CONTAMINANT HYDROGEOLOGY ONTOLOGY: INTEGRATION WITH THE LINKED OPEN DATA CLOUD

Large volumes of hydrogeochemical data originate from a diverse and distributed set of instruments that measure the geochemical constituents of surface and groundwater. These rapidly growing contaminant hydrogeology data are stored in isolated and networked databases. Accessing and integrating the heterogeneously structured data from these databases are difficult and time consuming. One way to overcome the data integration and software interoperation is to develop ontologies that formally and explicitly represent the knowledge in the domain (e.g., contaminant hydrogeology). The Contaminant Hydrogeology Ontology (CHO) is developed, applying the Protégé editor, to manage hydrological data for Georgia, USA.

CHO is a conceptual knowledge model for the contaminant hydrogeology domain in which concepts (classes, e.g. Contaminant, Aquifer), their relationships (properties, e.g. contaminates), and constraints are defined using the OWL constructs. CHO models the knowledge about the interaction of contaminants, their source, and how the aquifer properties affect the contaminant transport in the groundwater. CHO consists of top level classes (e.g., WaterQuality) that extend SWEET ontologies (e.g., phenEnvirImpact), from which additional concepts from the contaminant hydrogeology domain are created as subclasses (e.g., PrimaryContaminant). The consistency of the ontology was tested and maintained using the Pellet reasoner. Classes from CHO are associated with data from public databases (e.g., GA EPD’s HSI and LUST), allowing data to inherit the relationships established in CHO and making them discoverable on the Web by software agents.

Integration of CHO with the Linked Open Data (LOD) Cloud, which is presently underway, will enable access to the CHO represented data from mobile devices (e.g., smartphones and tablets) that use DBpedia Mobile client. This is done by integrating (linking) the class and property URIs from CHO with DBpedia ontology’s class and property URIs based on the Linked Open Data principles.