D7.4: Report summarising the impact of the oceanic and atmospheric boundary layer schemes in the coupled system

About this document This report presents the evaluation of key performance metrics in SR-ESMs with improved representation of the oceanic and atmospheric boundary layers with a focus on the tropics and on the Southern Ocean. Work package in charge WP7 Executive summary The main objective of WP6 has been to tune the existing ocean and atmospheric boundary layer schemes for the forced ocean and the atmosphere components of the ICON and IFSFESOM Storm-resolving ESMs (SREMS). This allowed for detailed investigations of mixing processes, because turbulence observations are available for a particular storm. The objective of the present D7.4 is now to evaluate how these boundary layer schemes affect the representation of climate in the nextGEMS coupled systems. As in WP6 we find that the details of the boundary layer schemes do not much affect the existing biases so that the ultimate choice of the boundary layer schemes in both, atmosphere and ocean, are determined by practical considerations: future IFS simulations will continue to use K-profile theory to parameterize boundary layer turbulence and eddy-diffusive mass-flux to parameterize turbulence and cloud formation in the boundary layer. This is motivated by the high accuracy, a key objective for operational agencies like ECMWF. future ICON-A simulations will avoid physics parameterizations beyond the Smagorinsky closureto dissipate energy at the grid scale. This makes the model physics and dynamics transparentand turns ICON-A into an ideal research tool. future IFS-FESOM simulations will be done with the K-profile paramterization (KPP), becauseit provides a realistic 20th century increase in global temperatures. The alternative scheme, theTurbulent Kinetic Energy (TKE) parameterization mixes too much, which results in unrealistic20th century warming. future ICON-O simulation will be done with the TKE scheme for two reasons: it is energy based(rather than the empirical KPP), which allows easy combination with the abyssal mixing schemeIDEMIX, and KPP is licensed under GNU, which would limit further applications.