Predicting ammonia volatilization from fertilized pastures used for grazing

Abstract Ammonia (NH3) volatilization from fertilised agricultural soils is driven by complex interactions between edaphic, climatic and plant canopy factors that can be difficult to measure or predict. We developed a simplified approach using default parameters in the DairyMod model to predict daily NH3 volatilization from urea applied to grazed dairy pastures. Several published datasets were used to validate the reliability of the model to reproduce key related soil processes in a whole farming systems framework. For the sites where monitoring for the experimental duration occurred, DairyMod simulated the main features of the observed NH3 emissions, with an overall predicted median of 4.1 kg/ha or 7% of applied N, compared to the measured median of 6.1 kg/ha for 12% of applied N fertilizer. There was an overall root mean square error (RMSE) of 0.9 kg NH3 N/ha/d and an overall mean prediction error (MPE) of 0.5 kg NH3 N /ha/d. However, there was high uncertainty in several of the datasets used which made it difficult to be conclusive about the validation. The simulation accuracy was improved using daily wind speed (collected on-site in field campaigns) as input to the evapotranspiration calculations. In cases of high certainty in the volatilization data, it was concluded that the model was useful for the analysis of N cycling in situations used for dairy farming without the need for a more complex mechanistic method with difficult-to-obtain parameters. DairyMod presents a simple but readily reproducible prediction of NH3 volatilization from urea application on pasture in intensive livestock farming systems compatible with the certainty of the model inputs and scale of model application. However, the collective understanding of NH3 volatilization in pasture based dairy systems is currently based on a limited number of often uncertain, short term plot studies in the absence of animals.