Large trees as key elements of carbon storage and dynamics after selective logging in the Eastern Amazon

The long term effect of Reduced-Impact Logging (RIL) on above-ground live biomass (AGB) dynamics was investigated in 18 1-ha logged over permanent sample plots set up in a terra firme rain forest in the Eastern Amazon (Brazil, Paragominas). Both tree survival and growth were investigated among three Diameter at Breath Height (DBH) classes (20-40, 40-60, P60 cm) to assess the contribution of each DBH class to the post-logging AGB recovery. Before logging, mean tree density and AGB per plot (dbhP20 cm) were 187 ± 14 trees ha_1 and 377.6 ± 62.8 Mg ha_1 respectively. Although big trees (dbhP60 cm) only represented 9.3% of the total tree density, they gathered almost half of total AGB. During the post-logging period (8 years), the mortality of large trees was found to drive the annual net changes and largely overcame the AGB gain in the smaller DBH classes. Indeed, plots with high post-logging mortality of large trees showed negative carbon balance t over the study period (8 years). The over mortality of large trees injured by logging contributed significantly to the annual AGB losses (up to 40%) in the first years after logging. Due to the overwhelming importance of this size class in carbon stocks and dynamic, reducing logging damages and intensity might have great impact in the post-logging biomass dynamics. We estimated that reducing logging intensity from 6 to 3 stems ha_1 would save 27.7 Mg C ha_1 for a 35 years rotation cycle. To compensate this loss of profits, compensatory payments of avoided CO2 emission should worth US $ 6.5/Mg of CO2. This price falls into the range of prices of the international carbon market. Sustainable forest management aiming at enhancing carbon stocks could therefore promote the preservation of the large trees. At our study site, we recommend the adoption of a maximum diameter cutting limit of 110 cm. (Résumé d'auteur