Received: 21 May 2018 – Discussion started: 09 Jul 2018
Abstract. To improve the model simulation through decisive coupling mechanisms, rather than blindly updating the parameterization schemes, it is necessary to compare model performances between the CGCMs with the same atmospheric or oceanic component model. Therefore, two new CGCMs have been developed with the same oceanic component model, namely ECHAM5-NEMO3.6 and ECHAM6-NEMO3.6. The MPI-ESM that consists of ECHAM6 and MPIOM has also been employed. Experiments are carried out with the same settings in coupler and individual component model if applicable, and the new models show substantial improvements in the simulation of SST, precipitation and ocean currents. Further analysis has made it clear that the primary cause of SST biases in ECHAM5-NEMO3.6 and ECHAM6-NEMO3.6 can be attributed to the momentum field, while oceanic dynamics and surface radiation budget are accountable for more SST deviations in the MPI-ESM. Inter-model comparison between the coupled models with the same oceanic model suggests that cumulus convection is in the central part of simulation differences, which finally influence the SST through WES feedback mechanism. Whereas the OGCM replacement shows that latent heat of evaporation plays a predominant role in changing SST and surface radiation budget, and eventually bringing about variations in air temperature and atmospheric circulation. The mechanisms revealed in this study provide a new perspective of bias genesis during model coupling, which can be helpful for tuning other climate models towards a more realistic simulation.
How to cite. Gui, S., Yang, R., and Cao, J.: Simulation Improvements of ECHAM5-NEMO3.6 and ECHAM6-NEMO3.6 Coupled Models Compared to MPI-ESM and the Corresponding Physical Mechanisms, Geosci. Model Dev. Discuss. [preprint], https://doi.org/10.5194/gmd-2018-130, 2018.
In this paper, two new coupled models have been developed, both of which show substantial improvements in the model simulation compared with the MPI-ESM model that is widely used in weather forecast and atmospheric research. Inter-model comparison suggests that cumulus convection and latent heat of evaporation over the sea surface are the two major factors that shape the model error of sea surface temperature. It implies a new vision of bias origin for coupled model development practices.
In this paper, two new coupled models have been developed, both of which show substantial...