Web 2.0 and Semantic Web Portal for Annotation and Discovery of Web Services in Geosciences Pearl Brazier 1 , Artem Chebotko 1 , Anthony Piazza 1 , Ann Q. Gates 2 , and Leonardo Salayandia 2 1 Department of Computer Science, University of Texas - Pan American, Edinburg, Texas, USA 2 Department of Computer Science, University of Texas at El Paso, El Paso, Texas, USA Abstract— With the rapid development of e-Science and scientific workflow technologies, geosciences Web portals are becoming even more important to support geoscien- tists in their research and discovery. While existing portals provide convenient means for humans to share and search for scientific data and services, they provide little support for automated agents like scientific workflow management systems. A pilot project, called GEO-SEED, addresses this drawback and supports both human and machine friendly interfaces for Web services annotation and discovery in geosciences. The architecture of the portal blends Web 2.0 and Semantic Web technologies, featuring an intuitive and collaborative work environment of a structured wiki and machine-interpretable metadata accessible via standard RDF, OWL, and SPARQL languages. This paper describes an ontology for Web Services annotation and elaborates on the wiki implementation and design of a relational RDF repository for the portal. Finally, it presents the results of the usability and performance studies for the project, report- ing on both human-effectiveness and machine-efficiency for knowledge collection, sharing, storing, and querying. Keywords: metadata; geosciences; Web service; repository; on- tology; wiki; RDF store 1. Introduction Geoscientists have been successful in collecting and shar- ing their research resources on the Web via specialized portals, such as the Geosciences Network (GEON) [1] and the Pan-American Center for Earth and Environmental Stud- ies (PACES) [2]. Using collaborative environments such as GEON, scientists can search for existing tools and datasets, as well as register their own research assets. The search and discovery capabilities of these repositories highly depend on metadata that describes available resources, its representa- tion, expressiveness, and storage and query interfaces. One of the drawbacks of many repositories is that they are largely intended for humans rather than for machines or automated agents. While a human-friendly interface for collecting and shar- ing information about scientific resources is a key require- ment for scientific Web portals, the importance of machine- interpretable metadata cannot be underestimated. Today, more and more scientists start to use workflow technologies to automate the steps they need to go through from raw datasets to potential scientific discovery. With the develop- ment of modern Scientific Workflow Management Systems (SWFMSs) that supports the specification, execution, and monitoring of scientific processes [3], [4], [5], automation and streamlining of many routine steps in workflow de- sign, composition, and execution are dependent on available machine-processable metadata. To enhance existing geosciences Web portals, a pilot project, called GEO-SEED [6] (GeoSciences WebService Discovery), was created to serve as a metadata repository for geosciences Web services, maintain rich annotation vocabu- lary, and provide both human and machine friendly interfaces for automatic Web services discovery. GEO-SEED is a new generation Web portal that uniquely builds on Web 2.0 and Semantic Web technologies to promote collaboration, integration, and interoperability of geosciences resources. In this new environment, Web services annotation and querying are driven by our Web Services Discovery Ontology that defines terminology for Web services description. GEO- SEED features a Web 2.0 structured wiki that provides an intuitive and interactive information venue to geoscientists and a Semantic Web knowledge management system that has RDF, OWL, and SPARQL interfaces for automated agents like SWFMSs. This paper extends a short paper by Brazier et al. [6] that presents GEO-SEED’s ontology and architecture with design and implementation sections for the structured wiki and RDF respository, as well as the usability and performance studies of the system. Among related projects that also adopt Web 2.0 and Se- mantic Web paradigms for scientific knowledge management are myExperiment [7] and BioCatalogue [8]. myExperiment focuses on building virtual research communities and sharing and using scientific workflows. While BioCatalogue and GEO-SEED promote a similar idea to provide a single reg- istration point for Web services, the projects target different domains – life and earth sciences, respectively. SWFMSs [3], [4], [5], [9], [10] also maintain metadata about workflow tasks, their implementation, and invocation in some type of a repository (see, e.g., task manager in [5]). Currently, scientists have to search for a Web service in a specialized metadata repository like GEON and then re-register it with a SWFMS, which doubles their effort. Machine-interpretable metadata supported by GEO-SEED is