Burning of Amazonian forest in Ariquemes, Rondonia, Brazil: biomass, charcoal formation and burning efficiency

Biomass burning in tropical forests-the normal practice to prepare land for agriculture and ranching-has been a major source of CO2 emitted to the atm osphere. Mass transformations by burning are still little studied in the tr opics, The present study estimated parameters, such as the stock of carbon contained in the biomass, burning efficiency and the formation of charcoal and ashes in a tropical moist forest. Two sets of plots arranged in the for m of 'stars' (720 m(2) total) were installed in a 3.5 ha area of forest tha t had been felled for planting pasture at Fazenda Nova Vida, Ariquemes, Ron donia. Each 'star' had six rays measuring 2 m x 30 m; alternating rays were designated for pre-burn and post-burn measurements, All above-ground bioma ss present in the plots was weighed directly before the burn in the pre-bum rays and after the burn in the post-burn rays. Pieces of wood with diamete r greater than or equal to 10 cm also had their biomasses estimated from vo lume estimates, using line-intersect sampling (LIS) in order to increase th e area of sampling and to allow volume loss to be estimated as an increment based on individual pieces measured before, and after, the burn at the sam e point (as opposed to inferring change as a difference between independent estimates of stocks). The initial above-ground biomass (dry weight) before the burn was estimated at 306.5 +/- 48.6 (mean +/- SE) Mg ha(-1), with an additional 4.5 Mg ha(-1) for trees left standing. Carbon stock in the initi al biomass (including trees left standing) was 141.3 (Mg C) ha-L. After bur ning, carbon stock was reduced by 36.8% (burning efficiency), The stocks of charcoal and ash formed in the burn were, respectively, 6.4 +/- 2.7 and 5. 7 +/- 1.0 Mg ha-The destructive and nondestructive (LIS) methods did not di ffer significantly (t-test, p > 0.05) in estimating post-burn stocks of woo d and charcoal. The results of this study contribute to improving the estim ates of parameters needed for global carbon calculations and point to ways in which estimates of these parameters could be further improved. (C) 1999 Published by Elsevier Science B.V. All rights reserved.