An intercomparison of models used to simulate the short-range atmospheric dispersion of agricultural ammonia emissions Mark R. Theobald a, b, * , Per Løfstrøm c , John Walker d , Helle V. Andersen c , Poul Pedersen e , Antonio Vallejo a , Mark A. Sutton b a Dept. of Agricultural Chemistry and Analysis, E.T.S.I. Agrónomos, Technical University of Madrid, Spain b Centre for Ecology & Hydrology, Edinburgh Research Station, Penicuik, United Kingdom c National Environmental Research Institute, University of Aarhus, Denmark d US EPA, National Risk Management Research Laboratory, Air Pollution Prevention and Control Division, USA e Dept. of Housing and Production Systems, Pig Research Centre, Danish Agriculture and Food Council, Denmark article info Article history: Received 26 July 2010 Received in revised form 6 March 2012 Accepted 7 March 2012 Available online 11 April 2012 Keywords: Atmospheric dispersion model Evaluation Validation Ammonia Agriculture abstract Ammonia emitted into the atmosphere from agricultural sources can have an impact on nearby sensitive ecosystems, either through elevated ambient concentrations or dry/wet deposition to vegetation and soil surfaces. Short-range atmospheric dispersion models are often used to assess these potential impacts on semi-natural ecosystems and a range of different models are used for these assessments. However, until now there has not been an intercomparison of the different models for the case of ammonia dispersion from agricultural sources and therefore it cannot be assumed that assessments are consistent. This paper presents an intercomparison of atmospheric concentration predictions made by a set of models commonly used for this type of assessment (ADMS; AERMOD; LADD and OPS-st). This intercomparison shows that there are differences between the concentration predictions of the models and some of these differences appear to be consistent and independent of the scenario modelled. The best model agree- ment was found for simple scenarios with area and volume sources, whereas the model agreement was worst for a scenario with elevated sources with exit velocities, for which ADMS predicted significantly smaller concentrations than the other models. The concentration predictions for the latter scenario depend strongly on the ability of the models to simulate the necessary processes, as well as the inter- action of these processes with pre-processor calculations of meteorological data. When applied to two case study farms in Denmark and the USA, the performance of all of the models is judged to be ‘acceptable’ according to a set of objective criteria, although the LADD model version used is currently not suitable for simulations with elevated sources with exit velocities. Ó 2012 Elsevier Ltd. All rights reserved. 1. Introduction Ammonia (NH 3 ) emitted into the atmosphere from agricultural sources can have an impact on nearby sensitive ecosystems either through elevated ambient concentrations or dry/wet deposition to vegetation and soil surfaces (Bobbink et al., 1998). Environmental impact assessments are often carried out using short-range atmo- spheric dispersion models to estimate mean annual atmospheric concentrations and total annual deposition of NH 3 at the ecosystem location. A range of different atmospheric dispersion models are used for these assessments, which have not, until now, been compared for dispersion of ammonia emissions from agricultural sources. For example, in the UK, modelling assessments for the disper- sion and deposition of agricultural NH 3 emissions normally use one of two ‘advanced’ Gaussian dispersion models (Environment Agency, 2010): the Atmospheric Dispersion Modelling System (ADMS, Carruthers et al., 1994) or the AMS/EPA Regulatory Model (AERMOD, Cimorelli et al., 2002). The term ‘advanced’ Gaussian dispersion model is used because the dispersion calculations are based on modified versions of the basic Gaussian plume equation taking into account vertical profiles of boundary layer parameters and continuous stability functions (Holmes and Morawska, 2006). These modifications improve the horizontal and vertical concen- tration distributions predicted by the basic equation (especially in * Corresponding author. Dpt. Agricultural Chemistry and Analysis, E.T.S.I. Agrónomos, Technical University of Madrid, Ciudad Universitaria s/n, 28040 Madrid, Spain. Tel.: þ34 91 336 3694; fax: þ34 91 336 5639. E-mail address: mrtheo@ceh.ac.uk (M.R. Theobald). Contents lists available at SciVerse ScienceDirect Environmental Modelling & Software journal homepage: www.elsevier.com/locate/envsoft 1364-8152/$ e see front matter Ó 2012 Elsevier Ltd. All rights reserved. doi:10.1016/j.envsoft.2012.03.005 Environmental Modelling & Software 37 (2012) 90e102