Identifying NH3 emission mitigation techniques from farm to field using a Bayesian network

This study addresses the challenge of reducing ammonia (NH3) emissions from agriculture by evaluating various mitigation techniques. The research utilized a Bayesian Belief Network (BBN) to integrate quantitative data on NH3 volatilization reduction with qualitative stakeholder perceptions, aiming to identify the best available techniques (BATs) that balance environmental, economic, and socio-cultural factors for farmers in the Veneto region of Italy. The BBN framework established probabilistic dependencies between variables related to livestock, crop type, manure storage, fertilization management, and pedo-climatic conditions. Stakeholder opinions were quantified through a value elicitation process and combined with the BBN to create an integrated Influence Diagram (ID). Results indicated that effective NH3 reduction requires a comprehensive approach across the entire agri-livestock supply chain. Based on the results obtained, no single technique clearly emerged as the primary focus, rather various areas would require improvement across the agri-livestock supply chain. However, if prioritizing techniques were necessary, efforts should concentrate on stable management of infirmary animals (HCInf), overcrowding reduction by decreasing the number of animals on densely populated farms (OC-Animal), and optimization of protein in animal ration (FDProt). These measures should be combined with effective manure application through slurry injection (INJSlu) in the field. Stakeholders showed reluctance towards more expensive or innovative methods, indicating that socio-cultural perceptions and economic feasibility can heavily influence the adoption of new technologies although they proved to be among the most environmentally effective. The primary insight from applying the BBNs was that selecting effective techniques necessitates a multi-perspective approach to foster consensus among stakeholders throughout the agri-livestock supply chain.