Archetype Analysis of golden eagle migration patterns using Bayesian Methods
Abstract
Animal migration has the potential to be a very good indicator of environmental changes that could affect us all. It also helps us understand the different species with whom we share this planet. In order to do the mentioned above, we analyzed a monthly golden eagle (Aquila Chrysaetos) location data. The main research question was whether covariates, such as age, could be a big factor on their migration routes. This exploration was possible through an archetype analysis, which is a statistical nonparametric approach that represents each individual as a mixture of multiple estimated archetypes. In addition to a traditional archetype analysis, we developed a new approach to archetype analysis, in which covariates are considered, and a subset of the archetypes is defined by existing golden eagles who exhibit known, interpretable behaviors, in order to be fitted using Bayesian methods and Markov Chain Monte Carlo simulations. This approach was developed using R with Machine Learning techniques and Bayesian Statistics. Once the analysis was complete, we were able to exhibit that covariates such as age influence birds’ behavior and their migration routes, concluding that the older they get, the more likely they will belong in a nonmigratory archetype. This novel approach showcases a new proposal for such databases and optimize processes in ecological research.
Full Article
Arbelaez