A probabilistic approach to exploring low-dimensional global dynamics

We demonstrate an approach to low-dimensional modeling of world population, carbon dioxide (CO2) emissions and gross domestic product (GDP) interactions in a way that explicitly characterizes the variability in the data informing model assumptions and the uncertainty in functional relationships. Our model choice was informed by the following considerations and choices. First, even a low-dimensional conceptualization of the interactions between these three global variables requires a model to illuminate the consequences of chains of cause and effect and feedback loops. Such interactions warrant analysis as they offer insights into influences on aggregate global dynamics. Second, rates are constrained to be consistent with world datasets where feasible thereby embedding a data driven philosophy into the dynamic model. Third, a probabilistic approach offers an effective way to deal with uncertain specification of functional relationships and the variability inherent in data informing such relationships. We use the model to highlight key features that result from the relative rates of change in the system and the nature of the feedback loops. Such an aggregated analysis offers a useful lens through which to study and interpret more detailed and realistic integrated models of human-biosphere dynamics.