Designing an API that will allow the Flexible Sampling of Multidimensional Data:

Peter. Trevelyan, and Richard Carne Met Office, Exeter, United Kingdom Designing an API that will allow the Flexible Sampling of Multidimensional Data: The problem with most of the geospatially enabled APIs is that they are really only designed to deal with 2D or at best 3D data. MetOcean data is inherently 4D and has even more dimensions if ensemble members are considered to be an extra dimension. This complexity may be appreciated by the MetOcean community, but other disciplines tend to abstract the vertical dimension by considering height as an attribute and not a “first class” axis. This restricts the use of MetOcean data as it limits the way the data may be sampled. By considering MetOcean data as a hyper cube (a n dimensional cube) then more complex sampling may be achieved e.g. full 4D trajectory analysis. This talk will outline the methodology behind the design of the MetOcean extensions to the OGC's WCS2.0 (Web Coverage Service). The starting point was the creation of a representative set of use cases that were examined to understand the sampling shape of the data needed to be extracted and consumed by the client. A summary of the use cases will be presented and the implications of design of the extension of the OGC's WCS2.0 core standard as well as the current progress through the OGC standardisation process. This approach has the advantage that only relevant data is passed to the client and this not only reduces the size of the data, but also computation on the client. The importance of this work is that it defines an international standard for a MetOcean responsive API that may be used by GIS tools so allowing the integration of MetOcean data with other disciplines.