METAFOR Common Metadata for Climate Modelling Digital Repositories

METAFOR Common Metadata for Climate Modelling Digital Repositories 11 partners Total budget of 2. 2 M€ (£ 1. 5 M, 3 M$) • NCAS, UK (Coordinator) Starting in March • BADC, UK Duration 3 years • CERFACS, France • Models and Data, MPI, Germany • IPSL, France • University of Manchester, UK 2008 • Met Office, UK • Administratia Nationala de Meterologie, Romania • Météo France, CNRM, France • CLIMPACT, France EU I 3 Infrastructure project • CICS, Princeton University, USA - The proposal

Vision of METAFOR An Information Model that is common for all stages of both production and the use of climate model data. CIM Common Information Model based on current, emerging and new metadata standards Tools that populate, create, manipulate, convert and exploit this Common Information Model to allow climate models and climate model data to be intercomparable and sharable. EU I 3 Infrastructure project - The proposal

The open standard developed in METAFOR will play a catalytic role in the way next generation climate data repositories, such as IPCC AR 5, are organised, preserved and accessed. The diagram below shows the nature of the current 4 th assessment of the Intergovernmental Panel on Climate Change (IPCC AR 4) data repository and shows the challenge for IPCC AR 5 which will be larger, more comprehensive and more complex. EU I 3 Infrastructure project - The proposal

METAFOR has 286 person months of effort WP 2 CIM Common Information Model WP 1 Admin Eric Guilyardi, NCAS 14 WP 2 CIM Lois Steenman. Clark, NCAS 101 WP 3 Verification Laurent Fairhead, IPSL 47 WP 4 Portals Bryan Lawrence, BADC 32 WP 5 Exploitation tools Allyn Treshansky, UKMO 41 WP 6 CIM creation tools Michael Lautenschlager, MPI 39 WP 7 Dissemination Eric Guilyardi, NCAS 12 WP 6 CIM creation tools WP 3 Verification WP 4 Data Portals WP 5 CIM exploitation tools

Oct 2007 Oct 2008 Oct 2009 Oct 2010 Oct 2011 Curator METAFOR Expt design Metadata requirements running expts Metadata collection AR 5 Portal Metadata delivery Timelines are tight AR 5 experiments could be running before metadata standards and tools are available in their final form Limited effort is available for accommodating modelling groups outside the current European/US METAFOR/Curator discussions AR 5

Oct 2007 Oct 2008 Oct 2009 Oct 2010 Oct 2011 Curator METAFOR AR 5 METAFOR for AR 5 • CIM based data portals design and implementation • CIM to encompass AR 5 metadata requirements Access to CIM tools and services • CIM verification • CIM tools for ingestion part of AR 5 quality control tools made available to modellers • CIM exploitation tools

METAFOR has to work closely with AR 5, as this is it’s key use case - definition of metadata requirements First 3 months of METAFOR metadata gap analysis Q: what metadata was missing from AR 4 that would have enhanced the shareability, the analysis and the understanding of the AR 4 data set - logistics of collecting data and metadata Q: can METAFOR build on Curator experience to enhance the metadata management issues? - sharing of metadata tools and services In designing tools and services can METAFOR be involved in - Playing a part in AR 5 quality control - Taking part in the CMOR re-design? - Delivery of data via the distributed data portals… BADC, WCDC will be CIM compliant data services

CIM Common Information Model based on current, emerging and new metadata standards NMM FLUME CF Key to success will be understanding the metadata need and requirements PRISM Curator Metadata Provenance international standards {ISO} community standards {CF} project standards {NMM} workflow standards {Curator, ESG, FLUME} hand generated {Numsim, DIF} externally imposed {CMOR, CERA} CIM metadata is common but an amalgam of different standards with diverse governance, each evolving and developing at different rates.

METAFOR gap analysis For the AR 5 use case. - Lessons learnt from AR 5, both data providers and data users as well as the data collection agents - What level of quality control on metadata will be required for AR 5 and how is this going to be managed? - Is the metadata going to be used for some element of quality control for the AR 5 data? - What level of 3 rd party involvement will be needed for the AR 5 metadata, is it expected to be used by the impacts community?