Decomposition rates of coarse woody debris in undisturbed Amazonian seasonally flooded and unflooded forests in the Rio Negro-Rio Branco Basin in Roraima, Brazil

Abstract

Estimates of carbon-stock changes in forest ecosystems require information on dead wood decomposition rates. In the Amazon, the lack of data is dramatic due to the small number of studies and the large range of forest types. The aim of this study was to estimate the decomposition rate of coarse woody debris (CWD) in two oligotrophic undisturbed forest formations of the northern Brazilian Amazon: seasonally flooded and unflooded. We analyzed 20 arboreal individuals (11 tree species and 3 palm species) with distinct wood-density categories. The mean annual decomposition rate of all samples independent of forest formation ranged from 0.044 to 0.963 yr−1, considering two observation periods (12 and 24 months). The highest rate (0.732 ± 0.206 [SD] yr−1) was observed for the lowest wood-density class of palms, whereas the lowest rate (0.119 ± 0.101 yr−1) was determined for trees with high wood density. In terms of forest formation, the rates values differ when weighted by the wood-density classes, indicating that unflooded forest (0.181 ± 0.083 [SE] yr−1; mean decay time 11–30 years) has a decomposition rate ∼19% higher than the seasonally flooded formations (0.152 ± 0.072 yr−1; 13–37 years). This result reflects the dominance of species with high wood density in seasonally flooded formations. In both formations 95% of the dead wood is expected to disappear within 30–40 years. Based on our results, we conclude that the CWD decomposition in the studied area is slower in forests on nutrient-poor seasonally flooded soils, where structure and species composition result in ∼40% of the aboveground biomass being in tree species with high wood density. Thus, it is estimated that CWD in seasonally flooded forest formations has longer residence time and slower carbon release by decomposition (respiration) than in unflooded forests. These results improve our ability to model stocks and fluxes of carbon derived from decomposition of dead wood in undisturbed oligotrophic forests in the Rio Negro-Rio Branco Basin, northern Brazilian Amazon. © 2017 The Authors