Mesoscale eddies are ubiquitous turbulent structures in the ocean, influencing the upper mixed layer. The mixed layer is the ocean surface layer mixed through air–sea exchanges. Using Argo profiling floats inside large Mediterranean anticyclones, we investigate the induced winter mixed-layer depth anomalies. Mixed-layer depth was observed to be greatly influenced by the eddy preexisting subsurface structure to which it possibly connects and can also create double-core anticyclones.

Altimeter-derived currents are used here to revisit the seasonal and interannual variability of all surface currents involved in the western tropical Atlantic circulation. A new approach based on the calculation of the current strengths and core positions is used to investigate the relationship between the currents, the remote wind variability, and the tropical Atlantic modes. The results show relationships at the seasonal and interannual timescale depending on the location of the currents.

Sea ice melt, together with other freshwater sources, has effects on the Arctic environment. Sea surface salinity (SSS) plays a key role in representing water mixing. Recently the satellite SSS from SMOS was developed in the Arctic region. In this study, we first evaluate the impact of assimilating these satellite data in an Arctic reanalysis system. It shows that SSS errors are reduced by 10–50 % depending on areas, encouraging its use in a long-time reanalysis to monitor the Arctic water cycle.

The Chukchi continental slope is one of the most energetic regions in the western Arctic Ocean as it is populated with strong boundary currents and mesoscale eddies. Using a set of experiments with an idealized primitive equation numerical model, we find that the ice friction can cause the loss of energy of both the Chukchi Slope Current and mesoscale eddies over a vertical scale of 100 m through Ekman pumping. Some scales for measuring the effects of ice friction are also provided.

Data assimilation techniques are important for combining observations with numerical models. Here, we approximate the adjoint of viscous-plastic dynamics (adjoint-VP) to replace the adjoint of free-drift dynamics (adjoint-FD) for developing an advanced Arctic Ocean and sea ice modeling and adjoint-based assimilation system. We find that adjoint-VP provides a better ocean and sea ice estimation than adjoint-FD, considering the residual errors and adjustments of the atmospheric states.