Predicting ammonia volatilization after field application of swine slurry: DNDC model development

Process-based biogeochemical models are needed for providing N loss predictions, but most lack a direct or accurate estimation of NH3 volatilization. This paper presents an improvement of the Canadian version of the denitrification-decomposition model (DNDC v.CAN) by integrating a sub-model capable of predicting NH3 volatilization after the field application of slurry. A validation dataset included NH3 flux measurements collected for 4-7d after 16 applications of swine slurry (May 1999-Sept 2005) to the surface of grassland and bare soil in Québec. Our model developments improved the ability of DNDC to predict NH3 loss and provided reasonable estimates for cumulative NH3 volatilization (i.e., average relative error of 5.7%). Consequently, lower errors of prediction for soil mineral N content were derived from the new DNDC v.CAN model compared to the previous version of DNDC. The new model was capable of demonstrating a significant influence of slurry pH and clay NH4+ adsorption on NH3 volatilization. Ultimately, the improved prediction of NH3 volatilization using DNDC v.CAN enhances the ability to accurately model N cycling and predict overall N loss from agro-ecosystems. To further refine NH3 volatilization predictions, we recommend DNDC developments focus on modelling soil unsaturated flow, slurry infiltration rates, and soil pH changes over time.