Authors: D. Hassell, S. Bartholomew, B.N.Lawrence.
Abstract
The Climate Forecast conventions (CF) provide a data model for the layout of weather and climate data in storage. Recent work has prepared the way for supporting finite element grids in CF via incorporation of UGRID into the CF data model. This task takes that conventions support and demonstrates that it is implementable by implementing full support for ICON, FESOM, LFRic as well as the DestinE generic state vector HEALPix mesh in the python implementation of the CF data model (cfdm) and a science library which builds upon that data model (cf-python). Part of this is to support an existing proposed CF extension for HEALPix through the CF process.
Metadata conventions, and tools that understand them, are a key requirement towards achieving FAIR datasets – mainly impacting the “I” (interoperability) aspect. This is particularly the case for complex grids of the types already mentioned. The structures of these grids are often not intuitive for the casual reader, but the CF conventions and cf-python infrastructure provides a guaranteed correct interpretation when using the data in analysis workflows, particularly when combined with datasets on other grids. This will be the case for the lifetime of the data (which is usually many years, or even decades), and future interoperability will only be achieved if the data have been encoded according to accepted metadata conventions, and the tools still exist that leverage those conventions.
