PS 39-188
Competitively-optimal plant carbon allocation increases the drought resistance of forest ecosystems in a next-generation dynamic global vegetation model

Tuesday, August 11, 2015
Exhibit Hall, Baltimore Convention Center
Tao Zhang, Department of Biology, University of Florida, Gainesville, FL
Jeremy W. Lichstein, Department of Biology, University of Florida, Gainesville, FL
Ensheng Weng, Princeton University, Princeton, NJ
Caroline E. Farrior, Ecology and Evolutionary Biology, Princeton University, Princeton, NJ
Sergey Malyshev, Princeton University
Elena Shevliakova, Ecology and Evolutionary Biology, Princeton University, Princeton, NJ
Ray Dybzinski, Princeton University, Princeton, NJ
Richard Birdsey, Forest Service, Newtown Square, PA
Stephen W. Pacala, Ecology and Evolutionary Biology, Princeton University, Princeton, NJ

Plant functional diversity and resource competition are represented in a simplistic manner in most of the dynamic global vegetation models (DGVMs) currently used to study interactions between the terrestrial biosphere and Earth’s climate. For example, most DGVMs represent global plant functional diversity with ~10 plant functional types (PFTs) with fixed properties. Thus, the vegetation in each biome is often represented by a single PFT, despite the large amount of functional diversity that in reality exists within each biome. Although it is widely believed that this simplification can lead to unrealistic model behavior, there have been few attempts to quantify the sensitivity of DGVM predictions to different representations of plant functional diversity. In this study, we used a next-generation DGVM, LM3-PPA, to explore how the representation of plant functional diversity affects the response of the forest carbon cycle to hypothetical severe drought in temperate forests in the northeastern U.S. In particular, we explored how drought sensitivity of late-successional forests depends on (1) the allocation of plant carbon to leaves, wood, and fine roots; and (2) how the identity of the dominant late-successional type in a given model grid cell was determined (competitively optimal strategy vs. biomass-maximizing strategy).


The competitively optimal late-successional strategy (i.e., the fractional allocation of carbon to leaves, wood, and fine roots that outcompetes all other strategies in the absence of disturbance) differed across model grid cells with distinct climate regimes. Biomass-maximizing strategies also varied across model grid cells and tended to have less fine root allocation than the competitively optimal strategy. These results suggest that (1) no single PFT can adequately represent vegetation dynamics, even within a single biome and region (temperate deciduous forest in the northeastern U.S.); (2) optimization algorithms that seek to maximize biomass do not accurately represent the outcome of competition; and (3) competition drives plants to allocate more carbon to fine roots compared to allocational strategies that would maximize terrestrial carbon storage under constant climatic conditions. However, when exposed to severe and extended drought, the relatively high fine-root densities of competitively optimal strategies resulted in greater drought resistance (i.e., smaller change in carbon storage and tree mortality rates) than biomass-maximizing strategies. Together, our results imply that competition drives trees to increase fine-root allocation, which decreases terrestrial carbon storage under a constant climate, but also confers resistance to severe drought.