SIGMAR: A Fuzzy Expert System for Critiquing Marine Forecasts

Meteorological information and knowledge are often uncertain, ambiguous, or vaguely defined. Fuzzy logic lets expert systems perform optimally with uncertain or ambiguous data and knowledge. With a fuzzy logic framework, one can efficiently implement linguistically expressed rules derived from experts. Operational meteorology is therefore treated as a fuzzy environment. An argument is made for the applicability of methods based on fuzzy logic for the optimal solution of problems related to the evaluation of meteorological data and forecasts. An expert system, SIGMAR, has been designed which uses fuzzy methods to interpret meteorological data. The system automatically evaluates the significance of actual wind reports. Two activities that challenge weather forecasters are coping with information overload and maintaining accuracy of forecasts. Both tasks can be performed more easily and consistently with SIGMAR. The system efficiently identifies significant information contained within huge amounts of data. Forecasters using the system can more consistently and easily monitor the accuracy of weather forecasts. Systems such as that described here are bound to become more common as time goes on.