Allometric equations for estimating stem biomass of Artocarpus chaplasha Roxb. in Sylhet hill forest of Bangladesh

Abstract

Accurate tree biomass estimation is crucial for management of forest stand either in term of conservation values or in sustainable management. The main objective of this study was to obtain the best-fit model for predicting stem biomass of Artocarpus Chaplasha in Sylhet hill forest region. In this study, 157 individual tree data from two separate national parks were used. The most widely used logarithmic allometric models were developed and compared. Commonly used parameters, such as R2, RSE, MAB, AICc and different statistical tests (such as Durbin–Watson for checking autocorrelation of residual, Shapiro–Wilk test for residual distribution) were used in model selection, where we found model 3 and model 4 having two predictor variables, i.e. tree diameter at 1.3 m (D) and tree height (H) as the best-fit models providing highest R2; lowest RSE, MAB and AICc values. The bias corrected best-fit models were stem biomass (kg) Y=0.0158×D1.7928H1.3198andY=0.0121×(DH)1.6342 which showed low RMSE% and MPE% values compared to previously published one. Though the best-fit models’ diagnostic results showed slight heteroscedasticity of its residuals distribution, they were normally distributed and there were no significant autocorrelation. The results of this study have implications on estimation of tree level biomass and carbon stocks of forests significant for forestry related mitigation options of climate change such as “Reducing Emissions from Deforestation and Forest Degradation (REDD+)” Program.

Keywords

• Allometric equations
• Stem biomass
• Log-linear model
• Non-destructive