The FAIR principles aim to promote sharing of science data, including, for example, so that data sets from different sources and/or locations can be combined to answer new science questions. It is relatively easy to achieve ‘FA’: adequate documentation in metadata and data portals will allow data to be found and accessed. It is much more difficult to accomplish the ‘IR’ because there are many complex and data-specific challenges in making data interoperable and re-useable. In this paper we describe the experience of exchanging underwater image datasets (images and annotations) between teams of researchers in Australia and France. We report on the process to achieve interoperability and re-use of a suite of shallow coastal and deep-sea datasets, by deriving consistent metrics of internationally-identified ‘Essential Biological Variables’ (EBV) from combined data.