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

There are calls to better incorporate humans into Earth system models. Some have attempted this using the concept of ‘carrying capacity’. But the concept of carrying capacity has been subject to forceful critique, as it vastly simplifies the many complex factors which affect human populations. This includes factors related to human behaviour, which is so unpredictable that it seems to resist formalisation altogether. Nevertheless, I suggest that the usefulness of a simplified concept like ‘carrying capacity’ depends on the purposes for which it is used. In toy models intended to elucidate a complex phenomenon, a simplified concept might indeed be useful. On the other hand, in forecast models intended for theory testing, a more complex concept may better approximate observed reality. I illustrate this point by looking at two recent toy models which use the concept of carrying capacity to elucidate idealised population dynamics, and at a series of forecast models which instead aim to estimate the population density of real groups in real conditions. In short, the usefulness of concepts depends on modelling choices, which in turn depend upon the purpose for which a given model has been constructed.