Community processes as emergent properties: Modelling multilevel interaction in small mammals communities

Abstract

Listen

Translate article

Community interactions of small rodents have attracted the attention of ecologists for many years due to their abrupt changes in population numbers, their impact on the whole biocoenosis and also because of immense damages to agricultural production and forestry. In particular, regularly oscillating rodent populations in Scandinavia have been subject of discussions among theoretically and empirically working ecologists for many decades. Spatial and temporal restrictions in empirical work led to various attempts to model these dynamics to understand large scale effects resulting from complex interactions in variable cause-effect networks of the numerous involved system components. The presented individual-based model for the first time described small rodent communities as a set of interacting autonomously acting agents with a detailed life history and behavioural repertoire in a food-web setup composed of three trophic levels (rodents, rodents food and predators). It thus allowed to integrate all relevant factors accounting for the dynamics of rodents which acted in a simulated environment containing the spatial arrangement of habitats and seasonal changing conditions. Due to the representation with interacting entities, the dynamics on higher levels resulted in a self-organisation process as emergent properties. This differentiation between the focal and the operational level allowed to investigate processes interacting between different integration levels and to adapt the model to different scenarios easily as well as to specify it for a large range of rodents species. Simulations have been executed for two different scenarios. The Bornhöved scenario simulating the situation of a Northern German rodent community in a beech forest represented bottom-up effects of mast events on population dynamics. The Scandinavian scenario which depicted the most important actors of these oscillating rodent communities, gave new insights into the processes causing the sudden decline of rodent populations. Both, lack of resources and predation, contributed to about 90% of mortality, but no pattern could be found when relating either cause with the properties of the respective cycle. Bottom up and top down control vary unpredictably and chaotically in the model. These results may explain considerable parts of contradicting empirical findings.