Conventional and additive models for estimating the biomass, carbon and nutrient stock in individual Shorea robusta Gaertn. f. tree of the Sal forests of Bangladesh

Abstract

Accurate tree biomass estimation is critical and crucial for calculating carbon stocking as well as for studying climate change, forest health, productivity, nutrient cycling and budget etc. A total of 50 individuals of Shorea robusta Gaertn. f. were harvested to assess the biomass of tree components (leaf, branch, bark and stem). Carbon and nutrients (N, P and K) content in the tree components were also measured. This study adopted component biomass models with cross-validation technique. Additive biomass models were developed following the modified Gaussian maximum likelihood aggregated approach using open source software R (version 4.0.1). Component and additive biomass model used D (Diameter at Breast Height) as a sole predictor performed satisfactorily, the inclusion of total tree height (H) in Da*Hb form showed its supremacy over all the models. The best fitted additive model (AGB = 0.002056*D2.923998*H−0.69278 + 0.00848*D2.3896*H0.29648 + 0.04224*D2.06986*H0.65549 + 0.00552*D2.06723*H0.70536) and conventional model (Ln (AGB) = -2.7977 + 2.1829*ln(D) + 0.5073*ln(H)) took the lowest AIC, MPE and MAE and the highest model efficiency values. The derived species-specific additive and non-additive model showed its superiority over the frequently used pan-tropical models and suggested using this model for estimating aboveground biomass of S. robusta in Bangladesh.

Keywords

• Additive model
• Biomass allocation
• Carbon
• Conventional model
• Maximum likelihood
• Nutrients