A Syntactic and Semantic Metadata Solution for Intelligent Applications in Earth Science

Different Earth science data collections are archived and distributed with different levels of metadata. Some files are metadata rich and thus allow applications to read this data while other files are metadata deficient and require the development of data specific reader modules for applications. This metadata heterogeneity leads to interoperability problems for scientists trying to use different types of data files with their applications. The Earth Science Markup Language (ESML) is designed as an elegant solution to this problem. ESML is an interchange technology that enables data interoperability by providing the additional metadata needed to allow applications to read the data file without requiring the development of new reader modules. Scientists can write external metadata files using the ESML Schema to describe the structure of their data files. Applications can then utilize the ESML Library to parse this ESML Description File and decode the data format. Software developers can now build data format independent scientific applications utilizing the ESML Library. Furthermore, the ESML Description File allows the addition of semantic metadata defined in different domain ontologies. Thus, the ESML Description File provides a complete syntactic and semantic metadata description of the data and links to the ontologies to provide a context for the semantic terms. This metadata description allows the development of intelligent applications that can now not only read the data but also understand and "use" the data. An agent based scientific data processing framework is one such intelligent application and will be presented in this paper.