Prediction of N₂O emissions under different field management practices and climate conditions

Due to the contributions of nitrous oxide (N ₂ O) to global climate change and stratospheric ozone destruction, it is important to understand how climate and agricultural management affect N ₂ O emissions. Although the process-based Denitrification Decomposition (DNDC) model is often used for quantifying emissions of N ₂ O, the accuracy of these predictions remains in question, and it is not clear which input variables, environmental or field management, have the greatest effect on model performance. In this study, DNDC was evaluated for prediction of N ₂ O fluxes from two climatically-different corn-field sites in the United States (a Colorado irrigated field and a Minnesota rainfed field). Besides climate, these sites offer the additional advantage that measurements are available for multiple field management practices, including fertilizer application, tillage, and crop rotation. This evaluation found that DNDC did not consistently, correctly predict daily-scale N ₂ O fluxes. Cumulative growing season N ₂ O fluxes were significantly under-predicted in Colorado and were both under- and over-predicted in Minnesota. Model calibration of four soil input parameters did not significantly improve N ₂ O emission predictions at either site or time scale. Modeled and measured N ₂ O fluxes and model error were all strongly correlated with precipitation. Over-predictions of N ₂ O fluxes were associated with heavy precipitation and high modeled denitrification. Based on our results, model improvements to decrease model error for corn cropping systems in temperate climate zones should focus on better accounting for the effects of precipitation on denitrification. Despite discrepancies in daily and cumulative growing season N ₂ O fluxes, DNDC correctly identified the only field management (fertilizer application rate) that significantly influenced the measured N ₂ O fluxes.