A Modeling Study of Direct and Indirect N2O Emissions From a Representative Catchment in the US Corn Belt

Indirect nitrous oxide (N2O) emissions from drainage ditches and headwater streams are poorly constrained. Few studies have monitored stream N2O emissions and fewer modeling studies have been conducted to simulate stream N2O emissions. In this study, we developed direct and indirect N2O emission modules and a corresponding calibration module for use in the Soil and Water Assessment Tool (SWAT) model, and implemented the expanded SWAT model (termed SWAT-N2O) to a representative fourthstream-order catchment (210 km(2)) and six first-order stream catchments (0.22-1.83 km(2)) in southeastern Minnesota. We simulated the spatial and temporal fluctuations of the indirect emissions from streams, identified emission "hot spots'' and "hot moments,'' and diagnosed the correlations between direct and indirect emissions. We showed that zero-order streams and first-order streams could contribute 0.034-0.066 and 0.011 nmol N2O m(-2) s(-1) (expressed on the basis of unit catchment area) to the total surface emissions, respectively. Emissions from zero-order and first-order streams equal 24-41% of direct emissions from soil, which may explain the emission gap between calculations using top-down and bottom-up methods. Clear spatial patterns were identified for both direct and indirect emissions and their spatial variations were negatively correlated. Our results suggest that the IPCC N2O emission factor for streams in the Corn Belt should be increased by 3.2-5.7 times. Increasing precipitation and streamflow in the Corn Belt may potentially increase frequencies of soil anoxic conditions and nitrate leaching to streams, and subsequently increase N2O emissions from both soils and streams.