An elegant conceptual model of long-term ecosystem development has emerged that focuses on the roles of nitrogen (N) and phosphorus (P) in shaping terrestrial ecosystem dynamics.  It states that N limits net primary production (NPP) on young soils because rocks contain P but negligible N, whereas P limits NPP on old soils because N inputs continue but P inputs from weathering eventually cease.  Although many data support this model, some seem not to (e.g., N limitation in old temperate forests), and there are good biological reasons that the picture might be more complicated.  Young ecosystems are more likely to be P limited than N limited if N fixers dominate, and old ecosystems might receive sufficient P inputs from dust to prevent the P-limited steady state.  Furthermore, organisms might adjust resource acquisition to be co-limited by N and P at any ecosystem age.  Under what conditions should N limitation, co-limitation, and P limitation prevail?  If there are transitions between these alternate ecosystem states, when do they occur?  What effect do different functional groups—such as N fixers versus nonfixers—have on these states and transitions?  We use a dynamical model and timescale separation techniques to investigate these questions.

Results/Conclusions

We derive equilibrium, quasi-equilibrium, and transient solutions at short (hours-months), medium (years-centuries), and long (millennia-) timescales.  The classical transition from N to P limitation at long timescales is but one of many possible ecosystem development trajectories.  It occurs when N fixer activity is constrained in young ecosystems and P inputs from dust deposition are sufficiently low.  Our calculations predict when the transition from N to P limitation occurs; this transition time depends most strongly on the weathering rate.  Alternative stable states are possible at many timescales, depending in particular on symbiotic N fixation.  N limitation at equilibrium or quasi-equilibrium is impossible when N fixation exceeds losses of N from plant-unavailable pools.  In the transient case N fixation can rapidly overcome N limitation if N fixers are present and active.  Therefore, the limiting nutrient depends critically on the regulation of N fixation.  Our results also reveal the potential for oscillations.  For example, a ground fire that combusts litter but also mineralizes N produces a pulse of plant-available N, followed by a drop below its pre-fire level due to plant uptake, followed by a slow rise to its pre-fire level as litter N builds back up.