Long-term nitrous oxide fluxes in annual and perennial agricultural and unmanaged ecosystems in the upper Midwest USA

Differences in soil nitrous oxide (N2O) fluxes among ecosystems are often difficult to evaluate and predict due to high spatial and temporal variabilities and few direct experimental comparisons. For 20 years, we measured N2O fluxes in 11 ecosystems in southwest Michigan USA: four annual grain crops (corn-soybean-wheat rotations) managed with conventional, no-till, reduced input, or biologically based/organic inputs; three perennial crops (alfalfa, poplar, and conifers); and four unmanaged ecosystems of different successional age including mature forest. Average N2O emissions were higher from annual grain and N-fixing cropping systems than from nonleguminous perennial cropping systems and were low across unmanaged ecosystems. Among annual cropping systems full-rotation fluxes were indistinguishable from one another but rotation phase mattered. For example, those systems with cover crops and reduced fertilizer N emitted more N2O during the corn and soybean phases, but during the wheat phase fluxes were similar to 40% lower. Likewise, no-till did not differ from conventional tillage over the entire rotation but reduced emissions similar to 20% in the wheat phase and increased emissions 30-80% in the corn and soybean phases. Greenhouse gas intensity for the annual crops (flux per unit yield) was lowest for soybeans produced under conventional management, while for the 11 other crop 9 management combinations intensities were similar to one another. Among the fertilized systems, emissions ranged from 0.30 to 1.33 kg N2O-N ha(-1) yr(-1) and were best predicted by IPCC Tier 1 and Delta EF emission factor approaches. Annual cumulative fluxes from perennial systems were best explained by soil NO3- pools (r(2) = 0.72) but not so for annual crops, where management differences overrode simple correlations. Daily soil N2O emissions were poorly predicted by any measured variables. Overall, long-term measurements reveal lower fluxes in nonlegume perennial vegetation and, for conservatively fertilized annual crops, the overriding influence of rotation phase on annual fluxes.