ALBERT

Description: The North Atlantic cold bias, associated with the misplacement of the North Atlantic Current (NAC) and typically extending from the surface to 1000 m depth, is a common problem in coupled models that compromises model fidelity. We investigate the use of a flow field correction (FFC) to adjust the path of the NAC and alleviate the cold bias. The FFC consists of three steps. First, climatological potential temperature (T) and salinity (S) fields for use with the model are produced using a three-dimensional restoring technique. Second, these T, S fields are used to modify the momentum equations of the ocean model. In the third stage, the correction term is diagnosed to construct a flow-independent correction. Results using the Kiel Climate Model show that the FFC allows the establishment of a northwest corner, substantially alleviating the subsurface cold bias. A cold bias remains at the surface but can be eliminated by additionally correcting the surface freshwater flux, without adjusting the surface heat flux seen by the ocean model. A model version in which only the surface fluxes of heat and freshwater are corrected continues to exhibit the incorrect path of the NAC and a strong subsurface bias. We also show that the bias in the atmospheric circulation is reduced in some corrected model versions. The FFC can be regarded as a way to correct for model error, e.g. associated with the deep water mass pathways and their impact on the large-scale ocean circulation, and unresolved processes such as eddy momentum flux convergence.

Description: The growth of El Niño/Southern Oscillation (ENSO) events is determined by the balance between ocean dynamics and thermodynamics. Here we quantify the contribution of the thermodynamic feedbacks to the sea surface temperature (SST) change during ENSO growth phase by integrating the atmospheric heat fluxes over the temporarily and spatially varying mixed layer to derive an offline “slab ocean” SST. The SST change due to ocean dynamics is estimated as the residual with respect to the total SST change. In observations, 1 K SST change in the Niño3.4 region is composed of an ocean dynamical SST forcing of + 2.6 K and a thermodynamic damping of − 1.6 K, the latter mainly by the shortwave-SST (− 0.9 K) and latent heat flux-SST feedback (− 0.7 K). Most climate models from the Coupled Model Intercomparison Project phase 5 (CMIP5) underestimate the SST change due to both ocean dynamics and net surface heat fluxes, revealing an error compensation between a too weak forcing by ocean dynamics and a too weak damping by atmospheric heat fluxes. In half of the CMIP5 models investigated in this study, the shortwave-SST feedback erroneously acts as an amplifying feedback over the eastern equatorial Pacific, resulting in a hybrid of ocean-driven and shortwave-driven ENSO dynamics. Further, the phase locking and asymmetry of ENSO is investigated in the CMIP5 model ensemble. The climate models with stronger atmospheric feedbacks tend to simulate a more realistic seasonality and asymmetry of the heat flux feedbacks, and they exhibit more realistic phase locking and asymmetry of ENSO. Moreover, the almost linear latent heat flux feedback contributes to ENSO asymmetry in the far eastern equatorial Pacific through an asymmetry in the mixed layer depth. This study suggests that the dynamic and thermodynamic ENSO feedbacks and their seasonality and asymmetries are important metrics to consider for improving ENSO representation in climate models.