Domain Ontology Linking to external data sets Liaison with Geonames

COIN-ARCTIC-SD 12 The spatial coordinates are defined for their part in the form of literals WKT Well Known Text associated with the object geometry by the property data owl: DataProperty geo: asWKT Figure 4.

6.2.1 Domain Ontology

We have extended our application ontology with a general ontology from hydrology used by the Consortium of Universities for the Advancement of Hydrologic Science CUAHSI Couch et al., 2014. Figure 6: An extract of the ontology of the CUAHSI hydrosphere. Only are developed in this hierarchy, the classes corresponding to the major metal concepts. This ontology defines a taxonomy that can structure hierarchically more than 4,000 words describing physical, chemical and biological measures related to water. It is used by the System Information CUAHSI CUAHSI-HIS Hydrologic Information System and consists of a set of servers and databases connected to client applications as web services to facilitate the discovery of time series data collected at a given point. We have taken this ontology defined in tabular form to translate it as an OWL class hierarchy Figure 6. The use of this ontology in our model is made by combining the type of corresponding measure in the terminology of CUAHSI with the descriptors measures identified during sampling. Figure 7: On the BKAI site at Banks Island a set of MS94 measures was taken. The M1881 12.10 measurement value is associated with a descriptor indicating that this is a measure of Calcium and its unit is mg l. This descriptor is connected to the domain ontology CUAHSI by a relationship rdf: type.

6.2.2 Linking to external data sets

A second enrichment of the initial data that allows their representation in the form of an RDF graph is the ability to link with other external data sets published respecting the principles of open and linked data. COIN-ARCTIC-SD 13 To demonstrate the potential of such enrichment we linked our data represented using our domain ontology with data from DBpedia Lehmann et al., 2015. We also linked our study data on regions with data from GeoNames.

6.2.3 Liaison with Geonames

The gazetteer GeoNames, which contains more than 10 million names, offers a free geographic database, accessible on the Internet under a Creative Commons license. Besides the geographical coordinates of each location listed, GeoNames offers data such as altitude, population, administrative subdivision, the postal code and links to corresponding Wikipedia pages in multiple languages. In addition to an access via web services, or a raw copy dump of the database, GeoNames publishes its data in RDF. Thus, each geographic feature of GeoNames is represented as a web resource. This web resource is identified by a stable URI, which, by negotiating the content, gives access to either the HTML page showing the entity on a map, or to a RDF description of the entity based upon a vocabulary defined by a OWL ontology Figure 8. Figure 8: Content Negotiation with GeoNames. Depending on the value of the “accept” header of an HTTP request, the resource server sends either an HTML page or an RDF description of this resource. For each of the regions in which the sites are located, we looked into GeoNames for an entry for the name used by filtering the results based on country and geolocation. This region of northern Canada is relatively poorly covered, sometimes we were driven to add new GeoNames entries this was the case for example of Bathurst Island. An owl link sameAs is then used to associate a region representation in GeoNames Figure 7. Figure 9: Data Binding to GeoNames and Dbpedia COIN-ARCTIC-SD 14

6.2.4 Liaison with Dbpedia