Agents on a Landscape: Simulating Spatial and Temporal Interactions in Economic and Ecological Systems

Modeling how communities benefit from common-property, depletable ecosystem services, such as non-timber forest product (NTFP) extraction, is challenging because it depends on agent proximity to resources and competition among agents. This challenge is greater when agents face complex economic decisions that depend on the state of the landscape and the actions of other agents. We address this complexity by developing an agent-based model, founded on standard economic theory, that defines household production and utility functions for millions of spatially-explicit economic agents. Inter-agent competition is directly modeled by defining how NTFP extraction of one agent changes the extraction efficiency and travel-time of nearby agents, thereby modifying agents’ profit functions and utility maximization. We demonstrate our simulation using Tanzania as a case study. Our application relies on estimates of NTFP stocks, local wages, and traversal times across a landscape network of grid-cells, which we derive using geospatial and household data. The results of our simulation provide spatially explicit and aggregate estimates of NTFP extraction and household profit. Our model provides a methodological advance for studies that require understanding the impacts of conservation policies on households that rely on natural capital from forests. More broadly, our model shows that agent-based approaches to spatial activity can incorporate valuable insights on decision-making from economics without simplifying the underlying theory, making strong assumptions on agent homogeneity, or ignoring spatial heterogeneity.