Spatial Representations of Meteorological Data for Intelligent Retrieval

Exploitation of the historical record of meteorological data has long been impeded by an inability to rapidly identify relevant data, and by the difficulties of managing the huge volume of data involved. This paper presents a new approach to archiving and retrieving meteorological data that addresses these problems. Techniques of case-based reasoning from artificial intelligence are employed to construct a knowledge base that summarizes the content of a much larger archive. This knowledge base facilitates intelligent retrieval of past data from high-level descriptions of their content. Once relevant data are retrieved, analyses and summary views of the data can be immediately displayed. Full data sets can also be retrieved from slower media such as CD-ROM or magnetic tape if a more detailed investigation is required. The approach is implemented in MetVUW Workbench, a system for intelligent retrieval and display of historical meteorological data. The system is intended to serve as a “memory amplifier” for meteorologists that allows them to rapidly locate, analyse, and display historical situations of interest.