Spatial heterogeneity can resolve the nitrogen paradox of tropical forests.

Many tropical forests are characterized by large losses of plant-available forms of nitrogen (N), indicating that they are N rich, and by an abundance of plants capable of symbiotic N fixation. These N-fixing plants can fix enough N to drive N richness. However, biological N fixation (BNF) is more expensive than using plant-available N, so sustained BNF in N-rich soils appears to be a paradox. Here, we use spatially explicit ecosystem models to analyze the conditions under which spatial heterogeneity can induce simultaneous BNF and loss of plant-available N (hereafter, we call this combination "N-rich BNF"). Spatial movement of litter to neighboring plants' rooting zones can maintain N-rich BNF under a variety of conditions. For example, when N-fixers have higher N demand than non-fixers, N-fixers export N-rich litter to non-fixers, inducing large losses of plant-available N from the ecosystem, and receive N-poor litter from non-fixers, inducing BNF. BNF and N loss fluxes increase in proportion to the ratio of N-fixer litter N:P to non-fixer litter N:P, and also in proportion to the fraction of litter transferred out of a tree's rooting zone. Stoichiometric variability augments N-rich BNF, as does increasing the fraction of the landscape occupied by N-fixers, at least when they are rare. On the contrary, greater root overlap between neighbors and clumping of N-fixers diminish N-rich BNF. Finally, we examined how spatial litter transfer interacts with another mechanism that can sustain N-rich BNF, incomplete down-regulation of BNF. Spatial transfer and incomplete down-regulation can both sustain N-rich BNF, but they are compensatory rather than additive. These mechanisms can be distinguished by examining where N losses occur. Incomplete down-regulation of BNF leads to greater N loss under N-fixing trees, whereas spatial litter transfer leads to greater N loss under non-fixing trees. Along with time lags in regulating BNF, these results comprise a series of hypotheses that could help understand the N paradox of tropical forests.