Tropical forests are thought to have high rates of nitrogen (N) fixation, both symbiotic, where specific tree species form relationships with bacteria to catalyze N-fixation in root nodules, and free-living, catalyzed by archaea and bacteria. Observations of the high abundance of symbiotic N-fixing trees and a limited number of empirical studies in tropical forests have led to the theory of high tropical N-fixation rates, especially after disturbance events. However, most tropical forests sit on highly weathered soils, depleted in the rock-derived nutrients molybdenum (Mo) and phosphorus (P) which can constrain N-fixation. In addition, most tropical N-fixation studies have been conducted in Hawaii, Costa Rica, and Panama, which receive relatively high atmospheric inputs of Mo and P from dust and sea-salt aerosols that can help offset losses from weathering.
The southeastern Amazon is a region isolated from atmospheric inputs and where soils are highly weathered. We hypothesized that at our study site, Tanguro Ranch in Mato Grosso, Brazil, 1) free-living N-fixation would be limited by Mo and P, and 2) fire disturbance would increase free-living and symbiotic N-fixation. We measured N-fixation after a fire disturbance, and conducted a series of block-design field experiments with Mo and P additions and measured free-living N-fixation.
Results/Conclusions
Across the mature control forest and the burned forest treatments, both symbiotic and free-living N-fixation rates totaled to less than 1 kg N ha-1 yr-1, a much lower estimate compared to observations in other tropical forests. Despite finding a higher proportion of recruitment by N-fixing trees in disturbed forests at our site, we still did not find evidence of higher associated symbiotic N-fixation rates, revealing that abundances of N-fixing trees do not always correlate with N-fixation rates. Surprisingly, while Mo and P concentrations in the soils were low, we did not find responses of free-living N-fixation to Mo and P additions either immediately, or in the weeks or months following the applications. Thus, limitation by Mo and P did not explain the low N-fixation rates found at our site. However, we did find high amounts of inorganic N in surface and deeper soils, which likely downregulates N-fixation. The discrepancies from our results and those reported from other studies may come from a lack of N-fixation measurements on these types of soils in the Amazon. Globally, tropical forests are highly heterogeneous, and our large-scale estimates of critical functions like ecosystem N-fixation are not well-captured.