Nitrogen oxide emissions from a banana plantation in the humid tropics

Use of nitrogen fertilizer is thought to contribute significantly to the increase of atmospheric nitrous oxide (N2O) and nitric oxide (NO). While the current increase of fertilizer use is concentrated in tropical areas, nearly all studies of nitrogen oxide emissions have been conducted in agricultural systems in temperate areas. We measured N2O and NO fluxes from a fertilized banana plantation in the humid tropics of Costa Rica, where 360 kg N ha(-1) yr(-1) is applied. Using chamber techniques, we sampled an Andisol and an Inceptisol on a monthly basis. Twice on each soil type, we sampled intensively in time following fertilizer applications. There is a strong spatial and temporal dependence of nitrogen oxide emissions on place and time of fertilizer application. We fmd greater mean N2O and NO emissions from the Andisol (31.4 ng N2O-N cm(-2) h(-1) and 55.6 ng NO-N cm(-2) h(-1)) than from the Inceptisol (9.3 ng N2O-N cm(-2) h(-1) and 41.1 ng NO-N cm(-2) h(-1)) under the plants where fertilizer is typically applied. The percentages of applied fertilizer-N that are converted into nitrogen oxide (''yield'') are between 1.26 and 2.91% for N2O and between 5.09 and 5.66% for NO depending on soil type. We consistently calculate higher nitrogen oxide yields based on intensive sampling versus monthly sampling. Temporal variation in nitrogen oxide emissions probably causes monthly sampling to underestimate mean annual fluxes. Our results suggest that in some tropical systems a higher percentage of applied nitrogen may be lost in gaseous form than in temperate agriculture. Current global estimates of N2O and NO sources from tropical agriculture are based on information from temperate areas and may cause an underestimate of the contribution of tropical agriculture to the budgets of these trace gases.