A Functional Ontology of Observation and Measurement Werner Kuhn Institute for Geoinformatics, University of Muenster Weselerstr. 253, 48151 Münster, Germany kuhn@uni-muenster.de Abstract. An ontology of observation and measurement is proposed, which models the relevant information processes independently of sensor technology. It is kept at a sufficiently general level to be widely applicable as well as compatible with a broad range of existing and evolving sensor and measurement standards. Its primary purpose is to serve as an extensible backbone for standards in the emerging semantic sensor web. It also provides a foundation for semantic reference systems by grounding the semantics of observations, as generators of data. In its current state, it does not yet deal with resolution and uncertainty, nor does it specify the notion of a semantic datum formally, but it establishes the ontological basis for these as well as other extensions. Keywords: observation, measurement, ontology, semantics, sensors. 1 Introduction Given that observation is the root of information, it is surprising how little we understand its ontology. Measurement theory, the body of literature on the mathematics of measurements, is only representational, treating questions of how to represent observed phenomena by symbols and how to manipulate these. Ontological questions like „what can be observed“ or „how do observations relate to reality?“ are not answered by it. Consequently, the semantics of information in general, and of observations in particular, rests on shaky ground. With sensor observations becoming ubiquitous and major societal decisions (concerning, for example, climate, security, or health) being taken based on them, an improved understanding of observations as information has become imperative. Answering some of the deepest and most pressing questions in geographic information science, such as how to model and monitor change, also requires progress in this direction. Furthermore, issues of scale, quality, trust, and reputation, are all intimately linked to observation processes. In response to these needs, this paper proposes a first cut at an ontology of observation. The ontology specifies observation as a process, not only as a result, and treats it as an information item with semantics that are independent of observation technology. The goal of this work is to understand the information processes involved in observations, not the details of physical, psychological, or technological processes.