A semantic annotation tool for hydrologic sciences

Abstract

Semantic annotations are playing an increasingly important role in the world of metadata, more specifically when dealing with semantic heterogeneities between information systems. The need to bring together disparate data sources (in terms of syntax and semantics) so they can be searched simultaneously from a single search environment has become one of the most challenging tasks in developing information systems that span multiple communities as is common in the geosciences. The key problem lies in the legacy information systems, in which, at the time of development, each system used (and continues to use) its own semantic framework to identify variable codes and names, as well as annotating the collected data with metadata. This lack of a common metadata framework as well the uncoordinated use of descriptors and controlled vocabularies has led to a situation in which synonyms and hyponyms abound. Experience has shown that a centralized system with just one vocabulary for all is not feasible. Rather, in order to overcome these discrepancies it is important to realize that heterogeneity is an inevitable aspect of the scientific data world that needs to be accommodated. This paper describes the development and end use of an application that is designed to connect arbitrary variable names to specific concepts in layered search ontology. We will demonstrate the utility of this application through its deployment for the Consortium for the Advancement of Hydrologic Sciences Inc. (CUAHSI) network of testbeds and report on the issues that emerged carrying out variable and concept tagging. These issues concern specificity of a concept, ancillary information needed when identifying proper ontology locations, and multiple appearances of variables at different locations.