Amazon Rainforests May Be More Resilient to Atmospheric Warming Than We Thought: A Cross-Site Analysis of Eddy Flux Data from Natural Forests and an Artificially Warmed Forest

Abstract:

Biomass losses in tropical forests due to rising temperatures could accelerate regional and global climate change via reduced land-surface feedbacks associated with the carbon cycle. However, the response of tropical vegetation to such increased temperatures is highly uncertain. Current vegetation models may fail to capture resilience in plant response as they do not appropriately account for photosynthetic acclimation and typically allow for relatively coarse changes in plant community composition. Laboratory, greenhouse, and experimental studies from other biomes provide evidence that acclimation and assembly changes occur, suggesting more plant- and community-organisational changes to temperature that create resilient functional shifts to this environmental driver. Experimental data on how tropical forests respond to increased temperatures is still limited. We assess the resilience of tropical forest carbon exchange to increased temperatures through experimental warming of a 0.2 ha artificial tropical forest: the Biosphere 2 tropical forest biome, B2-TF. The B2-TF experienced temperatures 5–7˚C higher than lowland Amazonia for over 20 years. We compare the temperature response of ecosystem carbon exchange in the B2-TF with several evergreen tropical forests in the Brazilian Amazon and a tropical dry forest in Mexico using eddy flux data. Our results show that the B2-TF sustains carbon assimilation at higher temperatures than in situ tropical forests. Whereas net ecosystem productivity at the Amazonian sites declines as temperatures increase above 28˚C, it is maintained to much higher temperatures in the B2-TF. The tropical dry forest also functions at higher temperatures than the evergreen forests in Brazil but carbon uptake declines at temperatures below those where uptake is sustained in the B2-TF biome. If the B2 forest function is a proxy for how real-world forests respond to temperature, this result suggests that these forests may have acclamation and assembly responses that allow them to tolerate greater temperature increases than models currently assume. Mechanism for this tolerance (not yet resolved in this analysis) could be due to differing responses to vapour pressure deficit inside B2, to true thermal acclimation of photosynthesis, or to shifts in tree community composition.