Light-driven intraspecific variation in leaf traits is critical for simulating ecosystem dynamics with process-based models in tropical forests

Background/Question/Methods

Plant functional diversity is critical for predicting ecosystem responses to climate variability in terrestrial biosphere models. Models often represent functional diversity across species by a few plant functional types (PFTs) defined by climate, life form, and phenology. However, such PFT-based approach ignores intraspecific trait variation in response to fine-scale environment heterogeneity. For closed-canopy forests, trait plasticity driven by within-canopy light gradient is particularly important because it can be essential for simulating the survival of understory plants and thus long-term ecosystem dynamics, particularly in dense tropical forests. Therefore, our main objectives are to 1) quantify light plasticity of mature trees in a tropical forest; 2) investigate how light plasticity influences vegetation demography; 3) discuss how the demographic effect scales up to long-term ecosystem functioning. We conducted model experiments using the Ecosystem Demography model version 2, which is a terrestrial biosphere model that simulates vegetation demography and land surface carbon, water, and energy balances. We used trait data of both canopy and understory leaves from species at Barro Colorado Island (BCI) to parameterize the light plasticity of leaf dark respiration (Rd), maximum carboxylation rate (Vcmax), specific leaf area (SLA), and leaf longevity. BCI forest census information served as model benchmark.

Results/Conclusions

Trait analysis demonstrates large variation in the magnitude of light-driven plasticity among BCI species. Rd and Rd light plasticity are positively correlated, and Vcmax plasticity increases with higher Vcmax, which are consistent with previously reported relationships in the Amazon. SLA experiences similar degree of plasticity as photosynthetic traits, but its plasticity is not related to any single trait. Compared to the default model, the model with light plasticity (ED2-plastic) increases growth rate and decreases mortality rate in small trees (DBH < 20cm). The basal area estimate of small trees is doubled in ED2-plastic and is more consistent with field observation. Light plasticity increases LAI by 33%, and leaf area profile indicates that this increase is primarily attributed to a much denser understory. Model experiments are still in progress and we will evaluate the long-term consequences of plasticity, including carbon dynamics and relative abundance of species. These preliminary results suggest that 1) photosynthetic trait values are reasonably good predictors of their light plasticity; 2) light plasticity is necessary for predicting plant demography and forest structure. We expect that the plasticity effect will be particularly pronounced in highly variable light environments, such as regrowing or post-disturbance forests.