The role of trophic interactions in shaping tropical tree communities

Abstract

Tropical rainforests contain exceptionally high biodiversity and account for >30% of the world’s carbon fixed by photosynthesis. Consequently, there are compelling reasons to deepen our understanding of the mechanisms that maintain these highly diverse forests and of the potential long-term threats to their preservation. An important process shaping tropical plant communities is negative density dependence (NDD). NDD occurs when plant performance is negatively impacted by increased neighbourhood density. Reduced performance at high neighbourhood density is thought to arise through ecological interactions between plants and their natural enemies. Thus in a healthy ecosystem, trophic interactions play vital roles as mechanisms driving NDD and are important as dispersers facilitating escape from NDD mortality. However, interruption to ecological processes caused by human activities, such as hunting, can perturb NDD interactions and cause cascading effects throughout an ecosystem. In my thesis I investigate the role of dispersal and mortality in NDD dynamics of tropical tree communities, as well as investigating local and global impacts of removing ecological interactions in tropical rainforests. In my thesis, I begin by addressing the presence and variation in strength of NDD among tree species and ontogenetic stages, the mechanisms driving NDD, and the role of trophic interactions in this process. The Janzen-Connell hypothesis predicts that host-specific natural enemies drive NDD by selectively reducing conspecific density, and increase diversity by suppressing competitive exclusion, thus allowing heterospecifics to persist. In chapters 2 and 3 of this thesis, I show that mortality driven by conspecific NDD is prevalent at the early life stages, and this effect is considerably stronger during the year after germination. Furthermore, this process is driven exclusively by host-specific fungal pathogens, which cause mortality selectively among conspecifics and drive diversity. As seedlings age beyond their first year, NDD interactions become less impacted by conspecifics but are impacted by closely related neighbours or by general neighbourhood density, representing changes in the mechanism driving NDD as seedlings age, and a decline in host-specificity of natural enemies. Equally, relative growth rates (RGR) are reduced under high neighbourhood density irrespective of species identity. Results suggest insect herbivores are the strongest driver of reduced RGR but not mortality under increased neighbourhood density. As a consequence of stronger inter than intra-specific NDD effects on RGR, insects had no impact on seedling diversity in the short term. This study supports assertions that regionally rare species experience stronger NDD than common species, accounting for the high variability in species relative abundance in the tropics. In the second part of my thesis, I address the role of large vertebrate dispersers in shaping tropical tree communities and the consequences of defaunation for tree assemblage and carbon storage. Dispersal allows seeds to escape NDD and persist to reproductive maturity and is therefore vital for the maintenance of diversity. Vertebrates disperse the seeds of more than 70% of neo-tropical tree species. However, many large vertebrates are becoming scarce due to widespread hunting. The decline of large vertebrates and their role as dispersers is predicted to alter tree community composition. Additionally, large vertebrates are responsible for the dispersal of large-seeded species, which are linked to species with high wood density. With wood density positively associated with carbon storage, there is a potential cascading influence of defaunation on global carbon storage. We investigate the consequences of declining large vertebrate mortality agents in chapter 3, and the consequences of declining large vertebrate dispersers in chapters 4 and 5. Although community composition is altered in a defaunated forest, species dispersed by extirpated fauna do not appear to drive this. In fact we find that many species thought to be heavily reliant on extirpated fauna manage to persist. Although it is thought that the simultaneous loss of seed predation from large terrestrial vertebrates may create compensatory effects, we found little support for this, with an absence of large terrestrial vertebrates driving only temporary changes to species diversity. Neither a loss of large frugivores or large-seeded species lead to declines in species with high wood density, but we detect a worrying decline in large stemmed species, which has negative implications for carbon storage. Overall, my thesis highlights the importance of NDD and trophic interactions, particularly fungal pathogens, at the early life stages in shaping tropical tree communities and in maintaining diversity. I provide evidence that the removal of trophic interactions among larger natural enemies and dispersers does not impact community assemblage in the directional manner found in previous studies. I provide evidence for the variability in response to trophic interactions among species and ontogenetic stages. I show disproportionate relative importance among natural enemies and dispersers in the maintenance of tropical tree assemblage, with implications for conservation and for assessing the consequences for tree diversity under the influence of degradation.