| Two methods to estimate aboveground biomass of individual trees from tropical forests are evaluated and refined. They are the scaling factor method and the derived local biomass equation method.; The scaling factor method assumes that stem wood density is constant for different tree sizes, and that the allometric relationship between crown and stem biomass is isometric. However, it was found that Vochysia ferruginia, a shade-intolerant species, does not have constant wood density across different tree sizes. It was also found that Tetragastris panamensis, a shade tolerant species, does not exhibit an isometric crown to stem biomass allometry. For this last species the scaling exponent (b{dollar}sb1{dollar}) in the constant allometric (CAR) model was greater than one, indicating that this species develops heavier crowns than expected.; The derived local biomass equation method requires a set of standard biomass equations to estimate tree biomass as a function of tree diameter and height. Several models were tested, and it was found that the best regression model was the Schumacher-Hall Model. This method also requires the development of a set of height-diameter functions for each forest stand in which the biomass equations are going to be applied. Several height-diameter models have been proposed with this objective. The model most frequently used assumes that the allometric relationship between tree height and diameter is constant. However, it was found that shade-tolerant and partial shade-tolerant tree species exhibit a variable height-diameter allometry, whereas Vochysia ferruginia, a shade-intolerant species has a constant height-diameter allometry.; It is concluded that only the derived local biomass equation method can be properly and practically applied to estimate aboveground biomass of individual trees in dense tropical wet forests.; This result is used to formulate a system of equations to estimate aboveground tree biomass for a tropical wet forest from the region of Sarapiqui, Costa Rica. A preliminary test of the system of equations was performed using the same data set from which the equations were developed. The bias percentage of the proposed system of equations is {dollar}-{dollar}0.185%. The standard deviation of the estimation errors is 383.99 kg/tree, which represents 24.0% of the mean of the observed values. |
