Dinitrogen emissions from agricultural soils increase over the growing season

Background/Question/Methods

Nitrogen gas losses from ecosystems are difficult to quantify because a large but unknown fraction is lost as N₂, which is difficult to detect against a high atmospheric background. It is commonly assumed that N₂ and N₂O emissions are roughly correlated across the growing season, but there is very little empirical evidence to support this assumption. We used the Nitrogen Free Air Recirculation Method to directly measure N₂ and N₂O emissions from intact soil cores (0-10cm) collected from experimental farms (in Delaware and Pennsylvania, USA). At each farm, five replicated treatments (n=3) were imposed on a cornfield: no fertilizer, urea, manure, composted-manure, and biocharred-manure (225kgN/ha in Delaware and 150kgN/ha in Pennsylvania). We collected 3 replicate soil cores per treatment after planting/fertilization (May/June), in mid-season (July) and pre-harvest (October/November). We measured soil moisture over the growing season, and pre-harvest N mineralization and C respiration. We hypothesized: 1) N₂ and N₂O emissions would be highest following fertilization (May/June), 2) N₂ and N₂O emissions would be correlated throughout the growing season, and 3) N₂ and N₂O emissions would be highest from manure fertilizer.

Results/Conclusions

Across all treatments, N₂:N₂O was lowest in the spring and highest in the fall (p=0.002), ranging from 9.2 ± 62 in spring to 280 ± 43 in fall. Soil moisture at the time of collection explained more variance (r²=0.36, p<0.0001) in N₂ production than did C availability (r²=0.13, p=0.005). N₂O emissions were positively correlated with N₂ following planting (r²=0.37, p=0.0006) and mid-season (r²=0.21, p=0.01), but not in fall (p=0.27) when N₂ emissions were greatest. All fertilizer treatment effects on N₂ emissions were small relative to seasonal differences, but we found significantly higher N₂ emissions from the biocharred-manure treatment in the spring (p=0.02) than from the other treatments. In contrast, N₂O emissions did not differ by treatment, despite field based chamber data that suggest higher emissions from manure fertilizer (p<0.0001). The high seasonal variation in N₂:N₂O emissions, and the unobserved correlation in the timing of N₂ and N₂O emissions, suggest that our understanding of how and when N gasses are lost from ecosystems needs to be refined by direct quantification of N₂ fluxes.