Australia's Integrated Marine Observing System (IMOS, imos.org.au) is research infrastructure to establish an enduring observing program for Australian oceanic waters and shelf seas. The observations cover physical, biological, and chemical variables to address themes of multi-decadal ocean change, climate variability and weather extremes, boundary currents and inter-basin flows, continental shelf processes and ecosystem responses.IMOS observations are collected by national facilities based on various platform types and operated by partner institutions around the country. In this paper we describe the infrastructure and workflows developed to manage and distribute the data to the public. We highlight the existing standards and open-source software we have adopted, and the contributions we have made. To demonstrate the value of this infrastructure we provide some illustrations of use and uptake.All IMOS data are freely and openly available to the public via the Ocean Portal (https://imos.aodn.org.au). All IMOSdeveloped software is open-source and accessible at https:// github.com/aodn.

The exchange of scientific datasets online and their subsequent use by service-centric applications requires semantic description of the data objects, or features, being transacted. This is particularly the case in the Earth Systems Sciences. Semantic repositories provide a partial answer to generating rich content. Ideally these repositories should be founded in a framework that permits cross-referencing between independently established semantic data-stores and which provides for a loose coupling between repositories and the agents or clients that will use them. We investigated the applicability of using an ISO 19110-based Feature Catalogue as a cross-domain, semantic repository for various Earth Systems Science communities of interest. Our aim was to develop a repository and a set of services capable of providing semantic content for consumption by smart clients. The constraint applied throughout the research was to develop a set of tools that would present a very low uptake barrier for programmers and domain specialists alike. To meet this challenge, we used Representational State Transfer (REST)-based services to expose content from an enhanced implementation of an ISO 19110-based Feature Catalogue. This article describes how the ISO 19110 conceptual model was augmented during implementation to cater for the requirements of two large multi-disciplinary science groups: the Scientific Committee on Antarctic Research and the Australian Ocean Data Network. The reasons for opting for a REST-based service pattern are discussed and the three REST service types that were developed are described.

ABSTRACT