Allometric equations remain essential to the estimation of forest biomass and carbon (C) stocks. However, such equations have mainly been developed for American and European forests, and transferring them to other species or eco-regions is problematic. There are large forested areas in the eastern Himalayas, but we lack all but the most rudimentary biomass equations and this hinders reliable estimates of C stocks. We destructively harvested 144 trees with diameters ranging from 10 to 77 cm, from five different forest types along an altitudinal gradient from 317 to 3300 m a.s.l. and used these to construct allometric equations. Out of six models, we identified two that could predict AGB across a range of trees, including those at the upper and the lower ends of the scale. The models were mainly a function of diameter at breast height (DBH) with height (H) as an additional factor for lower altitude forest types. Inclusion of specific gravity (SG) of wood improved models for three of the five forest types. We provide both types of models and argue that wood SG should be collected during forest inventories as it is an important predictive variable especially for mixed species forests. The average deviation of the measured and estimated AGB ranged between 15.9 and 38.5 per cent for the various models for the different forest types. With the best-fit models, estimated aboveground C stocks increased with altitude from 57 to 207 Mg C ha⁻¹. The use of measured C concentration rather than an assumption of 50 per cent of biomass reduced estimated AGB C stocks between 6.8 and 8.6 per cent.

Food security policies can enable the use and consumption of wild foods as a complementary source of food and nutrition.