气候变化领域集成服务门户

Due to the contributions of nitrous oxide (N2O) to global climate change and stratospheric ozone destruction, it is important to understand how climate and agricultural management affect N2O emissions. Although the process-based Denitrification Decomposition (DNDC) model is often used for quantifying emissions of N2O, 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 N2O 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 N2O fluxes. Cumulative growing season N2O 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 N2O emission predictions at either site or time scale. Modeled and measured N2O fluxes and model error were all strongly correlated with precipitation. Over-predictions of N2O 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 N2O fluxes, DNDC correctly identified the only field management (fertilizer application rate) that significantly influenced the measured N2O fluxes. (C) 2018 Elsevier B.V. All rights reserved.