Higher functional diversity improves modeling of Amazon forest carbon storage

Abstract

The impacts of reduced precipitation on plant functional diversity and how its components (richness, evenness, divergence and composition) modulate the Amazon carbon balance remain elusive. We present a novel trait-based approach, the CArbon and Ecosystem functional-Trait Evaluation (CAETÊ) model to investigate the role of plant trait diversity in representing vegetation carbon (C) storage and net primary productivity (NPP) in current climatic conditions. We assess impacts of plant functional diversity on vegetation C storage under low precipitation in the Amazon basin, by employing two approaches (low and high plant trait diversity, respectively): (i) a plant functional type (PFT) approach comprising three PFTs, and (ii) a trait-based approach representing 3000 plant life strategies (PLSs). The PFTs/PLSs are defined by combinations of six traits: C allocation and residence time in leaves, wood, and fine roots. We found that including trait variability improved the model's performance in representing NPP and vegetation C storage in the Amazon. When considering the whole basin, simulated reductions in precipitation caused vegetation C storage loss by ∼60% for both model approaches, while the PFT approach simulated a more widespread C loss and abrupt changes in neighboring grid cells. We found that owing to high trait variability in the trait-based approach, the plant community was able to functionally reorganize itself via changes in the relative abundance of different plant life strategies, which therefore resulted in the emergence of previously rare trait combinations in the model simulation. The trait-based approach yielded strategies that invest more heavily in fine roots to deal with limited water availability, which allowed the occupation of grid cells where none of the PFTs were able to establish. The prioritization of root investment at the expense of other tissues in response to drought has been observed in other studies. However, the higher investment in roots also had consequences: it resulted, for the trait-based approach, in a higher root:shoot ratio (a mean increase of 74.74%) leading to a lower vegetation C storage in some grid cells. Our findings highlight that accounting for plant functional diversity is crucial when evaluating the sensitivity of the Amazon forest to climate change, and therefore allow for a more mechanistic understanding of the role of biodiversity for tropical forest ecosystem functioning.