Using agent-based models to scale from individuals to populations

Abstract

All aggregate biological phenomena in the aerosphere are due to the behaviors of unique individuals acting according to their own rules of behavior and perceived external stimuli. An appealing characteristic of aeroecology is that we can observe both these aggregate behaviors of large groups (using tools such as radar and observational networks) as well as the behavior of individual animals (by employing animal tracking technology). Traditional population modeling efforts focus on equations that mimic natural populations in terms of overall population size and/or mean population parameters, often discounting mechanisms operating on the individual level that give rise to overall population dynamics. Concurrent advancements in computing capacity and animal tracking methodologies provide us with the opportunity to examine how the actions of individuals scale up to give rise to population-level phenomena in the aerosphere. More specifically, we can now model populations as a collection of individuals that behave independently, and we can validate the inferences from these agent-based models by tracking actual animals over the course of their annual cycle. In this chapter, we provide an example of how an agent-based model can be used to predict migration behavior across a species range based on a small set of actual migration tracks. The example provides a generalizable framework for using agent-based models as a link between data from individuals and broad-scale phenomena.