This is an ontology for health surveillance, assuming that surveillance data is recorded at three main levels: SAMPLE data - data about specific collected samples; OBSERVATIONS, which can aggregate several samples; and OBSERVATIONAL CONTEXT, which is the context in which the observations were made (for instance clinical observation, surveillance sampling, etc). HEALTH EVENT is the occurrence in the real World to be modeled. This occurrence is an abstract concept, as it generally cannot be determined what are its boundaries in time, space, or population units. We assume that several OBSERVATIONS are derived from each health event, and recorded in the same or different databases - for instance clinical journal, laboratory database, abattoir. At the same time or not. AHSO attempts to model these time specific observations. The ontology is focused on the modeling of the animal and health information connected to a single observation, made at a specific point on time. All the information recorded is relative to that specific point in time. For instance, an animal belongs to a herd, which belongs to an owner, AT THE TIME OF THE OBSERVATION. We do not attempt to model how that animal changed herds on time. Similarly, we do not model clinical progression. It is up to those using the data to create definitions of when two observations should be connected to the same event.