Why scaling up is harder than you think

Data and processes are tied to a specific scale. This has led to a quest to use theory to link together scales, which has proved challenging in ecology. I present a general framework for thinking about scale-spanning theories which leads to three generic scale-spanning approaches: summation, mean-field and macro-specific. I show that chaos and Jensen’s inequality usually make the first two difficult and/or inaccurate especially in the face of nonlinearities, noise and heterogeneity which are endemic in ecology. I talk about approaches to address these limitations and I dispute perceived limitations on the value of the macro-specific approach.