Predicting cattle grazing distributions: an agent-based modeling approach

Abstract

An agent-based model was designed which simulates foraging of yearling steers grazing in the short grass steppe region of Colorado, USA. Eleven hypotheses were analyzed that address different aspects of foraging behavior. Models tracked the grazing distributions of simulated steers, as well as their time spent grazing and amount of forage consumed. Model output was validated against grazing distributions and time spent grazing of real steers, observed using GPS-collars. Results indicate that in pastures containing sufficient heterogeneity, steers exhibit selective grazing behaviors in response to forage concentration and slope, as well as use reference memory to return to higher quality patches. In relatively homogenous pastures, cattle graze evenly, and over the course of multiple grazing seasons do not exhibit the selective foraging behaviors tested in this model. Future uses of this model include applying it to other range management scenarios to address differences in steer foraging behavior and pairing the agent-based model with a more elaborate ecosystem model to analyze relationships between steers and vegetation.