THE NEW MET OFFICE GLOBAL OCEAN FORECAST SYSTEM AT 1/12° RESOLUTION

• Main Authors: Aguiar, Ana, Waters, Jennifer, Price, Martin, Inverarity, Gordon, Pequignet, Christine, Maksymczuk, Jan, Smout-Day, Kerry, Martin, Matthew, Bell, Mike, King, Robert, While, James, Lea, Daniel, Siddorn, John
• Other Authors: UK Met Office, Shom, Ifremer, EuroGOOS AISBL
• Format: Conference Object
• Language: English
• Published: HAL CCSD 2021
• Subjects: Ocean model; eddy-resolving; ORCA12; neighbourhood verification methods; [SDE]Environmental Sciences; Sea ice
• Online Access: https://hal.archives-ouvertes.fr/hal-03339590; https://hal.archives-ouvertes.fr/hal-03339590/document; https://hal.archives-ouvertes.fr/hal-03339590/file/EuroGOOS2021_extended_abstract_Aguiar.pdf

International audience The Met Offi ce has recently upgraded its operational Forecasting Ocean Assimilation Model (FOAM) from an eddy permitting 1/4° tripolar grid (ORCA025) to the eddy resolving 1/12° ORCA12 confi guration. FOAM-ORCA12 uses NEMOv3.6 (GO6 confi guration) coupled to CICE (GSI8.1 confi guration) for the ocean and sea-ice components, respectively. It assimilates observations of sea surface temperature (SST), temperature and salinity profi les, altimeter sea level anomaly and sea ice concentration, via NEMOVAR which is a multivariate incremental 3DVar scheme that runs over a 1-day time window. Qualitatively FOAM-ORCA12 better represents the details of mesoscale features in SST and surface currents. Traditional statistical verifi cation methods suggest that the new system performs similarly or slightly worse than the equivalent 1/4° system. However, it is known that comparisons of models running at different resolutions suffer from a double penalty effect, whereby higher-resolution models are penalised more than lower-resolution models for features that are offset in time and space. Results are shown from neighbourhood verifi cation methods which use common spatial scales for a fairer comparison between confi gurations of different resolutions, applied to SST. We show that, as neighbourhood sizes increase, ORCA12 consistently has lower Continuous Ranked Probability Scores than ORCA025.