Co-Convenors: S. Allen Macklin, Bernard A. Megrey (U.S.A.) and Igor I. Shevchenko (Russia)
BackgroundA stated objective of PICES is to provide data in exchangeable formats to better enable the evaluation of North Pacific ecosystems status and trends, and to support other strategic pursuits. PICES scientists face challenges in managing and delivering data in a shareable way. Furthermore, a growing number of ocean observing systems require data management and communication methodologies that conform to rigid standards and protocols. For the most part, traditional science education of the past century offered little training in data management. Today's typical scientist, although supportive of data exchange, lacks the background to understand techniques to facilitate it.
Summary of presentationsThis session was designed to acquaint PICES scientists with state-of-the-art information about metadata description, data delivery and data browsing techniques, with emphasis on existing standards and web services recommended for ocean observing systems. Basic to advanced methods were presented. Presenters described successful systems and learned ways of mapping existing data structures into conformant, exchangeable formats using no-cost, open-source and/or commercial applications. Session topics covered database aggregation, real-time ship-shore data exchange protocols, cabled observatory data management, and GIS techniques.
The session's invited speaker established a foundation for the proceedings by discussing, with examples, three conceptual models for the exchange of data between systems. The first model, relying on a central data structure for passing data among nodes, is used commonly in meteorology and oceanography. A second, more formal model uses instances of a common data model. Nodes exchange data with an instance of a common database, with data replicated between the common-instance databases. The third conceptual model entails wrapper software that encapsulates the data asset. Applications query the data asset using an intermediate layer, sometimes called an integrator or mediator, to identify the required data asset. The mediator then deals with critical data issues such as consolidation of parameter codes, units, replicate data, metadata content and multiple structures. The resulting data is provided to the user as a coherent and internally consistent data set.
This session was the first scientific session organized by TCODE. Papers covered topics of real time data delivery between ships at sea and land based laboratories, marine data exchange protocol options and standards, merging and aggregating data sets and data streams from diverse sources, and metadata search and federation systems. Eleven papers were delivered and five posters. The session was received very well by PICES scientists and attendance was high, sometimes reaching 50 people, even though two other concurrent sessions were taking place.