GODAE is sponsored by
3.4 Validation and Intercomparison of Analysis and Forecast Products
Lead author: Fabrice Hernandez (Mercator-Ocean)
Author/co-authors: F. Hernandez1, L. Bertino2, G. Brassington3, J. Cummings4, L. Crosnier1, F. Davidson5, P. Hacker6, M. Kamachi7, K. A. Lisæter2, Ray Mahdon8, Matt Martin8
1Mercator-Ocean, France
2NERSC, Norway
3Bureau of Meteorology, BMRC, Australia
4NRL, USA
5DFO, Canada
6University of Hawaii, USA
7MRI, Tsukuba, Japan
8Met Office, UK
Abstract
One of the main benefits of the GODAE 10-year activity is the implementation of ocean forecasting systems in several countries. In 2008, several systems are operated routinely, at global or basin scale. Among them, the BlueLink (Australia), HYCOM (USA), MOVE/MRI.COM (Japan), Mercator (France), FOAM (United Kingdom) and TOPAZ (Norway) systems offered to demonstrate their operational feasibility by performing an intercomparison exercise during a three months period (February to April 2008). The objectives are: a) to show that operational ocean forecasting systems are operated routinely in different countries, and that they can interact; b) to perform in a similar way a scientific validation aimed to assess the quality of the ocean estimates, the performance, and forecasting capabilities of each system; and c) to learn from this intercomparison exercise to increase inter-operability and collaboration in real time.
A first international intercomparison exercise in the framework of the EU MERSEA Strand1 project has been performed in 2004, from which part of the assessment methodology has been kept. A common set of diagnostics, re-visited since MERSEA Strand1, has been adopted by all partners. These diagnostics, based on metrics, allow for each system: a) to verify if ocean estimates are consistent with the current general knowledge of the dynamics; b) to evaluate the accuracy of delivered products, compared to space and in-situ observations; c) to evaluate in target areas if major ocean processes are well reproduced; and d) to infer the forecasting skill of the system. Using the same diagnostics also allows one to intercompare the results from each system consistently. Moreover, a standardization of format outputs, and dedicated server and tools (i.e., NetCDF, OPeNDAP technologies) make possible direct comparison between products of different systems.
All the ocean forecasting systems provide daily fields describing at the mesoscale the ocean circulation (currents, sea level, mixed layer depth...) and the water mass changes (temperature, salinity). Some also include a representation of the sea-ice variations (concentration, thickness, drift, snow cover...). The intercomparison and validation first focus on the oceanic consistency of these products. General representations of tropical, subtropical and subpolar gyres together with differences to climatology are discussed. Then, in order to evaluate the accuracy, direct and synoptic comparisons are performed against satellite and in-situ observations (e.g., temperature and salinity profiles from Argo and other near real time delivery; SST products from GHRSST; surface currents derived from satellite altimetry and scatterometry; and sea-ice products from satellite radiometry).
(Last Updated: 16-10-2008)