A better understanding of the potential cumulative impacts of large-scale fish farming, could help marine aquaculture to become more environmentally sustainable. Risk assessment plays an important role in this process by elucidating the main challenges and associated risk factors. In Norway, impact of sea lice on wild salmonids, genetic introgression of escaped farmed salmon and poor fish welfare are identified as the main challenges to sustainable salmon production. However, also pathogen transmission to wild fish, organic emissions, use of drugs and anti-fouling agents as well as the use of wild caught wrasse for delousing plays a role in a comprehensive environmental risk assess ment of Norwegian aquaculture production.
Here we describe an approach to risk assessment using Bayesian belief networks to visualise overall casual structures (Figure 1). The risk assessment process is made up of three main steps 1) defining the top node and designing a graphical structure of underlying nodes in terms of risk sources (RS´), events (A´), and consequences (C´); 2) measuring the uncertainties related to risk sources, events and consequences in terms of subjective probabilities (P) and strength of knowledge (SoK); and 3) aggregating the uncertainty measurements of each node upwards until reaching the top node . The suggested methodology is anchored in the latest thinking in risk science and has been tested in a thorough study of environmental risk related to Norwegian aquaculture. The study shows that the new methodical approach has an immanent pedagogical potential and contributes to strengthening risk understanding and risk acknowledgement among stakeholders. In conclusion, the suggested risk assessment methodology has proven a valuable tool for marine scientists in analysing, evaluating, and communicating environmental risk.