The future provision of ecosystem services by tropical forests depends on the direct and indirect effects of changes in the physical environment, biotic interactions and human actions. We report the results of studies that aim to improve our ability to predict the impact of two important components of this system: firstly, the risk that increased disturbance rates pose for forest composition, and secondly, the risk that drought poses to the carbon stored in intact forest and to initiatives that aim to reduce carbon emissions from deforestation. Both studies were based on long-term plots maintained by the RAINFOR network.

To understand the effect of treefall disturbance events on composition, we compared differences in the composition of stems 2-10 cm diameter between paired, disturbed and non-disturbed, 20 x 20 m subplots of existing forest plots, in four sites. The effect of drought on aboveground biomass was examined by analysing inventory data of trees ≥10 cm diameter from before and after the 2005 Amazonian drought, across 55 plots. We also explored the interactions between future drought frequency, the sensitivity of above-ground carbon stocks to these events, and the emissions reductions from avoided deforestation that a REDD project in southern Peru aspires to achieve.

Results/Conclusions

Disturbance events caused a consistent decrease in community-wide wood density within all four sites, despite the very different species pools. However, in one site (southern Peru), there were was no overall change in mean wood density with disturbance because even the most ‘undisturbed’ patches of this highly dynamic forest often have a functional composition similar to areas that have been recently disturbed. This result implies that western Amazonia forests may be more resilient than eastern Amazon forests to reductions in carbon storage caused by compositional change from increasing disturbance rates. The 2005 drought caused a marked increase in tree mortality across Amazonia, reversing previous trends of an increase in the aboveground living biomass of these forests. By using these data to estimate the impact of a predicted increase in drought frequencies, we calculate that drought could reduce the predicted emissions reductions achieved by a REDD project in southern Peru by 15 %.

Feedbacks between compositional change and carbon storage, and the risk posed by increased drought are two threats faced by tropical forests. Broad-scale, field-based evaluations of these and similar processes will assist in developing a predictive framework of the impacts of environmental change on tropical forests.