An effective ontology architecture enables the development of a geospatial semantic system that forges multiple geospatial data sources into a powerful cross-discipline knowledge source. This paper suggests types of ontologies that could support a geospatial semantic system. Motivations of each of the ontology types will be expounded, as well as potential areas for standardization by the geospatial community. Finally, the use of this approach within the OGC GSW Interoperability Experiment will be discussed.