This paper will describe a knowledge synthesis engine being constructed to create curated Data Albums to support case study analysis and climatology studies. The technological challenges in building a reusable and scalable knowledge synthesis engine are several. First, how to encode domain knowledge in a machine usable form? This knowledge must capture what information and data resources are relevant and the semantic relationships between the various fragments of information and data. Second, how to extract semantic information from various heterogeneous sources including unstructured texts using the encoded knowledge? Finally, how to automatically design of the structured database from the encoded knowledge to store all information and data fragments to support querying? The structured database must allow both knowledge overviews of an event as well as drill down capability needed for detailed analysis. An application ontology driven framework is being used to design the knowledge synthesis engine.
The knowledge synthesis engine is being applied to build a portal for hurricane case studies at the Global Hydrology and Resource Center (GHRC), a NASA Data Center. This portal will auto-generate Data Albums for specific hurricane events, compiling information from distributed resources such as NASA field campaign collections, relevant data sets, storm reports, pictures, videos and other useful sources.