Development of an Ontology for Navigating and Discovering Hydrologic Data

Ontologies are increasingly emerging as a tool for bridging semantic heterogeneities, a problem that is prevalent in scientific data sets particularly across domains. The problem of semantic or descriptive heterogeneity is the result of historically independent annotation efforts by those that have collected or acted as steward for collected data responding to a specific mission statement without the realization that these data sets need to come together or at least complement each other in the long run. Because of this legacy approiach and the resulting status quo new ways need to be explored to overcome these annotation discrepancies not necessarily through a complete re-annotation but rather through tools that accept heterogeneity but try to mediate between the various existing description conventions. The hydrologic community is seeking to overcome these discrepancies for their constituency (through the Consortium of Universities for the Development of Hydrologic Sciences, CUAHSI, Hydrologic Information Systems, HIS, effort) by developing an information system in which disparate data sources can be accessed through a single search engine in which all data sources appear to be "one". To this end a data-discovery ontology for hydrologic data has been developed that permits registration of data sets to leaf concepts that are sufficiently detailed but one step more generic then what is typically used for data variable descriptions, for example Nitrate for all nitrate data collected. These leaf concepts originate from broader concept trees that can be navigated upwards through branches to more and more general concepts the top one of which is called HydroShpere. The ontology was in its first design meant to prove the concept and incorporated only a limited number of branches and leafs with detailed information only provided for the nutrients branch. Efforts are under way now to i) expose the ontology and its upper structure to wider audience vetting the approach and ii) to incorporate more branches to better capture a larger data spectrum that can be accessed for example in the USGS National Water Information System (NWIS), and EPA's STORET database.