Site-specific versus pantropical allometric equations: Which option to estimate the biomass of a moist central African forest?

Fewer data on tree biomass and allometric equations are available in African tropical moist forests than in the other tropical continents. When needing a biomass allometric equation, one thus faces the dilemma of using either a pantropical equation with the risk that it is biased for Africa, or a site-specific equation that is imprecise. Using a data set on aboveground biomass for 101 trees destructively measured at Zadie in northeastern Gabon, we fitted site-specific allometric equations and assessed the validity of ten existing equations. The best fitted model without height as a predictor was: B = exp[-4.0596 + 4.0624 in D - 0.228(InD)(2) + 1.4307 In rho], whereas the best fitted model with height was: B = exp[-2.5680 +0.95171n((DH)-H-2) + 1.1891 In rho], where B is the aboveground biomass in kg, D the diameter at breast height in cm, H the height in m, and rho the wood density in g cm(-3). Separate allometric equations for the stem, stump, foliage and branches were also fitted. Chave et al. (2005)'s pantropical equations for moist forests, that are currently the most commonly used allometric equations in central Africa, were not valid at Zadie with an overestimation of biomass of about 40%. The allometric equations of the same authors for wet forests were valid at Zadie, even though the climatic zone does not correspond. More data on tree biomass are needed in central Africa to explore the natural range of variability in tree biomass and identify the factors that influence variations among sites. (c) 2013 Elsevier B.V. All rights reserved.