Multi-disciplinary Lego-bricks: building an integrative metamodel for policy analysis by using Bayesian Networks

Bayesian networks (BNs) are often praised on their easy to update -characteristic. This is commonly understood as either updating the conditional probability tables when new data or knowledge appears or updating our prior knowledge by setting some of the variables to a "known" state. In addition to that, BNs are relatively easy to update in a sense that the structures can be modified to answer different research questions. In many cases, selected nodes and their defined mutual dependencies can be detached from the original model and linked to another BN as such. It is also possible to integrate whole BNs as submodels to larger entities – metamodels, which are useful e.g. for policy analysis where alternative management actions affecting different parts of the system should be evaluated and compared. We present a process of building a cross-disciplinary BN for minimizing the ecosystem risks caused by the increasing oil transport in the Gulf of Finland (GoF), North-Eastern Baltic Sea. This integrative metamodel enables searching for the best management actions in the light of current knowledge and uncertainties.