Agricultural management practices and environmental drivers of nitrous oxide emissions over a decade for an annual and an annual-perennial crop rotation

Nitrous oxide (N2O) emissions were measured over 4 fields in Manitoba, Canada, continuously from 2006 to 2016 using the micrometeorological flux-gradient technique. The fields were divided into 2 treatments: 2 fields with a rotation of annual crops (corn, faba bean, spring wheat, rapeseed, barley, spring wheat, corn, soybean, spring wheat, soybean, soybean); while 2 fields had a four-year period with a leguminous perennial crop in the rotation (corn, faba bean, alfalfa/grass, alfalfa/grass, alfalfa/grass, alfalfa/grass, corn, soybean, spring wheat, soybean, soybean). Nitrogen fertilizer was applied according to local practice, and was typically not applied in years when a legume was planted. N2O was emitted from the field in episodes during spring thaw and following nitrogen fertilizer application. Emissions were less during the perennial phase of the comparison (mean 1 kg N ha(-1) y(-1) from perennial fields; 4.7 kg N ha(-1) y(-1) from the annual fields; P < 0.001). However, over the 11-year period, the treatments were not different (P = 0.3), averaging 5 kg N ha(-1) y(-1) because of large emissions of about 20 kg N ha(-1) y(-1) in the year following termination of the perennial crop. Nitrogen fertilizer application rate explained 69% of the variability in emissions (P < 0.001), with only a marginal advantage including other environmental variables such as soil nitrate exposure or soil ammonium concentrations. The N2O flux data were combined with carbon dioxide flux measurements, fertilizer N manufacturing, and potential product end uses to estimate the net greenhouse gas budget. Net emissions were a source of 1.7 Mg CO2 equivalents ha(-1) y(-1), driven by 3 Mg CO2 equivalents ha(-1) y(-1) source from N2O emissions, net carbon exchange of 3 to 5 Mg CO2 equivalents ha(-1) y(-1) sink, and a consumption source of 6 to 7 Mg CO2 equivalents ha(-1) y(-1). High inter-annual variability in fluxes resulted in no significant difference (P = 0.87) in overall greenhouse gas emissions between treatments over the full period including N fertilizer inputs and end uses of crop products.