Validation and Uncertainty Analysis of an Ammonia Emission Model for Broiler Litter

Abstract

A mechanistic emission model was previously developed at laboratory scale to estimate ammonia emission fluxes from broiler litter. The overall model inputs include air temperature, air velocity, and litter properties such as litter nitrogen content, moisture content, and pH. The model outputs are predicted ammonia emission fluxes from litter. Considerable uncertainties may exist in measurement values of model inputs and outputs as well as model parameters. The purpose of this study was to perform model validation in the presence of measurement and model parameter uncertainties. A validation metric based on the mean and covariance in the measurement and in the model parameters was used to validate the ammonia emission model of broiler litter. The core model was validated given the uncertainties in the model prediction due to uncertainties of parameters (the Freundlich partition coefficient Kf and the mass transfer coefficient KG) and the uncertainties in the measurements. The significant level for the core model validation was 17.8%. The Kf submodel was validated at the given uncertainties of pH and temperature, and the significant levels were from 12.0% to 49.4%, which provided high confidence on the Kf submodel. At the given uncertainty levels of air velocity and temperature, the KG submodel passed the validation test (p > 0.05) when air velocities were low and failed the validation test (p < 0.05) when air velocities were high. The failure of the KG submodel at high air velocity levels may have been caused by significant loss of nitrogen and moisture content from the litter surface.