An adaptive fuzzy expert system for predicting structure, dynamics and distribution of herring shoals

For many of the worlds pelagic stocks, structure, dynamics and mesoscale distribution of fish shoals has considerable importance to central issues in fisheries management; stock structure, stock assessment, resilience and harvest control. A model is presented here, that attempts to bridge existing gaps in our basic understanding of the biological and ecological mechanisms underpinning the behavioural responses of herring, and how these govern spatial dynamics of shoals. The model, CLUPEX, is developed in the framework of an expert system and utilises fuzzy logic to capture and integrate scientific and local knowledge in the form of heuristic rules. Using input on biotic and abiotic environmental conditions, CLUPEX uses the rules to provide quantitative and qualitative predictions on the structure, dynamics and mesoscale distribution of shoals of migratory adult herring during different stages of their annual life cycle. Test predictions corresponded well with observed patterns, although accuracy for specific circumstances may be limited by the resolution of the knowledge. However, by adding specific local knowledge and adjusting weighting parameters, CLUPEX can be adapted to provide more accurate and precise predictions. The user interface combines hypertext and an explanation facility that is fully cross-referenced to a database, to provide an intuitive and transparent feel rarely found in more traditional analytical models.