Validation and Inter-comparison Against Observations of GODAE Ocean View Ocean Prediction Systems
Jinshan Xu1, Fraser J M Davidson1, Gregory C Smith2, Youyu Lu3, Fabrice Hernandez4, Charly Regnier5, Marie Drevillon5, Andrew Ryan6, Matthew Martin6, Todd D Spindler7, Gary B Brassington8 and Peter R Oke9, (1)Northwest Atlantic Fisheries Centre, Fisheries and Oceans Canada, St. John's, NF, Canada, (2)Environment Canada, Dorval, QC, Canada, (3)Bedford Inst of Oceanography, Dartmouth, NS, Canada, (4)IRD/Mercator Ocean, Ramonville St Agne, France, (5)Mercator Océan, Ramonville Saint Agne, France, (6)Met Office, Exeter, United Kingdom, (7)IMSG, College Park, MD, United States, (8)Bureau of Meteorology, Melbourne, VIC, Australia, (9)CSIRO, Hobart, Tasmania, SA, Australia
Abstract:
For weather forecasts, validation of forecast performance is done at the end user level as well as by the meteorological forecast centers. In the development of Ocean Prediction Capacity, the same level of care for ocean forecast performance and validation is needed. Herein we present results from a validation against observations of 6 Global Ocean Forecast Systems under the GODAE OceanView International Collaboration Network. These systems include the Global Ocean Ice Forecast System (GIOPS) developed by the Government of Canada, two systems PSY3 and PSY4 from the French Mercator-Ocean Ocean Forecasting Group, the FOAM system from UK met office, HYCOM-RTOFS from NOAA/NCEP/NWA of USA, and the Australian Bluelink-OceanMAPS system from the CSIRO, the Australian Meteorological Bureau and the Australian Navy.The observation data used in the comparison are sea surface temperature, sub-surface temperature, sub-surface salinity, sea level anomaly, and sea ice total concentration data. Results of the inter-comparison demonstrate forecast performance limits, strengths and weaknesses of each of the six systems. This work establishes validation protocols and routines by which all new prediction systems developed under the CONCEPTS Collaborative Network will be benchmarked prior to approval for operations. This includes anticipated delivery of CONCEPTS regional prediction systems over the next two years including a pan Canadian 1/12th degree resolution ice ocean prediction system and limited area 1/36th degree resolution prediction systems. The validation approach of comparing forecasts to observations at the time and location of the observation is called Class 4 metrics. It has been adopted by major international ocean prediction centers, and will be recommended to JCOMM-WMO as routine validation approach for operational oceanography worldwide.
