After cutting for timber or agriculture, forests often regrow naturally. As they regrow, there is typically a succession of different species that rise and fall in abundance over time, sometimes ending with a forest dominated by the same species of trees that were in the forest before cutting. However, some tropical forests seem not to recover this way: even after many years of natural regeneration, they fail to return to the same composition of species that would be found in a mature forest. This project will test what may prevent the recovery of tropical forests after clearing. Research will focus on two possible factors, the inability of some species to spread their seeds into the regrowing areas, and the inability of seedlings with certain traits to survive and mature under the conditions left after clearing. The project will measure the key traits of the dominant species in regrowing forests, termed secondary forests, and mature forests, and then observe how different traits may be favored in the different types of forest using annual censuses of trees made over 13 years in six secondary and two mature forests in Costa Rica. A field experiment will directly compare the survival and growth of seedlings with contrasting traits in pasture, secondary forest, and mature forest.

Besides helping to answer basic scientific questions about how the functional capabilities of species determine the ability of forests to recover after disturbance, this project is designed to improve forest restoration in the tropics. The research will show what types of species are missing from regenerating forests and test if planting these species can speed regeneration. Results will be widely communicated, and data on species traits will be incorporated into global databases. In addition, the project will employ and train both local Costa Rican assistants and undergraduate students from the U.S., and work with mentoring programs.

Project Report

Intellecutal Merit: Over half of the Neotropical forests have been damaged by human activity, often converted to pasture or agriculture. Under the right conditions, second-growth forests can naturally regenerate once these lands are abandoned, providing a means of recuperating biodiversity and ecosystem services. However, the mechanisms of community assembly during the process of natural regeneration are not fully understood. My dissertation research assessed the relative influences of environmental filtering and stochasticity in the regeneration of lowland wet tropical forests in Costa Rica. I measured leaf and stem functional traits (species-specific morphological and physiological attributes) to quantify the ecological characteristics of trees and palms. The National Science Foundation Doctoral Dissertation Improvement Grant supported three main research projects: 1) Exploring changes in species composition and functional trait composition over successional time among two old-growth and six second-growth stands ranging in age from 10–40 years after abandonment of pasture in the lowland wet tropical forest region. 2) Comparing patterns of taxonomic and functional beta diversity over successional time to assess if functional characteristics of the communities are converging at a different rate from the species composition. 3) Using an experimental approach to test the ability of species with different functional traits to survive and grow under three habitat treatments (pasture, second-growth forest, and old-growth forest) representing restoration scenarios. As predicted with environmental filtering, some of the community-level functional traits changed predictably across the successional gradient, demonstrating a decrease in traits related to faster growth rates, and an increase in traits conferring higher tissue defense. However, other functional traits did not demonstrate predictable changes with succession. For these traits, stand identity instead of stand age explained most of the variation among communities, supporting a stochastic model of community assembly. Functional turnover among plots was largely driven by differences in species composition, and did not show a directional trend with differences in stand structure (i.e. basal area difference). At these intermediate to late stages of succession in wet tropical forests, community assembly appears to be driven primarily by un-quantified site level differences (i.e. stochasticity), and not by changing environmental conditions (i.e. environmental filtering). The results of the tree seedling experiment in a wet tropical forest region demonstrated that across a range of environmental conditions, species commonly found in old-growth forests survive better than species commonly found in second-growth forests. Both types of species grow equally well across the different habitats. These findings suggest that old-growth specialists should be used for restoration (both direct planting and enrichment planting). Broader Impacts: This dissertation research provided employment and ecological research training for six local Costa Rican technicians, and eight US undergraduate students. Ph.D. candidate Vanessa Boukili closely mentored two of the US undergraduates in performing independent research projects. One of these students joined a Ph.D. program after her work with Boukili. Both of these students presented their research findings at an international ecology conference in the summer of 2013, and are currently preparing manuscripts for publication. In addition to writing 4 first-authored manuscripts of her dissertation research, Boukili presented her results at 3 national and 1 international scientific meetings. She was also a co-author on four additional manuscripts that have either been published or are in review, and three additional presentations. Boukili is an online scientist mentor for Planting Science (, for which she guides middle and high school students through the scientific process as they complete their own research projects.

National Science Foundation (NSF)
Division of Environmental Biology (DEB)
Standard Grant (Standard)
Application #
Program Officer
Peter Alpert
Project Start
Project End
Budget Start
Budget End
Support Year
Fiscal Year
Total Cost
Indirect Cost
University of Connecticut
United States
Zip Code