Discovery of agriculture data using federated catalogue service

With the widespread application of Agriculture GIS, an increasing requirement for discovering various agriculture data from multiple distributed data center has brought challenge to scientists. Data users need to interact with different data center to retrieve data of interest, and also have to solve the heterogeneity of both query interface and metadata model. For example, users need interact with NASA Reverb or other web portal to get MODIS data, for AVHRR data, users have to execute the query through NOAA CLASS system. As geospatial catalogues service are becoming accessible online through public query interfaces, building a standard-compliant web service to discover multi-source agriculture data is also becoming necessary for most data users. Since NOAA and NASA archived many Earth Observation data for agriculture research, this study investigates the feasibility of federating two distinct data center to discover agriculture data: NOAA Comprehensive Large Array-data Stewardship System (CLASS) and NASA Earth Observing System (EOS) Clearinghouse (ECHO). To implement federated catalogue, the OpenGIS Catalogue Service for web (CSW) is adopted to define the public query interfaces for data users. ISO profile and Core profile of CSW are implemented for define the metadata model for CSW response. The paper addresses two issues for implementing federated catalogue service: mapping querying interface and mapping metadata model. For query interface, a mapping relationship between OGC CSW core query predicate and target data center need to be established. This mapping table will translate the query parameters encoded in the request of CSW to the dedicated query language defined by back-end´s data center. As to the mapping relationship from native metadata model to the public model used in federated CSW. The attributes returned from distinct data center will be converted to OGC Core returnable elements and ISO elements. By implementing federated catalogue service, data us- rs only need to interface with public query language and metadata, which make it easy to find data from multiple data center by hiding the complicated low level details. Moreover, by using CSW-compliant web service, data users could integrate data discovering operation in more complex workflow.