Oral

IAPSO

JCMP10 - The atmosphere, cryosphere and oceans in Earth System Models

Characterization and causes of the Central North Atlantic cold bias in CMIP6 and HighResMIP simulations

The Central North Atlantic (CNA) sea surface temperature (SST) cold bias is a well-known issue in contemporary climate models, but the underlying causes are still not fully understood. In this study, we aim to characterize this CNA cold bias in the climate models that participated in the Coupled Model Intercomparison Project, phase 6 (CMIP6), and investigate the possible causes of the CNA cold bias through an energy budget analysis. Our findings reveal that the CNA cold bias in CMIP6 models is substantially larger than the range of observational uncertainty, ruling out a possible issue with the verification products themselves. The primary cause of the cold bias lies in an underestimation of the North Atlantic Current in the models, which results in an underestimation of the horizontal heat transport into the CNA region. Furthermore, it is shown that the CMIP6 models tend to overestimate CNA air-sea turbulent heat fluxes, but we suggest that this overestimation is a consequence rather than a direct cause of the CNA cold bias. We further investigate the impact of spatial resolution on the accuracy of modeling surface currents and CNA SST. Three climate models show improvement in the strength of the North Atlantic Current and CNA SST when transitioning from coarse (~1o) to eddy-permitting (~0.25o) resolutions. We note through a multi-model analysis that the spatial structure of CNA SST biases is very similar across models that share the POP or NEMO ocean model.