Simulating an Ecosystem - Investigating the Impact of Brown Bears’ Omnivorous Behavior in a Simulated Ecosystem

Abstract

This bachelor thesis investigates the omnivorous impact of brown bears on ecosystem dynamics through an agent based computer simulation. A digital ecosystem set in a Swedish forest was developed using Unity. Within this ecosystem there are three species: brown bears which are omnivores, wolves which are carnivores and moose which are herbivores. The simulation uses state machines to model each respective animal behavior and interactions, states and attributes were tracked during runtime. Three main scenarios were evaluated: a scenario where no bears were present, a scenario where an average amount of bears were present and then a scenario where many bears were present. Each scenario was run 10 times over 900 seconds with the same amount of resources, moose and wolves. The impact of brown bears during hibernation was also checked by taking the final populations from the average bear scenario and using them as the start values for a winter scenario. Results demonstrated that in the simulation brown bears significantly influence ecosystem dynamics. Increased bear populations correlate to reduced moose and wolf populations, increased moose fleeing behavior and decreased wolf hunting success due to bear interference. Statistical analysis revealed that 53 out of 68 tracked metrics showed significant change between the scenarios, indicating that brown bears worked as a disruptor for both herbivore and carnivore species. These results should however not be taken as predictive due to the fact that the simulation implements simplified behavioral models and limited environmental complexity compared to real ecosystems. The results should therefore be seen as illustrative.