Why is nutrient co-limitation not more common in land ecosystems (or perhaps it is)?

Background/Question/Methods Biogeochemical theory predicts that, over time, organisms should act to equilibrate resources to the point of co-limitation. However, a prevailing notion from empirical field experiments is that land ecosystems often are subject to single nutrient limitation. We here combine experimental observations across temperate and tropical land ecosystems with predictions from biogeochemical models to examine broad expectations of whether single or multiple nutrients ought to constrain biological processes. In particular we focus on three systems that we have studied extensively and that differ dramatically in functional biotic composition and disturbance regimes: unpolluted temperate forests in south America and New Zealand, unpolluted tropical forests in central america, and fire-disturbed savannas in Kruger National Park, South Africa. Our overall approach is to compare patterns of observed internal distribution and cycling of nutrients, and of responses to experimental manipulations, against predictions from simplified first-principle models and a more complex land earth system model capable of coupling to the earth climate system. Results/Conclusions Our overall findings show that co-limitation is not a trivial prediction for land ecosystems, and that co-limitation ought to be considered more frequently in combination with traditional perspectives that emphasize Liebig's law of the minimum. Consideration of the nitrogen (N) and phosphorus (P) system in forests and savannas suggests that co-limitation ought to be relatively common in stable and highly productive ecosystems in which nitrogen fixation is a prevalent feature. Similarly, consideration of the N, P and molybdenum (Mo) system in tropical soils suggests that it is sensitive to the stoichiometry of resource supply, such that co-limitation ought to be expected in certain conditions. We compare these predictions against our experiments and observations from field sites. We also derive a set of general rules that describe the mechanisms and conditions that favor co-limitation vs. single nutrient limitation, and where they are most likely to be found. These considerations are of fundamental importance to the fields of biogeochemistry and ecology, and also inform us about how to build the most appropriate land components within global earth models.