Development and validation of generalized biomass models for estimation of carbon stock in important agroforestry species

Estimation of biomass/wood production from timber species like Acacia nilotica, Dalbergia sissoo and Tectona grandis is required for research and merchant purposes. Accurate estimation of biomass can be done through destructive sampling, but that is cumbersome exercise. Alternatively, one can develop regression equations using easily measurable parameter like diameter at breast height (DBH). Usually biomass models are developed with the help of tree data for a particular location/area. Therefore, these models may not suitably be applied to other locations, as growth behavior of trees on other locations are not accounted for. In present study, generalized models for A. nilotica, D. sissoo and T. grandis have been developed and validated using available biomass equations, secondary data and primary data. The developed regression models were also validated on an independent dataset and found statistically good fit. In case of A. nilotica, model biomass of D. sissoo using D and D H as predictor variables and found Logistic and linear models as best B = 0.360 D^1.598 (R² = 0.926), for D. sissoo, exponential model B = 3.084 e^0.172D (R² = 0.924) and for T. grandis, parabolic model B= -22.262 + 2.845 D + 0.115 D² (R² = 0.951) were found good fit; where, B- biomass (kg tree⁻¹) and D- DBH (cm). On validation, these models gave an error of 0.536, 2.419 and 3.896 kg tree⁻¹, respectively in prediction of biomass. Hence, these models may be used for estimating biomass of A. nilotica, D. sissoo and T. grandis plantations in different regions.