Improving Predictions of Tropical Forest Response to Climate Change

Background

Tropical forests provide many globally important ecosystem services, including carbon uptake and storage, climate regulation and biodiversity maintenance, all of which are threatened by climate change. However, predictions of how tropical forests will respond to climate change are highly uncertain. These uncertainties include a) algorithmic uncertainty in how key process responses (e.g. photosynthesis, respiration, mortality) are represented in vegetation models, b) uncertainty in parameters that define important climate sensitivities (e.g. drought mortality thresholds) and c) lack of representation of the diversity of tropical forest responses to climate in models. This PhD project seeks to quantify and reduce these uncertainties through advanced statistical analytics and machine learning applied to novel datasets, new process-based modelling and new targeted field data collection. The project will thus enable more robust predictions of climate change impacts on tropical forests.

PhD Opportunity

This PhD project integrates novel datasets of pan-tropical forest climate resilience with state-of-the-art analytical methods and vegetation modelling to address one of the most pressing questions in ecosystem ecology: how will climate change impact the future of the tropical forest carbon sink? Vegetation models vary greatly in their predictions of climate change impacts on tropical forests, even when forced with the same climatic conditions – i.e. forests in some models are more resilient than in others. The core challenge in this PhD is to use advanced uncertainty quantification (UQ) methods such as Bayesian parameter estimation and Gaussian Process emulation applied to process-based ecosystem models to systematically quantify, attribute and reduce uncertainty in projections of tropical forest responses to climate change.

The student will have access to novel large-scale datasets of tropical tree climate resistance/risk (e.g. tree hydraulic and thermal safety margins), tropical forest inventory data and high-resolution data from an ecosystem-scale drought experiment in southern Amazonia. These datasets will be integrated with process-based modelling tools, including the Trait-based Forest Simulator, which simulates the responses of individual trees to environmental change as well as Dynamic Global Vegetation Models used to simulate large-scale ecosystem dynamics. Application of UQ techniques will establish a robust framework for linking data and models and enable rigorous evaluation of parameter and process uncertainty, enabling new insights into how climate resilience is structured across tropical forests and how it is shaped by functional diversity.

The student will be co-supervised by Prof. Emanuel Gloor (ecosystem modelling), Prof. Oliver Phillips (forest inventory analysis) and Dr. Chetan Deva and (climate, statistics). Although focus is on modelling and statistical analysis of large datasets, targeted field data collection is possible.

Applicant Profile

Students with strong quantitative skills who have an interest in applying those skills to address important ecological questions.

Other Information

THERMOS Project: https://environment.leeds.ac.uk/dir-record/research-projects/2126/thermos-thermal-safety-margins-of-earth-s-tropical-forests

Southern Amazon Drought Experiment: https://www.leeds.ac.uk/directory-record/22847/creating-a-drought-to-predict-the-amazon-tipping-point

Papers

Aguirre-Gutierrez J, …, Phillips O, Galbraith D, Malhi Y. 2025. Tropical forests in the Americas are changing too slowly to track climate change. Science 387:6738. (https://www.science.org/doi/10.1126/science.adl5414)

Tavares J, …, Gloor E, Galbraith D. 2023. Basin-wide variation in tree hydraulic safety margins predicts the carbon balance of Amazon forests. Nature 617:111-117. (https://www.nature.com/articles/s41586-023-05971-3)

Melnikova I, et al. 2024. Emergent constraints on future Amazon climate change-induced carbon loss using past global warming trends. Nature Communications 15:7263

Brienen R, Phillips, …, Gloor E, …, Galbraith D, … et al. 2015. Long-term decline of the Amazon carbon sink. Nature 519:344-348.