In the lowland Amazon, we observe substantial turnover in tree species composition across the landscape, especially among terra firme, flooded and white-sand forests. The predominant explanation for habitat specialization invokes allocation-based tradeoffs in response to different biotic and abiotic selection pressures. Along with international collaborators, we are testing the growth-defense trade-off for habitat specialization in Amazonian trees. For this project, we are working on characterizing insect community structure across wide gradients of soil resource availability in Peru and French Guiana. In addition, we are monitoring leaf production and herbivory rates on common species and focal tree genera across this same gradient. Further, we have established a reciprocal transplant experiment across habitats and countries that includes an herbivore exclusion treatment to conduct a controlled test of herbivory rates and their consequences for plant performance across the geographic and environmental gradient.

One of our major objectives is to describe functional strategies of tropical trees and their relationships with species performance and distributions. We want to better understand the processes contributing to community structure and ecosystem function. For this, we sampled leaves and wood from over 5000 trees and measured functional traits related to resource acquisition, growth and defense. This project represents the largest databases for DNA, functional traits and plant chemical defenses of any tropical forest region. The integration of these data has permitted rigorous tests of community assembly.

In collaboration with international teams, we are describing whole plant level variation in functional traits, including root and stem tissue densities, anatomy and chemistry. Using samples collected from permanent plots we have installed in lowland forests of Peru and French Guiana, we are seeking to test for coordinated allocation strategies among tissue types and the consistency of this relationship across habitats differing in resource availability and herbivore pressure. Further, in collaboration with teams from Berkley, we are expanding our investigations of plant chemical defenses beyond the presence-absence data of volatile molecules to a total metabolite approach for several widespread Amazonian tree lineages (Bombacoideae, Protieae, Swartzia, Inga, Micropholis, Eschweilera).

In collaboration with the University of Florida and a consortium of local partners, we have worked since 2005 to address impacts of the highway paving in the southwestern Amazon in Peru, Bolivia and Brazil. We are currently completing a large interdisciplinary project that examines social and ecological resilience in forest communities of the southwestern Amazon where the transnational highway has recently been paved. The project integrates remote sensing analyses with community and household inventories and descriptions of forest structure and plant diversity.

We work with botanical teams at three participating universities in Puerto Maldonado, Peru (UNAMAD); Cobija, Bolivia (CIPA-UAP); and Rio Branco, Brazil (UFAC) to build student teams that have now established more than 75 new permanent plots in the region. We are integrating this information with the remote sensing and interview data to test for effects of connectivity to population centers on forest structure and biomass, commercial resources, and species composition.