McGill University
Ph.D. candidate
Supervisor: Brian Leung
Fiona Soper, McGill University
Start: 2020-09-08

Project

Modelling Secondary Forest Growth in Latin America
Given the pressing climate crisis, regenerating deforested land has become an urgent matter. Secondary forests cover over half of tropical forests, and a third of the regrowing area is on soil previously abandoned by agriculture. Land use practices have been shown to greatly affect soil fertility and consequently, the regrowth rates of future forests that may establish on exhausted farmland. This factor, however, is currently unexplored by existing forest regrowth models. Accurately estimating carbon accumulation is critical to policymaking, the establishment of protected areas, and making climate projections. This research project focuses on modelling the future of carbon accumulation in Neotropical secondary forests, exploring the effects of land use history and biodiversity: The project encompasses three aims: 1. Modeling Forest Regrowth in the Brazilian Amazon: Utilizing remotely sensed land use maps encompassing 33 years of history and LiDAR-based global carbon storage data, I build a semi-mechanistic statistical model to predict forest regrowth in the Brazilian Amazon given environmental and anthropogenic predictors. By improving on existing regrowth models, we can better assist decision making on conservation strategies and improve future carbon sequestration projections. 2. Extending the Model to Central America: The model is expanded to Panama and Costa Rica, vital biodiversity hotspots. By classifying land use through machine learning, the project accounts for local agricultural practices and explores its application to different socioeconomic and ecological scenarios from its initial training in the Amazon basin. With this, we expand the usefulness of the model to management in other countries, and we open the possibility for future research in the Central American context. 3. The Biodiversity-Biomass Relationship: Biodiversity metrics, obtained through the S2BaK bias-adjusted biodiversity model for Panama and Costa Rica, are integrated into the model. We explore the effect of alpha diversity on the predictive ability of the model. I incorporate the presence of functional groups that have been shown to influence regrowth, such as nitrogen fixers and invasive grasses. This will add more evidence to a highly debated field, as well as explore the possibility of improving future carbon accumulation predictions by incorporating biodiversity metrics. The outcomes of this research improve on previous forest regrowth models and provide insight into biodiversity's impact on biomass. By addressing the complexities of regenerating deforested land, this project contributes to more accurate conservation strategies, sustainable land use policies, and comprehensive ecological understanding in the Neotropical region.

Keywords

secondary forest, regrowth, Statistical modelling, biodiversity, biodiversity/biomass relationship, land use history