Many tropical rainforests have the capacity to accumulate and export large quantities of bioavailable nitrogen at the ecosystem scale. Ultimately, N losses must be balanced by external inputs from atmospheric deposition and/or biological fixation but resolving the contribution of N fixation and how it is affected by the availability of other macro and micronutrients represents a considerable empirical challenge. Here, we take advantage of a unique set of old-growth forested watersheds on the island of Trinidad characterized by monodominance of Mora excelsa, a non-nodulating legume, to better understand the sources and dynamics of nitrogen losses from tropical forests and their dependency on phosphorus and micronutrient availability. We measured nutrients in stream waters and soils in small watershed forests in which Mora accounts for 85-95% of the tree canopy and 95% of trees greater than 2 m in diameter and compared these with locally paired botanically diverse and legume-rich forest watersheds lacking in Mora. We also analyzed natural abundance isotope distributions in rainfall and streams and compare results from our focal watersheds to a broader array of catchments distributed across a wide range of elevations and soil types in the Northern Range of Trinidad.
Results/Conclusions
Preliminary results indicate that both nitrate and phosphate losses in stream waters are significantly lower in Mora watersheds compared to non-Mora watersheds. Water isotope analysis indicates that hydrologic sourcing is indistinguishable between forest types, suggesting that differences in nutrient losses result from differences in external inputs or internal cycling rather than hydrology. Analysis of N and natural abundance isotopes of nitrate in rainfall suggests that atmospheric deposition may be insufficient to balance losses in either forest type, implying that biological fixation may account for differences in nutrient losses between watersheds. However, it remains unclear whether these differences can be attributed to symbiotic versus heterotrophic fixation. Results from our broader survey of rainforests in Trinidad revealed a strong positive but asymptotic relationship between phosphate and nitrate in stream losses, with Mora forests being lowest in both and nitrate increasing as a function of phosphate across forests but leveling off at the highest phosphate levels. This may imply that P availability constrains N accumulation and losses at low P levels but that something else limits N losses at higher levels of P. Results from soil analyses and micronutrient distributions will also be discussed.