The Effects Of Tree Interactions On Tropical Forest Structures And Functions

Tropical forests are characterized by harboring amazing number of species. Complex and diverse interaction networks among organisms make tropical forests integrated systems and different from their componential parts(e.g. trees). Forest ecosystem is a typical hierarchical complex system, which is organized by interacting individual trees at the lower level. Different hierarchical levels of forest system are not independent with each other but have tight relationships through complex energy, material and information flows. For example, the lower-level trees and tree interactions could play important roles in the upper-level structure and function of forests(e.g. species composition and primary productivity); while the upper-level forest structure and states(e.g. vertical structure of forest could change light distribution inside forests) could also modify tree interactions. This thesis applied multi-scale insights and approaches and focused on tree interactions and their impacts on forest structures and functions through both the bottom-up and top-down processes. This study explored the following three questions:(1) how tree interactions affect rare species formation and persistence in a tropical forest;(2) how tree interactions affect forest primary productivity; and(3) how drought-mediated forest vertical structure and resource changes modify tree interactions. To explore these questions, this thesis applied individual-based hierarchical Bayesian models to study tree growth and survival and their responses to neighborhood dissimilarity in functional traits and phylogeny in a 50-ha tropical forest plot on Barro Colorado Island, Panama.First, this thesis found that both functional traits and phylogeny were important in determining tree interactions. Focal trees grew faster with neighbors more dissimilar in functional traits and phylogeny. The strength of positive neighborhood dissimilarity(ND) effect was correlated with focal species abundance in the community, with stronger effect for dominant than rare species. This asymmetric positive ND effect allowed rare species, which is of disadvantage in average growth rate under community average environmental state, to attain growth advantage relative to dominant species when rare species became extremely rare. Thus, the asymmetric positive ND effects play an important role in fostering species coexistence and shaping species abundance distribution. These results suggest that trait and phylogeny mediated tree interactions, a specific bottom-up process, are crucial for the tropical forest structure. Second, this thesis found that the positive ND effect operated at local scale could be scaled up to ecosystem scale and affect primary productivity. Community average ND increases with species richness in communities, which further increases average tree growth rate and the whole ecosystem primary productivity. That is, the positive ND effect on tree growth acts as complementarity effect at neighborhood scale, and leads to positive biodiversity-productivity relationship in this forest. Thus, this study provides a new insight(neighborhood complementarity) for the biodiversity-ecosystem functioning relationships in natural tropical forests. It also implies the importance of trait and phylogeny mediated tree interactions in affecting tropical forest functions. The results of the above two sections together suggest the important roles of bottom-up processes(tree interactions) in forest structures and functions. Finally, the thesis studied how a severe drought happened in this forest affected forest vertical structure and tree interactions. It found that the drought led to high mortality rate of large trees, which further allowed more light get into the forest floor. Furthermore, droughtmediated light level changes at the forest floor could modify tree interactions: during drought census interval, light enriched at forest understory, which relieved tree competition for light. The thesis found that the positive neighborhood trait dissimilarity effect on tree survival in normal period disappeared in drought census interval. This result suggest the importance of trait-mediated tree competition to predicting forest responses to drought. It also implies that tropical forest community structures and functions are subjected to top-down controls.The thesis greatly promotes the understandings on some important ecological questions of species coexistence, species-abundance distribution, biodiversity-productivity relationships and indirect responses to climate change of tropical forests. It also used case studies to illustrate a systematic epistemology about forest ecosystems. To get more comprehensive understanding on forests, we need to consider different hierarchical levels of forests and their interactions in integrated views. In other words, we need to see the forests and the trees.Finally, the thesis suggest that future research may benefit from considering the effects of intraspecific trait variation on tree interactions and their impacts on forest structures and functions, and multiple demographics process(e.g. growth, survival and recruitment) at different ontogenetic stages.