NEPTUNE Canada

Data Interoperability Workshop

Nov 28, 2008

Specialists from several leading institutions gathered in early September at the University of Victoria to explore a way forward for ocean data interoperability.

Benoît Pirenne, Jason Marling and Paul Gaughan at the DMAS data interoperability workshop Sept, 2008.

The workshop afforded a great opportunity to grapple with key issues in data interoperability for Ocean Science. To help us put things in perspective, Markus Dolensky, formerly with the European Southern Observatory in Munich, Germany was invited. Markus has spent the last decade developing interoperability solutions in the field of astronomy. He delivered a number of presentations explaining some of the basic issues and solutions already in place in this field. Other participants described relevant work at their own institutes.

On day two, John Graybeal of the Monterey Bay Aquarium Research Institute spoke about ontologies and vocabularies. Jason Marling of Fisheries and Oceans Canada Integrated Science Data Management explained standards in use in his government agency. The rest of the day was devoted to discussions on issues such as registries and data services.

MBARI's John Graybeal discusses the finer aspects of data interoperability.

The workshop concluded with drafting of an initial list of sites to comprise a small consortium of data repositories for the North-East Pacific. In the next few months the Interoperability project will move forward with the selection of data exchange standards, the use of registries, etc.

Key Findings

Lorraine Brasseur of the Ocean Observatory Initiative. (Click to enlarge.)

• Ocean Observatories like NEPTUNE Canada and VENUS are tremendous platforms for integrated, collaborative, and cross-disciplinary research for two reasons: not only are the data streams long-term and continuous, but all streams and their sensors are also registered on the same spatial-temporal frame. This opens a great potential for analysis across different instruments and data types. When multiple data archives are registered to the same frame, this analysis potential expands significantly.
• The initial set of variables we're considering for interoperability include temperature, salinity, pressure (TSP) and their derived values. The ability to exchange multimedia data is very attractive as well. The focus of interoperability is not to grab data, but to show what is available and make it accessible.
• The experience of astronomy shows that heavily processed, multi-instrument focused surveys produce 20 times more publications than simple raw data. Facilities to help assemble such large datasets should be given priority. A suitable use case should be identified by the community.
• Interoperability should allow us to extend one or more dimensions (e.g., time) for a set of variables. This would immediately provide a more intensely observed area to the scientific community. The selection of variables such as TSP will illustrate the power of synergistic, highly observed areas. This can be achieved by providing data from global resources.

Robert Arko of Lamont Doherty Earth Observatory (left) and Bob Branton of the Ocean Tracking Network (right) during discussions at the data interoperability workshop.

Workshop attendees agreed on the value of pursuing interoperability over the next 18 months and to start with the exchange of standard variables (such as TSP) archived at our different data centres. The following table below shows what types of data participant centres have agreed to share in this preliminary phase. This list is expected to evolve as more players get involved and more data sources come on-line.

Institution	Variables: (TSP, others)	Other Data: Biology	Other data: Imaging
OTN/Dalhousie (OBIS)
MBARI
Bremen Univ. (MARUM)
ISDM/DFO
CHONe
LDEO

Next Steps

Our next steps include:

• drafting agreements
• identifying a suitable host for the registry of data and services (hopefully an existing one)
• reaching agreement on data exchange technologies (protocols, formats)
• implementation

We hope to reach the implementation phase by May 2009 with a first prototype demonstration shortly thereafter.

This workshop was very valuable in that it allowed a number of organizations with similar interests to evaluate possible initiatives that will benefit our respective user communities. A pragmatic approach to data interoperability was proposed whereby some of us agree to explore the exchange of a well-defined set of ocean data variables. Examples of how astronomy makes data from various origins interoperable were presented; these examples illustrate the tremendous potential of interoperability for empowering scientists. Finally, a roadmap is being prepared to achieve practical results in the near future.

Group photo of participants in dmas data interoperability workshop, Sept/08. (Click to enlarge.)

Return to News Home