Ecosystem condition emerges from an ecological equation of state applied to North American tree communities

Ecosystems represent the largest scales of biological organization and shape the ecology and evolution of genes, populations, species, and communities. Yet we lack an understanding of the key properties of ecosystems-the state variables-that must be tracked to predict changes in ecosystem condition through time,^1,2,3 instead commonly relying on reference states.^4,5 A recently published ecological equation of state demonstrated a strong relationship between biomass, species richness, organism abundance, and productivity, suggesting the untested possibility that this relationship may systematically vary under ecological disturbance (i.e., vary with ecosystem condition).⁶ To test this idea, we investigate how the performance of the ecological equation of state relates to expected ecosystem condition (derived from protected area data) using satellite-derived proxies for the forests of the conterminous USA. We found that, despite the noise introduced by the use of satellite-derived proxies, the performance of the ecological equation of state in predicting biomass varied systematically with expected ecosystem condition. Moreover, differences in equation performance could be used to identify areas with different expected ecosystem condition. This differential performance was stronger in the equation of state than in correlative models fit to the same data, though similar to patterns seen in the relationship between productivity and biomass. These findings suggest deeper underlying regularities linking ecosystem condition and state variables and the potential to break ecology's dependence on reference states. Further investigation of these relationships may reveal new principles of ecosystem dynamics, which are vital to informing global biodiversity conservation efforts.