New Statistical Tools for Ecological Modeling
Patterns in nature are usually the result of many interacting processes. The question, why are there roughly 1,100 bird species in the United States rather than 110 or 11,000, for example, cannot be answered by listing all the contributing factors. Researchers need to know which factors are more and less important. Similarly, when ecologists study an ecosystem altered by humans, they need to determine which factors are most crucial for preserving biodiversity and which are less critical. Because ecological questions often involve many processes operating over large spatial and temporal scales, experimental manipulations are not feasible. Inference must come from dynamic models fitted to empirical data.
Stephen P. Ellner, Ecology and Evolutionary Biology, and Giles Hooker, Statistics and Data Science/Computational Biology, are developing tools to identify factors that are most important in creating observed patterns, based on a general statistical method called Functional Analysis of Variance (fANOVA). Tools developed through this project will enable researchers to 1) extend ecological theories explaining how competing species can coexist; 2) determine which vital rates—for example, survival rates at different ages—contribute most to fluctuations in population abundance; and 3) identify which vital rates contribute most to large within-population variation in lifetime outcomes, such as total number of offspring, that cannot be explained by variation in observable traits.
This project advances statistical methods for assessing contributions from multiple processes that affect individuals, populations, and communities. The research aims to increase the information that can be extracted from complex ecological models, with the potential to inform more targeted and effective environmental policy and to optimize strategies for mitigating negative ecological impacts of human activity.