components model, D (diameter at breast height) alone performed best for leaves and branches while the product of D and H (total tree height) proved the better results for stem and bark. Our best-fit model (Biomass = 0.0389D2.3773 H0.4178 + 0.0492D2.3027 + 0.0112D1.1144 H1.4572 + 0.0306D1.8507) showed the highest model efficiency with the lowest AIC, RMSE%, MAE, and MPE values. The efficiency of our non-destructive model has shown that it is as effective as other widely used pan-tropical models. Our built models can therefore be used for accurate estimation of biomass and carbon stock in H. fomes of the Sundarbans Reserved Forest, Bangladesh.