Nitrogen and phosphorus fertilization alters biological nitrogen fixation in lowland tropical rainforest

Background/Question/Methods

Recent advances in estimates of biological nitrogen (N) fixation (BNF) have substantially changed our understanding of human impacts on the tropical N cycle. Anthropogenic N deposition may already more than double naturally occurring annual N inputs, and disruption of the tropical phosphorus (P) cycle may also be significant. Yet important questions remain, such as what feedbacks occur between altered nutrient cycling and BNF. Research from temperate ecosystems demonstrates that nutrient inputs may alter BNF, but the complexity of BNF in tropical forest makes cross-biome assumptions tenuous. First, BNF in tropical forests is the product of symbiotic and free-living sources. Second, BNF may be facultative (responding to altered nutrient cycling), or obligate (unresponsive to nutrient availability). When BNF is facultative, increased N may decrease BNF, while increased phosphorus cycling (P) may alleviate enzymatic P limitation and increase BNF. However, there is a paucity of in situtests of these hypotheses in lowland tropical rainforests. We measured free-living and symbiotic BNF using adaptive cluster sampling in a mature rainforest that received seven years of N and P additions in a factorial fertilization experiment. Putative N-fixing leguminous were dominant in this forest, located at EARTH University in Costa Rica.

Results/Conclusions

We measured low rates of BNF in unfertilized forest plots that were similar to other recent estimates from lowland rainforest: 5.8 kg N ha^-1 y^-1. Surprisingly, in half the unfertilized plots we did not detect any symbiotic BNF and, across all unfertilized plots, free-living BNF in leaf litter and soil was a significant component (56%) of total BNF. However, fertilization further reduced symbiotic BNF (p=0.052) and free-living BNF in the soil (p=0.013), but not free-living BNF in leaf litter (p=0.758). Consistent with our expectations, N fertilization nearly eliminated BNF (0.45 kg N ha^-1 y^-1), but rather than increasing BNF, P-fertilized plots had half the BNF of unfertilized plots (2.95 kg N ha^-1 y^-1). Fertilization increased the importance of free-living BNF relative to symbiotic BNF, because of obligate BNF in the litter layer. Free-living BNF was 99.2% of BNF in N-fertilized plots, 20.7% of BNF in P-fertilized plots, and 100% of BNF in N and P fertilized plots. While our results support recent investigations suggesting low rates of symbiotic BNF, they also demonstrate that altered biogeochemical cycling in tropical rainforests may have profound impacts on BNF.