Generating composite thematic maps from semantically-different collections of shapefiles and map services

The Geosciences Network (GEON) project is a large-scale collaborative effort aimed at creating cyberinfrastructure for the Earth Sciences. GEON focuses on consistent representation, query access, integration and analysis of multiple semantically-different distributed spatial datasets. The paper will demonstrate knowledge-based integration of heterogeneous spatial data, stored mostly in shapefiles, and their online visualization as composite thematic maps. Following design principles of Service-Oriented Architecture, we combined the Gridsphere portal framework with ArcIMS-based services. Shapefiles contributed by geoscience researchers and registered in the GEONgrid, are transferred on-demand into an “assembly” ArcIMS service. Once the service is instantiated via the Java connector API, it can be queried from the middleware or from map clients. To process user queries, the middleware consults dataset ontologies, creates a global map legend, and generates a composite thematic map with regard to this global legend. Color assignment is governed by ontology considerations. Finally, the paper proposes certain performance enhancements.