A non-destructive approach to develop tree-level allometric equations for estimating aboveground biomass in the forests of Tripura, Northeast India

Forest tree biomass estimation is a significant issue for the forest management system and to mitigate climate change. Aboveground biomass (AGB) is an indicator of the potential productivity of the ecosystem. However, an accurate estimation method of carbon stock is important for perfect carbon accounting. So, selecting a more appropriate method for biomass assessment is crucial to reaching the goal. Though allometric equations have been used to estimate AGB for both individual species as well as multiple species for regional and pan-tropical scales, species-specific models for a smaller area produce better accuracy. The current initiative has been taken to develop individual tree-level species-specific and generalized local allometric models in a non-destructive manner for the Tripura state of Northeast India, which was the main limitation of carbon stock estimation in the region. Five different forms of log-transformed linearized power equations were tested in the regression models, with AGB as the response variable. The best fit models were chosen based on adjusted R-2, F-statistic, Akaike information criterion, Breusch-Pagan test, Variance Inflation Factor, and three assumptions of linearity. A paired t test was conducted between the predicted and observed AGB values. The best fit generalized equation developed in this study was found to be AGB = 1.03 x exp (- 3.95 + 1.09 ln D-2 + 0.97 ln H - 0.20 N). Compared with some existing pan-tropical and regional models for their predictive accuracy, this equation expressed the highest R-2 value of 0.9089 in an independent testing dataset. When measuring the tree height is difficult due to a closed canopy, the equation AGB = 1.07 x exp (- 2.77 + 2.55 ln D) can be applied to obtain a quick estimation using just the DBH. The equations presented here may be the instrument for estimating carbon stock at the local level.