THE DEVELOPMENT OF A HOLISTIC EXPERT SYSTEM FOR INTEGRATED COASTAL ZONE MANAGEMENT

Abstract

Coastal data and information comprise a massive and complex resource, which is vital to the practice of Integrated Coastal Zone Management (ICZM), an increasingly important application. ICZM is just as complex, but uses the holistic paradigm to deal with the sophistication. The application domain and its resource require a tool of matching characteristics, which is facilitated by the current wide availability of high performance computing. An object-oriented expert system, COAMES, has been constructed to prove this concept. The application of expert systems to ICZM in particular has been flagged as a viable challenge and yet very few have taken it up. COAMES uses the Dempster- Shafer theory of evidence to reason with uncertainty and importantly introduces the power of ignorance and integration to model the holistic approach. In addition, object orientation enables a modular approach, embodied in the inference engine - knowledge base separation. Two case studies have been developed to test COAMES. In both case studies, knowledge has been successfully used to drive data and actions using metadata. Thus a holism of data, information and knowledge has been achieved. Also, a technological holism has been proved through the effective classification of landforms on the rapidly eroding Holderness coast. A holism across disciplines and CZM institutions has been effected by intelligent metadata management of a Fal Estuary dataset. Finally, the differing spatial and temporal scales that the two case studies operate at implicitly demonstrate a holism of scale, though explicit means of managing scale were suggested. In all cases the same knowledge structure was used to effectively manage and disseminate coastal data, information and knowledge.