Modelling uncertainty for decision making on ammonia mitigation with trees in the landscape (MUDMAT)

Ammonia pollution can have many harmful effects. Over 60% of the UK’s semi-natural habitats exceed threshold limits where damage (such as species composition changes and reduced biodiversity) is caused by nitrogen. Agricultural practices account for over 80% of nitrogen emissions, in the form of ammonia (NH3), within the UK.

Government guidance supports farmers to reduce their ammonia emissions, including via mitigation options like planting tree ‘shelter-belts’ around livestock housing to ‘scavenge’ ammonia. However, farmers need tools to help them quantify how options for planting tree shelter-belts will influence their NH3 pollution.

Principal Investigator David Cameron introduces the project in this short video


The aim of the MUDMAT project was to improve decision-making tools by bringing state-of-the-art statistical methods to quantify the uncertainty of predictions through the modelling of tree shelter belts.

There is already an existing web-based decision tool for modelling tree shelter belts, but it requires improvement. The very simple empirical model used in the web-tool has demonstrated proof of concept, but it is unlikely to give accurate predictions. Indeed, the accuracy of the predictions from the web-tool are not currently quantified. As a result, decision-makers such as farmers cannot use the tool to make important and potentially costly decisions.

The MUDMAT project created an explicit link from the web-tool to a model (MODDAS) that has demonstrated the ability to predict the effect of tree-belts in capturing ammonia. A statistical emulator of OpenFOAM-MODDAS was created which will allow model parameter uncertainty to be quantified and reduced in future model calibrations against observations. The uncertainty presented in the web tool was quantified by forwarding model parameter uncertainty through to model predictions of ammonia capture. Improving the model used in the web-tool and presenting uncertainty in the predictions will aid UK land managers in their decision-making in mitigating harmful ammonia emissions from agricultural sources.