A series of experiments was conducted using small wind tunnels to assess the influence of a range of environmental, manure and management variables on ammonia emissions following application of different manure types to grassland and arable land. Wind speed and dry matter content (for cattle slurry in particular) were identified as the parameters with greatest influence on ammonia emissions from slurries. For solid manures, rainfall was identified as the parameter with most influence on ammonia emissions. A Michaelis-Menten function was used to describe emission rates following manure application. Linear regression was then used to develop statistical models relating the Michaelis-Menten function parameters to the experimental variables for each manure type/land use combination. The fitted models accounted for between 62% and 94% of the variation in the data. Validation of the models for cattle slurry to grassland and pig slurry to arable land against independent data sets obtained from experiments using the micrometeorological mass balance measurement technique showed that the models overestimated losses, which was most probably due to inherent differences between the wind tunnel and the micrometerological mass balance measurement techniques. (C) 2004 Elsevier Ltd. All rights reserved.