Ontology-driven Automatic Geospatial-Processing Modeling based on Web-service Chaining

Earth System Science (ESS) research and applications often involve in collecting, analyzing and modeling with distributed heterogeneous geospatial data. Those data are processed step-by-step in geospatial analysis systems to extract information and knowledge products for applications and decision makings. Conceptually, such a step-by-step process forms a geospatial processing model that represents the knowledge of geospatial domain experts. This paper presents a study on ontology-driven automatic creation and execution of geospatial processing models in GeoBrain, a Web-service based geospatial knowledge system, to produce user-specific products. Web Services and Service-Oriented Architecture (SOA) provide a framework to support interoperable machine-to-machine interaction over a network. Web service chaining aimed to solve complex application tasks is changing the way of developing and deploying applications. By wrapping data and processes with Web services, it is easy to transform a geospatial processing model into a service chain. From design (knowledge), instantiation (information) to execution (data), this paper illustrates the whole life cycle of the geospatial-processing modeling and relevant implementation in GeoBrain. Ontology is usually used to capture domain knowledge. In this paper, a number of ontologies, including geospatial scientific ontology, geospatial data ontology and geospatial processing ontology, are introduced as the knowledge base to present geospatial domain terms and concepts, linkage between concepts and datasets, relationships among heterogeneous data, and associations between processes and data. By these ontologies, data and processes can be used for more effective discovery, automation, integration, and reuse across multiple diverse applications.