Abstract/Summary

Ocean reanalyses are becoming increasingly available and useful, and may eventually attract a similar applications base as atmospheric reanalyses. Here we look at how they are being evaluated against both assimilated and independent data, and emphasise that circulation and transport estimates are critical. The Ocean Reanalysis Intercomparison project, ORA-IP, has been comparing many products for consistency on a regional and global basis, including ocean heat content, air-sea fluxes, and recently polar properties including sea ice. The Atlantic meridional overturning circulation as measured by the RAPID array at 26N, is now a challenging new target for simulation. This chapter shows that reanalyses may represent interior ocean basin circulations well (better than free-running models) but they still fail to consistently constrain boundary currents, where most meridional heat transport takes place. There is new work ongoing to try to physically interpret observation increments in reanalysis products, and to look at how to best develop long period reanalysis in earlier years when ocean observations were scarce. Finally, we look at new coupled ocean-atmosphere reanalysis that, by always maintaining a coupled ocean-atmospheric boundary layer, may lead to reduced assimilation increments and air-sea fluxes across domains.