2017 ESA Annual Meeting (August 6 -- 11)

COS 185-5 - Integrating fire-related PFTs into predictions of large-scale forest responses to fire in Amazonia

Friday, August 11, 2017: 9:20 AM
E146, Oregon Convention Center
Carla Staver, Ecology and Evolutionary Biology, Yale University, New Haven, CT and Paulo M. Brando, Woods Hole Research Center, Falmouth, MA
Background/Question/Methods
Fire is an ecological process that is fundamental in determining global biome distributions and carbon dynamics. Some biomes – such as savanna and some Mediterranean ecosystems – are fire maintained; they are characterized by frequent or at least regular fires that have driven the evolution of fire-tolerant and fire-dependent flora. The role of fire is less often considered in other biomes, such as tropical forest, that are less flammable and where fires are less frequent. However, recent work has suggested that fires are increasing in prevalence in some tropical forests, including in the southern Amazon. These fires can have unparalleled impacts on forest community dynamics and on forest carbon stocks, but quantification of fire impacts on Amazonian forests is restricted to a few sites and estimates of their extent are thus largely anecdotal. Here, we examine how the fire-related functional traits of forest trees (from sites throughout the Amazon) influence the susceptibility of forests to fire-driven carbon losses. We combine these ecosystem-level functional trait evaluations with current and future estimates of fire risk (from remote sensing and global fire model predictions) to predict basin-wide estimates of fire impacts on the carbon cycle in the Amazon.

Results/Conclusions

Analyses suggest that bark thickness varies systematically across the Amazon with respect to climate (with drier sites having substantially thicker bark than wetter ones). Although bark is not solely a fire-related trait, past work has established that bark thickness is strongly predictor of tree mortality in forest-understory fires. Using these relationships, combined with maps of current and future projected fire risk, we find that explicitly considering functional traits impacts on our prediction of fire-driven carbon losses. This is especially true in drier forest, where forest communities are relatively tolerant of relatively mild understory fires. However, regions with elevated fire risk under current and future climate change scenarios (including the southern Amazon) are at significant risk of substantial fire-driven losses in fires.