How to carry out detailed modelling of ammonia emissions (GN 036)

Detailed modelling is carried out by utilising an air dispersion computer model. This is a specialist activity; we recommend you use a competent consultant to carry out the detailed modelling. Please ensure the consultant uses the information available on NRW website or the reports are likely to be rejected. 

You can find more detailed guidance available on the GOV.UK website 

When using computer models please follow the recommendations of the software developer. 

Model Source Type 

In order to model the dispersion of ammonia from intensive livestock farming it is important that the characteristics of the source, such as pollutant emission rate, temperature and velocity of exhaust gases, number and configuration of housing or manure storage facilities are properly represented in the model. 

If annual average emission factors are used in the modelling it will generally be an adequate approximation to assume that releases occur evenly over the year. However, it will be necessary to apportion them correctly between different sources, particularly if there are significant differences in emissions with location or over time. Although ammonia in its pure form is lighter than air, it is emitted as a minor constituent of air vented from development and will disperse in a similar manner to the air within which it is carried.  

If the dispersion model permits the density of the release to be specified this should be set to that for air rather than ammonia alone.  

The type of development will significantly affect the dispersion characteristics and it will be important that these are correctly represented in the model.  

For example, on a farm, housing types will vary widely, from sheds with high velocity roof fans to free-range operations with minimal housing in the field.  

The modeller will need to think carefully how these are to be represented in the model and the chosen approach should be properly justified in the modelling report.  

How typical housing types may be represented in a model

Sheds with roof vents should be modelled as a series of elevated point sources with appropriate release characteristics (volumetric flow, temperature etc.). Modellers should identify the mode of operation of fans. Some units may use variable speed fans which result in a reduced exit velocity at low fan speeds compared with ‘on/off’ type fans. Where there are a large number of vents they may need to be combined into a smaller number of composite sources distributed to represent the release scenario. When an effective diameter is used which is much larger than the actual physical diameter, the calculation of the plume rise may be inaccurate.

The use of an effective stack diameter may also have an impact on stack tip downwash modelling. We suggest that the stack diameter and exit velocity for an individual stack is used to represent the combined sources. However, the mass emission rate should be increased to represent the composite source. Sensitivity analysis needs to be carried out if necessary.

Some sheds may have tunnel ventilation with fans in the gable ends. If the line of gable ends is long or the shed is wide, they can be modelled as a low-level line source to account for the building downwash effect. For example, they can be modelled as a series of point sources with zero efflux velocity and including building downwash effects. Naturally ventilated sheds typically may be considered as ‘leaky boxes’ and modelled as volume sources.

For free-range animals, the appropriate emission factor will need to be apportioned between the free-range area and any housing in proportion to the time the animals spend in each location.  

We require the modellers assume that 20% of droppings will occur outside the house. This will only change if agreed with other UK regulators.

If using the free range emission factor of 0.045 for the ranging area, this must be applied to whole flock numbers.

Meteorological data and terrain impact  

The model should be run using 5 years (minimum 3 years if not possible) of representative meteorological data from the meteorological station nearest to the proposed site. Each year of meteorological data should be run separately and the highest predicted annual average from the 5 years used as the basis of the assessment.

If observed suitable meteorological data are not available, then high quality numerical weather prediction (NWP) data, e.g. Met Office high horizontal resolution (1.5 km) NWP data extracted at the proposed site, should be used. The effects of terrain should be considered in accordance with the current recommendation of the dispersion modelling software developer.  Where a submitted model deviates from the software developer’s recommendations, justification should be provided. 

Approach to model ammonia dispersion and deposition 

All the sensitive sites required for ammonia modelling in Wales can be found using our open data map 

ADMS and AERMOD are two air dispersion models commonly used in ammonia dispersion from poultry/pig farms for regulatory application. Both ADMS and AERMOD contain a deposition module for calculating the hourly dry deposition flux under the assumption that the dry deposition velocity is independent of the pollutant concentration.

The modules may not be fit for purpose in modelling ammonia from intensive farming emissions. Until the models deal adequately with this issue, we recommend a two-stage approach.

Find out about Stage 1 detailed modelling 

Find out about Stage 2 detailed modelling 


Last updated