Modeling Marine Animal Movements
The study of animal movement is being revolutionized by technologies such as satellite tags and global positioning systems that allow us to track the movements of individual animals over large distances. Such data provide insight into how and where the animals move. However, the rich behavioral record embedded in the data is, for the most part, beyond the reach of current statistical analysis.
Our first goal is to continue developing state-space techniques as an analytical tool for animal movement. So far, two-dimensional state-space model fitting to animal tracks has been limited largely to demonstration studies. The second goal is to develop general state-space movement models that provide real insight into the behavior and decision making processes of a wide range of free-ranging animals in their natural habitat. We hope to understand better how migrating and foraging behaviors are influenced by other biological and environmental variables. For instance, by incorporating fatty acid data (used to determine diet) we may be able to infer different patterns of foraging behavior associated with different prey species and environments. This project will build on an already successful collaboration between statisticians developing complex models and the biologists deploying satellite tags.
Research in this area has verified that state-space model fitting becomes nontrivial once one moves away from linearity and normal error assumptions - a conclusion validated during our inaugural NPCDS workshop. Research directions include development of efficient algorithms for model fitting (Monte Carlo approaches) and model selection methods (via particle filters), as well as methods for integrating environmental and diet composition (fatty acid) data.