Abstract: A total of 333 trees from China fir,masson pine,eucalyptus,Hard broad-leaf and soft broad-leaved species growing in Guangxi province were collected and their biomass was measured by harvest method.Three equations,W＝c1×D^c2,W=c1×(D²H)×c2 and W＝c1×D^c2×H^c3,were applied to above-ground biomass (AGB) fitting of each tree-species group,and the heteroscedasticity of equations was eliminated by weighted-regression.Roots biomass of some tree samples were employed to model the equations of root-to-shoot ratios.The results showed that:(1) the precision of AGB model based on W＝c1×D^c2×H^c3 was higher;(2) weighted regression slightly decreased the precision of AGB equations,but enhanced the stability of equations;(3) determination coefficients of AGB equations were greater than 0.9,and the total relative error (TRE) and mean systematic error (MSE) were basically controlled within ± 1%,then the estimate precision of biomass equations was about 94%.Therefore,combing with the equations of root-to-shoot ratios,these above-ground biomass equations were suitable to estimate the biomass of the major tree species of Guangxi in a certain range of diameter at breast height (DBH) class.

Key words: forest biomass, biomass equation, weighted regression, root-to-shoot ratio, Guangxi Province