Tree above-ground biomass allometries for carbon stocks estimation in the natural forests of Colombia

In this study, we analyzed the above-ground biomass data for 631 trees with a diameter >= 10 cm from different biogeographical regions in Colombia. The aims of this research were (1) to evaluate the accuracy of the most commonly employed pantropical allometric models for the estimation of above-ground biomass of natural forests in different sites located along a complex environmental gradient, (2) to develop new models that enable more precise estimations of current carbon stores in the above-ground biomass of natural forest ecosystems in Colombia, and (3) to evaluate the effect on allometric models of forest type classifications as determinants of above-ground biomass variation. The Brown et al. (1989) model for moist forests, which includes diameter, height, and wood density, showed the overall best performance in Colombian sites. The Type II models of Chave et al. (2005; hereafter Chave II), which include diameter and wood density but not height, tended to strikingly overestimate the above-ground biomass (54.7 +/- 135.7%) in the studied Colombian sites. The use of forest classification based on the life zone system systematically led to better statistical models to estimate AGB at the individual scale and site scale than the use of Chave's classification. Our results propose that Chave II models should be evaluated prior to their use for a given ecosystem. For Colombia, the new allometric models developed, which employed diameter, wood density, and height, could help improving our understanding of the carbon cycle. Forest type classification was found to be an important determinant of the above-ground biomass estimation when altitudinal and other complex environmental gradients are included. The new models presented here can be considered as an alternative option for assessing carbon stocks in the above-ground biomass of natural forests in neotropical countries. (C) 2011 Elsevier B.V. All rights reserved.