Ontologies have been used to support the interoperability of geospatial data by overcoming semantic problems related to semantic heterogeneities and to differences in data usage contexts. Ideally, to solve semantic heterogeneities, the data models involved in the interoperability process could be enriched, and the relationships between their elements could be defined based on a universal geospatial ontology. However, such ontology would encounter difficulties in achieving an efficient interoperability. This chapter aims to argue that universal ontology-based interoperability remains vulnerable to the risks of uncertain meaning of geospatial data that may go unnoticed during the interoperability process. The chapter discusses these risks and proposes a systematic approach to better support users dealing with them. The proposed approach identifies the risks, assesses their severity, and helps users to make decisions about them.
Nowadays, we observe a wide use of geospatial data which allows users to perceive geospatial1 properties of phenomena (e.g., position, shape, orientation, size), their geospatial relationships with other phenomena (e.g., adjacency, inclusion, distance), their geospatial distribution (e.g., regular, cluster, random), and their spatio-temporal properties and relationships (e.g., speed of expansion, minimum distance for a given period). Geospatial data are designed originally for specific uses. However, they are stored in different systems (e.g., GISs, DBMSs, multidimensional warehouses, Knowledge-Based systems) and typically end up being integrated with other data and used for new purposes. The use of geospatial data from different sources faces a challenge regarding the heterogeneity of data semantics (e.g., differences in meaning and in appropriateness for usage contexts) (Bishr, 1998; Brodeur, 2004). Dealing with such heterogeneity has been the principal aim of semantic interoperability research for the last two decades (Bishr, 1998; ISO/IEC, 1993; Brodeur, 2004; Kuhn, 2005; Giunchiglia, Maltese, Farazi, & Dutta, 2010). Semantic interoperability requires interpreting data in the context of use. The process of interpretation involves the comparison of different data descriptions and adapting them to the context of use (Brodeur, 2004; Vaccari, Shvaiko, & Marchese, 2009).
Comparing data descriptions could be enriched based on a universal geospatial ontology. Such ontology could be used to link two different data models and could thus resolve problems of semantic heterogeneity between geospatial data from different sources.
However, we believe that a universal geospatial ontology cannot resolve all the problems related to the semantic heterogeneity. Accordingly, the interoperability based on universal geospatial ontology still faces risks of misinterpreting data. In this chapter, we argue that universal ontology solution remains vulnerable to risks of data misinterpretation and we propose a systematic way to better support stakeholders (software agents and users) in dealing with such risks.
The first section of this chapter introduces geospatial data interoperability and discusses how universal geospatial ontology can facilitate the interpretation of data exchanged during the interoperability process. The second section discusses the risks of data misinterpretation related to the imperfection of universal geospatial ontology during the process of semantic interoperability for geospatial data; it defines such risks and determines their principal causes. The third section proposes a systematic approach to help users dealing with these risks. It also presents an example of application of the proposed approach and shows how this approach can help managing these risks. The last section presents future research directions and is followed by a conclusion.