GOSI9: UK Global Ocean and Sea Ice configurations

Abstract. The UK Global Ocean and Sea Ice configuration version 9 (GOSI9) is a new traceable hierarchy of three model configurations at 1°, 1/4° and 1/12° based on the version 4.0.4 of the NEMO code. GOSI9 has been developed as part of the UK's Joint Marine Modelling Programme (JMMP), a partnership between Met Office, National Oceanography Centre, British Antarctic Survey, and Centre for Polar Observation and Modelling. Following a seamless approach it will be used for a variety of applications across a wide range of spatial and temporal resolutions: short-range coupled NWP forecasts, ocean forecasts, seasonal and decadal forecasts, climate (including CMIP7) and Earth system modelling. The GOSI9 configurations are described in detail with special focus on the updates since the previous version (GO6-GSI8). The impacts of these updates are assessed with the 1/4° resolution configuration and results from 30-year ocean-ice integrations forced by CORE2 fluxes are presented for the three resolutions. The upgrade to NEMO 4.0.4 includes a new sea ice model SI³ (Sea Ice modelling Integrated Initiative) and faster integration achieved through improved performance and a significant increase in the length of the time step. Performances are overall improved compared to GO6-GSI8. The temperature and salinity drifts are largely reduced thanks to the upgrade to NEMO 4.0.4 and the adoption of 4^th order horizontal and vertical advections helping to reduce the numerical mixing. To improve the representation of the Southern Ocean, a scale-aware form of the Gent-McWilliams parametrisation and the application of a partial slip lateral boundary condition on momentum in the Southern Ocean have been added resulting in a stronger and more realistic Antarctic Circumpolar Current (ACC) transport and a reduction of the temperature and salinity biases along the shelf of Antarctica. In the Arctic, the representation of sea ice is improved leading to a reduction in surface temperature and salinity biases. In particular, the excessive and unrealistic Arctic summer sea ice melt in GO6-GSI8 is significantly improved in GOSI9 and can be attributed to the change in the sea ice model and to the higher albedos which increased sea ice thickness.