Interdisciplinary probabilistic network to examine the possibility to restore potential Baltic salmon rivers

Restoration projects of potential Atlantic salmon (Salmo salar) rivers in the Baltic Sea have known relatively limited success. In order to identify the possible factors currently limiting the re-establishment of salmon populations in potential salmon rivers, we compare a wild salmon river with potential salmon rivers in terms of the biological characteristics of the salmon population, the fishing pressure on the population and the rivers socio-economic and biological environment. The differences in biological and socio-economical circumstances between rivers can be examined through separate biological, economical and socio-economic studies. The results from the different disciplines can later be linked together within one single probabilistic or Bayesian network model. The probabilistic network has the advantage that the different conditional relationships can be obtained through different methodologies. In the case of Baltic salmon, the biological estimates within the Bayesian network are obtained through a probabilistic mark-recapture model of tagging data, the economic dependencies are obtained through bio-economic models and contingent valuation, while the social dependencies come from questionnaires and in-depth interviews with key people. Network modelling allows us to examine not only what has happened in the past but through the expression of causal relationships within the model structure, it also allows us to examine what would happen to the salmon populations if certain biological, economical and social conditions in potential salmon rivers would be different. It is therefore possible to examine the effect of different management actions on the fishing behaviour of the fishermen in the different rivers and to examine for each river which combination of management actions has the highest probability of successfully restoring the salmon stocks.