A cross-scale modeling framework for decision support on elephant management in Kruger National Park, South Africa

Abstract Conservation areas, like national parks, are hotspots of social-ecological and social-economic activities. The resulting interactions contribute to an inherent complexity of these systems, making simulation models a vital form of support for their management activities. These models are often unimodal, i.e., limited by design to only one particular question or a specific temporal and spatial scale. We implemented the cross-scale and multi-modal base model MARS KNP for the Kruger National Park, South Africa that combines the agent-based paradigm with a dynamic vegetation model. As a proof-of-concept, we developed an elephant movement model within MARS KNP to evaluate the base model's decision-support capabilities. The study was mainly focused on the underlying software mechanisms that allow easy integration of multi-scale spatio-temporal data objects. MARS agents can probe, interact with, and modify these objects. We found that this feature is essential for a cross-scale integration of different modeling approaches. Additionally, we propose a definition of the term ‘base model’ to shorten the provisioning time of decision-support tools. A spatio-temporal mechanism couples the large-scale multi-agent modeling and simulation framework MARS and the dynamic vegetation model aDGVM. We propose a definition of the term ‘base model’ to shorten the provisioning time of decision-support tools. A base model including prototypical elephant agents was created for the Kruger National Park, South Africa.