Is carrying capacity useful for integrating humans into Earth system models? On the purposes and limits of modelling

There are calls to better incorporate humans into Earth system models. As interdisciplinary research teams are assembled, this has led to interesting discussions about the purposes and limits of such models. In this perspective paper, I draw on a typology of Earth system models to discuss the limits of modelling, and look at how these limits might be affected by incorporating humans. For illustrative purposes, I look at a series of models of human population size as examples of the kinds of models that can be constructed. Two main types of model can be constructed for two different purposes: toy models are intended to elucidate a complex phenomenon and may use a simplified concept like carrying capacity for that purpose; on the other hand, forecast models intended for theory testing may require a more complex concept to better approximate observed reality. Many elements of observed reality are, however, unpredictable. And unlike other aspects of the Earth system, humans have both the capacity and the tendency to deliberately behave unpredictably, meaning that they often behave in ways that are by definition outside of any possible model. This paper therefore helps clarify why the human sciences have not yet produced a more complete model of human behaviour, why such a model might be impossible, and why the unusual unpredictability of human behaviour introduces new limits to Earth system models that incorporate humans.