Projecting the impacts of climate change in the Atlantic Ocean: a global ocean downscaling approach

The latest generation CMIP6-class Earth system models (ESMs) are a great tool for projecting climate variability on multi-centennial and global scales as they are designed to explicitly represent the process-coupling amongst the different Earth system components (atmosphere, ocean, land, cryosphere, biosphere) and prioritise system robustness such as minimisation of drift. However, CMIP6-ESMs do not accurately represent the fine-scale circulation and water-masses in ocean margins and shelf seas since by design: (i) their resolution is too coarse and so they only implicitly include regional-scale processes or even exclude these processes (particularly shelf-seas related processes); and (ii) their initialisation from a steady state leads to their divergence from reality and present-day conditions. To address these shortcomings and project the impacts of climate change in the Atlantic Ocean with focus on regional scales, we downscale globally an ensemble of future ocean projections with a NEMO-ERSEM coupled hydrodynamic-ecosystem model. Here, we discuss the design-methodology for our global ocean downscaling experiment: (i) selection of future scenarios, (ii) initialisation from real ocean conditions, (iii) selection of the CMIP6-ESMs atmospheric conditions to force our model based on their realism and uncertainty span, and (iv) treatment of the river runoffs as to impose both a realistic rivers state and a future trend consistent with CMIP6-ESMs. Comparisons of our global ocean downscaling simulations to CMIP6-ESMs during the historical period demonstrate their added value in terms of representation of physical ocean conditions and circulation in the Atlantic Ocean. We also present preliminary analysis in terms of future trends in temperature, salinity and circulation patterns in the Atlantic Ocean, with focus on regional features like changes in the Gulf Stream and trends in coastal regions.