An assessment of equatorial Atlantic interannual variability in OMIP simulations

The eastern equatorial Atlantic (EEA) seasonal cycle and interannual variability of the sea surface temperature strongly influence the climate of the surrounding continents. It is thus crucial that models used in both climate predictions and future climate projections are able to simulate them accurately. In that context, the EEA seasonal cycle and interannual variability are evaluated over the period 1985-2004 in models participating to the Ocean Model Intercomparison Project Phases 1 and 2 (OMIP1 and OMIP2). The main difference between OMIP1 and OMIP2 simulations is their atmospheric forcing: CORE-II and JRA55-do, respectively. Seasonal cycles of the equatorial Atlantic zonal winds, sea level anomaly and sea surface temper- ature in OMIP1 and OMIP2 are comparable to reanalysis datasets. Yet, some discrepancies exist in both OMIP ensembles: the thermocline is too diffusive and there is a lack of cooling during the development of the Atlantic cold tongue. In addition, the vertical ocean velocity in the eastern equatorial Atlantic in boreal summer is larger in OMIP1 than in OMIP2 simulations. The EEA interannual sea surface temperature variability in the OMIP1 ensemble mean is found to be 51% larger (0.62 ± 0.04 ˚C) than the OMIP2 ensemble mean (0.41 ± 0.03 ˚C). Sensitivity experiments demonstrate that the discrepancy in interannual sea surface temperature variability between OMIP1 and OMIP2 is mainly attributed to their wind forcing. While the April-May- June zonal wind variability in the western equatorial Atlantic is similar in both forcing, the zonal wind variability peaks in April for JRA55-do and in May for CORE-II. Differences in surface heat fluxes between the two atmospheric forcing datasets have no significant impacts on the simulated interannual SST variability.