Despite the well-recognized importance of tropical forests to climate and global hydrologic and carbon cycles, major uncertainties remain in our basic understanding of the nitrogen (N) cycle in this vast biome.  Many old-growth tropical forests accumulate, cycle, and export large quantities of bioavailable nitrate at levels that exceed those observed in N-saturated temperate forests.  While it is commonly assumed that tropical forests are not N-limited, it is unclear how such N-richness emerges and is sustained through time, particularly in the face of recent global change.  We here examine patterns of nitrogen concentration and natural abundance isotopes (15N/14N, 18O/16O) from 70 watershed rainforests distributed across three neo-tropical countries and spanning a diverse array of vegetation, soil, and climate types.  We focus particularly on a set of well-characterized small watershed rainforests in Costa Rica that display high loss rates of dissolved N in stream waters.  We use recent and historical measures of stream nutrients and the chemistry and isotopic distributions of soils and precipitation to probe the internal N cycle and examine putative mechanisms capable of sustaining high N losses.  We further evaluate our results in the context of a theoretical framework of the terrestrial N cycle. 

Results/Conclusions

Across all forests, dissolved N losses occurred predominantly as plant available nitrate.  Analysis of isotopes in rainfall and stream waters revealed that, despite exceedingly high N losses from these forests, virtually all (>98%) of the nitrate exported derived from internal cycling rather than direct flow-through from atmospheric N.  At the same time, nitrogen in stream waters and soils showed strong isotopic enrichment from denitrification and stream N levels changed little with changes in discharge, consistent with N-replete conditions.  For six focal watersheds we further provide evidence that N losses and their isotopic character have been stable for at least the past two decades, suggesting that N inputs and internal cycling have kept pace with any changes in plant growth over this period.  Together with evidence for high rates of internal N turnover, our results are consistent with model analysis showing that, in the absence of disturbance, high bioavailable N losses from tropical rainforests cannot be explained by leakiness under N limitation but rather require that N supply is sufficient at the ecosystem scale.  We discuss the implications of our findings in the context of our current understanding of atmospheric deposition, biological N fixation, and multiple element limitation in tropical forests.