Most approaches on ontology matching and mapping focus on finding equality or sameAs correspondences.
However, concepts of different ontologies may be related differently, e.g. according to an is-a or part-of relationship. Supporting such additional semantic correspondences can greatly improve the expressiveness of ontology mappings and their usefulness for tasks such as ontology merging and ontology evolution. We present a new approach called STROMA (SemanTic Refinement of Ontology MAppings) to determine semantic ontology mappings. In contrast to previous approaches, it follows a so-called enrichment strategy that refines the mappings determined with a state-of-the-art match tool such as COMA, GOMMA or AgreementMaker.STROMA employs several techniques including the use of background knowledge and linguistic approaches to identify the additional kinds of correspondences. In a comparative evaluation with previously proposed approaches STROMA shows a high effectiveness. We will also discuss the use of semantic mappings to better merge ontologies. Such an enhanced merge is supported by our tool ATOM (Automatic Target-driven Ontology Merging).