5 Conclusions

The main improvement of this new CNES-CLS22 MDT over the previous CNES-CLS18 MDT is in the Arctic, with better coverage, no artifacts and a more realistic solution. Globally, the new CNES-CLS22 solution is close to the CNES-CLS18 solution, both have better resolution of small scales than previous CNES-CLS MDTs but are potentially polluted by noise at very short scales. The CNES-CLS22 MDT has been evaluated against independent height and velocity data in comparison with the previous version, the CNES-CLS18. The new solution presents slightly better results, although not identical in all regions of the globe. In particular, the results are better in the Antarctic Circumpolar Current in terms of height, better off Japan and particularly in the Arctic in terms of geostrophic velocities. For geostrophic currents, those of the new CNES-CLS22 MDT in the Antarctic Circumpolar Current are slightly worse than those of the CNES-CLS18 in comparison with drifters.
Improvements to this new MDT include a new first-guess with the CNES-CLS22 MSS and the GOCO06s geoid to which optimal filtering has been applied, as well as Lagrangian filtering at the coast to reduce the intensity of normal currents at the coast, drifting buoy and T/S profile databases have been updated, as have updated processing to obtain synthetic mean geostrophic velocities and synthetic mean heights. In addition, a new type of data, HF radar data, was processed to extract physical content consistent with MDT in the Mid Atlantic Bight region. The study of this region in particular, showed the improvements of CNES-CLS22 MDT, but that there is still work to be done to obtain a more physical solution on the continental shelf. Indeed, on these continental shelves in general, the first guess is not always good because it’s close to the coast and today we only process T/S profiles with a depth greater than 200m; we therefore lack data on these shelves. HF radar data can provide some velocities information, as there are few drifting buoy data in these coastal regions, but processing is difficult, as complex and non-negligible ageostrophic currents have to be removed. In the case of the Mid Atlantic Bight, these HF radar data do not allow us to obtain a MDT with a geostrophic flow not crossing the coastline.
Thanks to the new first-guess and in particular the new CNES-CLS22 MSS, the Arctic area is covered in this MDT and the artifacts of CNES-CLS18 have disappeared. Looking more specifically at the European Arctic region, the new CNES-CLS22 solution presents structures in agreement with the literature in the Yermak Plateau area and in the St Anna Through region.
With the arrival of new swath observations from the SWOT (Surface Water and Ocean Topography) satellite launched in December 2022 (Fu et al. 2012), the continuous improvement of MDT accuracy and resolution is necessary. The inclusion of new coastal data, such as HF radar and SAR data, and shallower T/S profiles not currently taken into account, requires a clear separation of geostrophic and ageostrophic processes, and access to physical content consistent with MDT. These treatments call for new methods in order to best process continental shelf areas with varied oceanographic phenomena.