A data set with 315 trees with diameter at breast height (dbh) greater than 5 cm was used to test four statistical models - linear, non-linear and two logarithmics - to estimate aboveground biomass of standing trees. The data were collected destructively near Manaus, Central Amazonia, on a site covered by a typical dense "terra-flrme" moist forest on plateaus dominated by yellow oxisols. The difference between observed and estimated biomass was always below 5%. The logarithmic model using a single independent variable (dbh) produced results as consistent and precise as those with two variables (dbh and total height). Besides statistical models to estimate aboveground biomass, the following information are also presented in this paper: the contribution of each free component (stem, branch, twigs, leaves and flowers or fruits) to the total weight of a standing tree, water content to estimate the dry weight and carbon concentration of each tree component.
Carbon; forest management; statistical models; allometry